Occupational Exposure to Beryllium and Beryllium Compounds

AGENCY:

Occupational Safety and Health Administration (OSHA), Department of Labor.

ACTION:

Proposed rule; request for comments.

SUMMARY:

The Occupational Safety and Health Administration (OSHA) proposes to amend its existing exposure limits for occupational exposure in general industry to beryllium and beryllium compounds and promulgate a substance-specific standard for general industry regulating occupational exposure to beryllium and beryllium compounds. This document proposes a new permissible exposure limit (PEL), as well as ancillary provisions for employee protection such as methods for controlling exposure, respiratory protection, medical surveillance, hazard communication, and recordkeeping. In addition, OSHA seeks comment on a number of alternatives, including a lower PEL, that could affect construction and maritime, as well as general industry.

DATES:

Written comments. Written comments, including comments on the information collection determination described in Section IX of the preamble (OMB Review under the Paperwork Reduction Act of 1995), must be submitted (postmarked, sent, or received) by November 5, 2015.

Informal public hearings. The Agency will schedule an informal public hearing on the proposed rule if requested during the comment period. The location and date of the hearing, procedures for interested parties to notify the Agency of their intention to participate, and procedures for participants to submit their testimony and documentary evidence will be announced in the Federal Register if a hearing is requested.

ADDRESSES:

Written comments. You may submit comments, identified by Docket No. OSHA-H005C-2006-0870, by any of the following methods:

Electronically: You may submit comments and attachments electronically at http://www.regulations.gov, which is the Federal e-Rulemaking Portal. Follow the instructions on-line for making electronic submissions. When uploading multiple attachments into Regulations.gov, please number all of your attachments because www.Regulations.gov will not automatically number the attachments. This will be very useful in identifying all attachments in the beryllium rule. For example, Attachment 1—title of your document, Attachment 2—title of your document, Attachment 3—title of your document, etc. Specific instructions on uploading all documents are found in the Facts, Answer, Questions portion and the commenter check list on Regulations.gov Web page.

Fax: If your submissions, including attachments, are not longer than 10 pages, you may fax them to the OSHA Docket Office at (202) 693-1648.

Mail, hand delivery, express mail, messenger, or courier service: You may submit your comments to the OSHA Docket Office, Docket No. OSHA-H005C-2006-0870, U.S. Department of Labor, Room N-2625, 200 Constitution Avenue NW., Washington, DC 20210, telephone (202) 693-2350 (OSHA's TTY number is (877) 889-5627). Deliveries (hand, express mail, messenger, or courier service) are accepted during the Docket Office's normal business hours, 8:15 a.m.-4:45 p.m., E.S.T.

Instructions: All submissions must include the Agency name and the docket number for this rulemaking (Docket No. OSHA-H005C-2006-0870). All comments, including any personal information you provide, are placed in the public docket without change and may be made available online at http://www.regulations.gov. Therefore, OSHA cautions you about submitting personal information such as Social Security numbers and birthdates.

If you submit scientific or technical studies or other results of scientific research, OSHA requests (but is not requiring) that you also provide the following information where it is available: (1) Identification of the funding source(s) and sponsoring organization(s) of the research; (2) the extent to which the research findings were reviewed by a potentially affected party prior to publication or submission to the docket, and identification of any such parties; and (3) the nature of any financial relationships (e.g., consulting agreements, expert witness support, or research funding) between investigators who conducted the research and any organization(s) or entities having an interest in the rulemaking. If you are submitting comments or testimony on the Agency's scientific or technical analyses, OSHA requests that you disclose: (1) The nature of any financial relationships you may have with any organization(s) or entities having an interest in the rulemaking; and (2) the extent to which your comments or testimony were reviewed by an interested party before you submitted them. Disclosure of such information is intended to promote transparency and scientific integrity of data and technical information submitted to the record. This request is consistent with Executive Order 13563, issued on January 18, 2011, which instructs agencies to ensure the objectivity of any scientific and technological information used to support their regulatory actions. OSHA emphasizes that all material submitted to the rulemaking record will be considered by the Agency to develop the final rule and supporting analyses.

Docket: To read or download comments and materials submitted in response to this Federal Register notice, go to Docket No. OSHA-H005C-2006-0870 at http://www.regulations.gov,, or to the OSHA Docket Office at the address above. All comments and submissions are listed in the http://www.regulations.gov index; however, some information (e.g., copyrighted material) is not publicly available to read or download through that Web site. All comments and submissions are available for inspection at the OSHA Docket Office.

Electronic copies of this Federal Register document are available at http://www.regulations.gov. Copies also are available from the OSHA Office of Publications, Room N-3101, U.S. Department of Labor, 200 Constitution Avenue NW., Washington, DC 20210; telephone (202) 693-1888. This document, as well as news releases and other relevant information, is also available at OSHA's Web site at http://www.osha.gov.

OSHA has not provided the document ID numbers for all submissions in the record for this beryllium proposal. The proposal only contains a reference list for all submissions relied upon. The public can find all document ID numbers in an Excel spreadsheet that is posted on OSHA's rulemaking Web page (see www.osha.gov/berylliumrulemaking). The public will be able to locate submissions in the record in the public docked Web page: http://www.regulations.gov. To locate a particular submission contained in http://www.regulations.gov, the public should enter the full document ID number in the search bar.

FOR FURTHER INFORMATION CONTACT:

For general information and press inquiries, contact Frank Meilinger, Director, Office of Communications, Room N-3647, OSHA, U.S. Department of Labor, 200 Constitution Avenue NW., Washington, DC 20210; telephone: (202) 693-1999; email: meilinger.francis2@dol.gov . For technical inquiries, contact: William Perry or Maureen Ruskin, Directorate of Standards and Guidance, Room N-3718, OSHA, U.S. Department of Labor, 200 Constitution Avenue NW., Washington, DC 20210; telephone (202) 693-1955 or fax (202) 693-1678; email: perry.bill@dol.gov.

SUPPLEMENTARY INFORMATION:

The preamble to the proposed standard on occupational exposure to beryllium and beryllium compounds follows this outline:

Executive Summary

I. Issues and Alternatives

II. Pertinent Legal Authority

III. Events Leading to the Proposed Standards

IV. Chemical Properties and Industrial Uses

V. Health Effects

VI. Preliminary Risk Assessment

VII. Response to Peer Review

VIII. Significance of Risk

IX. Summary of the Preliminary Economic Analysis and Initial Regulatory Flexibility Analysis

X. OMB Review under the Paperwork Reduction Act of 1995

XI. Federalism

XII. State-Plan States

XIII. Unfunded Mandates Reform Act

XIV. Protecting Children from Environmental Health and Safety Risks

XV. Environmental Impacts

XVI. Consultation and Coordination with Indian Tribal Governments

XVII. Public Participation

XVIII. Summary and Explanation of the Proposed Standard

(a) Scope and Application

(b) Definitions

(c) Permissible Exposure Limits (PELs)

(d) Exposure Assessment

(e) Beryllium Work Areas and Regulated Areas

(f) Methods of Compliance

(g) Respiratory Protection

(h) Personal Protective Clothing and Equipment

(i) Hygiene Areas and Practices

(j) Housekeeping

(k) Medical Surveillance

(l) Medical Removal

(m) Communication of Hazards to Employees

(n) Recordkeeping

(o) Dates

XIX. References

Executive Summary

OSHA currently enforces permissible exposure limits (PELs) for beryllium in general industry, construction, and shipyards. These PELs were adopted in 1971, shortly after the Agency was created, and have not been updated since then. The time-weighted average (TWA) PEL for beryllium is 2 micrograms per cubic meter of air (μg/m³) as an 8-hour time-weighted average. OSHA is proposing a new TWA PEL of 0.2 μg/m³ in general industry. OSHA is also proposing other elements of a comprehensive health standard, including requirements for exposure assessment, preferred methods for controlling exposure, respiratory protection, personal protective clothing and equipment (PPE), medical surveillance, medical removal, hazard communication, and recordkeeping.

OSHA's proposal is based on the requirements of the Occupational Safety and Health Act (OSH Act) and court interpretations of the Act. For health standards issued under section 6(b)(5) of the OSH Act, OSHA is required to promulgate a standard that reduces significant risk to the extent that it is technologically and economically feasible to do so. See Section II of this preamble, Pertinent Legal Authority, for a full discussion of OSHA legal requirements.

OSHA has conducted an extensive review of the literature on adverse health effects associated with exposure to beryllium. The Agency has also assessed the risk of beryllium-related diseases at the current TWA PEL, the proposed TWA PEL and the alternative TWA PELs. These analyses are presented in this preamble at Section V, Health Effects, Section VI, Preliminary Risk Assessment, and Section VIII, Significance of Risk. As discussed in Section VIII of this preamble, Significance of Risk, the available evidence indicates that worker exposure to beryllium at the current PEL poses a significant risk of chronic beryllium disease (CBD) and lung cancer, and that the proposed standard will substantially reduce this risk.

Section 6(b) of the OSH Act requires OSHA to determine that its standards are technologically and economically feasible. OSHA's examination of the technological and economic feasibility of the proposed rule is presented in the Preliminary Economic Analysis and Initial Regulatory Flexibility Analysis (PEA) (OSHA, 2014), and is summarized in Section IX of this preamble, Summary of the Preliminary Economic Analysis and Initial Regulatory Flexibility Analysis. OSHA has preliminarily concluded that the proposed PEL of 0.2 μg/m³ is technologically feasible for all affected industries and application groups. Thus, OSHA preliminarily concludes that engineering and work practices will be sufficient to reduce and maintain beryllium exposures to the proposed PEL of 0.2 μg/m³ or below in most operations most of the time in the affected industries. For those few operations within an industry or application group where compliance with the proposed PEL cannot be achieved even when employers implement all feasible engineering and work practice controls, the proposed standard would require employers to supplement controls with respirators.

OSHA developed quantitative estimates of the compliance costs of the proposed rule for each of the affected industry sectors. The estimated compliance costs were compared with industry revenues and profits to provide a screening analysis of the economic feasibility of complying with the revised standard and an evaluation of the potential economic impacts. Industries with unusually high costs as a percentage of revenues or profits were further analyzed for possible economic feasibility issues. After performing these analyses, OSHA has preliminarily concluded that compliance with the requirements of the proposed rule would be economically feasible in every affected industry sector.

The Regulatory Flexibility Act, as amended by the Small Business Regulatory Enforcement Fairness Act (SBREFA), requires that OSHA either certify that a rule would not have a significant economic impact on a substantial number of small entities or prepare a regulatory flexibility analysis and hold a Small Business Advocacy Review (SBAR) Panel prior to proposing the rule. OSHA has determined that a regulatory flexibility analysis is needed and has provided this analysis in Chapter IX of the PEA (OSHA, 2014). A summary is provided in Section IX of this preamble, Summary of the Preliminary Economic Analysis and Initial Regulatory Flexibility Analysis. OSHA also previously held a SBAR Panel for this rule. The recommendations of the Panel and OSHA's response to them are summarized in Section IX of this preamble.

Executive Orders 13563 and 12866 direct agencies to assess all costs and benefits of available regulatory alternatives. Executive Order 13563 emphasizes the importance of quantifying both costs and benefits, of reducing costs, of harmonizing rules, and of promoting flexibility. This rule has been designated an economically significant regulatory action under section 3(f)(1) of Executive Order 12866. Accordingly, this proposed rule has been reviewed by the Office of Management and Budget. The remainder of this section summarizes the key findings of the analysis with respect to costs and benefits of the proposed standard, presents alternatives to the proposed standard, and requests comments on a number of issues.

Table I-1, which is derived from material presented in the PEA, provides a summary of OSHA's best estimate of the costs and benefits of this proposed rule. As shown, this proposed rule is estimated to prevent 96 fatalities and 50 non-fatal beryllium-related illnesses annually once it is fully effective, and the monetized annualized benefits of the proposed rule are estimated to be $576 million using a 3-percent discount rate and $255 million using a 7-percent discount rate. Also as shown in Table I-1, the estimated annualized cost of the rule is $37.6 million using a 3-percent discount rate and $39.1 million using a 7-percent discount rate. This proposed rule is estimated to generate net benefits of $538 million annually using a 3-percent discount rate and $216 million annually using a 7-percent discount rate. These estimates are for informational purposes only and have not been used by OSHA as the basis for its decision concerning the choice of a PEL or of other ancillary requirements for this proposed beryllium rule. The courts have ruled that OSHA may not use benefit-cost analysis or a criterion of maximizing net benefits as a basis for setting OSHA health standards.

Both the costs and benefits of Table I-1 reflect the incremental costs and benefits associated with achieving full compliance with the proposed standard. They do not include costs and benefits associated with employers' current exposure control measures or other aspects of the proposed standard they have already implemented. For example, for employers whose exposures are already below the proposed PEL, OSHA's estimated costs and benefits for the proposed standard do not include the costs of their exposure control measures or the benefits of these employers' compliance with the proposed PEL. The costs and benefits of Table I-1 also do not include costs and benefits associated with achieving compliance with existing requirements, to the extent that some employers may currently not be fully complying with applicable regulatory requirements.

I. Issues and Alternatives

In addition to the proposed standard itself, this preamble discusses more than two dozen regulatory alternatives, including various sub-alternatives, to the proposed standard and requests comments and information on a variety of topics pertinent to the proposed standard. The regulatory alternatives OSHA is considering include alternatives to the proposed scope of the standard, regulatory alternatives to the proposed TWA PEL of 0.2 μg/m³ and proposed STEL of 2 μg/m³, a regulatory alternative that would modify the proposed methods of compliance, and regulatory alternatives that affect proposed ancillary provisions. The Agency solicits comment on the proposed phase-in schedule for the various provisions of the standard. Additional requests for comments and information follow the summaries of regulatory alternatives, under the “Issues” heading.

Regulatory Alternatives

OSHA believes that inclusion of regulatory alternatives serves two important functions. The first is to explore the possibility of less costly ways (than the proposed standard) to provide an adequate level of worker protection from exposure to beryllium. The second is tied to the Agency's statutory requirement, which underlies the proposed standard, to reduce significant risk to the extent feasible. Each regulatory alternative presented here is described and analyzed more fully elsewhere in this preamble or in the PEA. Where appropriate, the alternative is included in this preamble at the end of the relevant section of Section XVIII, Summary and Explanation of the Proposed Standard, to facilitate comparison of the alternative to the proposed standard. For example, alternative PELs under consideration by the Agency are presented in the discussion of paragraph (c) in Section XVIII. In addition, all alternatives are discussed in the PEA, Chapter VIII: Regulatory Alternatives (OSHA, 2014). The costs and benefits of each regulatory alternative are presented both in Section IX of this preamble and in Chapter VIII of the PEA.

The more than two dozen regulatory alternatives, including various sub-alternatives regulatory alternatives under consideration are summarized below, and are organized into the following categories: alternatives to the proposed scope of the standard; alternatives to the proposed PELs; alternatives to the proposed methods of compliance; alternatives to the proposed ancillary provisions; and the timing of the standard.

Scope

OSHA has examined three alternatives that would alter the groups of employers and employees covered by this rulemaking. Regulatory Alternative #1a would expand the scope of the proposed standard to include all operations in general industry where beryllium exists only as a trace contaminant; that is, where the materials used contain no more than 0.1% beryllium by weight. Regulatory Alternative #1b is similar to Regulatory Alternative #1a, but exempts operations where the employer can show that employees' exposures will not meet or exceed the action level or exceed the STEL. Where the employer has objective data demonstrating that a material containing beryllium or a specific process, operation, or activity involving beryllium cannot release beryllium in concentrations at or above the proposed action level or above the proposed STEL under any expected conditions of use, that employer would be exempt from the proposed standard except for recordkeeping requirements pertaining to the objective data. Alternative #1a and Alternative #1b, like the proposed rule, would not cover employers or employees in construction or shipyards.

Regulatory Alternative #2a would expand the scope of the proposed standard to also include employers in construction and maritime. For example, this alternative would cover abrasive blasters, pot tenders, and cleanup staff working in construction and shipyards who have the potential for airborne beryllium exposure during blasting operations and during cleanup of spent media. Regulatory Alternative #2b would update §§ 1910.1000 Tables Z-1 and Z-2, 1915.1000 Table Z, and 1926.55 Appendix A so that the proposed TWA PEL and STEL would apply to all employers and employees in general industry, shipyards, and construction, including occupations where beryllium exists only as a trace contaminant. However, all other provisions of the standard would be in effect only for employers and employees that fall within the scope of the proposed rule. More detailed discussion of Regulatory Alternatives #1a, #1b, #2a, and #2b appears in Section IX of this preamble and in Chapter VIII of the PEA (OSHA, 2014). In addition, Section XVIII of this preamble, Summary and Explanation, includes a discussion of paragraph (a) that describes the scope of the proposed rule, issues with the proposed scope, and Regulatory Alternatives #1a, #1b, #2a, and #2b.

Another regulatory alternative that would impact the scope of affected industries, extending eligibility for medical surveillance to employees in shipyards, construction, and parts of general industry excluded from the scope of the proposed standard, is discussed along with other medical surveillance alternatives later in this section (Regulatory Alternative #21) and in the discussion of paragraph (k) in this preamble at Section XVIII, Summary and Explanation of the Proposed Standard.

Permissible Exposure Limits

OSHA has examined several regulatory alternatives that would modify the TWA PEL or STEL for the proposed rule. Under Regulatory Alternative #3, OSHA would adopt a STEL of 5 times the proposed PEL. Thus, this alternative STEL would be 1.0 μg/m³ if OSHA adopts a PEL of 0.2 μg/m³; it would be 0.5 μg/m³ if OSHA adopts a PEL of 0.1 μg/m³; and it would be 2.5 µg/m³ if OSHA adopts a PEL of 0.5 µg/m³ (see Regulatory Alternatives #4 and #5). Under Regulatory Alternative #4, the proposed PEL would be lowered from 0.2 μg/m³ to 0.1 μg/m³. Under Regulatory Alternative #5, the proposed PEL would be raised from 0.2 μg/m³ to 0.5 μg/m³. In addition, for informational purposes, OSHA examined a regulatory alternative that would maintain the TWA PEL at 2.0 μg/m³, but all of the other proposed provisions would be required with their triggers remaining the same as in the proposed rule. This alternative is not one OSHA could legally adopt because the absence of a more protective requirement for engineering controls would not be consistent with section 6(b)(5) of the OSH Act. More detailed discussion of these alternatives to the proposed PEL appears in Section IX of this preamble and in Chapter VIII of the PEA (OSHA, 2014). In addition, in Section XVIII of this preamble, Summary and Explanation of the Proposed Standard, the discussion of proposed paragraph (c) describes the proposed TWA PEL and STEL, issues with the proposed exposure limits, and Regulatory Alternatives #3, #4, and #5.

Methods of Compliance

The proposed standard would require employers to implement engineering and work practice controls to reduce employees' exposures to or below the TWA PEL and STEL. Where engineering and work practice controls are insufficient to reduce exposures to or below the TWA PEL and STEL, employers would still be required to implement them to reduce exposure as much as possible, and to supplement them with a respiratory protection program. In addition, for each operation where there is airborne beryllium exposure, the employer must ensure that one or more of the engineering and work practice controls listed in paragraph (f)(2) are in place, unless all of the listed controls are infeasible, or the employer can demonstrate that exposures are below the action level based on two samples taken seven days apart. Regulatory Alternative #6 would eliminate the engineering and work practice controls provision currently specified in paragraph (f)(2). This regulatory alternative does not eliminate the need for engineering controls to lower exposure levels to or below the TWA PEL and STEL; rather, it dispenses with the mandatory use of certain engineering controls that must be installed above the action level but at or below the TWA PEL.

More detailed discussion of Regulatory Alternative #6 appears in Section IX of this preamble and in Chapter VIII of the PEA (OSHA, 2014). In addition, the discussion of paragraph (f) in Section XVIII of this preamble, Summary and Explanation, provides a more detailed explanation of the proposed methods of compliance, issues with the proposed methods of com­pli­ance, and Regulatory Alternative #6.

Ancillary Provisions

The proposed rule contains several ancillary provisions, including requirements for exposure assessment, personal protective clothing and equipment (PPE), medical surveillance, medical removal, training, and regulated areas or access control. OSHA has examined a variety of regulatory alternatives involving changes to one or more of these ancillary provisions. OSHA has preliminarily determined that several of these ancillary provisions will increase the benefits of the proposed rule, for example, by helping to ensure the TWA PEL is not exceeded or by lowering the risks to workers given the significant risk remaining at the proposed TWA PEL. However, except for Regulatory Alternative #7 (involving the elimination of all ancillary provisions), OSHA did not estimate changes in monetized benefits for the regulatory alternatives that affect ancillary provisions. Two regulatory alternatives that involve all ancillary provisions are presented below (#7 and #8), followed by regulatory alternatives for exposure monitoring (#9, #10, and #11), for regulated areas (#12), for personal protective clothing and equipment (#13), for medical surveillance (#14 through #21), and for medical removal (#22).

All Ancillary Provisions

During the Small Business Regulatory Fairness Act (SBREFA) process conducted in 2007, the SBAR Panel recommended that OSHA analyze a PEL-only standard as a regulatory alternative. The Panel also recommended that OSHA consider applying ancillary provisions of the standard so as to minimize costs for small businesses where exposure levels are low (OSHA, 2008b). In response to these recommendations, OSHA analyzed Regulatory Alternative #7, a PEL-only standard, and Regulatory Alternative #8, which would only apply ancillary provisions of the beryllium standard at exposures above the proposed PEL of 0.2 µg/m³ or the proposed STEL of 2 µg/m³. Regulatory Alternative #7 would update the Z tables for § 1910.1000, so that the proposed TWA PEL and STEL would apply to all workers in general industry. All other provisions of the proposed standard would be dropped.

As indicated previously, OSHA has preliminarily determined that there is significant risk remaining at the proposed PEL of 0.2 μg/m³. However, the available evidence on feasibility suggests that 0.2 μg/m³ may be the lowest feasible PEL (see Chapter IV of the PEA, OSHA 2014). Therefore, the Agency believes that it is necessary to include ancillary provisions in the proposed rule to further reduce the remaining risk. In addition, the recommended standard provided to OSHA by representatives of the primary beryllium manufacturing industry and the Steelworkers Union further supports the importance of ancillary provisions in protecting workers from the harmful effects of beryllium exposure (Materion and USW, 2012).

Under Regulatory Alternative #8, several ancillary provisions that the current proposal would require under a variety of exposure conditions (e.g., dermal contact; any airborne exposure; exposure at or above the action level) would instead only apply where exposure levels exceed the TWA PEL or STEL. Regulatory Alternative #8 affects the following provisions of the proposed standard:

—Exposure monitoring. Whereas the proposed standard requires annual monitoring where exposure levels are at or above the action level and at or below the TWA PEL, Alternative #8 would require annual exposure monitoring only where exposure levels exceed the TWA PEL or STEL;

— Written exposure control plan. Whereas the proposed standard requires written exposure control plans to be maintained in any facility covered by the standard, Alternative #8 would require only facilities with exposures above the TWA PEL or STEL to maintain a plan;

—PPE. Whereas the proposed standard requires PPE for employees under a variety of conditions, such as exposure to soluble beryllium or visible contamination with beryllium, Alternative #8 would require PPE only for employees exposed above the TWA PEL or STEL;

—Housekeeping. Whereas the proposed standard's housekeeping requirements apply across a wide variety of beryllium exposure conditions, Alternative #8 would limit housekeeping requirements to areas with exposures above the TWA PEL or STEL.

—Medical Surveillance. Whereas the proposed standard's medical surveillance provisions require employers to offer medical surveillance to employees with signs or symptoms of beryllium-related health effects regardless of their exposure level, Alternative #8 would make surveillance available to such employees only if they were exposed above the TWA PEL or STEL.

More detailed discussions of Regulatory Alternatives #7 and #8, including a description of the considerations pertinent to these alternatives, appear in Section IX of this preamble and in Chapter VIII of the PEA (OSHA, 2014).

Exposure Monitoring

OSHA has examined three regulatory alternatives that would modify the proposed standard's provisions on exposure monitoring, which require periodic monitoring annually where exposures are at or above the action level and at or below the TWA PEL. Under Regulatory Alternative #9, employers would be required to perform periodic exposure monitoring every 180 days where exposures are at or above the action level or above the STEL, and at or below the TWA PEL. Under Regulatory Alternative #10, employers would be required to perform periodic exposure monitoring every 180 days where exposures are at or above the action level or above the STEL, including where exposures exceed the TWA PEL. Under Regulatory Alternative #11, employers would be required to perform periodic exposure monitoring every 180 days where exposures are at or above the action level or above the STEL, and every 90 days where exposures exceed the TWA PEL. More detailed discussions of Regulatory Alternatives #9, #10, and #11 appear in Section IX of this preamble and in Chapter VIII of the PEA (OSHA, 2014). In addition, the discussion of proposed paragraph (d) in Section XVIII of this preamble, Summary and Explanation of the Proposed Standard, provides a more detailed explanation of the proposed requirements for exposure monitoring, issues with exposure monitoring, and the considerations pertinent to Regulatory Alternatives #9, #10, and #11.

Regulated Areas

The proposed standard would require employers to establish and maintain two types of areas: beryllium work areas, wherever employees are, or can reasonably be expected to be, exposed to any level of airborne beryllium; and regulated areas, wherever employees are, or can reasonably be expected to be, exposed to airborne beryllium at levels above the TWA PEL or STEL. Employers are required to demarcate beryllium work areas, but are not required to restrict access to beryllium work areas or provide respiratory protection or other forms of PPE within work areas that are not also regulated areas. Employers must demarcate regulated areas, restrict access to them, post warning signs and provide respiratory protection and other PPE within regulated areas, as well as medical surveillance for employees who work in regulated areas for more than 30 days in a 12-month period. During the SBREFA process conducted in 2007, the SBAR Panel recommended that OSHA consider dropping or limiting the provision for regulated areas (OSHA, 2008b). In response to this recommendation, OSHA analyzed Regulatory Alternative #12, which would not require employers to establish regulated areas. More detailed discussion of Regulatory Alternative #12 appears in Section IX of this preamble and in Chapter VIII of the PEA (OSHA, 2014). In addition, the discussion of paragraph (e) in Section XVIII of this preamble, Summary and Explanation, provides a more detailed explanation of the proposed requirements for regulated areas, issues with regulated areas, and considerations pertinent to Regulatory Alternative #12.

Personal Protective Clothing and Equipment (PPE)

Regulatory Alternative #13 would modify the proposed requirements for PPE, which require PPE where exposure exceeds the TWA PEL or STEL; where employees' clothing or skin may become visibly contaminated with beryllium; and where employees may have skin contact with soluble beryllium compounds. The requirement to use PPE where work clothing or skin may become “visibly contaminated” with beryllium differs from prior standards that do not require contamination to be visible in order for PPE to be required. In the case of beryllium, which OSHA has preliminarily concluded can sensitize through dermal exposure, the exposure levels capable of causing adverse health effects and the PELs in effect are so low that beryllium surface contamination is unlikely to be visible (see this preamble at section V, Health Effects). OSHA is therefore considering Regulatory Alternative #13, which would require appropriate PPE wherever there is potential for skin contact with beryllium or beryllium-contaminated surfaces. More detailed discussion of Regulatory Alternative #13 is provided in Section IX of this preamble and in Chapter VIII of the PEA (OSHA, 2014). In addition, the discussion of paragraph (h) in Section XVIII of this preamble, Summary and Explanation, provides a more detailed explanation of the proposed requirements for PPE, issues with PPE, and the considerations pertinent to Regulatory Alternative #13.

Medical Surveillance

The proposed requirements for medical surveillance include: (1) Medical examinations, including a test for beryllium sensitization, for employees who are exposed to beryllium above the proposed PEL for 30 days or more per year, who are exposed to beryllium in an emergency, or who show signs or symptoms of CBD; and (2) low-dose helical tomography (low-dose computed tomography, hereafter referred to as “CT scans”), for employees who were exposed above the proposed PEL for more than 30 days in a 12-month period for 5 years or more. This type of CT scan is a method of detecting tumors, and is commonly used to diagnose lung cancer. The proposed standard would require periodic medical exams to be provided for employees in the medical surveillance program annually, while tests for beryllium sensitization and CT scans would be provided to eligible employees biennially.

OSHA has examined eight regulatory alternatives (#14 through #21) that would modify the proposed rule's requirements for employee eligibility, the types of exam that must be offered, and the frequency of periodic exams. Medical surveillance was a subject of special concern to SERs during the SBREFA process, and the SBREFA Panel offered many comments and recommendations related to medical surveillance for OSHA's consideration. Some of the Panel's concerns have been addressed in this proposal, which was modified since the SBREFA Panel was convened (see this preamble at Section XVIII, Summary and Explanation of the Proposed Standard, for more detailed discussion). Several of the alternatives presented here (#16, #18, and #20) also respond to recommendations by the SBREFA Panel to reduce burdens on small businesses by dropping or reducing the frequency of medical surveillance requirements. OSHA also seeks to ensure that the requirements of the final standard offer workers adequate medical surveillance while limiting the costs to employers. Thus, OSHA requests feedback on several additional alternatives and on a variety of issues raised later in this section of the preamble.

Regulatory Alternatives #14, #15, and #21 would expand eligibility for medical surveillance to a broader group of employees than would be eligible in the proposed standard. Under Regulatory Alternative #14, medical surveillance would be available to employees who are exposed to beryllium above the proposed PEL, including employees exposed for fewer than 30 days per year. Regulatory Alternative #15 would expand eligibility for medical surveillance to employees who are exposed to beryllium above the proposed action level, including employees exposed for fewer than 30 days per year. Regulatory Alternative #21 would extend eligibility for medical surveillance as set forth in proposed paragraph (k) to all employees in shipyards, construction, and general industry who meet the criteria of proposed paragraph (k)(1) (or any of the alternative criteria under consideration). However, all other provisions of the standard would be in effect only for employers and employees that fall within the scope of the proposed rule.

Regulatory Alternatives #16 and #17 would modify the proposed standard's requirements to offer beryllium sensitization testing to eligible employees. Under Regulatory Alternative #16, employers would not be required to offer employees testing for beryllium sensitization. Regulatory Alternative #17 would increase the frequency of periodic sensitization testing, from the proposed standard's biennial requirement to annual testing. Regulatory Alternatives #18 and #19 would similarly modify the proposed standard's requirements to offer CT scans to eligible employees. Regulatory Alternative #18 would drop the CT scan requirement from the proposed rule, whereas Regulatory Alternative #19 would increase the frequency of periodic CT scans from biennial to annual scans. Finally, under Regulatory Alternative #20, all periodic components of the medical surveillance exams would be available biennially to eligible employees. Instead of requiring employers to offer eligible employees a medical examination every year, employers would be required to offer eligible employees a medical examination every other year. The frequency of testing for beryllium sensitization and CT scans would also be biennial for eligible employees, as in the proposed standard.

More detailed discussions of Regulatory Alternatives #14, #15, #16, #17, #18, #19, #20, and #21 appear in Section IX of this preamble and in Chapter VIII of the PEA (OSHA, 2014). In addition, Section XVIII of this preamble, Summary and Explanation, paragraph (k) provides a more detailed explanation of the proposed requirements for medical surveillance, issues with medical surveillance, and the considerations pertinent to Regulatory Alternatives #14 through #21.

Medical Removal Protection (MRP)

The proposed requirements for medical removal protection provide an option for medical removal to an employee who is working in a job with exposure at or above the action level and is diagnosed with CBD or confirmed positive for beryllium sensitization. If the employee chooses removal, the employer must either remove the employee to comparable work in a work environment where exposure is below the action level, or if comparable work is not available, must place the employee on paid leave for 6 months or until such time as comparable work becomes available. In either case, the employer must maintain for 6 months the employee's base earnings, seniority, and other rights and benefits that existed at the time of removal. During the SBREFA process, the Panel recommended that OSHA give careful consideration to the impacts that an MRP requirement could have on small businesses (OSHA, 2008b). In response to this recommendation, OSHA analyzed Regulatory Alternative #22, which would not require employers to offer MRP. More detailed discussion of Regulatory Alternative #22 appears in Section IX of this preamble and in Chapter VIII of the PEA (OSHA, 2014). In addition, the discussion of paragraph (l) in section XVIII of this preamble, Summary and Explanation, provides a more detailed explanation of the proposed requirements for MRP, issues with MRP, and considerations pertinent to Regulatory Alternative #22.

Timing of the Standard

The proposed standard would become effective 60 days following publication of the final standard in the Federal Register. The effective date is the date on which the standard imposes compliance obligations on employers. However, the standard would not become enforceable by OSHA until 90 days following the effective date for exposure monitoring, work areas and regulated areas, written exposure control plan, respiratory protection, other personal protective clothing and equipment, hygiene areas and practices (except change rooms), housekeeping, medical surveillance, and medical removal. The proposed requirement for change rooms would not be enforceable until one year after the effective date, and the requirements for engineering controls would not be enforceable until two years after the effective date. In summary, employers will have some period of time after the standard becomes effective to come into compliance before OSHA will begin enforcing it: 90 days for most provisions, one year for change rooms, and two years for engineering controls. Beginning 90 days following the effective date, during periods necessary to install or implement feasible engineering controls where exposure exceed the TWA PEL or STEL, employers must provide employees with respiratory protection as described in the proposed standard under section (g), Respiratory Protection.

OSHA invites comment and suggestions for phasing in requirements for engineering controls, medical surveillance, and other provisions of the standard. A longer phase-in time would have several advantages, such as reducing initial costs of the standard or allowing employers to coordinate their environmental and occupational safety and health control strategies to minimize potential costs. However, a longer phase-in would also postpone and reduce the benefits of the standard. Suggestions for alternatives may apply to specific industries (e.g., industries where first-year or annualized cost impacts are highest), specific size-classes of employers (e.g., employers with fewer than 20 employees), combinations of these factors, or all firms covered by the rule.

OSHA requests comments on these regulatory alternatives, including the Agency's choice of regulatory alternatives (and whether there are other regulatory alternatives the Agency should consider) and the Agency's analysis of them. In addition, OSHA requests comments and information on a number of specific topics and issues pertinent to the proposed standard. These are summarized below.

Regulatory Issues

In this section, we solicit public feedback on issues associated with the proposed standard and request information that would help the Agency craft the final standard. In addition to the issues specified here, OSHA also raises issues for comment on technical questions and discussions of economic issues in the PEA (OSHA, 2014). OSHA requests comment on all relevant issues, including health effects, risk assessment, significance of risk, technological and economic feasibility, and the provisions of the proposed regulatory text. In addition, OSHA requests comments on all of the issues raised by the Small Business Advocacy Review (SBAR) Panel, as summarized in the SBAR report (OSHA, 2008b)

We present these issues and requests for information in the first chapter of the preamble to assist readers as they review the preamble and consider any comments they may want to submit. The issues are presented here in summary form. However, to fully understand the questions in this section and provide substantive input in response to them, the sections of the preamble relevant to these issues should be reviewed. These include: Section V, Health Effects; Section VI, the Preliminary Risk Assessment; Section VIII, Significance of Risk; Section IX, Summary of the Preliminary Economic Analysis and Initial Regulatory Flexibility Analysis; and Section XVIII, Summary and Explanation of the Proposed Standard.

OSHA requests that comments be organized, to the extent possible, around the following issues and numbered questions. Comment on particular provisions should contain a heading setting forth the section and the paragraph in the proposed standard that the comment addresses. Comments addressing more than one section or paragraph will have correspondingly more headings.

Submitting comments in an organized manner and with clear reference to the issue raised will enable all participants to easily see what issues the commenter addressed and how they were addressed. Many commenters, especially small businesses, are likely to confine their comments to the issues that affect them, and they will benefit from being able to quickly identify comments on these issues in others' submissions. The Agency welcomes comments concerning all aspects of this proposal. However, OSHA is especially interested in responses, supported by evidence and reasons, to the following questions:

Health Effects

1. OSHA has described a variety of studies addressing the major adverse health effects that have been associated with exposure to beryllium. Using currently available epidemiologic and experimental studies, OSHA has made a preliminary determination that beryllium presents risks of lung cancer; sensitization; CBD at 0.1 µg/m³; and at higher exposures acute beryllium disease, and hepatic, renal, cardiovascular and ocular diseases. Is this determination correct? Are there additional studies or other data OSHA should consider in evaluating any of these health outcomes?

2. Has OSHA adequately identified and documented all critical health impairments associated with occupational exposure to beryllium? If not, what other adverse health effects should be added? Are there additional studies or other data OSHA should consider in evaluating any of these health outcomes?

3. Are there any additional studies, other data, or information that would affect the information discussed or significantly change the determination of material health impairment?

Please submit any relevant information, data, or additional studies (or citations to studies), and explain your reasons for recommending any studies you suggest.

Risk Assessment and Significance of Risk

4. OSHA has developed an analysis of health risks associated with occupational beryllium exposure, including an analysis of sensitization and CBD based on a selection of recent studies in the epidemiological literature, a data set on a population of beryllium machinists provided by the National Jewish Medical Research Center (NJMRC), and an assessment of lung cancer risk using an analysis provided by NIOSH. Did OSHA rely on the best available evidence in its risk assessment? Are there additional studies or other data OSHA should consider in evaluating risk for these health outcomes? Please provide the studies, citations to studies, or data you suggest.

5. OSHA preliminarily concluded that there is significant risk of material health impairment (lung cancer or CBD) from a working lifetime of occupational exposure to beryllium at the current TWA PEL of 2 µg/m³, which would be substantially reduced by the proposed TWA PEL of 0.2 µg/m³ and the alternative TWA PEL of 0.1 µg/m³. OSHA's preliminary risk assessment also concludes that there is still significant risk of CBD and lung cancer at the proposed PEL and the alternative PELs, although substantially less than at the current PEL. Are these preliminary conclusions reasonable, based on the best available evidence? If not, please provide a detailed explanation of your position, including data to support your position and a detailed analysis of OSHA's risk assessment if appropriate.

6. Please provide comment on OSHA's analysis of risk for beryllium sensitization, CBD and lung cancer. Are there important gaps or uncertainties in the analysis, such that the Agency's preliminary conclusions regarding significance of risk at the current, proposed, and alternative PELs may be in error? If so, please provide a detailed explanation and suggestions for how OSHA's analysis should be corrected or improved.

7. OSHA has made a preliminary determination that the available data are not sufficient or suitable for risk analysis of effects other than beryllium sensitization, CBD and lung cancer. Do you have, or are you aware of, studies or data that would be suitable for a risk assessment for these adverse health effects? Please provide the studies, citations to studies, or data you suggest.

(a) Scope

8. Has OSHA defined the scope of the proposed standard appropriately? Does it currently include employers who should not be covered, or exclude employers who should be covered by a comprehensive beryllium standard? Are you aware of employees in construction or maritime, or in general industry who deal with beryllium only as a trace contaminant, who may be at significant risk from occupational beryllium exposure? Please provide the basis for your response and any applicable supporting information.

(b) Definitions

9. Has OSHA defined the Beryllium lymphocyte proliferation test appropriately? If not, please provide the definition that you believe is appropriate. Please provide rationale and citations supporting your comments.

10. Has OSHA defined CBD Diagnostic Center appropriately? In particular, should a CBD diagnostic center be required to analyze biological samples on-site, or should diagnostic centers be allowed to send samples off-site for analysis? Is the list of tests and procedures a CBD Diagnostic Center is required to be able to perform appropriate? Should any of the tests or procedures be removed from the definition? Should other tests or procedures be added to the definition? Please provide rationale and information supporting your comments.

(d) Exposure Monitoring

11. Do you currently monitor for beryllium exposures in your workplace? If so, how often? Please provide the reasoning for the frequency of your monitoring. If periodic monitoring is performed at your workplace for exposures other than beryllium, with what frequency is it repeated?

12. Is it reasonable to allow discontinuation of monitoring based on one sample below the action level? Should more than one result below the action level be required to discontinue monitoring?

(e) Work Areas and Regulated Areas

The proposed standard would require employers to establish and maintain two types of areas: beryllium work areas, wherever employees are, or can reasonably be expected to be, exposed to any level of airborne beryllium; and regulated areas, wherever employees are, or can reasonably be expected to be, exposed to airborne beryllium at levels above the TWA PEL or STEL. Employers are required to demarcate beryllium work areas, but are not required to restrict access to beryllium work areas or provide respiratory protection or other forms of PPE within work areas with exposures at or below the TWA PEL or STEL. Employers must also demarcate regulated areas, including posting warning signs; restrict access to regulated areas; and provide respiratory protection and other PPE within regulated areas.

13. Does your workplace currently have regulated areas? If so, how are regulated areas demarcated?

14. Please describe work settings where establishing regulated areas could be problematic or infeasible. If establishing regulated areas is problematic, what approaches might be used to warn employees in such work settings of high risk areas?

(f) Methods of Compliance

Paragraph (f)(2) of the proposed standard would require employers to implement engineering and work practice controls to reduce employees' exposures to or below the TWA PEL and STEL. Where engineering and work practice controls are insufficient to reduce exposures to or below the TWA PEL and STEL, employers would still be required to implement them to reduce exposure as much as possible, and to supplement them with a respiratory protection program. In addition, for each operation where there is airborne beryllium exposure, the employer must ensure that at least one of the engineering and work practice controls listed in paragraph (f)(2) is in place, unless all of the listed controls are infeasible, or the employer can demonstrate that exposures are below the action level based on no fewer than two samples taken seven days apart.

15. Do you usually use engineering or work practices controls (local exhaust ventilation, isolation, substitution) to reduce beryllium exposures? If so, which controls do you use?

16. Are the controls and processes listed in paragraph (f)(2)(i)(A) appropriate for controlling beryllium exposures? Are there additional controls or processes that should be added to paragraph (f)(2)(i)(A)?

(g) Respiratory Protection

17. OSHA's asbestos standard (CFR 1910.1001) requires employers to provide each employee with a tight-fitting, powered air-purifying respirator (PAPR) instead of a negative pressure respirator when the employee chooses to use a PAPR and it provides adequate protection to the employee. Should the beryllium standard similarly require employers to provide PAPRs (instead of allowing a negative pressure respirator) when requested by the employee? Are there other circumstances where a PAPR should be specified as the appropriate respiratory protection? Please provide the basis for your response and any applicable supporting information.

(h) Personal Protective Clothing and Equipment

18. Do you currently require specific PPE or respirators when employees are working with beryllium? If so, what type?

19. The proposal requires PPE wherever work clothing or skin may become visibly contaminated with beryllium; where employees' skin can reasonably be expected to be exposed to soluble beryllium compounds; or where employee exposure exceeds or can reasonably be expected to exceed the TWA PEL or STEL. The requirement to use PPE where work clothing or skin may become “visibly contaminated” with beryllium differs from prior standards which do not require contamination to be visible in order for PPE to be required. Is “visibly contaminated” an appropriate trigger for PPE? Is there reason to require PPE where employees' skin can be exposed to insoluble beryllium compounds? Please provide the basis for your response and any applicable supporting information.

(i) Hygiene Areas and Practices

20. The proposal requires employers to provide showers in their facilities if (A) Exposure exceeds or can reasonably be expected to exceed the TWA PEL or STEL; and (B) Beryllium can reasonably be expected to contaminate employees' hair or body parts other than hands, face, and neck. Is this requirement reasonable and adequately protective of beryllium-exposed workers? Should OSHA amend the provision to require showers in facilities where exposures exceed the PEL or STEL, without regard to areas of bodily contamination?

(j) Housekeeping

21. The proposed rule prohibits dry sweeping or brushing for cleaning surfaces in beryllium work areas unless HEPA-filtered vacuuming or other methods that minimize the likelihood and level of exposure have been tried and were not effective. Please comment on this provision. What methods do you use to clean work surfaces at your facility? Are HEPA-filtered vacuuming or other methods to minimize beryllium exposure used to clean surfaces at your facility? Have they been effective? Are there any circumstances under which dry sweeping or brushing are necessary? Please explain your response.

22. The proposed rule requires that materials designated for recycling that are visibly contaminated with beryllium particulate shall be cleaned to remove visible particulate, or placed in sealed, impermeable enclosures. However, small particles (<10 μg) may not be visible to the naked eye, and there are studies suggesting that small particles may penetrate the skin, beyond which beryllium sensitization can occur (Tinkle et al., 2003). OSHA requests feedback on this provision. Should OSHA require that all material to be recycled be decontaminated regardless of perceived surface cleanliness? Should OSHA require that all material disposed or discarded be in enclosures regardless of perceived surface cleanliness? Please provide explanation or data to support your comments.

(k) Medical Surveillance

The proposed requirements for medical surveillance include: (1) Medical examinations, including a test for beryllium sensitization, for employees who are exposed to beryllium above the proposed PEL for 30 days or more per year, who are exposed to beryllium in an emergency, or who show signs or symptoms of CBD; and (2) CT scans for employees who were exposed above the proposed PEL for more than 30 days in a 12-month period for 5 years or more. The proposed standard would require periodic medical exams to be provided for employees in the medical surveillance program annually, while tests for beryllium sensitization and CT scans would be provided to eligible employees biennially.

23. Is medical surveillance being provided for beryllium-exposed employees at your worksite? If so:

a. Do you provide medical surveillance to employees under another OSHA standard or as a matter of company policy? What OSHA standard(s) does the program address?

b. How many employees are included, and how do you determine which employees receive medical surveillance (e.g., by exposure level, other factors)?

c. Who administers and implements the medical surveillance (e.g., company doctor, nurse practitioner, physician assistant, or nurse; or outside doctor, nurse practitioner, physician assistant, or nurse)?

d. What examinations, tests, or evaluations are included in the medical surveillance program, and with what frequency are they administered? Does your program include a surveillance program specifically for beryllium-related health effects (e.g., the BeLPT or other tests for beryllium sensitization)?

e. If your facility offers the BeLPT, please provide feedback and data on your experience with the BeLPT, including the analytical or interpretive procedure you use and its role in your facility's exposure control program. Has identification of sensitized workers led to interventions to reduce exposures to sensitized individuals, or in the facility generally? If a worker is found to be sensitized, do you track worker health and possible progression of disease beyond sensitization? If so, how is this done?

f. What difficulties and benefits (e.g., health, reduction in absenteeism, or financial) have you experienced with your medical surveillance program? If applicable, please discuss benefits and difficulties you have experienced with the use of the BeLPT, providing detailed information or examples if possible.

g. What are the costs of your medical surveillance program? How do your costs compare with OSHA's estimated unit costs for the physical examination and employee time involved in the medical surveillance program? Are OSHA's baseline assumptions and cost estimates for medical surveillance consistent with your experiences providing medical surveillance to your employees?

24. Please review paragraph (k) of the proposed rule, Medical Surveillance, and comment on the frequency and contents of medical surveillance in the proposed rule. Is 30 days from initial assignment a reasonable time at which to provide a medical exam? Should there be a requirement for beryllium sensitization testing at time of employment? Should there be a requirement for beryllium sensitization testing at an employee's exit exam, regardless of when the employee's most recent sensitization test was administered? Are the tests required and the testing frequencies specified appropriate? Should sensitized employees have the opportunity to be examined at a CBD Diagnostic Center more than once following a confirmed positive BeLPT? Are there additional tests or alternate testing schedules you would suggest? Should the skin be examined for signs and symptoms of beryllium exposure or other medical issues, as well as for breaks and wounds? Please explain the basis for your position and provide data or studies if applicable.

25. Please provide comments on the proposed requirements regarding referral of a sensitized employee to a CBD diagnostic center, which specify referral to a diagnostic center “mutually agreed upon” by the employer and employee. Is this requirement for mutual agreement necessary and appropriate? How should a diagnostic center be chosen if the employee and employer cannot come to agreement? Should OSHA consider alternate language, such as referral for CBD evaluation at a diagnostic center in a reasonable location?

26. In the proposed rule, OSHA specifies that all medical examinations and procedures required by the standard must be performed by or under the direction of a licensed physician. Are physicians available in your geographic area to provide medical surveillance to workers who are covered by the proposed rule? Are other licensed health care professionals available to provide medical surveillance? Do you have access to other qualified personnel such as qualified X-ray technicians, and pulmonary specialists? Should the proposal be amended to allow examination by, or under the direction of, a physician or other licensed health care professional (PLHCP)? Please explain your position. Please note what you consider your geographic area in responding to this question.

27. The proposed standard requires the employer to obtain the Licensed Physician's Written Medical Opinion from the PLHCP within 30 days of the examination. Should OSHA revise the medical surveillance provisions of the proposed standard to allow employees to choose what, if any, medical information goes to the employer from the PLHCP? For example, the employer could instead be required to obtain a certification from the PLCHP within 30 days of the examination stating (1) when the examination took place, (2) that the examination complied with the standard, and (3) that the PLHCP provided the employee a copy of the Licensed Physician's Written Medical Opinion required by the standard. The PLHCP would need the employee's written consent to send the employer the Licensed Physician's Written Medical Opinion or any other medical information about the employee. This approach might lead to corresponding changes in proposed paragraphs (f)(1) (written exposure control program), (l) (medical removal) and (n) (recordkeeping) to reflect that employers will not automatically be receiving any medical information about employees as a result of the medical surveillance required by the proposed standard, but would instead only receive medical information the employee chooses to share with the employer. Please comment on the relative merits of the proposed standard's requirement that employers obtain the PLHCP's written opinion or an alternative that would provide employees with greater discretion over the information that goes to employers, and explain the basis for your position and the potential impact on the benefits of medical surveillance.

28. Appendix A to the proposed standard reviews procedures for conducting and interpreting the results of BeLPT testing for beryllium sensitization. Is there now, or should there be, a standard method for BeLPT laboratory procedure? If yes, please describe the existing or proposed method. Is there now, or should there be, a standard algorithm for interpreting BeLPT results to determine sensitization? Please describe the existing or proposed laboratory method or interpretation algorithm. Should OSHA require that BeLPTs performed to comply with the medical surveillance provisions of this rule adhere to the Department of Energy (DOE) analytical and interpretive specifications issued in 2001? Should interpretation of laboratory results be delegated to the employee's occupational physician or PLHCP?

29. Should OSHA require the clinical laboratories performing the BeLPT to be accredited by the College of American Pathologists or another accreditation organization approved under the Clinical Laboratory Improvement Amendments (CLIA)? What other standards, if any, should be required for clinical laboratories providing the BeLPT?

30. Are there now, or are there being developed, alternative tests to the BeLPT you would suggest? Please explain the reasons for your suggestion. How should alternative tests for beryllium sensitization be evaluated and validated? How should OSHA determine whether a test for beryllium sensitization is more reliable and accurate than the BeLPT? Please see Appendix A to the proposed standard for a discussion of the accuracy of the BeLPT.

31. The proposed rule requires employers to provide OSHA with the results of BeLPTs performed to comply with the medical surveillance provisions upon request, provided that the employer obtains a release from the tested employee. Will this requirement be unduly burdensome for employers? Are there alternative organizations that would be appropriate to send test results to?

(l) Medical Removal Protection

The proposed requirements for medical removal protection provide an option for medical removal to an employee who is working in a job with exposure at or above the action level and is diagnosed with CBD or confirmed positive for beryllium sensitization. If the employee chooses removal, the employer must remove the employee to comparable work in a work environment where exposure is below the action level, or if comparable work is not available, must place the employee on paid leave for 6 months or until such time as comparable work becomes available. In either case, the employer must maintain for 6 months the employee's base earnings, seniority, and other rights and benefits that existed at the time of removal.

32. Do you provide MRP at your facility? If so, please comment on the program's benefits, difficulties, and costs, and the extent to which eligible employees make use of MRP.

33. OSHA has included requirements for medical removal protection (MRP) in the proposed rule, which includes provisions for medical removal for employees with beryllium sensitization or CBD, and an extension of removed employees' rights and benefits for six months. Are beryllium sensitization and CBD appropriate triggers for medical removal? Are there other medical conditions or findings that should trigger medical removal? For what amount of time should a removed employee's benefits be extended?

(p) Appendices

34. Some OSHA health standards include appendices that address topics such as the hazards associated with the regulated substance, health screening considerations, occupational disease questionnaires, and PLHCP obligations. In this proposed rule, OSHA has included a non-mandatory appendix to describe and discuss the BeLPT (Appendix A), and a non-mandatory appendix presenting a non-exhaustive list of engineering controls employers may use to comply with paragraph (f) (Appendix B). What would be the advantages and disadvantages of including each appendix in the final rule? What would be the advantages and disadvantages of providing this information in guidance materials?

35. What additional information, if any, should be included in the appendices? What additional information, if any, should be provided in guidance materials?

General

36. The current beryllium proposal includes triggers that require employers to initiate certain provisions, programs, and activities to protect workers from beryllium exposure. All employers covered under an OSHA health standard are required to initiate certain activities such as initial monitoring to evaluate the potential hazard to employees. OSHA health standards typically include ancillary provisions with various triggers indicating when an employer covered under the standard would need to comply with a provision. The most common triggers are ones based an exposure level such as the PEL or action level. These exposure level triggers are sometimes combined with a minimum duration of exposure (e.g., ≥ 30 days per year). Other triggers may include reasonably anticipated exposure, medical surveillance findings, certain work activities, or simply the presence of the regulated substance in the workplace.

For the current Proposal, exposures to beryllium above the TWA PEL or STEL trigger the provisions for regulated areas, additional or enhanced engineering or work practice controls to reduce airborne exposures to or below the TWA PEL and STEL, personal protective clothing and equipment, medical surveillance, showers, and respiratory protection if feasible engineering and work practice controls cannot reduce airborne exposures to or below the TWA PEL and STEL. Exposures at or above the action level in turn trigger the provisions for periodic exposure monitoring, and medical removal eligibility (along with a diagnosis of CBD or confirmed positive for beryllium sensitization). Finally, an employer covered under the scope of the proposed standard must establish a beryllium work area where employees are, or can reasonably be expected to be, exposed to airborne beryllium regardless of the level of exposure. In beryllium work areas, employers must implement a written exposure control plan, provide washing facilities and change rooms (change rooms are only necessary if employees are required to remove their personal clothing), and follow housekeeping provisions. The employers must also implement at least one of the engineering and work practice controls listed in paragraph (f)(2) of the proposed standard. An employer is exempt from this requirement if he or she can demonstrate that such controls are not feasible or that exposures are below the action level.

Certain provisions are triggered by one condition and other provisions are triggered only if multiple conditions are present. For example, medical removal is only triggered if an employee has CBD or is confirmed positive AND the employee is exposed at or above the action level.

OSHA is requesting comment on the triggers in the proposed beryllium standard. Are the triggers OSHA has proposed appropriate? OSHA is also requesting comment on these triggers relative to the regulatory alternatives affecting the scope and PELs as described in this preamble in section I, Issues and Alternatives. For example, are the triggers in the proposed standard appropriate for Alternative #1a, which would expand the scope of the proposed standard to include all operations in general industry where beryllium exists only as a trace contaminant (less than 0.1% beryllium by weight)? Are the triggers appropriate for the alternatives that change the TWA PEL, STEL, and action level? Please specify the trigger and the alternative, if applicable, and why you agree or disagree with the trigger.

Relevant Federal Rules Which May Duplicate, Overlap, or Conflict With the Proposed Rule

37. In Section IX—Preliminary Economic Analysis under the Initial Regulatory Flexibility Analysis, OSHA identifies, to the extent practicable, all relevant Federal rules which may duplicate, overlap, or conflict with the proposed rule. One potential area of overlap is with the U.S. Department of Energy (DOE) beryllium program. In 1999, DOE established a chronic beryllium disease prevention program (CBDPP) to reduce the number of workers (DOE employees and DOE contractors) exposed to beryllium at DOE facilities (10 CFR part 850, published at 64 FR 68854-68914 (Dec. 8, 1999)). In establishing this program, DOE has exercised its statutory authority to prescribe and enforce occupational safety and health standards. Therefore pursuant to section 4(b)(1) of the OSH Act, 29 U.S.C. 653(b)(1), the DOE facilities are exempt from OSHA jurisdiction.

Nevertheless, under 10 CFR 850.22, DOE has included in its CBDPP regulation a requirement for compliance with the current OSHA permissible exposure limit (PEL), and any lower PEL that OSHA establishes in the future. Thus, although DOE has preempted OSHA's standard from applying at DOE facilities and OSHA cannot exercise any authority at those facilities, DOE relies on OSHA's PEL in implementing its own program. However, DOE's decision to tie its own standard to OSHA's PEL has little consequence to this rulemaking because the requirements in DOE's beryllium program (controls, medical surveillance, etc.) are triggered by DOE's action level of 0.2 µg/m³, which is much lower than DOE's existing PEL and the same as OSHA's proposed PEL. DOE's action level is not tied to OSHA's standard, so 10 CFR 850.22 would not require the CBDPP's action level or any non-PEL requirements to be automatically adjusted as a result of OSHA's rulemaking. For this reason, DOE has indicated to OSHA that OSHA's proposed rule would not have any impact on DOE's CBDPP, particularly since 10 CFR 850.25(b), Exposure reduction and minimization, requires DOE contractors to reduce exposures to below the DOE's action level of 0.2 µg/m³, if practicable.

DOE has expressed to OSHA that DOE facilities are already in compliance with 10 CFR 850 and its action level of 0.2 µg/m, so the only potential impact on DOE's CBDPP that could flow from OSHA's rulemaking would be if OSHA ultimately adopted a PEL of 0.1 µg/m³, as discussed in alternative #4, instead of the proposed PEL of 0.2 µg/m³, and DOE did not make any additional adjustments to its standards. Even in that hypothetical scenario, the impact would still be limited because of the odd result that DOE's PEL would drop below its own action level, while the action level would continue to serve as the trigger for most of DOE's program requirements.

DOE also has noted some potential overlap with a separate DOE provision in 10 CFR part 851, which requires its contractors to comply with DOE's CBDPP (10 CFR 851.23(a)(1)) and also with all OSHA standards under 29 CFR part 1910 except “Ionizing Radiation” (§ 1910.1096) (10 CFR 851.23(a)(3)). These requirements, which DOE established in 2006 (71 FR 6858 (February 9, 2006)), make sense in light of OSHA's current regulation because OSHA's only beryllium protection is a PEL, so compliance with 10 CFR 851.23(a)(1) and (3) merely make OSHA's current PEL the relevant level for purposes of the CBDPP. However, its function would be less clear if OSHA adopts a beryllium standard as proposed. OSHA's proposed beryllium standard would establish additional substantive protections beyond the PEL. Consequently, notwithstanding the CBDPP's preemptive effect on the OSHA beryllium standard as a result of 29 U.S.C. 653(b)(1), 10 CFR 851.23(a)(3) could be read to require DOE contractors to comply with all provisions in OSHA's proposal (if finalized), including the ancillary provisions, creating a dual regulatory scheme for beryllium protection at DOE facilities.

DOE officials have indicated that this is not their intent. Instead, their intent is that DOE contractors comply solely with the CBDPP provisions in 10 CFR part 850 for protection from beryllium. Based on its discussions with DOE officials, OSHA anticipates that DOE will clarify that its contractors do not need to comply with any ancillary provisions in a beryllium standard that OSHA may promulgate.

OSHA can envision several potential scenarios developing from its rulemaking, ranging from OSHA retaining the proposed PEL of 0.2 µg/m³ and action level of 0.1 µg/m³ in the final rule to adopting the PEL of 0.1 µg/m³, as discussed in alternative #4. Because OSHA's beryllium standard does not apply directly to DOE facilities, and the only impact of its rules on those facilities is the result of DOE's regulatory choices, there is also a range of actions that DOE could take to minimize any potential impact of any change to OSHA's rules, including (1) taking no action at all, (2) simply clarifying the CBDPP, as described above, to mean that OSHA's beryllium standard (other than its PEL) does not apply to contractors, or (3) revising both parts 850 and 851 to completely disassociate DOE's regulation of beryllium at DOE facilities from OSHA's regulation of beryllium.

OSHA is aware that, in the preamble to its 1999 CBDPP rule, DOE analyzed the costs for implementing the CBDPP for action levels of 0.1 µg/m³, 0.2 µg/m³, and 0.5 µg/m³ (64 FR 68875, December 8, 1999). DOE estimated costs for periodic exposure monitoring, notifying workers of the results of such monitoring, exposure reduction and minimization, regulated areas, change rooms and showers, respiratory protection, protective clothing, and disposal of protective clothing. All of these provisions are triggered by DOE's action level (64 FR 68874, December 8, 1999). Although DOE's rule is not identical to OSHA's proposed standard, OSHA believes that DOE's costs are sufficiently representative to form the basis of a preliminary estimate of the costs that could flow from OSHA's standard, if finalized.

Based on the range of potential scenarios and the prior DOE cost estimates, OSHA estimates that the annual cost impact on DOE facilities could range from $0 to $4,065,768 (2010 dollars). The upper end of the cost range would reflect the unlikely scenario in which OSHA promulgates a final PEL of 0.1 µg/m³, 10 CFR 851.23(a)(3) is found to compel DOE contractors to comply with OSHA's comprehensive beryllium standard in addition to DOE's CBDPP, and DOE takes no action to clarify that OSHA's beryllium standard does not apply to DOE contractors. The lower end of the cost range assumes OSHA promulgates its rule as proposed with a PEL of 0.2 µg/m³ and action level of 0.1 µg/m³, and DOE clarifies that it intends its contractors to follow DOE's CBDPP and not OSHA's beryllium standard, so that the ancillary provisions of OSHA's beryllium standard do not apply to DOE facilities. Additionally, OSHA assumes that DOE contractors are in compliance with DOE's current rule and therefore took the difference in cost between implementation of an action level of 0.2 µg/m³ and an action level of 0.1 µg/m³ for the above estimates. Finally, OSHA used the GDP price deflator to present the cost estimate in 2010 dollars.

OSHA requests comment on the potential overlap of DOE's rule with OSHA's proposed rule.

II. Pertinent Legal Authority

The purpose of the Occupational Safety and Health Act, 29 U.S.C. 651 et seq. (“the Act”), is to “. . . assure so far as possible every working man and woman in the nation safe and healthful working conditions and to preserve our human resources.” 29 U.S.C. 651(b).

To achieve this goal Congress authorized the Secretary of Labor (the Secretary) to promulgate and enforce occupational safety and health standards. 29 U.S.C. 654(b) (requiring employers to comply with OSHA standards), 655(a) (authorizing summary adoption of existing consensus and federal standards within two years of the Act's enactment), and 655(b) (authorizing promulgation, modification or revocation of standards pursuant to notice and comment).

The Act provides that in promulgating health standards dealing with toxic materials or harmful physical agents, such as this proposed standard regulating occupational exposure to beryllium, the Secretary, shall set the standard which most adequately assures, to the extent feasible, on the basis of the best available evidence that no employee will suffer material impairment of health or functional capacity even if such employee has regular exposure to the hazard dealt with by such standard for the period of his working life. See 29 U.S.C. 655(b)(5).

The Supreme Court has held that before the Secretary can promulgate any permanent health or safety standard, he must make a threshold finding that significant risk is present and that such risk can be eliminated or lessened by a change in practices. Industrial Union Dept., AFL-CIO v. American Petroleum Institute, 448 U.S. 607, 641-42 (1980) (plurality opinion) (“The Benzene case”). Thus, section 6(b)(5) of the Act requires health standards to reduce significant risk to the extent feasible. Id.

The Court further observed that what constitutes “significant risk” is “not a mathematical straitjacket” and must be “based largely on policy considerations.” The Benzene case, 448 U.S. at 655. The Court gave the example that if,

. . . the odds are one in a billion that a person will die from cancer . . . the risk clearly could not be considered significant. On the other hand, if the odds are one in one thousand that regular inhalation of gasoline vapors that are 2% benzene will be fatal, a reasonable person might well consider the risk significant. [Id.]

OSHA standards must be both technologically and economically feasible. United Steelworkers v. Marshall, 647 F.2d 1189, 1264 (D.C. Cir. 1980) (“The Lead I case”). The Supreme Court has defined feasibility as “capable of being done.” Am. Textile Mfrs. Inst. v. Donovan, 452 U.S. 490, 509-510 (1981) (“The Cotton Dust case”). The courts have further clarified that a standard is technologically feasible if OSHA proves a reasonable possibility,

. . . within the limits of the best available evidence . . . that the typical firm will be able to develop and install engineering and work practice controls that can meet the PEL in most of its operations. [See The Lead I case, 647 F.2d at 1272]

With respect to economic feasibility, the courts have held that a standard is feasible if it does not threaten massive dislocation to or imperil the existence of the industry. Id. at 1265. A court must examine the cost of compliance with an OSHA standard,

. . . in relation to the financial health and profitability of the industry and the likely effect of such costs on unit consumer prices . . . [T]he practical question is whether the standard threatens the competitive stability of an industry, . . . or whether any intra-industry or inter-industry discrimination in the standard might wreck such stability or lead to undue concentration. [Id. (citing Indus. Union Dep't, AFL-CIO v. Hodgson, 499 F.2d 467 (D.C. Cir. 1974))]

The courts have further observed that granting companies reasonable time to comply with new PELs may enhance economic feasibility. The Lead I case at 1265. While a standard must be economically feasible, the Supreme Court has held that a cost-benefit analysis of health standards is not required by the Act because a feasibility analysis is required. The Cotton Dust case, 453 U.S. at 509.

Finally, sections 6(b)(7) and 8(c) of the Act authorize OSHA to include among a standard's requirements labeling, monitoring, medical testing, and other information-gathering and -transmittal provisions. 29 U.S.C. 655(b)(7), 657(c).

III. Events Leading to the Proposed Standards

The first occupational exposure limit for beryllium was set in 1949 by the Atomic Energy Commission (AEC), which required that beryllium exposure in the workplaces under its jurisdiction be limited to 2 µg/m³ as an 8-hour time-weighted average (TWA), and 25 µg/m³ as a peak exposure never to be exceeded (Department of Energy, 1999). These exposure limits were adopted by all AEC installations handling beryllium, and were binding on all AEC contractors involved in the handling of beryllium.

In 1956, the American Industrial Hygiene Association (AIHA) published a Hygienic Guide which supported the AEC exposure limits. In 1959, the American Conference of Governmental Industrial Hygienists (ACGIH®) also adopted a Threshold Limit Value (TLV®) of 2 µg/m³ as an 8-hour TWA (Borak, 2006).

In 1971, OSHA adopted, under Section 6(a) of the Occupational Safety and Health Act of 1970, and made applicable to general industry, a national consensus standard (ANSI Z37.29-1970) for beryllium and beryllium compounds. The standard set a permissible exposure limit (PEL) for beryllium and beryllium compounds at 2 µg/m³ as an 8-hour TWA; 5 µg/m³ as an acceptable ceiling concentration; and 25 µg/m³ as an acceptable maximum peak above the acceptable ceiling concentration for a maximum duration of 30 minutes in an 8-hour shift (OSHA, 1971).

Section 6(a) stipulated that in the first two years after the effective date of the Act, OSHA was to promulgate “start-up” standards, on an expedited basis and without public hearing or comment, based on national consensus or established Federal standards that improved employee safety or health. Pursuant to that authority, in 1971, OSHA promulgated approximately 425 PELs for air contaminants, including beryllium, derived principally from Federal standards applicable to government contractors under the Walsh-Healey Public Contracts Act, 41 U.S.C. 35, and the Contract Work Hours and Safety Standards Act (commonly known as the Construction Safety Act), 40 U.S.C. 333. The Walsh-Healey Act and Construction Safety Act standards, in turn, had been adopted primarily from ACGIH®'s TLV®s.

The National Institute for Occupational Safety and Health (NIOSH) issued a document entitled Criteria for a Recommended Standard: Occupational Exposure to Beryllium (Criteria Document) in June 1972. OSHA reviewed the findings and recommendations contained in the Criteria Document along with the AEC control requirements for beryllium exposure. OSHA also considered existing data from animal and epidemiological studies, and studies of industrial processes of beryllium extraction, refinement, fabrication, and machining. In 1975, OSHA asked NIOSH to update the evaluation of the existing data pertaining to the carcinogenic potential of beryllium. In response to OSHA's request, the Director of NIOSH stated that, based on animal data and through all possible routes of exposure including inhalation, “beryllium in all likelihood represents a carcinogenic risk to man.”

In October 1975, OSHA proposed a new beryllium standard for all industries based on information that beryllium caused cancer in animal experiments (40 FR 48814 (October 17, 1975)). Adoption of this proposal would have lowered the 8-hour TWA exposure limit from 2 µg/m³ to 1 µg/m³. In addition, the proposal included ancillary provisions for such topics as exposure monitoring, hygiene facilities, medical surveillance, and training related to the health hazards from beryllium exposure. The rulemaking was never completed.

In 1977, NIOSH recommended an exposure limit of 0.5 µg/m³ and identified beryllium as a potential occupational carcinogen. In December 1998, ACGIH published a Notice of Intended Change for its beryllium exposure limit. The notice proposed a lower TLV of 0.2 µg/m³ over an 8-hour TWA based on evidence of CBD and sensitization in exposed workers.

In 1999, the Department of Energy (DOE) issued a Chronic Beryllium Disease Prevention Program (CBDPP) Final Rule for employees exposed to beryllium in its facilities (DOE, 1999). The DOE rule set an action level of 0.2 μg/m³, and adopted OSHA's PEL of 2 μg/m³ or any more stringent PEL OSHA might adopt in the future. The DOE action level triggers workplace precautions and control measures such as periodic monitoring, exposure reduction or minimization, regulated areas, hygiene facilities and practices, respiratory protection, protective clothing and equipment, and warning signs (DOE, 1999).

Also in 1999, OSHA was petitioned by the Paper, Allied-Industrial, Chemical and Energy Workers International Union (PACE) (OSHA, 2002) and by Dr. Lee Newman and Ms. Margaret Mroz, from the National Jewish Medical Research Center (NJMRC) (OSHA, 2002), to promulgate an Emergency Temporary Standard (ETS) for beryllium in the workplace. In 2001, OSHA was petitioned for an ETS by Public Citizen Health Research Group and again by PACE (OSHA, 2002). In order to promulgate an ETS, the Secretary of Labor must prove (1) that employees are exposed to grave danger from exposure to a hazard, and (2) that such an emergency standard is necessary to protect employees from such danger (29 U.S.C. 655(c)). The burden of proof is on the Department and because of the difficulty of meeting this burden, the Department usually proceeds when appropriate with 6(b) rulemaking rather than a 6(c) ETS. Thus, instead of granting the ETS requests, OSHA instructed staff to further collect and analyze research regarding the harmful effects of beryllium.

On November 26, 2002, OSHA published a Request for Information (RFI) for “Occupational Exposure to Beryllium” (OSHA, 2002). The RFI contained questions on employee exposure, health effects, risk assessment, exposure assessment and monitoring methods, control measures and technological feasibility, training, medical surveillance, and impact on small business entities. In the RFI, OSHA expressed concerns about health effects such as CBD, lung cancer, and beryllium sensitization. OSHA pointed to studies indicating that even short-term exposures below OSHA's PEL of 2 µg/m³ could lead to CBD. The RFI also cited studies describing the relationship between beryllium sensitization and CBD (67 FR at 70708). In addition, OSHA stated that beryllium had been identified as a carcinogen by organizations such as NIOSH, the International Agency for Research on Cancer (IARC), and the Environmental Protection Agency (EPA); and cancer had been evidenced in animal studies (67 FR at 70709).

On November 15, 2007, OSHA convened a Small Business Advocacy Review Panel for a draft proposed standard for occupational exposure to beryllium. OSHA convened this panel under Section 609(b) of the Regulatory Flexibility Act (RFA), as amended by the Small Business Regulatory Enforcement Fairness Act of 1996 (SBREFA) (5 U.S.C. 601 et seq.).

The Panel included representatives from OSHA, the Solicitor's Office of the Department of Labor, the Office of Advocacy within the Small Business Administration, and the Office of Information and Regulatory Affairs of the Office of Management and Budget. Small Entity Representatives (SERs) made oral and written comments on the draft rule and submitted them to the panel.

The SBREFA Panel issued a report which included the SERs' comments on January 15, 2008. SERs expressed concerns about the impact of the ancillary requirements such as exposure monitoring and medical surveillance. Their comments addressed potential costs associated with compliance with the draft standard, and possible impacts of the standard on market conditions, among other issues. In addition, many SERs sought clarification of some of the ancillary requirements such as the meaning of “routine” contact or “contaminated surfaces.”

The SBREFA Panel issued a number of recommendations, which OSHA carefully considered. In section XVIII of this preamble, Summary and Explanation, OSHA has responded to the Panel's recommendations and clarified the requirements about which SERs expressed confusion. OSHA also examined the regulatory alternatives recommended by the SBREFA Panel. The regulatory alternatives examined by OSHA are listed in section I of this preamble, Issues and Alternatives. The alternatives are discussed in greater detail in section XVIII of this preamble, Summary and Explanation, and in the PEA (OSHA, 2014). In addition, the Agency intends to develop interpretive guidance documents following the publication of a final rule.

In 2010, OSHA hired a contractor to oversee an independent scientific peer review of a draft preliminary beryllium health effects evaluation (OSHA, 2010a) and a draft preliminary beryllium risk assessment (OSHA, 2010b). The contractor identified experts familiar with beryllium health effects research and ensured that these experts had no conflict of interest or apparent bias in performing the review. The contractor selected five experts with expertise in such areas as pulmonary and occupational medicine, CBD, beryllium sensitization, the BeLPT, beryllium toxicity and carcinogenicity, and medical surveillance. Other areas of expertise included animal modeling, occupational epidemiology, biostatistics, risk and exposure assessment, exposure-response modeling, beryllium exposure assessment, industrial hygiene, and occupational/environmental health engineering.

Regarding the health effects evaluation, the peer reviewers concluded that the health effect studies were described accurately and in sufficient detail, and OSHA's conclusions based on the studies were reasonable. The reviewers agreed that the OSHA document covered the significant health endpoints related to occupational beryllium exposure. Peer reviewers considered the preliminary conclusions regarding beryllium sensitization and CBD to be reasonable and well presented in the draft health evaluation section. All reviewers agreed that the scientific evidence supports sensitization as a necessary condition in the development of CBD. In response to reviewers' comments, OSHA made revisions to more clearly describe certain sections of the health effects evaluation. In addition, OSHA expanded its discussion regarding the BeLPT.

Regarding the preliminary risk assessment, the peer reviewers were highly supportive of the Agency's approach and major conclusions. The peer reviewers stated that the key studies were appropriate and their selection clearly explained in the document. They regarded the preliminary analysis of these studies to be reasonable and scientifically sound. The reviewers supported OSHA's conclusion that substantial risk of sensitization and CBD were observed in facilities where the highest exposure generating processes had median full-shift exposures around 0.2 µg/m³ or higher, and that the greatest reduction in risk was achieved when exposures for all processes were lowered to 0.1 µg/m³ or below.

In February 2012 the Agency received for consideration a draft recommended standard for beryllium (Materion and USW, 2012). This draft proposal was the product of a joint effort between two stakeholders: Materion Corporation, a leading producer of beryllium and beryllium products in the United States, and the United Steelworkers, an international labor union representing workers who manufacture beryllium alloys and beryllium-containing products in a number of industries. The United Steelworkers and Materion sought to craft an OSHA-like model beryllium standard that would have support from both labor and industry. OSHA has considered this proposal along with other information submitted during the development of the Notice of Proposed Rulemaking for beryllium.

IV. Chemical Properties and Industrial Uses

Chemical and Physical Properties

Beryllium (Be; CAS Number 7440-41-7) is a silver-grey to greyish-white, strong, lightweight, and brittle metal. It is a Group IIA element with an atomic weight of 9.01, atomic number of 4, melting point of 1,287 °C, boiling point of 2,970°C, and a density of 1.85 at 20 °C (NTP 2014). It occurs naturally in rocks, soil, coal, and volcanic dust (ATSDR, 2002). Beryllium is insoluble in water and soluble in acids and alkalis. It has two common oxidation states, Be(0) and Be(+2). There are several beryllium compounds with unique CAS numbers and chemical and physical properties. Table IV-1 describes the most common beryllium compounds.

The physical and chemical properties of beryllium were realized early in the 20th century, and it has since gained commercial importance in a wide range of industries. Beryllium is lightweight, hard, spark resistant, non-magnetic, and has a high melting point. It lends strength, electrical and thermal conductivity, and fatigue resistance to alloys (NTP, 2014). Beryllium also has a high affinity for oxygen in air and water, which can cause a thin surface film of beryllium oxide to form on the bare metal, making it extremely resistant to corrosion. These properties make beryllium alloys highly suitable for defense, nuclear, and aerospace applications (IARC, 1993).

There are approximately 45 mineralized forms of beryllium. In the United States, the predominant mineral form mined commercially and refined into pure beryllium and beryllium alloys is bertrandite. Bertrandite, while containing less than 1% beryllium compared to 4% in beryl, is easily and efficiently processed into beryllium hydroxide (IARC, 1993). Imported beryl is also converted into beryllium hydroxide as the United States has very little beryl that can be economically mined (USGS, 2013a).

Industrial Uses

Materion Corporation, formerly called Brush Wellman, is the only producer of primary beryllium in the United States. Beryllium is used in a variety of industries, including aerospace, defense, telecommunications, automotive, electronic, and medical specialty industries. Pure beryllium metal is used in a range of products such as X-ray transmission windows, nuclear reactor neutron reflectors, nuclear weapons, precision instruments, rocket propellants, mirrors, and computers (NTP, 2014). Beryllium oxide is used in components such as ceramics, electrical insulators, microwave oven components, military vehicle armor, laser structural components, and automotive ignition systems (ATSDR, 2002). Beryllium oxide ceramics are used to produce sensitive electronic items such as lasers and satellite heat sinks.

Beryllium alloys, typically beryllium/copper or beryllium/aluminum, are manufactured as high beryllium content or low beryllium content alloys. High content alloys contain greater than 30% beryllium. Low content alloys are typically less than 3% beryllium. Beryllium alloys are used in automotive electronics (e.g., electrical connectors and relays and audio components), computer components, home appliance parts, dental appliances (e.g., crowns), bicycle frames, golf clubs, and other articles (NTP, 2014; Ballance et al., 1978; Cunningham et al., 1998; Mroz, et al., 2001). Electrical components and conductors are stamped and formed from beryllium alloys. Beryllium-copper alloys are used to make switches in automobiles (Ballance et al., 1978, 2002; Cunningham et al., 1998) and connectors, relays, and switches in computers, radar, satellite, and telecommunications equipment (Mroz et al., 2001). Beryllium-aluminum alloys are used in the construction of aircraft, high resolution medical and industrial X-ray equipment, and mirrors to measure weather patterns (Mroz et al., 2001). High content and low content beryllium alloys are precision machined for military and aerospace applications. Some welding consumables are also manufactured using beryllium.

Beryllium is also found as a trace metal in materials such as aluminum ore, abrasive blasting grit, and coal fly ash. Abrasive blasting grits such as coal slag and copper slag contain varying concentrations of beryllium, usually less than 0.1% by weight. The burning of bituminous and sub-bituminous coal for power generation causes the naturally occurring beryllium in coal to accumulate in the coal fly ash byproduct. Scrap and waste metal for smelting and refining may also contain beryllium. A detailed discussion of the industries and job tasks using beryllium is included in the Preliminary Economic Analysis (OSHA, 2014).

Occupational exposure to beryllium can occur from inhalation of dusts, fume, and mist. Beryllium dusts are created during operations where beryllium is cut, machined, crushed, ground, or otherwise mechanically sheared. Mists can also form during operations that use machining fluids. Beryllium fume can form while welding with or on beryllium components, and from hot processes such as those found in metal foundries.

Occupational exposure to beryllium can also occur from skin, eye, and mucous membrane contact with beryllium particulate or solutions.

V. Health Effects

Beryllium-associated health effects, including acute beryllium disease (ABD), beryllium sensitization (also referred to in this preamble as “sensitization”), chronic beryllium disease (CBD), and lung cancer, can lead to a number of highly debilitating and life-altering conditions including pneumonitis, loss of lung capacity (reduction in pulmonary function leading to pulmonary dysfunction), loss of physical capacity associated with reduced lung capacity, systemic effects related to pulmonary dysfunction, and decreased life expectancy (NIOSH, 1972).

This Health Effects section presents information on beryllium and its compounds, the fate of beryllium in the body, research that relates to its toxic mechanisms of action, and the scientific literature on the adverse health effects associated with beryllium exposure, including ABD, sensitization, CBD, and lung cancer. OSHA considers CBD to be a progressive illness with a continuous spectrum of symptoms ranging from no symptomatology at its earliest stage following sensitization to mild symptoms such as a slight almost imperceptible shortness of breath, to loss of pulmonary function, debilitating lung disease, and, in many cases, death. This section also discusses the nature of these illnesses, the scientific evidence that they are causally associated with occupational exposure to beryllium, and the probable mechanisms of action with a more thorough review of the supporting studies.

A. Beryllium and Beryllium Compounds

1. Particle Physical/Chemical Properties

Beryllium (Be; CAS No. 7440-41-7) is a steel-grey, brittle metal with an atomic number of 4 and an atomic weight of 9.01 (Group IIA of the periodic table). Because of its high reactivity, beryllium is not found as a free metal in nature; however, there are approximately 45 mineralized forms of beryllium. Beryllium compounds and alloys include commercially valuable metals and gemstones.

Beryllium has two oxidative states: Be(0) and Be(2⁺) Agency for Toxic Substance and Disease Registry (ATSDR) 2002). It is likely that the Be(2⁺) state is the most biologically reactive and able to form a bond with peptides leading to it becoming antigenic (Snyder et al., 2003). This will be discussed in more detail in the Beryllium Sensitization section below. Beryllium has a high charge-to-radius ratio and in addition to forming various types of ionic bonds, beryllium has a strong tendency for covalent bond formation (e.g., it can form organometallic compounds such as Be(CH₃)₂ and many other complexes) (ATSDR, 2002; Greene et al., 1998). However, it appears that few, if any, toxicity studies exist for the organometallic compounds. Additional physical/chemical properties for beryllium compounds that may be important in their biological response are summarized in Table 1 below. This information was obtained from their International Chemical Safety Cards (ICSC) (beryllium metal (ICSC 0226), beryllium oxide (ICSC 1325), beryllium sulfate (ICSC 1351), beryllium nitrate (ICSC 1352), beryllium carbonate (ICSC 1353), beryllium chloride (ICSC 1354), beryllium fluoride (ICSC 1355)) and from the hazardous substance data bank (HSDB) for beryllium hydroxide (CASRN: 13327-32-7), and beryllium phosphate (CASRN: 13598-15-7). Additional information on chemical and physical properties as well as industrial uses for beryllium can be found in this preamble at Section IV, Chemical Properties and Industrial Uses.

Beryllium shows a high affinity for oxygen in air and water, resulting in a thin surface film of beryllium oxide on the bare metal. If the surface film is disturbed, it may become airborne or dermal exposure may occur. The solubility, particle surface area, and particle size of some beryllium compounds are examined in more detail below. These properties have been evaluated in many toxicological studies. In particular, the properties related to the calcination (firing temperatures) and differences in crystal size and solubility are important aspects in their toxicological profile.

2. Factors Affecting Potency and Effect of Beryllium Exposure

The effect and potency of beryllium and its compounds, as for any toxicant, immunogen, or immunotoxicant, may be dependent upon the physical state in which they are presented to a host. For occupational airborne materials and surface contaminants, it is especially critical to understand those physical parameters in order to determine the extent of exposure to the respiratory tract and skin since these are generally the initial target organs for either route of exposure.

For example, large particles may have less of an effect in the lung than smaller particles due to reduced potential to stay airborne to be inhaled or be deposited along the respiratory tract. In addition, once inhalation occurs particle size is critical in determining where the particle will deposit along the respiratory tract. Solubility also has an important part in determining the toxicity and bioavailability of airborne materials as well. Respiratory tract retention and skin penetration are directly influenced by the solubility and reactivity of airborne material.

These factors may be responsible, at least in part, for the process by which beryllium sensitization progresses to CBD in exposed workers. Other factors influencing beryllium-induced toxicity include the surface area of beryllium particles and their persistence in the lung. With respect to dermal exposure, the physical characteristics of the particle are important as well since they can influence skin absorption and bioavailability. This section addresses certain physical characteristics (i.e., solubility, particle size, particle surface area) that are important in influencing the toxicity of beryllium materials in occupational settings.

a. Solubility

Solubility may be an important determinant of the toxicity of airborne materials, influencing the deposition and persistence of inhaled particles in the respiratory tract, their bioavailability, and the likelihood of presentation to the immune system. A number of chemical agents, including metals that contact and penetrate the skin, are able to induce an immune response, such as sensitization (Boeniger, 2003; Mandervelt et al., 1997). Similar to inhaled agents, the ability of materials to penetrate the skin is also influenced by solubility since dermal absorption may occur at a greater rate for soluble materials than insoluble materials (Kimber et al., 2011).

This section reviews the relevant information regarding solubility, its importance in a biological matrix and its relevance to sensitization and beryllium lung disease. The weight of evidence presented below suggests that both soluble and non-soluble forms of beryllium can induce a sensitization response and result in progression of lung disease.

Beryllium salts, including the chloride (BeCl₂), fluoride (BeF₂), nitrate (Be(NO₃)₂), phosphate (Be₃ (PO₄)₂), and sulfate (tetrahydrate) (BeSO₄ · 4H₂ O) salts, are all water soluble. However, soluble beryllium salts can be converted to less soluble forms in the lung (Reeves and Vorwald, 1967). Aqueous solutions of the soluble beryllium salts are acidic as a result of the formation of Be(OH₂)₄ 2⁺, the tetrahydrate, which will react to form insoluble hydroxides or hydrated complexes within the general physiological range of pH values (between 5 and 8) (EPA, 1998). This may be an important factor in the development of CBD since lower-solubility forms of beryllium have been shown to persist in the lung for longer periods of time and persistence in the lung may be needed in order for this disease to occur (NAS, 2008).

Beryllium oxide (BeO), hydroxide (Be(OH)₂), carbonate (Be₂ CO₃ (OH)₂), and sulfate (anhydrous) (BeSO₄) are either insoluble, slightly soluble, or considered to be sparingly soluble (almost insoluble or having an extremely slow rate of dissolution). The solubility of beryllium oxide, which is prepared from beryllium hydroxide by calcining (heating to a high temperature without fusing in order to drive off volatile chemicals) at temperatures between 500 and 1,750 °C, has an inverse relationship with calcination temperature. Although the solubility of the low-fired crystals can be as much as 10 times that of the high-fired crystals, low-fired beryllium oxide is still only sparingly soluble (Delic, 1992). In a study that measured the dissolution kinetics (rate to dissolve) of beryllium compounds calcined at different temperatures, Hoover et al., compared beryllium metal to beryllium oxide particles and found them to have similar solubilities. This was attributed to a fine layer of beryllium oxide that coats the metal particles (Hoover et al., 1989). A study conducted by Deubner et al., (2011) determined ore materials to be more soluble than beryllium oxide at pH 7.2 but similar in solubility at pH 4.5. Beryllium hydroxide was more soluble than beryllium oxide at both pHs (Deubner et al., 2011).

Investigators have also attempted to determine how biological fluids can dissolve beryllium materials. In two studies, insoluble beryllium, taken up by activated phagocytes, was shown to be ionized by myeloperoxidases (Leonard and Lauwerys, 1987; Lansdown, 1995). The positive charge resulting from ionization enabled the beryllium to bind to receptors on the surface of cells such as lymphocytes or antigen-presenting cells which could make it more biologically active (NAS, 2008). In a study utilizing phagolysosomal-simulating fluid (PSF) with a pH of 4.5, both beryllium metal and beryllium oxide dissolved at a greater rate than that previously reported in water or SUF (simulant fluid) (Stefaniak et al., 2006), and the rate of dissolution of the multi-constituent (mixed) particles was greater than that of the single-constituent beryllium oxide powder. The authors speculated that copper in the particles rapidly dissolves, exposing the small inclusions of beryllium oxide, which have higher specific surface areas (SSA) and therefore dissolve at a higher rate. A follow-up study by the same investigational team (Duling et al., 2012) confirmed dissolution of beryllium oxide by PSF and determined the release rate was biphasic (initial rapid diffusion followed by a latter slower surface reaction-driven release). During the latter phase, dissolution half-times were 1,400 to 2,000 days. The authors speculated this indicated bertrandite was persistent in the lung (Duling et al., 2012).

In a recent study investigating the dissolution and release of beryllium ions for 17 beryllium-containing materials (ore, hydroxide, metal, oxide, alloys, and processing intermediates) using artificial human airway epithelial lining fluid, Stefaniak et al., (2011) found release of beryllium ions within 7 days (beryl ore melter dust). The authors calculated dissolution half-times ranging from 30 days (reduction furnace material) to 74,000 days (hydroxide). Stefaniak et al., (2011) speculated that despite the rapid mechanical clearance, billions of beryllium ions could be released in the respiratory tract via dissolution in airway lining fluid (ALF). Under this scenario beryllium-containing particles depositing in the respiratory tract dissolving in ALF could provide beryllium ions for absorption in the lung and interact with immune cells in the respiratory tract (Stefaniak et al., 2011).

Huang et al., (2011) investigated the effect of simulated lung fluid (SLF) on dissolution and nanoparticle generation and beryllium-containing materials. Bertrandite-containing ore, beryl-containing ore, frit (a processing intermediate), beryllium hydroxide (a processing intermediate) and silica (used as a control), were equilibrated in SLF at two pH values (4.5 and 7.2) to reflect inter- and intra-cellular environments in the lung tissue. Concentrations of beryllium, aluminum, and silica ions increased linearly during the first 20 days in SLF, rose slowly thereafter, reaching equilibrium over time. The study also found nanoparticle formation (in the size range of 10-100 nm) for all materials (Huang et al., 2011).

In an in vitro skin model, Sutton et al., (2003) demonstrated the dissolution of beryllium compounds (insoluble beryllium hydroxide, soluble beryllium phosphate) in a simulated sweat fluid. This model showed beryllium can be dissolved in biological fluids and be available for cellular uptake in the skin. Duling et al., (2012) confirmed dissolution and release of ions from bertrandite ore in an artificial sweat model (pH 5.3 and pH 6.5).

b. Particle Size

The toxicity of beryllium as exemplified by beryllium oxide also is dependent, in part, on the particle size, with smaller particles (<10 μm) able to penetrate beyond the larynx (Stefaniak et al., 2008). Most inhalation studies and occupational exposures involve quite small (<1-2 μm) beryllium oxide particles that can penetrate to the pulmonary regions of the lung (Stefaniak et al., 2008). In inhalation studies with beryllium ores, particle sizes are generally much larger, with deposition occurring in several areas throughout the respiratory tract for particles <10 μm.

The temperature at which beryllium oxide is calcined influences its particle size, surface area, solubility, and ultimately its toxicity (Delic, 1992). Low-fired (500 °C) beryllium oxide is predominantly made up of poorly crystallized small particles, while higher firing temperatures (1000—1750 °C) result in larger particle sizes (Delic, 1992).

In order to determine the extent to which particle size plays a role in the toxicity of beryllium in occupational settings, several key studies are reviewed and detailed below. The findings on particle size have been related, where possible, to work process and biologically relevant toxicity endpoints of either sensitization or CBD.

Numerous studies have been conducted evaluating the particle size generated during basic industrial and machining operations. In a study by Cohen et al., (1983), a multi-cyclone sampler was utilized to measure the size mass distribution of the beryllium aerosol at a beryllium-copper alloy casting operation. Briefly, Cohen et al., (1983) found variable particle size generation based on the operations being sampled with particle size ranging from 3 to 16 μm. Hoover et al., (1990) also found variable particle sizes being generated based on operations. In general, Hoover et al., (1990) found that milling operations generated smaller particle sizes than sawing operations. Hoover et al., (1990) also found that beryllium metal generated higher concentrations than metal alloys. Martyny et al., (2000) characterized generation of particle size during precision beryllium machining processes. The study found that more than 50 percent of the beryllium machining particles collected in the breathing zone of machinists were less than 10 μm in aerodynamic diameter with 30 percent of that fraction being particles of less than 0.6 μm. A study by Thorat et al., (2003) found similar results with ore mixing, crushing, powder production and machining ranging from 5.0 to 9.5 μm. Kent et al., (2001) measured airborne beryllium using size-selective samplers in five furnace areas at a beryllium processing facility. A statistically significant linear trend was reported between the above alveolar-deposited particle mass concentration and prevalence of CBD and sensitization in the furnace production areas. The study authors suggested that the concentration of alveolar-deposited particles (e.g., <3.5 μm) may be a better predictor of sensitization and CBD than the total mass concentration of airborne beryllium.

A recent study by Virji et al. (2011) evaluated particle size distribution, chemistry and solubility in areas with historically elevated risk of sensitization and CBD at a beryllium metal powder, beryllium oxide, and alloy production facility. The investigators observed that historically, exposure-response relationships have been inconsistent when using mass concentration to identify process-related risk, possibly due to incomplete particle characterization. Two separate exposure surveys were conducted in March 1999 and June-August 1999 using multi-stage personal impactor samplers (to determine particle size distribution) and personal 37 mm closed face cassette (CFC) samplers, both located in workers' breathing zones. One hundred and ninety eight time-weighted-average (TWA) personal impactor samples were analyzed for representative jobs and processes. A total of 4,026 CFC samples were collected over the 5-month collection period and analyzed for mass concentration, particle size, chemical content and solubility and compared to process areas with high risk of sensitization and CBD. The investigators found that total beryllium concentration varied greatly between workers and among process areas. Analysis of chemical form and solubility also revealed wide variability among process areas, but high risk process areas had exposures to both soluble and insoluble forms of beryllium. Analysis of particle size revealed most process areas had particles ranging from 5-14 µm mass median aerodynamic diameter (MMAD). Rank order correlating jobs to particle size showed high overall consistency (Spearman r=0.84) but moderate correlation (Pearson r=0.43). The investigators concluded that consideration of relevant aspects of exposure such as particle size distribution, chemical form, and solubility will likely improve exposure assessments (Virji et al., 2011)

c. Particle Surface Area.

Particle surface area has been postulated as an important metric for beryllium exposure. Several studies have demonstrated a relationship between the inflammatory and tumorigenic potential of ultrafine particles and their increased surface area (Driscoll, 1996; Miller, 1995; Oberdorster et al., 1996). While the exact mechanism explaining how particle surface area influences its biological activity is not known, a greater particle surface area has been shown to increase inflammation, cytokine production, anti-oxidant defenses and apoptosis (Elder et al., 2005; Carter et al., 2006; Refsne et al., 2006).

Finch et al., (1988) found that beryllium oxide calcined at 500 °C had 3.3 times greater specific surface area (SSA) than beryllium oxide calcined at 1000 °C, although there was no difference in size or structure of the particles as a function of calcining temperature. The beryllium-metal aerosol (airborne beryllium particles), although similar to the beryllium oxide aerosols in aerodynamic size, had an SSA about 30 percent that of the beryllium oxide calcined at 1000 °C. As discussed above, a later study by Delic (1992) found calcining temperatures had an effect on SSA as well as particle size.

Several studies have investigated the lung toxicity of beryllium oxide calcined at different temperatures and generally had found that those calcined at lower temperatures have greater toxicity and effect than materials calcined at higher temperatures. This may be because beryllium oxide fired at the lower temperature has a loosely formed crystalline structure with greater specific surface area than the fused crystal structure of beryllium oxide fired at the higher temperature. For example, beryllium oxide calcined at 500 °C has been found to have stronger pathogenic effects than material calcined at 1,000 °C, as shown in several of the beagle dog, rat, mouse and guinea pig studies discussed in the section on CBD pathogenesis that follows (Finch et al., 1988; Polak et al., 1968; Haley et al., 1989; Haley et al., 1992; Hall et al., 1950). Finch et al. have also observed higher toxicity of beryllium oxide calcined at 500 °C, an observation they attribute to the greater surface area of beryllium particles calcined at the lower temperature (Finch et al., 1988). These authors found that the in vitro cytotoxicity to Chinese hamster ovary (CHO) cells and cultured lung epithelial cells of 500 °C beryllium oxide was greater than that of 1,000 °C beryllium oxide, which in turn was greater than that of beryllium metal. However, when toxicity was expressed in terms of particle surface area, the cytotoxicity of all three forms was similar. Similar results were observed in a study comparing the cytotoxicity of beryllium metal particles of various sizes to cultured rat alveolar macrophages, although specific surface area did not entirely predict cytotoxicity (Finch et al., 1991).

Stefaniak et al., (2003b) investigated the particle structure and surface area of particles (powder and process-sampled) of beryllium metal, beryllium oxide, and copper-beryllium alloy. Each of these samples was separated by aerodynamic size, and their chemical compositions and structures were determined with x-ray diffraction and transmission electron microscopy, respectively. In summary, beryllium-metal powder varied remarkably from beryllium oxide powder and alloy particles. The metal powder consisted of compact particles, in which SSA decreases with increasing surface diameter. In contrast, the alloys and oxides consisted of small primary particles in clusters, in which the SSA remains fairly constant with particle size. SSA for the metal powders varied based on production and manufacturing process with variations among samples as high as a factor of 37. Stefaniak et al. (2003b) found lesser variation in SSA for the alloys or oxides. This is consistent with data from other studies summarized above showing that process may affect particle size and surface area. Particle size and/or surface area may explain differences in the rate of BeS and CBD observed in some epidemiological studies. However, these properties have not been consistently characterized in most studies.

B. Kinetics and Metabolism of Beryllium

Beryllium enters the body by inhalation, ingestion, or absorption through the skin. For occupational exposure, the airways and the skin are the primary routes of uptake.

1. Exposure via the Respiratory System

The respiratory tract, especially the lung, is the primary target of inhalation exposure in workers. Inhaled beryllium particles are deposited along the respiratory tract in a size dependent manner. In general, particles larger than 10 μm tend to deposit in the upper respiratory tract or nasal region and do not appreciably penetrate lower in the tracheobronchial or pulmonary regions (Figure 1). Particles less than 10 μm increasingly penetrate and deposit in the tracheobronchial and pulmonary regions with peak deposition in the pulmonary region occurring below 5 μm in particle diameter. The CBD pathology of concern is found in the pulmonary region. For particles below 1 μm, regional deposition changes dramatically. Ultrafine particles (generally considered to be 100 nm or lower) have a higher rate of deposition along the entire respiratory system (ICRP model, 1994). Those particles depositing in the lung and along the entire respiratory tract may encounter immunologic cells or may move into the vascular system where they are free to leave the lung and can contribute to systemic beryllium concentrations.

Beryllium is removed from the respiratory tract by various clearance mechanisms. Soluble beryllium is removed from the respiratory tract via absorption. Sparingly soluble or insoluble beryllium may remain in the lungs for many years after exposure, as has been observed in workers (Schepers, 1962). Clearance mechanisms for sparingly soluble or insoluble beryllium particles include: In the nasal passage, sneezing, mucociliary transport to the throat, or dissolution; in the tracheobronchial region, mucociliary transport, coughing, phagocytosis, or dissolution; in the pulmonary or alveolar region, phagocytosis, movement through the interstitium (translocation), or dissolution (Schlesinger, 1997).

Clearance mechanisms may occur slowly in humans, which is consistent with some animal studies. For example, subjects in the Beryllium Case Registry (BCR), which identifies and tracks cases of acute and chronic beryllium diseases, had elevated concentrations of beryllium in lung tissue (e.g., 3.1 μg/g of dried lung tissue and 8.5 μg/g in a mediastinal node) more than 20 years after termination of short-term (generally between 2 and 5 years) occupational exposure to beryllium (Sprince et al., 1976).

Clearance rates may depend on the solubility, dose, and size of the beryllium particles inhaled as well as the sex and species of the animal tested. As reviewed in a WHO Report (2001), more soluble beryllium compounds generally tend to be cleared from the respiratory system and absorbed into the bloodstream more rapidly than less soluble compounds (Van Cleave and Kaylor, 1955; Hart et al., 1980; Finch et al., 1990). Animal inhalation or intratracheal instillation studies administering soluble beryllium salts demonstrated significant absorption of approximately 20 percent of the initial lung burden, while sparingly soluble compounds such as beryllium oxide demonstrated that absorption was slower and less significant (Delic, 1992). Additional animal studies have demonstrated that clearance of soluble and sparingly soluble beryllium compounds was biphasic: A more rapid initial mucociliary transport phase of particles from the tracheobronchial tree to the gastrointestinal tract, followed by a slower phase via translocation to tracheobronchial lymph nodes, alveolar macrophages uptake, and beryllium particles dissolution (Camner et al., 1977; Sanders et al., 1978; Delic, 1992; WHO, 2001). Confirmatory studies in rats have shown the half-time for the rapid phase between 1-60 days, while the slow phase ranged from 0.6-2.3 years. It was also shown that this process was influenced by the solubility of the beryllium compounds: Weeks/months for soluble compounds, months/years for sparingly soluble compounds (Reeves and Vorwald, 1967; Reeves et al., 1967; Zorn et al., 1977; Rhoads and Sanders, 1985). Studies in guinea-pigs and rats indicate that 40-50 percent of the inhaled soluble beryllium salts are retained in the respiratory tract. Similar data could not be found for the sparingly or less soluble beryllium compounds or metal administered by this exposure route. (WHO, 2001; ATSDR, 2002).

Evidence from animal studies suggests that greater amounts of beryllium deposited in the lung may result in slower clearance times. A comparative study of rats and mice using a single dose of inhaled aerosolized beryllium metal demonstrated that an acute inhalation exposure to beryllium metal can slow particle clearance and induce lung damage in rats (Haley et al., 1990) and mice (Finch et al., 1998a). In another study Finch et al. (1994) exposed male F344/N rats to beryllium metal at concentrations resulting in beryllium lung burdens of 1.8, 10, and 100 µg. These exposure levels resulted in an estimated clearance half-life ranging from 250-380 days for the three concentrations. For mice (Finch et al., 1998a), lung clearance half-lives were 91-150 days (for 1.7- and 2.6-μg lung burden groups) or 360-400 days (for 12- and 34-μg lung burden groups). While the lower exposure groups were quite different for rats and mice, the highest groups were similar in clearance half-lives for both species.

Beryllium absorbed from the respiratory system is mainly distributed to the tracheobronchial lymph nodes via the lymph system, bloodstream, and skeleton, which is the ultimate site of beryllium storage (Stokinger et al., 1953; Clary et al., 1975; Sanders et al., 1975; Finch et al., 1990). Trace amounts are distributed throughout the body (Zorn et al., 1977; WHO, 2001). Studies in rats have demonstrated accumulation of beryllium chloride in the skeletal system following intraperitoneal injection (Crowley et al., 1949; Scott et al., 1950) and accumulation of beryllium phosphate and beryllium sulfate in both nonparenchymal and parenchymal cells of the liver after intravenous administration in rats (Skilleter and Price, 1978). Studies have also demonstrated intracellular accumulation of beryllium oxide in bone marrow throughout the skeletal system after intravenous administration to rabbits (Fodor, 1977; WHO, 2001).

Systemic distribution of the more soluble compounds appears to be greater than that of the insoluble compounds (Stokinger et al., 1953). Distribution has also been shown to be dose dependent in research using intravenous administration of beryllium in rats; small doses were preferentially taken up in the skeleton, while higher doses were initially distributed preferentially to the liver. Beryllium was later mobilized from the liver and transferred to the skeleton (IARC, 1993). A half-life of 450 days has been estimated for beryllium in the human skeleton (ICRP, 1960). This indicates the skeleton may serve as a repository for beryllium that may later be reabsorbed by the circulatory system, making beryllium available to the immunological system.

2. Dermal Exposure

Beryllium compounds have been shown to cause skin irritation and sensitization in humans and certain animal models (Van Orstrand et al., 1945; de Nardi et al., 1953; Nishimura 1966; Epstein 1990; Belman, 1969; Tinkle et al., 2003; Delic, 1992). The Agency for Toxic Substances and Disease Registry (ATSDR) estimated that less than 0.1 percent of beryllium compounds are absorbed through the skin (ATSDR, 2002). However, even minute contact and absorption across the skin may directly elicit an immunological sensitization response (Deubner et al., 2001; Toledo et al., 2011). Recent studies by Tinkle et al. (2003) showed that penetration of beryllium oxide particles was possible ex vivo for human intact skin at particle sizes of ≤ 1μm, as confirmed by scanning electron microscopy. Using confocal microscopy, Tinkle et al. demonstrated that surrogate fluorescent particles up to 1 μm in size could penetrate the mouse epidermis and dermis layers in a model designed to mimic the flexing and stretching of human skin in motion. Other poorly soluble particles, such as titanium dioxide, have been shown to penetrate normal human skin (Tan et al., 1996) suggesting the flexing and stretching motion as a plausible mechanism for dermal penetration of beryllium as well. As earlier summarized, insoluble forms of beryllium can be solubilized in biological fluids (e.g., sweat) making them available for absorption through intact skin (Sutton et al., 2003; Stefaniak et al., 2011; Duling et al., 2012).

Although its precise role remains to be elucidated, there is evidence to indicate that dermal exposure can contribute to beryllium sensitization. As early as the 1940s it was recognized that dermatitis experienced by workers in primary beryllium production facilities was linked to exposures to the soluble beryllium salts. Except in cases of wound contamination, dermatitis was rare in workers whose exposures were restricted to exposure to poorly soluble beryllium-containing particles (Van Ordstrand et al., 1945). Further investigation by McCord in 1951 indicated that direct skin contact with soluble beryllium compounds, but not beryllium hydroxide or beryllium metal, caused dermal lesions (reddened, elevated, or fluid-filled lesions on exposed body surfaces) in susceptible persons. Curtis, in 1951, demonstrated skin sensitization to beryllium with patch testing using soluble and insoluble forms of beryllium in beryllium-naïve subjects. These subjects later developed granulomatous skin lesions with the classical delayed-type contact dermatitis following repeat challenge (Curtis, 1951). These lesions appeared after a latent period of 1-2 weeks, suggesting a delayed allergic reaction. The dermal reaction occurred more rapidly and in response to smaller amounts of beryllium in those individuals previously sensitized (Van Ordstrand et al., 1945). Contamination of cuts and scrapes with beryllium can result in the beryllium becoming embedded within the skin causing a granuloma to develop in the skin (Epstein, 1991). Introduction of soluble or insoluble beryllium compounds into or under the skin as a result of abrasions or cuts at work has been shown to result in chronic ulcerations with granuloma formation (Van Orstrand et al., 1945; Lederer and Savage, 1954). Beryllium absorption through bruises and cuts has been demonstrated as well (Rossman et al., 1991). In a study by Invannikov et al., (1982), beryllium chloride was applied directly to the skin of live animals with three types of wounds: abrasions (superficial skin trauma), cuts (skin and superficial muscle trauma), and penetration wounds (deep muscle trauma). The percentage of the applied dose absorbed into the systemic circulation during a 24-hour exposure was significant, ranging from 7.8 percent to 11.4 percent for abrasions, from 18.3 percent to 22.9 percent for cuts, and from 34 percent to 38.8 percent for penetration wounds (WHO, 2001).

A study by Deubner et al., (2001) concluded that exposure across damaged skin can contribute as much systemic loading of beryllium as inhalation (Deubner et al., 2001). Deubner et al., (2001) estimated dermal loading (amount of particles penetrating into the skin) in workers as compared to inhalation exposure. Deubner's calculations assumed a dermal loading rate for beryllium on skin of 0.43 μg/cm², based on the studies of loading on skin after workers cleaned up (Sanderson et al., 1999), multiplied by a factor of 10 to approximate the workplace concentrations and the very low absorption rate of 0.001 percent (taken from EPA estimates). It should be noted that these calculations did not take into account absorption of soluble beryllium salts that might occur across nasal mucus membranes, which may result from contact between contaminated skin and the nose (EPA, 1998).

A study conducted by Day et al. (2007) evaluated the effectiveness of a dermal protection program implemented in a beryllium alloy facility in 2002. The investigators evaluated levels of beryllium in air, on workplace surfaces, on cotton gloves worn over nitrile gloves, and on the necks and faces of workers over a six day period. The investigators found a good correlation between air samples and work surface contamination at this facility. The investigators also found measurable levels of beryllium on the skin of workers as a result of work processes even from workplace areas promoted as “visually clean” by the company housekeeping policy. Importantly, the investigators found that the beryllium contamination could be transferred from body region to body region (e.g., hand to face, neck to face). The investigators demonstrated multiple pathways of exposure which could lead to sensitization, increasing risk for developing CBD (Day, et al., 2007).

The same group of investigators (Armstrong et al., 2014) extended their work on investigating multiple exposure pathways contributing to sensitization and CBD. The investigators evaluated four different beryllium manufacturing and processing facilities to assess the contribution of various exposure pathways on worker exposure. Airborne, work surface and cotton glove beryllium concentrations were evaluated. The investigators found strong correlations between air-surface concentrations, glove-surface concentrations, and air-glove concentrations at this facility. This work confirms findings from Day et al. (2007) demonstrating the importance of airborne beryllium concentrations to surface contamination and dermal exposure even at exposures below the current OSHA PEL (Armstrong et al., 2014).

3. Oral and Gastrointestinal Exposure

According to the WHO Report (2001), gastrointestinal absorption of beryllium can occur by both the inhalation and oral routes of exposure. Through inhalation exposure, a fraction of the inhaled material is transported to the gastrointestinal tract by the mucociliary escalator or by the swallowing of the insoluble material deposited in the upper respiratory tract (WHO, 2001). Gastrointestinal absorption of beryllium can occur by both the inhalation and oral routes of exposure. In the case of inhalation, a portion of the inhaled material is transported to the gastrointestinal tract by the mucociliary escalator or by the swallowing of the insoluble material deposited in the upper respiratory tract (Schlesinger, 1997). Animal studies have shown oral administration of beryllium compounds to result in very limited absorption and storage (as reviewed by U.S. EPA, 1998). In animal ingestion studies using radio-labeled beryllium chloride in rats, mice, dogs, and monkeys, the vast majority of the ingested dose passed through the gastrointestinal tract unabsorbed and was excreted in the feces. In most studies, <1 percent of the administered radioactivity was absorbed into the bloodstream and subsequently excreted in the urine (Crowley et al., 1949; Furchner et al., 1973; LeFevre and Joel, 1986). Research using soluble beryllium sulfate has shown that as the compound passes into the intestine, which has a higher pH than the stomach (approximate pH of 6 to 8 for the intestine, pH of 1 or 2 for the stomach), the beryllium is precipitated as the insoluble phosphate and thus is no longer available for absorption (Reeves, 1965; WHO, 2001).

Urinary excretion of beryllium has been shown to correlate with the amount of occupational exposure (Klemperer et al., 1951). Beryllium that is absorbed into the bloodstream is excreted primarily in the urine (Crowley et al., 1949; Scott et al., 1950; Furchner et al., 1973; Stiefel et al., 1980), whereas excretion of unabsorbed beryllium is primarily via the fecal route (Hart et al., 1980; Finch et al., 1990). A far higher percentage of the beryllium administered parenterally in various animal species was eliminated in the urine than in the feces (Crowley et al., 1949; Scott et al., 1950; Furchner et al., 1973), confirming that beryllium found in the feces following oral exposure is primarily unabsorbed material. A study using percutaneous incorporation of soluble beryllium nitrate in rats similarly demonstrated that more than 90 percent of the beryllium in the bloodstream was eliminated via urine (Zorn et al., 1977; WHO, 2001). More than 99 percent of ingested beryllium chloride was excreted in the feces (Mullen et al., 1972). Elimination half-times of 890-1,770 days (2.4-4.8 years) were calculated for mice, rats, monkeys, and dogs injected intravenously with beryllium chloride (Furchner et al., 1973). Mean daily excretion of beryllium metal was 4.6 × 10⁻⁵ percent of the dose administered by intratracheal instillation in baboons and 3.1 × 10⁻⁵ percent in rats (Andre et al., 1987).

4. Metabolism

Beryllium and its compounds are not metabolized or biotransformed, but soluble beryllium salts may be converted to less soluble forms in the lung (Reeves and Vorwald, 1967). As stated earlier, solubility is an important factor for persistence of beryllium in the lung. Insoluble beryllium, engulfed by activated phagocytes, can be ionized by an acidic environment and by myeloperoxidases (Leonard and Lauwerys, 1987; Lansdown, 1995; WHO, 2001), and this positive charge could potentially make it more biologically reactive because it may allow the beryllium to bind to a peptide or protein and be presented to the T cell receptor or antigen-presenting cell (Fontenot, 2000).

5. Preliminary Conclusion for Particle Characterization and Kinetics of Beryllium

The forms and concentrations of beryllium across the workplace vary substantially based upon location, process, production and work task. Many factors influence the potency of beryllium including concentration, composition, structure, size and surface area of the particle.

Studies have demonstrated that beryllium sensitization can occur via the skin or inhalation from soluble or poorly soluble beryllium particles. Beryllium must be presented to a cell in a soluble form for activation of the immune system (NAS, 2008), and this will be discussed in more detail in the section to follow. Poorly soluble beryllium can be solubilized via intracellular fluid, lung fluid and sweat (Sutton et al., 2003; Stefaniak et al., 2011). For beryllium to persist in the lung it needs to be insoluble. However, soluble beryllium has been shown to precipitate in the lung to form insoluble beryllium (Reeves and Vorwald, 1967).

Some animal and epidemiological studies suggest that the form of beryllium may affect the rate of development of BeS and CBD. Beryllium in an inhalable form (either as soluble or insoluble particles or mist) can deposit in the respiratory tract and interact with immune cells located along the entire respiratory tract (Scheslinger, 1997). However, more study is needed to precisely determine the physiochemical characteristics of beryllium that influence toxicity and immunogenicity.

C. Acute Beryllium Diseases

Acute beryllium disease (ABD) is a relatively rapid onset inflammatory reaction resulting from breathing high airborne concentrations of beryllium. It was first reported in workers extracting beryllium oxide (Van Ordstrand et al., 1943). Since the Atomic Energy Commission's adoption of occupational exposure limits for beryllium beginning in 1949, cases of ABD have been rare. According to the World Health Organization (2001), ABD is generally associated with exposure to beryllium levels at or above 100 μg/m³ and may be fatal in 10 percent of cases. However, cases have been reported with beryllium exposures below 100 µg/m³ (Cummings et al., 2009). The disease involves an inflammatory reaction that may include the entire respiratory tract, involving the nasal passages, pharynx, bronchial airways and alveoli. Other tissues including skin and conjunctivae may be affected as well. The clinical features of ABD include a nonproductive cough, chest pain, cyanosis, shortness of breath, low-grade fever and a sharp drop in functional parameters of the lungs. Pathological features of ABD include edematous distension, round cell infiltration of the septa, proteinaceous materials, and desquamated alveolar cells in the lung. Monocytes, lymphocytes and plasma cells within the alveoli are also characteristic of the acute disease process (Freiman and Hardy, 1970).

Two types of acute beryllium disease have been characterized in the literature: a rapid and severe course of acute fulminating pneumonitis generally developing within 48 to 72 hours of a massive exposure, and a second form that takes several days to develop from exposure to lower concentrations of beryllium (still above the levels set by regulatory and guidance agencies) (Hall, 1950; DeNardi et al., 1953; Newman and Kreiss, 1992). Evidence of a dose-response relationship to the concentration of beryllium is limited (Eisenbud et al., 1948; Stokinger, 1950; Sterner and Eisenbud, 1951). Recovery from either type of ABD is generally complete after a period of several weeks or months (DeNardi et al., 1953). However, deaths have been reported in more severe cases (Freiman and Hardy, 1970). There have been documented cases of progression to CBD (ACCP, 1965; Hall, 1950) suggesting the possibility of an immune component to this disease (Cummings et al., 2009) as well. According to the BCR, in the United States, approximately 17 percent of ABD patients developed CBD (BCR, 2010). The majority of ABD cases occurred between 1932 and 1970 (Eisenbud, 1983; Middleton, 1998). ABD is extremely rare in the workplace today due to more stringent exposure controls implemented following occupational and environmental standards set in 1970-1972 (OSHA, 1971; ACGIH, 1971; ANSI, 1970) and 1974 (EPA, 1974).

D. Chronic Beryllium Disease

This section provides an overview of the immunology and pathogenesis of BeS and CBD, with particular attention to the role of skin sensitization, particle size, beryllium compound solubility, and genetic variability in individuals' susceptibility to beryllium sensitization and CBD.

Chronic beryllium disease (CBD), formerly known as “berylliosis” or “chronic berylliosis,” is a granulomatous disorder primarily affecting the lungs. CBD was first described in the literature by Hardy and Tabershaw (1946) as a chronic granulomatous pneumonitis. It was proposed as early as 1951 that CBD could be a chronic disease resulting from an immune sensitization to beryllium (Sterner and Eisenbud, 1951; Curtis, 1959; Nishimura, 1966). However, for a time, there remained some controversy as to whether CBD was a delayed-onset hypersensitivity disease or a toxicant-induced disease (NAS, 2008). Wide acceptance of CBD as a hypersensitivity lung disease did not occur until bronchoscopy studies and bronchoalveolar lavage (BAL) studies were performed demonstrating that BAL cells from CBD patients responded to beryllium challenge (Epstein et al., 1982; Rossman et al., 1988; Saltini et al., 1989).

CBD shares many clinical and histopathological features with pulmonary sarcoidosis, a granulomatous lung disease of unknown etiology. This includes such debilitating effects as airway obstruction, diminishment of physical capacity associated with reduced lung function, possible depression associated with decreased physical capacity, and decreased life expectancy. Without appropriate information, CBD may be difficult to distinguish from sarcoidosis. It is estimated that up to 6 percent of all patients diagnosed with sarcoidosis may actually have CBD (Fireman et al., 2003; Rossman and Kreiber, 2003). Among patients diagnosed with sarcoidosis in which beryllium exposure can be confirmed, as many as 40 percent may actually have CBD (Muller-Quernheim et al., 2006; Cherry et al., 2015).

Clinical signs and symptoms of CBD may include, but are not limited to, a simple cough, shortness of breath or dypsnea, fever, weight loss or anorexia, skin lesions, clubbing of fingers, cyanosis, night sweats, cor pulmonale, tachycardia, edema, chest pain and arthralgia. Changes or loss of pulmonary function also occur with CBD such as decrease in vital capacity, reduced diffusing capacity, and restrictive breathing patterns. The signs and symptoms of CBD constitute a continuum of symptoms that are progressive in nature with no clear demarcation between any stages in the disease (Rossman, 1996; NAS, 2008). Besides these listed symptoms from CBD patients, there have been reported cases of CBD that remained asymptomatic (Muller-Querheim, 2005; NAS, 2008).

Unlike ABD, CBD can result from inhalation exposure to beryllium at levels below the current OSHA PEL, can take months to years after initial beryllium exposure before signs and symptoms of CBD occur (Newman 1996, 2005 and 2007; Henneberger, 2001; Seidler et al., 2012; Schuler et al., 2012), and may continue to progress following removal from beryllium exposure (Newman, 2005; Sawyer et al., 2005; Seidler et al., 2012). Patients with CBD can progress to a chronic obstructive lung disorder resulting in loss of quality of life and the potential for decreased life expectancy (Rossman, et al., 1996; Newman et al., 2005). The NAS report (2008) noted the general lack of published studies on progression of CBD from an early asymptomatic stage to functionally significant lung disease (NAS, 2008). The report emphasized that risk factors and time course for clinical disease have not been fully delineated. However, for people now under surveillance, clinical progression from immunological sensitization and early pathological lesions (i.e., granulomatous inflammation) prior to onset of symptoms to symptomatic disease appears to be slow, although more follow-up is needed (NAS, 2008). A study by Newman (1996) emphasized the need for prospective studies to determine the natural history and time course from BeS and asymptomatic CBD to full-blown disease (Newman, 1996). Drawing from his own clinical experience, Newman was able to identify the sequence of events for those with symptomatic disease as follows: Initial determination of beryllium sensitization; gradual emergence of chronic inflammation of the lung; pathologic alterations with measurable physiologic changes (e.g., pulmonary function and gas exchange); progression to a more severe lung disease (with extrapulmonary effects such as clubbing and cor pulmonale in some cases); and finally death in some cases (reported between 5.8 to 38 percent) (NAS, 2008; Newman, 1996).

In contrast to some occupationally related lung diseases, the early detection of chronic beryllium disease may be useful since treatment of this condition can lead not only to regression of the signs and symptoms, but also may prevent further progression of the disease in certain individuals (Marchand-Adam, 2008; NAS, 2008). The management of CBD is based on the hypothesis that suppression of the hypersensitivity reaction (i.e., granulomatous process) will prevent the development of fibrosis. However, once fibrosis has developed, therapy cannot reverse the damage.

To date, there have been no controlled studies to determine the optimal treatment for CBD (Rossman, 1996; NAS 2008; Sood, 2009). Management of CBD is generally modeled after sarcoidosis treatment. Oral corticosteroid treatment can be initiated in patients with evidence of disease (either by bronchoscopy or other diagnostic measures before progression of disease or after clinical signs of pulmonary deterioration occur). This includes treatment with other anti-inflammatory agents (NAS, 2008; Maier et al., 2012; Salvator et al., 2013) as well. It should be noted, however, that treatment with corticosteroids has side-effects of their own that need to be measured against the possibility of progression of disease (Gibson et al., 1996; Zaki et al., 1987). Alternative treatments such as azathiopurine and infliximab, while successful at treating symptoms of CBD, have been demonstrated to have side-effects as well (Pallavicino et al., 2013; Freeman, 2012).

1. Development of Beryllium Sensitization

Sensitization to beryllium is an essential step for worker development of CBD. Sensitization to beryllium can result from inhalation exposure to beryllium (Newman et al., 2005; NAS, 2008), as well as from skin exposure to beryllium (Curtis, 1951; Newman et al., 1996; Tinkle et al., 2003). Sensitization is currently detected using a laboratory blood test described in Appendix A. Although there may be no clinical symptoms associated with BeS, a sensitized worker's immune system has been activated to react to beryllium exposures such that subsequent exposure to beryllium can progress to serious lung disease (Kreiss et al., 1996; Kreiss et al., 1997; Kelleher et al., 2001; and Rossman, 2001). Since the pathogenesis of CBD involves a beryllium-specific, cell-mediated immune response, CBD cannot occur in the absence of sensitization (NAS, 2008). Various factors, including genetic susceptibility, have been shown to influence risk of developing sensitization and CBD (NAS 2008) and will be discussed later in this section.

While various mechanisms or pathways may exist for beryllium sensitization, the most plausible mechanisms supported by the best available and most current science are discussed below. Sensitization occurs via the formation of a beryllium-protein complex (an antigen) that causes an immunological response. In some instances, onset of sensitization has been observed in individuals exposed to beryllium for only a few months (Kelleher et al., 2001; Henneberger et al., 2001). This suggests the possibility that relatively brief, short-term beryllium exposures may be sufficient to trigger the immune hypersensitivity reaction. Several studies (Newman et al., 2001; Henneberger et al., 2001; Rossman, 2001; Schuler et al., 2005; Donovan et al., 2007, Schuler et al., 2012) have detected a higher prevalence of sensitization among workers with less than one year of employment compared to some cross-sectional studies which, due to lack of information regarding initial exposure, cannot determine time of sensitization (Kreiss et al., 1996; Kreiss et al., 1997). While only very limited evidence has described humoral changes in certain patients with CBD (Cianciara et al., 1980), clear evidence exists for an immune cell-mediated response, specifically the T-cell (NAS, 2008). Figure 2 delineates the major steps required for progression from beryllium contact to sensitization to CBD.

Beryllium presentation to the immune system is believed to occur either by direct presentation or by antigen processing. It has been postulated that beryllium must be presented to the immune system in an ionic form for cell-mediated immune activation to occur (Kreiss et al., 2007). Some soluble forms of beryllium are readily presented, since the soluble beryllium form disassociates into its ionic components. However, for insoluble forms, dissolution may need to occur. A study by Harmsen et al. (1986) suggested that a sufficient rate of dissolution of small amounts of poorly soluble beryllium compounds might occur in the lungs to allow persistent low-level beryllium presentation to the immune system. Stefaniak et al. (2005 and 2012) reported that insoluble beryllium particles phagocytized by macrophages were dissolved in phagolysomal fluid (Stefaniak et al., 2005; Stefaniak et al., 2012) and that the dissolution rate stimulated by phagolysomal fluid was different for various forms of beryllium (Stefaniak et al., 2006; Duling et al., 2012). Several studies have demonstrated that macrophage uptake of beryllium can induce aberrant apoptotic processes leading to the continued release of beryllium ions which will continually stimulate T-cell activation (Sawyer et al., 2000; Sawyer et al., 2004; Kittle et al., 2002). Antigen processing can be mediated by antigen-presenting cells (APC). These may include macrophages, dendritic cells, or other antigen-presenting cells, although this has not been well defined in most studies (NAS, 2008).

Because of their strong positive charge, beryllium ions have the ability to haptenate and alter the structure of peptides occupying the antigen-binding cleft of major histocompatibility complex (MHC) class II on antigen-presenting cells (APC). The MHC class II antigen-binding molecule for beryllium is the human leukocyte antigen (HLA) with specific alleles (e.g.,HLA-DP, HLA-DR, HLA-DQ) associated with the progression to CBD (NAS, 2008; Yucesoy and Johnson, 2011). Several studies have also demonstrated that the electrostatic charge of HLA may be a factor in binding beryllium (Snyder et al., 2003; Bill et al., 2005; Dai et al., 2010). The strong positive ionic charge of the beryllium ion would have a strong attraction for the negatively charged patches of certain HLA alleles (Snyder et al., 2008; Dai et al., 2010). Alternatively, beryllium oxide has been demonstrated to bind to the MHC class II receptor in a neutral pH. The six carboxylates in the amino acid sequence of the binding pocket provide a stable bond with the Be-O-Be molecule when the pH of the substrate is neutral (Keizer et al., 2005). The direct binding of BeO may eliminate the biological requirement for antigen processing or dissolution of beryllium oxide to activate an immune response.

Next in sequence is the beryllium-MHC-APC complex binding to a T-cell receptor (TCR) on a naïve T-cell which stimulates the proliferation and accumulation of beryllium-specific CD4⁺ (cluster of differentiation 4⁺) T-cells (Saltini et al., 1989 and 1990; Martin et al., 2011) as depicted in Figure 3. Fontenot et al. (1999) demonstrated that diversely different variants of TCR were expressed by CD4⁺ T-cells in peripheral blood cells of CBD patients. However, the CD4⁺ T-cells from the lung were more homologous in expression of TCR variants in CBD patients, suggesting clonal expansion of a subset of T-cells in the lung (Fontenot et al., 1999). This may also indicate a pathogenic potential for subsets of T-cell clones expressing this homologous TCR (NAS, 2008). Fontenot et al. (2006) reported beryllium self-presentation by HLA-DP expressing BAL CD4⁺ T-cells. Self-presentation by BAL T-cells in the lung granuloma may result in activation-induced cell death, which may then lead to oligoclonality of the T-cell population characteristic of CBD (NAS, 2008).

As CD4⁺ T-cells proliferate, clonal expansion of various subsets of the CD4⁺ beryllium specific T-cells occurs (Figure 3). In the peripheral blood, the beryllium-specific CD4⁺ T cells require co-stimulation with a co-stimulant CD28 (cluster of differentiation 28). During the proliferation and differentiation process CD4⁺ T-cells secrete pro-inflammatory cytokines that may influence this process (Sawyer et al., 2004; Kimber et al., 2011).

2. Development of CBD

The continued persistence of residual beryllium in the lung leads to a T-cell maturation process. A large portion of beryllium-specific CD4⁺ T cells were shown to cease expression of CD28 mRNA and protein, indicating these cells no longer required co-stimulation with the CD28 ligand (Fontenot et al., 2003). This change in phenotype correlated with lung inflammation (Fontenot et al., 2003). The CD4⁺ independent cells continued to secrete cytokines necessary for additional recruitment of inflammatory and immunological cells; however, they were less proliferative and less susceptible to cell death compared to the CD28 dependent cells (Fontenot et al., 2005; Mack et al., 2008). These beryllium-specific CD4⁺ independent cells are considered to be mature memory effector cells (Ndejembi et al., 2006; Bian et al., 2005). Repeat exposure to beryllium in the lung resulting in a mature population of T cell development independent of co-stimulation by CD28 and development of a population of T effector memory cells (T_em cells) may be one of the mechanisms that lead to the more severe reactions observed specifically in the lung (Fontenot et al., 2005).

CD4⁺ T cells created in the sensitization process recognize the beryllium antigen, and respond by proliferating and secreting cytokines and inflammatory mediators, including IL-2, IFN-γ, and TNF-α (Tinkle et al., 1997a and b; Fontenot et al., 2002) and MIP-1α and GRO-1 (Hong-Geller, 2006). This also results in the accumulation of various types of inflammatory cells including mononuclear cells (mostly CD4⁺ T cells) in the bronchoalveolar lavage fluid (BAL fluid) (Saltini et al., 1989, 1990).

The development of granulomatous inflammation in the lung of CBD patients has been associated with the accumulation of beryllium responsive CD4⁺ T_em cells in BAL fluid (NAS, 2008). The subsequent release of pro-inflammatory cytokines, chemokines and reactive oxygen species by these cells may lead to migration of additional inflammatory/immune cells and the development of a microenvironment that contributes to the development of CBD (Sawyer et al., 2005; Tinkle et al., 1996; Hong-Geller et al., 2006; NAS, 2008).

The cascade of events described above results in the formation of a noncaseating granulomatous lesion. Release of cytokines by the accumulating T cells leads to the formation of granulomatous lesions that are characterized by an outer ring of histiocytes surrounding non-necrotic tissue with embedded multi-nucleated giant cells (Saltini et al., 1989, 1990).

Over time, the granulomas spread and can lead to lung fibrosis and abnormal pulmonary function, with symptoms including a persistent dry cough and shortness of breath (Saber and Dweik, 2000). Fatigue, night sweats, chest and joint pain, clubbing of fingers (due to impaired oxygen exchange), loss of appetite or unexplained weight loss, and cor pulmonale have been experienced in certain patients as the disease progresses (Conradi et al., 1971; ACCP, 1965; Kriebel et al., 1988a and b). While CBD primarily affects the lungs, it can also involve other organs such as the liver, skin, spleen, and kidneys (ATSDR, 2002).

As previously mentioned, the uptake of beryllium may lead to an aberrant apoptotic process with rerelease of beryllium ions and continual stimulation of beryllium-responsive CD4⁺ cells in the lung (Sawyer et al., 2000; Kittle et al., 2002; Sawyer et al., 2004). Several research studies suggest apoptosis may be one mechanism that enhances inflammatory cell recruitment, cytokine production and inflammation, thus creating a scenario for progressive granulomatous inflammation (Palmer et al., 2008; Rana, 2008). Macrophages and neutrophils can phagocytize beryllium particles in an attempt to remove the beryllium from the lung (Ding, et al., 2009). Multiple studies (Sawyer et al., 2004; Kittle et al., 2002) using BAL cells (mostly macrophages and neutrophils) from patients with CBD found that in vitro stimulation with beryllium sulfate induced the production of TNF-α (one of many cytokines produced in response to beryllium), and that production of TNF-α might induce apoptosis in CBD and sarcoidosis patients (Bost et al., 1994; Dai et al., 1999). The stimulation of CBD-derived macrophages by beryllium sulphate resulted in cells becoming apoptotic, as measured by propidium iodide. These results were confirmed in a mouse macrophage cell-line (p388D1) (Sawyer et al., 2000). However, other factors may influence the development of CBD and are outlined in the following section.

3. Genetic and Other Susceptibility Factors

Evidence from a variety of sources indicates genetic susceptibility may play an important role in the development of CBD in certain individuals, especially at levels low enough not to invoke a response in other individuals. Early occupational studies proposed that CBD was an immune reaction based on the high susceptibility of some individuals to become sensitized and progress to CBD and the lack of CBD in others who were exposed to levels several orders of magnitude higher (Sterner and Eisenbud, 1951). Additional in vitro human research has identified genes coding for specific protein molecules on the surface of their immune cells that place carriers at greater risk of becoming sensitized to beryllium and developing CBD (McCanlies et al., 2004). Recent studies have confirmed genetic susceptibility to CBD involves either HLA variants, T-cell receptor clonality, tumor necrosis factor (TNF-α) polymorphisms and/or transforming growth factor-beta (TGF-β) polymorphisms (Fontenot et al., 2000; Amicosante et al., 2005; Tinkle et al., 1996; Gaede et al., 2005; Van Dyke et al., 2011; Silveira et al., 2012).

Single Nucleotide Polymorphisms (SNPs) have been studied with regard to genetic variations associated with increased risk of developing CBD. SNPs are the most abundant type of human genetic variation. Polymorphisms in MHC class II and pro-inflammatory genes have been shown to contribute to variations in immune responses contributing to the susceptibility and resistance in many diseases including auto-immunity, and beryllium sensitization and CBD (McClesky et al., 2009). Specific SNPs have been evaluated as a factor in Glu69 variant from the HLA-DPB1 locus (Richeldi et al., 1993; Cai et al., 2000; Saltini et al., 2001; Silviera et al., 2012; Dai et al., 2013), HLA-DRPheβ47 (Amicosante et al., 2005).

HLA-DPB1 with a glutamic acid at amino position 69 (Glu 69) has been shown to confer increased risk of beryllium sensitization and CBD (Richeldi et al., 1993; Saltini et al., 2001; Amicosante et al., 2005; Van Dyke et al., 2011; Silveira et al., 2012). Fontenot et al. (2000) demonstrated that beryllium presentation by certain alleles of the class II human leukocyte antigen-DP (HLA-DP) to CD4+ T cells is the mechanism underlying the development of CBD. Richeldi et al. (1993) reported a strong association between the MHC class II allele HLA-DP 1 and the development of CBD in beryllium-exposed workers from a Tucson, AZ facility. This marker was found in 32 of the 33 workers who developed CBD, but in only 14 of 44 similarly exposed workers without CBD. The more common allele of the HLA-DP 1 variant is negatively charged at this site and could directly interact with the positively charged beryllium ion. The high percentage (~30 percent) of beryllium-exposed workers without CBD who had this allele indicates that other factors also contribute to the development of CBD (EPA, 1998). Additional studies by Amicosante et al. (2005) using blood lymphocytes derived from beryllium-exposed workers found a high frequency of this gene in those sensitized to beryllium. In a study of 82 CBD patients (beryllium-exposed workers), Stubbs et al. (1996) also found a relationship between the HLA-DP 1 allele and BeS. The glutamate-69 allele was present in 86 percent of sensitized subjects, but in only 48 percent of beryllium-exposed, non-sensitized subjects. Some variants of the HLA-DPB1 allele convey higher risk of BeS and CBD than others. For example, HLA-DPB1*0201 yielded an approximately 3-fold increase in disease outcome relative to controls; HLA-DPB1*1901 yielded an approximately 5-fold increase, and HLA-DPB1*1701 an approximately 10-fold increase (Weston et al., 2005; Snyder et al., 2008). By assigning odds ratios for specific alleles on the basis of previous studies discussed above, the researchers found a strong correlation (88 percent) between the reported risk of CBD and the predicted surface electrostatic potential and charge of the isotypes of the genes. They were able to conclude that the alleles associated with the most negatively charged proteins carry the greatest risk of developing beryllium sensitization and CBD. This confirms the importance of beryllium charge as a key factor in haptogenic potential.

In contrast, the HLA-DRB1 allele, which lacks Glu 69, has also been shown to increase the risk of developing sensitization and CBD (Amicosante et al., 2005; Maier et al., 2003). Bill et al. (2005) found that HLA-DR has a glutamic acid at position 71 of the β chain, functionally equivalent to the Glu 69 of HLA-DP (Bill et al., 2005). Associations with BeS and CBD have also been reported with the HLA-DQ markers (Amicosante et al., 2005; Maier et al., 2003). Stubbs et al. also found a biased distribution of the MHC class II HLA-DR gene between sensitized and non-sensitized subjects. Neither of these markers was completely specific for CBD, as each study found beryllium sensitization or CBD among individuals without the genetic risk factor. While there remains uncertainty as to which of the MHC class II genes interact directly with the beryllium ion, antibody inhibition data suggest that the HLA-DR gene product may be involved in the presentation of beryllium to T lymphocytes (Amicosante et al., 2002). In addition, antibody blocking experiments revealed that anti-HLA-DP strongly reduced proliferation responses and cytokine secretion by BAL CD4 T cells (Chou et al., 2005). In the study by Chou (2005), anti-HLA-DR ligand antibodies mainly affected beryllium-induced proliferation responses with little impact on cytokines other than IL-2, thus implying that nonproliferating BAL CD4 T cells may still contribute to inflammation leading to the progression of CBD (Chou et al., 2005).

TNF alpha (TNF-α) polymorphisms and TGF beta (TGF-β) polymorphisms have also been shown to confer a genetic susceptibility for developing CBD in certain individuals. TNF-α is a pro-inflammatory cytokine associated with a more severe pulmonary disease in CBD (NAS, 2008). Beryllium exposure has been shown to upregulate transcription factors AP-1 and NF-κB (Sawyer et al., 2007) inducing an inflammatory response by stimulating production of pro-inflammatory cytokines such as TNF-α by inflammatory cells. Polymorphisms in the 308 position of the TNF-α gene have been demonstrated to increase production of the cytokine and increase severity of disease (Maier et al., 2001; Saltini et al., 2001; Dotti et al., 2004). While a study by McCanlies et al. (2007) found no relationship between TNF-α polymorphism and BeS or CBD, the inconsistency may be due to misclassification, exposure differences or statistical power (NAS, 2008).

Other genetic variations have been shown to be associated with increased risk of beryllium sensitization and CBD (NAS, 2008). These include TGF-β (Gaede et al., 2005), angiotensin-1 converting enzyme (ACE) (Newman et al., 1992; Maier et al., 1999) and an enzyme involved in glutathione synthesis (glutamate cysteine ligase) (Bekris et al., 2006). McCanlies et al. (2010) evaluated the association between polymorphisms in a select group of interleukin genes (IL-1A; IL-1B, IL-1RN, IL-2, IL-9, IL-9R) due to their role in immune and inflammatory processes. The study evaluated SNPs in three groups of workers from large beryllium manufacturing facilities in OH and AZ. The investigators found a significant association between variants IL-1A-1142, IL-1A-3769 and IL-1A-4697 and CBD but not with beryllium sensitization. However, these still require confirmation in larger studies (NAS, 2008).

In addition to the genetic factors which may contribute to the susceptibility and severity of disease, other factors such as smoking and gender may play a role in the development of CBD (NAS, 2008). A recent longitudinal cohort study by Mroz et al. (2009) of 229 individuals identified with beryllium sensitization or CBD through workplace medical surveillance found that the prevalence of CBD among ever smokers was significantly lower than among never smokers (38.1 percent versus 49.4 percent, p=0.025). BeS subjects that never smoked were found to be more likely to develop CBD over the course of the study compared to current smokers (12.6 percent versus 6.4 percent, p=0.10). The authors suggested smoking may confer a protective effect against development of lung granulomas as has been demonstrated with hypersensitivity pneumonitis (Mroz et al., 2009).

4. Beryllium Sensitization and CBD in the Workforce

Sensitization to beryllium is currently detected in the workforce with the beryllium lymphocyte proliferation test (BeLPT), a laboratory blood test developed in the 1980s, also referred to as the LTT (Lymphocyte Transformation Test) or BeLT (Beryllium Lymphocyte Transformation Test). In this test, lymphocytes obtained from either bronchoalveolar lavage fluid (the BAL BeLPT) or from peripheral blood (the blood BeLPT) are cultured in vitro and exposed to beryllium sulfate to stimulate lymphocyte proliferation. The observation of beryllium-specific proliferation indicates beryllium sensitization. Hereafter, “BeLPT” generally refers to the blood BeLPT, which is typically used in screening for beryllium sensitization. This test is described in more detail in subsection D.5.b.

CBD can be detected at an asymptomatic stage by a number of techniques including bronchoalveolar lavage and biopsy (Cordeiro et al., 2007; Maier, 2001). Bronchoalveolar lavage is a method of “washing” the lungs with fluid inserted via a flexible fiberoptic instrument known as a bronchoscope, removing the fluid and analyzing the content for the inclusion of immune cells reactive to beryllium exposure, as described earlier in this section. Fiberoptic bronchoscopy can be used to detect granulomatous lung inflammation prior to the onset of CBD symptoms as well, and has been used in combination with the BeLPT to diagnose pre-symptomatic CBD in a number of recent screening studies of beryllium-exposed workers, which are discussed in the following section detailing diagnostic procedures. Of workers who were found to be sensitized and underwent clinical evaluation, 31-49 percent of them were diagnosed with CBD (Kreiss et al., 1993; Newman et al., 1996, 2005, 2007; Mroz, 2009), however some estimate that with increased surveillance the percent could be much higher (Newman, 2005; Mroz, 2009). It has been estimated from ongoing surveillance studies of sensitized individuals with an average follow-up time of 4.5 years that 31 percent of beryllium-sensitized employees were estimated to progress to CBD (Newman et al., 2005). A study of nuclear weapons facility employees enrolled in an ongoing medical surveillance program found that only 20 percent of sensitized workers employed less than 5 years eventually were diagnosed with CBD, while 40 percent of sensitized workers employed 10 years or more developed CBD (Stange et al., 2001). One limitation for all these studies is lack of long-term follow-up. It may be necessary to continue to monitor these workers in order to determine whether all BeS workers will develop CBD (Newman et al., 2005).

CBD has a clinical spectrum ranging from evidence of beryllium sensitization and granulomas in the lung with little symptomatology to loss of lung function and end stage disease which may result in the need for lung transplantation and decreased life expectancy. Unfortunately, there are very few published clinical studies describing the full range and progression of CBD from the beginning to the end stages and very few of the risk factors for progression of disease have been delineated (NAS, 2008). Clinical management of CBD is modeled after sarcoidosis where oral corticosteroid treatment is initiated in patients who have evidence of progressive lung disease, although progressive lung disease has not been well defined (NAS, 2008). In advanced cases of CBD, corticosteroids are the standard treatment (NAS, 2008). No comprehensive studies have been published measuring the overall effect of removal of workers from beryllium exposure on sensitization and CBD (NAS, 2008) although this has been suggested as part of an overall treatment regime for CBD (Mapel et al., 2002; Sood et al., 2004; Maier et al., 2006; Sood, 2009; Maier et al., 2012). Sood et al. reported that cessation of exposure can sometimes have beneficial effects on lung function (Sood et al., 2004). However, this was based on anecdotal evidence from six patients with CBD, so more research is needed to better determine the relationship between exposure duration and disease progression

5. Human Epidemiological Studies

This section describes the human epidemiological data supporting the mechanistic overview of beryllium-induced disease in workers. It has been divided into reviews of epidemiological studies performed prior to development and implementation of the BeLPT in the late 1980s and after wide use of the BeLPT for screening purposes. Use of the BeLPT has allowed investigators to screen for beryllium sensitization and CBD prior to the onset of clinical symptoms, providing a more sensitive and thorough analysis of the worker population. The discussion of the studies has been further divided by manufacturing processes that may have similar exposure profiles. Table A.1 in the Appendix summarizes the prevalence of beryllium sensitization and CBD, range of exposure measurements, and other salient information from the key epidemiological studies.

It has been well-established that beryllium exposure, either via inhalation or skin, may lead to beryllium sensitization, or, with inhalation exposure, may lead to the onset and progression of CBD. The available published epidemiological literature discussed below provides strong evidence of beryllium sensitization and CBD in workers exposed to airborne beryllium well below the current OSHA PEL of 2 μg/m³. Several studies demonstrate the prevalence of sensitization and CBD is related to the level of airborne exposure, including a cross-sectional survey of employees at a beryllium ceramics plant in Tucson, AZ (Henneberger et al., 2001), case-control studies of workers at the Rocky Flats nuclear weapons facility (Viet et al., 2000), and workers from a beryllium machining plant in Cullman, AL (Kelleher et al., 2001). The prevalence of beryllium sensitization also may be related to dermal exposure. An increased risk of CBD has been reported in workers with skin lesions, potentially increasing the uptake of beryllium (Curtis, 1951; Johnson et al., 2001; Schuler et al., 2005). Three studies describe comprehensive preventive programs, which included expanded respiratory protection, dermal protection, and improved control of beryllium dust migration, that substantially reduced the rate of beryllium sensitization among new hires (Cummings et al., 2007; Thomas et al., 2009; Bailey et al., 2010; Schuler et al., 2012).

Some of the epidemiological studies presented in this review suffer from challenges common to many published epidemiological studies: Limitations in study design (particularly cross-sectional); small sample size; lack of personal and/or short-term exposure data, particularly those published before the late 1990s; and incomplete information regarding specific chemical form and/or particle characterization. Challenges that are specific to beryllium epidemiological studies include: uncertainty regarding the contribution of dermal exposure; use of various BeLPT protocols; a variety of case definitions for determining CBD; and use of various exposure sampling/assessment methods (e.g., daily weighted average (DWA), lapel sampling). Even with these limitations, the epidemiological evidence presented in this section clearly demonstrates that beryllium sensitization and CBD are continuing to occur from present-day exposures below OSHA's PEL. The available literature also indicates that the rate of BeS can be substantially lowered by reducing inhalation exposure and minimizing dermal contact.

a. Studies Conducted Prior to the BeLPT

First reports of CBD came from studies performed by Hardy and Tabershaw (1946). Cases were observed in industrial plants that were refining and manufacturing beryllium metal and beryllium alloys and in plants manufacturing fluorescent light bulbs (NAS, 2008). From the late 1940s through the 1960s, clusters of non-occupational CBD cases were identified around beryllium refineries in Ohio and Pennsylvania, and outbreaks in family members of beryllium factory workers were assumed to be from exposure to contaminated clothes (Hardy, 1980). It had been established that the risk of disease among beryllium workers was variable and generally rose with the levels of airborne concentrations (Machle et al., 1948). And while there was a relationship between air concentrations of beryllium and risk of developing disease both in and surrounding these plants, the disease rates outside the plants were higher than expected and not very different from the rate of CBD within the plants (Eisenbud et al., 1949; Lieben and Metzner, 1959). There remained considerable uncertainty regarding diagnosis due to lack of well-defined cohorts, modern diagnostic methods, or inadequate follow-up. In fact, many patients with CBD may have been misdiagnosed with sarcoidosis (NAS, 2008).

The difficulties in distinguishing lung disease caused by beryllium from other lung diseases led to the establishment of the BCR in 1952 to identify and track cases of ABD and CBD. A uniform diagnostic criterion was introduced in 1959 as a way to delineate CBD from sarcoidosis. Patient entry into the BCR required either: documented past exposure to beryllium or the presence of beryllium in lung tissue as well as clinical evidence of beryllium disease (Hardy et al., 1967); or any three of the six criteria listed below (Hasan and Kazemi, 1974). Patients identified using the above criteria were registered and added to the BCR from 1952 through 1983 (Eisenbud and Lisson, 1983).

The BCR listed the following criteria for diagnosing CBD (Eisenbud and Lisson, 1983):

(1) Establishment of significant beryllium exposure based on sound epidemiologic history;

(2) Objective evidence of lower respiratory tract disease and clinical course consistent with beryllium disease;

(3) Chest X-ray films with radiologic evidence of interstitial fibronodular disease;

(4) Evidence of restrictive or obstructive defect with diminished carbon monoxide diffusing capacity (DL_CO) by physiologic studies of lung function;

(5) Pathologic changes consistent with beryllium disease on examination of lung tissue; and

(6) Presence of beryllium in lung tissue or thoracic lymph nodes.

Prevalence of CBD in workers during the time period between the 1940s and 1950s was estimated to be between 1-10% (Eisenbud and Lisson, 1983). In a 1969 study, Stoeckle et al. presented 60 case histories with a selective literature review utilizing the above criteria except that urinary beryllium was substituted for lung beryllium to demonstrate beryllium exposure. Stoeckle et al. (1969) were able to demonstrate corticosteroids as a successful treatment option in one case of confirmed CBD. This study also presented a 28 percent mortality rate from complications of CBD at the time of publication. However, even with the improved methodology for determining CBD based on the BCR criteria, these studies suffered from lack of well-defined cohorts, modern diagnostic techniques or adequate follow-up.

b. Criteria for Beryllium Sensitization and CBD Case Definition Following the Development of the BeLPT

The criteria for diagnosis of CBD have evolved over time as more advanced diagnostic technology, such as the (blood) BeLPT and BAL BeLPT, has become available. More recent diagnostic criteria have both higher specificity than earlier methods and higher sensitivity, identifying subclinical effects. Recent studies typically use the following criteria (Newman et al., 1989; Pappas and Newman, 1993; Maier et al., 1999):

(1) History of beryllium exposure;

(2) Histopathological evidence of noncaseating granulomas or mononuclear cell infiltrates in the absence of infection; and

(3) Positive blood or BAL BeLPT (Newman et al., 1989).

The availability of transbronchial lung biopsy facilitates the evaluation of the second criterion, by making histopathological confirmation possible in almost all cases.

A significant component for the identification of CBD is the demonstration of a confirmed abnormal BeLPT result in a blood or BAL sample (Newman, 1996). Since the development of the BeLPT in the 1980s, it has been used to screen beryllium-exposed workers for sensitization in a number of studies to be discussed below. The BeLPT is a non-invasive in vitro blood test which measures the beryllium antigen-specific T-cell mediated immune response and is the most commonly available diagnostic tool for identifying beryllium sensitization. The BeLPT measures the degree to which beryllium stimulates lymphocyte proliferation under a specific set of conditions, and is interpreted based upon the number of stimulation indices that exceed the normal value. The `cut-off' is based on the mean value of the peak stimulation index among controls plus 2 or 3 standard deviations. This methodology was modeled into a statistical method known as the “least absolute values” or “statistical-biological positive” method and relies on natural log modeling of the median stimulation index values (DOE, 2001; Frome, 2003). In most applications, two or more stimulation indices that exceed the cut-off constitute an abnormal test.

Early versions of the BeLPT test had high variability, but the use of tritiated thymidine to identify proliferating cells has led to a more reliable test (Mroz et al., 1991; Rossman et al., 2001). In recent years, the peripheral blood test has been found to be as sensitive as the BAL assay, although larger abnormal responses have been observed with the BAL assay (Kreiss et al., 1993; Pappas and Newman, 1993). False negative results have also been observed with the BAL BeLPT in cigarette smokers who have marked excess of alveolar macrophages in lavage fluid (Kreiss et al., 1993). The BeLPT has also been a useful tool in animal studies to identify those species with a beryllium-specific immune response (Haley et al., 1994).

Screenings for beryllium sensitization have been conducted using the BeLPT in several occupational surveys and surveillance programs, including nuclear weapons facilities operated by the Department of Energy (Viet et al., 2000; Strange et al., 2001; DOE/HSS Report, 2006), a beryllium ceramics plant in Arizona (Kreiss et al., 1996; Henneberger et al., 2001; Cummings et al., 2007), a beryllium production plant in Ohio (Kreiss et al., 1997; Kent et al., 2001), a beryllium machining facility in Alabama (Kelleher et al., 2001; Madl et al., 2007), a beryllium alloy plant (Schuler et al., 2005, Thomas et al., 2009), and another beryllium processing plant (Rosenman et al., 2005) in Pennsylvania. In most of these studies, individuals with an abnormal BeLPT result were retested and were identified as sensitized (i.e., confirmed positive) if the abnormal result was repeated.

There has been criticism regarding the reliability and specificity of the BeLPT as a screening tool (Borak et al., 2006). Stange et al. (2004) studied the reliability and laboratory variability of the BeLPT by splitting blood samples and sending samples to two laboratories simultaneously for BeLPT analysis. Stange et al. found the range of agreement on abnormal (positive BeLPT) results was 26.2—61.8 percent depending upon the labs tested (Stange et al., 2004). Borak et al. (2006) contended that the positive predictive value (PPV) (PPV is the portion of patients with positive test result correctly diagnosed) is not high enough to meet the criteria of a good screening tool. Middleton et al. (2008) used the data from the Stange et al. (2004) study to estimate the PPV and determined that the PPV of the BeLPT could be improved from 0.383 to 0.968 when an abnormal BeLPT result is confirmed with a second abnormal result (Middleton et al., 2008). However, an apparent false positive can occur in people not occupationally exposed to beryllium (NAS, 2008). An analysis of survey data from the general workforce and new employees at a beryllium manufacturer was performed to assess the reliability of the BeLPT (Donovan et al. 2007). Donovan et al. analyzed more than 10,000 test results from nearly 2400 participants over a 12-year period. Donovan et al. found that approximately 2 percent of new employees had at least one positive BeLPT at the time of hire and 1 percent of new hires with no known occupational exposure were confirmed positive at the time of hire with two BeLPTs. Since there are currently no alternatives to the BeLPT in a screening program many programs rely on a second test to confirm a positive result (NAS, 2008).

The epidemiological studies presented in this section utilized the BeLPT as either a surveillance tool or a screening tool for determining sensitization status and/or sensitization/CBD prevalence in workers for inclusion in the published studies. Most epidemiological studies have reported rates of sensitization and disease based on a single screening of a working population (‘cross-sectional' or 'population prevalence' rates). Studies of workers in a beryllium machining plant and a nuclear weapons facility have included follow-up of the population originally screened, resulting in the detection of additional cases of sensitization over several years (Newman et al., 2001, Stange et al., 2001). OSHA regards the BeLPT as a reliable medical surveillance tool. The BeLPT is discussed in more detail in Non-Mandatory Appendix A to the proposed standard, Immunological Testing for the Determination of Beryllium Sensitization.

c. Beryllium Mining and Extraction

Mining and extraction of beryllium usually involves the two major beryllium minerals, beryl (an aluminosilicate containing up to 4 percent beryllium) and bertrandite (a beryllium silicate hydrate containing generally less than 1 percent beryllium) (WHO, 2001). The United States is the world leader in beryllium extraction and also leads the world in production and use of beryllium and its alloys (WHO, 2001). Most exposures from mining and extraction come in the form of beryllium ore, beryllium salts, beryllium hydroxide (NAS 2008) or beryllium oxide (Stefaniak et al., 2008).

Deubner et al. published a study of 75 workers employed at a beryllium mining and extraction facility in Delta, UT (Deubner et al., 2001b). Of the 75 workers surveyed for sensitization with the BeLPT, three were identified as sensitized by an abnormal BeLPT result. One of those found to be sensitized was diagnosed with CBD. Exposures at the facility included primarily beryllium ore and salts. General area (GA), breathing zone (BZ), and personal lapel (LP) exposure samples were collected from 1970 to 1999. Jobs involving beryllium hydrolysis and wet-grinding activities had the highest air concentrations, with an annual median GA concentration ranging from 0.1 to 0.4 μg/m³. Median BZ concentrations were higher than either LP or GA. The average duration of exposure for beryllium sensitized workers was 21.3 years (27.7 years for the worker with CBD), compared to an average duration for all workers of 14.9 years. However, these exposures were less than either the Elmore, OH, or Tucson, AZ, facilities described below, which also had higher reported rates of BeS and CBD. A study by Stefaniak et al. (2008) demonstrated that beryllium was present at the mill in three forms: mineral, poorly crystalline oxide, and hydroxide.

There was no sensitization or CBD among those who worked only at the mine where exposure to beryllium resulted solely from working with bertrandite ore. The authors concluded that the results of this study indicated that beryllium ore and salts may pose less of a hazard than beryllium metal and beryllium hydroxide. These results are consistent with the previously discussed animal studies examining solubility and particle size.

d. Beryllium Metal Processing and Alloy Production

Kreiss et al. (1997) conducted a study of workers at a beryllium production facility in Elmore, OH. The plant, which opened in 1953 and initially specialized in production of beryllium-copper alloy, later expanded its operations to include beryllium metal, beryllium oxide, and beryllium-aluminum alloy production; beryllium and beryllium alloy machining; and beryllium ceramics production, which was moved to a different factory in the early 1980s. Production operations included a wide variety of jobs and processes, such as work in arc furnaces and furnace rebuilding, alloy melting and casting, beryllium powder processing, and work in the pebble plant. Non-production work included jobs in the analytical laboratory, engineering research and development, maintenance, laundry, production-area management, and office-area administration. While the publication refers to the use of respiratory protection in some areas, such as the pebble plant, the extent of its use across all jobs or time periods was not reported. Use of dermal PPE was not reported.

The authors characterized exposures at the plant using industrial hygiene (IH) samples collected between 1980 and 1993. The exposure samples and the plant's formulas for estimating workers' DWA exposures were used, together with study participants' work histories, to estimate their cumulative and average beryllium exposure levels. Exposure concentrations reflected the high exposures found historically in beryllium production and processing. Short-term BZ measurements had a median of 1.4, with 18.5 percent of samples exceeding OSHA's STEL of 5.0 μg/m³. Particularly high beryllium concentrations were reported in the areas of beryllium powder production, laundry, alloy arc furnace (approximately 40 percent of DWA estimates over 2.0 μg/m³) and furnace rebuild (28.6 percent of short-term BZ samples over the OSHA STEL of 5 μg/m³). LP samples (n = 179), which were available from 1990 to 1992, had a median value of 1 μg/m³.

Of 655 workers employed at the time of the study, 627 underwent BeLPT screening. Blood samples were divided and split between two labs for analysis, with repeat testing for results that were abnormal or indeterminate. Thirty-one workers had an abnormal blood test upon initial testing and at least one of two subsequent tests was classified as sensitized. These workers, together with 19 workers who had an initial abnormal result and one subsequent indeterminate result, were offered clinical evaluation for CBD including the BAL-BeLPT and transbronchial lung biopsy. Nine with an initial abnormal test followed by two subsequent normal tests were not clinically evaluated, although four were found to be sensitized upon retesting in 1995. Of 47 workers who proceeded with evaluation for CBD (3 of the 50 initial workers with abnormal results declined to participate), 24 workers were diagnosed with CBD based on evidence of granulomas on lung biopsy (20 workers) or on other findings consistent with CBD (4 workers) (Kreiss et al., 1997). After including five workers who had been diagnosed prior to the study, a total of 29 (4.6 percent) current workers were found to have CBD. In addition, the plant medical department identified 24 former workers diagnosed with CBD before the study.

Kreiss et al. reported that the highest prevalence of sensitization and CBD occurred among workers employed in beryllium metal production, even though the highest airborne total mass concentrations of beryllium were generally among employees operating the beryllium alloy furnaces in a different area of the plant (Kreiss et al., 1997). Preliminary follow-up investigations of particle size-specific sampling at five furnace sites within the plant determined that the highest respirable (e.g., particles <10 μm in diameter as defined by the authors) and alveolar-deposited (e.g., particles <1 μm in diameter as defined by the authors) beryllium mass and particle number concentrations, as collected by a general area impactor device, were measured at the beryllium metal production furnaces rather than the beryllium alloy furnaces (Kent et al., 2001; McCawley et al., 2001). A statistically significant linear trend was reported between the above alveolar-deposited particle mass concentration and prevalence of CBD and sensitization in the furnace production areas. The authors concluded that alveolar-deposited particles may be a more relevant exposure metric for predicting the incidence of CBD or sensitization than the total mass concentration of airborne beryllium.

Bailey et al. (2010) evaluated the effectiveness of a workplace preventive program in lowering BeS at the beryllium metal, oxide, and alloy production plant studied by Kreiss et al. (1997). The preventive program included use of administrative and PPE controls (e.g., improved training, skin protection and other PPE, half-mask or air-purified respirators, medical surveillance, improved housekeeping standards, clean uniforms) as well as engineering controls (e.g., migration controls, physical separation of administrative offices from production facilities) implemented over the course of five years.

In a cross-sectional/longitudinal hybrid study, Bailey et al. compared rates of sensitization in pre-program workers to those hired after the preventive program began. Pre-program workers were surveyed cross-sectionally in 1993-1994, and again in 1999 using the BeLPT to determine sensitization and CBD prevalence rates. The 1999 cross-sectional survey was conducted to determine if improvements in engineering and administrative controls were successful, however, results indicated no improvement in reducing rates of sensitization or CBD.

An enhanced preventive program including particle migration control, respiratory and dermal protection, and process enclosure was implemented in 2000, with continuing improvements made to the program in 2001, 2002-2004, and 2005. Workers hired during this period were longitudinally surveyed for sensitization using the BeLPT. Both the pre-program and program survey of worker sensitization status utilized split-sample testing to verify positive test results using the BeLPT. Of the total 660 workers employed at the production plant, 258 workers participated from the pre-program group while 290 participated from the program group (206 partial program, 84 full program). Prevalence comparisons of the pre-program and program groups (partial and full) were performed by calculating prevalence ratios. A 95 percent confidence interval (95 percent CI) was derived using a cohort study method that accounted for the variance in survey techniques (cross-sectional versus longitudinal) (Bailey et al., 2010). The sensitization prevalence of the pre-program group was 3.8 times higher (95 percent CI, 1.5-9.3) than the program group, 4.0 times higher (95 percent CI, 1.4-11.6) than the partial program subgroup, and 3.3 times higher (95 percent CI, 0.8-13.7) than the full program subgroup indicating that a comprehensive preventive program can reduce, but not eliminate, occurrence of sensitization among non-sensitized workers (Bailey et al., 2010).

Rosenman et al. (2005) studied a group of several hundred workers who had been employed at a beryllium production and processing facility that operated in eastern Pennsylvania between 1957 and 1978. Of 715 former workers located, 577 were screened for BeS with the BLPT and 544 underwent chest radiography to identify cases of BeS and CBD. Workers were reported to have exposure to beryllium dust and fume in a variety of chemical forms including beryl ore, beryllium metal, beryllium fluoride, beryllium hydroxide, and beryllium oxide.

Rosenman et al. used the plant's DWA formulas to assess workers' full-shift exposure levels, based on IH data collected between 1957-1962 and 1971-1976, to calculate exposure metrics including cumulative, average, and peak for each worker in the study. The DWA was calculated based on air monitoring that consisted of GA and short-term task-based BZ samples. Workers' exposures to specific chemical and physical forms of beryllium were assessed, including insoluble beryllium (metal and oxide), soluble beryllium (fluoride and hydroxide), mixed soluble and insoluble beryllium, beryllium dust (metal, hydroxide, or oxide), fume (fluoride), and mixed dust and fume. Use of respiratory or dermal protection by workers was not reported. Exposures in the plant were high overall. Representative task-based IH samples ranged from 0.9 μ g/m³ to 84 μ g/m³ in the 1960s, falling to a range of 0.5-16.7 μ g/m³ in the 1970s. A large number of workers' mean DWA estimates (25 percent) were above the OSHA PEL of 2.0 μ g/m³, while most workers had mean DWA exposures between 0.2 and 2.0 μ g/m³ (74 percent) or below 0.02 μ g/m³ (1 percent) (Rosenman et al., Table 11; revised erratum April, 2006).

Blood samples for the BeLPT were collected from the former workers between 1996 and 2001 and were evaluated at a single laboratory. Individuals with an abnormal test result were offered repeat testing, and were classified as sensitized if the second test was also abnormal. Sixty workers with two positive BeLPTs and 50 additional workers with chest radiography suggestive of disease were offered clinical evaluation, including bronchoscopy with bronchial biopsy and BAL-BeLPT. Seven workers met both criteria. Only 56 (51 percent) of these workers proceeded with clinical evaluation, including 57 percent of those referred on the basis of confirmed abnormal BeLPT and 47 percent of those with abnormal radiographs.

Of those workers who underwent bronchoscopy, 32 (5.5 percent) with evidence of granulomas were classified as “definite” CBD cases. Twelve (2.1 percent) additional workers with positive BAL-BeLPT or confirmed positive BeLPT and radiographic evidence of upper lobe fibrosis were classified as “probable” CBD cases. Forty workers (6.9 percent) without upper lobe fibrosis who had confirmed abnormal BeLPT, but who were not biopsied or who underwent biopsy with no evidence of granuloma, were classified as sensitized without disease. It is not clear how many of the 40 workers underwent biopsy. Another 12 (2.1 percent) workers with upper lobe fibrosis and negative or unconfirmed positive BeLPT were classified as “possible” CBD cases. Nine additional workers who were diagnosed with CBD before the screening were included in some parts of the authors' analysis.

The authors reported a total prevalence of 14.5 percent for CBD (definite and probable) and sensitization. This rate, considerably higher than the overall prevalence of sensitization and disease in several other worker cohorts as described earlier in this section, reflects in part the very high exposures experienced by many workers during the plant's operation in the 1950s, 1960s and 1970s. A total of 115 workers had mean DWAs above the OSHA PEL of 2 μ g/m³. Of those, 7 (6.0 percent) had definite or probable CBD and another 13 (11 percent) were classified as sensitized without disease. The true prevalence of CBD in the group may be higher than reported, due to the low rate of clinical evaluation among sensitized workers.

Although most of the workers in this study had high exposures, sensitization and CBD also were observed within the small subgroup of participants believed to have relatively low beryllium exposures. Thirty-three cases of CBD and 24 additional cases of sensitization occurred among 339 workers with mean DWA exposures below OSHA's PEL of 2.0 μ g/m³ (Rosenman et al., Table 11, erratum 2006). Ten cases of sensitization and five cases of CBD were found among office and clerical workers, who were believed to have low exposures (levels not reported).

Follow-up time for sensitization screening of workers in this study who became sensitized during their employment had a minimum of 20 years to develop CBD prior to screening. In this sense the cohort is especially well suited to compare the exposure patterns of workers with CBD and those sensitized without disease, in contrast to several other studies of workers with only recent beryllium exposures. Rosenman et al. characterized and compared the exposures of workers with definite and probable CBD, sensitization only, and no disease or sensitization using chi-squared tests for discrete outcomes and analysis of variance (ANOVA) for continuous variables (cumulative, mean, and peak exposure levels). Exposure-response relationships were further examined with logistic regression analysis, adjusting for potential confounders including smoking, age, and beryllium exposure from outside of the plant. The authors found that cumulative, peak, and duration of exposure were significantly higher for workers with CBD than for sensitized workers without disease (p <0.05), suggesting that the risk of progressing from sensitization to CBD is related to the level or extent of exposure a worker experiences. The risk of developing CBD following sensitization appeared strongly related to exposure to insoluble forms of beryllium, which are cleared slowly from the lung and increase beryllium lung burden more rapidly than quickly mobilized soluble forms. Individuals with CBD had higher exposures to insoluble beryllium than those classified as sensitized without disease, while exposure to soluble beryllium was higher among sensitized individuals than those with CBD.

Cumulative, mean, peak, and duration of exposure were found to be comparable for workers with CBD and workers without sensitization or CBD (“normal” workers). Cumulative, peak, and duration of exposure were significantly lower for sensitized workers without disease than for normal workers. Rosenman et al. suggested that genetic predisposition to sensitization and CBD may have obscured an exposure-response relationship in this study, and plan to control for genetic risk factors in future studies. Exposure misclassification from the 1950s and 1960s may have been another limitation in this study, introducing bias that could have influenced the lack of exposure response. It is also unknown if the 25 percent who died from CBD-related conditions may have had higher exposures.

A follow-up was conducted of the cross-sectional study of a population of workers first evaluated by Kreiss et al. (1997) and Rosenman et al. (2005) at a beryllium production and processing facility in eastern Pennsylvania by Schuler et al. (2012), and in a companion study by Virji et al. (2012). Schuler et al. evaluated the worker population employed in 1999 with six years or less work tenure in a cross-sectional study. The investigators evaluated the worker population by administering a work history questionnaire with a follow-up examination for sensitization and CBD. A job-exposure matrix (JEM) was combined with work histories to create individual estimates of average, cumulative, and highest-job-related exposure for total, respirable, and sub-micron beryllium mass concentration. Of the 291 eligible workers, 90.7 percent (264) participated in the study. Sensitization prevalence was 9.8 percent (26/264) with CBD prevalence of 2.3 percent (6/264). The investigators found a general pattern of increasing sensitization prevalence as the exposure quartile increased indicating an exposure-response relationship. The investigators found positive associations with both total and respirable mass concentration with sensitization (average and highest job) and CBD (cumulative). Increased sensitization prevalence was observed with metal oxide production alloy melting and casting, and maintenance. CBD was associated with melting and casting. The investigators summarized that both total and respirable mass concentration were relevant predictors of risk (Schuler et al., 2012).

In the companion study by Virji et al. (2012), the investigators reconstructed historical exposure from 1994 to 1999 utilizing the personal sampling data collected in 1999 as baseline exposure estimates (BEE). The study evaluated techniques for reconstructing historical data to evaluate exposure-response relationships for epidemiological studies. The investigators constructed JEMs using the BEE and estimates of annual changes in exposure for 25 different process areas. The investigators concluded these reconstructed JEMs could be used to evaluate a range of exposure parameters from total, respirable and submicron mass concentration including cumulative, average, and highest exposure. These two studies demonstrate that high-quality exposure estimates can be developed both for total mass and respirable mass concentrations.

e. Beryllium Machining Operations

Newman et al. (2001) and Kelleher et al. (2001) studied a group of 235 workers at a beryllium metal machining plant. Since the plant opened in 1969, its primary operations have been machining and polishing beryllium metal and high-beryllium content composite materials, with occasional machining of beryllium oxide/metal matrix (‘E-metal'), and beryllium alloys. Other functions include machining of metals other than beryllium; receipt and inspection of materials; acid etching; final inspection, quality control, and shipping of finished materials; tool making; and engineering, maintenance, administrative and supervisory functions (Newman et al., 2001; Madl et al., 2007). Machining operations, including milling, grinding, lapping, deburring, lathing, and electrical discharge machining (EDM), were performed in an open-floor plan production area. Most non-machining jobs were located in a separate, adjacent area; however, non-production employees had access to the machining area.

Engineering and administrative measures, rather than PPE, were primarily used to control beryllium exposures at the plant (Madl et al., 2007). Based on interviews with long-standing employees of the plant, Kelleher et al. reported that work practices were relatively stable until 1994, when a worker was diagnosed with CBD and a new exposure control program was initiated. Between 1995 and 1999 new engineering and work practice controls were implemented, including removal of pressurized air hoses and discouragement of dry sweeping (1995), enclosure of deburring processes (1996), mandatory uniforms (1997), and installation or updating of local exhaust ventilation (LEV) in EDM, lapping, deburring, and grinding processes (1998) (Madl et al., 2007). Throughout the plant's history, respiratory protection was used mainly for “unusually large, anticipated exposures” to beryllium (Kelleher et al., 2001), and was not routinely used otherwise (Newman et al., 2001).

All workers at the plant participated in a beryllium disease surveillance program initiated in 1994, and were screened for beryllium sensitization with the BeLPT beginning in 1995. A BeLPT result was considered abnormal if two or more of six stimulation indices exceeded the normal range (see section on BeLPT testing above), and was considered borderline if one of the indices exceeded the normal range. A repeat BeLPT was conducted for workers with abnormal or borderline initial results. Workers were identified as beryllium sensitized and referred for a clinical evaluation, including bronchoalveolar lavage (BAL) and transbronchial lung biopsy, if the repeat test was abnormal. CBD was diagnosed upon evidence of sensititization with granulomas or mononuclear cell infiltrates in the lung tissue (Newman et al., 2001). Following the initial plant-wide screening, plant employees were offered BeLPT testing at two-year intervals. Workers hired after the initial screening were offered a BeLPT within 3 months of their hire date, and at 2-year intervals thereafter (Madl et al., 2007).

Kelleher et al. performed a nested case-control study of the 235 workers evaluated in Newman et al. (2001) to evaluate the relationship between beryllium exposure levels and risk of sensitization and CBD (Kelleher et al., 2001). The authors evaluated exposures at the plant using IH samples they had collected between 1996 and 1999, using personal cascade impactors designed to measure the mass of beryllium particles less than 6 μ m, particles less than 1 μm in diameter, and total mass. The great majority of workers' exposures were below the OSHA PEL of 2 μ g/m³. However, a few higher levels were observed in machining jobs including deburring, lathing, lapping, and grinding. Based on a statistical comparison between their samples and historical data provided by the plant, the authors concluded that worker beryllium exposures across all time periods could be approximated using the 1996-1999 data. They estimated workers' cumulative and `lifetime weighted' (LTW) beryllium exposure based on the exposure samples they collected for each job in 1996-1999 and company records of each worker's job history.

Twenty workers with beryllium sensitization or CBD (cases) were compared to 206 workers (controls) for the case-control analysis from the study evaluating workers originally conducted by Newman et al. Thirteen workers were diagnosed with CBD based on lung biopsy evidence of granulomas and/or mononuclear cell infiltrates (11) or positive BAL results with evidence of lymphocytosis (2). Seven were evaluated for CBD and found to be sensitized only, thus twenty composing the case group. Nine of the remaining 215 workers first identified in original study (Newman et al., 2001) were excluded due to incomplete job history information, leaving 206 workers in the control group.

Kelleher et al.' s analysis included comparisons of the case and control groups' median exposure levels; calculation of odds ratios for workers in high, medium, and low exposure groups; and logistic regression testing of the association of sensitization or CBD with exposure level and other variables. Median cumulative exposures for total mass, particles <6 μ m, and particles <1 μm were approximately three times higher among the cases than controls, although the relationships observed were not statistically significant (p values ~ 0.2). No clear difference between cases and controls was observed for the median LTW exposures. Odds ratios with sensitization and CBD as outcomes were elevated in high (upper third) and intermediate exposure groups relative to low (lowest third) exposure groups for both cumulative and LTW exposure, though the results were not statistically significant (p > 0.1). In the logistic regression analysis, only machinist work history was a significant predictor of case status in the final model. Quantitative exposure measures were not significant predictors of sensitization or disease risk.

Citing an 11.5 percent prevalence of beryllium sensitization or CBD among machinists as compared with 2.9 percent prevalence among workers with no machinist work history, the authors concluded that the risk of sensitization and CBD is increased among workers who machine beryllium. Although differences between cases and controls in median cumulative exposure did not achieve conventional thresholds for statistical significance, the authors noted that cumulative exposures were consistently higher among cases than controls for all categories of exposure estimates and for all particle sizes, suggesting an effect of cumulative exposure on risk. The levels at which workers developed CBD and sensitization were predominantly below OSHA's current PEL of 2 μ g/m³, and no cases of sensitization or CBD were observed among workers with LTW exposure <0.02 μg/m³. Twelve (60 percent) of the 20 sensitized workers had LTW exposures > 0.20 μ g/m³.

In 2007, Madl et a l. published an additional study of 27 workers at the machining plant who were found to be sensitized or diagnosed with CBD between the start of medical surveillance in 1995 and 2005. As previously described, workers were offered a BeLPT in the initial 1995 screening (or within 3 months of their hire date if hired after 1995) and at 2-year intervals after their first screening. Workers with two positive BeLPTs were identified as sensitized and offered clinical evaluation for CBD, including bronchoscopy with BAL and transbronchial lung biopsy. The criteria for CBD in this study were somewhat stricter than those used in the Newman et al. study, requiring evidence of granulomas on lung biopsy or detection of X-ray or pulmonary function changes associated with CBD, in combination with two positive BeLPTs or one positive BAL-BeLPT.

Based on the history of the plant's control efforts and their analysis of historical IH data, Madl et al. identified three “exposure control eras”: A relatively uncontrolled period from 1980-1995; a transitional period from 1996 to 1999; and a relatively well-controlled “modern” period from 2000-2005. They found that the engineering and work practice controls instituted in the mid-1990s reduced workers' exposures substantially, with nearly a 15-fold difference in reported exposure levels between the pre-control and the modern period (Madl et al., 2007). Madl et al. estimated workers' exposures using LP samples collected between 1980 and 2005, including those collected by Kelleher et al., and work histories provided by the plant. As described more fully in the study, they used a variety of approaches to describe individual workers' exposures, including approaches designed to characterize the highest exposures workers were likely to have experienced. Their exposure-response analysis was based primarily on an exposure metric they derived by identifying the year and job of each worker's pre-diagnosis work history with the highest reported exposures. They used the upper 95th percentile of the LP samples collected in that job and year (in some cases supplemented with data from other years) to characterize the worker's upper-level exposures.

Based on their estimates of workers' upper level exposures, Madl et al. concluded that workers with sensitization or CBD were likely to have been exposed to airborne beryllium levels greater than 0.2 μg/m³ as an 8-hour TWA at some point in their history of employment in the plant. They also concluded that most sensitization and CBD cases were likely to have been exposed to levels greater than 0.4 μg/m³ at some point in their work at the plant. Madl et al. did not reconstruct exposures for workers at the plant who did not have sensitization or CBD and therefore could not determine whether non-cases had upper-bound exposures lower than these levels. They found that upper-bound exposure estimates were generally higher for workers with CBD than for those who were sensitized but not diagnosed with CBD at the conclusion of the study (Madl et al., 2007). Because CBD is an immunological disease and beryllium sensitization has been shown to occur within a year of exposure for some workers, Madl et al. argued that their estimates of workers' short-term upper-bound exposures may better capture the exposure levels that led to sensitization and disease than estimates of long-term cumulative or average exposures such as the LTW exposure measure constructed by Kelleher et al. (Madl et al., 2007).

f. Beryllium Oxide Ceramics

Kreiss et al. (1993) conducted a screening of current and former workers at a plant that manufactured beryllium ceramics from beryllium oxide between 1958 and 1975, and then transitioned to metalizing circuitry onto beryllium ceramics produced elsewhere. Of the plant's 1,316 current and 350 retired workers, 505 participated who had not previously been diagnosed with CBD or sarcoidosis, including 377 current and 128 former workers. Although beryllium exposure was not estimated quantitatively in this survey, the authors conducted a questionnaire to assess study participants' exposures qualitatively. Results showed that 55 percent of participants reported working in jobs with exposure to beryllium dust. Close to 25 percent of participants did not know if they had exposure to beryllium, and just over 20 percent believed they had not been exposed.

BeLPT tests were administered to all 505 participants in the 1989-1990 screening period and evaluated at a single lab. Seven workers had confirmed abnormal BeLPT results and were identified as sensitized; these workers were also diagnosed with CBD based on findings of granulomas upon clinical evaluation. Radiograph screening led to clinical evaluation and diagnosis of two additional CBD cases, who were among three participants with initially abnormal BeLPT results that could not be confirmed on repeat testing. In addition, nine workers had been previously diagnosed with CBD, and another five were diagnosed shortly after the screening period, in 1991-1992.

Eight (3.7 percent of the screening population) of the nine CBD cases identified in the screening population were hired before the plant stopped producing beryllium ceramics in 1975, and were among the 216 participants who had reported having been near or exposed to beryllium dust. Particularly high CBD rates of 11.1-15.8 percent were found among screening participants who had worked in process development/engineering, dry pressing, and ventilation maintenance jobs believed to have high or uncontrolled dust exposure. One case (0.6 percent) of CBD was diagnosed among the 171 study participants who had been hired after the plant stopped producing beryllium ceramics. Although this worker was hired eight years after the end of ceramics production, he had worked in an area later found to be contaminated with beryllium dust. The authors concluded that the study results suggested an exposure-response relationship between beryllium exposure and CBD, and recommended beryllium exposure control to reduce workers' risk of CBD.

Kreiss et al. later published a study of workers at a second ceramics plant located in Tucson, AZ (Kreiss et al., 1996), which since 1980 had produced beryllium ceramics from beryllium oxide powder manufactured elsewhere. IH measurements collected between 1981 and 1992, primarily GA or short-term BZ samples and a few (<100) LP samples, were available from the plant. Airborne beryllium exposures were generally low. The majority of area samples were below the analytical detection limit of 0.1 μg/m³, while LP and short-term BZ samples had medians of 0.3 μg/m³. However, 3.6 percent of short-term BZ samples and 0.7 percent of GA samples exceeded 5.0 μg/mg³, while LP samples ranged from 0.1 to 1.8 μg/m³. Machining jobs had the highest beryllium exposure levels among job tasks, with short-term BZ samples significantly higher for machining jobs than for non-machining jobs (median 0.6 μg/m³ vs. 0.3 μg/mg³, p = 0.0001). The authors used DWA formulas provided by the plant to estimate workers' full-shift exposure levels, and to calculate cumulative and average beryllium exposures for each worker in the study. The median cumulative exposure was 591.7 mg-days/m³ and the median average exposure was 0.35 μg/m³.

One hundred thirty-six of the 139 workers employed at the plant at the time of the Kreiss et al. (1996) study underwent BeLPT screening and chest radiographs in 1992. Blood samples were split between two laboratories. If one or both test results were abnormal, an additional sample was collected and split between the labs. Seven workers with an abnormal result on two draws were initially identified as sensitized. Those with confirmed abnormal BeLPTs or abnormal chest X-rays were offered clinical evaluation for CBD, including transbronchial lung biopsy and BAL BeLPT. CBD was diagnosed based on observation of granulomas on lung biopsy, in five of the six sensitized workers who accepted evaluation. An eighth case of sensitization and sixth case of CBD were diagnosed in one worker hired in October 1991 whose initial BeLPT was normal, but who was confirmed as sensitized and found to have lung granulomas less than two years later, after sustaining a beryllium-contaminated skin wound. The plant medical department reported 11 additional cases of CBD among former workers (Kreiss et al., 1996). The overall prevalence of sensitization in the plant was 5.9 percent, with a 4.4 percent prevalence of CBD.

Kreiss et al. reported that six (75 percent) of the eight sensitized workers were exposed as machinists during or before the period October 1985-March 1988, when measurements were first available for machining jobs. The authors reported that 14.3 percent of machinists were sensitized, compared to 1.2 percent of workers who had never been machinists (p <0.01). Workers' estimated cumulative and average beryllium exposures did not differ significantly for machinists and non-machinists, or for cases and non-cases. As in the previous study of the same ceramics plant published by Kreiss et al. in 1993, one case of CBD was diagnosed in a worker who had never been employed in a production job. This worker was employed in administration, a job with a median DWA of 0.1 μg/m³ (range 0.1-0.3).

In 1998, Henneberger et al. conducted a follow-up cross-sectional survey of 151 employees employed at the beryllium ceramics plant studied by Kreiss et al. (1996) (Henneberger et al., 2001). Employees were eligible who either had not participated in the Kreiss et al. survey (“short-term workers”—74 of those studied by Henneberger et al.), or who had participated and were not found to have sensitization or disease (“long-term workers”—77 of those studied by Henneberger et al.).

The authors estimated workers' cumulative, average, and peak beryllium exposures based on the plant's formulas for estimating job-specific DWA exposures, participants' work histories, and area and short-term task-specific BZ samples collected from the start of full production at the plant in 1981 to 1998. The long-term workers, who were hired before the 1992 study was conducted, had generally higher estimated exposures (median of average exposures—0.39 μg/m³; mean—14.9 μg/m³) than the short-term workers, who were hired after 1992 (median 0.28 μg/m³, mean 6.1 μg/m³).

Fifteen cases of sensitization were found, including eight among short-term and seven among long-term workers. Eight of the 15 workers were found to have CBD. Of the workers diagnosed with CBD, seven (88 percent) were long-term workers. One non-sensitized long-term worker and one sensitized long-term worker declined clinical examination.

Henneberger et al. reported a higher prevalence of sensitization among long-term workers with “high” (greater than median) peak exposures compared to long-term workers with “low” exposures; however, this relationship was not statistically significant. No association was observed for average or cumulative exposures. The authors reported higher prevalence of sensitization (but not statistically significant) among short-term workers with “high” (greater than median) average, cumulative, and peak exposures compared to short-term workers with “low” exposures of each type.

The cumulative incidence of sensitization and CBD was investigated in a cohort of 136 workers at the beryllium ceramics plant previously studied by the Kreiss and Henneberger groups (Schuler et al., 2008). The study cohort consisted of those who participated in the plant-wide BeLPT screening in 1992. Both current and former workers from this group were invited to participate in follow-up BeLPT screenings in 1998, 2000, and 2002-03. A total of 106 of the 128 non-sensitized individuals in 1992 participated in the 11-year follow-up. Sensitization was defined as a confirmed abnormal BeLPT based on the split blood sample-dual laboratory protocol described earlier. CBD was diagnosed in sensitized individuals based on pathological findings from transbronchial biopsy and BAL fluid analysis. The 11-year crude cumulative incidence of sensitization and CBD was 13 percent (14 of 106) and 8 percent (9 of 106) respectively. The cumulative prevalence was about triple the point prevalences determined in the initial 1992 cross-sectional survey. The corrected cumulative prevalences for those that ever worked in machining were nearly twice that for non-machinists. The data illustrate the value of longitudinal medical screening over time to obtain a more accurate estimate of the occurrence of sensitization and CBD among an exposed working population.

Following the 1998 survey, the company continued efforts to reduce exposures and risk of sensitization and CBD by implementing additional engineering, administrative, and PPE measures (Cummings et al., 2007). Respirator use was required in production areas beginning in 1999, and latex gloves were required beginning in 2000. The lapping area was enclosed in 2000, and enclosures were installed for all mechanical presses in 2001. Between 2000 and 2003, water-resistant or water-proof garments, shoe covers, and taped gloves were incorporated to keep beryllium-containing fluids from wet machining processes off the skin. The new engineering measures did not appear to substantially reduce airborne beryllium levels in the plant. LP samples collected between 2000 and 2003 had a median of 0.18 μg/m³, similar to the 1994-1999 samples. However, respiratory protection requirements to control workers' airborne beryllium exposures were instituted prior to the 2000 sample collections.

To test the efficacy of the new measures instituted after 1998, in January 2000 the company began screening new workers for sensitization at the time of hire and at 3, 6, 12, 24, and 48 months of employment. These more stringent measures appear to have substantially reduced the risk of sensitization among new employees. Of 126 workers hired between 2000 and 2004, 93 completed BeLPT testing at hire and at least one additional test at 3 months of employment. One case of sensitization was identified at 24 months of employment (1 percent). This worker had experienced a rash after an incident of dermal exposure to lapping fluid through a gap between his glove and uniform sleeve, indicating that he may have become sensitized via the skin. He was tested again at 48 months of employment, with an abnormal result.

A second worker in the 2000-2004 group had two abnormal BeLPT tests at the time of hire, and a third had one abnormal test at hire and a second abnormal test at 3 months. Both had normal BeLPTs at 6 months, and were not tested thereafter. A fourth worker had one abnormal BeLPT result at the time of hire, a normal result at 3 months, an abnormal result at 6 months, and a normal result at 12 months. Four additional workers had one abnormal result during surveillance, which could not be confirmed upon repeat testing.

Cummings et al. calculated two sensitization rates based on these screening results: (1) a rate using only the sensitized worker identified at 24 months, and (2) a rate including all four workers who had repeated abnormal results. They reported a sensitization incidence rate (IR) of 0.7 per 1,000 person-months to 2.7 per 1,000 person-months for the workers hired between 2000 and 2004, using the sum of sensitization-free months of employment among all 93 workers as the denominator.

The authors also estimated an incidence rate (IR) of 5.6 per 1,000 person-months for workers hired between 1993 and the 1998 survey. This estimated IR was based on one BeLPT screening, rather than BeLPTs conducted throughout the workers' employment. The denominator in this case was the total months of employment until the 1998 screening. Because sensitized workers may have been sensitized prior to the screening, the denominator may overestimate sensitization-free time in the legacy group, and the actual sensitization IR for legacy workers may be somewhat higher than 5.6 per 1,000 person-months. Based on comparison of the IRs, the authors concluded that the addition of respirator use, dermal protection, and housekeeping improvements appeared to have reduced the risk of sensitization among workers at the plant, even though airborne beryllium levels in some areas of the plant had not changed significantly since the 1998 survey.

g. Copper-Beryllium Alloy Processing and Distribution

Schuler et al. (2005) studied a group of 152 workers at a facility processing copper-beryllium alloys and small quantities of nickel-beryllium alloys, and converting semi-finished alloy strip and wire into finished strip, wire and rod. Production activities included annealing, drawing, straightening, point and chamfer, rod and wire packing, die grinding, pickling, slitting, and degreasing. Periodically in the plant's history, they also did salt baths, cadmium plating, welding and deburring. Since the late 1980s, rod and wire production processes were physically segregated from strip metal production. Production support jobs included mechanical maintenance, quality assurance, shipping and receiving, inspection, and wastewater treatment. Administration was divided into staff primarily working within the plant and personnel who mostly worked in office areas (Schuler, et al., 2005). Workers' respirator use was limited, mostly to occasional tasks where high exposures were anticipated.

Following the 1999 diagnosis of a worker with CBD, the company surveyed the workforce, offering all current employees BeLPT testing in 2000 and offering sensitized workers clinical evaluation for CBD, including BAL and transbronchial biopsy. Of the facility's 185 employees, 152 participated in the BeLPT screening. Samples were split between two laboratories, with additional draws and testing for confirmation if conflicting tests resulted in the initial draw. Ten participants (7 percent) had at least two abnormal BeLPT results. The results of nine workers who had abnormal BeLPT results from only one laboratory were not included because the authors believed it was experiencing technical problems with the test (Schuler et al., 2005). CBD was diagnosed in six workers (4 percent) on evidence of pathogenic abnormalities (e.g., granulomas) or evidence of clinical abnormalities consistent with CBD based on pulmonary function testing, pulmonary exercise testing, and/or chest radiography. One worker diagnosed with CBD had been exposed to beryllium during previous work at another copper-beryllium processing facility.

Schuler et al. evaluated airborne beryllium levels at the plant using IH samples collected between 1969 and 2000, including 4,524 GA samples, 650 LP samples and 815 short-duration (3-5 min) high volume (SD-HV) BZ task-specific samples. Occupational exposures to airborne beryllium were generally low. Ninety-nine percent of all LP measurements were below the current OSHA PEL of 2.0 μg/m³ (8-hr TWA); 93 percent were below the DOE action level of 0.2 μg/m³; and the median value was 0.02 μg/m³. The SD-HV BZ samples had a median value of 0.44 μg/m³, with 90 percent below the OSHA Short-Term Exposure Limit (STEL) of 5.0 μg/m³. The highest levels of beryllium were found in rod and wire production, particularly in wire annealing and pickling, the only production job with a median personal sample measurement greater than 0.1 μg/m³ (median 0.12 μg/m³; range 0.01-7.8 μg/m³) (Schuler et al., Table 4). These concentrations were significantly higher than the exposure levels in the strip metal area (median 0.02, range 0.01-0.72 μg/m³), in production support jobs (median 0.02, range <0.01-0.33 μg/m³), plant administration (median 0.02, range <0.01-0.11 μg/m³), and office administration jobs (median 0.01, range <0.01-0.06 μg/m³).

The authors reported that eight of the ten sensitized employees, including all six CBD cases, had worked in both major production areas during their tenure with the plant. The 7 percent prevalence (6 of 81 workers) of CBD among employees who had ever worked in rod and wire was statistically significantly elevated compared with employees who had never worked in rod and wire (p <0.05), while the 6 percent prevalence (6 of 94 workers) among those who had worked in strip metal was not significantly elevated compared to non-strip metal workers (p > 0.1). Based on these results, together with the higher exposure levels reported for the rod and wire production area, Schuler et al. concluded that work in rod and wire was a key risk factor for CBD in this population. Schuler et al. also found a high prevalence (13 percent) of sensitization among workers who had been exposed to beryllium for less than a year at the time of the screening, a rate similar to that found by Henneberger et al. among beryllium ceramics workers exposed for one year or less (16 percent, Henneberger et al., 2001). All four workers who were sensitized without disease had been exposed 5 years or less; conversely, all six of the workers with CBD had first been exposed to beryllium at least five years prior to the screening (Schuler et al., Table 2).

As has been seen in other studies, beryllium sensitization and CBD were found among workers who were typically exposed to low time-weighted average airborne concentrations of beryllium. While jobs in the rod and wire area had the highest exposure levels in the plant, the median personal sample value was only 0.12 μg/m³. However, workers may have occasionally been exposed to higher beryllium levels for short periods during specific tasks. A small fraction of personal samples recorded in rod and wire were above the OSHA PEL of 2.0 μg/m³, and half of workers with sensitization or CBD reported that they had experienced a “high-exposure incident” at some point in their work history (Schuler et al., 2005). The only group of workers with no cases of sensitization or CBD, a group of 26 office administration workers, was the group with the lowest recorded exposures (median personal sample 0.01 μg/m³, range <0.01-0.06 μg/m³).

After the BeLPT screening was conducted in 2000, the company began implementing new measures to further reduce workers' exposure to beryllium (Thomas et al., 2009). Requirements designed to minimize dermal contact with beryllium, including long-sleeve facility uniforms and polymer gloves, were instituted in production areas in 2000. In 2001 the company installed LEV in die grinding and polishing. LP samples collected between June 2000 and December 2001 show reduced exposures plant-wide. Of 2,211 exposure samples collected, 98 percent were below 0.2 μg/m³, and 59 percent below the limit of detection (LOD), which was either 0.02 µg/m³ or 0.2 µg/m³ depending on the method of sample analysis (Thomas et al., 2009). Median values below 0.03 μg/m³ were reported for all processes except the wire annealing and pickling process. Samples for this process remained somewhat elevated, with a median of 0.1 μg/m³. In January 2002, the plant enclosed the wire annealing and pickling process in a restricted access zone (RAZ), requiring respiratory PPE in the RAZ and implementing stringent measures to minimize the potential for skin contact and beryllium transfer out of the zone. While exposure samples collected by the facility were sparse following the enclosure, they suggest exposure levels comparable to the 2000-01 samples in areas other than the RAZ. Within the RAZ, required use of powered air-purifying respirators indicates that respiratory exposure was negligible.

To test the efficacy of the new measures in preventing sensitization and CBD, in June 2000 the facility began an intensive BeLPT screening program for all new workers. The company screened workers at the time of hire; at intervals of 3, 6, 12, 24, and 48 months; and at 3-year intervals thereafter. Among 82 workers hired after 1999, three (3.7 percent) cases of sensitization were found. Two (5.4 percent) of 37 workers hired prior to enclosure of the wire annealing and pickling process were found to be sensitized within 3 and 6 months of beginning work at the plant. One (2.2 percent) of 45 workers hired after the enclosure was confirmed as sensitized.

Thomas et al. calculated a sensitization IR of 1.9 per 1,000 person-months for the workers hired after the exposure control program was initiated in 2000 (“program workers”), using the sum of sensitization-free months of employment among all 82 workers as the denominator (Thomas et al., 2009). They calculated an estimated IR of 3.8 per 1,000 person-months for 43 workers hired between 1993 and 2000 who had participated in the 2000 BeLPT screening (“legacy workers”). This estimated IR was based on one BeLPT screening, rather than BeLPTs conducted throughout the legacy workers' employment. The denominator in this case is the total months of employment until the 2000 screening. Because sensitized workers may have been sensitized prior to the screening, the denominator may overestimate sensitization-free time in the legacy group, and the actual sensitization IR for legacy workers may be somewhat higher than 3.8 per 1,000 person-months. Based on comparison of the IRs and the prevalence rates discussed previously, the authors concluded that the combination of dermal protection, respiratory protection, housekeeping improvements and engineering controls implemented beginning in 2000 appeared to have reduced the risk of sensitization among workers at the plant. However, they noted that the small size of the study population and the short follow-up time for the program workers suggested that further research is needed to confirm the program's efficacy (Thomas et al., 2009).

Stanton et al. (2006) conducted a study of workers in three different copper-beryllium alloy distribution centers in the United States. The distribution centers, including one bulk products center established in 1963 and strip metal centers established in 1968 and 1972, sell products received from beryllium production and finishing facilities and small quantities of copper-beryllium, aluminum-beryllium, and nickel-beryllium alloy materials. Work at distribution centers does not require large-scale heat treatment or manipulation of material typical of beryllium processing and machining plants, but involves final processing steps that can generate airborne beryllium. Slitting, the main production activity at the two strip product distribution centers, generates low levels of airborne beryllium particles, while operations such as tensioning and welding used more frequently at the bulk products center can generate somewhat higher levels. Non-production jobs at all three centers included shipping and receiving, palletizing and wrapping, production-area administrative work, and office-area administrative work.

The authors estimated workers' beryllium exposures using IH data from company records and job history information collected through interviews conducted by a company occupational health nurse. Stanton et al. evaluated airborne beryllium levels in various jobs based on 393 full-shift LP samples collected from 1996 to 2004. Airborne beryllium levels at the plant were generally very low, with 54 percent of all samples at or below the LOD, which ranged from 0.02 to 0.1 μg/m³. The authors reported a median of 0.03 μg/m³ and an arithmetic mean of 0.05 μg/m³ for the 393 full-shift LP samples, where samples below the LOD were assigned a value of half the applicable LOD. Median and geometric mean values for specific jobs ranged from 0.01-0.07 and 0.02-0.07 µg/m³, respectively. All measurements were below the OSHA PEL of 2.0 μg/m³ and 97 percent were below the DOE action level of 0.2 μg/m³. The paper does not report use of respiratory or skin protection. Exposure conditions may have changed somewhat over the history of the plant due to changes in exposure control measures, including improvements to product and container cleaning practices instituted during the 1990s.

Eighty-eight of the 100 workers (88 percent) employed at the three centers at the time of the study participated in screening for beryllium sensitization. Blood samples were collected between November 2000 and March 2001 by the company's medical staff. Samples collected from employees of the strip metal centers were split and evaluated at two laboratories, while samples from the bulk product center workers were evaluated at a single laboratory. Participants were considered to be “sensitized” to beryllium if two or more BeLPT results, from two laboratories or from repeat testing at the same laboratory, were found to be abnormal. One individual was found to be sensitized and was offered clinical evaluation, including BAL and fiberoptic bronchoscopy. He was found to have lung granulomas and was diagnosed with CBD.

The worker diagnosed with CBD had been employed at a strip metal distribution center from 1978 to 2000 as a shipper and receiver, loading and unloading trucks delivering materials from a beryllium production facility and to the distribution center's customers. Although the LP samples collected for his job between 1996 and 2000 were generally low (n = 35, median 0.01, range < 0.02-0.13 µg/m³), it is not clear whether these samples adequately characterize his exposure conditions over the course of his work history. He reported that early in his work history, containers of beryllium oxide powder were transported on the trucks he entered. While he did not recall seeing any breaks or leaks in the beryllium oxide containers, some containers were known to have been punctured by forklifts on trailers used by the company during the period of his employment, and could have contaminated trucks he entered. With 22 years of employment at the facility, this worker had begun beryllium-related work earlier and performed it longer than about 90 percent of the study population (Stanton et al., 2006).

h. Nuclear Weapons Production Facilities & Cleanup of Former Facilities

Primary exposure from nuclear weapons production facilities comes from beryllium metal and beryllium alloys. A study conducted by Kreiss et al. (1989) documented sensitization and CBD among beryllium-exposed workers in the nuclear industry. A company medical department identified 58 workers with beryllium exposure among a work force of 500, of whom 51 (88 percent) participated in the study. Twenty-four workers were involved in research and development (R&D), while the remaining 27 were production workers. The R&D workers had a longer tenure with a mean time from first exposure of 21.2 years, compared to a mean time since first exposure of 5 years among the production workers. The number of workers with abnormal BeLPT readings was 6, with 4 being diagnosed with CBD. This resulted in an estimated 11.8 percent prevalence of sensitization.

Kreiss et al. (1993) expanded the work of Kreiss et al. (1989) by performing a cross-sectional study of 895 (current and former) beryllium workers in the same nuclear weapons plant. Participants were placed in qualitative exposure groups (“no exposure,” “minimal exposure,” “intermittent exposure,” and “consistent exposure”) based on questionnaire responses. The number of workers with abnormal BeLPT totaled 18 with 12 being diagnosed with CBD. Three additional workers with sensitization developed CBD over the next 2 years. Sensitization occurred in all of the qualitatively defined exposure groups. Individuals who had worked as machinists were statistically overrepresented among beryllium-sensitized cases, compared with non-cases. Cases were more likely than non-cases to report having had a measured overexposure to beryllium (p = 0.009), a factor which proved to be a significant predictor of sensitization in logistic regression analyses, as was exposure to beryllium prior to 1970. Beryllium sensitized cases were also significantly more likely to report having had cuts that were delayed in healing (p = 0.02). The authors concluded that individual variability and susceptibility along with exposure circumstances are important factors in developing beryllium sensitization and CBD.

In 1991, the Beryllium Health Surveillance Program (BHSP) was established at the Rocky Flats Nuclear Weapons Facility to offer BLPT screening to current and former employees who may have been exposed to beryllium (Stange et al., 1996). Participants received an initial BeLPT and follow-ups at one and three years. Based on histologic evidence of pulmonary granulomas and a positive BAL-BeLPT, Stange et al. published a study of 4,397 BHSP participants tested from June 1991 to March 1995, including current employees (42.8 percent) and former employees (57.2 percent). Twenty-nine cases of CBD and 76 cases of sensitization were identified. The sensitization rate for the population was 2.43 percent. Available exposure data included fixed airhead (FAH) exposure samples collected between 1970 and 1988 (mean concentration 0.016 µg/m³) and personal samples collected between 1984 and 1987 (mean concentration 1.04 µg/m³). Cases of CBD and sensitization were noted in individuals in all jobs classifications, including those believed to involve minimal exposure to beryllium. The authors recommended ongoing surveillance for workers in all jobs with potential for beryllium exposure.

Stange et al. (2001) extended the previous study, evaluating 5,173 participants in the Rocky Flats BHSP who were tested between June 1991 and December 1997. Three-year serial testing was offered to employees who had not been tested for three years or more and did not show beryllium sensitization during the previous study. This resulted in 2,891 employees being tested. Of the 5,173 workers participating in the study, 172 were found to have abnormal BeLPT. Ninety-eight (3.33 percent) of the workers were found to be sensitized (confirmed abnormal BeLPT results) in the initial screening, conducted in 1991. Of these workers 74 were diagnosed with CBD (history of beryllium exposure, evidence of non-caseating granulomas or mononuclear cell infiltrates on lung biopsy, and a positive BeLPT or BAL-BeLPT). A follow-up survey of 2,891 workers three years later identified an additional 56 sensitized workers and an additional seven cases of CBD. Sensitization and CBD rates were analyzed with respect to gender, building work locations, and length of employment. Historical employee data included hire date, termination date, leave of absences, and job title changes. Exposure to beryllium was determined by job categories and building or work area codes. Personal beryllium air monitoring results were used, when available, from employees with the same job title or similar job. However, no quantitative information was presented in the study. The authors conclude that for some individuals, exposure to beryllium at levels less that the OSHA PEL could cause sensitization and CBD.

Viet et al. (2001) conducted a case-control study of the Rocky Flats worker population studied by Stange et al. (1996 and 2001) to examine the relationship between estimated beryllium exposure level and risk of sensitization or CBD. The worker population included 74 beryllium-sensitized workers and 50 workers diagnosed with CBD. Beryllium exposure levels were estimated based on FAH airhead samples from one building, the beryllium machine shop. These were collected away from the BZ of the machine operator and likely underestimated exposure. To estimate levels in other locations, these air sample concentrations were used to construct a job exposure matrix that included the determination of the Building 444 exposure estimates for a 30-year period; each subject's work history by job location, task, and time period; and assignment of exposure estimates to each combination of job location, task, and time period as compared to Building 444 machinists. The authors adjusted the levels observed in the machine shop by factors based on interviews with former workers. Workers' estimated mean exposure concentrations ranged from 0.083 µg/m³ to 0.622 µg/m³. Estimated maximum air concentrations ranged from 0.54 µg/m³ to 36.8 µg/m³. Cases were matched to controls of the same age, race, gender, and smoking status (Viet et al., 2001).

Estimated mean and cumulative exposure levels and duration of employment were found to be significantly higher for CBD cases than for controls. Estimated mean exposure levels were significantly higher for sensitization cases than for controls. No significant difference was observed for estimated cumulative exposure or duration of exposure. Similar results were found using logistic regression analysis, which identified statistically significant relationships between CBD and both cumulative and mean estimated exposure, but did not find significant relationships between estimated exposure levels and sensitization without CBD. Comparing CBD with sensitization cases, Viet et al. found that workers with CBD had significantly higher estimated cumulative and mean beryllium exposure levels than workers who were sensitized, but did not have CBD.

Johnson et al. (2001) conducted a review of personal sampling records and medical surveillance reports at an atomic weapons establishment in Cardiff, United Kingdom. The study evaluated airborne samples collected over the 36-year period of operation for the plant. Data included 367,757 area samples and 217,681 personal lapel samples from 194 workers over the time period from 1981-1997. Data was available prior to this time period but was not analyzed since this data was not available electronically. The authors estimated that over the 17 years of measurement data analyzed, airborne beryllium concentrations did exceed 2.0 µg/m³, however, due to the limitations with regard to collection times it is difficult to assess the full reliability of this estimate. The authors noted that in the entire plant's history, only one case of CBD had been diagnosed. It was also noted that BeLPT has not been routinely conducted among any of the workers at this facility.

Armojandi et al. (2010) conducted a cross-sectional study of workers at a nuclear weapons research and development (R&D) facility to determine the risk of developing CBD in sensitized workers at facilities with exposures much lower than production plants. Of the 1875 current or former workers at the R&D facility, 59 were determined to be sensitized based on at least two positive BeLPTs (i.e., samples drawn on two separate occasions or on split samples tested in two separate DOE-approved laboratories) for a sensitization rate of 3.1 percent. Workers found to have positive BeLPTs were further evaluated in an Occupational Medicine Clinic between 1999 through 2005. Armojandi et al. (2010) evaluated 50 of the sensitized workers who also had medical and occupational histories, physical examination, chest imaging with high-resolution computed tomography (HRCT) (N = 49), and pulmonary function testing (nine of the 59 workers refused physical examinations so were not included in this study). Forty of the 50 workers chosen for this study underwent bronchoscopy for bronchoalveolar lavage and transbronchial biopsies in additional to the other testing. Five of the 49 workers had CBD at the time of evaluation (based on histology or high-resolution computed tomography); three others had evidence of probable CBD; however, none of these cases were classified as severe at the time of evaluation. The rate of CBD at the time of study among sensitized individuals was 12.5 percent (5/40) for those using pathologic review of lung tissue, and 10.2 percent (5/49) for those using HRCT as a criteria for diagnosis. The rate of CBD among the entire population (5/1875) was 0.3 percent.

The mean duration of employment at the facility was 18 years, and the mean latency period (from first possible exposure) to time of evaluation and diagnosis was 32 years. There was no available exposure monitoring in the breathing zone of workers at the facility but the beryllium levels were believed to be relatively low (possibly less than 0.1 μg/m³ for most jobs). There was not an apparent exposure-response relationship for sensitization or CBD. The sensitization prevalence was similar and the CBD prevalence higher among workers with the lower-exposure jobs. The authors concluded that these sensitized workers, who were subjected to an extended duration of low potential beryllium exposures over a long latency period, had a low prevalence of CBD (Armojandi et al., 2010).

i. Aluminum Smelting

Bauxite ore, the primary source of aluminum, contains naturally occurring beryllium. Worker exposure to beryllium can occur at aluminum smelting facilities where aluminum extraction occurs via electrolytic reduction of aluminum oxide into aluminum metal. Characterization of beryllium exposures and sensitization prevalence rates were examined by Taiwo et al. (2010) in a study of nine aluminum smelting facilities from four different companies in the U.S., Canada, Italy and Norway.

Of the 3,185 workers determined to be potentially exposed to beryllium, 1,932 agreed to participate in a medical surveillance program between 2000 and 2006 (60 percent participation rate). The medical surveillance program included serum BeLPT analysis, confirmation of an abnormal BeLPT with a second BeLPT, and follow-up of all confirmed positive responses by a pulmonary physician to evaluate for progression to CBD.

Eight-hour TWAs were assessed utilizing 1,345 personal samples collected from the 9 smelters. The personal beryllium samples obtained showed a range of 0.01-13.00 μg/m³ time-weighted average with an arithmetic mean of 0.25 μg/m³ and geometric mean of 0.06 μg/m³. Exposure levels to beryllium observed in aluminum smelters are similar to those seen in other industries that utilize beryllium. Of the 1,932 workers surveyed by BeLPT, nine workers were diagnosed with sensitization (prevalence rate of 0.47 percent, 95% confidence interval = 0.21-0.88 percent) with 2 of these workers diagnosed with probable CBD after additional medical evaluations.

The authors concluded that compared with beryllium-exposed workers in other industries, the rate of sensitization among aluminum smelter workers appears lower. The authors speculated that this lower observed rate could be related to a more soluble form of beryllium found in the aluminum smelting work environment as well as the consistent use of respiratory protection. However, the authors also speculated that the 60 percent participation rate may have underestimated the sensitization rate in this worker population.

A study by Nilsen et al. (2010) also found a low rate of sensitization among aluminum workers in Norway. Three-hundred sixty-two workers and thirty-one control individuals were tested for beryllium sensitization based on the BeLPT. The results found that one (0.28%) of the smelter workers had been sensitized. No borderline results were reported. The exposure estimated in this plant was 0.1 µg/m³ to 0.31 µg/m³ (Nilsen et al., 2010).

6. Animal Models of CBD

This section reviews the relevant animal studies supporting the mechanisms outlined above. Researchers have attempted to identify animal models with which to further investigate the mechanisms underlying the development of CBD. A suitable animal model should exhibit major characteristics of CBD, including the demonstration of a beryllium-specific immune response, the formation of immune granulomas following inhalation exposure to beryllium, and mimicking the progressive nature of the human disease. While exposure to beryllium has been shown to cause chronic granulomatous inflammation of the lung in animal studies using a variety of species, most of the granulomatous lesions were formed by foreign-body reactions, which result from persistent irritation and consist predominantly of macrophages and monocytes, and small numbers of lymphocytes. Foreign-body granulomas are distinct from the immune granulomas of CBD, which are caused by antigenic stimulation of the immune system and contain large numbers of lymphocytes. Animal studies have been useful in providing biological plausibility for the role of immunological alterations and lung inflammation and in clarifying certain specific mechanistic aspects of beryllium disease. However, the lack of a dependable animal model that mimics all facets of the human response combined with study limitations in terms of single dose experiments, few animals, or abbreviated observation periods have limited the utility of the data. Currently, no single model has completely mimicked the disease process as it progresses in humans. The following is a discussion of the most relevant animal studies regarding the mechanisms of sensitization and CBD development in humans. Table A.2 in the Appendix summarizes species, route, chemical form of beryllium, dose levels, and pathological findings of the key studies.

Harmsen et al. performed a study to assess whether the beagle dog could provide an adequate model for the study of beryllium-induced lung diseases (Harmsen et al., 1986). One group of dogs served as a control group (air inhalation only) and four other groups received high (approximately 50 μg/kg) and low (approximately 20 μg/kg) doses of beryllium oxide calcined at 500 °C or 1,000° C, administered as aerosols in a single exposure. As discussed above, calcining temperature controls the solubility and SSA of beryllium particles. Those particles calcined at higher temperatures (e.g., 1,000° C) are less soluble and have lower SSA than particles calcined at lower temperatures (e.g., 500 °C). Solubility and SSA are factors in determining the toxic potential of beryllium compounds or materials.

Cells were collected from the dogs by BAL at 30, 60, 90, 180, and 210 days after exposure, and the percentages of neutrophils and lymphocytes were determined. In addition, the mitogenic responses of blood lymphocytes and lavage cells collected at 210 days were determined with either phytohemagglutinin or beryllium sulfate as mitogen. The percentage of neutrophils in the lavage fluid was significantly elevated only at 30 days with exposure to either dose of 500 °C beryllium oxide. The percentage of lymphocytes in the fluid was significantly elevated in samples across all times with exposure to the high dose of this beryllium oxide form. Beryllium oxide calcined at 1,000° C elevated lavage lymphocytes only in high dose at 30 days. No significant effect of 1,000° C beryllium oxide exposure on mitogenic response of any lymphocytes was seen. In contrast, peripheral blood lymphocytes from the 500 °C beryllium oxide exposed groups were significantly stimulated by beryllium sulfate compared with the phytohemagglutinin exposed cells. The investigators in this study were able to replicate some of the same findings as those observed in human studies—specifically, that beryllium in soluble and insoluble forms can be mitogenic to immune cells, an important finding for progression of sensitization and proliferation of immune cells to developing full-blown CBD.

In another beagle study Haley et al. also found that the beagle dog appears to model some aspects of human CBD (Haley et al., 1989). The authors monitored lung pathologic effects, particle clearance, and immune sensitization of peripheral blood leukocytes following a single exposure to beryllium oxide aerosol generated from beryllium oxide calcined at 500 °C or 1,000° C. The aerosol was administered to the dogs perinasally to attain initial lung burdens of 6 or 18 μg beryllium/kg body weight. Granulomatous lesions and lung lymphocyte responses consistent with those observed in humans with CBD were observed, including perivascular and peribronchiolar infiltrates of lymphocytes and macrophages, progressing to microgranulomas with areas of granulomatous pneumonia and interstitial fibrosis. Beryllium specificity of the immune response was demonstrated by positive results in the BeLPT, although there was considerable inter-animal variation. The lesions declined in severity after 64 days post-exposure. Thus, while this model was able to mimic the formation of Be-specific immune granulomas, it was not able to mimic the progressive nature of disease.

This study also provided an opportunity to compare the effects of beryllium oxide calcination temperature on granulomatous disease in the beagle respiratory system. Haley et al. found an increase in the percentage and numbers of lymphocytes in BAL fluid at 3 months post-exposure in dogs exposed to either dose of beryllium oxide calcined at 500 °C, but not in dogs exposed to the material calcined at the higher temperature. Although there was considerable inter-animal variation, lesions were generally more severe in the dogs exposed to material calcined at 500 °C. Positive BeLPT results were observed with BAL lymphocytes only in the group with a high initial lung burden of the material calcined at 500 °C, but positive results with peripheral blood lymphocytes were observed at both doses with material calcined at both temperatures.

The histologic and immunologic responses of canine lungs to aerosolized beryllium oxide were investigated in another Haley et al. (1989) study. Beagle-dogs were exposed in a single exposure to high dose (50 µg/kg of body weight) or low dose (l7 µg/kg) levels of beryllium oxide calcined at either 500° or 1000° C. One group of dogs was examined up to 365 days after exposure for lung histology and biochemical assay to determine the fate of inhaled beryllium oxide. A second group underwent BAL for lung lymphocyte analysis for up to 22 months after exposure. Histopathologic examination revealed peribronchiolar and perivascular lymphocytic histiocytic inflammation, peaking at 64 days after beryllium oxide exposure. Lymphocytes were initially well differentiated, but progressed to lymphoblastic cells and aggregated in lymphofollicular nodules or microgranulomas over time. Alveolar macrophages were large, and filled with intracytoplasmic material. Cortical and paracortical lymphoid hyperplasia of the tracheobronchial nodes was found. Lung lymphocyte concentrations were increased at 3 months and returned to normal in both dose groups given 500 °C treated beryllium chloride. No significant elevations in lymphocyte concentrations were found in dogs given 1,000° C treated beryllium oxide. Lung retention was higher in the 500 °C treated beryllium oxide group. The lesions found in dog lungs closely resembled those found in humans with CBD: severe granulomas, lymphoblast transformation, increased pulmonary lymphocyte concentrations and variation in beryllium sensitivity. It was concluded that the canine model for berylliosis may provide insight into this disease.

In a follow-up experiment, control dogs and those exposed to beryllium oxide calcined at 500 °C were allowed to rest for 2.5 years, and then re-exposed to filtered air (controls) or beryllium oxide calcined at 500 °C for an initial lung burden (ILB) target of 50 μg beryllium oxide/kg body weight (Haley et al., 1992). Immune responses of blood and BAL lymphocytes, and lung lesions in dogs sacrificed 210 days post-exposure, were compared with results following the initial exposure. The severity of lung lesions was comparable under both conditions, suggesting that a 2.5-year interval was sufficient to prevent cumulative pathologic effects. Conradi et al. (1971) found no exposure-related histological alterations in the lungs of six beagle dogs exposed to a range of 3,300-4,380 μg Be/m³ as beryllium oxide calcined at 1,400° C for 30 min, once per month for 3 months. Because the dogs were sacrificed 2 years post-exposure, the long time period between exposure and response may have allowed for the reversal of any beryllium-induced changes (EPA, 1998).

A 1994 study by Haley et al. showed that intra-bronchiolar instillation of beryllium induced immune granulomas and sensitization in monkeys. Haley et al. (1994) exposed male cynomolgus monkeys to either beryllium metal or beryllium oxide calcined at 500 °C by intrabronchiolar instillation as a saline suspension. Lymphocyte counts in BAL fluid were observed, and were found to be significantly increased in monkeys exposed to beryllium metal on post-exposure days 14 to 90, and on post-exposure day 60 in monkeys exposed to beryllium oxide. The lungs of monkeys exposed to beryllium metal had lesions characterized by interstitial fibrosis, Type II cell hyperplasia, and lymphocyte infiltration. Some monkeys also exhibited immune granulomas. Similar lesions were observed in monkeys exposed to beryllium oxide, but the incidence and severity were much less. BAL lymphocytes from monkeys exposed to beryllium metal, but not from monkeys exposed to beryllium oxide, proliferated in response to beryllium sulfate in the BeLPT (EPA, 1998).

In an experiment similar to the one conducted with dogs, Conradi et al. (1971) found no effect in monkeys (Macaca irus) exposed via whole-body inhalation for three 30-minute monthly exposures to a range of 3,300-4,380 μg Be/m³ as beryllium oxide calcined at 1,400° C. The lack of effect may have been related to the long period (2 years) between exposure and sacrifice, or to low toxicity of beryllium oxide calcined at such a high temperature.

As discussed earlier in this Health Effects section, at the cellular level, beryllium dissolution must occur for either a dendritic cell or a macrophage to present beryllium as an antigen to induce the cell-mediated CBD immune reactions (Stefaniak et al., 2006). Several studies have shown that low-fired beryllium oxide, which is predominantly made up of poorly crystallized small particles, is more immunologically reactive than beryllium oxide calcined at higher firing temperatures that result in less reactivity due to increasing crystal size. As discussed previously, Haley et al. (1989a) found more severe lung lesions and a stronger immune response in beagle dogs receiving a single inhalation exposure to beryllium oxide calcined at 500 °C than in dogs receiving an equivalent initial lung burden of beryllium oxide calcined at 1,000° C. Haley et al. found that beryllium oxide calcined at 1,000° C elicited little local pulmonary immune response, whereas the much more soluble beryllium oxide calcined at 500 °C produced a beryllium-specific, cell-mediated immune response in dogs (Haley et al., 1991).

In a later study, beryllium metal appeared to induce a greater toxic response than beryllium oxide following intrabronchiolar instillation in cynomolgus monkeys, as evidenced by more severe lung lesions, a larger effect on BAL lymphocyte counts, and a positive response in the BeLPT with BAL lymphocytes only after exposure to beryllium metal (Haley et al., 1994). Because an oxide layer may form on beryllium-metal surfaces after exposure to air (Mueller and Adolphson, 1979; Harmsen et al., 1986) dissolution of small amounts of poorly soluble beryllium compounds in the lungs might be sufficient to allow persistent low-level beryllium presentation to the immune system (NAS, 2008).

Genetic studies in humans led to the creation of an animal model containing different human HLA-DP alleles inserted into FVB/N mice for mechanistic studies of CBD. Three strains of genetically engineered mice (transgenic mice) were created that conferred different risks for developing CBD based on human studies (Weston et al., 2005; Snyder et al., 2008): (1) the HLDPB1*401 transgenic strain, where the transgene codes for lysine residue at the 69th position of the B-chain conferred low risk of CBD; (2) the HLA-DPB1*201 mice, where the transgene codes for glutamic acid residue at the 69th position of the B-chain and glycine residues at positions 84 and 85 conferred medium risk of CBD; and (3) the HLA-DPB1*1701 mice, where the transgene codes for glutamic acid at the 69th position of the B-chain and aspartic acid and glutamic acid residues at positions 84 and 85, respectively, conferred high risk of CBD (Tarantino-Hutchinson et al., 2009).

In order to validate the transgenic model, Tarantino-Hutchison et al. challenged the transgenic mice along with seven different inbred mouse strains to determine the susceptibility and sensitivity to beryllium exposure. Mice were dermally exposed with either saline or beryllium, then challenged with either saline or beryllium (as beryllium sulfate) using the MEST protocol (mouse ear-swelling test). The authors determined that the high risk HLA-DPB1*1701 transgenic strain responded 4 times greater (as measured via ear swelling) than control mice and at least 2 times greater than other strains of mice. The findings correspond to epidemiological study results reporting an enhanced CBD odds ratio for the HLA-DPB1*1701 in humans (Weston et al., 2005; Snyder et al., 2008). Transgenic mice with the genes corresponding to the low and medium odds ratio study did not respond significantly over the control group. The authors concluded that while HLA-DPB1*1701 is important to beryllium sensitization and progression to CBD, other genetic and environmental factors contribute to the disease process as well.

7. Preliminary Beryllium Sensitization and CBD Conclusions

It is well-established that skin and inhalation exposure to beryllium may lead to sensitization and that inhalation exposure, or skin exposure coupled with inhalation exposure, may lead to the onset and progression of CBD. This is supported by extensive human studies. While all facets of the biological mechanism for this complex disease have yet to be fully elucidated, many of the key events in the disease sequence have been identified and described in the previous sections. Sensitization is a necessary first step to the onset of CBD (NAS, 2008). Sensitization is the process by which the immune system recognizes beryllium as a foreign substance and responds in a manner that may lead to development of CBD. It has been documented that a substantial proportion of sensitized workers exposed to airborne beryllium progress to CBD (Rosenman et al., 2005; NAS, 2008; Mroz et al., 2009). Animal studies, particularly in dogs and monkeys, have provided supporting evidence for T-cell lymphocyte proliferation in the development of granulomatous lung lesions after exposure to beryllium (Harmsen et al., 1986; Haley et al., 1989, 1992, 1994). The animal studies have also provided important insights into the roles of chemical form, genetic susceptibility, and residual lung burden in the development of beryllium lung disease (Harmsen et al., 1986; Haley et al., 1992; Tarantino-Hutchison et al., 2009). OSHA has made a preliminary determination to consider sensitization and CBD to be adverse events along the pathological continuum in the disease process, with sensitization being the necessary first step in the progression to CBD.

The epidemiological evidence presented in this section demonstrates that sensitization and CBD are continuing to occur from present-day exposures below OSHA's PEL (Rosenman, 2005 with erratum published 2006). The available literature discussed above shows that disease prevalence can be reduced by reducing inhalation exposure (Thomas et al., 2009). However, the available epidemiological studies also indicate that it may be necessary to minimize skin exposure to further reduce the incidence of sensitization (Bailey et al., 2010). The preliminary risk assessment further discusses the effectiveness of interventions to reduce beryllium exposures and the risk of sensitization and CBD (see section VI, Preliminary Risk Assessment).

Studies have demonstrated there remains a prevalence of sensitization and CBD in facilities with exposure levels below the current OSHA PEL (Rosenman et al., 2005; Thomas et al., 2009), that risk of sensitization and CBD appears to vary across industries and processes (Deubner et al., 2001; Kreiss et al., 1997; Newman et al., 2001; Henneberger et al., 2001; Schuler et al., 2005; Stange et al., 2001; Taiwo et al., 2010), and that efforts to reduce exposure have succeeded in reducing the frequency of beryllium sensitization and CBD (Bailey et al., 2010) (See Table A-1 in the Appendix).

Of workers who were found to be sensitized and underwent clinical evaluation, 20-49 percent were diagnosed with CBD (Kreiss et al., 1993; Newman, 1996, 2005 and 2007; Stange et al., 2001). Overall prevalence of CBD in cross-sectional screenings ranges from 0.6 to 8 percent (Kreiss et al., 2007). A study by Newman (2005) estimated from ongoing surveillance of sensitized individuals, with an average follow-up time of 6 years, that 31 percent of beryllium-exposed employees progressed to CBD (Newman, 2005). However, Newman (2005) went on to suggest that if follow-up times were increased the rate of progression from sensitization to CBD could be much higher. A study of nuclear weapons facility employees enrolled in an ongoing medical surveillance program found that only about 20 percent of sensitized individuals employed less than five years eventually were diagnosed with CBD, while 40 percent of sensitized employees employed ten years or more developed CBD (Stange et al., 2001) indicating length of exposure may play a role in further development of the disease. In addition, Mroz et al. (2009) conducted a longitudinal study of individuals clinically evaluated at National Jewish Health (between 1982 and 2002) who were identified as having sensitization and CBD through workforce medical surveillance. The authors identified 171 cases of CBD and 229 cases of sensitization; all individuals were identified through workplace screening using the BeLPT (Mroz et al., 2009). Over the 20-year study period, 8.8 percent (i.e., 22 cases out 251 sensitized) of individuals with sensitization went on to develop CBD. The findings from this study indicated that on the average span of time from initial beryllium exposure to CBD diagnosis was 24 years (Mroz et al., 2009).

E. Beryllium Lung Cancer Section

Beryllium exposure has been associated with a variety of adverse health effects including lung cancer. The potential for beryllium and its compounds to cause cancer has been previously assessed by various other agencies (EPA, ATSDR, NAS, NIEHS, and NIOSH) with each agency identifying beryllium as a potential carcinogen. In addition, the International Agency for Research on Cancer (IARC) did an extensive evaluation in 1993 and reevaluation in April 2009 (IARC, 2012). In brief, IARC determined beryllium and its compounds to be carcinogenic to humans (Group 1 category), while EPA considers beryllium to be a probable human carcinogen (EPA, 1998), and the National Toxicology Program (NTP) has determined beryllium and its compounds to be known carcinogens (NTP, 2014). OSHA has conducted an independent evaluation of the carcinogenic potential of beryllium and these compounds as well. The following is a summary of the studies used to support the Agency findings that beryllium and its compounds are human carcinogens.

1. Genotoxicity Studies

Genotoxicity can be an important indicator for screening the potential of a material to induce cancer and an important mechanism leading to tumor formation and carcinogenesis. In a review conducted by the National Academy of Science, beryllium and its compounds have tested positively in nearly 50 percent of the genotoxicity studies conducted without exogenous metabolic activity. However, they were found to be non-genotoxic in most bacterial assays (NAS, 2008).

Gene mutations have been observed in mammalian cells cultured with beryllium chloride in a limited number of studies (EPA, 1998; ATSDR, 2002; Gordon and Bowser, 2003). Culturing mammalian cells with beryllium chloride, beryllium sulfate, or beryllium nitrate has resulted in clastogenic alterations. However, most studies have found that beryllium chloride, beryllium nitrate, beryllium sulfate, and beryllium oxide did not induce gene mutations in bacterial assays with or without metabolic activation. In the case of beryllium sulfate, all mutagenicity studies (Ames (Simmon, 1979; Dunkel et al., 1984; Arlauskas et al., 1985; Ashby et al., 1990); E. coli pol A (Rosenkranz and Poirer, 1979); E. coli WP2 uvr A (Dunkel et al., 1984) and Saccharomyces cerevisiae (Simmon, 1979)) were negative with the exception of results reported for Bacillus subtilis rec assay (Kada et al., 1980; Kanematsu et al., 1980; EPA, 1998). Beryllium sulfate did not induce unscheduled DNA synthesis in primary rat hepatocytes and was not mutagenic when injected intraperitoneally in adult mice in a host-mediated assay using Salmonella typhimurium (Williams et al., 1982).

Beryllium nitrate was negative in the Ames assay (Tso and Fung, 1981; Kuroda et al., 1991) but positive in a Bacillus subtilis rec assay (Kuroda et al., 1991). Beryllium chloride was negative in a variety of studies (Ames (Ogawa et al., 1987; Kuroda et al., 1991); E. coli WP2 uvr A (Rossman and Molina, 1984); and Bacillus subtilis rec assay (Nishioka, 1975)). In addition, beryllium chloride failed to induce SOS DNA repair in E. coli (Rossman et al., 1984). However, positive results were reported for Bacillus subtilis rec assay using spores (Kuroda et al., 1991), E. coli KMBL 3835; lacI gene (Zakour and Glickman, 1984), and hprt locus in Chinese hamster lung V79 cells (Miyaki et al., 1979). Beryllium oxide was negative in the Ames assay and Bacillus subtilis rec assays (Kuroda et al., 1991; EPA, 1998).

Gene mutations have been observed in mammalian cells (V79 and CHO) cultured with beryllium chloride (Miyaki et al., 1979; Hsie et al., 1979a, b), and culturing of mammalian cells with beryllium chloride (Vegni-Talluri and Guiggiani, 1967), and beryllium sulfate (Brooks et al., 1989; Larramendy et al., 1981) has resulted in clastogenic alterations—producing breakage or disrupting chromosomes (EPA, 1998). Beryllium chloride evaluated in a mouse model indicated increased DNA strand breaks and the formation of micronuclei in bone marrow (Attia et al., 2013).

Data on the in vivo genotoxicity of beryllium are limited to a single study that found beryllium sulfate (1.4 and 2.3 g/kg, 50 percent and 80 percent of median lethal dose) administered by gavage did not induce micronuclei in the bone marrow of CBA mice. However, a marked depression of erythropoiesis (red blood cell production) was suggestive of bone marrow toxicity which was evident 24 hours after dosing. No mutations were seen in p53 or c-raf-1 and only weak mutations were detected in K-ras in lung carcinomas from F344/N rats given a single nose-only exposure to beryllium metal (Nickell-Brady et al., 1994). The authors concluded that the mechanisms for the development of lung carcinomas from inhaled beryllium in the rat do not involve gene dysfunctions commonly associated with human non-small-cell lung cancer (EPA, 1998).

2. Human Epidemiological Studies

This section reviews in greater detail the studies used to support the mechanistic findings for beryllium-induced cancer. Table A.3 in the Appendix summarizes the important features and characteristics of each study.

a. Beryllium Case Registry (BCR).

Two studies evaluated participants in the BCR (Infante et al., 1980; Steenland and Ward, 1991). Infante et al. (1980) evaluated the mortality patterns of white male participants in the BCR diagnosed with non-neoplastic respiratory symptoms of beryllium disease. Of the 421 cases evaluated, 7 of the participants had died of lung cancer. Six of the deaths occurred more than 15 years after initial beryllium exposure. The duration of exposure for 5 of the 7 participants with lung cancer was less than 1 year, with the time since initial exposure ranging from 12 to 29 years. One of the participants was exposed for 4 years with a 26-year interval since the initial exposure. Exposure duration for one participant diagnosed with pulmonary fibrosis could not be determined; however, it had been 32 years since the initial exposure. Based on BCR records, the participants were classified as being in the acute respiratory group (i.e., those diagnosed with acute respiratory illness at the time of entry in the registry) or the chronic respiratory group (i.e., those diagnosed with pulmonary fibrosis or some other chronic lung condition at the time of entry into the BCR). The 7 participants with lung cancer were in the BCR because of diagnoses of acute respiratory illness. For only one of those individuals was initial beryllium exposure less than 15 years prior. Only 1 of the 6 (with greater than 15 years since initial exposure to beryllium) had been diagnosed with chronic respiratory disease. The study did not report exposure concentrations or smoking habits. The authors concluded that the results of this cohort agreed with previous animal studies and with epidemiological studies demonstrating an increased risk of lung cancer in workers exposed to beryllium.

Steenland and Ward (1991) extended the work of Infante et al. (1980) to include females and to include 13 additional years of follow-up. At the time of entry in the BCR, 93 percent of the women in the study, but only 50 percent of the men, had been diagnosed with CBD. In addition, 61 percent of the women had worked in the fluorescent tube industry and 50 percent of the men had worked in the basic manufacturing industry. A total of 22 males and 6 females died of lung cancer. Of the 28 total deaths from lung cancer, 17 had been exposed to beryllium for less than 4 years and 11 had been exposed for greater than 4 years. The study did not report exposure concentrations. Survey data collected in 1965 provided information on smoking habits for 223 cohort members (32 percent), on the basis of which the authors suggested that the rate of smoking among workers in the cohort may have been lower than U.S. rates. The authors concluded that there was evidence of increased risk of lung cancer in workers exposed to beryllium and diagnosed with beryllium disease.

b. Beryllium Manufacturing and/or Processing Plants (Extraction, Fabrication, and Processing)

Several epidemiological cohort studies have reported excess lung cancer mortality among workers employed in U.S. beryllium production and processing plants during the 1930s to 1960s. The largest and most comprehensive study investigated the mortality experience of 9,225 workers employed in seven different beryllium processing plants over a 30-year period (Ward et al., 1992). The workers at the two oldest facilities (i.e., Lorain, OH, and Reading, PA) were found to have significant excess lung cancer mortality relative to the U.S. population. Of the seven plants in the study, these two plants were believed to have the highest exposure levels to beryllium. A different analysis of the lung cancer mortality in this cohort using various local reference populations and alternate adjustments for smoking generally found smaller, non-significant rates of excess mortality among the beryllium employees (Levy et al., 2002). Both cohort studies are limited by a lack of job history and air monitoring data that would allow investigation of mortality trends with beryllium exposure. The majority of employees at the Lorain, OH, and Reading, PA, facilities were employed for a relatively short period of less than one year.

Bayliss et al. (1971) performed a nested cohort study of more than 7,000 former workers from the beryllium processing industry employed from 1942-1967. Information for the workers was collected from the personnel files of participating companies. Of the more than 7,000 employees, a cause of death was known for 753 male workers. The number of observed lung cancer deaths was 36 compared to 34.06 expected for a standardized mortality ratio (SMR) of 1.06. When evaluated by the number of years of employment, 24 of the 36 men were employed for less than 1 year in the industry (SMR = 1.24), 8 were employed for 1 to 5 years (SMR 1.40), and 4 were employed for more than 5 years (SMR = 0.54). Half of the workers who died from lung cancer began employment in the beryllium production industry prior to 1947. When grouped by job classification, over two thirds of the workers with lung cancer were in production-related jobs while the rest were classified as office workers. The authors concluded that while the lung cancer mortality rates were the highest of all other mortality rates, the SMR for lung cancer was still within range of the expected based on death rates in the United States. The limitations of this study included the lack of information regarding exposure concentrations, smoking habits, and the age and race of the participants.

Mancuso (1970, 1979, 1980) and Mancuso and El-Attar (1969) performed a series of occupational cohort studies on a group of over 3,685 workers (primarily white males) employed in the beryllium manufacturing industry during 1937-1948. The beryllium production facilities were located in Ohio and Pennsylvania and the records for the employees, including periods of employment, were obtained from the Social Security Administration. These studies did not include analyses of mortality by job title or exposure category. In addition, there were no exposure concentrations estimated or adjustments for smoking. The estimated duration of employment ranged from less than 1 year to greater than 5 years. In the most recent study (Mancuso, 1980), employees from the viscose rayon industry served as a comparison population. There was a significant excess of lung cancer deaths based on the total number of 80 observed lung cancer mortalities at the end of 1976 compared to an expected number of 57.06 based on the comparison population resulting in an SMR of 1.40 (p < 0.01) (Mancuso, 1980). There was a statistically significant excess in lung cancer deaths for the shortest duration of employment (< 12 months, p < 0.05) and the longest duration of employment (> 49 months, p < 0.01). Based on the results of this study, the author concluded that the ability of beryllium to induce cancer in workers does not require continuous exposure and that it is reasonable to assume that the amount of exposure required to produce lung cancer can occur within a few months of exposure regardless of the length of employment.

Wagoner et al. (1980) expanded the work of Mancuso (1970; 1979; 1980) using a cohort of 3,055 white males from the beryllium extraction, processing, and fabrication facility located in Reading, Pennsylvania. The men included in the study worked at the facility sometime between 1942 and 1968, and were followed through 1976. The study accounted for length of employment. Other factors accounted for included age, smoking history, and regional lung cancer mortality. Forty-seven members of the cohort died of lung cancer compared to an expected 34.29 based on U.S. white male lung cancer mortality rates (p < .05). The results of this cohort showed an excess risk of lung cancer in beryllium-exposed workers at each duration of employment (< 5 years and ≥ 5 years), with a statistically significant excess noted at < 5 years durations of employment and a ≥ 25-year interval since the beginning of employment (p < 0.05). The study was criticized by several epidemiologists (MacMahon, 1978, 1979; Roth, 1983), by a CDC Review Committee appointed to evaluate the study, and by one of the study's coauthors (Bayliss, 1980) for inadequate discussion of possible alternative explanations of excess lung cancer in the cohort. The specific issues identified include the use of 1965-1967 U.S. white male lung cancer mortality rates to generate expected numbers of lung cancers in the period 1968-1975 and inadequate adjustment for smoking.

Ward et al. (1992) performed a retrospective mortality cohort study of 9,225 male workers employed at seven beryllium processing facilities, including the Ohio and Pennsylvania facilities studied by Mancuso and El-Attar (1969), Mancuso (1970; 1979; 1980), and Wagoner et al. (1980). The men were employed for no less than 2 days between January 1940 and December 1988. At the end of the study 61.1 percent of the cohort was known to be living and 35.1 percent was known to be deceased. The duration of employment ranged from 1 year or less to greater than 10 years with the largest percentage of the cohort (49.7 percent) employed for less than one year, followed by 1 to 5 years of employment (23.4 percent), greater than 10 years (19.1 percent), and 5 to 10 years (7.9 percent). Of the 3,240 deaths, 280 observed deaths were caused by lung cancer compared to 221.5 expected deaths, yielding a statistically significant SMR of 1.26 (p < 0.01). Information on the smoking habits of 15.9 percent of the cohort members, obtained from a 1968 Public Health Service survey conducted at four of the plants, was used to calculate a smoking-adjusted SMR of 1.12, which was not statistically significant. The number of deaths from lung cancer was also examined by decade of hire. The authors reported a relationship between earlier decades of hire and increased lung cancer risk.

The EPA Integrated Risk Information System (IRIS), IARC, and California EPA Office of Environmental Health Hazard Assessment (OEHHA) have all based their cancer assessment on the Ward et al. 1992 study, with supporting data concerning exposure concentrations from Eisenbud and Lisson (1983) and NIOSH (1972), who estimated that the lower-bound estimate of the median exposure concentration exceeded 100 µg/m³ and found that concentrations in excess of 1,000 µg/m³ were common. The IRIS cancer risk assessment recalculated expected lung cancers based on U.S. white male lung cancer rates (including the period 1968-1975) and used an alternative adjustment for smoking. In addition, one individual with lung cancer, who had not worked at the plant, was removed from the cohort. After these adjustments were made, an elevated rate of lung cancer was still observed in the overall cohort (46 cases vs. 41.9 expected cases). However, based on duration of employment or interval since beginning of employment, neither the total cohort nor any of the subgroups had a statistically significant excess in lung cancer (EPA, 1987). Based on their evaluation of this and other epidemiological studies, the EPA characterized the human carcinogenicity data then available as “limited” but “suggestive of a causal relationship between beryllium exposure and an increased risk of lung cancer” (IRIS database). This report includes quantitative estimates of risk that were derived using the information presented in Wagoner et al. (1980), the expected lung cancers recalculated by the EPA, and bounds on presumed exposure levels.

Levy et al. (2002) questioned the results of Ward et al. (1992) and performed a reanalysis of the Ward et al. data. The Levy et al. reanalysis differed from the Ward et al. analysis in the following significant ways. First, Levy et al. (2002) examined two alternative adjustments for smoking, which were based on (1) a different analysis of the American Cancer Society (ACS) data used by Ward et al. (1992) for their smoking adjustment, or (2) results from a smoking/lung cancer study of veterans (Levy and Marimont, 1998). Second, Levy et al. (2002) also examined the impact of computing different reference rates derived from information about the lung cancer rates in the cities in which most of the workers at two of the plants lived. Finally, Levy et al. (2002) considered a meta-analytical approach to combining the results across beryllium facilities. For all of the alternatives Levy et al. (2002) considered, except the meta-analysis, the facility-specific and combined SMRs derived were lower than those reported by Ward et al. (1992). Only the SMR for the Lorain, OH, facility remained statistically significantly elevated in some reanalyses. The SMR obtained when combining over the plants was not statistically significant in eight of the nine approaches they examined, leading Levy et al. (2002) to conclude that there was little evidence of statistically significant elevated SMRs in those plants.

One occupational nested case-control study evaluated lung cancer mortality in a cohort of 3,569 male workers employed at a beryllium alloy production plant in Reading, PA, from 1940 to 1969 and followed through 1992 (Sanderson et al., 2001). There were a total of 142 known lung cancer cases and 710 controls. For each lung cancer death, 5 age- and race-matched controls were selected by incidence density sampling. Confounding effects of smoking were evaluated. Job history and historical air measurements at the plant were used to estimate job-specific beryllium exposures from the 1930s to 1990s. Calendar-time-specific beryllium exposure estimates were made for every job and used to estimate workers' cumulative, average, and maximum exposure. Because of the long period of time required for the onset of lung cancer, an “exposure lag” was employed to discount recent exposures less likely to contribute to the disease.

The cumulative, average, and maximum beryllium exposure concentration estimates for the 142 known lung cancer cases were 46.06 ± 9.3µg/m³-days, 22.8 ± 3.4 µg/m³, and 32.4 ± 13.8 µg/m³, respectively. The lung cancer mortality rate was 1.22 (95 percent CI = 1.03 − 1.43). Exposure estimates were lagged by 10 and 20 years in order to account for exposures that did not contribute to lung cancer because they occurred after the induction of cancer. In the 10- and 20-year lagged exposures the geometric mean tenures and cumulative exposures of the lung cancer mortality cases were higher than the controls. In addition, the geometric mean and maximum exposures of the workers were significantly higher than controls when the exposure estimates were lagged 10 and 20 years (p < 0.01).

Results of a conditional logistic regression analysis indicated that there was an increased risk of lung cancer in workers with higher exposures when dose estimates were lagged by 10 and 20 years. There was also a lack of evidence that confounding factors such as smoking affected the results of the regression analysis. The authors noted that there was considerable uncertainty in the estimation of exposure in the 1940's and 1950's and the shape of the dose-response curve for lung cancer. Another analysis of the study data using a different statistical method did not find a significantly greater relative risk of lung cancer with increasing beryllium exposures (Levy et al., 2007). The average beryllium air levels for the lung cancer cases were estimated to be an order of magnitude above the current 8-hour OSHA TWA PEL (2 μg/m³) and roughly two orders of magnitude higher than the typical air levels in workplaces where beryllium sensitization and pathological evidence of CBD have been observed. IARC evaluated this reanalysis in 2012 and found the study introduced a downward bias into risk estimates (IARC, 2012).

Schubauer-Berigan et al. reanalyzed data from the nested case-control study of 142 lung cancer cases in the Reading, PA, beryllium processing plant (Schubauer-Berigan et al., 2008). This dataset was reanalyzed using conditional (stratified by case age) logistic regression. Independent adjustments were made for potential confounders of birth year and hire age. Average and cumulative exposures were analyzed using the values reported in the original study. The objective of the reanalysis was to correct for the known differences in smoking rates by birth year. In addition, the authors evaluated the effects of age at hire to determine differences observed by Sanderson et al. in 2001. The effect of birth cohort adjustment on lung cancer rates in beryllium-exposed workers was evaluated by adjusting in a multivariable model for indicator variables for the birth cohort quartiles.

Unadjusted analyses showed little evidence of lung cancer risk associated with beryllium occupational exposure using cumulative exposure until a 20-year lag was used. Adjusting for either birth cohort or hire age attenuated the risk for lung cancer associated with cumulative exposure. Using a 10- or 20-year lag in workers born after 1900 also showed little evidence of lung cancer risk, while those born prior to 1900 did show a slight elevation in risk. Unlagged and lagged analysis for average exposure showed an increase in lung cancer risk associated with occupational exposure to beryllium. The finding was consistent for either workers adjusted or unadjusted for birth cohort or hire age. Using a 10-year lag for average exposure showed a significant effect by birth cohort.

The authors stated that the reanalysis indicated that differences in the hire ages among cases and controls, first noted by Deubner et al. (2001) and Levy et al. (2007), were primarily due to the fact that birth years were earlier among controls than among cases, resulting from much lower baseline risk of lung cancer for men born prior to 1900 (Schubauer-Berigan et al., 2008). The authors went on to state that the reanalysis of the previous NIOSH case-control study suggested the relationship observed previously between cumulative beryllium exposure and lung cancer was greatly attenuated by birth cohort adjustment.

Hollins et al. (2009) re-examined the weight of evidence of beryllium as a lung carcinogen in a recent publication (Hollins et al., 2009). Citing more than 50 relevant papers, the authors noted the methodological shortcomings examined above, including lack of well-characterized historical occupational exposures and inadequacy of the availability of smoking history for workers. They concluded that the increase in potential risk of lung cancer was observed among those exposed to very high levels of beryllium and that beryllium's carcinogenic potential in humans at these very high exposure levels were not relevant to today's industrial settings. IARC performed a similar re-evaluation in 2009 (IARC, 2012) and found that the weight of evidence for beryllium lung carcinogenicity, including the animal studies described below, still warranted a Group I classification, and that beryllium should be considered carcinogenic to humans.

Schubauer-Berigan et al. (2010) extended their analysis from a previous study estimating associations between mortality risk and beryllium exposure to include workers at 7 beryllium processing plants. The study (Schubauer-Berigan et al., 2010) followed the mortality incidences of 9,199 workers from 1940 through 2005 at the 7 beryllium plants. JEMs were developed for three plants in the cohort: The Reading plant, the Hazleton plant, and the Elmore plant. The last is described in Couch et al. 2010. Including these JEMs substantially improved the evidence base for evaluating the carcinogenicity of beryllium and, and this change represents more than an update of the beryllium cohort. Standardized mortality ratios (SMRs) were estimated based on US population comparisons for lung, nervous system and urinary tract cancers, chronic obstructive pulmonary disease (COPD), chronic kidney disease, and categories containing chronic beryllium disease (CBD) and cor pulmonale. Associations with maximum and cumulative exposure were calculated for a subset of the workers.

Overall mortality in the cohort compared with the US population was elevated for lung cancer (SMR 1.17; 95% CI 1.08 to 1.28), COPD (SMR 1.23; 95% CI 1.13 to 1.32), and the categories containing CBD (SMR 7.80; 95% CI 6.26 to 9.60) and cor pulmonale (SMR 1.17; 95% CI 1.08 to 1.26). Mortality rates for most diseases of interest increased with time-since-hire. For the category including CBD, rates were substantially elevated compared to the US population across all exposure groups. Workers whose maximum beryllium exposure was ≥ 10 μg/m³ had higher rates of lung cancer, urinary tract cancer, COPD and the category containing cor pulmonale than workers with lower exposure. These studies showed strong associations for cumulative exposure (when short-term workers were excluded), maximum exposure or both. Significant positive trends with cumulative exposure were observed for nervous system cancers (p = 0.0006) and, when short-term workers were excluded, lung cancer (p = 0.01), urinary tract cancer (p = 0.003) and COPD (p < 0.0001).

The authors concluded the findings from this reanalysis reaffirmed that lung cancer and CBD are related to beryllium exposure. The authors went on to suggest that beryllium exposures may be associated with nervous system and urinary tract cancers and that cigarette smoking and other lung carcinogens were unlikely to explain the increased incidences in these cancers. The study corrected an error that was discovered in the indirect smoking adjustment initially conducted by Ward et al., concluding that cigarette smoking rates did not differ between the cohort and the general U.S. population. No association was found between cigarette smoking and either cumulative or maximum beryllium exposure, making it very unlikely that smoking was a substantial confounder in this study (Schubauer-Berigan et al., 2010).

3. Animal Cancer Studies

This section reviews the animal literature used to support the findings for beryllium-induced lung cancer. Lung tumors have been induced via inhalation and intratracheal administration of beryllium to rats and monkeys, and osteosarcomas have been induced via intravenous and intramedullary (inside the bone) injection of beryllium in rabbits and possibly in mice. The chronic oral studies did not report increased incidences of tumors in rodents, but these were conducted at doses below the maximum tolerated dose (MTD) (EPA, 1998).

Early animal studies revealed that some beryllium compounds are carcinogenic when inhaled (ATSDR, 2002). Animal experiments have shown consistent increases in lung cancers in rats, mice and rabbits chronically exposed to beryllium and beryllium compounds by inhalation or intratracheal instillation. In addition to lung cancer, osteosarcomas have been produced in mice and rabbits exposed to various beryllium salts by intravenous injection or implantation into the bone (NTP, 1999).

In an inhalation study assessing the potential tumorigenicity of beryllium, Schepers et al. (1957) exposed 115 albino Sherman and Wistar rats (male and female) via inhalation to 0.0357 mg beryllium/m (1 γ beryllium/ft) as an aqueous aerosol of beryllium sulfate for 44 hours/week for 6 months, and observed the rats for 18 months after exposure. Three to four control rats were killed every two months for comparison purposes. Seventy-six lung neoplasms, including adenomas, squamous-cell carcinomas, acinous adenocarcinomas, papillary adenocarcinomas, and alveolar-cell adenocarcinomas, were observed in 52 rats exposed to beryllium sulfate aerosol. Adenocarcinomata were the most numerous. Pulmonary metastases tended to localize in areas with foam cell clustering and granulomatosis. No neoplasia was observed in any of the control rats. The incidence of lung tumors in exposed rats is presented in the following Table 2:

Schepers (1962) reviewed 38 existing beryllium studies that evaluated seven beryllium compounds and seven mammalian species. Beryllium sulfate, beryllium fluoride, beryllium phosphate, beryllium alloy (BeZnMnSiO₄), and beryllium oxide were proven to be carcinogenic and have remarkable pleomorphic neoplasiogenic proclivities. Ten varieties of tumors were observed, with adenocarcinoma being the most common variety.

In another study, Vorwald and Reeves (1959) exposed Sherman albino rats via the inhalation route to aerosols of 0.006 mg beryllium/m³ as beryllium oxide and 0.0547 mg beryllium/m³ as beryllium sulfate for 6 hours/day, 5 days/week for an unspecified duration. Lung tumors (single or multifocal) were observed in the animals sacrificed following 9 months of daily inhalation exposure. The histologic pattern of the cancer was primarily adenomatous; however, epidermoid and squamous cell cancers were also observed. Infiltrative, vascular, and lymphogenous extensions often developed with secondary metastatic growth in the tracheobronchial lymph nodes, the mediastinal connective tissue, the parietal pleura, and the diaphragm.

In the first of two articles, Reeves et al. (1967a) investigated the carcinogenic process in lungs resulting from chronic (up to 72 weeks) beryllium sulfate inhalation. One hundred fifty male and female Sprague Dawley C.D. strain rats were exposed to beryllium sulfate aerosol at a mean atmospheric concentration of 34.25 μg beryllium/m³ (with an average particle diameter of 0.12 µm). Prior to initial exposure and again during the 67-68 and 75-76 weeks of life, the animals received prophylactic treatments of tetracycline-HCl to combat recurrent pulmonary infections.

The animals entered the exposure chamber at 6 weeks of age and were exposed 7 hours per day/5 days per week for up to 2,400 hours of total exposure time. An equal number of unexposed controls were held in a separate chamber. Three male and three female rats were sacrificed monthly during the 72-week exposure period. Mortality due to respiratory or other infections did not appear until 55 weeks of age, and 87 percent of all animals survived until their scheduled sacrifices.

Average lung weight towards the end of exposure was 4.25 times normal with progressively increasing differences between control and exposed animals. The increase in lung weight was accompanied by notable changes in tissue texture with two distinct pathological processes—inflammatory and proliferative. The inflammatory response was characterized by marked accumulation of histiocytic elements forming clusters of macrophages in the alveolar spaces. The proliferative response progressed from early epithelial hyperplasia of the alveolar surfaces, through metaplasia (after 20-22 weeks of exposure), anaplasia (cellular dedifferentiation) (after 32-40 weeks of exposure), and finally to lung tumors.

Although the initial proliferative response occurred early in the exposure period, tumor development required considerable time. Tumors were first identified after nine months of beryllium sulfate exposure, with rapidly increasing rates of incidence until tumors were observed in 100 percent of exposed animals by 13 months. The 9-to-13-month interval is consistent with earlier studies. The tumors showed a high degree of local invasiveness. No tumors were observed in control rats. All 56 tumors studied appeared to be alveolar adenocarcinomas and 3 “fast-growing” tumors that reached a very large size comparatively early. About one-third of the tumors showed small foci where the histologic pattern differed. Most of the early tumor foci appeared to be alveolar rather than bronchiolar, which is consistent with the expected pathogenesis, since permanent deposition of beryllium was more likely on the alveolar epithelium rather than on the bronchiolar epithelium. Female rats appeared to have an increased susceptibility to beryllium exposure. Not only did they have a higher mortality (control males [n = 8], exposed males [n = 9] versus control females [n = 4], exposed females [n = 17]) and body weight loss than male rats, but the three “fast-growing” tumors only occurred in females.

In the second article, Reeves et al. (1967b) described the rate of accumulation and clearance of beryllium sulfate aerosol from the same experiment (Reeves et al., 1967a). At the time of the monthly sacrifice, beryllium assays were performed on the lungs, tracheobronchial lymph nodes, and blood of the exposed rats. The pulmonary beryllium levels of rats showed a rate of accumulation which decreased during continuing exposure and reached a plateau (defined as equilibrium between deposition and clearance) of about 13.5 μg beryllium for males and 9 μg beryllium for females in whole lungs after approximately 36 weeks. Females were notably less efficient than males in utilizing the lymphatic route as a method of clearance, resulting in slower removal of pulmonary beryllium deposits, lower accumulation of the inhaled material in the tracheobronchial lymph nodes, and higher morbidity and mortality.

There was no apparent correlation between the extent and severity of pulmonary pathology and total lung load. However, when the beryllium content of the excised tumors was compared with that of surrounding nonmalignant pulmonary tissues, the former showed a notable decrease (0.50 ± 0.35 μg beryllium/gram versus 1.50 ± 0.55 μg beryllium/gram). This was believed to be largely a result of the dilution factor operating in the rapidly growing tumor tissue. However, other factors, such as lack of continued local deposition due to impaired respiratory function and enhanced clearance due to high vascularity of the tumor, may also have played a role. The portion of inhaled beryllium retained in the lungs for a longer duration, which is in the range of one-half of the original pulmonary load, may have significance for pulmonary carcinogenesis. This pulmonary beryllium burden becomes localized in the cell nuclei and may be an important factor in eliciting the carcinogenic response associated with beryllium inhalation.

Groth et al. (1980) conducted a series of experiments to assess the carcinogenic effects of beryllium, beryllium hydroxide, and various beryllium alloys. For the beryllium metal/alloys experiment, 12 groups of 3-month-old female Wistar rats (35 rats/group) were used. All rats in each group received a single intratracheal injection of either 2.5 or 0.5 mg of one of the beryllium metals or beryllium alloys as described in Table 3 below. These materials were suspended in 0.4 cc of isotonic saline followed by 0.2 cc of saline. Forty control rats were injected with 0.6 cc of saline. The geometric mean particle sizes varied from 1 to 2 µm. Rats were sacrificed and autopsied at various intervals ranging from 1 to 18 months post-injection.

Lung tumors were observed only in rats exposed to beryllium metal, passivated beryllium metal, and beryllium-aluminum alloy. Passivation refers to the process of removing iron contamination from the surface of beryllium metal. As discussed, metal alloys may have a different toxicity than beryllium alone. Rats exposed to 100 percent beryllium exhibited relatively high mortality rates, especially in the groups where lung tumors were observed. Nodules varying from 1 to 10 mm in diameter were also observed in the lungs of rats exposed to beryllium metal, passivated beryllium metal, and beryllium-aluminum alloy. These nodules were suspected of being malignant.

To test this hypothesis, transplantation experiments involving the suspicious nodules were conducted in nine rats. Seven of the nine suspected tumors grew upon transplantation. All transplanted tumor types metastasized to the lungs of their hosts. Lung tumors were observed in rats injected with both the high and low doses of beryllium metal, passivated beryllium metal, and beryllium-aluminum alloy. No lung tumors were observed in rats injected with the other compounds. From a total of 32 lung tumors detected, most were adenocarcinomas and adenomas; however, two epidermoid carcinomas and at least one poorly differentiated carcinoma were observed. Bronchiolar alveolar cell tumors were frequently observed in rats injected with beryllium metal, passivated beryllium metal, and beryllium-aluminum alloy. All stages of cuboidal, columnar, and squamous cell metaplasia were observed on the alveolar walls in the lungs of rats injected with beryllium metal, passivated beryllium metal, and beryllium-aluminum alloy. These lesions were generally reduced in size and number or absent from the lungs of animals injected with the other alloys (BeCu, BeCuCo, BeNi).

The extent of alveolar metaplasia could be correlated with the incidence of lung cancer. The incidences of lung tumors in the rats that received 2.5 mg of beryllium metal, and 2.5 and 0.5 mg of passivated beryllium metal, were significantly different (p ≤ 0.008) from controls. When autopsies were performed at the 16-to-19-month interval, the incidence (2/6) of lung tumors in rats exposed to 2.5 mg of beryllium-aluminum alloy was statistically significant (p = 0.004) when compared to the lung tumor incidence (0/84) in rats exposed to BeCu, BeNi, and BeCuCo alloys, which contained much lower concentrations of Be (Groth et al., 1980).

Finch et al. (1998b) investigated the carcinogenic effects of inhaled beryllium on heterozygous TSG-p53 knockout mice (p53^+/−) and wild-type (p53+/+) mice. Knockout mice can be valuable tools in determining the role of specific genes on the toxicity of a material of interest, in this case, beryllium. Equal numbers of approximately 10-week-old male and female mice were used for this study. Two exposure groups were used to provide dose-response information on lung carcinogenicity. The maximum initial lung burden (ILB) target of 60 μg beryllium was based on previous acute inhalation exposure studies in mice. The lower exposure target level of 15 μg was selected to provide a lung burden significantly less than the high-level group, but high enough to yield carcinogenic responses. Mice were exposed in groups to beryllium metal or to filtered air (controls) via nose-only inhalation. The specific exposure parameters are presented in Table 4 below. Mice were sacrificed 7 days post exposure for ILB analysis, and either at 6 months post exposure (n = 4-5 mice per group per gender) or when 10 percent or less of the original population remained (19 months post exposure for p53^+/− knockout and 22.5 months post exposure for p53+/+ wild-type mice). The sacrifice time was extended in the study because a significant number of lung tumors were not observed at 6 months post exposure.

Lung burdens of beryllium measured in wild-type mice at 7 days post exposure were approximately 70-90 percent of target levels. No exposure-related effects on body weight were observed in mice; however, lung weights and lung-to-body-weight ratios were somewhat elevated in 60 μg target ILB p53^+/− knockout mice compared to controls (0.05 < p < 0.10). In general, p53+/+ wild-type mice survived longer than p53^+/− knockout mice and beryllium exposure tended to decrease survival time in both groups. The incidence of beryllium-induced lung tumors was marginally higher in the 60 μg target ILB p53^+/− knockout mice compared to 60 μg target ILB p53+/+ wild-type mice (p = 0.056). The incidence of lung tumors in the 60 μg target ILB p53^+/− knockout mice was also significantly higher than controls (p = 0.048). No tumors developed in the control mice, 15 μg target ILB p53^+/− knockout mice, or 60 μg target ILB p53+/+ wild-type mice throughout the length of the study. Most lung tumors in beryllium-exposed mice were squamous cell carcinomas, three of four of which were poorly circumscribed and all were associated with at least some degree of granulomatous pneumonia. The study results suggest that having an inactivated p53 allele is associated with lung tumor progression in p53^+/− knockout mice. This is based on the significant difference seen in the incidence of beryllium-induced lung neoplasms for the p53^+/− knockout mice compared with the p53^+/+ wild-type mice. The authors conclude that since there was a relatively late onset of tumors in the beryllium-exposed p53^+/− knockout mice, a 6-month bioassay in this mouse strain might not be an appropriate model for lung carcinogenesis (Finch et al., 1998b).

Nickell-Brady et al. (1994) investigated the development of lung tumors in 12-week-old F344/N rats after a single nose-only inhalation exposure to beryllium aerosol, and evaluated whether beryllium lung tumor induction involves alterations in the K-ras, p53, and c-raf−1 genes. Four groups of rats (30 males and 30 females per group) were exposed to different mass concentrations of beryllium (Group 1: 500 mg/m³ for 8 min; Group 2: 410 mg/m³ for 30 min; Group 3: 830 mg/m³ for 48 min; Group 4: 980 mg/m³ for 39 min). The beryllium mass median aerodynamic diameter was 1.4 μm (σ_g = 1.9). The mean beryllium lung burdens for each exposure group were 40, 110, 360, and 430 μg, respectively.

To examine genetic alterations, DNA isolation and sequencing techniques (PCR amplification and direct DNA sequence analysis) were performed on wild-type rat lung tissue (i.e., control samples) along with two mouse lung tumor cell lines containing known K-ras mutations, 12 carcinomas induced by beryllium (i.e., experimental samples), and 12 other formalin-fixed specimens. Tumors appeared in beryllium-exposed rats by 14 months, and 64 percent of exposed rats developed lung tumors during their lifetime. Lungs frequently contained multiple tumor sites, with some of the tumors greater than 1 cm. A total of 24 tumors were observed. Most of the tumors (n = 22) were adenocarcinomas exhibiting a papillary pattern characterized by cuboidal or columnar cells, although a few had a tubular or solid pattern. Fewer than 10 percent of the tumors were adenosquamous (n = 1) or squamous cell (n = 1) carcinomas.

No transforming mutations of the K-ras gene (codons 12, 13, or 61) were detected by direct sequence analysis in any of the lung tumors induced by beryllium. However, using a more sensitive sequencing technique (PCR enrichment restriction fragment length polymorphism (RFLP) analysis) resulted in the detection of K-ras codon 12 GGT to GTT transversions in 2 of 12 beryllium-induced adenocarcinomas. No p53 and c-raf-1 alterations were observed in any of the tumors induced by beryllium exposure (i.e., no differences observed between beryllium-exposed and control rat tissues). The authors note that the results suggest that activation of the K-ras proto-oncogene is both a rare and late event, possibly caused by genomic instability during the progression of beryllium-induced rat pulmonary adenocarcinomas. It is unlikely that the K-ras gene plays a role in the carcinogenicity of beryllium. The results also indicate that p53 mutation is unlikely to play a role in tumor development in rats exposed to beryllium.

Belinsky et al. (1997) reviewed the findings by Nickell-Brady et al. (1994) to further examine the role of the K-ras and p53 genes in lung tumors induced in the F344 rat by non-mutagenic (non-genotoxic) exposures to beryllium. Their findings are discussed along with the results of other genomic studies that look at carcinogenic agents that are either similarly non-mutagenic or, in other cases, mutagenic. The authors conclude that the identification of non-ras transforming genes in rat lung tumors induced by non-mutagenic exposures, such as beryllium, as well as mutagenic exposures will help define some of the mechanisms underlying cancer induction by different types of DNA damage.

The inactivation of the p16^INK4a (p16) gene is a contributing factor in disrupting control of the normal cell cycle and may be an important mechanism of action in beryllium-induced lung tumors. Swafford et al. (1997) investigated the aberrant methylation and subsequent inactivation of the p16 gene in primary lung tumors induced in F344/N rats exposed to known carcinogens via inhalation. The research involved a total of 18 primary lung tumors that developed after exposing rats to five agents, one of which was beryllium. In this study, only one of the 18 lung tumors was induced by beryllium exposure; the majority of the other tumors were induced by radiation (x-rays or plutonium-239 oxide). The authors hypothesized that if p16 inactivation plays a central role in development of non-small-cell lung cancer, then the frequency of gene inactivation in primary tumors should parallel that observed in the corresponding cell lines. To test the hypothesis, a rat model for lung cancer was used to determine the frequency and mechanism for inactivation of p16 in matched primary lung tumors and derived cell lines. The methylation-specific PCR (MSP) method was used to detect methylation of p16 alleles. The results showed that the presence of aberrant p16 methylation in cell lines was strongly correlated with absent or low expression of the gene. The findings also demonstrated that aberrant p16 CpG island methylation, an important mechanism in gene silencing leading to the loss of p16 expression, originates in primary tumors.

Building on the rat model for lung cancer and associated findings from Swafford et al. (1997), Belinsky et al. (2002) conducted experiments in 12-week-old F344/N rats (male and female) to determine whether beryllium-induced lung tumors involve inactivation of the p16 gene and estrogen receptor α (ER) gene. Rats received a single nose-only inhalation exposure to beryllium aerosol at four different exposure levels. The mean lung burdens measured in each exposure group were 40, 110, 360, and 430 μg. The methylation status of the p16 and ER genes was determined by MSP. A total of 20 tumors detected in beryllium-exposed rats were available for analysis of gene-specific promoter methylation. Three tumors were classified as squamous cell carcinomas and the others were determined to be adenocarcinomas. Methylated p16 was present in 80 percent (16/20), and methylated ER was present in one-half (10/20), of the lung tumors induced by exposure to beryllium. Additionally, both genes were methylated in 40 percent of the tumors. The authors noted that four tumors from beryllium-exposed rats appeared to be partially methylated at the p16 locus. Bisulfite sequencing of exon 1 of the ER gene was conducted on normal lung DNA and DNA from three methylated, beryllium-induced tumors to determine the density of methylation within amplified regions of exon 1 (referred to as CpG sites). Two of the three methylated, beryllium-induced lung tumors showed extensive methylation, with more than 80 percent of all CpG sites methylated.

The overall findings of this study suggest that inactivation of the p16 and ER genes by promoter hypermethylation are likely to contribute to the development of lung tumors in beryllium-exposed rats. The results showed a correlation between changes in p16 methylation and loss of gene transcription. The authors hypothesize that the mechanism of action for beryllium-induced p16 gene inactivation in lung tumors may be inflammatory mediators that result in oxidative stress. The oxidative stress damages DNA directly through free radicals or indirectly through the formation of 8-hydroxyguanosine DNA adducts, resulting primarily in a single-strand DNA break.

Wagner et al. (1969) studied the development of pulmonary tumors after intermittent daily chronic inhalation exposure to beryllium ores in three groups of male squirrel monkeys. One group was exposed to bertrandite ore, a second to beryl ore, and the third served as unexposed controls. Each of these three exposure groups contained 12 monkeys. Monkeys from each group were sacrificed after 6, 12, or 23 months of exposure. The 12-month sacrificed monkeys (n = 4 for bertrandite and control groups; n = 2 for beryl group) were replaced by a separate replacement group to maintain a total animal population approximating the original numbers and to provide a source of confirming data for biologic responses that might arise following the ore exposures. Animals were exposed to bertrandite and beryl ore concentrations of 15 mg/m³, corresponding to 210 μg beryllium/m³ and 620 μg beryllium/m³ in each exposure chamber, respectively. The parent ores were reduced to particles with geometric mean diameters of 0.27 μm (± 2.4) for bertrandite and 0.64 μm (± 2.5) for beryl. Animals were exposed for approximately 6 hours/day, 5 days/week. The histological changes in the lungs of monkeys exposed to bertrandite and beryl ore exhibited a similar pattern. The changes generally consisted of aggregates of dust-laden macrophages, lymphocytes, and plasma cells near respiratory bronchioles and small blood vessels. There were, however, no consistent or significant pulmonary lesions or tumors observed in monkeys exposed to either of the beryllium ores. This is in contrast to the findings in rats exposed to beryl ore and to a lesser extent bertrandite, where atypical cell proliferation and tumors were frequently observed in the lungs. The authors hypothesized that the rats' greater susceptibility may be attributed to the spontaneous lung disease characteristic of rats, which might have interfered with lung clearance.

As previously described, Conradi et al. (1971) investigated changes in the lungs of monkeys and dogs two years after intermittent inhalation exposure to beryllium oxide calcined at 1,400 °C. Five adult male and female monkeys (Macaca irus) weighing between 3 and 5.75 kg were used in the study. The study included two control monkeys. Beryllium concentrations in the atmosphere of whole-body exposed monkeys varied between 3.30 and 4.38 mg/m³. Thirty-minute exposures occurred once a month for three months, with beryllium oxide concentrations increasing at each exposure interval. Lung tissue was investigated using electron microscopy and morphometric methods. Beryllium content in portions of the lungs of five monkeys was measured two years following exposure by emission spectrography. The reported concentrations in monkeys (82.5, 143.0, and 112.7 μg beryllium per 100 gm of wet tissue in the upper lobe, lower lobe, and combined lobes, respectively) were higher than those in dogs. No neoplastic or granulomatous lesions were observed in the lungs of any exposed animals and there was no evidence of chronic proliferative lung changes after two years.

4. In vitro Studies

The exact mechanism by which beryllium induces pulmonary neoplasms in animals remains unknown (NAS 2008). Keshava et al. (2001) performed studies to determine the carcinogenic potential of beryllium sulfate in cultured mammalian cells. Joseph et al. (2001) investigated differential gene expression to understand the possible mechanisms of beryllium-induced cell transformation and tumorigenesis. Both investigations used cell transformation assays to study the cellular/molecular mechanisms of beryllium carcinogenesis and assess carcinogenicity. Cell lines were derived from tumors developed in nude mice injected subcutaneously with non-transformed BALB/c-3T3 cells that were morphologically transformed in vitro with 50-200 μg beryllium sulfate/ml for 72 hours. The non-transformed cells were used as controls.

Keshava et al. (2001) found that beryllium sulfate is capable of inducing morphological cell transformation in mammalian cells and that transformed cells are potentially tumorigenic. A dose-dependent increase (9-41 fold) in transformation frequency was noted. Using differential polymerase chain reaction (PCR), gene amplification was investigated in six proto-oncogenes (K-ras, c-myc, c-fos, c-jun, c-sis, erb-B2) and one tumor suppressor gene (p53). Gene amplification was found in c-jun and K-ras. None of the other genes tested showed amplification. Additionally, Western blot analysis showed no change in gene expression or protein level in any of the genes examined. Genomic instability in both the non-transformed and transformed cell lines was evaluated using random amplified polymorphic DNA fingerprinting (RAPD analysis). Using different primers, 5 of the 10 transformed cell lines showed genomic instability when compared to the non-transformed BALB/c-3T3 cells. The results indicate that beryllium sulfate-induced cell transformation might, in part, involve gene amplification of K-ras and c-jun and that some transformed cells possess neoplastic potential resulting from genomic instability.

Using the Atlas mouse 1.2 cDNA expression microarrays, Joseph et al. (2001) studied the expression profiles of 1,176 genes belonging to several different functional categories. Compared to the control cells, expression of 18 genes belonging to two functional groups (nine cancer-related genes and nine DNA synthesis, repair, and recombination genes) was found to be consistently and reproducibly different (at least 2-fold) in the tumor cells. Differential gene expression profile was confirmed using reverse transcription-PCR with primers specific to the differentially expressed genes. Two of the differentially expressed genes (c-fos and c-jun) were used as model genes to demonstrate that the beryllium-induced transcriptional activation of these genes was dependent on pathways of protein kinase C and mitogen-activated protein kinase and independent of reactive oxygen species in the control cells. These results indicate that beryllium-induced cell transformation and tumorigenesis are associated with up-regulated expression of the cancer-related genes (such as c-fos, c-jun, c-myc, and R-ras) and down-regulated expression of genes involved in DNA synthesis, repair, and recombination (such as MCM4, MCM5, PMS2, Rad23, and DNA ligase I).

5. Preliminary Lung Cancer Conclusions

OSHA has preliminarily determined that the weight of evidence indicates that beryllium compounds should be regarded as potential occupational lung carcinogens. Other scientific organizations, including the International Agency for Research on Cancer (IARC), the National Toxicology Program (NTP), the U.S. Environmental Protection Agency (EPA), the National Institute for Occupational Safety and Health (NIOSH), and the American Conference of Governmental Industrial Hygienists (ACGIH) have reached similar conclusions with respect to the carcinogenicity of beryllium.

While some evidence exists for direct-acting genotoxicity as a possible mechanism for beryllium carcinogenesis, the weight of evidence suggests a possible indirect mechanism may be responsible for most tumorigenic activity of beryllium in animal models and possibly humans (EPA, 1998). Inflammation has been postulated to be a key contributor to many different forms of cancer (Jackson et al., 2006; Pikarsky et al., 2004; Greten et al., 2004; Leek, 2002). In fact, chronic inflammation may be a primary factor in the development of up to one-third of all cancers (Ames et al., 1990; NCI, 2010).

In addition to a T-cell mediated response beryllium has been demonstrated to produce an inflammatory response in animal models similar to other particles (Reeves et al., 1967; Swafford et al., 1997; Wagner et al., 1969) possibly contributing to its carcinogenic potential. Animal studies, as summarized above, have demonstrated a consistent scenario of beryllium exposure resulting in chronic pulmonary inflammation. Studies conducted in rats have demonstrated that chronic inhalation of materials similar in solubility to beryllium result in increased pulmonary inflammation, fibrosis, epithelial hyperplasia, and, in some cases, pulmonary adenomas and carcinomas (Heinrich et al., 1995; Nikula et al., 1995; NTP, 1993; Lee et al., 1985; Warheit et al., 1996). This response is generally referred to as an “overload” response or threshold effect. Substantial data indicate that tumor formation in the rat after exposure to some sparingly soluble particles at doses causing marked, chronic inflammation is due to a secondary mechanism unrelated to the genotoxicity (or lack thereof) of the particle itself.

It has been hypothesized that the recruitment of neutrophils during the inflammatory response and subsequent release of oxidants from these cells have been demonstrated to play an important role in the pathogenesis of rat lung tumors (Borm et al., 2004; Carter and Driscoll, 2001; Carter et al., 2006; Johnston et al., 2000; Knaapen et al., 2004; Mossman, 2000). Inflammatory mediators, as characterized in many of the studies summarized above, have been shown to play a significant role in the recruitment of cells responsible for the release of reactive oxygen and hydrogen species. These species have been determined to be highly mutagenic themselves as well as mitogenic, inducing a proliferative response (Feriola and Nettesheim, 1994; Jetten et al., 1990; Moss et al., 1994; Coussens and Werb, 2002). The resultant effect is an environment rich for neoplastic transformations and the progression of fibrosis and tumor formation. This finding does not imply no risk at levels below an inflammatory response; rather, the overall weight of evidence is suggestive of a mechanism of an indirect carcinogen at levels where inflammation is seen. While tumorigenesis secondary to inflammation is one reasonable mode of action, other plausible modes of action independent of inflammation (e.g., epigenetic, mitogenic, reactive oxygen mediated, indirect genotoxicity, etc.) may also contribute to the lung cancer associated with beryllium exposure.

Epidemiological studies indicate excess risk of lung cancer mortality from occupational beryllium exposure levels at or below the current OSHA PEL (Schubauer-Berigan et al., 2010; Table 4).

F. Other Health Effects

Past studies on other health effects have been thoroughly reviewed by several scientific organizations (NTP, 1999; EPA, 1998; ATSDR, 2002; WHO, 2001; HSDB, 2010). These studies include summaries of animal studies, in vitro studies, and human epidemiological studies associated with cardiovascular, hematological, hepatic, renal, endocrine, reproductive, ocular and mucosal, and developmental effects. High-dose exposures to beryllium have been shown to have an adverse effect upon a variety of organs and tissues in the body, particularly the liver. The adverse systemic effects from human exposures mostly occurred prior to the introduction of occupational and environmental standards set in 1970-1972 (OSHA, 1971; ACGIH, 1971; ANSI, 1970) and 1974 (EPA, 1974) and therefore are less relevant today than in the past. The available data is fairly limited. The hepatic, cardiovascular, renal, and ocular and mucosal effects are briefly summarized below. Health effects in other organ systems listed above were only observed in animal studies at very high exposure levels and are, therefore, not discussed here.

1. Hepatic Effects

Beryllium has been shown to accumulate in the liver and a correlation has been demonstrated between beryllium content and hepatic damage. Different compounds have been shown to distribute differently within the hepatic tissues. For example, beryllium phosphate had accumulated almost exclusively within sinusoidal (Kupffer) cells of the liver, while the beryllium derived from beryllium sulfate was found mainly in parenchymal cells. Conversely, beryllium sulphosalicylic acid complexes were rapidly excreted (Skillteter and Paine, 1979).

According to a few autopsies, beryllium-laden liver had central necrosis, mild focal necrosis as well as congestion, and occasionally beryllium granuloma.

Residents near a beryllium plant may have been exposed by inhaling trace amounts of beryllium powder, and different beryllium compounds may have induced different toxicant reactions (Yian and Yin, 1982).

2. Cardiovascular Effects

There is very limited evidence of cardiovascular effects of beryllium and its compounds in humans. Severe cases of chronic beryllium disease can result in cor pulmonale, which is hypertrophy of the right heart ventricle. In a case history study of 17 individuals exposed to beryllium in a plant that manufactured fluorescent lamps, autopsies revealed right atrial and ventricular hypertrophy (Hardy and Tabershaw, 1946). It is not likely that these cardiac effects were due to direct toxicity to the heart, but rather were a response to impaired lung function. However, an increase in deaths due to heart disease or ischemic heart disease was found in workers at a beryllium manufacturing facility (Ward et al., 1992).

Animal studies performed in monkeys indicate heart enlargement after acute inhalation exposure to 13 mg beryllium/m³ as beryllium hydrogen phosphate, 0.184 mg beryllium/m³ as beryllium fluoride, or 0.198 mg beryllium/m³ as beryllium sulfate (Schepers 1964). Decreased arterial oxygen tension was observed in dogs exposed to 30 mg beryllium/m³ as beryllium oxide for 15 days (HSDB, 2010), 3.6 mg beryllium/m³ as beryllium oxide for 40 days (Hall et al., 1950), or 0.04 mg beryllium/m³ as beryllium sulfate for 100 days (Stokinger et al., 1950). These are expected to be indirect effects on the heart due to pulmonary fibrosis and toxicity which can increase arterial pressure and restrict blood flow.

3. Renal Effects

Renal calculi (stones) were unusually prevalent in severe cases that resulted from high levels of beryllium exposure. Renal stones containing beryllium occurred in about 10 percent of patients affected by high exposures (Barnett, et al., 1961). Kidney stones were observed in 10 percent of the CBD cases collected by the BCR up to 1959 (Hall et al., 1959). In addition, an excess of calcium in the blood and urine has been seen frequently in patients with chronic beryllium disease (ATSDR, 2002).

4. Ocular and Mucosal Effects

Both the soluble, sparingly soluble, and insoluble beryllium compounds have been shown to cause ocular irritation in humans (Van Orstrand et al., 1945; De Nardi et al., 1953; Nishimura, 1966; Epstein, 1990; NIOSH, 1994). In addition, beryllium compounds (soluble, sparingly soluble, or insoluble) have been demonstrated to induce acute conjunctivitis with corneal maculae and diffuse erythema (HSDB, 2010).

The mucosa (mucosal membrane) is the moist lining of certain tissues/organs including the eyes, nose, mouth, lungs, and the urinary and digestive tracts. Soluble beryllium salts have been shown to be directly irritating to mucous membranes (HSDB, 2010).

G. Summary of Preliminary Conclusions Regarding Health Effects

Through careful analysis of the current best available scientific information outlined in this Health Effects Section V, OSHA has preliminarily determined that beryllium and beryllium-containing compounds are able to cause sensitization, chronic beryllium disease (CBD) and lung cancer below the current OSHA PEL of 2 μg/m³. The Agency has preliminarily determined through the studies outlined in section V.A.2 of this health effects section that skin and inhalation exposure to beryllium can lead to sensitization; and inhalation exposure, or skin exposure coupled with inhalation, can cause onset and progression of CBD. In addition, the Agency has preliminarily determined through studies outlined in section V.E. of this health effects section that inhalation exposure to beryllium and beryllium containing materials causes lung cancer.

1. Beryllium Causes Sensitization Below the Current PEL and Sensitization is a Precursor to CBD

Through the biological and immunological processes outlined in section V.B. of the Health Effects, the Agency believes that the scientific evidence supports the following mechanism for the development of sensitization and CBD.

• Inhaled beryllium and beryllium-containing materials able to be retained and solubilized in the lungs initiate sensitization and facilitate CBD development (Section V.B.5).
• Beryllium compounds that dissolve in biological fluids, such as sweat, can penetrate intact skin and initiate sensitization (section V.A.2; V.B). Phagosomal fluid and lung fluid have been demonstrated to dissolve beryllium compounds in the lung (section V.A.2a).
• Sensitization occurs through a CD4+ T-cell mediated process with both soluble and insoluble beryllium and beryllium-containing compounds through direct antigen presentation or through further antigen processing (section V.D.1) in the skin or lung. T-cell mediated responses, such as sensitization, are generally regarded as long-lasting (e.g., not transient or readily reversible) immune conditions.
• Beryllium sensitization and CBD are adverse events along a pathological continuum in the disease process with sensitization being the necessary first step in the progression to CBD (section V.D).

○ Animal studies have provided supporting evidence for T-cell proliferation in the development of granulomatous lung lesions after beryllium exposure (section V.D.2; V.D.6).

○ Since the pathogenesis of CBD involves a beryllium-specific, cell-mediated immune response, CBD cannot occur in the absence of beryllium sensitization (V.D.1). While no clinical symptoms are associated with sensitization, a sensitized worker is at risk of developing CBD upon subsequent inhalation exposure to beryllium.

○ Epidemiological evidence that covers a wide variety of different beryllium compounds and industrial processes demonstrates that sensitization and CBD are continuing to occur at present-day exposures below OSHA's PEL (section V.D.4; V.D.5).

• OSHA considers CBD to be a progressive illness with a continuous spectrum of symptoms ranging from its earliest asymptomatic stage following sensitization through to full-blown CBD and death (section V.D.7).
• Genetic variabilities may enhance risk for developing sensitization and CBD in some groups (section V.D.3).

In addition, epidemiological studies outlined in section V.D.5 have demonstrated that efforts to reduce exposures have succeeded in reducing the frequency of sensitization and CBD.

2. Evidence Indicates Beryllium is a Human Carcinogen

OSHA has conducted an evaluation of the current available scientific information of the carcinogenic potential of beryllium and beryllium-containing compounds (section V.E). Based on weight of evidence and plausible mechanistic information obtained from in vitro and in vivo animal studies as well as clinical and epidemiological investigations, the Agency has preliminarily determined that beryllium and beryllium-containing materials should be regarded as human carcinogens. This information is in accordance with findings from IARC, NTP, EPA, NIOSH, and ACGIH (section V.E).

• Lung cancer is an irreversible and frequently fatal disease with an extremely poor 5-year survival rate (NCI, 2009).
• Epidemiological cohort studies have reported statistically significant excess lung cancer mortality among workers employed in U.S. beryllium production and processing plants during the 1930s to 1970s (Section V.E.2).
• Significant positive associations were found between lung cancer mortality and both average and cumulative beryllium exposures when appropriately adjusted for birth cohort and short-term work status (Section V.E.2).
• Studies in which large amounts of different beryllium compounds were inhaled or instilled in the respiratory tracts of experimental animals resulted in an increased incidence of lung tumors (Section V.E.3).
• Authoritative scientific organizations, such as the IARC, NTP, and EPA, have classified beryllium as a known or probable human carcinogen.

While OSHA has preliminarily determined there is sufficient evidence of beryllium carcinogenicity, the exact tumorigenic mechanism for beryllium is unclear and a number of mechanisms are plausibly involved, including chronic inflammation, genotoxicity, mitogenicity oxidative stress, and epigenetic changes (section V.E.3).

• Studies of beryllium exposed animals have consistently demonstrated chronic pulmonary inflammation after exposure (section V.E.3).

○ Substantial data indicate that tumor formation in certain animal models after inhalation exposure to sparingly soluble particles at doses causing marked, chronic inflammation is due to a secondary mechanism unrelated to the genotoxicty of the particle (section V.E.5).

• A review conducted by the NAS (2008) found that beryllium and beryllium-containing compounds tested positive for genotoxicity in nearly 50 percent of studies without exogenous metabolic activity, suggesting a possible direct-acting mechanism may exist (section V.E.1) as well as the potential for epigenetic changes (section V.E.4).

Other health effects have been summarized in sections F of the Health Effects Section and include hepatic, cardiovascular, renal, ocular, and mucosal effects. The adverse systemic effects from human exposures mostly occurred prior to the introduction of occupational and environmental standards set in 1970-1972 (OSHA, 1971; ACGIH, 1971; ANSI, 1970) and 1974 (EPA, 1974) and therefore are less relevant today than in the past.

APPENDIX

VI. Preliminary Beryllium Risk Assessment

The Occupational Safety and Health (OSH) Act and court cases arising under it have led OSHA to rely on risk assessment to support the risk determinations required to set a permissible exposure limit (PEL) for a toxic substance in standards under the OSH Act. Section 6(b)(5) of the OSH Act states that “The Secretary [of Labor], in promulgating standards dealing with toxic materials or harmful physical agents under this subsection, shall set the standard which most adequately assures, to the extent feasible, on the basis of the best available evidence, that no employee will suffer material impairment of health or functional capacity even if such employee has regular exposure to the hazard dealt with by such standard for the period of his working life” (29 U.S.C. 655(b)(5)).

In Industrial Union Department, AFL-CIO v. American Petroleum Institute, 448 U.S. 607 (1980) (Benzene), the United States Supreme Court ruled that the OSH Act requires that, prior to the issuance of a new standard, a determination must be made that there is a significant risk of material impairment of health at the existing PEL and that issuance of a new standard will significantly reduce or eliminate that risk. The Court stated that “before [the Secretary] can promulgate any permanent health or safety standard, the Secretary is required to make a threshold finding that a place of employment is unsafe—in the sense that significant risks are present and can be eliminated or lessened by a change in practices” (Id. at 642). The Court also stated “that the Act does limit the Secretary's power to requiring the elimination of significant risks” (488 U.S. at 644 n.49), and that “OSHA is not required to support its finding that a significant risk exists with anything approaching scientific certainty” (Id. at 656).

OSHA's approach for the risk assessment incorporates both a review of the recent literature on populations of workers exposed to beryllium below the current Permissible Exposure Limit (PEL) of 2 μg/m³ and a statistical exposure-response analysis. OSHA evaluated risk at several alternate PELs under consideration by the Agency: 2 μg/m³, 1 μg/m³, 0.5 μg/m³, 0.2 μg/m³, and 0.1 μg/m³. A number of recently published epidemiological studies evaluate the risk of sensitization and CBD for workers exposed at and below the current PEL and the effectiveness of exposure control programs in reducing risk. OSHA also conducted a statistical analysis of the exposure-response relationship for sensitization and CBD at the current PEL and alternate PELs the Agency is considering. For this analysis, OSHA used data provided by National Jewish Medical and Research Center (NJMRC) on a population of workers employed at a beryllium machining plant in Cullman, AL. The review of the epidemiological studies and OSHA's own analysis show substantial risk of sensitization and CBD among workers exposed at and below the current PEL of 2 μg/m³. They also show substantial reduction in risk where employers have implemented a combination of controls, including stringent control of airborne beryllium levels and additional measures such as respirators, dermal personal protective equipment (PPE), and strict housekeeping to protect workers against dermal and respiratory beryllium exposure. To evaluate lung cancer risk, OSHA relied primarily on a quantitative risk assessment published in 2011 by NIOSH. This risk assessment was based on an update of the Reading cohort analyzed by Sanderson et al., as well as workers from two smaller plants (Schubauer-Berigan et al., 2011) where workers were exposed to lower levels of beryllium and worked for longer periods than at the Reading plant. The authors found that lung cancer risk was strongly and significantly related to mean, cumulative, and maximum measures of workers' exposure; they predicted substantial risk of lung cancer at the current PEL, and substantial reductions in risk at the alternate PELs OSHA considered for the proposed rule (Schubauer-Berigan et al., 2011).

A. Review of Epidemiological Literature on Sensitization and Chronic Beryllium Disease From Occupational Exposure

As discussed in the Health Effects section, studies of beryllium-exposed workers conducted using the beryllium lymphocyte proliferation test (BeLPT) have found high rates of beryllium sensitization and CBD among workers in many industries, including at some facilities where exposures were primarily below OSHA's PEL of 2 μg/m³ (Kreiss et al., 1993; Henneberger et al., 2001; Schuler et al., 2005; Schuler et al., 2012). In the mid-1990s, some facilities using beryllium began to aggressively monitor and reduce workplace exposures. Four plants where several rounds of BeLPT screening were conducted before and after implementation of new exposure control methods provide the best currently available evidence on the effectiveness of various exposure control measures in reducing the risk of sensitization and CBD. The experiences of these plants—a copper-beryllium processing facility in Reading, PA, a beryllia ceramics facility in Tucson, AZ; a beryllium processing facility in Elmore, OH; and a machining facility in Cullman, AL—show that efforts to prevent sensitization and CBD by using engineering controls to reduce workers' beryllium exposures to median levels at or around 0.2 μg/m³ and did not emphasize PPE and stringent housekeeping methods, had only limited impact on risk. However, exposure control programs implemented more recently, which drastically reduced respiratory exposure to beryllium via a combination of engineering controls and respiratory protection, controlled dermal contact with beryllium using PPE, and employed stringent housekeeping methods to keep work areas clean and prevent transfer of beryllium between work areas, sharply curtailed new cases of sensitization among newly-hired workers. There is additional, but more limited, information available on the occurrence of sensitization and CBD among aluminum smelter workers with low-level beryllium exposures (Taiwo et al., 2008; Taiwo et al., 2010; Nilsen et al., 2010). A discussion of the experiences at these plants follows.

The Health Effects section also discussed the role of particle characteristics and beryllium compound solubility in the development of sensitization and CBD among beryllium-exposed workers. Respirable particles small enough to reach the deep lung are responsible for CBD. However, larger inhalable particles that deposit in the upper respiratory tract may lead to sensitization. The weight of evidence indicates that both soluble and insoluble forms of beryllium are able to induce sensitization and CBD. Insoluble forms of beryllium that persist in the lung for longer periods may pose greater risk of CBD while soluble forms may more easily trigger immune sensitization. Although these factors potentially influence the toxicity of beryllium, the available data are too limited to reliably account for solubility and particle size in the Agency estimates of risk. The qualitative impact on conclusions and uncertainties with regard to risk are discussed in a later section.

1. Reading, PA, Plant

Schuler et al. conducted a study of workers at a copper-beryllium processing facility in Reading, PA, screening 152 workers with the BeLPT (Schuler et al., 2005). Exposures at this plant were believed to be low throughout its history due to the low percentage of beryllium in the metal alloys used, and the relatively low exposures found in general area samples collected starting in 1969 (sample median ≤ 0.1 μg/m³, 97% < 0.5 μg/m³). The reported prevalences of sensitization (6.5 percent) and CBD (3.9 percent) showed substantial risk at this facility, even though airborne exposures were primarily below OSHA's current PEL of 2 μg/m³.

Personal lapel samples were collected in production and production support jobs between 1995 and May 2000. These samples showed primarily very low airborne beryllium levels, with a median of 0.073 μg/m. The wire annealing and pickling process had the highest personal lapel sample values, with a median of 0.149 μg/m³. Despite these low exposure levels, cases of sensitization continued to occur among workers whose first exposures to beryllium occurred in the 1990s. Five (11.5 percent) workers of 43 hired after 1992 who had no prior beryllium exposure became sensitized, including four in production work and one in production support (Thomas et al., 2009; evaluation for CBD not reported). Two (13 percent) of these sensitized workers were among 15 workers in this group who had been hired less than a year before the screening.

After the BeLPT screening was conducted in 2000, the company began implementing new measures to further reduce workers' exposure to beryllium. Requirements designed to minimize dermal contact with beryllium, including long-sleeve facility uniforms and polymer gloves, were instituted in production areas in 2000. In 2001 the company installed local exhaust ventilation (LEV) in die grinding and polishing. Personal lapel samples collected between June 2000 and December 2001 show reduced exposures plant-wide. Of 2,211 exposure samples collected during this “pre-enclosure program” period, 98 percent were below 0.2 μg/m³ (Thomas et al., 2009, p. 124). Median, arithmetic mean, and geometric mean values ≤ 0.03 μg/m³ were reported in this period for all processes except the wire annealing and pickling process. Samples for this process remained elevated, with a median of 0.1 μg/m³ (arithmetic mean of 0.127 μg/m³, geometric mean of 0.083 μg/m³). In January 2002, the plant enclosed the wire annealing and pickling process in a restricted access zone (RAZ), required respiratory PPE in the RAZ, and implemented stringent measures to minimize the potential for skin contact and beryllium transfer out of the zone. While exposure samples collected by the facility were sparse following the enclosure, they suggest exposure levels comparable to the 2000-01 samples in areas other than the RAZ. Within the RAZ, required use of powered air-purifying respirators (PAPRs) indicates that respiratory exposure was negligible. A 2009 publication on the facility reported that outside the RAZ, “the vast majority of employees do not wear any form of respiratory protection due to very low airborne beryllium concentrations” (Thomas et al., 2009, p. 122).

To test the efficacy of the new measures in preventing sensitization and CBD, in June 2000 the facility began an intensive BeLPT screening program for all new workers. The company screened workers at the time of hire; at intervals of 3, 6, 12, 24, and 48 months; and at 3-year intervals thereafter. Among 82 workers hired after 1999, three cases of sensitization were found (3.7 percent). Two (5.4 percent) of 37 workers hired prior to enclosure of the wire annealing and pickling process were found to be sensitized within 3 and 6 months of beginning work at the plant. One (2.2 percent) of 45 workers hired after the enclosure was confirmed as sensitized. Among these early results, it appears that the greatest reduction in sensitization risk was achieved after median exposures in all areas of the plant were reduced to below 0.1 μg/m³and PPE to prevent dermal contact was instituted.

2. Tucson, AZ, Plant

Kreiss et al. conducted a study of workers at a beryllia ceramics plant, screening 136 workers with the BeLPT in 1992 (Kreiss et al., 1996). Full-shift area samples collected between 1983 and 1992 showed primarily low airborne beryllium levels at this facility. Of 774 area samples, 76 percent were at or below 0.1 μg/m³ and less than 1 percent exceeded 2 μg/m³. A small set (75) of personal lapel samples collected at the plant beginning in 1991 had a median of 0.2 μg/m³ and ranged from 0.1 to 1.8 μg/m³ (arithmetic and geometric mean values not reported) (Kreiss et al., 1996, p. 19). However, area samples and short-term breathing zone samples also showed occasional instances of very high beryllium exposure levels, with extreme values of several hundred μg/m³ and 3.6 percent of short-term breathing zone samples in excess of 5 μg/m³.

Kreiss et al. reported that eight (5.9 percent) of 136 workers tested were sensitized, six (4.4 percent) of whom were diagnosed with CBD. Seven of the eight sensitized employees had worked in machining, where general area samples collected between October 1985 and March 1988 had a median of 0.3 μg/m³, in contrast to a median value of less than 0.1 μg/m³ in other areas of the plant (Kreiss et al., 1996, p. 20; mean values not reported). Short-term breathing zone measurements associated with machining had a median of 0.6 μg/m³, double the median of 0.3 μg/m³ for breathing zone measurements associated with other processes (id., p. 20; mean values not reported). One sensitized worker was one of 13 administrative workers screened, and was among those diagnosed with CBD. Exposures of administrative workers were not well-characterized, but were believed to be among the lowest in the plant. Of three personal lapel samples reported for administrative staff during the 1990s, all were below the then detection limit of 0.2 μg/m³ (Cummings et al., 2007, p.138).

Following the 1992 screening, the facility reduced exposures in machining areas by enclosing machines and installing HEPA filter exhaust systems. Personal samples collected between 1994 and 1999 had a median of 0.2 μg/m³ in production jobs and 0.1 μg/m³ in production support (geometric means 0.21 μg/m³ and 0.11 μg/m³, respectively; arithmetic means not reported. Cummings et al., 2007, p. 138). In 1998, a second screening found that 9 percent of tested workers hired after the 1992 screening were sensitized, of whom one was diagnosed with CBD. All of the sensitized workers had been employed at the plant for less than two years (Henneberger et al., 2001).

Following the 1998 screening, the company continued efforts to reduce exposures and risk of sensitization and CBD by implementing additional engineering and administrative controls and PPE. Respirator use was required in production areas beginning in 1999, and latex gloves were required beginning in 2000. The lapping area was enclosed in 2000, and enclosures were installed for all mechanical presses in 2001. Between 2000 and 2003, water-resistant or water-proof garments, shoe covers, and taped gloves were incorporated to keep beryllium-containing fluids from wet machining processes off the skin. The new engineering measures did not appear to substantially reduce airborne beryllium levels in the plant. Personal lapel samples collected in production processes between 2000 and 2003 had a median and geometric mean of 0.18 μg/m³, similar to the 1994-1999 samples (Cummings et al., 2007, p. 138). However, respiratory protection requirements were instituted in 2000 to control workers' airborne beryllium exposures.

To test the efficacy of the new measures instituted after 1998, in January 2000 the company began screening new workers for sensitization at the time of hire and at 3, 6, 12, 24, and 48 months of employment (Cummings et al., 2007). These more stringent measures appear to have substantially reduced the risk of sensitization among new employees. Of 97 workers hired between 2000 and 2004, one case of sensitization was identified (1 percent). This worker had experienced a rash after an incident of dermal exposure to lapping fluid through a gap between the glove and uniform sleeve, indicating that sensitization may have occurred via skin exposure.

3. Elmore, OH, Plant

Kreiss et al., Schuler et al., and Bailey et al. conducted studies of workers at a beryllium metal, alloy, and oxide production plant. Workers participated in BeLPT surveys in 1992 (Kreiss et al., 1997) and in 1997 and 1999 (Schuler et al., 2012). Exposure levels at the plant between 1984 and 1993 were characterized by a mixture of general area, short-term breathing zone, and personal lapel samples. Kreiss et al. reported that the median area samples for various work areas ranged from 0.1 to 0.7 μg/m³, with the highest values in the alloy arc furnace and alloy melting-casting areas (other measures of central tendency not reported). Personal lapel samples were available from 1990-1992, and showed high exposures overall (median value of 1.0 μg/m³) with very high exposures for some processes. The authors reported median sample values of 3.8 μg/m³ for beryllium oxide production, 1.75 μg/m³ for alloy melting and casting, and 1.75 μg/m³ for the arc furnace.

Kreiss et al. reported that 43 (6.9 percent) of 627 workers tested in 1992 were sensitized, six of whom were diagnosed with CBD (4.4 percent). Workers with less than one year tenure at the plant were not tested in this survey (Bailey et al., 2010, p. 511). The work processes that appeared to carry the highest risk for sensitization and CBD (e.g., ceramics) were not those with the highest reported exposure levels (e.g., arc furnace and melting-casting). The authors noted several possible reasons for this, including factors such as solubility, particle size/number, and particle surface area that could not be accounted for in their analysis (Kreiss et al., 1997).

In 1996-1999, the company took steps to reduce workers' beryllium exposures: some high-exposure processes were enclosed, special restricted-access zones were set up, HEPA filters were installed in air handlers, and some ventilation systems were updated. In 1997 workers in the pebble plant restricted access zone were required to wear half-face air-purifying respirators, and beginning in 1999 all new employees were required to wear loose-fitting powered air-purifying respirators (PAPR) in manufacturing buildings (Bailey et al., 2010, p. 506). Skin protection became part of the protection program for new employees in 2000, and glove use was required in production areas and for handling work boots beginning in 2001. Also beginning in 2001, either half-mask respirators or PAPRs were required in the production facility (type determined by airborne beryllium levels), and respiratory protection was required for roof work and during removal of work boots (Bailey et al., 2010, p. 506). Respirator use was reported to be used on about half or less of industrial hygiene sample records for most processes in 1990-1992 (Kreiss et al., 1996).

Beginning in 2000, workers were offered periodic BeLPT testing to evaluate the effectiveness of a new exposure control program implemented by the company. Bailey et al. (2010) reported on the results of this surveillance for 290 workers hired between February 21, 2000 and December 18, 2006. They compared the occurrence of beryllium sensitization and disease among 258 employees who began work at the Elmore plant between January 15, 1993 and August 9, 1999 (the `pre-program group') and among 290 employees who were hired between February 21, 2000 and December 18, 2006 and were tested at least once after hire (the `program group'). They found that, as of 1999, 23 (8.9 percent) of the pre-program group were sensitized to beryllium. Six (2.1 percent) of the program group had confirmed abnormal results on their final round of BeLPTs, which occurred in different years for different employees. In addition, another five employees had confirmed abnormal BeLPT results at some point during the testing period, followed by at least one instance of a normal test result. One of these employees had a confirmed abnormal baseline BeLPT at hire, and had two subsequent normal BeLPT results at 6 and 12 months after hire. Four others had confirmed abnormal BeLPT results at 3 or 6 months after hire, later followed by a normal test. Including these four in the count of sensitized workers, there were a total of ten (3.5 percent) workers sensitized after hire in the program group. It is not clear whether the occurrence of a normal result following an abnormal result reflects an error in one of the test results, a change in the presence or level of memory T-cells circulating in the worker's blood, or other possibilities. Because most of the workers in the study had been employed at the facility for less than two years, Bailey et al. did not report the incidence of CBD among the sensitized workers (Bailey et al., 2010, p. 511).

In addition, Bailey et al. divided the program group into the `partial program subgroup' (206 employees hired between February 21, 2000 and December 31, 2003) and the `full program subgroup' (84 employees hired between January 1, 2004 and December 18, 2006) to account for the greater effectiveness of the exposure control program after the first three years of implementation (Bailey et al., pp 506-507). Four (1.9 percent) of the partial program group were found to be sensitized on their final BeLPT (excluding one with a confirmed abnormal BeLPT from their baseline test at hire). Two (2.4 percent) of the full program group were found to be sensitized on their final BeLPT (Bailey et al., 2010, p. 509). An additional three employees in the partial program group and one in the full program group were confirmed sensitized at 3 or 6 months after hire, then later had a single normal BeLPT (Bailey et al., 2010, p. 509).

Schuler et al. (2012) published a study examining beryllium sensitization and CBD among short-term workers at the Elmore, OH plant, using exposure estimates created by Virji et al. (2012). The study population included 264 workers employed in 1999 with up to six years tenure at the plant (91 percent of the 291 eligible workers). By including only short-term workers, Virji et al. were able to construct participants' exposures with more precision than was possible in studies involving workers exposed for longer durations and in time periods with less exposure sampling. Each participant completed a work history questionnaire and was tested for beryllium sensitization. The overall prevalence of sensitization was 9.8 percent (26/264). Sensitized workers were offered further evaluation for CBD. Twenty-two sensitized workers consented to clinical testing for CBD via transbronchial biopsy. Six of those sensitized were diagnosed with CBD (2.3 percent, 6/264).

Exposure estimates were constructed using two exposure surveys conducted in 1999: a survey of total mass exposures (4022 full-shift personal samples) and a survey of size-separated impactor samples (198 samples). The 1999 exposure surveys and work histories were used to estimate long-term lifetime weighted (LTW) average, cumulative, and highest-job-worked exposure for total, respirable, and submicron beryllium mass concentrations. Schuler et al. (2012) found no cases of sensitization among workers with total mass LTW average exposures below 0.09 μg/m³, among workers with total mass cumulative exposures below 0.08 μg/m³-yr, or among workers with total mass highest job worked exposures below 0.12 μg/m³. Twenty-four percent, 16 percent, and 25 percent of the study population were exposed below those levels, respectively. Both total and respirable beryllium mass concentration estimates were positively associated with sensitization (average and highest job), and CBD (cumulative) in logistic regression models.

4. Cullman, AL, Plant

Newman et al. conducted a series of BeLPT screenings of workers at a precision machining facility between 1995 and 1999 (Newman et al., 2001). A small set of personal lapel samples collected in the early 1980s and in 1995 suggests that exposures in the plant varied widely during this time period. In some processes, such as engineering, lapping, and electrical discharge machining (EDM), exposures were apparently low (≤ 0.1 μg/m). Madl et al. reported that personal lapel samples from all machining processes combined had a median of 0.33 μg/m, with a much higher arithmetic mean of 1.63 μg/m (Madl et al., 2007, Table IV, p. 457). The majority of these samples were collected in the high-exposure processes of grinding (median of 1.05 μg/m, mean of 8.48 μg/m), milling (median of 0.3 μg/m, mean of 0.82 μg/m), and lathing (median of 0.35 μg/m, mean of 0.88 μg/m) (Madl et al., 2007, Table IV, p. 457). As discussed in greater detail in the background document, the data set of machining exposure measurements included a few extremely high values (41-73 μg/m³) that a NIOSH researcher identified as probable errors, and that appear to be included in Madl et al.' s arithmetic mean calculations. Because high single-data point exposure errors influence the arithmetic mean far more than the median value of a data range, OSHA believes the median values reported by Madl et al. are more reliable than the arithmetic means they reported.

After a sentinel case of CBD was diagnosed at the plant in 1995, the company began BeLPT screenings to identify workers at increased risk of CBD and implemented engineering and administrative controls and PPE designed to reduce workers' beryllium exposures in machining operations. Newman et al. reported 22 (9.4 percent) sensitized workers among 235 tested, 13 of whom were diagnosed with CBD within the study period. Between 1995 and 1997, the company built enclosures and installed or updated local exhaust ventilation (LEV) for several machining departments, removed pressurized air hoses, and required the use of company uniforms. Madl et al. reported that historically, engineering and work process controls, rather than personal protective equipment, were used to limit workers' exposure to beryllium; respirators were used only in cases of high exposure, such as during sandblasting (Madl et al., 2007, p. 450). In contrast to the Reading and Tucson plants, gloves were not required at this plant.

Personal lapel samples collected extensively between 1996 and 1999 in machining jobs have an overall median of 0.16 μg/m³, showing that the new controls achieved a marked reduction in machinists' exposures during this period. Nearly half of the samples were collected in milling (median = 0.18 μg/m³). Exposures in other machining processes were also reduced, including grinding (median of 0.18 μg/m³) and lathing (median of 0.13 μg/m³). However, cases of sensitization and CBD continued to occur.

At the time that Newman et al. reviewed the results of BeLPT screenings conducted in 1995-1999, a subset of 60 workers had been employed at the plant for less than a year. Four (6.7 percent) of these workers were found to be sensitized, of whom two were diagnosed with CBD and one with probable CBD (Newman et al., 2001). All four had been hired in 1996. Two (one CBD case, one sensitized only) had worked only in milling, and had worked for approximately 3-4 months (0.3-0.4 yrs) at the time of diagnosis. One of those diagnosed with CBD worked only in EDM, where lapel samples collected between 1996 and 1999 had a median of 0.03 μg/m³. This worker was diagnosed with CBD in the same year that he began work at the plant. The last CBD case worked as a shipper, where exposures in 1996-1999 were similarly low, with a median of 0.09 μg/m³.

Beginning in 2000, exposures in all jobs at the machining facility were reduced to extremely low levels. Personal lapel samples collected in machining processes between 2000 and 2005 had a median of 0.09 μg/m³, where more than a third of samples came from the milling process (n = 765, median of 0.09 μg/m³). A later publication on this plant by Madl et al. reported that only one worker hired after 1999 became sensitized. This worker had been employed for 2.7 years in chemical finishing, where exposures were roughly similar to other machining processes (n = 153, median of 0.12 μg/m³). Madl et al. did not report whether this worker was evaluated for CBD.

5. Aluminum Smelting Plants

Taiwo et al. (2008) studied a population of 734 employees at four aluminum smelters located in Canada (2), Italy (1), and the United States (1). In 2000, a beryllium exposure limit of 0.2 μg/m³ 8-hour TWA (action level 0.1 μg/m³) and a short-term exposure limit (STEL) of 1.0 μg/m³ (15-minute sample) were instituted at these plants. Sampling to determine compliance with the exposure limit began at all smelters in 2000. Table VI-1 below, adapted from Taiwo et al. (2008), shows summary information on samples collected from the start of sampling through 2005.

All employees potentially exposed to beryllium levels at or above the action level for at least 12 days per year, or exposed at or above the STEL 12 or more times per year, were offered medical surveillance including the BeLPT (Taiwo et al., 2008, p. 158). Table VI-2 below, adapted from Taiwo et al. (2008), shows test results for each facility between 2001 and 2005.

The two workers with confirmed beryllium sensitization were offered further evaluation for CBD. Both were diagnosed with CBD, based on broncho-alveolar lavage (BAL) results in one case and pulmony function tests, respiratory symptoms, and radiographic evidence in the other.

In 2010, Taiwo et al. published a study of beryllium-exposed workers from smelters at four companies, including some of the workers from the 2008 publication. 3,185 workers were determined to be “significantly exposed” to beryllium and invited to participate in BeLPT screening. Each company used different criteria to determine “significant” exposure, which appeared to vary considerably (p. 570). About 60 percent of invited workers participated in the program between 2000 and 2006, of whom nine were determined to be sensitized (see Table VI-3 below). The authors state that all nine workers were referred to a respiratory physician for further evaluation for CBD. Two were diagnosed with CBD, as described above (Taiwo et al., 2008). The authors do not report the details of other sensitized workers' evaluation for CBD.

In general, there appeared to be a low level of sensitization and CBD among employees at the aluminum smelters studied by Taiwo et al. This is striking in light of the fact that many of the employees tested had worked at the smelters long before the institution of exposure limits for beryllium at some smelters in 2000. However, the authors note that respiratory protection had long been used at these plants to protect workers from other hazards. The results are roughly consistent with the observed prevalence of sensitization following the institution of respiratory protection at the Tucson beryllium ceramics plant discussed previously. A study by Nilsen et al. (2010) also found a low rate of sensitization among aluminum workers in Norway. Three-hundred sixty-two workers and thirty-one control individuals received BeLPT testing for beryllium sensitization. The authors found one sensitized worker (0.28 percent). No borderline results were reported. The authors reported that current exposures in this plant ranged from 0.1 µg/m³ to 0.31 µg/m³ (Nilsen et al., 2010) and that respiratory protection was in use, as is the case in the smelters studied by Taiwo et al. (2008, 2010).

B. Preliminary Conclusions

The published literature on beryllium sensitization and CBD shows that risk of both can be substantial in workplaces in compliance with OSHA's current PEL (Kreiss et al., 1993; Schuler et al., 2005). The experiences of several facilities in developing effective industrial hygiene programs have shown that minimizing both airborne and dermal exposure, using a combination of engineering and administrative controls, respiratory protection, and dermal PPE, has substantially lowered workers' risk of beryllium sensitization. In contrast, risk-reduction programs that relied primarily on engineering controls to reduce workers' exposures to median levels in the range of 0.1-0.2 µg/m³, such as those implemented in Tucson following the 1992 survey and in Cullman during 1996-1999, had only limited impact on reducing workers' risk of sensitization. The prevalence of sensitization among workers hired after such controls were installed at the Cullman plant remained high (Newman et al. (6.7 percent) and Henneberger et al. (9 percent)). A similar prevalence of sensitization was found in the screening conducted in 2000 at the Reading plant, where the available sampling data show median exposure levels of less than 0.2 µg/m³ (6.5 percent). The risk of sensitization was found to be particularly high among newly-hired workers (≤1 year of beryllium exposure) in the Reading 2000 screening (13 percent) and the Tucson 1998 screening (16 percent).

Cases of CBD have also continued to develop among workers in facilities and jobs where exposures were below 0.2 µg/m³. One case of CBD was found in the Tucson 1998 screening among nine sensitized workers hired less than two years previously (Henneberger et al., 2001). At the Cullman plant, at least two cases of CBD were found among four sensitized workers screened in 1995-1999 and hired less than a year previously (Newman et al., 2001). These results suggest a substantial risk of progression from sensitization to CBD among workers exposed at levels well below the current PEL, especially considering the extremely short time of exposure and follow-up for these workers. Six of 10 sensitized workers identified at Reading in the 2000 screening were diagnosed with CBD. The four sensitized workers who did not have CBD at their last clinical evaluation had been hired between one and five years previously; therefore, the time may have been too short for CBD to develop.

In contrast, more recent exposure control programs that have used a combination of engineering controls, PPE, and stringent housekeeping measures to reduce workers' airborne and dermal exposures have substantially lowered risk of sensitization among newly-hired workers. Of 97 workers hired between 2000 and 2004 in Tucson, where respiratory and skin protection was instituted for all workers in production areas, only one (1 percent) worker became sensitized, and in that case the worker's dermal protection had failed during wet-machining work (Thomas et al., 2009). In the aluminum smelters discussed by Taiwo et al., where available exposure samples indicated median beryllium levels of about 0.1 μg/m³ or below (measured as an 8-hour TWA) and workers used respiratory and dermal protection, confirmed cases of sensitization were rare (zero or one case per location). Sensitization was also rare among workers at a Norwegian aluminum smelter (Nilsen et al., 2010), where estimated exposures in the plant ranged from 0.1 μg/m³ to 0.3 μg/m³ and respiratory protection was regularly used. In Reading, where in 2000-2001 airborne exposures in all jobs were reduced to a median of 0.1 μg/m³ or below (measured as an 8-hour TWA) and dermal protection was required for production-area workers, two (5.4 percent) of 37 newly hired workers became sensitized (Thomas et al., 2009). After the process with the highest exposures (median of 0.1 μg/m³) was enclosed in 2002 and workers in that process were required to use respiratory protection, the remaining jobs had very low exposures (medians ~ 0.03 μg/m³). Among 45 workers hired after the enclosure, one was found to be sensitized (2.2 percent). In Elmore, where all workers were required to wear respirators and skin PPE in production areas beginning in 2000-2001, the estimated prevalence of sensitization among workers hired after these measures were put in place was around 2-3 percent (Bailey et al., 2010). In addition, Schuler et al. (2012) found no cases of sensitization among short-term Elmore workers employed in 1999 who had total mass LTW average exposures below 0.09 μg/m³, among workers with total mass cumulative exposures below 0.08 μg/m³-yr, or among workers with total mass highest job worked exposures below 0.12 μg/m³.

Madl et al. reported one case of sensitization among workers at the Cullman plant hired after 2000. The median personal exposures were about 0.1 μg/m³ or below for all jobs during this period. Several changes in the facility's exposure control methods were instituted in the late 1990s that were likely to have reduced dermal as well as respiratory exposure to beryllium. For example, the plant installed change/locker rooms for workers entering the production facility, instituted requirements for work uniforms and dedicated work shoes for production workers, implemented annual beryllium hazard awareness training that encouraged glove use, and purchased high efficiency particulate air (HEPA) filter vacuum cleaners for workplace cleanup and decontamination.

The results of the Reading, Tucson, and Elmore studies show that reducing airborne exposures to below 0.1 μg/m³ and protecting workers from dermal exposure, in combination, have achieved a substantial reduction in sensitization risk among newly-hired workers. Because respirator use, dermal protection, and engineering changes were often implemented concurrently at these plants, it is difficult to attribute the reduced risk to any single control measure. The reduction is particularly evident when comparing newly-hired workers in the most recent Reading screenings (2.2-5.4 percent), and the rate of sensitization found among workers hired within the year before the 2000 screening (13 percent). There is a similarly striking difference between the rate of prevalence found among newly-hired workers in the most recent Tucson study (1 percent) and the rate found among workers hired within the year before the 1998 screening at that plant (16 percent). These results are echoed in the Cullman facility, which combined engineering controls to reduce airborne exposures to below 0.1 μg/m³ with measures such as housekeeping improvements and worker training to reduce dermal exposure.

The studies on recent programs to reduce workers' risk of sensitization and CBD were conducted on populations with very short exposure and follow-up time. Therefore, they could not address the question of how frequently workers who become sensitized in environments with extremely low airborne exposures (median <0.1 μg/m³) develop CBD. Clinical evaluation for CBD was not reported for sensitized workers identified in the most recent Tucson, Reading, and Elmore studies. In Cullman, however, two of the workers with CBD had been employed for less than a year and worked in jobs with very low exposures (median 8-hour personal sample values of 0.03-0.09 μg/m³). The body of scientific literature on occupational beryllium disease also includes case reports of workers with CBD who are known or believed to have experienced minimal beryllium exposure, such as a worker employed only in shipping at a copper-beryllium distribution center (Stanton et al., 2006), and workers employed only in administration at a beryllium ceramics facility (Kreiss et al., 1996).

Arjomandi et al. published a study of 50 sensitized workers from a nuclear weapons research and development facility (Arjomandi et al., 2010). Occupational and medical histories including physical examination and chest imaging were available for the great majority (49) of these individuals. Forty underwent testing for CBD via bronchoscopy and transbronchial biopsies. In contrast to the studies of low-exposure populations discussed previously, this group had much longer follow-up time (mean time since first exposure = 32 years) and length of employment at the facility (mean of 18 years). Quantitative exposure estimates for the workers were not presented; however, the authors characterized their probable exposures as “low” (13 workers), “moderate” (28 workers), or “high” (nine workers) based on the jobs they performed at the facility.

Five of the 50 sensitized workers (10 percent) were diagnosed with CBD based on histology or high-resolution computed tomography. An additional three (who had not undergone full clinical evaluation for CBD) were identified as probable CBD cases, bringing the total prevalence of CBD and probable CBD in this group to 16 percent. As discussed in the epidemiology section of the Health Effects chapter, the prevalence of CBD among worker populations regularly exposed at higher levels (e.g., median > 0.1 μg/m³) is typically much greater, approaching 80-100% in several studies. The lower prevalence of CBD in this group of sensitized workers, who were believed to have primarily low exposure levels, suggests that controlling respiratory exposure to beryllium may reduce risk of CBD among sensitized workers as well as reducing risk of CBD via prevention of sensitization. However, it also demonstrates that some workers in low-exposure environments can become sensitized and go on to develop CBD. The next section discusses an additional source of information on low-level beryllium exposure and CBD: studies of community-acquired CBD in residential areas surrounding beryllium production facilities.

C. Review of Community-Acquired CBD Literature

The literature on community-acquired chronic beryllium disease (CA-CBD) documents cases of CBD among individuals exposed to airborne beryllium at concentrations below the proposed PEL. OSHA notes that these case studies do not provide information on how frequently individuals exposed to very low airborne levels develop CBD and that reconstructed exposure estimates for CA-CBD cases are less reliable than exposure estimates for working populations reviewed in the previous sections. In addition, the cumulative exposure that an occupationally exposed person would accrue at any given exposure concentration is far less than would typically accrue from long-term environmental exposure. The literature on CA-CBD thus has important limitations and is not used as a basis for quantitative risk assessment for CBD from low-level beryllium exposure. Nevertheless, these case reports and the broader CA-CBD literature indicate that individuals exposed to airborne beryllium below the proposed PEL can develop CBD.

Cases of CA-CBD were first reported among residents of Lorain, OH, and Reading, PA, who lived in the vicinity of beryllium plants. More recently, BeLPT screening has been used to identify additional cases of CA-CBD in Reading.

1. Lorain, OH

In 1948, the State of Ohio Department of Public Health conducted an X-ray program surveying more than 6,000 people who lived within 1.5 miles of a Lorain beryllium plant (Eisenbud, 1949; Eisenbud, 1982; Eisenbud, 1998). This survey, together with a later review of all reported cases of CBD in the area, found 13 cases of CBD. All of the residents who developed CBD lived within 0.75 miles of the plant, and none had occupational exposure or lived with beryllium-exposed workers. Among the population of 500 people living within 0.25 miles of the plant, seven residents (1.4 percent) were diagnosed with CBD. Five cases were diagnosed among residents living between 0.25 and 0.5 miles from the plant, one case was diagnosed among residents living between 0.5 and 0.75 miles from the plant, and no cases were found among those living farther than 0.75 miles from the plant (total populations not reported) (Eisenbud, 1998).

Beginning in January 1948, air sampling was conducted using a mobile sampling station to measure atmospheric beryllium downwind from the plant. An approximate concentration of 0.2 μg/m³ was measured at 0.25 miles from the plant's exhaust stack, and concentrations decreased with greater distance from the plant, to 0.003 μg/m³ at a distance of 5 miles (Eisenbud, 1982). A 10-week sampling program was conducted using three fixed monitoring stations within 700 feet of the plant and one station 7,000 feet from the plant. Interpolating the measurements collected at these locations, Eisenbud and colleagues estimated an average airborne beryllium concentration of between 0.004 and 0.02 μg/m³ at a distance of 0.75 miles from the plant. Accounting for the possibility that previous exposures may have been higher due to production level fluctuations and greater use of rooftop emissions, they concluded that the lowest airborne beryllium level associated with CA-CBD in this community was somewhere between 0.01 μg/m³ and 0.1 μg/m³ (Eisenbud, 1982).

2. Reading, PA

Thirty-two cases of CA-CBD were reported in a series of papers published in 1959-1969 concerning a beryllium refinery in Reading (Lieben and Metzner, 1959; Metzner and Lieben, 1961; Dattoli et al., 1964; Lieben and Williams, 1969). The plant, which opened in 1935, manufactured beryllium oxide, alloys and metal, and beryllium tools and metal products (Maier et al., 2008; Sanderson et al., 2001b). In a follow-up study, Maier et al. presented eight additional cases of CA-CBD who had lived within 1.5 miles of the plant (Maier et al., 2008). Individuals with a history of occupational beryllium exposure and those who had resided with occupationally exposed workers were not classified as having CA-CBD.

The Pennsylvania Department of Health conducted extensive environmental sampling in the area of the plant beginning in 1958. Based on samples collected in 1958, Maier et al. stated that most cases identified in their study would typically have been exposed to airborne beryllium at levels between 0.0155 and 0.028 μg/m³ on average, with the potential for some excursions over 0.35 μg/m³ (Maier et al 2008, p. 1015). To characterize exposures to cases identified in the earlier publications, Lieben and Williams cited a sampling program conducted by the Department of Health between January and July 1962, using nine sampling stations located between 0.2 and 4.8 miles from the plant. They reported that 72 percent of 24-hour samples collected were below 0.01 μg/m³. Of samples that exceeded 0.01 μg/m³, most were collected at close proximity to the plant (e.g., 0.2 miles from the plant).

In the early series of publications, cases of CA-CBD were reported among people living both close to the plant (Maier et al., 2008; Dutra, 1948) and up to several miles away. Of new cases identified in the 1968 update, all lived between 3 and 7.5 miles from the plant. Lieben and Williams suggested that some cases of CA-CBD found among more distant residents might have resulted from working or visiting a graveyard closer to the plant (Lieben and Williams, 1969). For example, a milkman who developed CA-CBD had a route in the neighborhood of the plant. Another resident with CA-CBD had worked as a cleaning woman in the area of the plant, and a third worked within a half-mile of the plant.

At the time of the final follow-up study (1968), 11 residents diagnosed with CA-CBD were alive and 21 were deceased. Among those who had died, berylliosis was listed as the cause of death for three, including a 10-year-old girl and two women in their sixties. Fibrosis, granuloma or granulomatosis, and chronic or fibrous pneumonitis were listed as the cause of death for eight more of those deceased. Histologic evidence of CBD was reported for nine of 12 deceased individuals who had been evaluated for it. In addition to showing radiologic abnormalities associated with CBD, all living cases were dyspneic.

Following the 1969 publication by Liebman and Williams, no additional CA-CBD cases were reported in the Reading area until 1999, when a new case was diagnosed. The individual was a 72-year-old woman who had had abnormal chest x-rays for the previous six years (Maier et al., 2008). After the diagnosis of this case, Maier et al. reviewed medical records and/or performed medical evaluations, including BeLPT results for 16 community residents who were referred by family members or an attorney.

Among those referred, eight cases of definite or probable CBD were identified between 1999 and 2002. All eight were women who lived between 0.1 and 1.05 miles from the plant, beginning between 1943-1953 and ending between 1956-2001. Five of the women were considered definite cases of CA-CBD, based on an abnormal blood or lavage cell BeLPT and granulomatous inflammation on lung biopsy. Three probable cases of CA-CBD were identified. One had an abnormal BeLPT and radiography consistent with CBD, but granulomatous disease was not pathologically proven. Two met Beryllium Case Registry epidemiologic criteria for CBD based on radiography, pathology and a clinical course consistent with CBD, but both died before they could be tested for beryllium sensitization. One of the probable cases, who could not be definitively diagnosed with CBD because she died before she could be tested, was the mother of both a definite case and the probable case who had an abnormal BeLPT but did not show granulomatous disease.

The individuals with CA-CBD identified in this study suffered significant health impacts from the disease, including obstructive, restrictive, and gas exchange pulmonary defects in the majority of cases. All but two had abnormal pulmonary physiology. Those two were evaluated at early stages of disease following their mother's diagnosis. Six of the eight women required treatment with prednisone, a step typically reserved for severe cases due to the adverse side effects of steroid treatment. Despite treatment, three had died of respiratory impairment from CBD as of 2002 (Maier et al., 2008). The authors concluded that “low levels of exposures with significant disease latency can result in significant morbidity and mortality” (id., p. 1017).

OSHA notes that compared with the occupational studies discussed in the previous section, there is comparatively sparse information on exposure levels of Lorain and Reading residents. There remains the possibility that some individuals with CA-CBD may have had higher exposures than were known and reported in these studies, or have had unreported exposure to beryllium dust via contact with beryllium-exposed workers. Nevertheless, the studies conducted in Lorain and Reading demonstrate that long-term exposure to the apparent low levels of airborne beryllium, with sufficient disease latency, can lead to serious or fatal CBD. Genetic susceptibility may play a role in cases of CBD among individuals with very low or infrequent exposures to beryllium. The role of genetic susceptibility in the CBD disease process is discussed in detail in section V.D.3.

D. Exposure-Response Literature on Beryllium Sensitization and CBD

To further examine the relationship between exposure level and risk of both sensitization and disease, we next review exposure-response studies in the CBD literature. Many publications have reported that exposure levels correlate with risk, including a small number of exposure-response analyses. Most of these studies examined the association between job-specific beryllium air measurements and prevalence of sensitization and CBD. This section focuses on studies at three facilities that included a more rigorous historical reconstruction of individual worker exposures in their exposure-response analyses.

1. Rocky Flats, CO, Facility

In 2000, Viet et al. published a case-control study of participants in the Rocky Flats Beryllium Health Surveillance Program (BHSP), which was established in 1991 to screen workers at the Department of Energy's Rocky Flats, CO, nuclear weapons facility for beryllium sensitization and evaluate sensitized workers for CBD (Viet et al., 2000). The program, which at the time of publication had tested over 5,000 current and former Rocky Flats employees, had identified a total of 127 sensitized individuals as of 1994 when Viet et al. initiated their study.

Workers were considered sensitized if two BeLPT results were positive, either from two blood draws or from a single blood draw analyzed by two different laboratories. All sensitized individuals were offered clinical evaluation, and 51 were diagnosed with CBD based on positive lung LPT and evidence of noncaseating granulomas upon lung biopsy. The number of sensitized individuals who declined clinical evaluation was not reported. Two cases, one with CBD and one who was sensitized but not diagnosed with CBD, were excluded from the case-control analysis due to reported or potential prior beryllium exposure at a ceramics plant. Another sensitized individual who had not been diagnosed with CBD was excluded because she could not be matched by the study's criteria to a non-sensitized control within the BHSP database. Viet et al. matched a total of 50 CBD cases to 50 controls who were negative on the BeLPT and had the same age (± 3 years), gender, race and smoking status, and were otherwise randomly selected from the database. Using the same matching criteria, 74 sensitized workers who were not diagnosed with CBD were age-, gender-, race-, and smoking status-matched to 74 control individuals who tested negative by the BeLPT from the BHSP database.

Viet et al. developed exposure estimates for the cases and controls based on daily beryllium air samples collected in one of 36 buildings where beryllium was used at Rocky Flats, the Building 444 Beryllium Machine Shop. Over half of the approximately 500,000 industrial hygiene samples collected at Rocky Flats were taken from this building. Air monitoring in other buildings was reported to be limited and inconsistent and, thus, not utilized in the exposure assessment. The sampling data used to develop worker exposure estimates were exclusively Building 444 fixed airhead (FAH) area samples collected at permanent fixtures placed around beryllium work areas and machinery.

Exposure estimates for jobs in Building 444 were constructed for the years 1960-1988 from this database. Viet et al. worked with Rocky Flats industrial hygienists and staff to assign a “building area factor” (BAF) to each of the other buildings, indicating the likely level of exposure in a building relative to exposures in Building 444. Industrial hygienists and staff similarly assigned a job factor (JF) to all jobs, representing the likely level of beryllium exposure relative to the levels experienced by beryllium machinists. A JF of 1 indicated the lowest exposures, and a JF of 10 indicated the highest exposures. For example, administrative work and vehicle operation were assigned a JF of 1, while machining, mill operation, and metallurgical operation were each assigned a JF of 10. Estimated FAH values for each combination of job, building and year in the study subjects' work histories were generated by multiplying together the job and building factors and the mean annual FAH exposure level. Using data collected by questionnaire from each BHSP participant, Viet et al. reconstructed work histories for each case and control, including job title and building location in each year of their employment at Rocky Flats. These work histories and the estimated FAH values were used to generate a cumulative exposure estimate (CEE) for each case and control in the study. A long-term mean exposure estimate (MEE) was generated by dividing each CEE by the individual's number of years employed at Rocky Flats.

Viet et al.' s statistical analysis of the resulting data set included conditional logistic regression analysis, modeling the relationship between risk of each health outcome and log-transformed CEE and MEE. They found highly statistically significant relationships between log-CEE and risk of CBD (coef = 0.837, p = 0.0006) and between log-MEE (coef = 0.855, p = 0.0012) and risk of CBD, indicating that risk of CBD increases with exposure level. These coefficients correspond to odds ratios of 6.9 and 7.2 per 10-fold increase in exposure, respectively. Risk of sensitization without CBD did not show a statistically significant relationship with log-CEE (coef = 0.111, p = 0.32), but showed a nearly-significant relationship with log-MEE (coef = 0.230, p = 0.097).

2. Cullman, AL, Facility

The Cullman, AL, precision machining facility discussed previously was the subject of a case-control study published by Kelleher et al. in 2001. After the diagnosis of an index case of CBD at the plant in 1995, NJMRC researchers worked with the plant to conduct a medical surveillance program using the BeLPT to screen workers biennially for beryllium sensitization and CBD. Of 235 employees screened between 1995 and 1999, 22 (9.4 percent) were found to be sensitized, including 13 diagnosed with CBD (Newman et al., 2001). Concurrently, research was underway by Martyny et al. to characterize the particle size distribution of beryllium exposures generated by processes at this plant (Martyny et al., 2000). The exposure research showed that the machining operations during this time period generated respirable particles (10 μm or less) at the worker breathing zone that made up greater than 50 percent of the beryllium mass. Kelleher et al. used the dataset of 100 personal lapel samples collected by Martyny et al. and other NJMRC researchers in 1996, 1997, and 1999 to characterize exposures for each job in the plant. Following a statistical analysis comparing the samples collected by NJMRC with earlier samples collected at the plant, Kelleher et al. concluded that the 1996-1999 data could be used to represent job-specific exposures from earlier periods.

Detailed work history information gathered from plant data and worker interviews was used in combination with job exposure estimates to characterize cumulative and LTW average beryllium exposures for workers in the surveillance program. In addition to cumulative and LTW exposure estimates based the total mass of beryllium reported in their exposure samples, Kelleher et al. calculated cumulative and LTW estimates based specifically on exposure to particles < 6 μm and particles < 1 μm in diameter.

To analyze the relationship between exposure level and risk of sensitization and CBD, Kelleher et al. performed a case-control analysis using measures of both total beryllium exposure and particle size-fractionated exposure. The analysis included sensitization cases identified in the 1995-1999 surveillance and 206 controls from the group of 215 non-sensitized workers. For nine workers, the researchers could not reconstruct complete job histories. Logistic regression models using categorical exposure variables showed positive associations between risk of sensitization and the six exposure measures tested: Total CEE, total MEE, and variations of CEE and MEE constructed based on particles < 6 μm and < 1 μm in diameter. None of the associations were statistically significant (p < 0.05); however, the authors noted that the dataset was relatively small, with limited power to detect a statistically significant exposure-response relationship.

Although the Viet et al. and Kelleher et al. exposure-response analyses provide valuable insight into exposure-response for beryllium sensitization and CBD, both studies have limitations that affect their suitability as a basis for quantitative risk assessment. Their limitations primarily involve the exposure data used to estimate workers' exposures. Viet et al.' s exposure reconstruction was based on area samples from a single building within a large, multi-building facility. Where possible, OSHA prefers to base risk estimates on exposure data collected in the breathing zone of workers rather than area samples, because data collected in the breathing zone more accurately represent workers' exposures. Kelleher's analysis, on the other hand, was based on personal lapel samples. However, the samples Kelleher et al. used were collected between 1996 and 1999, after the facility had initiated new exposure control measures in response to the diagnosis of a case of CBD in 1995. OSHA believes that industrial hygiene samples collected at the Cullman plant prior to 1996 better characterize exposures prior to the new exposure controls. In addition, since the publication of the Kelleher study, the population has continued to be screened for sensitization and CBD. Data collected on workers hired in 2000 and later, after most exposure controls had been completed, can be used to characterize risk at lower levels of exposure than have been examined in many previous studies.

To better characterize the relationship between exposure level and risk of sensitization and CBD, OSHA developed an independent exposure-response analysis based on a dataset maintained by NJMRC on workers at the Cullman, AL, machining plant. The dataset includes exposure samples collected between 1980 and 2005, and has updated work history and screening information for several hundred workers through 2003. OSHA's analysis of the NJMRC data set is presented in the next section, E. OSHA's Exposure-Response Analysis.

3. Elmore, OH, Facility

After OSHA completed its analysis of the NJMRC data set, Schuler et al. (2012) published a study examining beryllium sensitization and CBD among 264 short-term workers employed at the previously described Elmore, OH plant in 1999. The analysis used a high-quality exposure reconstruction by Virji et al. (2012) and presented a regression analysis of the relationship between beryllium exposure levels and beryllium sensitization and CBD in the short-term worker population. By including only short-term workers, Virji et al. were able to construct participants' exposures with more precision than was possible in studies involving workers exposed for longer durations and in time periods with less exposure sampling. In addition, the focus on short-term workers allowed more precise knowledge of when sensitization and CBD occurred than had been the case for previously published cross-sectional studies of long-term workers. Each participant completed a work history questionnaire and was tested for beryllium sensitization, and sensitized workers were offered further evaluation for CBD. The overall prevalence of sensitization was 9.8 percent (26/264). Twenty-two sensitized workers consented to clinical testing for CBD via transbronchial biopsy. Six of those sensitized were diagnosed with CBD (2.3 percent, 6/264).

Schuler et al. (2012) used logistic regression to explore the relationship between estimated beryllium exposure and sensitization and CBD, using estimates of total, respirable, and submicron mass concentrations. Exposure estimates were constructed using two exposure surveys conducted in 1999: a survey of total mass exposures (4,022 full-shift personal samples) and a survey of size-separated impactor samples (198 samples). The 1999 exposure surveys and work histories were used to estimate long-term lifetime weighted (LTW) average, cumulative, and highest-job-worked exposure for total, respirable, and submicron beryllium mass concentrations.

For beryllium sensitization, logistic models showed elevated odds ratios for average (OR 1.48) and highest job (OR 1.37) exposure for total mass exposure; the OR for cumulative exposure was smaller (OR 1.23) and borderline statistically significant (95 percent CI barely included unity). Relationships between sensitization and respirable exposure estimates were similarly elevated for average (OR 1.37) and highest job (OR 1.32). Among the submicron exposure estimates, only highest job (OR 1.24) had a 95 percent CI that just included unity for sensitization. For CBD, elevated odds ratios were observed only for the cumulative exposure estimates and were similar for total mass and respirable exposure (total mass OR 1.66, respirable (OR 1.68). Cumulative submicron exposure showed an elevated, borderline significant odds ratio (OR 1.58). The odds ratios for average exposure and highest-exposed job were not statistically significantly elevated. Schuler et al. concluded that both total and respirable mass concentrations of beryllium exposure were relevant predictors of risk for beryllium sensitization and CBD.

E. OSHA's Exposure-Response Analysis

OSHA evaluated exposure and health outcome data on a population of workers employed at the Cullman machining facility. NJMRC researchers, with consent and information provided by the facility, compiled a dataset containing employee work histories, medical diagnoses, and air sampling results and provided it to OSHA for analysis. OSHA's contractors from Eastern Research Group (ERG) gathered additional information from (1) two surveys of the Cullman plant conducted by OSHA's contractor (ERG, 2003 and ERG, 2004a), (2) published articles of investigations conducted at the plant by researchers from NJMRC (Kelleher et al., 2001; Madl et al., 2007; Martyny et al., 2000; and Newman et al., 2001), (3) a case file from a 1980 OSHA complaint inspection at the plant, (4) comments submitted to the OSHA docket office in 1976 and 1977 by representatives of the metal machining plant regarding their beryllium control program, and (5) personal communications with the plant's current industrial hygienist (ERG, 2009b) and an industrial hygiene researcher at NJMRC (ERG, 2009a).

1. Plant Operations

The Cullman plant is a leading fabricator of precision-machined and processed materials including beryllium and its alloys, titanium, aluminum, quartz, and glass (ERG, 2009b). The plant has approximately 210 machines, primarily mills and lathes, and processes large quantities of beryllium on an annual basis. The plant provides complete fabrication services including ultra-precision machining; ancillary processing (brazing, ion milling, photo etching, precision cleaning, heat treating, stress relief, thermal cycling, mechanical assembly, and chemical milling/etching); and coatings (plasma spray, anodizing, chromate conversion coating, nickel sulfamate plate, nickel plate, gold plate, black nickel plate, copper plate/strike, passivation, and painting). Most of the plant's beryllium operations involve machining beryllium metal and high beryllium content composite materials (beryllium metal/beryllium oxide metal composites called E-Metal or E-Material), with occasional machining of beryllium oxide/metal matrix (such as AlBeMet, aluminum beryllium matrix) and beryllium-containing alloys. E-Materials such as E-20 and E-60 are currently processed in the E-Cell department.

The 120,000 square-foot plant has two main work areas: a front office area and a large, open production shop. Operations in the production shop include inspection of materials, machining, polishing, and quality assurance. The front office is physically separated from the production shop. Office workers enter through the front of the facility and have access to the production shop through a change room where they must don laboratory coats and shoe covers to enter the production area. Production workers enter the shop area at the rear of the facility where a change/locker room is available to change into company uniforms and work shoes. Support operations are located in separate areas adjacent to the production shop and include management and administration, sales, engineering, shipping and receiving, and maintenance. Management and administrative personnel include two groups: those primarily working in the front offices (front office management) and those primarily working on the shop floor (shop management).

In 1974, the company moved its precision machining operations to the plant's current location in Cullman. Workplace exposure controls reportedly did not change much until the diagnosis of an index case of CBD in 1995. Prior to 1995, exposure controls for machining operations primarily included a low volume/high velocity (LVHV) central exhaust system with operator-adjusted exhaust pickups and wet machining methods. Protective clothing, gloves, and respiratory protection were not required. After the diagnosis, the facility established an in-house target exposure level of 0.2 μg/m³, installed change/locker rooms for workers entering the production facility, eliminated pressurized air hoses, discouraged the use of dry sweeping, initiated biennial medical surveillance using the BeLPT, and implemented annual beryllium hazard awareness training.

In 1996, the company instituted requirements for work uniforms and dedicated work shoes for production workers, eliminated dry sweeping in all departments, and purchased high-efficiency particulate air (HEPA) filter vacuum cleaners for workplace cleanup and decontamination. Major engineering changes were also initiated in 1996, including the purchase of a new local exhaust ventilation (LEV) system to exhaust machining operations producing finer aerosols (e.g., dust and fume versus metal chips). The facility also began installing mist eliminators for each machine. Departments affected by these changes included cutter grind (tool and die), E-cell, electrical discharge machining (EDM), flow lines, grind, lapping, and optics. Dry machining operations producing chips were exhausted using the existing LVHV exhaust system (ERG, 2004a). In the course of making the ventilation system changes, old ductwork and baghouses were dismantled and new ductwork and air cleaning devices were installed. The company also installed Plexiglas enclosures on machining operations in 1996-1997, including the lapping, deburring, grinding, EDM, and tool and die operations. In 1998, LEV was installed in EDM and modified in the lap, deburr, and grind departments.

Most exposure controls were reportedly in place by 2000 (ERG, 2009a). In 2004, the plant industrial hygienist reported that all machines had LEV and about 65 percent were also enclosed with either partial or full enclosures to control the escape of machining coolant (ERG, 2004b). Over time, the facility has built enclosures for operations that consistently produce exposures greater than 0.2 μg/m³. The company has never required workers to use gloves or other PPE.

2. Air Sampling Database and Job Exposure Matrix (JEM)

The NJMRC dataset includes industrial hygiene sampling results collected by the plant (1980-1984 and 1995-2005) and NJMRC researchers (June 1996 to February 1997 and September 1999), including 4,370 breathing zone (personal lapel) samples and 712 area samples (ERG, 2004b). Limited air sampling data is available before 1980 and no exposure data appears to be available for the 10-year time period 1985 through 1994. A review of the NJMRC air sampling database from 1995 through 2005 shows a significant increase in the number of air samples collected beginning in 2000, which the plant industrial hygienist attributes to an increase in the number of air sampling pumps (from 5 to 23) and the purchase of an automated atomic absorption spectrophotometer.

ERG used the personal breathing zone sampling results contained in the sample database to quantify exposure levels for each year and for several-year periods. Separate exposure statistics were calculated for each job included in the job history database. For each job included in the job history database, ERG estimated the arithmetic mean, geometric mean, median, minimum, maximum, and 95th percentile value for the available exposure samples. Prior to generating these statistics ERG made several adjustments. After consultation with researchers at NJMRC, four particularly high exposures were identified as probably erroneous and excluded from calculations. In addition, a 1996 sample for the HS (Health and Safety) process was removed from the sample calculations after ERG determined it was for a non-employee researcher visiting the facility.

Most samples in the sample database for which sampling times were recorded were long-term samples: 2,503 of the 2,557 (97.9 percent) breathing zone samples with sampling time recorded had times greater than or equal to 400 minutes. No adjustments were made for sampling time, except in the case of four samples for the “maintenance” process for 1995. These results show relatively high values and exceptionally short sampling times consistent with the nature of much maintenance work, marked by short-term exposures and periods of no exposure. The four 1995 maintenance samples were adjusted for an eight-hour sampling time assuming that the maintenance workers received no further beryllium exposure over the rest of their work shift.

OSHA examined the database for trends in exposure by reviewing sample statistics for individual years and grouping years into four time periods that correspond to stages in the plant's approach to beryllium exposure control. These were: 1980-1995, a period of relatively minimal control prior to the 1995 discovery of a case of CBD among the plant's workers; 1996-1997, a period during which some major engineering controls were in the process of being installed on machining equipment; 1998-1999, a period during which most engineering controls on the machining equipment had been installed; and 2000-2003, a period when installation of all exposure controls on machining equipment was complete and exposures very low throughout the plant. Table VI-4 below summarized the available data for each time period. As the four probable sampling errors identified in the original data set are excluded here, arithmetic mean values are presented.

Reviewing the revised statistics for individual years for different groupings, OSHA noted that exposures in the 1996-1997 period were for some machining jobs equivalent to, or even higher than, exposure levels recording during the 1980-1995 period. During 1996-1997, major engineering controls were being installed, but exposure levels were not yet consistently reduced.

Table VI-5 below summarizes exposures for the four time periods in jobs other than beryllium machining. These include jobs such as administrative work, health and safety, inspection, toolmaking (`Tool' and `Cgrind'), and others. A description of jobs by title is available in the risk assessment background document.

From Table VI-5, it is evident that exposure samples are not available for many non-machining jobs prior to 2000. Where samples are available before 2000, sample numbers are small, particularly prior to 1998. In jobs for which exposure values are available in 1998-1999 and 2000-2003, exposures appear either to decline from 1998-1999 to 2000-2003 (Assembly, Chem, Inspection, Maintenance) or to be roughly equivalent (Administration, Cgrind, Engineering, Msupp, PCIC, and Spec). Among the jobs with exposure samples prior to 1998, most had very few (1-5) samples, with the exception of Ecell (13 samples in 1996-1997). Based on this limited information, it appears that exposures declined from the period before the first dentification of a CBD case to the period in which exposure controls were introduced.

Because exposure results from 1996-1997 were not found to be consistently reduced in comparison to the 1985-1995 period in primary machining jobs, these two periods were grouped together in the JEM. Exposure monitoring for jobs other than the primary machining operations were represented by a single mean exposure value for 1980-2003. As respiratory protection was not routinely used at the plant, there was no adjustment for respiratory protection in workers' exposure estimates. The job exposure matrix is presented in full in the background document for the quantitative risk assessment.

3. Worker Exposure Reconstruction

The work history database contains job history records for 348 workers, including start years, duration of employment, and percentage of worktime spent in each job. One hundred ninety-eight of the workers had been employed at some point in primary machining jobs, including deburring, EDM, grinding, lapping, lathing, and milling. The remainder worked only in non-primary machining jobs, such as administration, engineering, quality control, and shop management. The total number of years worked at each job are presented as integers, leaving some uncertainty regarding the worker's exact start and end date at the job.

Based on these records and the JEM described previously, ERG calculated cumulative and average exposure estimates for each worker in the database. Cumulative exposure was calculated as, Σ_i^e_i^t_i, where e(i) is the exposure level for job (i), and t(i) is the time spent in job (i). Cumulative exposure was divided by total exposure time to estimate each worker's long-term average exposure. These exposures were computed in a time-dependent manner for the statistical modeling. For workers with beryllium sensitization or CBD, exposure estimates excluded exposures following diagnosis.

Workers who were employed for long time periods in jobs with low-level exposures tend to have low average and cumulative exposures due to the way these measures are constructed, incorporating the worker's entire work history. As discussed in the Health Effects chapter, higher-level exposures or short-term peak exposures such as those encountered in machining jobs may be highly relevant to risk of sensitization. Unfortunately, because it is not possible to continuously monitor individuals' beryllium exposure levels and sensitization status, it is not known exactly when workers became sensitized or what their “true” peak exposures leading up to sensitization were. Only a rough approximation of the upper levels of exposure a worker experienced is possible. ERG constructed a third type of exposure estimate reflecting the exposure level associated with the highest-exposure job (HEJ) and time period experienced by each worker. This exposure estimate (HEJ), the cumulative exposure estimate, and the average exposure were used in the quartile analysis and statistical analyses.

4. Prevalence of Sensitization and CBD

In the database provided to OSHA, seven workers were reported as sensitized only. Sixteen workers were listed as sensitized and diagnosed with CBD upon initial clinical evaluation. Three workers, first shown to be sensitized only, were later diagnosed with CBD. Tables VI-6, VI-7, and VI-8 below present the prevalence of sensitization and CBD cases across several categories of lifetime-weighted (LTW) average, cumulative, and highest-exposed job (HEJ) exposure. Exposure values were grouped by quartile. Note that all workers with CBD are also sensitized. Thus, the columns “Total Sensitized” and “Total %” refer to all sensitized workers in the dataset, including workers with and without a diagnosis of CBD.

Table VI-6 shows increasing prevalence of total sensitization and CBD with increasing LTW average exposure, measured both as average and cumulative exposure. The lowest prevalence of sensitization and CBD was observed among workers with average exposure levels less than or equal to 0.08 μg/m³, where two sensitized workers (2.2 percent) including one case of CBD (1.0 percent) were found. The sensitized worker in this category without CBD had worked at the facility as an inspector since 1972, one of the lowest-exposed jobs at the plant. Because the job was believed to have very low exposures, it was not sampled prior to 1998. Thus, estimates of exposures in this job are based on data from 1998-2003 only. It is possible that exposures earlier in this worker's employment history were somewhat higher than reflected in his estimated average exposure. The worker diagnosed with CBD in this group had been hired in 1996 in production control, and had an estimated average exposure of 0.08 μg/m³. He was diagnosed with CBD in 1997.

The second quartile of LTW average exposure (0.081—0.18 μg/m³) shows a marked rise in overall prevalence of beryllium-related health effects, with six workers sensitized (8.2 percent), of whom four (5.5 percent) were diagnosed with CBD. Among six sensitized workers in the third quartile (0.19—0.50 μg/m³), all were diagnosed with CBD (7.8 percent). Another increase in prevalence is seen from the third to the fourth quartile, with 12 cases of sensitization (15.4 percent), including eight (10.3 percent) diagnosed with CBD.

The quartile analysis of cumulative exposure also shows generally increasing prevalence of sensitization and CBD with increasing exposure. As shown in Table VI-7, the lowest prevalences of CBD and sensitization are in the first two quartiles of cumulative exposure (0.0-0.147 μg/m³-yrs, 0.148-1.467 μg/m³-yrs). The upper bound on this cumulative exposure range, 1.467 μg/m³-yrs, is the cumulative exposure that a worker would have if exposed to beryllium at a level of 0.03 μg/m³ for a working lifetime of 45 years; 0.15 μg/m³ for ten years; or 0.3 μg/m³ for five years.

A sharp increase in prevalence of sensitization and CBD and total sensitization occurs in the third quartile (1.468-7.008 μg/m³-yrs), with roughly similar levels of both in the highest group (7.009-61.86 μg/m³-yrs). Cumulative exposures in the third quartile would be experienced by a worker exposed for 45 years to levels between 0.03 and 0.16 μg/m³, for 10 years to levels between 0.15 and 0.7 μg/m³, or for five years to levels between 0.3 and 1.4 μg/m³.

When workers' exposures from their highest-exposed job are considered, the exposure-response pattern is similar to that for LTW average exposure in the lower quartiles (Table VI-8). The lowest prevalence is observed in the first quartile (0.0-0.86 μg/m³), with sharply rising prevalence from first to second and second to third exposure quartiles. The prevalence of sensitization and CBD in the top quartile (0.954-2.213 μg/m³) decreases relative to the third, with levels similar to the overall prevalence in the dataset. Many workers in the highest exposure quartiles are long-time employees, who were hired during the early years of the shop when exposures were highest. One possible explanation for the drop in prevalence in the highest exposure quartiles is that highly-exposed workers from early periods may have developed CBD and left the plant before sensitization testing began in 1995.

It is of some value to compare the prevalence analysis of the Cullman (NJMRC) data set with the results of the Reading and Tucson studies discussed previously. An exact comparison is not possible, in part because the Reading and Tucson exposure values are associated with jobs and the NJMRC values are estimates of lifetime weighted average, cumulative, and highest-exposed job (HEJ) exposures for individuals in the data set. Nevertheless, OSHA believes it is possible to very roughly compare the results of the Reading and Tucson studies and the results of the NJMRC prevalence analysis presented above. As discussed in detail below, OSHA found a general consistency between the prevalence of sensitization and CBD in the quartiles of average exposure in the NJMRC data set and the prevalence of sensitization and CBD at the Reading and Tucson plants for similar exposure values.

Personal lapel samples collected at the Reading plant between 1995 and 2000 were relatively low overall (median of 0.073 μg/m³), with higher exposures (median of 0.149 μg/m³) concentrated in the wire annealing and pickling process (Schuler et al., 2005). Exposures in the Reading plant in this time period were similar to the second-quartile average (Table VI-6-0.081-0.18 μg/m³). The prevalence of sensitization observed in the NJMRC second quartile was 8.2 percent and appears roughly consistent with the prevalence of sensitization among Reading workers in the mid-1990s (11.5 percent). The reported prevalence of CBD (3.9 percent) among the Reading workforce was also consistent with that observed in the second NJMRC quartile (5.5 percent), After 2000, exposure controls reduced exposures in most Reading jobs to median levels below 0.03 μg/m³, with a median value of 0.1 μg/m³ for the wire annealing and pickling process. The wire annealing and pickling process was enclosed and stringent respirator and skin protection requirements were applied for workers in that area after 2002, essentially eliminating airborne and dermal exposures for those workers. Thomas et al. (2009) reported that one of 45 workers (2.2 percent) hired after the enclosure in 2002 was confirmed as sensitized, a value in line with the sensitization prevalence observed in the lowest quartiles of average exposure (2.2 percent, 0.0-0.08 μg/m³).

As with Reading, the prevalence of sensitization observed at Tucson and in the NJMRC data set are not exactly comparable due to the different natures of the exposure estimates. Nevertheless, in a rough sense the results of the Tucson study and the NJMRC prevalence analysis appear similar. In Tucson, a 1998 BeLPT screening showed that 9.5 percent of workers hired after 1992 were sensitized (Henneberger et al., 2001). Personal full-shift exposure samples collected in production jobs between 1994 and 1999 had a median of 0.2 μg/m³ (0.1 μg/m³ for non-production jobs). In the NJMRC data set, a sensitization prevalence of 8.2 percent was seen among workers with average exposures between 0.081 and 0.18 μg/m³. At the time of the 1998 screening, workers hired after 1992 had a median one year since first beryllium exposure and, therefore, CBD prevalence was only 1.4 percent. This prevalence is likely an underestimate since CBD often requires more than a year to develop. Longer-term workers at the Tucson plant with a median 14 years since first beryllium exposure had a 9.1 percent prevalence of CBD. There was a 5.5 percent prevalence of CBD among the entire workforce (Henneberger et al., 2001). As with the Reading plant employees, this reported prevalence is reasonably consistent with the 5.5 percent CBD prevalence observed in the second NJMRC quartile.

Beginning in 1999, the Tucson facility instituted strict requirements for respiratory protection and other PPE, essentially eliminating airborne and dermal exposure for most workers. After these requirements were put in place, Cummings et al. (2007) reported only one case of sensitization (1 percent; associated with a PPE failure) among 97 workers hired between 2000 and 2004. This appears roughly in line with the sensitization prevalence of 2.2 percent observed in the lowest quartiles of average exposure (0.0-0.08 μg/m³) in the NJMRC data set.

While the literature analysis presented here shows a clear reduction in risk with well-controlled airborne exposures (≤ 0.1 μg/m³ on average) and protection from dermal exposure, the level of detail presented in the published studies limits the Agency's ability to characterize risk at all the alternate PELs OSHA is considering. To better understand these risks, OSHA used the NJMRC dataset to characterize risk of sensitization and CBD among workers exposed to each of the alternate PELs under consideration in the proposed beryllium rule.

F. OSHA's Statistical Modeling

OSHA's contractor performed a complementary log-log proportional hazards model using the NJMRC data set. The proportional hazards model is a generalization of logistic regression that allows for time-dependent exposures and differential time at risk. The proportional hazards model accounts for the fact that individuals in the dataset are followed for different amounts of time, and that their exposures change over time. The proportional hazards model provides hazards ratios, which estimate the relative risk of disease at a specified time for someone with exposure level 1 compared to exposure level 2. To perform this analysis, OSHA's contractor constructed exposure files with time-dependent cumulative and average exposures for each worker in the data set in each year that a case of sensitization or CBD was identified. Workers were included in only those years after they started working at the plant and continued to be followed. Sensitized cases were not included in analysis of sensitization after the year in which they were identified as being sensitized, and CBD cases were not included in analyses of CBD after the year in which they were diagnosed with CBD. Follow-up is censored after 2002 because work histories were deemed to be less reliable after that date.

The results of the discrete proportional hazards analyses are summarized in Tables VI-9-12 below. All coefficients used in the models are displayed, including the exposure coefficient, the model constant for diagnosis in 1995, and additional exposure-independent coefficients for each succeeding year (1996-1999 for sensitization and 1996-2002 for CBD) of diagnosis that are fit in the discrete time proportional hazards modeling procedure. Model equations and variables are explained more fully in the companion risk assessment background document.

Relative risk of sensitization increased with cumulative exposure (p = 0.05). A positive, but not statistically significant, association was observed with LTW average exposure (p = 0.09). The association was much weaker for exposure duration (p = 0.31), consistent with the expected biological action of an immune hypersensitivity response where onset is believed to be more dependent on the concentration of the sensitizing agent at the target site rather than the number of years of occupational exposure. The association was also much weaker for highest-exposed job (HEJ) exposure (p = 0.3).

The proportional hazards models for the CBD endpoint (Tables VI-13 through 16 below) showed positive relationships with cumulative exposure (p = 0.09) and duration of exposure (p = 0.10). However, the association with the cumulative exposure metric was not as strong as that for sensitization, probably due to the smaller number of CBD cases. LTW average exposure and HEJ exposure were not closely related to relative risk of CBD (p-values > 0.5).

In addition to the models reported above, comparable models were fit to the upper 95 percent confidence interval of the HEJ exposure; log-transformed cumulative exposure; log-transformed LTW average exposure; and log-transformed HEJ exposure. Each of these measures was positively but not significantly associated with sensitization.

OSHA used the proportional hazards models based on cumulative exposure, shown in Tables VI-9 and VI-13, to derive quantitative risk estimates. Of the metrics related to exposure level, the cumulative exposure metric showed the most consistent association with sensitization and CBD in these models. Table VI-17 summarizes these risk estimates for sensitization and the corresponding 95 percent confidence intervals separately for 1995 and 1999, the years with the highest and lowest baseline rates, respectively. The estimated risks for CBD are presented in VI-18. The expected number of cases is based on the estimated conditional probability of being a case in the given year. The models provide time-specific point estimates of risk for a worker with any given exposure level, and the corresponding interval is based on the uncertainty in the exposure coefficient (i.e., the predicted values based on the 95 percent confidence limits for the exposure coefficient).

Each estimate represents the number of sensitized workers the model predicts in a group of 1000 workers at risk during the given year with an exposure history at the specified level and duration. For example, in the exposure scenario where 1000 workers are occupationally exposed to 2 μg/m³ for 10 years in 1995, the model predicts that about 56 (55.7) workers would be sensitized that year. The model for CBD predicts that about 42 (41.9) workers would be diagnosed with CBD that year.

The statistical modeling analysis predicts high risk of both sensitization (96-394 cases per 1000, or 9.6-39.4 percent) and CBD (44-313 cases per 1000, or 4.4-31.3 percent) at the current PEL of 2 μg/m³ for an exposure duration of 45 years (90 μg/m³-yr). The predicted risks of < 8.2-39.9 per 1000 (0.8-3.9 percent) cases of sensitization or 3.6 to 30.0 per 1000 (0.4-3 percent) cases of CBD are substantially less for a 45-year exposure at the proposed PEL, 0.2 μg/m³ (9 μg/m³-yr).

The model estimates are not directly comparable to prevalence values discussed in previous sections. They assume a group without turnover and are based on a comparison of unexposed and hypothetically exposed workers at specific points in time, whereas the prevalence analysis simply reports the percentage of workers at the Cullman plant with sensitization or CBD in each exposure category. Despite the difficulty of direct comparison, the level of risk seen in the prevalence analysis and predicted in the modeling analysis appear roughly similar at low exposures. In the second quartile of cumulative exposure (0.148-1.467 μg/m³-yr), prevalence of sensitization and CBD was 2.5 percent. This is roughly congruent with the model predictions for workers with cumulative exposures between 0.5 and 1 μg/m³-yr: 6.3-31.3 cases of sensitization per 1000 workers (0.6-3.1 percent) and 2.8 to 23.5 cases of CBD per 1000 workers (0.28-2.4 percent). As discussed in the background document for this analysis, most workers in the data set had low cumulative exposures (roughly half below 1.5 μg/m³-years). It is difficult to make any statement about the results at higher levels, because there were few workers with high exposure levels and the higher quartiles of cumulative exposure include an extremely wide range of exposures. For example, the highest quartile of cumulative exposure was 7.009-61.86 μg/m³-yr. This quartile, which showed an 11.3 percent prevalence of sensitization and 8.8 percent prevalence of CBD, includes the cumulative exposure that a worker exposed for 45 years at the proposed PEL would experience (9 μg/m³-yr) near its lower bound. Its upper bound approaches the cumulative exposure that a worker exposed for 45 years at the current PEL would experience (90 μg/m³-yr).

Due to limitations including the size of the dataset, relatively limited exposure data from the plant's early years, study size-related constraints on the statistical analysis of the dataset, and limited follow-up time on many workers, OSHA must interpret the model-based risk estimates presented in Tables VI-17 and VI-18 with caution. The Cullman study population is a relatively small group and can support only limited statistical analysis. For example, its size precludes inclusion of multiple covariates in the exposure-response models or a two-stage exposure-response analysis to model both sensitization and the subsequent development of CBD within the subpopulation of sensitized workers. The limited size of the Cullman dataset is characteristic of studies on beryllium-exposed workers in modern, low-exposure environments, which are typically small-scale processing plants (up to several hundred workers, up to 20-30 cases). However, these recent studies also have important strengths: They include workers hired after the institution of stringent exposure controls, and have extensive exposure sampling using full-shift personal lapel samples. In contrast, older studies of larger populations tend to have higher exposures, less exposure data, and exposure data collected in short-term samples or outside of workers' breathing zones.

Another limitation of the Cullman dataset, which is common to recent low-exposure studies, is the short follow-up time available for many of the workers. While in some cases CBD has been known to develop in short periods (< 2 years), it more typically develops over a longer time period. Sensitization occurs in a typically shorter time frame, but new cases of sensitization have been observed in workers exposed to beryllium for many years. Because the data set is limited to individuals then working at the plant, the Cullman data set cannot capture CBD occurring among workers who retire or leave the plant. OSHA expects that the dataset does not fully represent the risk of sensitization, and is likely to particularly under-represent CBD among workers exposed to beryllium at this facility. The Agency believes the short follow-up time to be a significant source of uncertainty in the statistical analysis, a factor likely to lead to underestimation of risk in this population.

A common source of uncertainty in quantitative risk assessment is the series of choices made in the course of statistical analysis, such as model type, inclusion or exclusion of additional explanatory variables, and the assumption of linearity in exposure-response. Sensitivity analyses and statistical checks were conducted to test the validity of the choices and assumptions in the exposure-response analysis and the impact of alternative choices on the end results. These analyses did not yield substantially different results, adding to OSHA's confidence in the conclusions of its preliminary risk assessment.

OSHA's contractor examined whether smoking and age were confounders in the exposure-response analysis by adding them as variables in the discrete proportional hazards model. Neither smoking status nor age was a statistically significant predictor of sensitization or CBD. The model coefficients, 95 percent confidence intervals, and p values can be found in the background document. A sensitivity analysis was done using the standard Cox model that treats survival time as continuous rather than discrete. The model coefficients with the standard Cox using cumulative exposure were 0.025 and very similar to the 0.03 reported in Tables VI-9 and VI-13 above. The interaction between exposure and follow-up time was not significant in these models, suggesting that the proportional hazard assumption should not be rejected. The proportional hazards model assumes a linear relationship between exposure level and relative risk. The linearity assumption was assessed using a fractional polynomial approach. For both sensitization and CBD, the best-fitting fractional polynomial model did not fit significantly better than the linear model. This result supports OSHA's use of the linear model to estimate risk. The details of these statistical analyses can be found in the background document.

The possibility that the number of times a worker has been tested for sensitization might influence the probability of a positive test was examined (surveillance bias). Surveillance bias could occur if workers were tested because they showed some sign of disease, and not tested otherwise. It is also possible that the original analysis included erroneous assumptions about the dates of testing for sensitization and CBD. OSHA's contractor performed a sensitivity analysis, modifying the original analysis to gauge the effect of different assumptions about testing dates. In the sensitivity analysis, the exposure coefficients increased for all four indices of exposure when the sensitization analysis was restricted to times when cohort members were assumed to be tested. The exposure coefficient was statistically significant for duration of exposure but not for cumulative, LTW average, or HEJ exposure. The increase in exposure coefficients suggests that the original models may have underestimated the exposure-response relationship for sensitization and CBD.

Errors in exposure measurement are a common source of uncertainty in quantitative risk assessments. Because errors in high exposures can heavily influence modeling results, OSHA's contractor performed sensitivity analyses excluding the highest 5 percent of cumulative exposures (those above 25.265 μg/m³-yrs) and the highest 10 percent of cumulative exposures (those above 18.723 μg/m³-yrs). As discussed in more detail in the background document, exposure coefficients were not statistically significant when these exposures were dropped. This is not surprising, given that the exclusion of high exposure values reduced the size of the data set. Prior to excluding high exposure values, the data set was already relatively small and many of the exposure coefficients were non-significant or weakly significant in the original analyses. As a result, the sensitivity analyses did not provide much information about uncertainty due to exposure measurement error and its effects on the modeling analysis.

Particle size, particle surface area, and beryllium compound solubility are believed to be important factors influencing the risk of sensitization and CBD among beryllium-exposed workers. The workers at the Cullman machining plant were primarily handling insoluble beryllium compounds, such as beryllium metal and beryllium metal/beryllium oxide composites. Particle size distributions from a limited number of airborne beryllium samples collected just after the 1996 installation of engineering controls indicate worker exposure to a substantial proportion of respirable particulates. There was no available particle size data for the 1980 to 1995 period prior to installation of engineering controls when total beryllium mass exposure levels were greatest. Particle size data was also lacking from 1998 to 2003 when additional control measures were in place and total beryllium mass exposures were lowest. For these reasons, OSHA was not able to quantitatively account for the influence of particle size and solubility in developing the risk estimates based on the Cullman data set. However, it is not unreasonable to expect the CBD experienced by this cohort to generally reflect the risk from exposure to beryllium that is relatively insoluble and enriched with respirable particles. As explained previously, the role of particle size and surface area on risk of sensitization is more difficult to predict.

Additional uncertainty is introduced when extrapolating the quantitative estimates presented above to operations that process beryllium compounds that have different solubility and particle characteristics than those encountered at the Cullman machining plant. OSHA does not have sufficient information to quantitatively assess the degree to which risks of beryllium sensitization and CBD based on the NJMRC data may be impacted in workplaces where such beryllium forms and processes are used. However, OSHA does not expect this uncertainty to alter its qualitative conclusions with regard to the risk at the current PEL and at alternate PELs as low as 0.1 μg/m³. The existing studies provide clear evidence of sensitization and CBD risk among workers exposed to a number of beryllium forms as a result of different processes such as beryllium machining, beryllium-copper alloy production, and beryllium ceramics production. The Agency believes all of these forms of beryllium exposure contribute to the overall risk of sensitization and CBD among beryllium-exposed workers.

G. Lung Cancer

OSHA considers lung cancer to be an important health endpoint for beryllium-exposed workers. The International Agency for Research on Cancer (IARC), National Toxicology Program (NTP), and American Conference of Governmental Industrial Hygienists (ACGIH) have all classified beryllium as a known human carcinogen. The National Academy of Sciences (NAS), Environmental Protection Agency, the Agency for Toxic Substances and Disease Registry (ATSDR), the National Institute of Occupational Safety and Health (NIOSH), and other reputable scientific organizations have reviewed the scientific evidence demonstrating that beryllium is associated with an increased incidence of cancer. OSHA also has performed an extensive review of the scientific literature regarding beryllium and cancer. This includes an evaluation of human epidemiological, animal cancer, and mechanistic studies described in the Health Effects section of this preamble. Based on the weight of evidence, the Agency has preliminarily determined beryllium to be an occupational carcinogen.

Although epidemiological and animal evidence supports a conclusion of beryllium carcinogenicity, there is considerable uncertainty surrounding the mechanism of carcinogenesis for beryllium. The evidence for direct genotoxicity of beryllium and its compounds has been limited and inconsistent (NAS, 2008; IARC, 1993; EPA, 1998; NTP, 2002; ATSDR, 2002). One plausible pathway for beryllium carcinogenicity described in the Health Effects section of this preamble includes a chronic, sustained neutrophilic inflammatory response that induces epigenetic alterations leading to the neoplastic changes necessary for carcinogenesis. The National Cancer Institute estimates that nearly one-third of all cancers are caused by chronic inflammation (NCI, 2009). This mechanism of action has also been hypothesized for crystalline silica and other agents that are known to be human carcinogens but have limited evidence of genotoxicity.

OSHA's review of epidemiological studies of lung cancer mortality among beryllium workers found that most did not characterize exposure levels sufficiently for exposure-response analysis. However, one NIOSH study evaluated the association between beryllium exposure and lung cancer mortality based on data from a beryllium processing plant in Reading, PA (Sanderson et al., 2001a). As discussed in the Health Effects section of this preamble, this case-control study evaluated lung cancer incidence in a cohort of workers employed at the plant from 1940 to 1969 and followed through 1992. For each lung cancer victim, 5 age- and race-matched controls were selected by incidence density sampling, for a total of 142 lung cancer cases and 710 controls.

Between 1971 and 1992, the plant collected close to 7,000 high volume filter samples consisting of both general area and short-term, task-based breathing zone measurements for production jobs and exclusively area measurements for office, lunch, and laboratory areas (Sanderson et al., 2001b). In addition, a few (< 200) impinger and high-volume filter samples were collected by other organizations between 1947 and 1961, and about 200 6-to-8-hour personal samples were collected in 1972 and 1975. Daily-weighted-average (DWA) exposure calculations based on the impinger and high-volume samples collected prior to the 1960s showed that exposures in this period were extremely high. For example, about half of production jobs had estimated DWAs ranging between 49 and 131 μg/m³ in the period 1935-1960, and many of the “lower-exposed” jobs had DWAs of approximately 20-30 μg/m³ (Table II, Sanderson et al., 2001b). Exposures were reported to have decreased between 1959 and 1962 with the installation of ventilation controls and improved housekeeping and following the passage of the OSH Act in 1970. While no exposure measurements were available from the period 1961-1970, measurements from the period 1971-1980 showed a dramatic reduction in exposures plant-wide. Estimated DWAs for all jobs in this period ranged from 0.1 μg/m³ to 1.9 μg/m³. Calendar-time-specific beryllium exposure estimates were made for every job based on the DWA calculations and were used to estimate workers' cumulative, average, and maximum exposures. Exposure estimates were lagged by 10 and 20 years in order to account for exposures that did not contribute to lung cancer because they occurred after the induction of cancer.

Results of a conditional logistic regression analysis showed an increased risk of lung cancer in workers with higher exposures when dose estimates were lagged by 10 and 20 years (Sanderson et al., 2001a). The authors noted that there was considerable uncertainty in the estimation of exposure in the 1940s and 1950s and the shape of the dose-response curve for lung cancer. NIOSH later reanalyzed the data, adjusting for potential confounders of hire age and birth year (Schubauer-Berigan et al., 2008). The study reported a significant increasing trend (p<0.05) in the odds ratio when increasing quartiles of average (log transformed) exposure were lagged by 10 years. However, it did not find a significant trend when quartiles of cumulative (log transformed) exposure were lagged by 0, 10, or 20 years.

OSHA is interested in lung cancer risk estimates from a 45-year (i.e., working lifetime) exposure to beryllium levels between 0.1 μg/m³ and 2 μg/m³. The majority of case and control workers in the Sanderson et al. case-control analysis were first hired during the 1940s when exposures were extremely high (estimated DWAs > 20 μg/m³ for most jobs). The cumulative, average, and maximum beryllium exposure concentration estimates for the 142 known lung cancer cases were: 46.06 ± 9.3μg/m³-days, 22.8 ± 3.4 μg/m³, and 32.4 ± 13.8 μg/m³, respectively. About two-thirds of cases and half of controls worked at the plant for less than a year. Thus, a risk assessment based on this exposure-response analysis would need to extrapolate from very high to very low exposures, based on a working population with extremely short tenure. While OSHA risk assessments must often make extrapolations to estimate risk within the range of exposures of interest, the Agency acknowledges that these issues of short tenure and extremely high exposures would create substantial uncertainty in a risk assessment based on this study population.

In addition, the relatively high exposures of even the least-exposed workers in the NIOSH study may create methodological issues for the lung cancer case-control study design. Mortality risk is expressed as an odds ratio that compares higher exposure quartiles to the lowest quartile. It is preferable that excess risks attributable to occupational beryllium be determined relative to an unexposed or minimally exposed reference population. However, in the NIOSH study workers in the lowest quartile were exposed well above the OSHA PEL (average exposure <11.2 μg/m³) and may have had a significant lung cancer risk. This issue would introduce further uncertainty in lung cancer risks estimated from this epidemiological study.

In 2010, researchers at NIOSH published a quantitative risk assessment based on an update of the Reading cohort analyzed by Sanderson et al., as well as workers from two smaller plants (Schubauer-Berigan et al., 2010b). This new risk assessment addresses several of OSHA's concerns regarding the Sanderson et al. analysis. The new cohort was exposed, on average, to lower levels of beryllium and had fewer short-term workers. Finally, the updated cohorts followed the populations through 2005, increasing the length of follow-up time overall by an additional 17 years of observation. For these reasons, OSHA considers the Schubauer-Berigan risk analysis more appropriate than the Sanderson et al. analysis for its preliminary risk assessment.

The cohort studied by Schubauer-Berigan et al. included 5,436 male workers who had worked for at least two days at the Reading facility and beryllium processing plants at Hazleton PA and Elmore OH prior to 1970. The authors developed job-exposure matrices (JEMs) for the three plants based on extensive historical exposure data, primarily short-term general area and personal breathing zone samples, collected on a quarterly basis from a wide variety of operations. These samples were used to create daily weighted average (DWA) estimates of workers' full-shift exposures, using records of the nature and duration of tasks performed by workers during a shift. Details on the JEM and DWA construction can be found in Sanderson et al. (2001a), Chen et al. (2001), and Couch et al. (2010).

Workers' cumulative exposures (μg/m³-days) were estimated by summing daily average exposures (assuming five workdays per week). To estimate mean exposure (μg/m³), cumulative exposure was divided by exposure time (in days). Maximum exposure (μg/m³) was estimated as the highest annual DWA on record for a worker prior to the study cutoff date of December 31, 2005 and accounting where appropriate for lag time. Exposure estimates were lagged by 5, 10, 15, and 20 years in order to account for exposures that may not have contributed to lung cancer because of the long latency required for manifestation of the disease. The authors also fit models with no lag time. As shown in Table VI-19 below, estimated exposure levels for workers from the Hazleton and Elmore plants were on average far lower than those for workers from the Reading plant. The median worker from Hazleton had a mean exposure across his tenure of less than 2 µg/m³, while the median worker from Elmore had a mean exposure of less than 1 µg/m³. The Elmore and Hazleton worker populations also had fewer short-term workers than the Reading population. This was particularly evident at Hazleton where the median value for cumulative exposure among cases was higher than at Reading despite the much lower mean and maximum exposure levels.

Schubauer-Berigan et al. analyzed the data set using a variety of exposure-response modeling approaches, including categorical analyses and continuous-variable piecewise log-linear and power models, described in Schubauer-Berigan et al. (2011). All models adjusted for birth cohort and plant. As exposure values were log-transformed for the power model analyses, the authors added small values to exposures of 0 in lagged analyses (0.05 µg/m³ for mean and maximum exposure, 0.05 µg/m³-days for cumulative exposure). The authors used restricted cubic spline models to assess the shape of the exposure-response curve and suggest appropriate parametric model forms. The Akaike Information Criterion (AIC) value was used to evaluate the fit of different model forms and lag times.

Because smoking information was available for only about 25 percent of the cohort, smoking could not be controlled for directly in the models. The authors reported that within the subset with smoking information, there was little difference in smoking by cumulative or maximum exposure category (p. 6), suggesting that smoking was unlikely to act as a confounder in the cohort. In addition to models based on the full cohort, Schubauer-Berigan et al. also prepared risk estimates based on models excluding professional workers and workers believed to have asbestos exposure. These models were intended to mitigate the potential impact of smoking and asbestos as confounders. If professional workers had both lower beryllium exposures and lower smoking rates than production workers, smoking could be a confounder in the cohort comprising both production and professional workers. However, the authors reasoned that smoking was unlikely to be correlated with beryllium exposure among production workers, and would therefore probably not act as a confounder in a cohort excluding professional workers.

The authors found that lung cancer risk was strongly and significantly related to mean, cumulative, and maximum measures of workers' exposure (all models reported in Schubauer-Berigan et al., 2011). They selected the best-fitting categorical, power, and monotonic piecewise log-linear (PWL) models with a 10-year lag to generate hazard ratios for male workers with a mean exposure of 0.5 µg/m (the current NIOSH Recommended Exposure Limit for beryllium). To estimate excess lifetime risk of cancer, they multiplied this hazard ratio by the 2004-2006 background lifetime lung cancer rate among U.S. males who had survived, cancer-free, to age 30. In addition, they estimated the mean exposure that would be associated with an excess lifetime risk of one in 1000, a value often used as a benchmark for significant risk in OSHA regulations. At OSHA's request, they also estimated excess lifetime risks for workers with mean exposures at the current PEL of 2 μg/m³ each of the other alternate PELs under consideration: 1 μg/m³, 0.2 μg/m³, and 0.1 μg/m³ (Schubauer-Berigan, 4/22/11). The resulting risk estimates are presented in Table VI-20 below.

Schubauer-Berigan et al. discuss several strengths, weaknesses, and uncertainties of their analysis. Strengths include long (> 30 years) follow-up time for members of the cohort and the extensive exposure and work history data available for the development of exposure estimates for workers in the cohort. Among the weaknesses and uncertainties of the study are the limited information available on workers' smoking habits: smoking information was available only for workers employed in 1968, about 25 percent of the cohort. In addition, the JEMs used did not account for possible respirator use among workers in the cohort. The authors note that workers' exposures may therefore have been overestimated, and that overestimation may have been especially severe for workers with high estimated exposures. They suggest that overestimation of exposures for workers in highly exposed positions may have caused attenuation of the exposure-response curve in some models at higher exposures.

The NIOSH publication did not discuss the reasons for basing risk estimates on mean exposure rather than cumulative exposure that is more commonly used for lung cancer risk analysis. OSHA believes the decision may involve the nonmonotonic relationship NIOSH observed between cancer risk and cumulative exposure level. As discussed previously, workers from the Reading plant frequently had very short tenures and high exposures yielding lower cumulative exposures compared to cohort workers from other plants with longer employment. Despite the low estimated cumulative exposures among the short-term Reading workers, they may be at high risk of lung cancer due to the tendency of beryllium to persist in the lung for long periods. This exposure misclassification could lead to the appearance of a nonmonotonic relationship between cumulative exposure and lung cancer risk. It is possible that a dose-rate effect may exist for beryllium, such that the risk from a cumulative exposure gained by long-term, low-level exposure is not equivalent to the risk from a cumulative exposure gained by very short-term, high-level exposure. In this case, mean exposure level may better correlate with the risk of lung cancer than cumulative exposure level. For these reasons OSHA considers the NIOSH choice of mean exposure metric to be appropriate and scientifically defensible for this particular dataset.

H. Preliminary Conclusions

As described above, OSHA's risk assessment for beryllium sensitization and CBD relied on two approaches: (1) review of the literature and (2) analysis of a dataset provided by NJRMC. First, the Agency reviewed the scientific literature to ascertain whether there is substantial risk to workers exposed at and below the current PEL and to characterize the expected impact of more stringent controls on workers' risk of sensitization and CBD. This review focused on facilities where exposures were primarily below the current PEL, and where several rounds of BeLPT and CBD screening had been conducted to evaluate the effectiveness of various exposure control measures. Second, OSHA investigated the exposure-response relationship for beryllium sensitization and CBD by analyzing a dataset that NJMRC provided on workers at a prominent, long-established beryllium machining facility. Although exposure-response studies have been published on sensitization and CBD, OSHA believes the nature and quality of their exposure data significantly limits their value for the Agency's risk assessment. Therefore, OSHA developed an independent exposure-response analysis using the NJMRC dataset, which was recently updated, includes workers exposed at low levels, and includes extensive exposure data collected in workers' breathing zones, as is preferred by OSHA.

OSHA's review of the scientific literature found substantial risk of both sensitization and CBD in workplaces in compliance with OSHA's current PEL (e.g., Kreiss et al., 1992; Schuler et al., 2000; Madl et al., 2007). At these plants, including a copper-beryllium processing facility, a beryllia ceramics facility, and a beryllium machining facility, exposure reduction programs that primarily used engineering controls to reduce airborne exposures to median levels at or around 0.2 μg/m³ had only limited impact on workers' risk. Cases of sensitization continued to occur frequently among newly hired workers, and some of these workers developed CBD within the short follow-up time.

In contrast, industrial hygiene programs that minimized both airborne and dermal exposure substantially lowered workers' risk of sensitization in the first years of employment. Programs that drastically reduced respiratory exposure via a combination of engineering controls and respiratory protection, minimized the potential for skin exposure via dermal PPE, and employed stringent housekeeping methods to keep work areas clean and prevent transfer of beryllium between areas sharply curtailed new cases of sensitization among newly-hired workers. For example, studies conducted at copper-beryllium processing, beryllium production, and beryllia ceramics facilities show that reduction of exposures to below 0.1 μg/m³ and protection from dermal exposure, in combination, achieved a substantial reduction in sensitization risk among newly-hired workers. However, even these stringent measures did not protect all workers from sensitization.

The most recent epidemiological literature on programs that have been successful in reducing workers' risk of sensitization have had very short follow-up time; therefore, they cannot address the question of how frequently workers sensitized in very low-exposure environments develop CBD. Clinical evaluation for CBD was not reported for workers at the copper-beryllium processing, beryllium production, and ceramics facilities. However, cases of CBD among workers exposed at low levels at a machining plant and cases of CA-CBD demonstrate that individuals exposed to low levels of airborne beryllium can develop CBD, and over time, can progress to severe disease. This conclusion is also supported by case reports within the literature of workers with CBD who may have been minimally exposed to beryllium, such as a worker employed only in administration at a beryllium ceramics facility (Kreiss et al., 1996).

The Agency's analysis of the Cullman dataset provided by NJMRC showed strong exposure-response trends using multiple analytical approaches, including examination of sensitization and disease prevalence by exposure categories and a proportional hazards modeling approach. In the prevalence analysis, cases of sensitization and disease were evident at all levels of exposure. The lowest prevalence of sensitization (2.0 percent) and CBD (1.0 percent) was observed among workers with LTW average exposure levels below 0.1 μg/m³, while those with LTW average exposure between 0.1-0.2 μg/m³ showed a marked increase in overall prevalence of sensitization (9.8 percent) and CBD (7.3 percent). Prevalence of sensitization and CBD also increased with cumulative exposure.

OSHA's proportional hazards analysis of the Cullman dataset found increasing risk of sensitization with both cumulative exposure and average exposure. OSHA also found a positive relationship between risk of CBD and cumulative exposure, but not between CBD and average exposure. The Agency used the cumulative exposure model results to estimate hazards ratios and risk of sensitization and CBD at the current PEL of 2 μg/m³ and each of the alternate PELs under consideration: 1 μg/m³, 0.5 μg/m³, 0.2 μg/m³, and 0.1 μg/m³. To estimate risk of CBD from a working lifetime of exposure, the Agency calculated the cumulative exposure associated with 45 years of exposure at each level, for total cumulative exposures of 90, 45, 22.5, 9, and 4.5 μg/m³-years. The risk estimates for sensitization and CBD ranged from 100-403 and 40-290 cases, respectively, per 1000 workers exposed at the current PEL of 2 μg/m³. The risks are projected to be substantially lower for both sensitization and CBD at 0.1 μg/m³ and range from 7.2-35 cases per 1000 and 3.1-26 cases per 1000, respectively. In these ways, the modeling results are similar to results observed from published studies of the Reading, Tucson, and Cullman plants and the OSHA analysis of sensitization and CBD prevalence within the Cullman plant.

OSHA has a high level of confidence in the finding of substantial risk of sensitization and CBD at the current PEL, and the Agency believes that a standard requiring a combination of more stringent controls on beryllium exposure will reduce workers' risk of both sensitization and CBD. Programs that have reduced median levels to below 0.1 μg/m³, tightly controlled both respiratory and dermal exposure, and incorporated stringent housekeeping measures have substantially reduced risk of sensitization within the first years of exposure. These conclusions are supported by the results of several studies conducted in state-of-the-art facilities dealing with a variety of production activities and physical forms of beryllium. In addition, these conclusions are supported by OSHA's statistical analysis of a dataset with highly detailed exposure and work history information on several hundred beryllium workers. While there is uncertainty regarding the precision of model-derived risk estimates, they provide further evidence that there is substantial risk of sensitization and CBD associated with exposure at the current PEL, and that this risk can be substantially lessened by stringent measures to reduce workers' beryllium exposure levels.

Furthermore, OSHA believes that beryllium-exposed workers' risk of lung cancer will be reduced by more stringent control of airborne beryllium exposures. The risk estimates from NIOSH's recent lung cancer study, described above, range from 33 to 140 excess lung cancers per 1000 workers exposed at the current PEL of 2 μg/m³. The NIOSH risk assessment's six best-fitting models each predict substantial reductions in risk with reduced exposure, ranging from 3 to 19 excess lung cancers per 1000 workers exposed at the proposed PEL of 0.1 μg/m³. The evidence of lung cancer risk from NIOSH's risk assessment provides additional support for OSHA's preliminary conclusions regarding the significance of risk to workers exposed to beryllium levels at and below the current PEL. However, the lung cancer risks require a sizable low dose extrapolation below beryllium exposure levels experienced by workers in the NIOSH study. As a result, there is a greater uncertainty in the lung cancer risk estimates and lesser confidence in their significance of risk below the current PEL than with beryllium sensitization and CBD. The preliminary conclusions with regard to significance of risk are presented and further discussed in section VIII of the preamble.

VII. Expert Peer Review of Health Effects and Preliminary Risk Assessment

In 2010, Eastern Research Group, Inc. (ERG), under contract to the Occupational Safety and Health Administration (OSHA) , conducted an independent, scientific peer review of (1) a draft Preliminary Beryllium Health Effects Evaluation (OSHA, 2010a), (2) a draft Preliminary Beryllium Risk Assessment (OSHA, 2010b), and (3) two NIOSH study manuscripts (Schubauer-Berigan et al., 2011 and 2011a). This section of the preamble describes the review process and summarizes peer reviewers' comments and OSHA's responses.

ERG conducted a search for nationally recognized experts in the areas of occupational epidemiology, occupational medicine, toxicology, immunology, industrial hygiene/exposure assessment, and risk assessment/biostatistics as requested by OSHA. ERG sought experts familiar with beryllium health effects research and who had no conflict of interest (COI) or apparent bias in performing the review. Interested candidates submitted evidence of their qualifications and responded to detailed COI questions. ERG also searched the Internet to determine whether qualified candidates had made public statements or declared a particular bias regarding beryllium regulation.

From the pool of qualified candidates, ERG selected five experts to conduct the review, based on:

○ Their qualifications, including their degrees, years of relevant experience, number of related peer-reviewed publications, experience serving as a peer reviewer for OSHA or other government organizations, and committee and association memberships related to the review topic;

○ Lack of any actual, potential, or perceived conflict of interest; and

○ The need to ensure that the panel collectively was sufficiently broad and diverse to fairly represent the relevant scientific and technical perspectives and fields of knowledge appropriate to the review.

OSHA reviewed the qualifications of the candidates proposed by ERG to verify that they collectively represented the technical areas of interest. ERG then contracted the following experts to perform the review.

(1) John Balmes, MD, Professor of Medicine, University of California-San Francisco

Expertise: pulmonary and occupational medicine, CBD, occupational lung disease, epidemiology, occupational exposures, medical surveillance.

(2) Patrick Breysse, Ph.D., Professor, Johns Hopkins University Bloomberg School of Public Health

Expertise: industrial hygiene, occupational/environmental health engineering, exposure monitoring/analysis, biomarkers, beryllium exposure assessment

(3) Terry Gordon, Ph.D., Professor, New York University School of Medicine.

Expertise: inhalation toxicology, pulmonary disease, beryllium toxicity and carcinogenicity, CBD genetic susceptibility, mode of action, animal models.

(4) Milton Rossman, MD, Professor of Medicine, Hospital of the University of Pennsylvania School of Medicine.

Expertise: pulmonary and clinical medicine, immunology, beryllium sensitization, BeLPT, clinical diagnosis for CBD.

(5) Kyle Steenland, Ph.D., Professor, Emory University, Rollins School of Public Health.

Expertise: occupational epidemiology, biostatistics, risk and exposure assessment, lung cancer, CBD, exposure-response models.

Reviewers were provided with the Technical Charge and Instructions (see ERG, 2010), a Request for Peer Review of NIOSH Manuscripts (see ERG, 2010), the draft Preliminary OSHA Health Effects Evaluation (OSHA, 2010a), the draft Preliminary Beryllium Risk Assessment (OSHA, 2010b), and access to relevant references. Each reviewer independently provided comments on the Health Effects, Risk Assessment, and NIOSH documents. A briefing call was held early in the review to ensure that reviewers understood the peer review process. ERG organized the call and OSHA representatives were available to respond to technical questions of clarification. Reviewers were invited to submit any subsequent questions of clarification.

The written comments from each reviewer were received and organized by ERG by charge questions. The unedited individual and reorganized comments were submitted to OSHA and the reviewers in preparation for a follow-up conference call. The conference call, organized and facilitated by ERG, provided an opportunity for OSHA to clarify individual reviewer's comments. After the call, reviewers were given the opportunity to revise their written comments to include the clarifications or additional information provided on the call. ERG submitted the revised comments to OSHA organized by both individual reviewer and by charge question. A final peer review report is available in the docket (ERG, 2010). Section VII.A of this preamble summarizes the comments received on the draft health effects document and OSHA's responses to those comments. Section VII.B summarizes comments received on the draft Preliminary Risk Assessment and the OSHA response.

A. Peer Review of Draft Health Effects Evaluation

The Technical Charge to peer reviewers posed general questions on the draft health effects document as well as specific questions pertaining to particle/chemical properties, kinetics and metabolism, acute beryllium disease, development of beryllium sensitization and CBD, genetic susceptibility, epidemiological studies of sensitization and CBD, animal models of chronic beryllium disease, genotoxicity, lung cancer epidemiological studies, animal cancer studies, other health effects, and preliminary conclusions drawn by OSHA.

OSHA asked the peer reviewers to generally comment on whether the draft health effects evaluation included the important studies, appropriately addressed their strengths and limitations, accurately described the results, and drew scientifically sound conclusions. Overall, the reviewers felt that the studies were described in sufficient detail, the interpretations accurate, and the conclusions reasonable. They agreed that the OSHA document covered the significant health endpoints related to occupational beryllium exposure. However, several reviewers requested that additional studies and other specific information be included in various sections of the document and these are discussed further below.

The reviewers had similar suggestions to improve the section V.A of this preamble on physical/chemical properties and section V.B on kinetics/metabolism. Dr. Balmes requested that physical and chemical characteristics of beryllium more clearly relate to development of sensitization and progression to CBD. Dr. Gordon requested greater consistency in the terminology used to describe particle characteristics, sampling methodologies, and the particle deposition in the respiratory tract. Dr. Breysse agreed and requested that the respiratory deposition discussion be better related to the onset of sensitization and CBD. Dr. Rossman suggested that the discussion of particle/chemical characteristics might be better placed after section V.D on the immunobiology of sensitization and CBD.

OSHA made a number of revisions to sections V.A and V.B to address the peer review comments above. Terminology used to describe particle characteristics in various studies was modified to be more consistent and better reflect the authors' intent in the published research articles. Section V.B.1 on respiratory kinetics of inhaled beryllium was modified to more clearly describe particle deposition in the different regions of the respiratory tract and their influence on CBD. At the recommendation of Dr. Gordon, a confusing figure was removed since it did not portray particle deposition in a clear manner. Rather than relocate the entire discussion of particle/chemical characteristics, a new section V.B.5 was added to specifically address the influence of beryllium particle characteristics and chemical form on the development of sensitization and CBD. Other section areas were shortened to remove information that was not necessarily relevant to the overall disease process. Statements were added on the effect of pre-existing diseases and smoking on beryllium clearance from the lung. It was made clear that the precise role of dermal exposure in beryllium sensitization is not completely understood. These smaller changes were made at the request of individual reviewers.

There were a couple of comments from reviewers pertaining to acute beryllium disease (ABD). Dr. Rossman commented that ABD did not make the development of CBD more likely. He requested that the document include a reference to the Van Ordstrand et al. (1943) article that first reported ABD in the U.S. Dr. Balmes pointed out that pathologists, rather than clinicians, interpret ABD pathology from lung tissue biopsy. Dr. Gordon commented that ABD is of lesser importance than CBD to the risk assessment and suggested that discussion of ABD be moved later in the document.

The Van Ordstrand reference was included in section V.C on acute beryllium diseases and statements were modified to address the peer review comments above. While OSHA agrees that ABD does not have a great impact on the Agency risk findings, the Agency believes the current organization does not create confusion on this point and decided not to move the ABD section later in the document. A statement that ABD is only relevant at exposures higher than the current PEL has been added to section V.C. Other reviewers did not feel the ABD discussion needed to be moved to a later section.

Most reviewers found the description of the development and pathogenesis of CBD in section V.D to be accurate and understandable. Dr. Breysse felt the section could better delineate the steps in disease development (e.g., development of beryllium sensitization, CBD progression) and recommended the 2008 National Academy of Sciences report as a model. He and Dr. Gordon felt the section overemphasized the role of apoptosis in CBD development. Dr. Breysse and Dr. Balmes recommended avoiding the phrase `subclinical' to describe sensitization and asymptomatic CBD, preferring the term `early stage' as a more appropriate description. Dr. Balmes requested clarification regarding accumulation of inflammatory cells in the bronchoalveolar lavage (BAL) fluid during CBD development. Dr. Rossman suggested some additional description of beryllium binding with the HLA-class II receptor and subsequent interaction with the naïve CD4⁺ T cells in the development of sensitization.

OSHA extensively reorganized section V.D to clearly delineate the disease process in a more linear fashion starting with the formation of beryllium antigen complex, its interaction with naïve T-cells to trigger CD4⁺ T-cell proliferation, and development of beryllium sensitization. This is presented in section V.D.1. A figure has been added that schematically presents this process in its entirety and the steps at which dermal exposure and genetic factors are believed to influence disease development (Figure 2 in section V.D). Section V.D.2 describes how subsequent inhalation and the persistent residual presence of beryllium in the lung leads to CD4⁺ T cell differentiation, cytokine production, accumulation of inflammatory cells in the alveolar region, granuloma formation, and progression of CBD. The section was modified to present apoptosis as only one of the plausible mechanisms for development/progression of CBD. The `early stage' terminology was adopted and the role of inflammatory cells in BAL was clarified.

While peer reviewers felt genetic susceptibility was adequately characterized, Dr. Rossman, Dr. Gordon, and Dr. Breysse suggested that additional study data be discussed to provide more depth on the subject, particularly the role genetic polymorphisms in providing a negatively charged HLA protein binding site for the positively charged beryllium ion. Section V.D.3 on genetic susceptibility now includes more information on the importance of gene-environment interaction in the development of CBD in low-exposed workers. The section expands on HLA-DPB1 alleles that influence beryllium-hapten binding and its impact on CBD risk.

All reviewers found the definition of CBD to be clear and understandable. However, several reviewers commented on the document discussion of the BeLPT which operationally defines beryllium sensitization. Drs. Balmes and Rossman requested a more clear statement that two abnormal blood BeLPT results were generally necessary to confirm sensitization. Dr. Balmes and Dr. Breysse requested more discussion of historical changes in the BeLPT method that have led to improvement in test performance and reductions in interlaboratory variability. These comments were addressed in an expanded document section V.D.5.b on criteria for sensitization and CBD case definition following development of the BeLPT.

Reviewers made suggestions to improve presentation of the many epidemiological studies of sensitization and CBD in the draft health effects document. Dr. Breysse and Dr. Gordon recommended that common weaknesses that apply to multiple studies be more rigorously discussed. Dr. Gordon requested that the discussion of the Beryllium Case Registry be modified to clarify the case inclusion criteria. Most reviewers called for the addition of tables to assist in summarizing the epidemiological study information.

A paragraph has been added near the beginning of section V.D.5 that identifies the common challenges to interpreting the epidemiological evidence that supports the occurrence of sensitization and CBD at occupational beryllium exposures below the current PEL. These include studies with small numbers of subjects and CBD cases, potential exposure misclassification resulting from lack of personal and short-term exposure data prior to the late 1990s, and uncertain dermal contribution among other issues. Table A.1 summarizing the key sensitization and CBD epidemiological studies was added to this preamble in appendix A of section V. Subsection V.D.5.a on studies conducted prior to the BeLPT has been reorganized to more clearly present the need for the Registry prior to listing the inclusion criteria.

Several reviewers requested that the draft health effects document discuss additional occupational studies on sensitization and CBD. Dr. Balmes suggested including Bailey et al. (2010) on reduction in sensitization at a beryllium production plant and Arjomandi et al. (2010) on CBD among workers in a nuclear weapons facility. Dr. Breysse recommended adding a brief discussion of Taiwo et al. (2008) on sensitization in aluminum smelter workers. Dr. Gordon and Dr. Rossman suggested mention of Curtis, (1951) on cutaneous hypersensitivity to beryllium as important for the role of dermal exposure. Dr. Rossman also provided a reference to a number of other sensitization and CBD articles of historical significance.

The above studies have been incorporated in several subsections of V.D.5 on human epidemiological evidence. The 1951 Curtis study is mentioned in the introduction to section V.D.5 as evidence of sensitization from dermal exposure. The Bailey et al. (2010) study is discussed in subsection V.D.5.d on beryllium metal processing and alloy production. The Arjomandi et al. (2010) study is discussed subsection V.D.5.h on nuclear weapons facilities and cleanup of former facilities. The Taiwo et al. (2008) study is discussed in subsection V.D.5.i on aluminum smelting. The other historical studies of historical significance are referenced in subsection V.D.5.a on studies conducted prior to the BeLPT.

Dr. Gordon suggested that the draft health effects document make clear that limitations in study design and lack of an appropriate model limited extrapolation of animal findings to the human immune-based respiratory disease. Dr. Rossman also remarked on the lack of a good animal model that consistently demonstrates a specific cell-mediated immune response to beryllium. Section V.D.6 was modified to include a statement that lack of a dependable animal model combined with studies that used single doses, few animals or abbreviated observation periods have limited the utility of the data. Table A.2 was added that summarizes important information on key animal studies of beryllium-induced immune response and lung inflammation.

In general, peer reviewers considered the preliminary conclusions with regard to sensitization and CBD to be reasonable and well presented in the draft health effects evaluation. All reviewers agreed that the scientific evidence supports sensitization as a necessary condition and an early endpoint in the development of CBD. The peer reviewers did not consider the presented evidence to convincingly show lung burden to be an important dose metric. Dr. Gordon explained that some animal studies in dogs have indicated that lung dose does influence granuloma formation but the importance of dose relative to genetic susceptibility, and physical/chemical form is unclear. He suggested the document indicate that many factors, including lung burden, affect the pulmonary tissue response to beryllium particles in the workplace.

There were other suggested improvements to the preliminary conclusion section of the draft document. Dr. Breysse felt that presenting the range of observed prevalence from occupational studies would help support the Agency findings. He also recommended that the preliminary conclusions make clear that CBD is a very complex disease and certain steps involved in the onset and progression are not yet clearly understood. Dr. Rossman pointed out that a report from Mroz et al. (2009) updated information on the rate at which beryllium sensitized individuals progress to CBD.

A statement has been added to section V.D.7 on the preliminary sensitization and CBD conclusions to indicate that all facets of development and progression of sensitization and CBD are not fully understood. Study references and prevalence ranges were provided to support the conclusion that epidemiological evidence demonstrates that sensitization and CBD occur from present-day exposures below OSHA's PEL. Statements were modified to indicate animal studies provide important insights into the roles of chemical form, genetic susceptibility, and residual lung burden in the development of beryllium lung disease. Updated information on rate of progression from sensitization to CBD was also included.

Reviewers made suggestions to improve presentation of the epidemiological studies of lung cancer that were similar to their comments on the CBD studies. Dr. Steenland requested that a table summarizing the lung cancer studies be added. He also recommended that more emphasis be placed on the SMR results from the Ward et al. (1992) study. Dr. Balmes felt that more detail was presented on the animal cancer studies than necessary to convey the relevant message. All reviewers thought that the Schubauer-Berigan et al. (2010) cohort mortality study that addressed some of the shortcomings of earlier lung cancer mortality studies should be discussed in the health effects document.

The recent Schubauer-Berigan et al. (2010) study conducted by the NIOSH Division of Surveillance, Hazard Evaluations, and Field Studies is now described and discussed in section V.E.2 on human epidemiology studies. Table A.3 summarizing the range of exposure measurements, study strengths and limitations, and other key lung cancer epidemiological study information was added to the health effects preamble. Section V.E.3 on the animal cancer studies already contained several tables that present study data so OSHA decided a summary table was not needed in this section.

Reviewers were asked two questions regarding the OSHA preliminary conclusions on beryllium-induced lung cancer: was the inflammation mechanism presented in the lung cancer section reasonable; and were there other mechanisms or modes of action to be considered? All reviewers agreed that inflammation was a reasonable mechanistic presentation as outlined in the document. Dr. Gordon requested OSHA clarify that inflammation may not be the sole mechanism for carcinogenicity. OSHA inserted statements in section V.E.5 on the preliminary lung cancer conclusions clarifying that tumorigenesis secondary to inflammation is a reasonable mechanism of action but other plausible mechanisms independent of inflammation may also contribute to the lung cancer associated with beryllium exposure.

There were a few comments from reviewers on health effects other than sensitization/CBD and lung cancer in the draft document. Dr. Balmes requested that the term “beryllium poisoning” not be used when referring to the hepatic effects of beryllium. He also offered language to clarify that the cardiovascular mortality among beryllium production workers in the Ward study cohort was probably due to ischemic heart disease and not the result of impaired lung function. Dr. Gordon requested removal of references to hepatic studies from in vitro and intravenous administration done at very high dose levels of little relevance to the occupational exposures of interest to OSHA. These changes were made to section V.F on other health effects.

B. Peer Review of the Draft Preliminary Risk Assessment

The Technical Charge to peer reviewers for review of the draft preliminary risk assessment was to ensure OSHA selected appropriate study data, assessed the data in a scientifically credible manner, and clearly explained its analysis. Specific charge questions were posed regarding choice of data sets, risk models, and exposure metrics; the role of dermal exposure and dermal protection; construction of the job exposure matrix; characterization of the risk estimates and their uncertainties; and whether a quantitative assessment of lung cancer risk, in addition to sensitization and CBD, was warranted.

Overall, the peer reviewers were highly supportive of the Agency's approach and major conclusions. They offered valuable suggestions for revisions and additional analysis to improve the clarity and certain technical aspects of the risk assessment. These suggestions and the steps taken by OSHA to address them are summarized here. A final peer review report (ERG, 2010c) and a risk assessment background document (OSHA, 2014a) are available in the docket.

OSHA asked peer reviewers a series of questions regarding its selection of surveys from a beryllium ceramics facility, a beryllium machining facility, and a beryllium alloy processing facility as the critical studies that form the basis of the preliminary risk assessment. Research showed that these workplaces had well characterized and relatively low beryllium exposures and underwent plant-wide screenings for sensitization and CBD before and after implementation of exposure controls. The reviewers were requested to comment on whether the study discussions were clearly presented, whether the role of dermal exposure and dermal protection were adequately addressed, and whether the preliminary conclusions regarding the observed exposure-related prevalence and reduction in risk were reasonable and scientifically credible. They were also asked to identify other studies that should be reviewed as part of the sensitization/CBD risk assessment.

Every peer reviewer felt the key studies were appropriate and their selection clearly explained in the document. Every peer reviewer regarded the preliminary conclusions from the OSHA review of these studies to be reasonable and scientifically sound. This conclusion stated that substantial risk of sensitization and CBD were observed in facilities where the highest exposed processes had median full-shift beryllium exposures around 0.2 μg/m³ or higher and that the greatest reduction in risk was achieved when exposures for all processes were lowered to 0.1 μg/m³ or below.

The reviewers suggested that three additional studies be added to the risk assessment review of the epidemiological literature. Dr. Balmes felt the document would be strengthened by including the Bailey et al. (2010) investigation of sensitization in a population of workers at the beryllium metal, alloy, and oxide production plant in Elmore, OH and the Arjomandi et al. (2010) publication on a group of 50 sensitized workers from a nuclear plant. Dr. Breysse suggested the study by Taiwo et al. (2008) on sensitization among workers in four aluminum smelters be considered.

A new subsection VI.A.3 was added to the preliminary risk assessment that describes the changes in beryllium exposure measurements, prevalence of sensitization and CBD, and implementation of exposure controls between 1992 and 2006 at the Elmore plant. This subsection includes a discussion of the Bailey et al. study. A summary of the Taiwo et al. (2008) study was added as subsection VI.A.5. A discussion of the Arjomandi et al. (2010) study was added in subsection VI.B as evidence that sensitized workers with primarily low beryllium exposure go on to develop CBD. However, the low rates of CBD among this group of sensitized workers also suggest that low beryllium exposure may reduce CBD risk when compared to worker populations with higher exposure levels.

While the majority of reviewers stated that OSHA adequately addressed the role of dermal exposure in sensitization and the importance of dermal protection for workers, a few had additional suggestions for OSHA's discussion. Dr. Breysse and Dr. Gordon pointed out that because the beryllium exposure control programs featured steps to reduce both skin contact and inhalation, it was difficult to distinguish between the effects of reducing airborne and dermal exposure. A statement was added to subsection VI.B that concurrent implementation of respirator use, dermal protection and engineering changes made it difficult to attribute reduced risk to any single control measure. Since the Cullman plant did not require glove use, OSHA believes it to be the best data set available for evaluating the effects of airborne exposure control on risk of sensitization.

Dr. Breysse requested additional discussion of the role of respiratory protection in achieving reduction in risk. Dr. Gordon suggested some additional clarification regarding mean and median exposure measures. Additional information on respiratory programs and exposure measures (e.g., median, arithmetic and geometric means), where available, were presented for each of the studies discussed in subsection VI.A.

The peer reviewers generally agreed that it was reasonable to conclude that community-acquired CBD (CA-CBD) resulted from low beryllium exposures. Drs. Breysse, Balmes and others noted that higher short-term excursions could not be ruled out. Dr. Gordon suggested that genetic susceptibility may have a role in cases of CA-CBD. Dr. Rossman raised the possibility that some CA-CBD cases could occur from contact with beryllium workers. All these points were added to subsection VI.C.

OSHA asked the peer reviewers to evaluate the choice of the National Jewish Medical and Research Center (NJMRC) data set on the Cullman, AL machinist population as a basis for exposure-response analysis and the reliance on cumulative exposure as the basis for the exposure-response analysis of sensitization and CBD. All peer reviewers indicated that the choice of the NJMRC data set for exposure-response analysis was clearly explained and reasonable and that they knew of no better data set for the analysis. Dr. Rossman commented that the NJMRC data set was an excellent source of exposures to different levels of beryllium and testing and evaluation of the workers. Dr. Steenland and Dr. Gordon suggested that the results from the OSHA analysis of the NJMRC data be compared with the available data from the studies of other beryllium facilities discussed in the epidemiological literature analysis. While a rigorous quantitative comparison (e.g., meta analysis) is difficult due to differences in the study designs and data types available for each study, subsection VI.E.4 compares the results of OSHA's prevalence analysis from the Cullman data with results from studies of the Tucson and Reading facilities.

OSHA asked the peer reviewers to evaluate methods used to construct the job exposure matrix (JEM) and to estimate beryllium exposure for each worker in the NJMRC data set. The JEM procedure was briefly summarized in the review document and described in detail as part of a risk assessment technical background document made available to the reviewers (OSHA, 2014a). Dr. Balmes felt that a more thorough discussion of the JEM would strengthen the preamble document. Dr. Gordon requested information about values assigned exposures below the limit of detection. Dr. Steenland requested that both the preamble and technical background document contain additional information on aspects of the JEM construction such as the job categories, job-specific exposure values, how jobs were grouped, and how non-machining jobs were handled in the JEM. He suggested the entire JEM be included in the technical background document. OSHA greatly expanded subsection VI.E.2 on air sampling and JEM to include more detailed discussion of the JEM construction. Exposure values for machining and non-machining job titles were provided in Tables VI-4 and VI-5. The procedures and rationale for grouping job-specific measurements into four time periods was explained. Jobs were not grouped in the JEM; rather, individual exposure estimates were created for each job in the work history data set. The technical background document further clarifies the JEM construction and the full JEM is included as an appendix to the revised background document (OSHA, 2014a). Subsection VI.E.3 on worker exposure reconstruction contains further detail about the work histories.

Peer reviewers fully supported OSHA's choice of the cumulative exposure metric to estimate risk of CBD from the NJMRC data set. As explained by Dr. Steenland, “cumulative exposure is often the choice for many chronic diseases as opposed to average or highest exposure.” He pointed out that the cumulative exposure metric also fit the CBD data better than other metrics. The reviewers generally felt that short-term peak exposure was probably the measure of airborne exposure most relevant to risk of beryllium sensitization. However, peer reviewers agreed that data required to capture workers' short-term peak exposures and to relate the peak exposure levels to sensitization were not available. Dr. Breysse explained that “short-term (hrs to minutes) peak exposures may be important to sensitization risk, while long term averages are more important for CBD risk. Unfortunately data for short-term peak exposures may not exist.” Dr. Steenland explained that of the available metrics “cumulative exposure fits the sensitization data better than the two alternatives, and hence is the best metric.” Statements were added to subsection VI.E.3 to indicate that while short-term exposures may be highly relevant to risk of sensitization, the individual peak exposures leading up to onset of sensitization was not able to be determined in the NJRMC Cullman study.

Peer reviewers found the methods used in the statistical exposure-response analysis to be clearly described. With the exception of Dr. Steenland, reviewers believed that a detailed critique of the statistical approach was beyond their level of expertise. Dr. Steenland supported OSHA's overall approach to the risk modeling and recommended additional analyses to explore the sensitivity of OSHA's results to alternate choices and to test the validity of aspects of the analysis. Dr. Steenland recommended that the logistic regression used by OSHA as a preliminary first analysis be dropped as an inappropriate model for a situation where it is important to account for changing exposures and case onset over time. Instead, he suggested a sensitivity analysis in which exposure-response coefficients generated using a traditional Cox proportionate hazards model be compared to the discrete time Cox model analog (i.e., complementary log-log Cox model) used by OSHA. The sensitivity analysis would facilitate examination of the proportional hazard assumption implied by the use of these models. Dr. Steenland advocated that OSHA include a table that displayed the mean number of BeLPT tests for the study population in order to address whether the number of sensitization tests introduced a potential bias. He inquired about the possibility of determining a sensitization incidence rate using cumulative or average exposure. Dr. Steenland suggested that the model control for additional potential confounders, such as age, smoking status, race and gender. He wanted a more complete explanation of the model constant for the year of diagnosis in Tables VI-9 through VI-12 to be included in the preamble as it was in the technical background document. Dr. Steenland recommended a sensitivity analysis that excludes the highest 5 to 10 percent of cumulative exposures which might address potential model uncertainty at the high end exposures. He requested that the results of statistical tests for non-linearity be included and confidence intervals for the risk estimates in Tables VI-17 and VI-18 be determined.

Many of Dr. Steenland's comments were addressed in subsection VI.F on the statistical modeling. The logistic regression analysis was removed from the section. A sensitivity analysis using the standard Cox model that treats survival time as continuous rather than discrete was added to the risk assessment background document and results were described in subsection VI.F. The interaction between exposure and follow-up time was not significant in the models suggesting that the proportional hazard assumption should not be rejected. The model coefficients using the standard Cox model were similar to model coefficients for the discrete model. Given this, OSHA did not feel it necessary to further estimate risks using the continuous Cox model at specific exposure levels.

A table of the mean number of BeLPT tests across the study population was added to the risk assessment background document. Subsection VI.F describes the table results and its impact on the statistical modeling. Smoking status and age were included in the discrete Cox proportional hazards model and not found to be significant predictors of beryllium sensitization. However, the available study population composition did not allow a confounder analysis of race and gender. OSHA chose not to include a detailed explanation of the model constant for the year of diagnosis in the preamble section. OSHA agrees with Dr. Steenland that the risk assessment background document adequately describes the model terms. For that reason, OSHA prefers that the risk assessment preamble focus on the results and major points of the analysis and refer the reader to the more technical background document for an explanation of model parameters. The linearity assumption was assessed using a fractional polynomial approach. The best fitting polynomials did not fit significantly better than the linear model. The details of the analysis were included in the risk assessment background document. Tables VI-17 and VI-18 now include the upper 95 percent confidence limits on the model-predicted cases of sensitization and CBD for the current and alternative PELs.

Most peer reviewers felt the major uncertainties of the risk assessment were clearly and adequately discussed in the documents they reviewed. Dr. Breysse requested that the risk assessment cover potential underestimation of risk from exposure misclassification bias. He requested further discussion of the degree to which the risk estimates from the Cullman machining plant could be extrapolated to workplaces that use other physical (e.g., particle size) and chemical forms of beryllium. He went on to question the strength of evidence that insoluble forms of beryllium cause CBD. Dr. Breysse also suggested that the assumptions used in the risk modeling be consolidated and more clearly presented. Dr. Steenland felt that there was potential underestimation of CBD risk resulting from exclusion of former workers and case status of current workers after employment.

Discussion of these uncertainties was added in the final paragraphs of section VI.F. The section was modified to more clearly identify assumptions with regard to the risk modeling such as an assumed linearity in exposure-response and cumulative dose equivalency when extrapolating risks over a 45-year working lifetime. Section VI.F recognizes the uncertainties in risk that can result from reconstructing individual exposures with very limited sampling data prior to 1994. The potential exposure misclassification can limit the strength of exposure-response relationships and result in the underestimation of risk. A more technical discussion of modeling assumptions and exposure measurement error are provided in the risk assessment background document. Section VI.F points out that the NJMRC data set does not capture CBD that occurred among workers who retired or left the Cullman plant. This and the short follow-up time is a source of uncertainty that likely leads to underestimation of risk. The section indicates that it is not unreasonable to expect the risk estimates to generally reflect onset of sensitization and CBD from exposure to beryllium forms that are relatively insoluble and enriched with respirable particles as encountered at the Cullman machining plant. Additional uncertainty is introduced when extrapolating the risk estimates to beryllium compounds of vastly different solubility and particle characteristics. OSHA does not agree with the comment suggesting that the association between CBD and insoluble forms of beryllium is weak. The principle sources of beryllium encountered at the Cullman machining plant, the Reading copper beryllium processing plant and the Tucson ceramics plant where excessive CBD was observed are insoluble forms of beryllium, such as beryllium metal, beryllium alloy, and beryllium oxide.

Finally, OSHA asked the peer reviewers to evaluate its treatment of lung cancer in the earlier draft preliminary risk assessment (OSHA, 2010b). When that document was prepared, OSHA had elected not to conduct a lung cancer risk assessment. The Agency believed that the exposure-response data available to conduct a lung cancer risk assessment from a Sanderson et al. study of a Reading, PA beryllium plant by was highly problematic. The Sanderson study primarily involved workers with extremely high and short-term exposures above airborne exposure levels of interest to OSHA (2 μg/m³ and below).

Just prior to arranging the peer review, a NIOSH study was published by Schubauer-Berigan et al. updating the Reading, PA cohort studied by Sanderson et al. and adding cohorts from two additional plants in Elmore, OH and Hazleton, PA (Schubauer-Berigan, 2011). At OSHA's request, the peer reviewers reviewed this study to determine whether it could provide a better basis for lung cancer risk analysis than the Sanderson et al. study. The reviewers found that the NIOSH update addressed the major concerns OSHA had expressed about the Sanderson study. In particular, they pointed out that workers in the Elmore and Hazleton cohorts had longer tenure at the plants and experienced lower exposures than those at the Reading, PA plant. Dr. Steenland recommended that “OSHA consider the new NIOSH data and develop risk estimates for lung cancer as well as sensitization and CBD.” Dr. Breysse believed that the NIOSH data “suggest that a risk assessment for lung cancer should be conducted by OSHA and the results be compared to the CBD/sensitization risk assessment before recommending an appropriate exposure concentration.” While acknowledging the improvements in the quality of the data, other reviewers were more restrained in their support for quantitative estimates of lung cancer risk. Dr. Gordon stated that despite improvements, there was “still uncertainty associated with the paucity of data below the current PEL of 2 μg/m³.” Dr. Rossman noted that the NIOSH study “did not address the problem of the uncertainty of the mechanism of beryllium carcinogenicity.” He felt that the updated NIOSH lung cancer mortality data “should not change the Agency's rationale for choosing to establish its risk findings for the proposed rule on its analysis for beryllium sensitization and CBD.” Dr. Balmes agreed that “the agency will be on firmer ground by focusing on sensitization and CBD.”

The preliminary risk assessment preamble subsection VI.G on lung cancer includes a discussion of the quantitative lung cancer risk assessment published by NIOSH researchers in 2010 (Schubauer-Berigan, 2011). The discussion describes the lower exposure levels, longer tenure, fewer short-term workers and additional years of observation that make the data more suitable for risk assessment. NIOSH relied on several modeling approaches to show that lung cancer risk was significantly related to both mean and cumulative beryllium exposure. Subsection VI.G provides the excess lifetime lung cancer risks predicted from several best-fitting NIOSH models at beryllium exposures of interest to OSHA (Table VI-20). Using the piecewise log-linear proportional hazards model favored by NIOSH, there is a projected drop in excess lifetime lung cancer risks from approximately 61 cases per 1000 exposed workers at the current PEL of 2.0 μg/m³ to approximately 6 cases per 1000 at the proposed PEL of 0.2 μg/m³. Subsection VI.H on preliminary conclusions indicates that these projections support a reduced risk of lung cancer from more stringent control of beryllium exposures but that the lung cancer risk estimates are more uncertain than those for sensitization and CBD.

VIII. Significance of Risk

To promulgate a standard that regulates workplace exposure to toxic materials or harmful physical agents, OSHA must first determine that the standard reduces a “significant risk” of “material impairment.” The first part of this requirement, “significant risk,” refers to the likelihood of harm, whereas the second part, “material impairment,” refers to the severity of the consequences of exposure.

The Agency's burden to establish significant risk is based on the requirements of the OSH Act (29 U.S.C. 651 et seq). Section 3(8) of the Act requires that workplace safety and health standards be “reasonably necessary or appropriate to provide safe or healthful employment” (29 U.S.C. 652(8)). The Supreme Court, in the Benzene decision, interpreted section 3(8) to mean that “before promulgating any standard, the Secretary must make a finding that the workplaces in question are not safe” (Industrial Union Department, AFL-CIO v. American Petroleum Institute, 448 U.S. 607, 642 (1980) (plurality opinion)). Examining section 3(8) more closely, the Court described OSHA's obligation to demonstrate significant risk:

“[S]afe” is not the equivalent of “risk-free.” A workplace can hardly be considered “unsafe” unless it threatens the workers with a significant risk of harm. Therefore, before the Secretary can promulgate any permanent health or safety standard, he must make a threshold finding that the place of employment is unsafe in the sense that significant risks are present and can be eliminated or lessened by a change in practices (Id).

As the Court made clear, the Agency has considerable latitude in defining significant risk and in determining the significance of any particular risk. The Court did not specify a means to distinguish significant from insignificant risks, but rather instructed OSHA to develop a reasonable approach to making a significant risk determination. The Court stated that “it is the Agency's responsibility to determine in the first instance what it considers to be a 'significant' risk,” (448 U.S. at 655) and it did not express “any opinion on the . . . difficult question of what factual determinations would warrant a conclusion that significant risks are present which make promulgation of a new standard reasonably necessary or appropriate” (448 U.S. at 659). The Court also stated that, while OSHA's significant risk determination must be supported by substantial evidence, the Agency “is not required to support the finding that a significant risk exists with anything approaching scientific certainty” (448 U.S. at 656). Furthermore:

A reviewing court [is] to give OSHA some leeway where its findings must be made on the frontiers of scientific knowledge . . . . [T]he Agency is free to use conservative assumptions in interpreting the data with respect to carcinogens, risking error on the side of overprotection rather than underprotection [so long as such assumptions are based on] a body of reputable scientific thought (448 U.S. at 656).

Thus, to make the significance of risk determination for a new or proposed standard, OSHA uses the best available scientific evidence to identify material health impairments associated with potentially hazardous occupational exposures and to evaluate exposed workers' risk of these impairments.

The OSH Act also requires that the Agency make a finding that the toxic material or harmful physical agent at issue causes material impairment to worker health. In that regard, the Act directs the Secretary of Labor to set standards based on the available evidence where no employee, over his/her working life time, will suffer from material impairment of health or functional capacity, even if such employee has regular exposure to the hazard, to the exent feasible (29 U.S.C. 655(b)(5)).

As with significant risk, what constitutes material impairment in any given case is a policy determination for which OSHA is given substantial leeway. “OSHA is not required to state with scientific certainty or precision the exact point at which each type of [harm] becomes a material impairment” (AFL-CIO v. OSHA, 965 F.2d 962, 975 (11th Cir. 1992)). Courts have also noted that OSHA should consider all forms and degrees of material impairment—not just death or serious physical harm—and that OSHA may act with a “pronounced bias towards worker safety” (Id; Bldg & Constr. Trades Dep't v. Brock, 838 F.2d 1258, 1266 (D.C. Cir. 1988)). OSHA's long-standing policy is to consider 45 years as a “working life,” over which it must evaluate material impairment and risk.

In formulating this proposed beryllium standard, OSHA has reviewed the best available evidence pertaining to the adverse health effects of occupational beryllium exposure, including lung cancer and chronic beryllium disease (CBD), and has evaluated the risk of these effects from exposures allowed under the current standard as well as the expected impact of the proposed standard on risk. Based on its review of extensive epidemiological and experimental research, OSHA has preliminarily determined that long-term exposure at the current Permissible Exposure Limit (PEL) would pose a significant risk of material impairment to workers' health, and that adoption of the new PEL and other provisions of the proposed rule will substantially reduce this risk.

A. Material Impairment of Health

In this preamble at section V, Health Effects, OSHA reviewed the scientific evidence linking occupational beryllium exposure to a variety of adverse health effects, including CBD and lung cancer. Based on this review, OSHA preliminarily concludes that beryllium exposure causes these effects. The Agency's preliminary conclusion was strongly supported by a panel of independent peer reviewers, as discussed in section VII.

Here, OSHA discusses its preliminary conclusion that CBD and lung cancer constitute material impairments of health, and briefly reviews other adverse health effects that can result from beryllium exposure. Based on this preliminary conclusion and on the scientific evidence linking beryllium exposure to both CBD and lung cancer, OSHA concludes that occupational exposure to beryllium causes “material impairment of health or functional capacity” within the meaning of the OSH Act.

1. Chronic Beryllium Disease

CBD is a respiratory disease in which the body's immune system reacts to the presence of beryllium in the lung, causing a progression of pathological changes including chronic inflammation and tissue scarring. CBD can also impair other organs such as the liver, skin, spleen, and kidneys and cause adverse health effects such as granulomas of the skin and lymph nodes and cor pulmonale (i.e., enlargement of the heart) (Conradi et al., 1971; ACCP, 1965; Kriebel et al., 1988a and b). In early, asymptomatic stages of CBD, small granulomatous lesions and mild inflammation occur in the lungs. Early stage CBD among some workers has been observed to progress to more serious disease even after the worker is removed from exposure (Mroz, 2009), probably because common forms of beryllium have slow clearance rates and can remain in the lung for years after exposure. Sood et al. has reported that cessation of exposure can sometimes have beneficial effects on lung function (Sood et al., 2004). However, this was based on a small study of six patients with CBD, and more research is needed to better determine the relationship between exposure duration and disease progression. In general, progression of CBD from early to late stages is understood to vary widely, responding differently to exposure cessation and treatment for different individuals (Sood, 2009; Mroz, 2009).

Over time, the granulomas can spread and lead to lung fibrosis (scarring) and moderate to severe loss of pulmonary function, with symptoms including a persistent dry cough and shortness of breath (Saber and Dweik, 2000). Fatigue, night sweats, chest and joint pain, clubbing of fingers (due to impaired oxygen exchange), loss of appetite, and unexplained weight loss may occur as the disease progresses. Corticosteroid therapy, in workers whose beryllium exposure has ceased, has been shown to control inflammation, ease symptoms (e.g., difficulty breathing, fever, cough, and weight loss) and in some cases prevent the development of fibrosis (Marchand-Adam et al., 2008). Thus early treatment can lead to CBD regression in some patients, although there is no cure (Sood, 2004). Other patients have shown short-term improvements from corticosteroid treatment, but then developed serious fibrotic lesions (Marchand-Adam et al., 2008). Once fibrosis has developed in the lungs, corticosteroid treatment cannot reverse the damage (Sood, 2009). Persons with late-stage CBD experience severe respiratory insufficiency and may require supplemental oxygen (Rossman, 1991). Historically, late-stage CBD often ended in death (NAS, 2008).

While the use of steroid therapy has mitigated CBD mortality, treatment with corticosteroids has side effects that need to be measured against the possibility of progression of disease (Trikudanathan and McMahon, 2008; Lipworth, 1999; Gibson et al., 1996; Zaki et al., 1987). Adverse effects associated with long-term corticosteroid use include, but are not limited to, increased risk of opportunistic infections (Lionakis and Kontoyiannis, 2003; Trikudanathan and McMahon, 2008); accelerated bone loss or osteoporosis leading to increased risk of fractures or breaks (Hamida et al., 2011; Lehouck et al., 2011; Silva et al., 2011; Sweiss et al., 2011; Langhammer et al., 2009); psychiatric effects including depression, sleep disturbances, and psychosis (Warrington and Bostwick, 2006; Brown, 2009); adrenal suppression (Lipworth, 1999; Frauman, 1996); ocular effects including cataracts, ocular hypertension, and glaucoma (Ballonzolli and Bourchier, 2010; Trikudanathan and McMahon, 2008; Lipworth, 1999); an increase in glucose intolerance (Trikudanathan and McMahon, 2008); excessive weight gain (McDonough et al., 2008; Torres and Nowson, 2007; Dallman et al., 2007; Wolf, 2002; Cheskin et al., 1999); increased risk of atherosclerosis and other cardiovascular syndromes (Franchimont et al., 2002); skin fragility (Lipworth, 1999); and poor wound healing (de Silva and Fellows, 2010). Studies relating the long-term effect of corticosteroid use for the treatment of CBD need to be undertaken to evaluate the treatment's overall effectiveness against the risk of adverse side effects from continued usage.

OSHA considers late-stage CBD to be a material impairment of health, as it involves permanent damage to the pulmonary system, causes additional serious adverse health effects, can have adverse occupational and social consequences, requires treatment associated with severe and lasting side effects, and may in some cases be life-threatening. Furthermore, OSHA believes that material impairment begins prior to the development of symptoms of the disease.

Although there are no symptoms associated with early-stage CBD, during which small lesions and inflammation appear in the lungs, the Agency has preliminarily concluded that the earliest stage of CBD is material impairment of health. OSHA bases this conclusion on evidence showing that early-stage CBD is a measurable change in the state of health which, with and sometimes without continued exposure, can progress to symptomatic disease. Thus, prevention of the earliest stages of CBD will prevent development of more serious disease. The OSHA Lead Standard established the Agency's position that a `subclinical' health effect may be regarded as a material impairment of health. In the preamble to that standard, the Agency said:

OSHA believes that while incapacitating illness and death represent one extreme of a spectrum of responses, other biological effects such as metabolic or physiological changes are precursors or sentinels of disease which should be prevented . . . Rather than revealing beginnings of illness the standard must be selected to prevent an earlier point of measurable change in the state of health which is the first significant indicator of possibly more severe ill health in the future. The basis for this decision is twofold—first, pathophysiologic changes are early stages in the disease process which would grow worse with continued exposure and which may include early effects which even at early stages are irreversible, and therefore represent material impairment themselves. Secondly, prevention of pathophysiologic changes will prevent the onset of the more serious, irreversible and debilitating manifestations of disease. (43 FR 52952, 52954, November 14, 1978)

Since the Lead rulemaking, OSHA has also found other non-symptomatic health conditions to be material impairments of health. In the Bloodborne Pathogens (BP) rulemaking, OSHA maintained that material impairment includes not only workers with clinically “active” hepatitis from the hepatitis B virus (HBV) but also includes asymptomatic HBV “carriers” who remain infectious and are able to put others at risk of serious disease through contact with body fluids (e.g., blood, sexual contact) (56 FR 64004, December 6, 1991). OSHA stated: “Becoming a carrier [of Hepatitis B] is a material impairment of health even though the carrier may have no symptoms. This is because the carrier will remain infectious, probably for the rest of his or her life, and any person who is not immune to HBV who comes in contact with the carrier's blood or certain other body fluids will be at risk of becoming infected” (56 FR 64004, 64036).

OSHA preliminarily finds that early-stage CBD is the type of asymptomatic health effect the Agency determined to be a material impairment of health in the lead standard. Early stage CBD involves lung tissue inflammation without symptomatology that can worsen with—or without—continued exposure. The lung pathology progresses over time from a chronic inflammatory response to tissue scarring and fibrosis accompanied by moderate to severe loss in pulmonary function. Early stage CBD is clearly a precursor of advanced clinical disease, prevention of which will prevent symptomatic disease. OSHA argued in the Lead standard that such precursor effects should be considered material health impairments in their own right, and that the Agency should act to prevent them when it is feasible to do so. Therefore, OSHA preliminarily finds all stages of CBD to be material impairments of health.

2. Lung Cancer

OSHA considers lung cancer, a frequently fatal disease, to be a material impairment of health. OSHA's finding that inhaled beryllium causes lung cancer is based on the best available epidemiological data, reflects evidence from animal and mechanistic research, and is consistent with the conclusions of other government and public health organizations (see this preamble at section V, Health Effects). For example, the International Agency for Research on Cancer (IARC), National Toxicology Program (NTP), and American Conference of Governmental Industrial Hygienists (ACGIH) have all classified beryllium as a known human carcinogen (IARC, 2009).

The Agency's epidemiological evidence comes from multiple studies of U.S. beryllium workers (Sanderson et al., 2001a; Ward et al., 1992; Wagoner et al., 1980; Mancuso et al., 1979). Most recently, a NIOSH cohort study found significantly increased lung cancer mortality among workers at seven beryllium processing facilities (Schubauer-Berigan et al., 2011). The cohort was exposed, on average, to lower levels of beryllium than those in most previous studies, had fewer short-term workers, and had sufficient follow-up time to observe lung cancer in the population. OSHA considers the Schubauer-Berigan study to be the best available epidemiological evidence regarding the risk of lung cancer from beryllium at exposure levels near the PEL.

Supporting evidence of beryllium carcinogenicity comes from various animal studies as well as in vitro genotoxicity and other studies (EPA, 1998; ATSDR, 2002; Gordon and Bowser, 2003; NAS, 2008; Nickell-Brady et al., 1994; NTP, 1999 and 2005; IARC, 1993 and 2009). Multiple mechanisms may be involved in the carcinogenicity of beryllium, and factors such as epigenetics, mitogenicity, reactive oxygen-mediated indirect genotoxicity, and chronic inflammation may contribute to the lung cancer associated with beryllium exposure, although the results of studies testing the direct genotoxicity of beryllium are mixed (EPA summary, 1998). While there is uncertainty regarding the exact mechanism of carcinogenesis for beryllium, the overall weight of evidence for the carcinogenicity of beryllium is strong. Therefore, the Agency has preliminarily determined beryllium to be an occupational carcinogen.

3. Other Impairments

While OSHA has relied primarily on the relationship between occupational beryllium exposure and CBD and lung cancer to demonstrate the necessity of the standard, the Agency has also determined that several other adverse health effects can result from exposure to beryllium. Inhalation of high airborne concentrations of beryllium (well above the 2 μg/m³ OSHA PEL) can cause acute beryllium disease, a severe (sometimes fatal), rapid-onset inflammation of the lungs. Hepatic necrosis, damage to the heart and circulatory system, chronic renal disease, mucosal irritation and ulceration, and urinary tract cancer have also reportedly been associated with occupational exposures well above the current PEL (see this preamble at section V, Health Effects, subsection E, Epidemiological Studies, and subsection F, Other Health Effects). These adverse systemic effects and acute beryllium disease mostly occurred prior to the introduction of occupational and environmental standards set in 1970-1972 (OSHA, 1971; ACGIH, 1971; ANSI, 1970) and 1974 (EPA, 1974) and therefore are less relevant today than in the past. Because they occur only rarely in current-day occupational environments, they are not addressed in OSHA's risk analysis or significance of risk determination.

The Agency has also determined that beryllium sensitization, a precursor which occurs before early stage CBD and is an essential step for worker development of the disease, can result from exposure to beryllium. The Agency takes no position at this time on whether sensitization constitutes a material impairment of health, because it was unnecessary to do so as part of this rulemaking. As discussed in Section V, Health Effects, only sensitized individuals can develop CBD (NAS, 2008). OSHA's risk assessment for sensitization informs the Agency's understanding of what exposure control measures have been successful in preventing sensitization, which in turn prevents development of CBD. Therefore sensitization is considered in the next section on significance of risk. In AFL-CIO v. Marshall, 617 F.2d 636, 654 n.83 (D.C. Cir. 1979) (Cotton Dust), the D.C. Circuit upheld OSHA's authority to regulate to prevent precursors to a material impairment of health without deciding whether the precursors themselves constituted material impairment of health.

B. Significance of Risk and Risk Reduction

To evaluate the significance of the health risks that result from exposure to hazardous chemical agents, OSHA relies on the best available epidemiological, toxicological, and experimental evidence. The Agency uses both qualitative and quantitative methods to characterize the risk of disease resulting from workers' exposure to a given hazard over a working lifetime at levels of exposure reflecting compliance with current standards and compliance with the new standards being proposed.

As discussed above, the Agency's characterization of risk is guided in part by the Benzene decision. In Benzene, the Court broadly describes the range of risks OSHA might determine to be significant:

It is the Agency's responsibility to determine in the first instance what it considers to be a “significant” risk. Some risks are plainly acceptable and others are plainly unacceptable. If, for example, the odds are one in a billion that a person will die from cancer by taking a drink of chlorinated water, the risk clearly could not be considered significant. On the other hand, if the odds are one in a thousand that regular inhalation of gasoline vapors that are 2 percent benzene will be fatal, a reasonable person might well consider the risk significant and take the appropriate steps to decrease or eliminate it (Benzene, 448 U.S. at 655).

The Court further stated, “The requirement that a 'significant' risk be identified is not a mathematical straitjacket. . . . Although the Agency has no duty to calculate the exact probability of harm, it does have an obligation to find that a significant risk is present before it can characterize a place of employment as 'unsafe', “and proceed to promulgate a regulation (Id.).

In this preamble at section VI, Preliminary Risk Assessment, OSHA finds that the available epidemiological data are sufficient to evaluate risk for beryllium sensitization, CBD, and lung cancer among beryllium-exposed workers. The preliminary findings from this assessment are summarized below.

1. Risk of Beryllium Sensitization and CBD

OSHA's preliminary risk assessment for CBD and beryllium sensitization relies on studies conducted at a Tucson, AZ beryllium ceramics plant (Kreiss et al., 1996; Henneberger et al., 2001; Cummings et al., 2006); a Reading, PA alloy processing plant (Schuler et al., 2005; Thomas et al., 2009); a Cullman, AL beryllium machining plant (Kelleher et al., 2001; Madl et al., 2007); and an Elmore, OH metal, alloy, and oxide production plant (Kreiss et al., 1997; Bailey et al., 2010; Schuler et al., 2012). The Agency uses these studies to demonstrate the significance of risk at the current PEL and the significant reduction in risk expected with reduction of the PEL. In addition to the effects OSHA anticipates from reduction of airborne beryllium exposure, the Agency expects that dermal protection provisions in the proposed rule will further reduce risk. Studies conducted in the 1950s by Curtis et al. showed that soluble beryllium particles could penetrate the skin and cause beryllium sensitization (Curtis 1951, NAS 2008). Tinkle et al. established that 0.5- and 1.0-μm particles can penetrate intact human skin surface and reach the epidermis, where beryllium particles would encounter antigen-presenting cells and initiate sensitization (Tinkle et al., 2003). Tinkle et al. further demonstrated that beryllium oxide and beryllium sulfate, applied to the skin of mice, generate a beryllium-specific, cell-mediated immune response similar to human beryllium sensitization (Tinkle et al., 2003). In the epidemiological studies discussed below, the exposure control programs that most effectively reduced the risk of beryllium sensitization and CBD incorporated both respiratory and dermal protection. OSHA has preliminarily determined that an effective exposure control program should incorporate both airborne exposure reduction and dermal protection provisions.

In the Tucson ceramics plant, 4,133 short-term breathing zone measurements collected between 1981 and 1992 had a median of 0.3 μg/m³. Kreiss et al. reported that eight (5.9 percent) of 136 workers tested for beryllium sensitization in 1992 were sensitized, six (4.4 percent) of whom were diagnosed with CBD. Exposure control programs were initiated in 1992 to reduce workers' airborne beryllium exposure, but the programs did not address dermal exposure. Full-shift personal samples collected between 1994 and 1999 showed a median beryllium exposure of 0.2 μg/m³ in production jobs and 0.1 μg/m³ in production support (Cummings et al., 2007). In 1998, a second screening found that 6, (9 percent) of 69 tested workers hired after the 1992 screening, were sensitized, of whom 1 was diagnosed with CBD. All of the sensitized workers had been employed at the plant for less than 2 years (Henneberger et al., 2001), too short a time period for most people to develop CBD following sensitization. Of the 77 Tucson workers hired prior to 1992 who were tested in 1998, 8 (10.4 percent) were sensitized and all but 1 of these (9.7 percent) were diagnosed with CBD (Henneberger et al., 2001).

Kreiss et al., studied workers at a beryllium metal, alloy, and oxide production plant in Elmore, OH. Workers participated in a BeLPT survey in 1992 (Kreiss et al., 1997). Personal lapel samples collected during 1990-1992 had a median value of 1.0 μg/m³. Kreiss et al. reported that 43 (6.9 percent) of 627 workers tested in 1992 were sensitized, 6 of whom were diagnosed with CBD (4.4 percent).

Newman et al. conducted a series of BeLPT screenings of workers at a Cullman, AL precision machining facility between 1995 and 1999 (Newman et al., 2001). Personal lapel samples collected at this plant in the early 1980s and in 1995 from all machining processes combined had a median of 0.33 μg/m³ (Madl et al., 2007). After a sentinel case of CBD was diagnosed at the plant in 1995, the company implemented engineering and administrative controls and PPE designed to reduce workers' beryllium exposures in machining operations. Personal lapel samples collected extensively between 1996 and 1999 in machining jobs have an overall median of 0.16 μg/m³, showing that the new controls reduced machinists' exposures during this period. However, the results of BeLPT screenings conducted in 1995-1999 showed that the exposure control program initiated in 1995 did not sufficiently protect workers from beryllium sensitization and CBD. In a group of 60 workers who had been employed at the plant for less than a year, and thus would not have been working there prior to 1995, 4 (6.7 percent) were found to be sensitized. Two of these workers (3.35 percent) were diagnosed with CBD. (Newman et al., 2001).

Sensitization and CBD were studied in a population of workers at a Reading, PA copper beryllium plant, where alloys containing a low level of beryllium were processed (Schuler et al., 2005). Personal lapel samples were collected in production and production support jobs between 1995 and May 2000. These samples showed primarily very low airborne beryllium levels, with a median of 0.073 μg/m³. The wire annealing and pickling process had the highest personal lapel sample values, with a median of 0.149 μg/m³. Despite these low exposure levels, a BeLPT screening conducted in 2000 showed that 5, (11.5 percent) workers of 43 hired after 1992 were sensitized (evaluation for CBD not reported). Two of the sensitized workers had been hired less than a year before the screening (Thomas et al., 2009).

In summary, the epidemiological literature on beryllium sensitization and CBD that OSHA's risk assessment relied on show sensitization prevalences ranging from 6.5 percent to 11.5 percent and CBD prevalences ranging from 1.3 percent to 9.7 percent among workers who had full-shift exposures well below the current PEL and median full-shift exposures at or below the proposed PEL, and whose follow-up time was less than 45 years. As referenced earlier, OSHA is interested in the risk associated with a 45-year (i.e., working lifetime) exposure. Because CBD often develops over the course of years following sensitization, the risk of CBD that would result from 45 years' occupational exposure to airborne beryllium is likely to be higher than the prevalence of CBD observed among these workers. In either case, based on these studies, the risks to workers appear to be significant.

The available epidemiological evidence shows that reducing workers' levels of airborne beryllium exposure can substantially reduce risk of beryllium sensitization and CBD. The best available evidence on effective exposure control programs comes partly from studies of programs introduced around 2000 at Reading, Tucson, and Elmore that used a combination of engineering controls, dermal and respiratory PPE, and stringent housekeeping measures to reduce workers' dermal exposures and airborne exposures to levels well below the proposed PEL of 0.2 μg/m³. These programs have substantially lowered the risk of sensitization among new workers. As discussed earlier, prevention of beryllium sensitization prevents subsequent development of CBD.

In the Reading, PA copper beryllium plant, full-shift airborne exposures in all jobs were reduced to a median of 0.1 μg/m³ or below and dermal protection was required for production-area workers beginning in 2000-2001 (Thomas et al., 2009). After these adjustments were made, 2 (5.4 percent) of 37 newly hired workers became sensitized. Thereafter, in 2002, the process with the highest exposures (median 0.1 μg/m³) was enclosed and workers involved in that process were required to use respiratory protection. As a result, the remaining jobs had very low exposures (medians ~ 0.03 μg/m³). Among 45 workers hired after the enclosure was built and respiratory protection instituted, 1 was found to be sensitized (2.2 percent). This is a sharp reduction in sensitization from the 11.5 percent of 43 workers, discussed above, who were hired after 1992 and had been sensitized by the time of testing in 2000.

In the Tucson beryllium ceramics plant, respiratory and skin protection was instituted for all workers in production areas in 2000. BeLPT testing done in 2000-2004 showed that only 1 (1 percent) worker had been sensitized out of 97 workers hired during that time period (Cummings et al., 2007; testing for CBD not reported). This contrasts with the prevalence of sensitization in the 1998 Tucson BeLPT screening, which found that 6 (9 percent) of 69 workers hired after 1992 were sensitized (Cummings et al., 2007).

The modern Elmore facility provides further evidence that combined reductions in respiratory exposure (via respirator use) and dermal exposure are effective in reducing risk of beryllium sensitization. In Elmore, historical beryllium exposures were higher than in Tucson, Reading, and Cullman. Personal lapel samples collected at Elmore in 1990-1992 had a median of 1.0 µg/m³. In 1996-1999, the company took steps to reduce workers' beryllium exposures, including engineering and process controls (Bailey et al., 2010; exposure levels not reported). Skin protection was not included in the program until after 1999. Beginning in 1999 all new employees were required to wear loose-fitting powered air-purifying respirators (PAPR) in manufacturing buildings (Bailey et al., 2010). Skin protection became part of the protection program for new employees in 2000, and glove use was required in production areas and for handling work boots beginning in 2001. Bailey et al., (2010) compared the occurrence of beryllium sensitization and CBD in 2 groups of workers: 1) 258 employees who began work at the Elmore plant between January 15, 1993 and August 9, 1999 (the “pre-program group”) and were tested in 1997 and 1999, and 2) 290 employees who were hired between February 21, 2000 and December 18, 2006 and underwent BeLPT testing in at least one of frequent rounds of testing conducted after 2000 (the “program group”). They found that, as of 1999, 23 (8.9 percent) of the pre-program group were sensitized to beryllium. The prevalence of sensitization among the “program group” workers, who were hired after the respiratory protection and PPE measures were put in place, was around 2-3 percent. Respiratory protection and skin protection substantially reduced, but did not eliminate, risk of sensitization. Evaluation of sensitized workers for CBD was not reported.

OSHA's preliminary risk assessment also includes analysis of a data set provided to OSHA by the National Jewish Research and Medical Center (NJMRC). The data set describes a population of 319 beryllium-exposed workers at a Cullman, AL machining facility. It includes exposure samples collected between 1980 and 2005, and has updated work history and screening information for over three hundred workers through 2003. Seven (2.2 percent) workers in the data set were reported as sensitized only. Sixteen (5.0 percent) workers were listed as sensitized and diagnosed with CBD upon initial clinical evaluation. Three (1.0 percent) workers, first shown to be sensitized only, were later diagnosed with CBD. The data set includes workers exposed at airborne beryllium levels near the proposed PEL, and extensive exposure data collected in workers' breathing zones, as is preferred by OSHA. Unlike the Tucson, Reading, and Elmore facilities, respirator use was not generally required for workers at the Cullman facility. Thus, analysis of this data set shows the risk associated with varying levels of airborne exposure, rather than the virtual elimination of airborne exposure via respiratory PPE. Also unlike the Tucson, Elmore, and Reading facilities, glove use was not reported to be mandatory in the Cullman facility. Thus, OSHA believes reductions in risk at the Cullman facility to be the result of airborne exposure control, rather than the combination of airborne and dermal exposure controls at the Tucson, Elmore, and Reading facilities.

OSHA analyzed the prevalence of beryllium sensitization and CBD among workers at the Cullman facility who were exposed to airborne beryllium levels at and below the current PEL of 2 µg/m³. In addition, a statistical modeling analysis of the NJMRC Cullman data set was conducted under contract with Dr. Roslyn Stone of the University of Pittsburgh Graduate School of Public Heath, Department of Biostatistics. OSHA summarizes these analyses briefly below, and in more detail in this preamble at section VI, Preliminary Risk Assessment.

Tables 1 and 2 below present the prevalence of sensitization and CBD cases across several categories of lifetime-weighted (LTW) average and highest-exposed job (HEJ) exposure at the Cullman facility. The HEJ exposure is the exposure level associated with the highest-exposure job and time period experienced by each worker. The columns “Total” and “Total percent” refer to all sensitized workers in the dataset, including workers with and without a diagnosis of CBD.

The current PEL of 2 μg/m is close to the upper bound of the highest quartile of LTW average (0.51-2.15 μg/m) and HEJ (0.954-2.213) exposure levels. In the highest quartile of LTW average exposure, there were 12 cases of sensitization (15.4 percent), including 8 (10.3 percent) diagnosed with CBD. Notably, the Cullman workers had been exposed to beryllium dust for considerably less than 45 years at the time of testing. A high prevalence of sensitization (9.2 percent) and CBD (5.3 percent) is seen in the top quartile of HEJ exposure as well, with even higher prevalences in the third quartile (0.387-0.691 μg/m).

The proposed PEL of 0.2 μg/m³ is close to the upper bound of the second quartile of LTW average (0.81-0.18 μg/m³) and HEJ (0.091-0.214 μg/m³) exposure levels and to the lower bound of the third quartile of LTW average (0.19-0.50 μg/m³) exposures. The second quartile of LTW average exposure shows a high prevalence of beryllium-related health effects, with six workers sensitized (8.2 percent), of whom four (5.5 percent) were diagnosed with CBD. The second quartile of HEJ exposure also shows a high prevalence of beryllium-related health effects, with seven workers sensitized (8.6 percent), of whom 6 (7.4 percent) were diagnosed with CBD. Among six sensitized workers in the third quartile of LTW average exposures, all were diagnosed with CBD (7.8 percent). The prevalence of CBD among workers in these quartiles was approximately 5-8 percent, and overall sensitization (including workers with and without CBD) was about 8 percent. OSHA considers these rates as evidence that the risk of developing CBD is significant among workers exposed at and below the current PEL, even down to the proposed PEL. Much lower prevalences of sensitization and CBD were found among workers with exposure levels less than or equal to about 0.08 μg/m³. Two sensitized workers (2.2 percent), including 1 case of CBD (1.0 percent), were found among workers with LTW average exposure levels and HEJ exposure levels less than or equal to 0.08 μg/m³ and 0.086 μg/m³, respectively. Strict control of airborne exposure to levels below 0.1 μg/m³ can, therefore, significantly reduce risk of sensitization and CBD. Although OSHA recognizes that maintaining exposure levels below 0.1 μg/m³ may not be feasible in some operations (see this preamble at section IX, Summary of the Preliminary Economic Analysis and Initial Regulatory Flexibility Analysis), the Agency believes that workers in facilities that meet the proposed action level of 0.1 μg/m³ will be at less risk of sensitization and CBD than workers in facilities that cannot meet the action level.

Table 3 below presents the prevalence of sensitization and CBD cases across cumulative exposure quartiles, based on the same Cullman data used to derive Tables 1 and 2. Cumulative exposure is the sum of a worker's exposure across the duration of his employment.

A 45-year working lifetime of occupational exposure at the current PEL would result in 90 μg/m ³-years, a value far higher than the cumulative exposures of workers in this data set, who worked for periods of time less than 45 years and whose exposure levels were mostly well below the PEL. Workers with 45 years of exposure to the proposed PEL would have a cumulative exposure (9 μg/m ³-years) in the highest quartile for this worker population. As with the average and HEJ exposures, the greatest risk of sensitization and CBD appears at high exposure levels (> 1.468 μg/m ³-years). The third cumulative quartile, at which a sharp increase in sensitization and CBD appears, is bounded by 1.468 and 7.008 μg/m ³-years. This is equivalent to 0.73-3.50 years of exposure at the current PEL of 2 μg/m ³, or 7.34-35.04 years of exposure at the proposed PEL of 0.2 μg/m ³. Prevalence of both sensitization and CBD is substantially lower in the second cumulative quartile (0.148-1.467 μg/m ³-years). This is equivalent to approximately 0.7 to 7 years at the proposed PEL of 0.2 μg/m ³, or 1.5 to 15 years at the proposed action level of 0.1 μg/m ³. This supports that maintaining exposure levels below the proposed PEL, where feasible, will help to protect long-term workers against risk of beryllium sensitization and early stage CBD.

As discussed in the Health Effects section (V.D), CBD often worsens with increased time and level of exposure. In a longitudinal study, workers initially identified as beryllium sensitized through workplace surveillance developed early stage CBD defined by granulomatous inflammation but no apparent physiological abnormalities (Newman et al., 2005). A study of workers with this early stage CBD showed significant declines in breathing capacity and gas exchange over the 30 years from first exposure (Mroz et al., 2009). Many of the workers went on to develop more severe disease that required immunosuppressive therapy despite being removed from exposure. While precise beryllium exposure levels were not available on the individuals in these studies, most started work in the 1980s and 1990s and were likely exposed to average levels below the current 2 μg/m ³ PEL. The evidence for time-dependent disease progression indicates that the CBD risk estimates for a 45-year lifetime exposure at the current PEL will include a higher proportion of individuals with advanced clinical CBD than found among the workers in the NJMRC data set.

Studies of community-acquired (CA) CBD support the occurrence of advanced clinical CBD from long-term exposure to airborne beryllium (Eisenbud, 1998; Maier et al., 2008). A discussion of the study findings can be found in this preamble at section VI.C, Preliminary Risk Assessment. For example, one study evaluated 16 potential cases of CA-CBD in individuals that resided near a beryllium production facility in the years between 1943 and 2001 (Maier et al., 2008). Five cases of definite CBD and three cases of probable CBD were found. Two of the subjects with probable cases died before they could be confirmed with the BeLPT; the third had an abnormal BeLPT and radiography consistent with CBD, but granulomatous disease was not pathologically proven. The individuals with CA-CBD identified in this study suffered significant health impacts from the disease, including obstructive, restrictive, and gas exchange pulmonary defects. Six of the eight cases required treatment with prednisone, a step typically reserved for severe cases due to the adverse side effects of steroid treatment. Despite treatment, three had died of respiratory impairment as of 2002. There was insufficient information to estimate exposure to the individuals, but the limited amount of ambient air sampling in the 1950s suggested that average beryllium levels in the area where the cases resided were below 2 μg/m ³. The authors concluded that “low levels of exposures with significant disease latency can result in significant morbidity and mortality” (Maier et al., 2008, p. 1017).

OSHA believes that the literature review, prevalence analysis, and the evidence for time-dependent progression of CBD described above provide sufficient information to draw preliminary conclusions about significance of risk, and that further quantitative analysis of the NJMRC data set is not necessary to support the proposed rule. The studies OSHA used to support its preliminary conclusions regarding risk of beryllium sensitization and CBD were conducted at modern industrial facilities with exposure levels in the range of interest for this rulemaking, so a model is not needed to extrapolate risk estimates from high to low exposures, as has often been the case in previous rules. Nevertheless, the Agency felt further quantitative analysis might provide additional insight into the exposure-response relationship for sensitization and CBD.

Using the NJMRC data set, Dr. Stone ran a complementary log-log proportional hazards model, an extension of logistic regression that allows for time-dependent exposures and differential time at risk. Relative risk of sensitization increased with cumulative exposure (p = 0.05). A positive, but not statistically significant association was observed with LTW average exposure (p = 0.09). There was little association with highest-exposed job (HEJ) exposure (p = 0.3). Similarly, the proportional hazards models for the CBD endpoint showed positive relationships with cumulative exposure (p = 0.09), but LTW average exposure and HEJ exposure were not closely related to relative risk of CBD (p-values > 0.5). Dr. Stone used the cumulative exposure models to generate risk estimates for sensitization and CBD.

Tables 4 and 5 below present risk estimates from these models, assuming 5, 10, 20, and 45 years of beryllium exposure. The tables present sensitization and CBD risk estimates based on year-specific intercepts, as explained in the section on Risk Assessment and the accompanying background document. Each estimate represents the number of sensitized workers the model predicts in a group of 1000 workers at risk during the given year with an exposure history at the specified level and duration. For example, in the exposure scenario for 1995, if 1000 workers were occupationally exposed to 2 μg/m ³ for 10 years, the model predicts that about 56 (55.7) workers would be identified as sensitized. The model for CBD predicts that about 42 (41.9) workers would be diagnosed with CBD that year. The year 1995 shows the highest risk estimates generated by the model for both sensitization and CBD, while 1999 and 2002 show the lowest risk estimates generated by the model for sensitization and CBD, respectively. The corresponding 95 percent confidence intervals are based on the uncertainty in the exposure coefficient.

As shown in Tables 4 and 5, the exposure-response models Dr. Stone developed based on the Cullman data set predict a high risk of both sensitization (about 96-394 cases per 1000 exposed workers) and CBD (about 44-313 cases per 1000) at the current PEL of 2 μg/m³ for an exposure duration of 45 years (90 μg/m³-yr). For a 45-year exposure at the proposed PEL of 0.2 μg/m³, risk estimates for sensitization (about 8-40 cases per 1000 exposed workers) and CBD (about 4-30 per 1000 exposed workers) are substantially reduced. Thus, the model predicts that the risk of sensitization and CBD at a PEL of 0.2 μg/m³ will be about 10 percent of the risk at the current PEL of 2 μg/m³.

OSHA does not believe the risk estimates generated by these exposure-response models to be highly accurate. Limitations of the analysis include the size of the dataset, relatively sparse exposure data from the plant's early years, study size-related constraints on the statistical analysis of the dataset, and limited follow-up time on many workers. The Cullman study population is a relatively small group and can support only limited statistical analysis. For example, its size precludes inclusion of multiple covariates in the exposure-response models or a two-stage exposure-response analysis to model both sensitization and the subsequent development of CBD within the subpopulation of sensitized workers. The limited size of the Cullman dataset is characteristic of studies on beryllium-exposed workers in modern, low-exposure environments, which are typically small-scale processing plants (up to several hundred workers, up to 20-30 cases).

Despite these issues with the statistical analysis, OSHA believes its main policy determinations are well supported by the best available evidence, including the literature review and careful examination of the prevalence of sensitization and CBD among workers with exposure levels comparable to the current and proposed PELs in the NJMRC data set. The previously described literature analysis and prevalence analysis demonstrate that workers with occupational exposure to airborne beryllium at the current PEL face a risk of becoming sensitized to beryllium and progressing to both early and advanced stages of CBD that far exceeds the value of 1 in 1000 used by OSHA as a benchmark of clearly significant risk. Furthermore, OSHA's preliminary risk assessment indicates that risk of beryllium sensitization and CBD can be significantly reduced by reduction of airborne exposure levels, along with respiratory and dermal protection measures, as demonstrated in facilities such as the Tucson ceramics plant, the Elmore beryllium production facility, and the Reading copper beryllium facility described in the literature review.

OSHA's preliminary risk assessment also indicates that despite the reduction in risk expected with the proposed PEL, the risk to workers with average exposure levels of 0.2 μg/m³ is still clearly significant (see this preamble at section VI). In the prevalence analysis, workers with LTW average or HEJ exposures close to 0.2 μg/m³ experienced high levels of sensitization and CBD. This finding is corroborated by the literature analysis, which showed that workers exposed to mean plant-wide airborne exposures between 0.1 and 0.5 μg/m³ had a similarly high prevalence of sensitization and CBD. Given the significant risk at these levels of exposure, the Agency believes that the proposed action level of 0.1 μg/m³, dermal protection requirements, and other ancillary provisions of the proposed rule are key to reducing the risk of beryllium sensitization and CBD among exposed workers. OSHA preliminarily concludes that the proposed standard, including the PEL of 0.2 μg/m³, the action level of 0.1 μg/m³, and provisions to limit dermal exposure to beryllium, together will significantly reduce workers' risk of beryllium sensitization and CBD from occupational beryllium exposure.

2. Risk of Lung Cancer

OSHA's review of epidemiological studies of lung cancer mortality among beryllium workers found that most did not characterize exposure levels sufficiently to characterize risk of lung cancer at the current and proposed PELs. However, as discussed in this preamble at section V, Health Effects and section VI, Preliminary Risk Assessment, NIOSH recently published a quantitative risk assessment based on beryllium exposure and lung cancer mortality among 5436 male workers employed at beryllium processing plants in Reading, PA; Elmore, OH; and Hazleton, PA, prior to 1970 (Schubauer-Berigan et al., 2010b). This new risk assessment addresses important sources of uncertainty for previous lung cancer analyses, including the sole prior exposure-response analysis for beryllium and lung cancer, conducted by Sanderson et al. (2001) on workers from the Reading plant alone. Workers from the Elmore and Hazleton plants who were added to the analysis by Schubauer-Berigan et al. were, in general, exposed to lower levels of beryllium than those at the Reading plant. The median worker from Hazleton had a mean exposure across his tenure of less than 2 μg/m³, while the median worker from Elmore had a mean exposure of less than 1 μg/m³. The Elmore and Hazleton worker populations also had fewer short-term workers than the Reading population. Finally, the updated cohorts followed the worker populations through 2005, increasing the length of follow-up time compared to the previous exposure-response analysis. For these reasons, OSHA based its preliminary risk assessment for lung cancer on the Schubauer-Berigan risk analysis.

Schubauer-Berigan et al. (2011) analyzed the data set using a variety of exposure-response modeling approaches, described in this preamble at section VI, Preliminary Risk Assessment. The authors found that lung cancer mortality risk was strongly and significantly related to mean, cumulative, and maximum measures of workers' exposure to beryllium (all models reported in Schubauer-Berigan et al., 2011). They selected the best-fitting models to generate risk estimates for male workers with a mean exposure of 0.5 μg/m³ (the current NIOSH Recommended Exposure Limit for beryllium). In addition, they estimated the mean exposure that would be associated with an excess lung cancer mortality risk of one in one thousand. At OSHA's request, the authors also estimated excess risks for workers with mean exposures at each of the other alternate PELs under consideration: 1 μg/m³, 0.2 μg/m³, and 0.1 μg/m³. Table 6 presents the estimated excess risk of lung cancer mortality associated with various levels of beryllium exposure allowed under the current rule, based on the final models presented in Schubauer-Berigan et al' s risk assessment.

The lowest estimate of excess lung cancer deaths from the six final models presented by Schubauer-Berigan et al. is 33 per 1000 workers exposed at a mean level of 2 μg/m³, the current PEL. Risk estimates as high as 200 lung cancer deaths per 1000 result from the other five models presented. Regardless of the model chosen, the excess risk of about 33 to 200 per 1000 workers is clearly significant, falling well above the level of risk the Supreme Court indicated a reasonable person might consider acceptable (See Benzene, 448 U.S. at 655). The proposed PEL of 0.2 μg/m³ is expected to reduce these risks significantly, to somewhere between 2.7-30 excess lung cancer deaths per 1000 workers. These risk estimates still fall above the threshold of 1 in 1000 that OSHA considers clearly significant. However, the Agency believes the lung cancer risks should be regarded with a greater degree of uncertainty than the risk estimates for CBD discussed previously. While the risk estimates for CBD at the proposed PEL were determined from exposure levels observed in occupational studies, the lung cancer risks are extrapolated from much higher exposure levels.

C. Conclusions

As discussed above, OSHA used the best available scientific evidence to identify adverse health effects of occupational beryllium exposure, and to evaluate exposed workers' risk of these impairments. The Agency reviewed extensive epidemiological and experimental research pertaining to adverse health effects of occupational beryllium exposure, including lung cancer, immunological sensitization to beryllium, and CBD, and has evaluated the risk of these effects from exposures allowed under the current and proposed standards. The Agency has, additionally, reviewed previous policy determinations and case law regarding material impairment of health, and has preliminarily determined that CBD, in all stages, and lung cancer constitute material health impairments. Furthermore, OSHA has preliminarily determined that long-term exposure to beryllium at the current PEL would pose a risk of CBD and lung cancer greater than the risk of 1 per 1000 exposed workers the Agency considers clearly significant. OSHA's risk assessment for beryllium indicates that adoption of the new PEL, action level, and dermal protection provisions of the proposed rule will significantly reduce this risk. OSHA therefore believes it has met the statutory requirements pertaining to significance of risk, consistent with the OSH Act, Supreme Court precedent, and the Agency's previous policy decisions.

IX. Summary of the Preliminary Economic Analysis and Initial Regulatory Flexibility Analysis

A. Introduction and Summary

OSHA's Preliminary Economic Analysis and Initial Regulatory Flexibility Analysis (PEA) addresses issues related to the costs, benefits, technological and economic feasibility, and the economic impacts (including impacts on small entities) of this proposed respirable beryllium rule and evaluates regulatory alternatives to the proposed rule. Executive Orders 13563 and 12866 direct agencies to assess all costs and benefits of available regulatory alternatives and, if regulation is necessary, to select regulatory approaches that maximize net benefits (including potential economic, environmental, and public health and safety effects; distributive impacts; and equity), unless a statute requires another regulatory approach. Executive Order 13563 emphasized the importance of quantifying both costs and benefits, of reducing costs, of harmonizing rules, and of promoting flexibility. The full PEA has been placed in OSHA rulemaking docket OSHA-H005C-2006-0870. This rule is an economically significant regulatory action under Sec. 3(f)(1) of Executive Order 12866 and has been reviewed by the Office of Information and Regulatory Affairs in the Office of Management and Budget, as required by executive order.

The purpose of the PEA is to:

• Identify the establishments and industries potentially affected by the proposed rule;
• Estimate current exposures and the technologically feasible methods of controlling these exposures;
• Estimate the benefits resulting from employers coming into compliance with the proposed rule in terms of reductions in cases of lung cancer and chronic beryllium disease;
• Evaluate the costs and economic impacts that establishments in the regulated community will incur to achieve compliance with the proposed rule;
• Assess the economic feasibility of the proposed rule for affected industries; and
• Assess the impact of the proposed rule on small entities through an Initial Regulatory Flexibility Analysis (IRFA), to include an evaluation of significant regulatory alternatives to the proposed rule that OSHA has considered.

The PEA contains the following chapters:

Chapter I. Introduction

Chapter II. Assessing the Need for Regulation

Chapter III. Profile of Affected Industries

Chapter IV. Technological Feasibility

Chapter V. Costs of Compliance

Chapter VI. Economic Feasibility Analysis and Regulatory Flexibility Determination

Chapter VII. Benefits and Net Benefits

Chapter VIII. Regulatory Alternatives

Chapter IX. Initial Regulatory Flexibility Analysis

The PEA includes all of the economic analyses OSHA is required to perform, including the findings of technological and economic feasibility and their supporting materials required by the OSH Act as interpreted by the courts (in Chapters III, IV, V, and VI); those required by EO 12866 and EO 13563 (primarily in Chapters III, V, and VII, though these depend on material in other chapters); and those required by the Regulatory Flexibility Act (in Chapters VI, VIII, and IX, though these depend, in part, on materials presented in other chapters).

Key findings of these chapters are summarized below and in sections IX.B through IX.I of this PEA summary.

Profile of Affected Industries

This proposed rule would affect employers and employees in many different industries across the economy. As described in Section IX.C and reported in Table IX-2 of this preamble, OSHA estimates that a total of 35,051 employees in 4,088 establishments are potentially at risk from exposure to beryllium.

Technological Feasibility

As described in more detail in Section IX.D of this preamble and in Chapter IV of the PEA, OSHA assessed, for all affected sectors, the current exposures and the technological feasibility of the proposed PEL of 0.2 μg/m³.

Tables IX-5 in section IX.D of this preamble summarizes all nine application groups (industry sectors and production processes) studied in the technological feasibility analysis. The technological feasibility analysis includes information on current exposures, descriptions of engineering controls and other measures to reduce exposures, and a preliminary assessment of the technological feasibility of compliance with the proposed PELs.

The preliminary technological feasibility analysis shows that for the majority of the job groups evaluated, exposures are either already at or below the proposed PEL, or can be adequately controlled with additional engineering and work practice controls. Therefore, OSHA preliminarily concludes that the proposed PEL of 0.2 μg/m³ is technologically feasible for most operations most of the time.

Based on the currently available evidence, it is more difficult to determine whether an alternative PEL of 0.1 μg/m³ would also be feasible in most operations. For some application groups, a PEL of 0.1 μg/m³ would almost certainly be feasible. In other application groups, a PEL of 0.1 μg/m³ appears feasible, except for establishments working with high beryllium content alloys. For application groups with the highest exposure, the exposure monitoring data necessary to more fully evaluate the effectiveness of exposure controls adopted after 2000 are not currently available to OSHA, which makes it difficult to determine the feasibility of achieving exposure levels at or below 0.1 μg/m³.

OSHA also evaluated the feasibility of a STEL of 2.0 μg/m³. The majority of the available short-term measurements are below 2.0 μg/m³; therefore OSHA preliminarily concludes that the proposed STEL of 2.0 μg/m³ can be achieved for most operations most of the time. OSHA recognizes that for a small number of tasks, short-term exposures may exceed the proposed STEL, even after feasible control measures to reduce TWA exposure to below the proposed PEL have been implemented, and therefore assumes that the use of respiratory protection will continue to be required for some short-term tasks. It is more difficult based on the currently available evidence to determine whether the alternative STEL of 1.0 μg/m³ would also be feasible in most operations based on lack of detail in the activities of the workers presented in the data. OSHA expects additional use of respiratory protection would be required for tasks in which peak exposures can be reduced to less than 2.0 μg/m³ but not less than 1.0 μg/m³. Due to limitations in the available sampling data and the higher detection limits for short term measurements, OSHA could not determine the percentage of the STEL measurements that are less than or equal to 0.5 μg/m³.

Costs of Compliance

As described in more detail in Section IX.E and reported, by application group and NAICS code, in Table IX-7 of this preamble, the total annualized cost of compliance with the proposed standard is estimated to be about $37.6 million. The major cost elements associated with the revisions to the standard are housekeeping ($12.6 million), engineering controls ($9.5 million), training ($5.8 million), and medical surveillance ($2.9 million).

The compliance costs are expressed as annualized costs in order to evaluate economic impacts against annual revenue and annual profits, to be able to compare the economic impact of the rulemaking with other OSHA regulatory actions, and to be able to add and track Federal regulatory compliance costs and economic impacts in a consistent manner. Annualized costs also represent a better measure for assessing the longer-term potential impacts of the rulemaking. The annualized costs were calculated by annualizing the one-time costs over a period of 10 years and applying a discount rate of 3 percent (and an alternative discount rate of 7 percent).

The estimated costs for the proposed beryllium standard represent the additional costs necessary for employers to achieve full compliance. They do not include costs associated with current compliance that has already been achieved with regard to the new requirements or costs necessary to achieve compliance with existing beryllium requirements, to the extent that some employers may currently not be fully complying with applicable regulatory requirements.

Economic Impacts

To assess the nature and magnitude of the economic impacts associated with compliance with the proposed rule, OSHA developed quantitative estimates of the potential economic impact of the new requirements on entities in each of the affected industry sectors. The estimated compliance costs were compared with industry revenues and profits to provide an assessment of the economic feasibility of complying with the revised standard and an evaluation of the potential economic impacts.

As described in greater detail in Section IX.F of this preamble and in Chapter VI of the PEA, the costs of compliance with the proposed rulemaking are not large in relation to the corresponding annual financial flows associated with each of the affected industry sectors. The estimated annualized costs of compliance represent about 0.11 percent of annual revenues and about 1.52 percent of annual profits, on average, across all affected firms. Compliance costs do not represent more than 1 percent of revenues or more than 16.25 percent of profits in any affected industry.

Based on its analysis of the relative inelasticity of demand for beryllium-containing inputs and products and of possible international trade effects, OSHA concluded that most or all costs arising from this proposed beryllium rule would be passed on in higher prices rather than absorbed in lost profits and that any price increases would result in minimal loss of business to foreign competition.

Given the minimal potential impact on prices or profits in the affected industries, OSHA has preliminarily concluded that compliance with the requirements of the proposed rulemaking would be economically feasible in every affected industry sector.

Benefits, Net Benefits, and Cost-Effectiveness

As described in more detail in Section VIII.G of this preamble, OSHA estimated the benefits, net benefits, and incremental benefits of the proposed beryllium rule. That section also contains a sensitivity analysis to show how robust the estimates of net benefits are to changes in various cost and benefit parameters. A full explanation of the derivation of the estimates presented there is provided in Chapter VII of the PEA for the proposed rule.

OSHA estimated the benefits associated with the proposed beryllium PEL of 0.2 μg/m³ and, for analytical purposes to comply with OMB Circular A-4, with alternative beryllium PELs of .1 μg/m³ and .5 μg/m³ by applying the dose-response relationship developed in the Agency's preliminary risk assessment—summarized in Section VI of this preamble—to current exposure levels. OSHA determined current exposure levels by first developing an exposure profile for industries with workers exposed to beryllium, using OSHA inspection and site-visit data, and then applying this exposure profile to the total current worker population. The industry-by-industry exposure profile is summarized in Table IX-3 in Section IX.C of this preamble.

By applying the dose-response relationship to estimates of current exposure levels across industries, it is possible to project the number of cases of the following diseases expected to occur in the worker population given current exposure levels (the “baseline”):

• fatal cases of lung cancer,
• fatal cases of chronic beryllium disease (CBD), and
• morbidity related to chronic beryllium disease.

Table IX-1 provides a summary of OSHA's best estimate of the costs and benefits of the proposed rule. As shown, the proposed rule, once it is fully effective, is estimated to prevent 96 fatalities and 50 non-fatal beryllium-related illnesses annually, and the monetized annualized benefits of the proposed rule are estimated to be $575.8 million using a 3-percent discount rate and $255.3 million using a 7-percent discount rate. Also as shown in Table IX-1, the estimated annualized cost of the rule is $37.6 million using a 3-percent discount rate and $39.1 million using a 7-percent discount rate. The proposed rule is estimated to generate net benefits of $538.2 million annually using a 3-percent discount rate and $216.2 million annually using a 7-percent discount rate. The estimated costs and benefits of the proposed rule, disaggregated by industry sector, were previously presented in Table I-1 in this preamble.

Initial Regulatory Flexibility Analysis

OSHA has prepared an Initial Regulatory Flexibility Analysis (IRFA) in accordance with the requirements of the Regulatory Flexibility Act, as amended in 1996. Among the contents of the IRFA are an analysis of the potential impact of the proposed rule on small entities and a description and discussion of significant alternatives to the proposed rule that OSHA has considered. The IRFA is presented in its entirety both in Chapter IX of the PEA and in Section IX.I of this preamble.

The remainder of this section (Section IX) of the preamble is organized as follows:

B. The Need for Regulation

C. Profile of Affected Industry

D. Technological Feasibility Analysis

E. Costs of Compliance

F. Economic Feasibility Analysis and Regulatory Flexibility Determination

G. Benefits and Net Benefits

H. Regulatory Alternatives

I. Initial Regulatory Flexibility Analysis.

B. Need for Regulation

Employees in work environments addressed by the proposed beryllium rule are exposed to a variety of significant hazards that can and do cause serious injury and death. As described in Chapter II of the PEA in support of the proposed rule, the risks to employees are excessively large due to the existence of various types of market failure, and existing and alternative methods of overcoming these negative consequences—such as workers' compensation systems, tort liability options, and information dissemination programs—have been shown to provide insufficient worker protection.

After carefully weighing the various potential advantages and disadvantages of using a regulatory approach to improve upon the current situation, OSHA preliminarily concludes that, in the case of beryllium exposure, the proposed mandatory standards represent the best choice for reducing the risks to employees. In addition, rulemaking is necessary in this case in order to replace older existing standards with updated, clear, and consistent health standards.

C. Profile of Affected Industries

1. Introduction

Chapter III of the PEA presents a profile of industries that use beryllium, beryllium oxide, and/or beryllium alloys. The discussion below summarizes the findings in that chapter. For each industry sector identified, the Agency describes the uses of beryllium and estimates the number of establishments and employees that may be affected by this proposed rulemaking. Employee exposure to beryllium can also occur as a result of certain processes such as welding that are found in many industries. OSHA uses the umbrella term “application group” to refer either to an industrial sector or a cross-industry group with a common process. These groups are all mutually exclusive and are analyzed in separate sections in Chapter III of the PEA. These sections briefly describe each application group and then explain how OSHA estimated the number of establishments working with beryllium and the number of employees exposed to beryllium. Beryllium is rarely used by all establishments in any particular application group because its unique properties and relatively high cost typically result in only very specific and limited usage within a portion of a group.

The information in Chapter III of the PEA is based on reports prepared under task order by Eastern Research Group (ERG), an OSHA contractor; information collected during OSHA's Small Business Advocacy Review Panel (OSHA 2008b); and Agency research and analysis. Technological feasibility reports (summarized in Chapter IV of the PEA) for each beryllium-using application group provide a detailed presentation of processes and occupations with beryllium exposure, including available sampling exposure measurements and estimates of how many employees are affected in each specific occupation.

OSHA has identified nine application groups that would be potentially affected by the proposed beryllium standard:

1. Beryllium Production

2. Beryllium Oxide Ceramics and Composites

3. Nonferrous Foundries

4. Secondary Smelting, Refining, and Alloying

5. Precision Turned Products

6. Copper Rolling, Drawing, and Extruding

7. Fabrication of Beryllium Alloy Products

8. Welding

9. Dental Laboratories

These application groups are broadly defined, and some include establishments in several North American Industrial Classification System (NAICS) codes. For example, the Copper Rolling and Drawing, and Extruding application group is made up both of NAICS 331421 Copper Rolling, Drawing, and Extruding and NAICS 331422 Copper Wire Drawing. While an application group may contain numerous NAICS six-digit industry codes, in most cases only a fraction of the establishments in any individual six-digit NAICS industry use beryllium and would be affected by the proposed rule. For example, not all companies in the above application group work with copper that contains beryllium.

One application group, welding, reflects industrial activities or processes that take place in various industry sectors. All of the industries in which a given activity or process may result in worker exposure to beryllium are identified in the sections on the application group. The section on each application group describes the production processes where occupational contact with beryllium can occur and contains estimates of the total number of firms, employees, affected establishments, and affected employees.

Chapter III of the PEA presents formulas in the text, usually in parentheses, to help explain the derivation of estimates. Because the values used in the formulas shown in the text are sometimes rounded, while the actual spreadsheet formulas used to create final costs are not, the calculation using the presented formula will sometimes differ slightly from the total presented in the text—which is the actual total as shown in the tables.

At the end of Chapter III in the PEA, OSHA discusses other industry sectors that have reportedly used beryllium in the past or for which there are anecdotal or informal reports of beryllium use. The Agency was unable to verify beryllium use in these sectors that would be affected by the proposed standard, and seeks further information in this rulemaking on these or other industries where there may be significant beryllium use and employee exposure.

2. Summary of Affected Establishments and Employers

As shown in Table IX-2, OSHA estimates that a total of 35,051 workers in 4,088 establishments will be affected by the proposed beryllium standard. Also shown are the estimated annual revenues for these entities.

3. Beryllium Exposure Profile of At-Risk Workers

The technological feasibility analyses presented in Chapter IV of the PEA contain data and discussion of worker exposures to beryllium throughout industry. Exposure profiles, by job category, were developed from individual exposure measurements that were judged to be substantive and to contain sufficient accompanying description to allow interpretation of the circumstance of each measurement. The resulting exposure profiles show the job categories with current overexposures to beryllium and, thus, the workers for whom beryllium controls would be implemented under the proposed rule.

Table IX-3 summarizes, from the exposure profiles, the number of workers at risk from beryllium exposure and the distribution of 8-hour TWA respirable beryllium exposures by affected job category and sector. Exposures are grouped into the following ranges: Less than 0.1 μg/m³; ≥ 0.1 μg/m³ and ≤ 0.2 μg/m³; > 0.2 μg/m³ and ≤ 0.5 μg/m³; > 0.5 μg/m³ and ≤ 1.0 μg/m³; > 1.0 μg/m³ and ≤ 2.0 μg/m³; and greater than 2.0 μg/m³. These frequencies represent the percentages of production employees in each job category and sector currently exposed at levels within the indicated range.

Table IX-4 presents data by NAICS code on the estimated number of workers currently at risk from beryllium exposure, as well as the estimated number of workers at risk of beryllium exposure above 0 μg/m³, at or above 0.1 μg/m³, at or above 0.2 μg/m³, at or above 0.5 μg/m³, at or above 1.0 μg/m³, and at or above 2.0 μg/m³. As shown, an estimated 12,101 workers currently have beryllium exposures at or above the proposed action level of 0.1 μg/m³; and an estimated 8,091 workers currently have beryllium exposures above the proposed PEL of 0.2 μg/m³.

D. Technological Feasibility Analysis of the Proposed Permissible Exposure Limit to Beryllium Exposures

This section summarizes the technological feasibility analysis presented in Chapter IV of the PEA (OSHA, 2014). The technological feasibility analysis includes information on current exposures, descriptions of engineering controls and other measures to reduce exposures, and a preliminary assessment of the technological feasibility of compliance with the proposed standard, including a reduction in OSHA's permissible exposure limits (PELs) in nine affected application groups. The current PELs for beryllium are 2.0 μg/m³ as an 8-hour time weighted average (TWA), and 5.0 μg/m³ as an acceptable ceiling concentration. OSHA is proposing a PEL of 0.2 μg/m³ as an 8-hour TWA and is additionally considering alternative TWA PELs of 0.1 and 0.5 μg/m³. OSHA is also proposing a 15-minute short-term exposure limit (STEL) of 2.0 μg/m³, and is considering alternative STELs of 0.5, 1.0 and 2.5 μg/m³.

The technological feasibility analysis includes nine application groups that correspond to specific industries or production processes that OSHA has preliminarily determined fall within the scope of the proposed standard. Within each of these application groups, exposure profiles have been developed that characterize the distribution of the available exposure measurements by job title or group of jobs. Descriptions of existing engineering controls for operations that create sources of beryllium exposure, and of additional engineering and work practice controls that can be used to reduce exposure are also provided. For each application group, a preliminary determination is made regarding the feasibility of achieving the proposed permissible exposure limits. For application groups in which the median exposures for some jobs exceed the proposed TWA PEL, a more detailed analysis is presented by job or group of jobs within the application group. The analysis is based on the best information currently available to the Agency, including a comprehensive review of the industrial hygiene literature, National Institute for Occupational Safety and Health (NIOSH) Health Hazard Evaluations and case studies of beryllium exposure, site visits conducted by an OSHA contractor (Eastern Research Group (ERG)), submissions to OSHA's rulemaking docket, and inspection data from OSHA's Integrated Management Information System (IMIS). OSHA also obtained information on production processes, worker exposures, and the effectiveness of existing control measures from the primary beryllium producer in the United States, Materion Corporation, and from interviews with industry experts.

The nine application groups included in this analysis were identified based on information obtained during preliminary rulemaking activities that included a SBRFA panel, a comprehensive review of the published literature, stakeholder input, and an analysis of IMIS data collected during OSHA workplace inspections where detectable airborne beryllium was found. The nine application groups and their corresponding section numbers in Chapter IV of the PEA are:

• Section 3—Beryllium Production,
• Section 4—Beryllium Oxide Ceramics and Composites,
• Section 5—Nonferrous Foundries,
• Section 6—Secondary Smelting, Refining, and Alloying,
• Section 7—Precision Turned Products,
• Section 8—Copper Rolling, Drawing, and Extruding,
• Section 9—Fabrication of Beryllium Alloy Products,
• Section 10—Welding, and
• Section 11—Dental Laboratories.

OSHA developed exposure profiles by job or group of jobs using exposure data at the application, operation or task level to the extent that such data were available. In those instances where there were insufficient exposure data to create a profile, OSHA used analogous operations to characterize the operations. The exposure profiles represent baseline conditions with existing controls for each operation with potential exposure. For job groups where exposures were above the proposed TWA PEL of 0.2 μg/m³, OSHA identified additional controls that could be implemented to reduce employee exposures to beryllium. These included engineering controls, such as process containment, local exhaust ventilation and wet methods for dust suppression, and work practices, such as improved housekeeping and the prohibition of compressed air for cleaning beryllium-contaminated surfaces.

For the purposes of this technological feasibility assessment, these nine application groups can be divided into three general categories based on current exposure levels:

(1) application groups in which current exposures for most jobs are already below the proposed PEL of 0.2 μg/m³;

(2) application groups in which exposures for most jobs are below the current PEL, but exceed the proposed PEL of 0.2 μg/m³, and therefore additional controls would be required; and

(3) application groups in which exposures in one or more jobs routinely exceed the current PEL, and therefore substantial reductions in exposure would be required to achieve the proposed PEL.

The majority of exposure measurements taken in the application groups in the first category are already at or below the proposed PEL of 0.2 μg/m³, and most of the jobs with exposure to beryllium in these four application groups have median exposures below the alternative PEL of 0.1 μg/m³ (See Table IX-5). These four application groups include rolling, drawing, and extruding; fabrication of beryllium alloy products; welding; and dental laboratories.

The two application groups in the second category include: precision turned products and secondary smelting. For these two groups, the median exposures in most jobs are below the current PEL, but the median exposure levels for some job groups currently exceed the proposed PEL. Additional exposure controls and work practices could be implemented that the Agency has preliminarily concluded would reduce exposures to or below the proposed PEL for most jobs most of the time. One exception is furnace operations in secondary smelting, in which the median exposure exceeds the current PEL. Furnace operations involve high temperatures that produce significant amounts of fumes and particulate that can be difficult to contain. Therefore, the proposed PEL may not be feasible for most furnace operations involved with secondary smelting, and in some cases, respiratory protection would be required to adequately protect furnace workers when exposures exceed 0.2 μg/m³ despite the implementation of all feasible controls.

Exposures in the third category of application groups routinely exceed the current PEL for several jobs. The three application groups in this category include: Beryllium production, beryllium oxide ceramics production, and nonferrous foundries. The individual job groups for which exposures exceed the current PEL are discussed in the application group specific sections later in this summary, and described in greater detail in the PEA. For the jobs that routinely exceed the current PEL, OSHA identified additional exposure controls and work practices that the Agency preliminarily concludes would reduce exposures to or below the proposed PEL most of the time, with three exceptions: Furnace operations in primary beryllium production and nonferrous foundries, and shakeout operations at nonferrous foundries. For these jobs, OSHA recognizes that even after installation of feasible controls, respiratory protection may be needed to adequately protect workers.

In conclusion, the preliminary technological feasibility analysis shows that for the majority of the job groups evaluated, exposures are either already at or below the proposed PEL, or can be adequately controlled with additional engineering and work practice controls. Therefore, OSHA preliminarily concludes that the proposed PEL of 0.2 μg/m³ is feasible for most operations most of the time. The preliminary feasibility determination for the proposed PEL is also supported by Materion Corporation, the sole primary beryllium production company in the U.S., and by the United Steelworkers, who jointly submitted a draft proposed standard that specified an exposure limit of 0.2 μg/m³ to OSHA (Materion and USW, 2012). The technological feasibility analysis conducted for each application group is briefly summarized below, and a more detailed discussion is presented in Sections 3 through 11 of Chapter IV of the PEA (OSHA, 2014).

Based on the currently available evidence, it is more difficult to determine whether an alternative PEL of 0.1 μg/m³ would also be feasible in most operations. For some application groups, such as fabrication of beryllium alloy products, a PEL of 0.1 μg/m³ would almost certainly be feasible. In other application groups, such as precision turned products, a PEL of 0.1 μg/m³ appears feasible, except for establishments working with high beryllium content alloys. For application groups with the highest exposure, the exposure monitoring data necessary to more fully evaluate the effectiveness of exposure controls adopted after 2000 are not currently available to OSHA, which makes it difficult to determine the feasibility of achieving exposure levels at or below 0.1 μg/m³.

OSHA also evaluated the feasibility of a STEL of 2.0 μg/m³, and alternative STELs of 0.5 and 1.0 μg/m³. An analysis of the available short-term exposure measurements indicates that elevated exposures can occur during short-term tasks such as those associated with the operation and maintenance of furnaces at primary beryllium production facilities, at nonferrous foundries, and at secondary smelting operations. Peak exposure can also occur during the transfer and handling of beryllium oxide powders. OSHA believes that in many cases, reducing short-term exposures will be necessary to reduce workers' TWA exposures to or below the proposed PEL. The majority of the available short-term measurements are below 2.0 μg/m³, therefore OSHA preliminarily concludes that the proposed STEL of 2.0 μg/m³ can be achieved for most operations most of the time. OSHA recognizes that for a small number of tasks, short-term exposures may exceed the proposed STEL, even after feasible control measures to reduce TWA exposure to below the proposed PEL have been implemented, and therefore assumes that the use of respiratory protection will continue to be required for some short-term tasks. It is more difficult based on the currently available evidence to determine whether the alternative STEL of 1.0 μg/m³ would also be feasible in most operations based on lack of detail in the activities of the workers presented in the data. OSHA expects additional use of respiratory protection would be required for tasks in which peak exposures can be reduced to less than 2.0 μg/m³ but not less than 1.0 μg/m³. Due to limitations in the available sampling data and the higher detection limits for short term measurements, OSHA could not determine the percentage of the STEL measurements that are less than or equal to 0.5 μg/m³. A detailed discussion of the STELs being considered by OSHA is presented in Section 12 of Chapter IV of the PEA (OSHA, 2014).

OSHA requests available exposure monitoring data and comments regarding the effectiveness of currently implemented control measures and the feasibility of the PELs under consideration, particularly the proposed TWA PEL of 0.2 μg/m³, the alternative TWA PEL of 0.1 μg/m³, the proposed STEL of 2.0 μg/m³, and the alternative STEL of 1.0 μg/m³ to inform the Agency's final feasibility determinations.

Application Group Summaries

This section summarizes the technological feasibility analysis for each of the nine application groups affected by the proposed standard. Chapter IV of the PEA, Technological Feasibility Analysis, identifies specific jobs or job groups with potential exposure to beryllium, and presents exposure profiles for each of these job groups (OSHA, 2014). Control measures and work practices that OSHA believes can reduce exposures are described along with preliminary conclusions regarding the feasibility of the proposed PEL. Table IX-5, located at the end of this summary, presents summary statistics for the personal breathing zone samples taken to measure full-shift exposures to beryllium in each application group. For the five application groups in which the median exposure level for at least one job group exceeds the proposed PEL, the sampling results are presented by job group. Table IX-5 displays the number of measurements; the range, the mean and the median of the measurement results; and the percentage of measurements less than 0.1 μg/m³, less than or equal to the proposed PEL of 0.2 μg/m³, and less than or equal to the current PEL of 2.0 μg/m³. A more detailed discussion of exposure levels by job or job group for each application group is provided in Chapter IV of the PEA, sections 3 through 11, along with a description of the available exposure measurement data, existing controls, and additional controls that would be required to achieve the proposed PEL.

Beryllium Production

Only one primary beryllium production facility is currently in operation in the United States, a plant owned and operated by Materion Corporation, located in Elmore, Ohio. OSHA identified eight job groups at this facility in which workers are exposed to beryllium. These include: Chemical operations, powdering operations, production support, cold work, hot work, site support, furnace operations, and administrative work.

The Agency developed an exposure profile for each of these eight job groups to analyze the distribution of exposure levels associated with primary beryllium production. The job exposure profiles are based primarily on full-shift personal breathing zone (PBZ) (lapel-type) sample results from air monitoring conducted by Brush Wellman's primary production facility in 1999 (Brush Wellman, 2004). Starting in 2000, the company developed the Materion Worker Protection Program (MWPP), a multi-faceted beryllium exposure control program designed to reduce airborne exposures for the vast majority of workers to less than an internally established exposure limit of 0.2 μg/m³. According to information provided by Materion, a combination of engineering controls, work practices, and housekeeping were used together to reduce average exposure levels to below 0.2 μg/m³ for the majority of workers (Materion Information Meeting, 2012). Also, two operations with historically high exposures, the wet plant and pebble plants, were decommissioned in 2000, thereby reducing average exposure levels. Therefore, the samples taken prior to 2000 may overestimate current exposures.

Additional exposure samples were taken by NIOSH at the Elmore facility from 2007 through 2008 (NIOSH, 2011). This dataset, which was made available to OSHA by Materion, contains fewer samples than the 1999 survey. OSHA did not incorporate these samples into the exposure profile due to the limited documentation associated with the sampling data. The lack of detailed information for individual samples has made it difficult for OSHA to correlate job classifications and identify the working conditions associated with the samples. Sampling data provided by Materion for 2007 and 2008 were not incorporated into the exposure profiles because the data lacked specific information on jobs and workplace conditions. In a meeting in May 2012 held between OSHA and Materion Corporation at the Elmore facility, the Agency was able to obtain some general information on the exposure control modifications that Materion Corporation made between 1999 and 2007, but has been unable to determine what specific controls were in place at the time NIOSH conducted sampling (Materion Information Meeting, 2012).

In five of the primary production job groups (i.e., hot work, cold work, production support, site support, and administrative work), the baseline exposure profile indicates that exposures are already lower than the proposed PEL of 0.2 μg/m³. Median exposure values for these job groups range from nondetectable to 0.08 μg/m³.

For three of the job groups involved with primary beryllium production, (i.e., chemical operations, powdering, and furnace operations), the median exposure level exceeds the proposed PEL of 0.2 μg/m³. Median exposure values for these job groups are 0.47, 0.37, and 0.68 μg/m³ respectively, and only 17 percent to 29 percent of the available measurements are less than or equal to 0.2 μg/m³. Therefore, additional control measures for these job groups would be required to achieve compliance with the proposed PEL. OSHA has identified several engineering controls that the Agency preliminarily concludes can reduce exposures in chemical processes and powdering operations to less than or equal to 0.2 μg/m³. In chemical processes, these include fail-safe drum-handling systems, full enclosure of drum-handling systems, ventilated enclosures around existing drum positions, automated systems to prevent drum overflow, and automated systems for container cleaning and disposal such as those designed for hazardous powders in the pharmaceutical industry. Similar engineering controls would reduce exposures in powdering operations. In addition, installing remote viewing equipment (or other equally effective engineering controls) to eliminate the need for workers to enter the die-loading hood during die filling will reduce exposures associated with this powdering task and reduce powder spills. Based on the availability of control methods to reduce exposures for each of the major sources of exposure in chemical operations, OSHA preliminarily concludes that exposures at or below the proposed 0.2 μg/m³ PEL can be achieved in most chemical and powdering operations most of the time. OSHA believes furnace operators' exposures can be reduced using appropriate ventilation, including fume capture hoods, and other controls to reduce overall beryllium levels in foundries, but is not certain whether the exposures of furnace operators can be reduced to the proposed PEL with currently available technology. OSHA requests additional information on current exposure levels and the effectiveness of potential control measures for primary beryllium production operations to further refine this analysis.

Beryllium Oxide Ceramics Production

OSHA identified seven job groups involved with beryllium oxide ceramics production. These include: Material preparation operator, forming operator, machining operator, kiln operator, production support, metallization, and administrative work. Four of these jobs (material preparation, forming operator, machining operator and kiln operator) work directly with beryllium oxides, and therefore these jobs have a high potential for exposure. The other three job groups (production support work, metallization, and administrative work) have primarily indirect exposure that occurs only when workers in these jobs groups enter production areas and are exposed to the same sources to which the material preparation, forming, machining and kiln operators are directly exposed. However, some production support and metallization activities do require workers to handle beryllium directly, and workers performing these tasks may at times be directly exposed to beryllium.

The Agency developed exposure profiles for these jobs based on air sampling data from four sources: (1) Samples taken between 1994 and 2003 at a large beryllium oxide ceramics facility, (2) air sampling data obtained during a site visit to a primary beryllium oxide ceramics producer, (3) a published report that provides information on beryllium oxide ceramics product manufacturing for a slightly earlier time period, and (4) exposure data from OSHA's Integrated Management Information System (OSHA, 2009). The exposure profile indicates that the three job groups with mostly indirect exposure (production support work, metallization, and administrative work) already achieve the proposed PEL of 0.2 μg/m³. Median exposure sample values for these job groups did not exceed 0.06 μg/m³.

The four job groups with direct exposure had higher exposures. In forming operations and machining operations, the median exposure levels of 0.18 and 0.15 ug/m³, respectively, are below the proposed PEL, while the median exposure levels for material preparation and kiln operations of 0.41 μg/m³ and 0.25 μg/m³, respectively, exceed the proposed PEL.

The profile for the directly exposed jobs may overestimate exposures due to the preponderance of data from the mid-1990s, a time period prior to the implementation of a variety of exposure control measures introduced after 2000. In forming operations, 44 percent of sample values in the exposure profile exceeded 0.2 ug/m³. However, the median exposure levels for some tasks, such as small-press and large-press operation, based on sampling conducted in 2003 were below 0.1 μg/m³. The exposure profile for kiln operation was based on three samples taken from a single facility in 1995, and are all above 0.2 ug/m³. Since then, exposures at the facility have declined due to changes in operations that reduced the amount of time kiln operators spend in the immediate vicinity of the kilns, as well as the discontinuation of a nearby high-exposure process. More recent information communicated to OSHA suggests that current exposures for kiln operators at the facility are currently below 0.1 ug/m³. Exposures in machining operations, most of which were already below 0.2 ug/m³ during the 1990s, may have been further reduced since then through improved work practices and exposure controls (PEA Chapter IV, Section 7). For forming, kiln, and machining operations, OSHA preliminarily concludes that the installation of additional controls such as machine interlocks (for forming) and improved enclosures and ventilation will reduce exposures to or below the proposed PEL most of the time. OSHA requests information on recent exposure levels and controls in beryllium oxide forming and kiln operations to help the Agency evaluate the effectiveness of available exposure controls for this application group.

In the exposure profile for material preparation, 73 percent of sample values exceeded 0.2 ug/m. As with other parts of the exposure profile, exposure values from the mid-1990s may overestimate airborne beryllium levels for current operations. During most material preparation tasks, such as material loading, transfer, and spray drying, OSHA preliminarily concludes that exposures can be reduced to or below 0.2 μg/m with process enclosures, ventilation hoods, and improved housekeeping procedures. However, OSHA acknowledges that peak exposures from some short-term tasks such as servicing of the spray chamber might continue to drive the TWA exposures above 0.2 μg/m on days when these material preparation tasks are performed. Respirators may be needed to protect workers from exposures above the proposed TWA PEL during these tasks. OSHA notes that material preparation for production of beryllium oxide ceramics currently takes place at only two facilities in the United States.

Nonferrous Foundries

OSHA identified eight job groups in aluminum and copper foundries with beryllium exposure: Molding, material handling, furnace operation, pouring, shakeout operation, abrasive blasting, grinding/finishing, and maintenance. The Agency developed exposure profiles based on an air monitoring survey conducted by NIOSH in 2007, a Health Hazard Evaluation (HHE) conducted by NIOSH in 1975, a site visit by ERG in 2003, a site visit report from 1999 by the California Cast Metals Association (CCMA); and two sets of data from air monitoring surveys obtained from Materion in 2004 and 2010.

The exposure profile indicates that in foundries processing beryllium alloys, six of the eight job groups have median exposures that exceed the proposed PEL of 0.2 μg/m³ with baseline working conditions. One exception is grinding/finishing operations, where the median value is 0.12 μg/m³ and 73 percent of exposure samples are below 0.2 μg/m³. The other exception is abrasive blasting. The samples for abrasive blasting used in the exposure profile were obtained during blasting operations using enclosed cabinets, and all 5 samples were below 0.2 μg/m³. Exposures for other job groups ranged from just below to well above the proposed PEL, including molder (all samples above 0.2 μg/m³), material handler (1 sample total, above 0.2 μg/m³), furnace operator (81.8 percent of samples above 0.2 μg/m³), pouring operator (60 percent of samples above 0.2 μg/m³), shakeout operator (1 sample total, above 0.2 μg/m³), and maintenance worker (50 percent of samples above 0.2 μg/m³).

In some of the foundries at which the air samples included in the exposure profile were collected, there are indications that the ventilation systems were not properly used or maintained, and dry sweeping or brushing and the use of compressed air systems for cleaning may have contributed to high dust levels. OSHA believes that exposures in foundries can be substantially reduced by improving and properly using and maintaining the ventilation systems; switching from dry brushing, sweeping and compressed air to wet methods and use of HEPA-filtered vacuums for cleaning molds and work areas; enclosing processes; automation of high-exposure tasks; and modification of processes (e.g., switching from sand-based to alternative casting methods). OSHA preliminarily concludes that these additional engineering controls and modified work practices can be implemented to achieve the proposed PEL most of the time for molding, material handling, maintenance, abrasive blasting, grinding/finishing, and pouring operations at foundries that produce aluminum and copper beryllium alloys.

The Agency is less confident that exposure can be reliably reduced to the proposed PEL for furnace and shakeout operators. Beryllium concentrations in the proximity of the furnaces are typically higher than in other areas due to the fumes generated and the difficulty of controlling emissions during furnace operations. The exposure profile for furnace operations shows a median beryllium exposure level of 1.14 μg/m³. OSHA believes that furnace operators' exposures can be reduced using local exhaust ventilation and other controls to reduce overall beryllium levels in foundries, but it is not clear that they can be reduced to the proposed PEL with currently available technology. In foundries that use sand molds, the shakeout operation typically involves removing the freshly cast parts from the sand mold using a vibrating grate that shakes the sand from castings. The shakeout equipment generates substantial amounts of airborne dust that can be difficult to contain, and therefore shakeout operators are typically exposed to high dust levels. During casting of beryllium alloys, the dust may contain beryllium and beryllium oxide residues dislodged from the casting during the shakeout process. The exposure profile for the shakeout operations contains only one result of 1.3 μg/m³. This suggests that a substantial reduction would be necessary to achieve compliance with a proposed PEL of 0.2 μg/m³. OSHA requests additional information on recent employee exposure levels and the effectiveness of dust controls for shakeout operations for copper and aluminum alloy foundries.

Secondary Smelting, Refining, and Alloying

OSHA identified two job groups in this application group with exposure to beryllium: Mechanical process operators and furnace operations workers. Mechanical operators handle and treat source material, and furnace operators run heating processes for refining, melting, and casting metal alloy. OSHA developed exposure profiles for these jobs based on exposure data from ERG site visits to a precious/base metals recovery facility and a facility that melts and casts beryllium-containing alloys, both conducted in 2003. The available exposure data for this application group are limited, and therefore, the exposure profile is supplemented in part by summary data presented in secondary sources of information on beryllium exposures in this application group.

The exposure profile for mechanical processing operators indicates low exposures (3 samples less than 0.2 μg/m³), even though these samples were collected at a facility where the ventilation system was allowing visible emissions to escape exhaust hoods. Summary data from studies and reports published in 2005-2009 showed that mechanical processing operator exposures averaged between 0.01 and 0.04 μg/m³ at facilities where mixed or electronic waste including beryllium alloy parts were refined. Based on these results, OSHA preliminarily concludes that the proposed PEL is already achieved for most mechanical processing operations most of the time, and exposures could be further reduced through improved ventilation system design and other measures, such as process enclosures.

As with furnace operations examined in other application groups, the exposure profile indicates higher worker exposures for furnace operators in the secondary smelting, refining, and alloying application group (six samples with a median of 2.15 μg/m³, and 83.3 percent above 0.2 μg/m³). The two lowest samples in this job's exposure profile (0.03 and 0.5 μg/m³) were collected at a facility engaged in recycling and recovery of precious metals where work with beryllium-containing material is incidental. At this facility, the furnace is enclosed and fumes are ducted into a filtration system. The four higher samples, ranging from 1.92 to 14.08 μg/m³, were collected at a facility engaged primarily in beryllium alloying operations, where beryllium content is significantly higher than in recycling and precious metal recovery activities, the furnace is not enclosed, and workers are positioned directly in the path of the exhaust ventilation over the furnace. OSHA believes these exposures could be reduced by enclosing the furnace and repositioning the worker, but is not certain whether the reduction achieved would be enough to bring exposures down to the proposed PEL. Based on the limited number of samples in the exposure profile and surrogate data from furnace operations, the proposed PEL may not be feasible for furnace work in beryllium recovery and alloying, and respirators may be necessary to protect employees performing these tasks.

Precision Turned Products

OSHA's preliminary feasibility analysis for precision turned products focuses on machinists who work with beryllium-containing alloys. The Agency also examined the available exposure data for non-machinists and has preliminarily concluded that, in most cases, controlling the sources of exposures for machinists will also reduce exposures for other job groups with indirect exposure when working in the vicinity of machining operations.

OSHA developed exposure profiles based on exposure data from four NIOSH surveys conducted between 1976 and 2008; ERG site visits to precision machining facilities in 2002, 2003, and 2004; case study reports from six facilities machining copper-beryllium alloys; and exposure data collected between 1987 and 2001 by the U.S. Navy Environmental Health Center (NEHC). Analysis of the exposure data showed a substantial difference between the median exposure level for workers machining pure beryllium and/or high-beryllium alloys compared to workers machining low-beryllium alloys. Most establishments in the precision turned products application group work only with low-beryllium alloys, such as copper-beryllium. A relatively small number of establishments (estimated at 15) specialize in precision machining of pure beryllium and/or high-beryllium alloys.

The exposure profile indicates that machinists working with low-beryllium alloys have mostly low exposure to airborne beryllium. Approximately 85 percent of the 80 exposure results are less than or equal to 0.2 μg/m³, and 74 percent are less than or equal to 0.1 μg/m³. Some of the results below 0.1 μg/m³ were collected at a facility where machining operations were enclosed, and metal cutting fluids were used to control the release of airborne contaminants. Higher results (0.1 μg/m³-1.07 μg/m³) were found at a facility where cutting and grinding operations were conducted in partially enclosed booths equipped with LEV, but some LEV was not functioning properly. A few very high results (0.77 μg/m³-24 μg/m³) were collected at a facility where exposure controls were reportedly inadequate and poor work practices were observed (e.g., improper use of downdraft tables, use of compressed air for cleaning). Based on these results, OSHA preliminarily concludes that exposures below 0.2 μg/m³ can be achieved most of the time for most machinists at facilities dealing primarily with low-beryllium alloys. OSHA recognizes that higher exposures may sometimes occur during some tasks where exposures are difficult to control with engineering methods, such as cleaning, and that respiratory protection may be needed at these times.

Machinists working with high-beryllium alloys have higher exposure than those working with low-beryllium alloys. This difference is reflected in the exposure profile for this job, where the median of exposure is 0.31 μg/m³ and 75 percent of samples exceed the proposed PEL of 0.2 μg/m³. The exposure profile was based on two machining facilities at which LEV was used and machining operations were performed under a liquid coolant flood. Like most facilities where pure beryllium and high-beryllium alloys are machined, these facilities also used some combination of full or partial enclosures, as well as work practices to minimize exposure such as prohibiting the use of compressed air and dry sweeping and implementing dust migration control practices to prevent the spread of beryllium contamination outside production areas. At one facility machining high-beryllium alloys, where all machining operations were fully enclosed and ventilated, exposures were mostly below 0.1 μg/m³ (median 0.035 μg/m³, range 0.02-0.11 μg/m³). Exposures were initially higher at the second facility, where some machining operations were not enclosed, existing LEV system were in need of upgrades, and some exhaust systems were improperly positioned. Samples collected there in 2003 and 2004 were mostly below the proposed PEL in 2003 (median 0.1 μg/m³) but higher in 2004 (median 0.25 μg/m³), and high exposure means in both years (1.65 and 0.68 μg/m³ respectively) show the presence of high exposure spikes in the facility. However, the facility reported that measures to reduce exposure brought almost all machining exposures below 0.2 μg/m³ in 2006. With the use of fully enclosed machines and LEV and work practices that minimize worker exposures, OSHA preliminarily concludes that the proposed PEL is feasible for the vast majority of machinists working with pure beryllium and high-beryllium alloys. OSHA recognizes that higher exposures may sometimes occur during some tasks where exposures are difficult to control with engineering methods, such as machine cleaning and maintenance, and that respiratory protection may be needed at these times.

Copper Rolling, Drawing, and Extruding

OSHA's exposure profile for copper rolling, drawing, and extruding includes four job groups with beryllium exposure: strip metal production, rod and wire production, production support, and administrative work. Exposure profiles for these jobs are based on personal breathing zone lapel sampling conducted at the Brush Wellman Reading, Pennsylvania, rolling and drawing facility from 1977 to 2000.

Prior to 2000, the Reading facility had limited engineering controls in place. Equipment in use included LEV in some operations, HEPA vacuums for general housekeeping, and wet methods to control loose dust in some rod and wire production operations. The exposure profile shows very low exposures for all four job groups. All had median exposure values below 0.1 μg/m³, and in strip metal production, production support, and administrative work, over 90 percent of samples were below 0.1 μg/m³. In rod and wire production, 70 percent of samples were below 0.1 μg/m³.

To characterize exposures in extrusion, OSHA examined the results of an industrial hygiene survey of a copper-beryllium extruding process conducted in 2000 at another facility. The survey reported eight PBZ samples, which were not included in the exposure profile because of their short duration (2 hours). Samples for three of the four jobs involved with the extrusion process (press operator, material handler, and billet assembler) were below the limit of detection (LOD) (level not reported). The two samples for the press operator assistant, taken when the assistant was buffing, sanding, and cleaning extrusion tools, were very high (1.6 and 1.9 μg/m³). Investigators recommended a ventilated workstation to reduce exposure during these activities.

In summary, exposures at or below 0.2 μg/m³ have already been achieved for most jobs in rolling, drawing, and extruding operations, and OSHA preliminarily concludes that the proposed PEL of 0.2 μg/m³ is feasible for this application group. For jobs or tasks with higher exposures, such as tool refinishing, use of exposure controls such as local exhaust ventilation can help reduce workers' exposures. The Agency recognizes the limitations of the available data, which were drawn from two facilities and did not include full-shift PBZ samples for extrusion. OSHA requests additional exposure data from other facilities in this application group, especially data from facilities where extrusion is performed.

Fabrication of Beryllium Alloy Products

This application group includes the fabrication of beryllium alloy springs, stampings, and connectors for use in electronics. The exposure profile is based on a study conducted at four precision stamping companies; a NIOSH report on a spring and stamping company; an ERG site visit to a precision stamping, forming, and plating establishment; and exposure monitoring results from a stamping facility presented at the American Industrial Hygiene Conference and Exposition in 2007. The exposure profiles for this application group include three jobs: chemical processing operators, deburring operators, and assembly operators. Other jobs for which all samples results were below 0.1 μg/m³ are not shown in the profile.

For the three jobs in the profile, the majority of exposure samples were below 0.1 μg/m³ (deburring operators, 79 percent; chemical processing operators, 81 percent; assembly operators, 93 percent). Based on these results, OSHA preliminarily concludes that the proposed PEL is feasible for this application group. The Agency notes that a few exposures above the proposed PEL were recorded for the chemical processing operator (in plating and bright cleaning) and for deburring (during corn cob deburring in an open tumbling mill). OSHA believes the use of LEV, improved housekeeping, and work practice modifications would reduce the frequency of excursions above the proposed PEL.

Welding

Most of the samples in OSHA's exposure profile for welders in general industry were collected between 1994 and 2001 at two of Brush Wellman's alloy strip distribution centers, and in 1999 at Brush Wellman's Elmore facility. At these facilities, tungsten inert gas (TIG) welding was conducted on beryllium alloy strip. Seven samples in the exposure profile came from a case study conducted at a precision stamping facility, where airborne beryllium levels were very low (see previous summary, Fabrication of Beryllium Alloy Products). At this facility, resistance welding was performed on copper-beryllium parts, and welding processes were automated and enclosed.

Most of the sample results in the welding exposure profile were below 0.2 μg/m³. Of the 44 welding samples in the profile, 75 percent were below 0.2 μg/m³ and 64 percent were below 0.1 μg/m³, with most values between 0.01 and 0.05 μg/m³. All but one of the 16 exposure samples above 0.1 μg/m³ were collected in Brush Wellman's Elmore facility in 1999. According to company representatives, these higher exposure levels may have been due to beryllium oxide that can form on the surface of the material as a result of hot rolling. All seven samples from the precision stamping facility were below the limit of detection. Based on these results, OSHA preliminarily concludes that the proposed PEL of 0.2 μg/m³ is feasible for most welding operations in general industry.

Dental Laboratories

OSHA's exposure profile for dental technicians includes sampling results from a site visit conducted by ERG in 2003; a study of six dental laboratories published by Rom et al. in 1984; a data set of exposure samples collected between 1987 and 2001, on dental technicians working for the U.S. Navy; and a docket submission from CMP Industries including two samples from a large commercial dental laboratory using nickel-beryllium alloy. Information on exposure controls in these facilities suggests that controls in some cases may have been absent or improperly used.

The exposure profile indicates that 52 percent of samples are less than or equal to 0.2 μg/m³. However, the treatment of nondetectable samples in the feasibility analysis may overestimate many of the sample values in the exposure profile. Twelve of the samples in the profile are nondetectable for beryllium. In the exposure profile, these were assigned the highest possible value, the limit of detection (LOD). For eight of the nondetectable samples, the LOD was reported as 0.2 μg/m³. For the other four nondetectable samples, the LOD was between 0.23 and 0.71 μg/m³. If the true values for these four nondetectable samples are actually less than or equal to the assigned value of 0.2 μg/m³, then the true percentage of profile sample values less than or equal to 0.2 μg/m³ is between 52 and 70 percent. Of the sample results with detectable beryllium above 0.2 μg/m³, some were collected in 1984 at facilities studied by Rom et al., who reported that they occurred during grinding with LEV that was improperly used or, in one case, not used at all. Others were collected at facilities where little contextual information was available to determine what control equipment or work practices might have reduced exposures.

Based on this information, OSHA preliminarily concludes that beryllium exposures for most dental technicians are already below 0.2 μg/m³ most of the time. OSHA furthermore believes that exposure levels can be reduced to or below 0.1 μg/m³ most of the time via material substitution, engineering controls, and work practices. Beryllium-free alternatives for casting dental appliances are readily available from commercial sources, and some alloy suppliers have stopped carrying alloys that contain beryllium. For those dental laboratories that continue to use beryllium alloys, exposure control options include properly designed, installed, and maintained LEV systems (equipped with HEPA filters) and enclosures; work practices that optimize LEV system effectiveness; and housekeeping methods that minimize beryllium contamination in the workplace. In summary, OSHA preliminarily concludes that the proposed PEL is feasible for dental laboratories.

E. Costs of Compliance

Chapter V of the PEA in support of the proposed beryllium rule provides a detailed assessment of the costs to establishments in all affected application groups of reducing worker exposures to beryllium to an eight-hour time-weighted average (TWA) permissible exposure limit (PEL) of 0.2 μg/m³ and to the proposed short-term exposure limit (STEL) of 2.0 μg/m³, as well as of complying with the proposed standard's ancillary provisions. OSHA describes its methodology and sources in more detail in Chapter V. OSHA's preliminary cost assessment is based on the Agency's technological feasibility analysis presented in Chapter IV of the PEA; analyses of the costs of the proposed standard conducted by OSHA's contractor, Eastern Research Group (ERG); and the comments submitted to the docket in response to the request for information (RFI) and as part of the SBREFA process.

As shown in Table IX-7 at the end of this section, OSHA estimates that the proposed standard would have an annualized cost of $37.6 million. All cost estimates are expressed in 2010 dollars and were annualized using a discount rate of 3 percent, which—along with 7 percent—is one of the discount rates recommended by OMB. Annualization periods for expenditures on equipment are based on equipment life, and one-time costs are annualized over a 10-year period.

The estimated costs for the proposed beryllium rule represent the additional costs necessary for employers to achieve full compliance. They do not include costs associated with current compliance that may already have been achieved with regard to existing beryllium requirements or costs necessary to achieve compliance with existing beryllium requirements, to the extent that some employers may currently not be fully complying with applicable regulatory requirements.

Throughout this section and in the PEA, OSHA presents cost formulas in the text, usually in parentheses, to help explain the derivation of cost estimates for individual provisions. Because the values used in the formulas shown in the text are shown only to the second decimal place, while the actual spreadsheet formulas used to create final costs are not limited to two decimal places, the calculation using the presented formula will sometimes differ slightly from the presented total in the text, which is the actual and mathematically correct total as shown in the tables.

1. Compliance With the Proposed PEL/STEL

OSHA's estimate of the costs for affected employers to comply with the proposed PEL of 0.2 μg/m³ and the proposed STEL of 2.0 μg/m³ consists of two parts. First, costs are estimated for the engineering controls, additional studies and custom design requirements to implement those controls, work practices, and specific training required for those work practices (as opposed to general training in compliance with the rule) needed for affected employers to meet the proposed PEL and STEL, as well as opportunity costs (lost productivity) that may result from working with some of the new controls. In most cases, the PEA breaks out these costs, but in other instances some or all of the costs are shortened simply to “engineering controls” in the text, for convenience. Second, for employers unable to meet the proposed PEL and STEL using engineering controls and work practices alone, costs are estimated for respiratory protection sufficient to reduce worker exposure to the proposed PEL and STEL or below.

In the technological feasibility analysis presented in Chapter IV of the PEA, OSHA concluded that implementing all engineering controls and work practices necessary to reach the proposed PEL will, except for a small residual group (accounting for about 6 percent of all exposures above the STEL), also reduce exposures below the STEL. However, based on the nature of the processes this residual group is likely to be engaged in, the Agency expects that employees would already be using respirators to comply with the PEL under the proposed standard. Therefore, with the proposed STEL set at ten times the proposed PEL, the Agency has preliminarily determined that engineering controls, work practices, and (when needed) respiratory protection sufficient to meet the proposed PEL are also sufficient to meet the proposed STEL. For that reason, OSHA has taken no additional costs for affected employers to meet the proposed STEL. The Agency invites comment and requests that the public provide data on this issue.

a. Engineering Controls

For this preliminary cost analysis, OSHA estimated the necessary engineering controls and work practices for each affected application group according to the exposure profile of current exposures by occupation presented in Chapter III of the PEA. Under the requirements of the proposed standard, employers would be required to implement engineering or work practice controls whenever beryllium exposures exceed the proposed PEL of 0.2 μg/m³ or the proposed STEL of 2.0 μg/m³.

In addition, even if employers are not exposed above the proposed PEL or proposed STEL, paragraph (f)(2) of the proposed standard would require employers at or above the action level to use at least one engineering or work practice control to minimize worker exposure. Based on the technological feasibility analysis presented in Chapter IV of the PEA, OSHA has determined that, for only two job categories in two application groups—chemical process operators in the Stamping, Spring and Connection Manufacture application group and machinists in the Machining application group—do the majority of facilities at or above the proposed action level, but below the proposed PEL, lack the baseline engineering or work controls required by paragraph (f)(2). Therefore, OSHA has estimated costs, where appropriate, for employers in these two application groups to comply with paragraph (f)(2).

By assigning controls based on application group, the Agency is best able to identify those workers with exposures above the proposed PEL and to design a control strategy for, and attribute costs specifically to, these groups of workers. By using this approach, controls are targeting those specific processes, emission points, or procedures that create beryllium exposures. Moreover, this approach allows OSHA to assign costs for technologies that are demonstrated to be the most effective in reducing exposures resulting from a particular process.

In developing cost estimates, OSHA took into account the wide variation in the size or scope of the engineering or work practice changes necessary to minimize beryllium exposures based on technical literature, judgments of knowledgeable consultants, industry observers, and other sources. The resulting cost estimates reflect the representative conditions for the affected workers in each application group and across all work settings. In all but a handful of cases (with the exceptions noted in the PEA), all wage costs come from the 2010 Occupational Employment Statistics (OES) of the Bureau of Labor Statistics (BLS, 2010a) and utilize the median wage for the appropriate occupation. The wages used include a 30.35 percent markup for fringe benefits as a percentage of total compensation, which is the average percentage markup for fringe benefits for all civilian workers from the 2010 Employer Costs for Employee Compensation of the BLS (BLS, 2010b). All descriptions of production processes are drawn from the relevant sections of Chapter IV of the PEA.

The specific engineering costs for each of the applications groups, and the NAICS industries that contain those application groups, are discussed in Chapter V of the PEA. Like the industry profile and technological feasibility analysis presented in other PEA chapters, Chapter V of the PEA presents engineering control costs for the following application groups:

Beryllium Production

Beryllium Oxide, Ceramics & Composites Production

Nonferrous Foundries

Stamping, Spring and Connection Manufacture

Secondary Smelting, Refining, and Alloying

Copper Rolling, Drawing, and Extruding

Secondary Smelting, Refining, and Alloying

Precision Machining

Welding

Dental Laboratories

The costs within these application groups are estimated by occupation and/or operation. One application group could have multiple occupations, operations, or activities where workers are exposed to levels of beryllium above the proposed PEL, and each will need its own set of controls. The major types of engineering controls needed to achieve compliance with the proposed PEL include ventilation equipment, pharmaceutical-quality high-containment isolators, decontainment chambers, equipment with controlled water sprays, closed-circuit remote televisions, enclosed cabs, conveyor enclosures, exhaust hoods, and portable local-exhaust-ventilation (LEV) systems. Capital costs and annual operation and maintenance (O&M) costs, as well as any other annual costs, are estimated for the set of engineering controls estimated to be necessary for limiting beryllium exposures for each occupation or operation within each application group.

Tables V-2 through V-10 in Chapter V of the PEA summarize capital, maintenance, and operating costs for each application group disaggregated by NAICS code. Table IX-7 at the end of this section breaks out the costs of engineering controls/work practices by application group and NAICS code.

Some engineering control costs are estimated on a per-worker basis and then multiplied by the estimated number of affected workers—as identified in Chapter III: Profile of Affected Industries in the PEA—to arrive at a total cost for a particular control within a particular application group. This worker-based method is necessary because—even though OSHA has data on the number of firms in each affected industry, the occupations and industrial activities that result in worker exposure to beryllium, and the exposure profile of at-risk occupations—the Agency does not have a way to match up these data at the firm level. Nor does the Agency have establishment-specific data on worker exposure to beryllium for all establishments, or even establishment-specific data on the level of activity involving worker exposure to beryllium. Thus, OSHA could not always directly estimate per-affected-establishment costs, but instead first had to estimate aggregate compliance costs (using an estimated per-worker cost multiplied by the number of affected workers) and then calculate the average per-affected-establishment costs by dividing those aggregate costs by the number of affected establishments. This method, while correct on average, may under- or over-state costs for certain firms. For other controls that are implemented on a fixed-cost basis per establishment (e.g., creating a training program, writing a control program), the costs are estimated on an establishment basis, and these costs were multiplied by the number of affected establishments in the given application group to obtain total control costs.

In developing cost estimates, the Agency sometimes had to make case-specific judgments about the number of workers affected by each engineering control. Because work environments vary within occupations and across establishments, there are no definitive data on how many workers are likely to have their exposures reduced by a given set of controls. In the smallest establishments, especially those that might operate only one shift per day, some controls would limit exposures for only a single worker in one specific affected occupation. More commonly, however, several workers are likely to benefit from each enhanced engineering control. Many controls were judged to reduce exposure for employees in multi-shift work or where workstations are used by more than one worker per shift.

In general, improving work practices involves operator training, actual work practice modifications, and better enforcement or supervision to minimize potential exposures. The costs of these process improvements consist of the supervisor and worker time involved and would include the time spent by supervisors to develop a training program.

Unless otherwise specified, OSHA viewed the extent to which exposure controls are already in place to be reflected in the distribution of exposures at levels above the proposed PEL among affected workers. Thus, for example, if 50 percent of workers in a given occupation are found to be exposed to beryllium at levels above the proposed PEL, OSHA judged this equivalent to 50 percent of facilities lacking adequate exposure controls. The facilities may have, for example, the correct equipment installed but without adequate ventilation to provide protection to workers exposed to beryllium. In this example, the Agency would expect that the remaining 50 percent of facilities to either have installed the relevant controls to reduce beryllium exposures below the PEL or that they engage in activities that do not require that the exposure controls be in place (for example, they do not perform any work with beryllium-containing materials). To estimate the need for incremental controls on a per-worker basis, OSHA used the exposure profile information as the best available data. OSHA recognizes that a very small percentage of facilities might have all the relevant controls in place but are still unable, for whatever reason, to achieve the proposed PEL through controls alone. ERG's review of the industrial hygiene literature and other source materials (ERG, 2007b), however, suggest that the large majority of workplaces where workers are exposed to high levels of beryllium lack at least some of the relevant controls. Thus, in estimating the costs associated with the proposed standard, OSHA has generally assumed that high levels of exposure to beryllium occur due to the absence of suitable controls. This assumption likely results in an overestimate of costs since, in some cases, employers may not need to install and maintain new controls in order to meet the proposed PEL but merely need to upgrade or better maintain existing controls, or to improve work practices.

b. Respiratory Protection Costs

Based on the findings of the technological feasibility analysis, a small subset of employees working with a few processes in a handful of application groups will need to use respirators, in addition to required engineering controls and improved work practices, to reduce employee exposures to meet the proposed PEL. Specifically, furnace operators—both in non-ferrous foundries (both sand and non-sand) and in secondary smelting, refining, and alloying—as well as welders in a few other processes, will need to wear half-mask respirators. In beryllium production, workers who rebuild or otherwise maintain furnaces and furnace tools will need to wear full-face powered air-purifying respirators. Finally, the Agency recognizes the possibility that, after all feasible engineering and other controls are in place, there may still be a residual group with potential exposure above the proposed PEL and/or STEL. To account for these residual cases, OSHA estimates that 10 percent of the workers, across all application groups and job categories, who are above the proposed PEL before the beryllium proposed standard is in place (according to the baseline exposure profile presented in Chapter III of the PEA), would still be above the PEL after all feasible controls are implemented and, hence, would need to use half-mask respirators to achieve compliance with the proposed PEL.

There are five primary costs for respiratory protection. First, there is a cost per establishment to set up a written respirator program in accordance with the respiratory protection standard (29 CFR 1910.134). The respiratory protection standard requires written procedures for the proper selection, use, cleaning, storage, and maintenance of respirators. As derived in the PEA, OSHA estimates that, when annualized over 10 years, the annualized per-establishment cost for a written respirator program is $207.

For reasons unrelated to the proposed standard, certain establishments will already have a respirator program in place. Table V-11 in Chapter V of the PEA presents OSHA's estimates, by application group, of current levels of compliance with the respirator program provision of the proposed rule.

The four other major costs of respiratory protection are the per-employee costs for all aspects of respirator use: equipment, training, fit-testing, and cleaning. Table V-12 of Chapter V in the PEA breaks out OSHA's estimate of the unit costs for the two types of respirators needed: A half-mask respirator and a full-face powered air-purifying respirator. As derived in the PEA, the annualized per-employee cost for a half-mask respirator would be $524 and the annualized per-employee cost for a full-face powered air-purifying respirator would be $1,017.

Table V-13 in Chapter V of the PEA presents the number of additional employees, by application group and NAICS code, that would need to wear respirators to comply with the proposed standard and the cost to industry to comply with the respirator protection provisions in the proposed rule. OSHA judges that only workers in Beryllium Production work with processes that would require a full-face respirator and estimates that there are 23 of those workers. Three hundred and eighteen workers in other assorted application groups are estimated to need half-mask respirators. A total of 341 employees would need to wear some type of respirator, resulting in a total annualized cost of $249,684 for affected industries to comply with the respiratory protection requirements of the proposed standard. Table IX-7 at the end of this section breaks out the costs of respiratory protection by application group and NAICS code.

2. Ancillary Provisions

This section presents OSHA's estimated costs for ancillary beryllium control programs required under the proposed rule. Based on the program requirements contained in the proposed standard, OSHA considered the following cost elements in the following employer duties: (a) Assess employees' exposure to airborne beryllium, (b) establish regulated areas, (c) develop a written exposure control plan, (d) provide protective work clothing, (e) establish hygiene areas and practices, (f) implement housekeeping measures, (g) provide medical surveillance, (h) provide medical removal for employees who have developed CBD or been confirmed positive for beryllium sensitization, and (i) provide appropriate training.

The worker population affected by each program element varies by several criteria discussed in detail in each subsection below. In general, some elements would apply to all workers exposed to beryllium at or above the action level. Other elements would apply to a smaller set of workers who are exposed above the PEL. The training requirements would apply to all employees who work in a beryllium work area (e.g., an area with any level of exposure to airborne beryllium). The regulated area program elements triggered by exposures exceeding the proposed PEL of 0.2 μ g/m³ would apply to those workers for whom feasible controls are not adequate. In the earlier discussion of respiratory protection, OSHA estimated that 10 percent of all affected workers with current exposures above the proposed PEL would fall in this category.

Costs for each program requirement are aggregated by employment and by industry. For the most part, unit costs do not vary by industry, and any variations are specifically noted. The estimated compliance rate for each provision of the proposed standard by application group is presented in Table V-15 of the PEA.

a. Exposure Assessment

Most establishments wishing to perform exposure monitoring would require the assistance of an outside consulting industrial hygienist (IH) to obtain accurate results. While some firms might already employ or train qualified staff, OSHA judged that the testing protocols are fairly challenging and that few firms have sufficiently skilled staff to eliminate the need for outside consultants.

The proposed standard requires that, after receiving the results of any exposure monitoring where exposures exceed the TWA PEL or STEL, the employer notify each such affected employee in writing of suspected or known sources of exposure, and the corrective action(s) being taken to reduce exposure to or below the PEL. Those workers exposed at or above the action level and at or below the PEL must have their exposure levels monitored annually.

For costing purposes, OSHA estimates that, on average, there are four workers per work area. OSHA interpreted the initial exposure assessment as requiring first-year testing of at least one worker in each distinct job classification and work area who is, or may reasonably be expected to be, exposed to airborne concentrations of beryllium at or above the action level.

The proposed standard requires that whenever there is a change in the production, process, control equipment, personnel, or work practices that may result in new or additional exposures, or when the employer has any reason to suspect that a change may result in new or additional exposures, the employer must conduct additional monitoring. The Agency has estimated that this provision would require an annual sampling of 10 percent of the affected workers.

OSHA estimates that an industrial hygienist (IH) would spend 1 day each year to sample 2 workers, for a per worker IH fee of $257. This exposure monitoring requires that three samples be taken per worker: One TWA and two STEL for an annual IH fee per sample of $86. Based on the 2000 EMSL Laboratory Testing Catalog (ERG, 2007b), OSHA estimated that analysis of each sample would cost $137 in lab fees. When combined with the IH fee, OSHA estimated the annual cost to obtain a TWA sample to be $223 per sampled worker and the annual cost to obtain the two STEL samples to be $445 per sampled worker. The direct exposure monitoring unit costs are summarized in Table V-16 in Chapter V of the PEA.

The cost of the sample also incorporates a productivity loss due to the additional time for the worker to participate in the sampling (30 minutes per worker sampled) as well as for the associated recordkeeping time incurred by a manager (15 minutes per worker sampled). The STEL samples are assumed to be taken along with the TWA sample and, thus, labor costs were not added to both unit costs. Including the costs related to lost productivity, OSHA estimates the total annual cost of a TWA sample to be $251, and 2 STEL samples, $445. The total annual cost per worker for all sampling taken is then $696. OSHA estimates the total annualized cost of this provision to be $2,208,950 for all affected industries. The annualized cost of this provision for each affected NAICS industry is shown in Table IX-6.

b. Beryllium Work Areas and Regulated Areas

The proposed beryllium standard requires the employer to establish and maintain a regulated area wherever employees are, or can reasonably expected to be, exposed to airborne beryllium at levels above the TWA PEL or STEL. Regulated areas require specific provisions that both limit employee exposure within its boundaries and curb the migration of beryllium outside the area. The Agency judged, based on the preliminary findings of the technological feasibility analysis, that companies can reduce establishment-wide exposure by ensuring that only authorized employees wearing proper protective equipment have access to areas of the establishment where such higher concentrations of beryllium exist, or can be reasonably expected to exist. Workers in other parts of the establishment are also likely to see a reduction in beryllium exposures due to these measures since fewer employees would be traveling through regulated areas and subsequently carrying beryllium residue to other work areas on their clothes and shoes.

Requirements in the proposed rule for a regulated area include: Demarcating the boundaries of the regulated area as separate from the rest of the workplace, limiting access to the regulated area, providing an appropriate respirator to each person entering the regulated area and other protective clothing and equipment as required by paragraph (g) and paragraph (h), respectively.

OSHA estimated that the total annualized cost per regulated area, including set-up costs ($76), respirators ($1,768) and protective clothing ($4,500), is $6,344.

When establishments are in full compliance with the standard, regulated areas would be required only for those workers for whom controls could not feasibly reduce their exposures to or below the 0.2 μg/m³ TWA PEL and the 2 μg/m³ STEL. Based on the findings of the technological feasibility analysis, OSHA estimated that 10 percent of the affected workers would be exposed above the TWA PEL or STEL after implementation of engineering controls and thus would require regulated areas (with one regulated area, on average, for every four workers exposed above the proposed TWA PEL or STEL).

The proposed standard requires that all beryllium work areas are adequately established and demarcated. ERG estimated that one work area would need to be established for every 12 at-risk workers. OSHA estimates that the annualized cost would be $33 per work area.

OSHA estimates the total annualized cost of the regulated areas and work areas is $629,031 for all affected industries. The cost for each affected application group and NAICS code is shown in Table IX-6.

c. Written Exposure Control Plan

The proposed standard requires that employers must establish and maintain a written exposure control plan for beryllium work areas. The written program must contain:

1. An inventory of operations and job titles reasonably expected to have exposure.

2. An inventory of operations and job titles reasonably expected to have exposure at or above the action level.

3. An inventory of operations and job titles reasonably expected to have exposure above the TWA PEL or STEL.

4. Procedures for minimizing cross-contamination, including but not limited to preventing the transfer of beryllium between surfaces, equipment, clothing, materials and articles within beryllium work areas.

5. Procedures for keeping surfaces in the beryllium work area free as practicable of beryllium.

6. Procedures for minimizing the migration of beryllium from beryllium work areas to other locations within or outside the workplace.

7. An inventory of engineering and work practice controls required by paragraph (f)(2) of this standard.

8. Procedures for removal, laundering, storage, cleaning, repairing, and disposal of beryllium-contaminated personal protective clothing and equipment, including respirators.

The unit cost estimates take into account the judgment that (1) most establishments have an awareness of beryllium risks and, thus, should be able to develop or modify existing safeguards in an expeditious fashion, and (2) many operations have limited beryllium activities and these establishments need to make only modest changes in procedures to create the necessary exposure control plan. ERG's experts estimated that managers would spend eight hours per establishment to develop and implement such a written exposure control plan, yielding a total cost per establishment to develop and implement the written control plan of $563.53 and an annualized cost of $66. In addition, because larger firms with more affected workers will need to develop more complicated written control plans, the development of a plan would require an extra thirty minutes of a manager's time per affected employee, for a cost of $35 per affected employee and an annualized cost of $4 per employee. Managers would also need 12 minutes (0.2 hours) per affected employee per quarter, or 48 minutes per affected employee per year to review and update the plan, for a recurring cost of $56 per affected employee per year to maintain and update the plan. Five minutes of clerical time would also be needed per employee for providing each employee with a copy of the written exposure control plan—yielding an annualized cost of $2 per employee. The total annual per-employee cost for development, implementation, review, and update of a written exposure control plan is then $62. The Agency estimates the total annualized cost of this provision to be $1,769,506 for all affected establishments. The breakdown of these costs by application group and NAICS code is presented in Table IX-6.

d. Personal Protective Clothing and Equipment

The proposed standard requires personal protective clothing and equipment for workers:

1. Whose exposure can reasonably be expected to exceed the TWA PEL or STEL.

2. When work clothing or skin may become visibly contaminated with beryllium, including during maintenance and repair activities or during non-routine tasks.

3. Where employees' skin can reasonably be expected to be exposed to soluble beryllium compounds.

OSHA has determined that it would be necessary for employers to provide reusable overalls and/or lab coats at a cost of $284 and $86, respectively, for operations in the following application groups:

Beryllium Production

Beryllium Oxide, Ceramics & Composites

Nonferrous Foundries

Fabrication of Beryllium Alloy Products

Copper Rolling, Drawing & Extruding

Secondary Smelting, Refining and Alloying

Precision Turned Products

Dental Laboratories

Chemical process operators in the spring and stamping application group would require chemical resistant protective clothing at an annual cost of $849. Gloves and/or shoe covers would be required when performing operations in several different application groups, depending on the process being performed, at an annual cost of $50 and $78, respectively.

The proposed standard requires that all reusable protective clothing and equipment be cleaned, laundered, repaired, and replaced as needed to maintain their effectiveness. This includes such safeguards as transporting contaminated clothing in sealed and labeled impermeable bags and informing any third party businesses coming in contact with such materials of the risks associated with beryllium exposure. OSHA estimates that the lowest cost alternative to satisfy this provision is for an employer to rent and launder reusable protective clothing—at an estimated annual cost per employee of $49. Ten minutes of clerical time would also be needed per establishment with laundry needs to notify the cleaners in writing of the potentially harmful effects of beryllium exposure and how the protective clothing and equipment must be handled in accordance with this standard—at a per establishment cost of $3.

The Agency estimates the total annualized cost of this provision to be $1,407,365 for all affected establishments. The breakdown of these costs by application group and NAICS code is shown in Table IX-6.

e. Hygiene Areas and Practices

The proposed standard requires employers to provide readily accessible washing facilities to remove beryllium from the hands, face, and neck of each employee working in a beryllium work area and also to provide a designated change room in workplaces where employees would have to remove their personal clothing and don the employer-provided protective clothing. The proposed standard also requires that employees shower at the end of the work shift or work activity if the employee reasonably could have been exposed to beryllium at levels above the PEL or STEL, and if those exposures could reasonably be expected to have caused contamination of the employee's hair or body parts other than hands, face, and neck.

In addition to other forms of PPE costed previously, for processes where hair may become contaminated, head coverings can be purchased at an annual cost of $28 per employee. This could satisfy the requirement to avoid contaminated hair. If workers are covered by protective clothing such that no body parts (including their hair where necessary, but not including their hands, face, and neck) could reasonably be expected to have been contaminated by beryllium, and they could not reasonably be expected to be exposed to beryllium while removing their protective clothing, they would not need to shower at the end of a work shift or work activity. OSHA notes that some facilities already have showers, and the Agency judges that all employers either already have showers where needed or will have sufficient measures in place to ensure that employees could not reasonably be expected to be exposed to beryllium while removing protective clothing. Therefore, OSHA has preliminarily determined that employers will not need to provide any new shower facilities to comply with the standard.

The Agency estimated the costs for the addition of a change room and segregated lockers based on the costs for acquisition of portable structures. The change room is presumed to be used in providing a transition zone from general working areas into beryllium-using regulated areas. OSHA estimated that portable building, adequate for 10 workers per establishment can be rented annually for $3,251, and that lockers could be procured for a capital cost of $407—or $48 annualized—per establishment. This results in an annualized cost of $3,299 per facility to rent a portable change room with lockers. OSHA estimates that the 10 percent of affected establishments unable to meet the proposed TWA PEL would require change rooms. The Agency estimated that a worker using a change room would need 2 minutes per day to change clothes. Assuming 250 days per year, this annual time cost for changing clothes is $185 per employee.

The Agency estimates the total annualized cost of the provision on hygiene areas and practices to be $389,241 for all affected establishments. The breakdown of these costs by application group and NAICS code can be seen in Table IX-6.

f. Housekeeping

The proposed rule specifies requirements for cleaning and disposing of beryllium-contaminated wastes. The employer shall maintain all surfaces in beryllium work areas as free as practicable of accumulations of beryllium and shall ensure that all spills and emergency releases of beryllium are cleaned up promptly, in accordance with the employer's written exposure control plan and using a HEPA-filtered vacuum or other methods that minimize the likelihood and level of exposure. The employer shall not allow dry sweeping or brushing for cleaning surfaces in beryllium work areas unless HEPA-filtered vacuuming or other methods that minimize the likelihood and level of exposure have been tried and were not effective.

ERG's experts estimated that each facility would need to purchase a single vacuum at a cost of $2,900 for every five affected employees in order to successfully integrate housekeeping into their daily routine. The per-employee cost would be $580, resulting in an annualized cost of $68 per worker. ERG's experts also estimated that all affected workers would require an additional five minutes per work day (.083 hours) to complete vacuuming tasks and to label and dispose of beryllium-contaminated waste. While this allotment is modest, OSHA judged that the steady application of this incremental additional cleaning, when combined with currently conducted cleaning, would be sufficient in average establishments to address dust or surface contamination hazards. Assuming that these affected workers would be working 250 days per year, OSHA estimates that the annual labor cost per employee for additional time spent cleaning in order to comply with this provision is $462.

The proposed standard requires each disposal bag with contaminated materials to be properly labeled. ERG estimated a cost of 10 cents per label with one label needed per day for every five workers. With the disposal of one labeled bag each day and 250 working days, the per-employee annual cost would be $5. The annualized cost of a HEPA-filtered vacuum, combined with the additional time needed to perform housekeeping and the labeling of disposal bags, results in a total annualized cost of $535 per employee.

The Agency estimates the total annualized cost of this provision to be $12,574,921 for all affected establishments. The breakdown of these costs by application group and NAICS code is shown in Table IX-6.

g. Medical Surveillance

The proposed standard requires the employer to make medical surveillance available at no cost to the employee, and at a reasonable time and place, for the following employees:

1. Employees who have worked in a regulated area for more than 30 days in the last 12 months

2. Employees showing signs or symptoms of chronic beryllium disease (CBD)

3. Employees exposed to beryllium during an emergency; and

4. Employees exposed to airborne beryllium above 0.2 μg/m³ for more than 30 days in a 12-month period for 5 years or more.

As discussed in the regulated areas section of this analysis of program costs, the Agency estimates that approximately 10 percent of affected employees would have exposure in excess of the PEL after the standard goes into effect and would therefore be placed in regulated areas. The Agency further estimates that a very small number of employees will be affected by emergencies in a given year, likely less than 0.1 percent of the affected population, representing a small additional cost. The number of workers who would suffer signs and symptoms of CBD after the rule takes effect is difficult to estimate, but would likely substantially exceed those with actual cases of CBD.

While the symptoms of CBD vary greatly, the first to appear are usually chronic dry cough (generally defined as a nonproductive cough, without phlegm or sputum, lasting two months or more) and shortness of breath during exertion. Ideally, in developing these costs estimates, OSHA would first estimate the percent of affected workers who might be presenting with a chronic cough and/or experiencing shortness of breath.

Studies have found the prevalence of a chronic cough ranging from 10 to 38 percent across various community populations, with smoking accounting for up to 18 percent of cough prevalence (Irwin, 1990; Barbee, 1991). However, these studies are over 20 years old, and the number of smokers has decreased substantially since then. It's also not clear whether the various segments of the U.S. population studied are similar enough to the population of workers exposed to beryllium such that results of these studies could be generalized to the affected worker population.

A more recent study from a plant in Cullman, Alabama that works with beryllium alloy found that about five percent of employees said they were current smokers, with roughly 52 percent saying they were previous smokers and approximately 43 percent stating they had never smoked (Newman et al., 2001). This study does not, however, report on the prevalence of chronic cough in this workplace.

OSHA was unable to identify any studies on the general prevalence of the other common early symptom of CBD, shortness of breath. Lacking any better data to base an estimate on, the Agency used the studies cited above (Irwin, 1990; Barbee, 1991) showing the prevalence of chronic cough in the general population, adjusted to account for the long term decrease in smoking prevalence (and hence, the amount of overall cases of chronic cough), and estimated that 15 percent of the worker population with beryllium exposure would exhibit a chronic cough or other sign or symptom of CBD that would trigger medical surveillance. The Agency welcomes comment and further data on this question.

According to the proposed rule, the initial (baseline) medical examination would consist of the following:

1. A medical and work history, with emphasis on past and present exposure, smoking history and any history of respiratory system dysfunction;

2. A physical examination with emphasis on the respiratory tract;

3. A physical examination for skin breaks and wounds;

4. A pulmonary function test;

5. A standardized beryllium lymphocyte proliferation test (BeLPT) upon the first examination and within every two years from the date of the first examination until the employee is confirmed positive for beryllium sensitization;

6. A CT scan, offered every two years for the duration of the employee's employment, if the employee was exposed to airborne beryllium at levels above 0.2 μg/m³ for more than 30 days in a 12-month period for at least 5 years. This obligation begins on the start-up date of this standard, or on the 15th year after the employee's first exposure above for more than 30 days in a 12-month period, whichever is later; and

7. Any other test deemed appropriate by the Physician or other Licensed Health Care Professional (PLHCP).

Table V-17 in Chapter V of the PEA lists the direct unit costs for initial medical surveillance activities including: Work and medical history, physical examination, pulmonary function test, BeLPT, CT scan, and costs of additional tests. In OSHA's cost model, all of the activities will take place during an employee's initial visit and on an annual basis thereafter and involve a single set of travel costs, except that: (1) The BeLPT tests will only be performed at two-year intervals after the initial test, but will be conducted in conjunction with the annual general examination (no additional travel costs); and (2) the CT scans will typically involve different specialists and are therefore treated as separate visits not encompassed by the general exams (therefore requiring separate travel costs). Not all employees would require CT scans, and employers would only be required to offer them every other year.

In addition to the fees for the annual medical exam, employers may also incur costs for lost work time when their employees are unavailable to perform their jobs. This includes time for traveling, a health history review, the physical exam, and the pulmonary function test. Each examination would require 15 minutes (or 0.25 hours) of a human resource manager's time for recording the results of the exam and tests and the PLHCP's written opinion for each employee and any necessary post-exam consultation with the employee. There is also a cost of 15 minutes of supervisor time to provide information to the physician, five minutes of supervisor time to process a licensed physician's written medical opinion, and five minutes for an employee to receive a licensed physician's written medical opinion. The total unit annual cost for the medical examinations and tests, excluding the BeLPT test, and the time required for both the employee and the supervisor is $297.

The estimated fee for the BeLPT is $259. With the addition of the time incurred by the worker to undergo the test, the total cost for a BeLPT is $261. The standard requires a biennial BeLPT for each employee covered by the medical surveillance provision, so most workers would receive between two and five BeLPT tests over a ten year period (including the BeLPT performed during the initial examination), depending on whether the results of these tests were positive. OSHA therefore estimates a net present value (NPV) of $1,417 for all five tests. This NPV annualized over a ten year period is $166.

Together, the annualized net present value of the BeLPT and the annualized cost of the remaining medical surveillance produce an annual cost of $436 per employee.

The proposed standard requires that a helical tomography (CT scan) be offered to employees exposed to airborne beryllium above 0.2 μg/m³ for more than 30 days in a 12-month period, for a period of 5 years or more. The five years do not need to be consecutive, and the exposure does not need to occur after the effective date of the standard. The CT scan shall be offered every 2 years starting on the 15th year after the first year the employee was exposed above 0.2 μg/m³ for more than 30 days in a 12-month period, for the duration of their employment. The total yearly cost for biennial CT scans consists of medical costs totaling $1,020, comprised of a $770 fee for the scan and the cost of a specialist to review the results, which OSHA estimates would cost $250. The Agency estimates an additional cost of $110 for lost work time, for a total of $1,131. The annualized yearly cost for biennial CT scans is $574.

Based on OSHA's estimates explained earlier in this section, all workers in regulated areas, workers exposed in emergencies, and an estimated 15 percent of workers not in regulated areas who exhibit signs and symptoms of CBD will be eligible for medical surveillance other than CT scans. The estimate for the number of workers eligible to receive CT scans is 25 percent of workers who are exposed above 0.2 in the exposure profile. The estimate of 25 percent is based on the facts that roughly this percentage of workers have 15-plus years of job tenure in the durable manufacturing sector and the estimate that all those with 15-plus years of job tenure and current exposure over 0.2 would have had at least 5 years of such exposure in the past.

The costs estimated for this provision are likely to be significantly overestimated, since not all affected employees offered medical surveillance would necessarily accept the offer. At Department of Energy facilities, only about 50 percent of eligible employees participate in the voluntary medical surveillance tests, and a report on an initial medical surveillance program at four aluminum manufacture facilities found participation rates to be around 57 percent (Taiwo et al., 2008). Where employers already offer equivalent health surveillance screening, no new costs are attributable to the proposed standard.

Within 30 days after an employer learns that an employee has been confirmed positive for beryllium sensitization, the employer's designated licensed physician shall consult with the employee to discuss referral to a CBD diagnostic center that is mutually agreed upon by the employer and the employee. If, after this consultation, the employee wishes to obtain a clinical evaluation at a CBD diagnostic center, the employer must provide the evaluation at no cost to the employee. OSHA estimates this consultation will take 15 minutes, with an estimated total cost of $33.

Table V-18 in Chapter V of the PEA lists the direct unit costs for a clinical evaluation with a specialist at a CBD diagnostic center. To estimate these costs, ERG contacted a healthcare provider who commonly treats patients with beryllium-related disease, and asked them to provide both the typical tests given and associated costs of an initial examination for a patient with a positive BeLPT test, presented in Table V-18 in Chapter V of the PEA. Their typical evaluation includes bronchoscopy with lung biopsy, a pulmonary stress test, and a chest CAT scan. The total cost for the entire suite of tests is $6,305.

In addition, there are costs for lost productivity and travel. The Agency has estimated the clinical evaluation would take three days of paid time for the worker to travel to and from one of two locations: Penn Lung Center at the Cleveland Clinic Foundation in Cleveland, Ohio or National Jewish Medical Center in Denver, Colorado. OSHA estimates lost work time is 24 hours, yielding total cost for the 3 days of $532.

OSHA estimates that roundtrip air-fare would be available for most facilities at $400, and the cost of a hotel room would be approximately $100 per night, for a total cost of $200 for the hotel room. OSHA estimates a per diem cost of $50 for three days, for a total of $150. The total cost per trip for traveling expenses is therefore $750.

The total cost of a clinical evaluation with a specialist at a CBD diagnostic center is equal to the cost of the examination plus the cost of lost work-time and the cost for the employee to travel to the CBD diagnostic center, or $7,620.

Based on the data from the exposure profile and the prevalence of beryllium sensitization observed at various levels of cumulative exposure, OSHA estimated the number of workers eligible for BeLPT testing (4,181) and the percentage of workers who will be confirmed positive for sensitization (two positive BeLPT tests, as specified in the proposed standard) and referred to a CBD diagnostic center. During the first year that the medical surveillance provisions are in effect, OSHA estimates that 9.4 percent of the workers who are tested for beryllium sensitization will be identified as sensitized. This percentage is an average based on: (1) The number of employees in the baseline exposure profile that are in a given cumulative exposure range; (2) the expected prevalence for a given cumulative exposure range (from Table VI-6 in Section VI of the preamble); and (3) an assumed even distribution of employees by cumulative years of exposure at a given level—20 percent having exposures at a given level for 5 years, 20 percent for 15 years, 20 percent for 25 years, 20 percent for 30 years, and 20 percent for 40 years.

OSHA did not assume that all workers with confirmed sensitization would choose to undergo evaluation at a CBD diagnostic center, which may involve invasive procedures and/or travel. For purposes of this cost analysis, OSHA estimates that approximately two-thirds of workers who are confirmed positive for beryllium sensitization will choose to undergo evaluation for CBD. OSHA requests comment on the CBD evaluation participation rate. OSHA estimates that about 264 of all non-dental lab workers will go to a diagnostic center for CBD evaluation in the first year.

The calculation method described above applies to all workers except dental technicians, who were analyzed with one modification. The rates for dental technicians are calculated differently due to the estimated 75 percent beryllium-substitution rate at dental labs, where the 75 percent of labs that eliminate all beryllium use are those at higher exposure levels. None of the remaining labs affected by this standard had exposures above 0.1 μg/m³. For the dental labs, the same calculation was done as presented in the previous paragraph, but only the remaining 25 percent of employees (2,314) who would still face beryllium exposures were included in the baseline cumulative exposure profile. With that one change, and all other elements of the calculation the same, OSHA estimates that 9 percent of dental lab workers tested for beryllium sensitization will be identified as sensitized. The predicted prevalence of sensitization among those dental lab workers tested in the first year after the standard takes effect is slightly lower than the predicted prevalence among all other tested workers combined. This slightly lower rate is not surprising because only dental lab workers with exposures below 0.1 μg/m³ are included (after adjusting for substitution), and OSHA's exposure profile indicates that the vast majority of non-dental workers exposed to beryllium are also exposed at 0.1 μg/m³ or lower. OSHA estimates that 20 dental lab workers (out of 347 tested for sensitization) would go to a diagnostic center for CBD evaluation in the first year.

In each year after the first year, OSHA relied on a 10 percent worker turnover rate in a steady state (as discussed in Chapter VII of the PEA) to estimate that the annual sensitization incidence rate is 10 percent of the first year's incidence rate. Based on that rate and the number of workers in the medical surveillance program, the CBD evaluation rate for workers other than those in dental labs would drop to 0.63 percent (.063 × .10). The evaluation rate for dental labs technicians is similarly estimated to drop to 0.58 percent (.058 × .10).

Based on these unit costs and the number of employees requiring medical surveillance estimated above, OSHA estimates that the medical surveillance and referral provisions would result in an annualized total cost of $2,882,706. These costs are presented by application group and NAICS code in Table IX-7.

h. Medical Removal Provision

Once a licensed physician diagnoses an employee with CBD or the employee is confirmed positive for sensitization to beryllium, that employee is eligible for medical removal and has two choices:

(a) Removal from current job, or

(b) Remain in a job with exposure above the action level while wearing a respirator pursuant to 29 CFR 1910.134.

To be eligible for removal, the employee must accept comparable work if such is available, but if not available the employer would be required to place the employee on paid leave for six months or until such time as comparable work becomes available, whichever comes first. During that six-month period, whether the employee is re-assigned or placed on paid leave, the employer must continue to maintain the employee's base earnings, seniority and other rights, and benefits that existed at the time of the first test.

For purposes of this analysis, OSHA has conservatively estimated the costs as if all employees will choose removal, rather than remaining in the current job while wearing a respirator. In practice, many workers may prefer to continue working at their current job while wearing a respirator, and the employer would only incur the respirator costs identified earlier in this chapter. The removal costs are significantly higher over the same six-month period, so this analysis likely overestimates the total costs for this provision.

OSHA estimated that the majority of firms would be able to reassign the worker to a job at least at the clerical level. The employer will often incur a cost for re-assigning the worker because this provision requires that, regardless of the comparable work the medically removed worker is performing, the employee must be paid the full base earnings for the previous position for six months. The cost per hour of reassigning a worker to a clerical job is based on the wage difference of a production worker of $22.16 and a clerical worker of $19.97, for a difference of $2.19. Over the six-month period, the incremental cost of reassigning a worker to a clerical position would be $2,190 per employee. This estimate is based on the employee remaining in a clerical position for the entire 6-month period, but the actual cost would be lower if there is turnover or if the employee is placed in any alternative position (for any part of the six-month period) that is compensated at a wage closer to the employee's previous wage.

Some firms may not have the ability to place the worker in an alternate job. If the employee chooses not to remain in the current position, the additional cost to the employer would be at most the cost of equipping that employee with a respirator, which would be required if the employee would continue to face exposures at or above the action level. Based on the earlier discussion of respirator costs, that option would be significantly cheaper than the alternative of providing the employee with six months of paid leave. Therefore, in order to estimate the maximum potential economic cost of the remaining alternatives, the Agency has conservatively estimated the cost per worker based on the cost of 6 months paid leave.

Using the wage rate of a production worker of $22.16 for 6 months (or 8 hours a day for 125 days), the total per-worker cost for this provision when a firm cannot place a worker in an alternate job is $22,161.

OSHA has estimated an average medical removal cost per worker assuming 75 percent of firms are able to find the employee an alternate job, and the remaining 25 percent of firms would not. The weighted average of these costs is $7,183. Based on these unit costs, OSHA estimates that the medical removal provision would result in an annualized total cost of $148,826. The breakdown of these costs by application group and NAICS code is shown in Table IX-6.

i. Training

As specified in the proposed standard and existing OSHA standard 29 CFR 1910.1200 on hazard communication, training is required for all employees where there is potential exposure to beryllium. In addition, newly hired employees would require training before starting work.

OSHA anticipates that training in accordance with the requirements of the proposed rule, which includes hazard communication training, would be conducted by in-house safety or supervisory staff with the use of training modules or videos. ERG estimated that this training would last, on average, eight hours. (Note that this estimate does not include the time taken for hazard communication training that is already required by 29 CFR 1910.1200.) The Agency judged that establishments could purchase sufficient training materials at an average cost of $2 per worker, encompassing the cost of handouts, video presentations, and training manuals and exercises. For initial and periodic training, ERG estimated an average class size of five workers with one instructor over an eight hour period. The per-worker cost of initial training totals to $239.

Annual retraining of workers is also required by the standard. OSHA estimates the same unit costs as for initial training, so retraining would require the same per-worker cost of $239.

Finally, to calculate training costs, the Agency needs the turnover rate of affected workers to know how many workers are receiving initial training versus retraining. Based on a 26.3 percent new hire rate in manufacturing, OSHA calculated a total net present value (NPV) of ten years of initial and annual retraining of $2,101 per employee. Annualizing this NPV gives a total annual cost for training of $246.

Based on these unit costs, OSHA estimates that the training requirements in the standard would result in an annualized total cost of $5,797,535. The breakdown of these costs by application group and NAICS code is presented in Table IX-6.

Total Annualized Cost

As shown in Table IX-7, the total annualized cost of the proposed rule is estimated to be about $37.6 million. As shown, at $27.8 million, the program costs represent about 74 percent of the total annualized costs of the proposed rule. The annualized cost of complying with the PEL accounts for the remaining 26 percent, almost all of which is for engineering controls and work practices. Respiratory protection, at about $237,600, represents only 3 percent of the annualized cost of complying with the PEL and less than 1 percent of the annualized cost of the proposed rule.

F. Economic Feasibility Analysis and Regulatory Flexibility Determination

Chapter VI of the PEA, summarized here, investigates the economic impacts of the proposed beryllium rule on affected employers. This impact investigation has two overriding objectives: (1) To establish whether the proposed rule is economically feasible for all affected application groups/industries, and (2) to determine if the Agency can certify that the proposed rule will not have a significant economic impact on a substantial number of small entities.

In the discussion below, OSHA first presents its approach for achieving these objectives and next applies this approach to industries with affected employers. The Agency invites comment on any aspect of the methods, data, or preliminary findings presented here or in Chapter VI of the PEA.

1. Analytic Approach

a. Economic Feasibility

Section 6(b)(5) of the OSH Act directs the Secretary of Labor to set standards based on the available evidence where no employee, over his/her working life time, will suffer from material impairment of health or functional capacity, even if such employee has regular exposure to the hazard, “to the exent feasible” (29 U.S.C. 655(b)(5)). OSHA interpreted the phrase “to the extent feasible” to encompass economic feasibility and was supported in this view by the U.S. Court of Appeals for the D.C. Circuit, which has long held that OSHA standards would satisfy the economic feasibility criterion even if they imposed significant costs on regulated industries and forced some marginal firms out of business, so long as they did not cause massive economic dislocations within a particular industry or imperil the existence of that industry. Am. Iron and Steel Inst. v. OSHA, 939 F.2d 975, 980 (D.C. Cir. 1991); United Steelworkers of Am., AFL-CIO-CLC v. Marshall, 647 F.2d 1189, 1265 (D.C. Cir. 1980); Indus. Union Dep't v. Hodgson, 499 F.2d 467 (D.C. Cir. 1974).

b. The Price Elasticity of Demand and Its Relationship to Economic Feasibility

In practice, the economic burden of an OSHA standard on an industry—and whether the standard is economically feasible for that industry—depends on the magnitude of compliance costs incurred by establishments in that industry and the extent to which they are able to pass those costs on to their customers. That, in turn, depends, to a significant degree, on the price elasticity of demand for the products sold by establishments in that industry.

The price elasticity of demand refers to the relationship between the price charged for a product and the demand for that product: The more elastic the relationship, the less an establishment's compliance costs can be passed through to customers in the form of a price increase and the more the establishment has to absorb compliance costs in the form of reduced profits. When demand is inelastic, establishments can recover most of the costs of compliance by raising the prices they charge; under this scenario, profit rates are largely unchanged and the industry remains largely unaffected. Any impacts are primarily on those customers using the relevant product. On the other hand, when demand is elastic, establishments cannot recover all compliance costs simply by passing the cost increase through in the form of a price increase; instead, they must absorb some of the increase from their profits. Commonly, this will mean reductions both in the quantity of goods and services produced and in total profits, though the profit rate may remain unchanged. In general, “[w]hen an industry is subjected to a higher cost, it does not simply swallow it; it raises its price and reduces its output, and in this way shifts a part of the cost to its consumers and a part to its suppliers,” in the words of the court in Am. Dental Ass'n v. Sec'y of Labor (984 F.2d 823, 829 (7th Cir. 1993)).

The court's summary is in accord with microeconomic theory. In the long run, firms can remain in business only if their profits are adequate to provide a return on investment that ensures that investment in the industry will continue. Over time, because of rising real incomes and productivity increases, firms in most industries are able to ensure an adequate profit. As technology and costs change, however, the long-run demand for some products naturally increases and the long-run demand for other products naturally decreases. In the face of additional compliance costs (or other external costs), firms that otherwise have a profitable line of business may have to increase prices to stay viable. Increases in prices typically result in reduced quantity demanded, but rarely eliminate all demand for the product. Whether this decrease in the total production of goods and services results in smaller output for each establishment within the industry or the closure of some plants within the industry, or a combination of the two, is dependent on the cost and profit structure of individual firms within the industry.

If demand is perfectly inelastic (i.e., the price elasticity of demand is zero), then the impact of compliance costs that are one percent of revenues for each firm in the industry would be a one percent increase in the price of the product, with no decline in quantity demanded. Such a situation represents an extreme case, but might be observed in situations in which there were few, if any, substitutes for the product in question, or if the products of the affected sector account for only a very small portion of the revenue or income of its customers.

If the demand is perfectly elastic (i.e., the price elasticity of demand is infinitely large), then no increase in price is possible and before-tax profits would be reduced by an amount equal to the costs of compliance (net of any cost savings—such as reduced workers' compensation insurance premiums—resulting from the proposed standard) if the industry attempted to maintain production at the same level as previously. Under this scenario, if the costs of compliance are such a large percentage of profits that some or all plants in the industry could no longer operate in the industry with hope of an adequate return on investment, then some or all of the firms in the industry would close. This scenario is highly unlikely to occur, however, because it can only arise when there are other products—unaffected by the proposed rule—that are, in the eyes of their customers, perfect substitutes for the products the affected establishments make.

A commonly-discussed intermediate case would be a price elasticity of demand of one (in absolute terms). In this situation, if the costs of compliance amount to one percent of revenues, then production would decline by one percent and prices would rise by one percent. As a result, industry revenues would remain the same, with somewhat lower production, but with similar profit rates per unit of output (in most situations where the marginal costs of production net of regulatory costs would fall as well). Customers would, however, receive less of the product for their (same) expenditures, and firms would have lower total profits; this, as the court described in Am. Dental Ass'n v. Sec'y of Labor, is the more typical case.

c. Variable Costs Versus Fixed Costs

A decline in output as a result of an increase in price may occur in a variety of ways: individual establishments could each reduce their levels of production; some marginal plants could close; or, in the case of an expanding industry, new entry may be delayed until demand equals supply. In some situations, there could be a combination of these three effects. Which possibility is most likely depends on the form that the costs of the regulation take. If the costs are variable costs (i.e., costs that vary with the level of production at a facility), then economic theory suggests that any reductions in overall output will be the result of reductions in output at each affected facility, with few, if any, plant closures. If, on the other hand, the costs of a regulation primarily take the form of fixed costs (i.e., costs that do not vary with the level of production at a facility), then reductions in overall output are more likely to take the form of plant closures or delays in new entry.

Most of the costs of this regulation, as estimated in Chapter V of the PEA, are variable costs in the sense that they will tend to vary by production levels and/or employment levels. Almost all of the major costs of program elements, such as medical surveillance and training, will vary in proportion to the number of employees (which is a rough proxy for the amount of production). Exposure monitoring costs will vary with the number of employees, but do have some economies of scale to the extent that a larger firm need only conduct representative sampling rather than sample every employee. Finally, the costs of operating and maintaining engineering controls tend to vary by usage—which typically closely tracks the level of production and are not fixed costs in the strictest sense.

This leaves two kinds of costs that are, in some sense, fixed costs—capital costs of engineering controls and certain initial costs. The capital costs of engineering controls due to the standard—many of which are scaled to production and/or employment levels—constitute a relatively small share of the total costs, representing 10 percent of total annualized costs (or approximately $870 per year per affected establishment).

Some ancillary provisions require initial costs that are fixed in the sense that they do not vary by production activity or the number of employees. Some examples are the costs to develop a training plan for general training not currently required and to develop a written exposure control plan.

As a result of these considerations, OSHA expects it to be quite likely that any reductions in total industry output would be due to reductions in output at each affected facility rather than as a result of plant closures. However, closures of some marginal plants or poorly performing facilities are always possible.

d. Economic Feasibility Screening Analysis

To determine whether a rule is economically feasible, OSHA begins with two screening tests to consider minimum threshold effects of the rule under two extreme cases: (1) All costs are passed through to customers in the form of higher prices (consistent with a price elasticity of demand of zero), and (2) all costs are absorbed by the firm in the form of reduced profits (consistent with an infinite price elasticity of demand).

In the former case, the immediate impact of the rule would be observed in increased industry revenues. While there is no hard and fast rule, in the absence of evidence to the contrary, OSHA generally considers a standard to be economically feasible for an industry when the annualized costs of compliance are less than a threshold level of one percent of annual revenues. Retrospective studies of previous OSHA regulations have shown that potential impacts of such a small magnitude are unlikely to eliminate an industry or significantly alter its competitive structure, particularly since most industries have at least some ability to raise prices to reflect increased costs, and normal price variations for products typically exceed three percent a year.

In the latter case, the immediate impact of the rule would be observed in reduced industry profits. OSHA uses the ratio of annualized costs to annual profits as a second check on economic feasibility. Again, while there is no hard and fast rule, in the absence of evidence to the contrary, OSHA generally considers a standard to be economically feasible for an industry when the annualized costs of compliance are less than a threshold level of ten percent of annual profits. In the context of economic feasibility, the Agency believes this threshold level to be fairly modest, given that normal year-to-year variations in profit rates in an industry can exceed 40 percent or more. OSHA also considered whether this threshold would be adequate to assure that upfront costs would not create major credit problems for affected employers. To do this, OSHA examined a worst case scenario in which annualized costs were ten percent of profits and all of the annualized costs were the result of upfront costs. In this scenario, assuming a three percent discount rate and a ten year life of equipment, total costs would be 85 percent of profits in the year in which these upfront costs were incurred. Because upfront costs would be less than one year's profits in the year they were incurred, this means that an employer could pay for all of these costs from that year's profits and would not necessarily have to incur any new borrowing. As a result, it is unlikely that these costs would create a credit crunch or other major credit problems. It would be true, however, that paying regulatory costs from profits might reduce investment from profits in that year. OSHA's choice of a threshold level of ten percent of annual profits is low enough that even if, in a hypothetical worst case, all compliance costs were upfront costs, then upfront costs—assuming a three percent discount rate and a ten-year time period—would be no more than 85 percent of first-year profits and thus would be affordable from profits without resort to credit markets. If the threshold level were first-year costs of ten percent of annual profits, firms could even more easily expect to cover first-year costs at the threshold level out of current profits without having to access capital markets and otherwise being threatened with short-term insolvency.

In general, because it is usually the case that firms would be able to pass on to their customers some or all of the costs of the proposed rule in the form of higher prices, OSHA will tend to give much more weight to the ratio of industry costs to industry revenues than to the ratio of industry costs to industry profits. However, if costs exceed either the threshold percentage of revenue or the threshold percentage of profits for an industry, or if there is other evidence of a threat to the viability of an industry because of the proposed standard, OSHA will examine the effect of the rule on that industry more closely. Such an examination would include market factors specific to the industry, such as normal variations in prices and profits, and any special circumstances, such as close domestic substitutes of equal cost, which might make the industry particularly vulnerable to a regulatory cost increase.

The preceding discussion focused on the economic viability of the affected industries in their entirety. However, even if OSHA found that a proposed standard did not threaten the survival of affected industries, there is still the question of whether the industries' competitive structure would be significantly altered. For example, if the annualized costs of an OSHA standard were equal to 10 percent of an industry's annual profits, and the price elasticity of demand for the products in that industry were equal to one, then OSHA would not expect the industry to go out of business. However, if the increase in costs were such that most or all small firms in that industry would have to close, it might reasonably be concluded that the competitive structure of the industry had been altered. For this reason, OSHA also calculates compliance costs by size of firm and conducts its economic feasibility screening analysis for small and very small entities.

e. Regulatory Flexibility Screening Analysis

The Regulatory Flexibility Act (RFA), Public Law 96-354, 94 Stat. 1164 (codified at 5 U.S.C. 601), requires Federal agencies to consider the economic impact that a proposed rulemaking will have on small entities. The RFA states that whenever a Federal agency is required to publish general notice of proposed rulemaking for any proposed rule, the agency must prepare and make available for public comment an initial regulatory flexibility analysis (IRFA). 5 U.S.C. 603(a). Pursuant to section 605(b), in lieu of an IRFA, the head of an agency may certify that the proposed rule will not have a significant economic impact on a substantial number of small entities. A certification must be supported by a factual basis. If the head of an agency makes a certification, the agency shall publish such certification in the Federal Register at the time of publication of general notice of proposed rulemaking or at the time of publication of the final rule. 5 U.S.C. 605(b).

To determine if the Assistant Secretary of Labor for OSHA can certify that the proposed beryllium rule will not have a significant economic impact on a substantial number of small entities, the Agency has developed screening tests to consider minimum threshold effects of the proposed rule on small entities. These screening tests do not constitute hard and fast rules and are similar in concept to those OSHA developed above to identify minimum threshold effects for purposes of demonstrating economic feasibility.

There are, however, two differences. First, for each affected industry, the screening tests are applied, not to all establishments, but to small entities (defined as “small business concerns” by SBA) and also to very small entities (as defined by OSHA as businesses with fewer than 20 employees). Second, although OSHA's regulatory flexibility screening test for revenues also uses a minimum threshold level of annualized costs equal to one percent of annual revenues, OSHA has established a minimum threshold level of annualized costs equal to five percent of annual profits for the average small entity or very small entity. The Agency has chosen a lower minimum threshold level for the profitability screening analysis and has applied its screening tests to both small entities and very small entities in order to ensure that certification will be made, and an IRFA will not be prepared, only if OSHA can be highly confident that a proposed rule will not have a significant economic impact on a substantial number of small entities or very small entities in any affected industry.

Furthermore, certification will not be made, and an IRFA will be prepared, if OSHA believes the proposed rule might otherwise have a significant economic impact on a substantial number of small entities, even if the minimum threshold levels are not exceeded for revenues or profitability for small entities or very small entities in all affected industries.

2. Impacts on Affected Industries

In this section, OSHA applies its screening criteria and other analytic methods, as needed, to determine (1) whether the proposed rule is economically feasible for all affected industries within the scope of this proposed rule, and (2) whether the Agency can certify that the proposed rule will not have a significant economic impact on a substantial number of small entities.

a. Economic Feasibility Screening Analysis: All Establishments

To determine whether the proposed rule's projected costs of compliance would threaten the economic viability of affected industries, OSHA first compared, for each affected industry, annualized compliance costs to annual revenues and profits per (average) affected establishment. The results for all affected establishments in all affected industries are presented in Table IX-8. Shown in the table for each affected industry are the total number of establishments, the total number of affected establishments, annualized costs per affected establishment, annual revenues per establishment, the profit rate, annual profits per establishment, annualized compliance costs as a percentage of annual revenues, and annualized compliance costs as a percentage of annual profits.

The annualized costs per affected establishment for each affected industry were calculated by distributing the industry-level (incremental) annualized compliance costs among all affected establishments in the industry, where annualized compliance costs reflect a 3 percent discount rate. The annualized cost of the proposed rule for the average affected establishment is estimated at $9,197 in 2010 dollars. It is clear from Table IX-8 that the estimates of the annualized costs per affected establishment vary widely from industry to industry. These estimates range from $1,257,214 for NAICS 331419 (Beryllium Production) and $120,372 for NAICS 327113a (Porcelain Electrical Supply Manufacturing (primary)) to $1,636 for NAICS 621210 (Offices of Dentists) and $1,632 for NAICS 339116 (Dental Laboratories).

As previously discussed, OSHA has established a minimum threshold level of annualized costs equal to one percent of annual revenues—and, secondarily, annualized costs equal to 10 percent of annual profits—below which the Agency has concluded that costs are unlikely to threaten the economic viability of an affected industry. The results of OSHA's threshold tests for all affected establishments are displayed in Table IX-8. For all affected establishments, the estimated annualized cost of the proposed rule is, on average, equal to 0.11 percent of annual revenue and 1.52 percent of annual profit.

As Table IX-8 shows, there are no industries in which the annualized costs of the proposed rule exceed one percent of annual revenues. However there are three six-digit NAICS industries where annualized costs exceed ten percent of annual profits.

NAICS 331525 (Copper foundries except die-casting) has the highest cost impact as a percentage of profits. NAICS 331525 is made up of two types of copper foundries: sand casting foundries and non-sand casting foundries, incurring an annualized cost as a percent of profit of 16.25 percent and 14.92 percent, respectively. The other two six-digit NAICS industries where annualized costs exceed ten percent of annual profits are NAICS 331534: Aluminum foundries (except die-casting), 13.65 percent; and NAICS 811310: Commercial and industrial machinery and equipment repair, 10.19 percent.

OSHA believes that the beryllium-containing inputs used by these industries have a relatively inelastic demand for three reasons. First, beryllium has rare and unique characteristics, including low mass, high melting temperature, dimensional stability over a wide temperature range, strength, stiffness, light weight, and high elasticity (“springiness”) that can significantly improve the performance of various alloys. These characteristics cannot easily be replicated by other materials. In economic terms, this means that the elasticity of substitution between beryllium and non-beryllium inputs will be low. Second, products which contain beryllium or beryllium-alloy components typically have high-performance applications (whose performance depends on the use of higher-cost beryllium). The lack of available competing products with these performance characteristics suggests that the price elasticity of demand for products containing beryllium or beryllium-alloy components will be low. Third, components made of beryllium or beryllium-containing alloys typically account for only a small portion of the overall cost of the finished goods that these parts are used to make. For example, the cost of brakes made of a beryllium-alloy used in the production of a jet airplane represents a trivial percentage of the overall cost to produce that airplane. As economic theory indicates, the elasticity of derived demand for a factor of production (such as beryllium) varies directly with the elasticity of substitution between the input in question and other inputs; the price elasticity of demand for the final product that the input is used to produce; and, in general, the share of the cost of the final product that the input accounts for. Applying these three conditions to beryllium points to the relative inelastic derived demand for this factor of production and the likelihood that cost increases resulting from the proposed rule would be passed on to the consumer in the form of higher prices.

A secondary point is that the establishments in an industry that use beryllium may be more profitable than those that don't. This follows from the prior arguments about beryllium's rare and desirable characteristics and its valuable applications. For example, of the 208 establishments that make up NAICS 331525, OSHA estimated that 45 establishments (or 21 percent) work with beryllium. Of the 394 establishments that make up NAICS 331524, OSHA estimated that only 7 establishments (less than 2 percent) work with beryllium. Of the 21,960 establishments that make up NAICS 811310, OSHA estimated that 143 (0.7 percent) work with beryllium. However, when OSHA calculated the cost-to-profit ratio, it used the average profit per firm for the entire NAICs industry, not the average profit per firm for firms working with beryllium.

(1) Normal Year-to-Year Variations in Prices and Profit Rates

The United States has a dynamic and constantly changing economy in which an annual percentage increase in industry revenues or prices of one percent or more are common. Examples of year-to-year changes in an industry that could cause such an increase in revenues or prices include increases in fuel, material, real estate, or other costs; tax increases; and shifts in demand.

To demonstrate the normal year-to-year variation in prices for all the manufacturers in general industry affected by the proposed rule, OSHA developed in the PEA year-to-year producer price indices and year-to-year percentage changes in producer prices, by industry, for the years 1999-2010. For all of the industries estimated to be affected by this proposed standard over the 12-year period, the average change in producer prices was 4.4 percent a year—which is over 4 times as high as OSHA's 1 percent cost-to-revenue threshold. For the industries found to have the largest estimated potential annual cost impact as a percentage of revenue shown in Chapter VI of the PEA are—NAICS 331524: Aluminum Foundries (except Die-Casting), (0.71 percent); NAICS 331525(a and b): Copper Foundries (except Die-Casting) (average of 0.81 percent); NAICS 332721a: Precision Turned Product Manufacturing of high content beryllium (0.49 percent); and NAICS 811310: Commercial and Industrial Machinery and Equipment (Except Automotive and Electronic) Repair and Maintenance (0.55 percent)—the average annual changes in producer prices in these industries over the 12-year period analyzed were 3.1 percent, 8.2 percent, 3.6 percent and 2.3 percent, respectively.

Based on these data, it is clear that the potential price impacts of the proposed rule in affected industries are all well within normal year-to-year variations in prices in those industries. The maximum cost impact of the proposed rule as a percentage of revenue in any affected industry is 0.84 percent, while, as just noted, the average annual change in producer prices for affected industries was 4.4 percent for the period 1999-2010. In fact, Chapter VI of the PEA shows two of the industries within the secondary smelting, refining, and alloying group, for example, the prices rose over 60 percent in one year without imperiling the existence of those industries. Thus, OSHA preliminarily concludes that the potential price impacts of the proposal would not threaten the economic viability of any industries affected by this proposed standard.

Profit rates are also subject to the dynamics of the U.S. economy. A recession, a downturn in a particular industry, foreign competition, or the increased competitiveness of producers of close domestic substitutes are all easily capable of causing a decline in profit rates in an industry of well in excess of ten percent in one year or for several years in succession.

To demonstrate the normal year-to-year variation in profit rates for all the manufacturers affected by the proposed rule, OSHA presented data in the PEA on year-to-year profit rates and year-to-year percentage changes in profit rates, by industry, for the years 2002-2009. For the industries that OSHA has estimated will be affected by this proposed standard over the 8-year period, the average change in profit rates is calculated to be 39 percent per year. For the industries with the largest estimated potential annual cost impacts as a percentage of profit—NAICS 331524: Aluminum foundries (except die-casting), (14 percent); NAICS 331525(a and b): Copper foundries (except die-casting) (16 percent); NAICS 332721a: Precision Turned Product Manufacturing of high content beryllium (8 percent); and NAICS 811310 Commercial and Industrial Machinery and Equipment (Except Automotive and Electronic) Repair and Maintenance (10 percent)—the average annual changes in profit rates in these industries over the eight-year period were 35 percent, 35 percent, 11 percent, and 5 percent, respectively.

A longer-term loss of profits in excess of 10 percent a year could be more problematic for some affected industries and might conceivably, under sufficiently adverse circumstances, threaten an industry's economic viability. However, as previously discussed, OSHA's analysis indicates that affected industries would generally not absorb the costs of the proposed rule in reduced profits but, instead, would be able to pass on most or all of those costs in the form of higher prices (due to the relative price inelasticity of demand for beryllium and beryllium-containing inputs). It is possible that such price increases will result in some reduction in output, and the reduction in output might be met through the closure of a small percentage of the plants in the industry. The only realistic circumstance where an entire industry would be significantly affected by small potential price increases would be where there is a very close or perfect substitute product available not subject to OSHA regulation. In most cases where beryllium is used, there is no substitute product that could be used in place of beryllium and achieve the same level of performance. The main potential concern would be substitution by foreign competition, but the following discussion reveals why such competition is not likely.

(2) International Trade Effects

World production of beryllium is a thin market, with only a handful of countries known to process beryllium ores and concentrates into beryllium products, and characterized by a high degree of variation and uncertainty. The United States accounts for approximately 65 percent of world beryllium deposits and 90 percent of world production, but there is also a significant stockpiling of beryllium materials in Kazakhstan, Russia, China, and possibly other countries (USGS, 2013a). For the individual years 2008-2012, the United States' net import reliance as a percentage of apparent consumption (that is, imports minus exports net of industry and government stock adjustments) ranged from 10 percent to 61 percent (USGS, 2013b). To assure an adequate stockpile of beryllium materials to support national defense interests, the U.S. Department of Defense, in 2005, under the Defense Production Act, Title III, invested in a public-private partnership with the leading U.S. beryllium producer to build a new $90.4 million primary beryllium facility in Elmore, Ohio. Construction of that facility was completed in 2011 (USGS, 2013b).

One factor of importance to firms working with beryllium and beryllium alloys is to have a reliable supply of beryllium materials. U.S. manufacturers can have a relatively high confidence in the availability of beryllium materials relative to manufacturers in many foreign countries, particularly those that do not have economic or national security partnerships with the United States.

Firms using beryllium in production must consider not just the cost of the chemical itself but also the various regulatory costs associated with the use, transport, and disposal of the material. For example, for marine transport, metallic beryllium powder and beryllium compounds are classified by the International Maritime Organization (IMO) as poisonous substances, presenting medical danger. Beryllium is also classified as flammable. The United Nations classification of beryllium and beryllium compounds for the transport of dangerous goods is “poisonous substance” and, for packing, a “substance presenting medium danger” (WHO, 1990). Because of beryllium's toxicity, the material is subject to various workplace restrictions as well as international, national, and State requirements and guidelines regarding beryllium content in environmental media (USGS, 2013a).

As the previous discussion indicates, the production and use of beryllium and beryllium alloys in the United States and foreign markets appears to depend on the availability of production facilities; beryllium stockpiles; national defense and political considerations; regulations limiting the shipping of beryllium and beryllium products; international, national, and State regulations and guidelines regarding beryllium content in environmental media; and, of course, the special performance properties of beryllium and beryllium alloys in various applications. Relatively small changes in the price of beryllium would seem to have a minor effect on the location of beryllium production and use. In particular, as a result of this proposed rule, OSHA would expect that, if all compliance costs were passed through in the form of higher prices, a price increase of 0.11 percent, on average, for firms manufacturing or using beryllium in the United States—and not exceeding 1 percent in any affected industry—would have a negligible effect on foreign competition and would therefore not threaten the economic viability of any affected domestic industries.

(b) Economic Feasibility Screening Analysis: Small and Very Small Businesses

The preceding discussion focused on the economic viability of the affected industries in their entirety. Even though OSHA found that the proposed standard did not threaten the survival of these industries, there is still the possibility that the competitive structure of these industries could be significantly altered such as by small entities exiting from the industry as a result of the proposed standard.

To address this possibility, OSHA examined the annualized costs of the proposed standard per affected small entity, and per affected very small entity, for each affected industry. Again, OSHA used a minimum threshold level of annualized compliance costs equal to one percent of annual revenues—and, secondarily, annualized compliance costs equal to ten percent of annual profits—below which the Agency has concluded that the costs are unlikely to threaten the survival of small entities or very small entities or, consequently, to alter the competitive structure of the affected industries.

Based on the results presented in Table IX-9, the annualized cost of compliance with the proposed rule for the average affected small entity is estimated to be $8,108 in 2010 dollars. Based on the results presented in Table IX-10, the annualized cost of compliance with the proposed rule for the average affected very small entity is estimated to be $1,955 in 2010 dollars. These tables also show that there are no industries in which the annualized costs of the proposed rule for small entities or very small entities exceed one percent of annual revenues. NAICS 331525b: Sand Copper Foundries (except die-casting) has the highest estimated cost impact as a percentage of revenues for small entities, 0.95 percent, and NAICS 336322b: Other motor vehicle electrical and electronic equipment has the highest estimated cost impact as a percentage of revenues for very small entities, 0.70 percent.

Small entities in four industries—NAICS 331525: Sand and non-sand foundries (except die-casting); NAICS 331524(a and b): Aluminum foundries (except die-casting); NAICS 811310: Commercial and Industrial Machinery and Equipment; and NAICS 331522: Nonferrous (except aluminum) die-casting foundries—have annualized costs in excess of 10 percent of annual profits (17.45 percent, 16.12 percent, 11.68 percent, and 10.64 percent, respectively). Very small entities in 7 industries are estimated to have annualized costs in excess of 10 percent of annual profit; NAICS 336322b: Other motor vehicle electrical and electronic equipment (38.49 percent); NAICS 336322a: Other motor vehicle electrical and electronic equipment, (18.18 percent); NAICS 327113: Porcelain electrical Supply Manufacturing (13.82 percent); NAICS 811310: Commercial and Industrial Machinery and Equipment (Except Automotive and Electronic) Repair and Maintenance (12.76 percent); NAICS 332721a: Precision turned product manufacturing (10.50 percent); NAICS 336214: Travel trailer and camper manufacturing (10.75 percent); and NAICS 336399: All other motor vehicle parts manufacturing (10.38 percent).

In general, cost impacts for affected small entities or very small entities will tend to be somewhat higher, on average, than the cost impacts for the average business in those affected industries. That is to be expected. After all, smaller businesses typically suffer from diseconomies of scale in many aspects of their business, leading to less revenue per dollar of cost and higher unit costs. Small businesses are able to overcome these obstacles by providing specialized products and services, offering local service and better service, or otherwise creating a market niche for themselves. The higher cost impacts for smaller businesses estimated for this rule—other than very small entities in NAICS 336322b: Other motor vehicle electrical and electronic equipment—generally fall within the range observed in other OSHA regulations and, as verified by OSHA's lookback reviews, have not been of such a magnitude to lead to the economic failure of regulated small businesses.

The ratio of annualized costs to annual profit is a sizable 38.49 percent in NAICS 336322b: Other motor vehicle electrical and electronic equipment. However, OSHA believes that the actual ratio is significantly lower. There are 386 very small entities in NAICS 336322, of which only 6, or 1.5 percent, are affected entities using beryllium. When OSHA calculated the cost-to-profit ratio, it used the average profit per firm for the entire NAICs industry, not the average profit rate for firms working with beryllium. The profit rate for all establishments in NAICS 336322b was estimated at 1.83 percent. If, for example, the average profit rate for a very small entity in NAICS 336322b were equal to 5.95 percent, the average profit rate for its application group, Beryllium Oxide Ceramics and Composites, then the ratio of the very small entity's annualized cost of the proposed rule to its annual profit would actually be 11.77 percent. OSHA tentatively concludes the 6 establishments in the NAICS specializing in beryllium production will have a higher than average profit rate and will be able to pass much of the cost onto the consumer for three main reasons: (1) The absence of substitutes containing the rare performance characteristics of beryllium; (2) the relative price insensitivity of (other) motor vehicles containing the special performance characteristics of beryllium and beryllium alloys; and (3) the fact that electrical and electronic components made of beryllium or beryllium-containing alloys typically account for only a small portion of the overall cost of the finished (other) motor vehicles. The annualized compliance cost to annual revenue ratio for NAICS 336332b is 0.70 percent, 0.30 percent below the 1 percent threshold. Based on OSHA's experience, price increases of this magnitude have not historically been associated with the economic failure of small businesses.

(c) Regulatory Flexibility Screening Analysis

To determine if the Assistant Secretary of Labor for OSHA can certify that the proposed beryllium standard will not have a significant economic impact on a substantial number of small entities, the Agency has developed screening tests to consider minimum threshold effects of the proposed standard on small entities. The minimum threshold effects for this purpose are annualized costs equal to one percent of annual revenues, and annualized costs equal to five percent of annual profits, applied to each affected industry. OSHA has applied these screening tests both to small entities and to very small entities. For purposes of certification, the threshold level cannot be exceeded for affected small entities or very small entities in any affected industry.

Tables IX-9 and Table IX-10, presented above, show that the annualized costs of the proposed standard do not exceed one percent of annual revenues for affected small entities or affected very small entities in any affected industry. These tables also show that the annualized costs of the proposed standard exceed five percent of annual profits for affected small entities in 12 industries and for affected very small entities in 30 industries. OSHA is therefore unable to certify that the proposed standard will not have a significant economic impact on a substantial number of small entities and must prepare an Initial Regulatory Flexibility Analysis (IRFA). The IRFA is presented in Chapter IX of the PEA and is reproduced in Section IX.I of this preamble.

G. Benefits and Net Benefits

In this section, OSHA presents a summary of the estimated benefits and net benefits of the proposed beryllium rule. This section proceeds in five steps. The first step estimates the numbers of diseases and deaths prevented by comparing the current (baseline) situation to a world in which the proposed PEL is adopted in a final standard to a world in which employees are exposed at the level of the proposed PEL throughout their working lives. The second step also assumes that the proposed PEL is adopted, but uses the results from the first step to estimate what would happen under a more realistic scenario in which employees have been exposed for varying periods of time to the baseline situation and will thereafter be exposed to the new PEL.

The third step covers the monetization of benefits. Then, in the fourth step, OSHA estimates the net benefits and incremental benefits of the proposed rule by comparing the monetized benefits to the costs presented in Chapter V of the PEA. The models underlying each step inevitably need to make a variety of assumptions based on limited data. In the fifth step, OSHA provides a sensitivity analysis to explore the robustness of the estimates of net benefits with respect to many of the assumptions made in developing and applying the underlying models. A full explanation of the derivation of the estimates presented here is provided in Chapter VII of the PEA for the proposed rule. OSHA invites comments on any aspect of the data and methods used to estimate the benefits and net benefits of this proposed rule. Because dental labs constitute a significant source of both costs and benefits to the rule (over 40 percent), OSHA is particularly interested in comments regarding the appropriateness of the model, assumptions, and data to estimating the benefits to workers in that industry.

OSHA has added to the docket the spreadsheets used to calculate the estimates of benefits outlined below (OSHA, 2015a). Those interested in exploring the details and methodology of OSHA's benefits analysis, such as how the life table referred to below was developed and applied, should consult those spreadsheets.

Step 1—Estimation of the Steady-State Number of Beryllium-Related Diseases Avoided

Methods of Estimation

The first step in OSHA's development of the benefits analysis compares the situation in which employees continue to be at baseline exposure levels for their entire working lives to the situation in which all employees have been exposed at a given PEL for their entire working lives. This is a comparison of two steady-state situations. To do this, OSHA must estimate both the risk associated with the baseline exposure levels and the risk following the promulgation of a new beryllium standard. OSHA's approach assumes for inputs such as the turnover rate and the exposure response function that they are similar across all workers exposed to beryllium, regardless of industry.

An exposure-response model, discussed below, is used to estimate a worker's risk of beryllium-related disease based on the worker's cumulative beryllium exposure. The Agency used a lifetime risk model to estimate the baseline risk and the associated number of cases for the various disease endpoints. A lifetime risk model explicitly follows a worker each year, from work commencement onwards, accumulating the worker's beryllium exposure in the workplace and estimating outcomes each year for the competing risks that can occur. To go from exposure to number of cases, the Agency needs to estimate an exposure-response relationship, and this is discussed below. The possible outcomes are no change, or the various health endpoints OSHA has considered (beryllium sensitization, CBD, lung cancer, and the mortality associated with these endpoints). As part of the estimation discussion, OSHA will mention specific parameters used in some of the estimation methods, but will further discuss how these parameters were derived later in this section.

The baseline lifetime risk model is the most complicated part of the analysis. The Agency only needs to make relatively simple adjustments to this model to reflect changes in activities and conditions due to the standard, which, working through the model, then lead to changes in relevant health outcomes. There are three channels by which the standard generates benefits. First are estimated benefits due to the lowering of the PEL. Second are estimated benefits with further exposure reductions from the substitution of non-beryllium for beryllium-containing materials, ending workers' beryllium exposures entirely. This potential source of benefits is particularly significant with respect to OSHA's assumptions for how dental labs are likely to reduce exposures (see below). Finally, the model estimates benefits due to the ancillary programs that are required by the proposed standard. The last channel affects CBD and sensitization, endpoints which may be mitigated or prevented with the help of ancillary provisions such as dermal protection and medical surveillance for early detection, and for which the Agency has some information on the effects on risk of ancillary provisions. The benefits of ancillary provisions are not estimated for lung cancer because the benefits from reducing lung cancer are considered to be the result of reducing airborne exposure only and thus the ancillary provisions will have no separable effect on airborne exposures. The discussion here will concentrate on CBD as being the most important and complex endpoint, and most illustrative of other endpoints: The structure for other endpoints is the same; only the exposure response functions are different. Here OSHA will discuss first the exposure-response model, then the structure of the year-to-year changes for a worker, then the estimated exposure distribution in the affected population and the risk model with the lowering of the PEL, and, last, the other adjustments for the ancillary benefits and the substitution benefits.

The exposure response model is designed to translate beryllium exposure to risk of adverse health endpoints. In the case of beryllium sensitization and CBD, the Agency uses the cumulative exposure data from a beryllium manufacturing facility. Specifically, OSHA uses the quartile data from the Cullman plant that is presented in Table VI-7 of the Preliminary Risk Assessment in the preamble. The raw data from this study show cases of CBD with cumulative exposures that would represent an average exposure level of less than 0.1 µg/m³ if exposed for 10 years; show cases of CBD with exposures lasting less than one year; and show cases of CBD with actual average exposure of less than 0.1 µg/m³.

Prevalence is defined as the percentage of persons with a condition in a population at a given point in time. The quartile data in Table VI-7 of the Preliminary Risk Assessment are prevalence percentages (the number of cases of illness documented over several years in the 319 person cohort from the Cullman plant) at different cumulative exposure levels. The Cullman data do not cover persons who left the work force or what happened to persons who remained in the workforce after the study was completed. For the lifetime risk model, the prevalence percentages will be translated into incidence percentages—the estimated number of new cases predicted to occur each year. For this purpose OSHA assumed that the incidence for any given cumulative exposure level is constant from year to year and continues after exposure ceases.

To calculate incidence from prevalence, OSHA assumed a steady state in which both the size of the beryllium-exposed affected population, exposure concentrations during employment and prevalence are constant over time. If these conditions are met, and turnover among workers with a condition is equal to turnover for workers without a condition, then the incidence rate will be equal to the turnover rate multiplied by the prevalence rate. If the turnover rate among persons with a condition is higher than the turnover rate for workers without the condition, then this assumption will underestimate incidence. This might happen if, in addition to other reasons for leaving work, persons with a condition leave a place of employment more frequently because their disabilities cause them to have difficulty continuing to do the work. If the turnover rate among persons with a condition is lower than the turnover rate for workers without the condition, then this assumption will overestimate incidence. This could happen if an employer provides special benefits to workers with the condition, and the employer would cease to provide these benefits if the employee left work.

To illustrate, if 10 percent of the work force (including 10 percent of those with the condition) leave each year and if the overall prevalence is at 20 percent, then a 2 percent (10 percent times 20 percent) incidence rate will be needed in order to keep a steady 20 percent group prevalence rate each year. OSHA's model assumes a constant 10 percent turnover rate (see later in this section for the rationale for this particular turnover rate). While turnover rates are not available for the specific set of employees in question, for manufacturing as a whole, the turnover rates are greater than 20 percent, and greater than 30 percent for the economy as a whole (BLS, 2013). For this analysis, OSHA assumed an effective turnover rate of 10 percent. Different turnover rates will result in different incidence rates. The lower the turnover rate the lower the estimated incidence rate. This is a conservative assumption for the industries where turnover rates may be higher. However, some occupations/industries, such as dental lab technicians, may have lower turnover rates than manufacturing workers. Additionally, the typical dental technician even if leaving one workplace, has significant likelihood of continuing to work as a dental technician and going to another workplace that uses beryllium. OSHA welcomes comments on its turnover estimates and on sectors, such as dental laboratories, where turnover may be lower than ten percent.

Using Table VI-7 of the Preliminary Risk Assessment, when a worker's cumulative exposure is below 0.147 (μg/m³-years), the prevalence of CBD is 2.5 percent and so the derived annual risk would be 0.25 percent (0.10 × 2.5 percent). It will stay at this level until the worker has reached a cumulative exposure of 1.468, where it will rise to 0.80 percent.

The model assumes a maximum 45-year (250 days per year) working life (ages 20 through 65 or age of death or onset of CBD, whichever is earlier) and follows workers after retirement through age 80. The 45-year working life is based on OSHA's legal requirements and is longer than the working lives of most exposed workers. A shorter working life will be examined later in this section. While employed, the worker accumulates beryllium exposure at a rate depending on where the worker is in the empirical exposure profile presented in Chapter IV of the PEA (i.e., OSHA calculates a general risk model which depends on the exposure level and then plug in our empirical exposure distribution to estimate the final number of cases of various health outcomes). Following a worker's retirement, there is no increased exposure, just a constant annual risk resulting from the worker's final cumulative exposure.

OSHA's model follows the population of workers each year, keeping track of cumulative exposure and various health outcomes. Explicitly, each year the model calculates: The increased cumulative exposure level for each worker versus last year, the incidence at the new exposure level, the survival rate for this age bracket, and the percentage of workers who have not previously developed CBD in earlier years.

For any individual year, the equation for predicting new cases of CBD for workers at age t is:

New CBD cases rate(t) = modeled incidence rate(t) * survival rate(t) * (1- currently have CBD rate(t)), where the variables used are:

New CBD cases rate(t) is the output variable to be calculated;

cumulative exposure(t) = cumulative exposure(t-1) + current exposure;

modeled incidence rate(t) is a function of cumulative exposure; and

survival rate(t) is the background survival rate from mortality due to other causes in the national population.

Then for the next year the model updates the survival rate (due to an increase in the worker's age), incidence rate (due to any increased cumulative exposure), and the rate of those currently having CBD, which increases due to the new CBD case rate of the year before. This process then repeats for all 60 years.

It is important to note that this model is based on the assumption that prevalence is explained by an underlying constant incidence, and as a result, prevalence will be different depending on the average number of years of exposure in the population examined and (though a sensitivity analysis is provided later) on the assumption of a maximum of 45 years of exposure. OSHA also examined (OSHA 2015c) a model in which prevalence is constant at the levels shown in Table VI-7 of the preliminary risk assessment, with a population age (and thus exposure) distribution estimated based on an assumed constant turnover rate. OSHA solicits comment on this and other alternative approaches to using the available prevalence data to develop an exposure-response function for this benefits analysis.

In the next step, OSHA uses its model to take into account the adoption of the lower proposed PEL. OSHA uses the exposure profile for workers as estimated in Chapter IV of the PEA for each of the various application groups. These exposure profiles estimate the number of workers at various exposure levels, specifically the ranges less than 0.1 μg/m³, 0.1 to 0.2, 0.2 to 0.5, 0.5 to 1.0, 1.0 to 2.0, and greater than 2.0 μg/m³. Translating these ranges into exposure levels for the risk model, the model assumes an average exposure equal to the midpoint of the range, except for the lower end, where it was assumed to be equal to 0.1 μg/m³, and the upper end, where it was assumed to be equal to 2.0 μg/m³.

The model increases the workers' cumulative exposure each year by these midpoints and then plugs these new values into the new case equation. This alters the incidence rate as cumulative exposure crosses a threshold of the quartile data. So then using the exposure profiles by application group from Chapter IV of the PEA, the baseline exposure flows through the life time risk model to give us a baseline number of cases. Next OSHA calculated the number of cases estimated to occur after the implementation of the proposed PEL of 0.2 μg/m³. Here OSHA simply takes the number of workers with current average exposure above 0.2 μg/m³ and set their exposure level at 0.2 μg/m³; all exposures for workers exposed below 0.2 μg/m³ stay the same. After adjusting the worker exposure profile in this way, OSHA goes through all the same calculations and obtains a post-standard number of CBD cases. Subtracting estimated post-standard CBD cases from estimated pre-standard CBD cases gives us the number of CBD cases that would be averted due to the proposed change in the PEL.

Based on these methods, OSHA's estimate of benefits associated with the proposed rule does not include benefits associated with current compliance that have already been achieved with regard to the new requirements, or benefits obtained from future compliance with existing beryllium requirements. However, available exposure data indicate that few employees are currently exposed above the existing standard's PEL of 2.0 μg/m³. To achieve consistency with the cost estimation method in chapter V, all employees in the exposure profile that are above 2.0 μg/m³ are assumed to be at the 2.0 μg/m³ level.

There is also a component that applies only to dental labs. OSHA has preliminarily assumed, based on the estimates of higher costs for engineering controls than using substitutes presented in the cost chapter, that rather than incur the costs of compliance with the proposed standard, many dental labs are likely to stop using beryllium-containing materials after the promulgation of the proposed standard. OSHA estimated earlier in this PEA that, for the baseline, only 25 percent of dental lab workers still work with beryllium. OSHA estimates that, if OSHA adopts the proposed rule, 75 percent of the 25 percent still using beryllium will stop working with beryllium; their beryllium exposure level will therefore drop to zero. OSHA estimates that the 75 percent of workers will not be a random sample of the dental lab exposure profile but instead will concentrate among workers who are currently at the highest exposure levels because it would cost more to reduce those higher exposures into compliance with the proposed PEL. Under this judgment OSHA is estimating that the rule would eliminate all cases of CBD in the 75 percent of dental lab workers with the highest exposure levels. As discussed in the sensitivity analysis below, dental labs constitute a significant source of both costs and benefits to the rule (over 40 percent), and the extent to which dental laboratories substitute other materials for beryllium has significant effects on the benefits and costs of the rule. To derive its baseline estimate of cases of CBD in dental laboratories, OSHA (1) estimated baseline cases of CBD using the existing rate of beryllium use in dental labs without a projection of further substitution; (2) estimated cases of CBD with the proposed regulation using an estimate that 75 percent of the dental labs with higher exposure would switch to other materials and thus eliminate exposure to beryllium; and (3) estimated that the turnover rate in the industry is 10 percent. OSHA welcomes comments on all aspects of the analysis of substitution away from beryllium in the dental laboratories sector.

Estimation results for both dental labs and non-dental workplaces appear in the table below.

In contrast to this PEL component of the benefits, both the ancillary program benefits calculation and the substitution benefits calculation are relatively simple. Both are percentages of the lifetime-risk-model CBD cases that still occur in the post-standard world. OSHA notes that in the context of existing CBD prevention programs, some ancillary-provision programs similar to those included in OSHA's proposal have eliminated a significant percentage of the remaining CBD cases (discussed later in this chapter). If the ancillary provisions reduce remaining CBD cases by 90 percent for example, and if the estimated baseline contains 120 cases of CBD, and post-standard compliance with a lower PEL reduces the total to 100 cases of CBD, then 90 of those remaining 100 cases of CBD would be averted due to the ancillary programs.

OSHA assumed, based on the clinical experience discussed further below, that approximately 65 percent of CBD cases ultimately result in death. Later in this chapter, OSHA provides a sensitivity analysis of the effects of different values for assuming this percentage at 50 percent and 80 percent on the number of CBD deaths prevented. OSHA welcomes comment on this assumption. OSHA's exposure-response model for lung cancer is based on lung cancer mortality data. Thus, all of the estimated cases of lung cancer in the benefits analysis are cases of premature death from beryllium-related lung cancer.

Finally, in recognition of the uncertainty in this aspect of these models, OSHA presents a “high” estimate, a “low” estimate, and uses the midpoint of these two as our “primary” estimate. The low estimate is simply those CBD fatalities prevented due to everything except the ancillary provisions, i.e., both the reduction in the PEL and the substitution by dental labs. The high estimate includes both of these factors plus all the ancillary benefits calculated at an effectiveness rate of 90 percent in preventing cases of CBD not averted by the reduction of the PEL. The midpoint is the combination of reductions attributed to adopting the proposed PEL, substitution by dental labs, and the ancillary provisions calculated at an effectiveness rate of only 45 percent.

a. Chronic Beryllium Disease

CBD is a respiratory disease in which the body's immune system reacts to the presence of beryllium in the lung, causing a progression of pathological changes including chronic inflammation and tissue scarring. Immunological sensitization to beryllium (BeS) is a precursor that occurs before early-stage CBD. Only sensitized individuals can go on to develop CBD. In early, asymptomatic stages of CBD, small granulomatous lesions and mild inflammation occur in the lungs. As CBD progresses, the capacity and function of the lungs decrease, which eventually affects other organs and bodily functions as well. Over time the spread of lung fibrosis (scarring) and loss of pulmonary function cause symptoms such as: A persistent dry cough, shortness of breath, fatigue, night sweats, chest and join pain, clubbing of fingers due to impaired oxygen exchange, and loss of appetite. In these later stages CBD can also impair the liver, spleen, and kidneys, and cause health effects such as granulomas of the skin and lymph nodes, and cor pulmonale (enlargement of the heart). The speed and extent of disease progression may be influenced by the level and duration of exposure, treatment with corticosteroids, and genetics, but these effects are not fully understood.

Corticosteroid therapy, in workers whose beryllium exposure has ceased, has been shown to control inflammation, ease symptoms, and in some cases prevent the development of fibrosis. However, corticosteroid use can have adverse effects, including increased risk of infections; accelerated bone loss or osteoporosis; psychiatric effects such as depression, sleep disturbances, and psychosis; adrenal suppression; ocular effects; glucose intolerance; excessive weight gain; increased risk of cardiovascular disease; and poor wound healing. The effects of CBD, and of common treatments for CBD, are discussed in detail in this preamble at Section V, Health Effects, and Section VIII, Significance of Risk.

OSHA's review of the literature on CBD suggests three broad types of CBD progression (see this preamble at Section V, Health Effects). In the first, individuals progress relatively directly toward death related to CBD. They suffer rapidly advancing disability and their death is significantly premature. Medical intervention is not applied, or if it is, does little to slow the progression of disease. In the second type, individuals live with CBD for an extended period of time. The progression of CBD in these individuals is naturally slow, or may be medically stabilized. They may suffer significant disability, in terms of loss of lung function—and quality of life—and require medical oversight their remaining years. They would be expected to lose some years of normal lifespan. As discussed previously, advanced CBD can involve organs and systems beyond the respiratory system; thus, CBD can contribute to premature death from other causes. Finally, individuals with the third type of CBD progression do not die prematurely from causes related to CBD. The disease is stabilized and may never progress to a debilitating state. These individuals nevertheless may experience some disability or loss of lung function, as well as side effects from medical treatment, and may be affected by the disease in many areas of their lives: Work, recreation, family, etc.

In the analysis that follows, OSHA assumes, based on the clinical experience discussed below, that 35 percent of workers who develop CBD experience the third type of progression and do not die prematurely from CBD. The remaining 65 percent were estimated to die prematurely, whether from rapid disease progression (type 1) or slow (type 2). Although the proportion of CBD patients who die prematurely as a result of the disease is not well understood or documented at this time, OSHA believes this assumption is consistent with the information submitted in response to the RFI. Newman et al. (2003) presented a scenario for what they considered to be the “typical” CBD patient:

We have included an example of a life care plan for a typical clinical case of CBD. In this example, the hypothetical case is diagnosed at age 40 and assumed to live an additional 33.7 years (approximately 5% reduced life expectancy in this model). In this hypothetical example, this individual would be considered to have moderate severity of chronic beryllium disease at the time of initial diagnosis. They require treatment with prednisone and treatment for early cor pulmonale secondary to CBD. They have experienced some, but not all, of the side effects of treatment and only the most common CBD-related health effects.

In short, most workers diagnosed with CBD are expected to have shortened life expectancy, even if they do not progress rapidly and directly to death. It should be emphasized that this represents the Agency's best estimate of the mortality related to CBD based upon the current available evidence. As described in Section V, Health Effects, there is a substantial degree of uncertainty as to the prognosis for those contracting CBD, particularly as the relatively less severe cases are likely not to be studied closely for the remainder of their lives.

As mentioned previously, OSHA used the Cullman data set for empirical estimates of beryllium sensitization and CBD prevalence in its exposure response model, which translates beryllium exposure to risk of adverse health endpoints for the purpose of determining the benefits that could be achieved by preventing those adverse health endpoints.

OSHA chose the cumulative exposure quartile data as the basis for this benefits analysis. The choice of cumulative quartiles was based in part on the need to use the cumulative exposure forecast developed in the model, and in part on the fact that in statistically fitted models for CBD, the cumulative exposure tended to fit the CBD data better than other exposure variables. OSHA also chose the quartile model because the outside expert who examined the logistic and proportional hazards models believed statistical modeling of the data set to be unreliable due to its small size. In addition, the proportional hazards model with its dummy variables by year of detection is difficult to interpret for purposes of this section. Of course regression analyses are often useful in empirical analysis. They can be a useful compact representation of a set of data, allow investigations of various variable interactions and possible causal relationships, have added flexibility due to covariate transformations, and under certain conditions can be shown to be statistically “optimal.” However, they are only useful when used in the proper setting. The possibility of misspecification of functional form, endogeneity, or incorrect distributional assumptions are just three reasons to be cautious about using regression analyses.

On the other hand, the use of results produced by a quartile analysis as inputs in a benefits assessment implies that the analytic results are being interpreted as evidence of an exposure-response causal relationship. Regression analysis is a more sophisticated approach to estimating causal relationships (or even correlations) than quartile or other quantile analysis, and any data limitations that may apply to a particular regression-based exposure-response estimation also apply to exposure-response estimation conducted with a quartile analysis using the same data set. In this case, OSHA adopted the quartile analysis because the logistic regression analysis yielded extremely high prevalence rates for higher level of exposure over long time periods that some might not find credible. Use of the quartile analysis serves to show that there are significant benefits even without using an extremely high estimate of prevalence for long periods of exposure at high levels. As a check on the quartile model, the Agency performed the same benefits calculation using the logit model estimated by the Agency's outside expert, and these benefit results are presented in a separate OSHA background document (OSHA, 2015b). The difference in benefits between the two models is slight, and there is no qualitative change in final outcomes. The Agency solicits comment on these issues.

(1) Number of CBD Cases Prevented by the Proposed PEL

To examine the effect of simply changing the PEL, including the effect of the standard on some dental labs to discontinue their use of beryllium, OSHA compared the number of CBD-related deaths (mortality) and cases of non-fatal CBD (morbidity) that would occur if workers were exposed for a 45-year working life to PELs of 0.1, 0.2, or 0.5 μg/m³ to the number of cases that would occur at levels of exposure at or below the current PEL. The number of avoided cases over a hypothetical working life of exposure for the current population at a lower PEL is then equal to the difference between the number of cases at levels of exposure at or below the current PEL for that population minus the number of cases at the lower PEL. This approach represents a steady-state comparison based on what would hypothetically happen to workers who received a specific average level of occupational exposure to beryllium during an entire working life. (Chapter VII in the PEA modifies this approach by introducing a model that takes into account the timing of benefits before steady state is reached.)

As indicated in Table IX-11, the Agency estimates that there would be 16,240 cases of beryllium sensitization, from which there would be 11,017, or about 70 percent, progressing to CBD. The Agency arrived at these estimates by using the CBD and BeS prevalence values from the Agency's preliminary risk analysis, the exposure profile at current exposure levels (under an assumption of full, or fixed, compliance with the existing beryllium PEL), and the model outlined in the previous methods of estimation section after a working lifetime of exposure. Applying the prior midpoint estimate, as explained above, that 65 percent of CBD cases cause or contribute to premature death, the Agency predicts a total of 7,161 cases of mortality and 3,856 cases of morbidity from exposure at current levels; this translates, annually, to 165 cases of mortality and 86 cases of morbidity. At the proposed PEL, OSHA's base model estimates that, due to the airborne factor only, a total of 2,563 CBD cases would be avoided from exposure at current levels, including 1,666 cases of mortality and 897 cases of morbidity—or an average of 37 cases of mortality and 20 cases of morbidity annually. OSHA has not estimated the quantitative benefits of sensitization cases avoided.

OSHA requests comment on this analysis, including feedback on the data relied on and the approach and assumptions used. As discussed earlier, based on information submitted in response to the RFI, the Agency estimates that most of the workers with CBD will progress to an early death, even if it comes after retirement, and has quantified those cases prevented. However, given the evolving nature of science and medicine, the Agency invites public comment on the current state of CBD-related mortality.

The proposed standard also includes provisions for medical surveillance and removal. The Agency believes that to the extent the proposal provides medical surveillance sooner and to more workers than would have been the case in the absence of the proposed standard, workers will be more likely to receive appropriate treatment and, where necessary, removal from beryllium exposure. These interventions may lessen the severity of beryllium-related illnesses, and possibly prevent premature death. The Agency requests public comment on this issue.

(2) CBD Cases Prevented by the Ancillary Provisions of the Proposed Standard

The nature of the chronic beryllium disease process should be emphasized. As discussed in this preamble at Section V, Heath Effects, the chronic beryllium disease process involves two steps. First, workers become sensitized to beryllium. In most epidemiological studies of CBD conducted to date, a large percentage of sensitized workers have progressed to CBD. A certain percentage of the population has an elevated risk of this occurring, even at very low exposure levels, and sensitization can occur from dermal as well as inhalation exposure to beryllium. For this reason, the threat of beryllium sensitization and CBD persist to a substantial degree, even at very low levels of airborne beryllium exposure. It is therefore desirable not only to significantly reduce airborne beryllium exposure, but to avoid nearly any source of beryllium exposure, so as to prevent beryllium sensitization.

The analysis presented above accounted only for CBD-prevention benefits associated with the proposed reduction of the PEL, from 2 ug/m³ to 0.2 ug/m³. However, the proposed standard also includes a variety of ancillary provisions—including requirements for respiratory protection, personal protective equipment (PPE), housekeeping procedures, hygiene areas, medical surveillance, medical removal, and training—that the Agency believes would further reduce workers' risk of disease from beryllium exposure. These provisions were described in Chapter I of the PEA and discussed extensively in Section XVIII of this preamble, Summary and Explanation of the Proposed Standard.

The leading manufacturer of beryllium in the U.S., Materion Corporation (Materion), has implemented programs including these types of provisions in several of its plants and has worked with NIOSH to publish peer-reviewed studies of their effectiveness in reducing workers' risk of sensitization and CBD. The Agency used the results of these studies to estimate the health benefits associated with a comprehensive standard for beryllium.

The best available evidence on comprehensive beryllium programs comes from studies of programs introduced at Materion plants in Reading, PA; Tucson, AZ; and Elmore, OH. These studies are discussed in detail in this preamble at Section VI, Preliminary Risk Assessment, and Section VIII, Significance of Risk. All three facilities were in compliance with the current PEL prior to instituting comprehensive programs, and had taken steps to reduce airborne levels of beryllium below the PEL, but their medical surveillance programs continued to identify cases of sensitization and CBD among their workers. Beginning around 2000, these facilities introduced comprehensive beryllium programs that used a combination of engineering controls, dermal and respiratory PPE, and stringent housekeeping measures to reduce workers' dermal exposures and airborne exposures. These comprehensive beryllium programs have substantially lowered the risk of sensitization among workers. At the times that studies of the programs were published, insufficient follow-up time had elapsed to report directly on the results for CBD. However, since only sensitized workers can develop CBD, reduction of sensitization risk necessarily reduces CBD risk as well.

In the Reading, PA copper beryllium plant, full-shift airborne exposures in all jobs were reduced to a median of 0.1 ug/m³ or below, and dermal protection was required for production-area workers, beginning in 2000-2001 (Thomas et al., 2009). In 2002, the process with the highest exposures (with a median of 0.1 ug/m³) was enclosed, and workers involved in that process were required to use respiratory protection. Among 45 workers hired after the enclosure was built and respiratory protection instituted, one was found to be sensitized (2.2 percent). This is more than an 80 percent reduction in sensitization from a previous group of 43 workers hired after 1992, 11.5 percent of whom had been sensitized by the time of testing in 2000.

In the Tucson beryllium ceramics plant, respiratory and skin protection was instituted for all workers in production areas in 2000 (Cummings et al., 2007). BeLPT testing in 2000-2004 showed that only 1 (1 percent) of 97 workers hired during that time period was sensitized to beryllium. This is a 90 percent reduction from the prevalence of sensitization in a 1998 BeLPT screening, which found that 6 (9 percent) of 69 workers hired after 1992 were sensitized.

In the Elmore, OH beryllium production and processing facility, all new workers were required to wear loose-fitting powered air-purifying respirators (PAPRs) in manufacturing buildings, beginning in 1999 (Bailey et al., 2010). Skin protection became part of the protection program for new workers in 2000, and glove use was required in production areas and for handling work boots, beginning in 2001. Bailey et al. (2010) found that 23 (8.9 percent) of 258 workers hired between 1993 and 1999, before institution of respiratory and dermal protection, were sensitized to beryllium. The prevalence of sensitization among the 290 workers who were hired after the respiratory protection and PPE measures were put in place was about 2 percent, close to an 80 percent reduction in beryllium sensitization.

In a response to OSHA's 2002 Request for Information (RFI), Lee Newman et al. from National Jewish Medical and Research Center (NJMRC) summarized results of beryllium program effectiveness from several sources. Said Dr. Newman (in response to Question #33):

Q. 33. What are the potential impacts of reducing occupational exposures to beryllium in terms of costs of controls, costs for training, benefits from reduction in the number or severity of illnesses, effects on revenue and profit, changes in worker productivity, or any other impact measures than you can identify?

A: From experience in [the Tucson, AZ facility discussed above], one can infer that approximately 90 percent of beryllium sensitization can be eliminated. Furthermore, the preliminary data would suggest that potentially 100 percent of CBD can be eliminated with appropriate workplace control measures.

In a study by Kelleher 2001, Martyny 2000, Newman, JOEM 2001) in a plant that previously had rates of sensitization as high as 9.7 percent, the data suggests that when lifetime weighted average exposures were below 0.02 μg per cu meter that the rate of sensitization fell to zero and the rate of CBD fell to zero as well.

In an unpublished study, we have been conducting serial surveillance including testing new hires in a precision machining shop that handles beryllium and beryllium alloys in the Southeast United States. At the time of the first screening with the blood BeLPT of people tested within the first year of hire, we had a rate of 6.7 percent (4/60) sensitization and with 50 percent of these individuals showing CBD at the time of initial clinical evaluation. At that time, the median exposures in the machining areas of the plant was 0.47 μg per cu meter. Subsequently, efforts were made to reduce exposures, further educate the workforce, and increase monitoring of exposure in the plant. Ongoing testing of newly hired workers within the first year of hire demonstrated an incremental decline in the rate of sensitization and in the rate of CBD. For example, at the time of most recent testing when the median airborne exposures in the machining shop were 0.13 μg per cu meter, the percentage of newly hired workers found to have beryllium sensitization or CBD was now 0 percent (0/55). Notably, we also saw an incremental decline in the percentage of longer term workers being detected with sensitization and disease across this time period of exposure reduction and improved hygiene practices.

Thus, in calculating the potential economic benefit, it's reasonable to work with the assumption that with appropriate efforts to control exposures in the work place, rates of sensitization can be reduced by over 90 percent. (NJMRC, RFI Ex. 6-20)

OSHA has reviewed these papers and is in agreement with Dr. Newman's testimony. OSHA judges Dr. Newman's estimate to be an upper bound of the effectiveness of ancillary programs and examined the results of using Dr. Newman's estimate that beryllium ancillary programs can reduce BeS by 90 percent, and potentially eliminate CBD where sensitization is reduced, because CBD can only occur where there is sensitization. OSHA applied this 90 percent reduction factor to all cases of CBD remaining after application of the reductions due to lowering the PEL alone. OSHA applied this reduction broadly because the proposed standard would require housekeeping and PPE related to skin exposure (18,000 of 28,000 employees will need PPE because of possible skin exposure) to apply to all or most employees likely to come in contact with beryllium and not just those with exposure above the action level. Table IX-11 shows that there are 11,017 baseline cases of CBD and that the proposed PEL of 0.2 µg/m³ would prevent 2,563 cases through airborne prevention alone. The remaining number of cases of CBD is then 8,454 (11,017 minus 2,563). If OSHA applies the full ninety percent reduction factor to account for prevention of skin exposure (“non-airborne” protections), then 7,609 (90 percent of 8,454 cases) additional cases of CBD would be prevented.

The Agency recognizes that there are significant differences between the comprehensive programs discussed above and the proposed standard. While the proposed standard includes many of the same elements, it is generally less stringent. For example, the proposed standard's requirements for respiratory protection and PPE are narrower, and many provisions of the standard apply only to workers exposed above the proposed TWA PEL or STEL. However, many provisions, such as housekeeping and beryllium work areas, apply to all employers covered by the proposed standard. To account for these differences, OSHA has provided a range of benefits estimates (shown in Table IX-11), first, assuming that there are no ancillary provisions to the standard, and, second, assuming that the comprehensive standard achieves the full 90-percent reduction in risk documented in existing programs. The Agency is taking the midpoint of these two numbers as its main estimate of the benefits of avoided CBD due to the ancillary provisions of the proposed standard. The results in Table IX-11 suggest that approximately 60 percent of the beryllium sensitization cases and the CBD cases avoided would be attributable to the ancillary provisions of the standard. OSHA solicits comment on all aspects of this approach to analyzing ancillary provisions and solicits additional data that might serve to make more accurate estimates of the effects of ancillary provisions. OSHA is interested in the extent of the effects of ancillary provisions and whether these apply to all exposed employees or only those exposed above or below a given exposure level.

(3) Morbidity Only Cases

As previously indicated, the Agency does not believe that all CBD cases will ultimately result in premature death. While currently strong empirical data on this are lacking, the Agency estimates that approximately 35 percent of cases would not ultimately be fatal, but would result in some pain and suffering related to having CBD, and possible side effects from steroid treatment, as well as the dread of not knowing whether the disease will ultimately lead to premature death. These would be described as “mild” cases of CBD relative to the others. These are the residual cases of CBD after cases with premature mortality have been counted. As indicated in Table IX-11, the Agency estimates the standard will prevent 2,228 such cases (midpoint) over 45 years, or an estimated 50 cases annually.

b. Lung Cancer

In addition to the Agency's determinations with respect to the risk of chronic beryllium disease, the Agency has preliminarily determined that chronic beryllium exposure at the current PEL can lead to a significantly elevated risk of (fatal) lung cancer. OSHA used the estimation methodology outlined at the beginning of this section. However, unlike with chronic beryllium disease, the underlying data were based on incidence of lung cancer and thus there was no need to address the possible limitations of prevalence data. The Agency also used lifetime excess risk estimates of lung cancer mortality, presented in Table VI-20 in Section VI of this preamble, Preliminary Risk Assessment, to estimate the benefits of avoided lung cancer mortality. The lung cancer risk estimates are derived from one of the best-fitting models in a recent, high-quality NIOSH lung cancer study, and are based on average exposure levels. The estimates of excess lifetime risk of lung cancer were taken from the line in Table VI-20 in the risk assessment labeled PWL (piecewise log-linear) not including professional and asbestos workers. This model avoids possible confounding from asbestos exposure and reduces the potential for confounding due to smoking, as smoking rates and beryllium exposures can be correlated via professional worker status. Of the three estimates in the NIOSH study that excluded professional workers and those with asbestos exposure, this model was chosen because it was at the midpoint of risk results.

Table IX-11 shows the number of avoided fatal lung cancers for PELs of 0.2 μg/m³, 0.1 μg/m³, and 0.5 μg/m³. At the proposed PEL of 0.2 μg/m³, an estimated 180 lung cancers would be prevented over the lifetime of the current worker population. This is the equivalent of 4.0 cases avoided annually, given a 45-year working life of exposure.

Combining the two major fatal health endpoints—for lung cancer and CBD-related mortality—OSHA estimates that the proposed PEL would prevent between 1,846 and 6,791 premature fatalities over the lifetime of the current worker population, with a midpoint estimate of 4,318 fatalities prevented. This is the equivalent of between 41 and 151 premature fatalities avoided annually, with a midpoint estimate of 96 premature fatalities avoided annually, given a 45-year working life of exposure.

Note that the Agency based its estimates of reductions in the number of beryllium-related diseases over a working life of constant exposure for workers who are employed in a beryllium-exposed occupation for their entire working lives, from ages 20 to 65. In other words, workers are assumed not to enter or exit jobs with beryllium exposure mid-career or to switch to other exposure groups during their working lives. While the Agency is legally obligated to examine the effect of exposures from a working lifetime of exposure and set its standard accordingly, in an alternative analysis purely for informational purposes, using the same underlying risk model for CBD, the Agency examined, in Chapter VII of the PEA, the effect of assuming that workers are exposed for a maximum of only 25 working years, as opposed to the 45 years assumed in the main analysis. While all workers are assumed to have less cumulative exposure under the 25-years-of-exposure assumption, the effective exposed population over time is proportionately increased.

A comparison of exposures over a maximum of 25 working years versus over a potentially 45-year working life shows variations in the number of estimated prevented cases by health outcome. For chronic beryllium disease, there is a substantial increase in the number of estimated baseline and prevented cases if one assumes that the typical maximum exposure period is 25 years, as opposed to 45. This reflects the relatively flat CBD risk function within the relevant exposure range, given varying levels of airborne beryllium exposure—shortening the average tenure and increasing the exposed population over time translates into larger total numbers of people sensitized to beryllium. This, in turn, results in larger populations of individuals contracting CBD. Since the lung cancer model itself is based on average, as opposed to cumulative, exposure, it is not adaptable to estimate exposures over a shorter period of time. As a practical matter, however, over 90 percent of illness and mortality attributable to beryllium exposure in this analysis comes from CBD.

Overall, the 45-year-maximum-working-life assumption yields smaller estimates of the number of cases of avoided fatalities and illnesses than does the maximum-25-years-of-exposure assumption. For example, the midpoint estimates of the number of avoided fatalities and illnesses related to CBD under the proposed PEL of 0.2 μg/m increases from 92 and 50, respectively, under the maximum-45-year-working-life assumption to 145 and 78, respectively, under the maximum-25-year-working-life assumption—or approximately a 57 to 58 percent increase.

Step 2—Estimating the Stream of Benefits Over Time

Risk assessments in the occupational environment are generally designed to estimate the risk of an occupationally related illness over the course of an individual worker's lifetime. As demonstrated previously in this section, the current occupational exposure profile for a particular substance for the current cohort of workers can be matched up against the expected profile after the proposed standard takes effect, creating a “steady state” estimate of benefits. However, in order to annualize the benefits for the period of time after the beryllium rule takes effect, it is necessary to create a timeline of benefits for an entire active workforce over that period.

While there are various approaches that could be taken for modeling the workforce, there seem to be two polar extremes. At one extreme, one could assume that none of the benefits occur until after the worker retires, or at least 45 years in the future. In the case of lung cancer, that period would effectively be at least 55 years, since the 45 years of exposure must be added to a 10-year latency period during which it is assumed that lung cancer does not develop. At the other extreme, one could assume that the benefits occur immediately, or at least immediately after a designated lag. However, based on the various risk models discussed in this preamble at Section VI, Risk Assessment, which reflect real-world experience with development of disease over an extended period of time, it appears that the actual pattern occurs at some point between these two extremes.

At first glance, the simplest intermediate approach would be to follow the pattern of the risk assessments, which are based in part on life tables, and observe that typically the risk of the illness grows gradually over the course of a working life and into retirement. Thus, the older the person exposed to beryllium, the higher the odds that that person will have developed the disease.

However, while this is a good working model for an individual exposed over a working life, it is not very descriptive of the effect of lowering exposures for an entire working population. In the latter case, in order to estimate the benefits of the standard over time, one has to consider that workers currently being exposed to beryllium are going to vary considerably in age. Since the calculated health risks from beryllium exposure depend on a worker's cumulative exposure over a working lifetime, the overall benefits of the proposed standard will phase in over several decades, as the cumulative exposure gradually falls for all age groups, until those now entering the workforce reach retirement and the annual stream of beryllium-related illnesses reaches a new, significantly lowered “steady state.” That said, the near-term impact of the proposed rule estimated for those workers with similar current levels of cumulative exposure will be greater for workers who are now middle-aged or older. This conclusion follows in part from the structure of the relative risk model used for lung cancer in this analysis and the fact that the background mortality rates for lung cancer increase with age.

In order to characterize the magnitude of benefits before the steady state is reached, OSHA created a linear phase-in model to reflect the potential timing of benefits. Specifically, OSHA estimated that, for all non-cancer cases, while the number of cases of beryllium-related disease would gradually decline as a result of the proposed rule, they would not reach the steady-state level until 45 years had passed. The reduction in cases estimated to occur in any given year in the future was estimated to be equal to the steady-state reduction (the number of cases in the baseline minus the number of cases in the new steady state) times the ratio of the number of years since the standard was implemented and a working life of 45 years. Expressed mathematically:

N_t = (C−S) × (t/45),

Where N_t is the number of non-malignant beryllium-related diseases avoided in year t; C is the current annual number of non-malignant beryllium-related diseases; S is the steady-state annual number of non-malignant beryllium-related diseases; and t represents the number of years after the proposed standard takes effect, with t ≤ 45.

In the case of lung cancer, the function representing the decline in the number of beryllium-related cases as a result of the proposed rule is similar, but there would be a 10-year lag before any reduction in cancer cases would be achieved. Expressed mathematically, for lung cancer:

L_t = (C_m−S_m) × ((t−10)/45)),

Where 10 ≤ t ≤ 55 and L_t is the number of lung cancer cases avoided in year t as a result of the proposed rule; C_m is the current annual number of beryllium-related lung cancers; and S_m is the steady-state annual number of beryllium-related lung cancers.

This model was extended to 60 years for all the health effects previously discussed in order to incorporate the 10-year lag, in the case of lung cancer, and a maximum-45-year working life, as well as to capture some occupationally-related disease that manifests itself after retirement. As a practical matter, however, there is no overriding reason for stopping the benefits analysis at 60 years. An internal analysis by OSHA indicated that, both in terms of cases prevented, and even with regard to monetized benefits, particularly when lower discount rates are used, the estimated benefits of the standard are larger on an annualized basis if the analysis extends further into the future. The Agency welcomes comment on the merit of extending the benefits analysis beyond the 60-years analyzed in the PEA.

In order to compare costs to benefits, OSHA assumes that economic conditions remain constant and that annualized costs—and the underlying costs—will repeat for the entire 60-year time horizon used for the benefits analysis (as discussed in Chapter V of the PEA). OSHA welcomes comments on the assumption for both the benefit and cost analysis that economic conditions remain constant for sixty years. OSHA is particularly interested in what assumptions and time horizon should be used instead and why.

Separating the Timing of Mortality

In previous sections, OSHA modeled the timing and incidence of morbidity. OSHA's benefit estimates are based on an underlying CBD-related mortality rate of 65 percent. However, this mortality is not simultaneous with the onset of morbidity. Although mortality from CBD has not been well studied, OSHA believes, based on discussions with experienced clinicians, that the average lag for a larger population has a range of 10 to 30 years between morbidity and mortality. The Agency's review of Workers Compensation data related to beryllium exposure from the Office of Worker Compensation Programs (OWCP) Division of Energy Employees Occupational Illness Compensation is consistent with this range. Hence, for the purposes of this proposal, OSHA estimates that mortality occurs on average 20 years after the onset of CBD morbidity. Thus, for example, the prevented deaths that would have occurred in year 21 after the promulgation of the rule are associated with the CBD morbidity cases prevented in year one. OSHA requests comment on this estimate and range.

The Agency invites comment on each of these elements of the analysis, particularly on the estimates of the expected life expectancy of a patient with CBD.

Step 3—Monetizing the Benefits of the Proposed Rule

To estimate the monetary value of the reductions in the number of beryllium-related fatalities, OSHA relied, as OMB recommends, on estimates developed from the willingness of affected individuals to pay to avoid a marginal increase in the risk of fatality. While a willingness-to-pay (WTP) approach clearly has theoretical merit, it should be noted that an individual's willingness to pay to reduce the risk of fatality would tend to underestimate the total willingness to pay, which would include the willingness of others—particularly the immediate family—to pay to reduce that individual's risk of fatality.

For estimates using the willingness-to-pay concept, OSHA relied on existing studies of the imputed value of fatalities avoided based on the theory of compensating wage differentials in the labor market. These studies rely on certain critical assumptions for their accuracy, particularly that workers understand the risks to which they are exposed and that workers have legitimate choices between high- and low-risk jobs. These assumptions are far from obviously met in actual labor markets. A number of academic studies, as summarized in Viscusi & Aldy (2003), have shown a correlation between higher job risk and higher wages, suggesting that employees demand monetary compensation in return for a greater risk of injury or fatality. The estimated trade-off between lower wages and marginal reductions in fatal occupational risk—that is, workers' willingness to pay for marginal reductions in such risk—yields an imputed value of an avoided fatality: The willingness-to-pay amount for a reduction in risk divided by the reduction in risk.

OSHA has used this approach in many recent proposed and final rules. Although this approach has been criticized for yielding results that are less than statistically robust (see, for example, Hintermann, Alberini and Markandya, 2010), a more recent WTP analysis, by Kniesner et al. (2012), of the trade-off between fatal job risks and wages, using panel data, seems to address many of the earlier econometric criticisms by controlling for measurement error, endogeneity, and heterogeneity. In conclusion, the Agency views the WTP approach as the best available and will rely on it to monetize benefits. OSHA welcomes comments on the use of willingness-to-pay measures and estimates based on compensating wage differentials.

Viscusi & Aldy (2003) conducted a meta-analysis of studies in the economics literature that use a willingness-to-pay methodology to estimate the imputed value of life-saving programs and found that each fatality avoided was valued at approximately $7 million in 2000 dollars. Using the GDP Deflator (U.S. BEA, 2010), this $7 million base number in 2000 dollars yields an estimate of $8.7 million in 2010 dollars for each fatality avoided.

In addition to the benefits that are based on the implicit value of fatalities avoided, workers also place an implicit value on occupational injuries or illnesses avoided, which reflect their willingness to pay to avoid monetary costs (for medical expenses and lost wages) and quality-of-life losses as a result of occupational illness. Chronic beryllium disease and lung cancer can adversely affect individuals for years, or even decades, in non-fatal cases, or before ultimately proving fatal. Because measures of the benefits of avoiding these illnesses are rare and difficult to find, OSHA has included a range based on a variety of estimation methods.

For both CBD and lung cancer, there is typically some permanent loss of lung function and disability, on-going medical treatments, side effects of medicines, and major impacts on one's ability to work, marry, enjoy family life, and quality of life.

While diagnosis with CBD is evidence of material impairment of health, placing a precise monetary value on this condition is difficult, in part because the severity of symptoms may vary significantly among individuals. For that reason, for this preliminary analysis, the Agency employed a broad range of valuation, which should encompass the range of severity these individuals may encounter.

Using the willingness-to-pay approach, discussed in the context of the imputed value of fatalities avoided, OSHA has estimated a range in valuations (updated and reported in 2010 dollars) that runs from approximately $62,000 per case—which reflects estimates developed by Viscusi and Aldy (2003), based on a series of studies primarily describing simple accidents—to upwards of $5 million per case—which reflects work developed by Magat, Viscusi, and Huber (1996) for non-fatal cancer. The latter number is based on an approach that places a willingness-to-pay value to avoid serious illness that is calibrated relative to the value of an avoided fatality. OSHA previously used this approach in the Preliminary Economic Analysis (PEA) supporting its respirable crystalline silica proposal (2013) and in the Final Economic Analysis (FEA) supporting its hexavalent chromium final rule (2006), and EPA (2003) used this approach in its Stage 2 Disinfection and Disinfection Byproducts Rule concerning regulation of primary drinking water. Based on Magat, Viscusi, and Huber (1996), EPA used studies on the willingness to pay to avoid nonfatal lymphoma and chronic bronchitis as a basis for valuing a case of nonfatal cancer at 58.3 percent of the value of a fatal cancer. OSHA's estimate of $5 million for an avoided case of non-fatal cancer is based on this 58.3 percent figure.

The Agency believes this range of estimates, between $62,000 and $5 million, is descriptive of the value of preventing morbidity associated with moderate to severe CBD that ultimately results in premature death.

While the Agency has estimated that 65 percent of CBD cases will result in premature mortality, the Agency has also estimated that approximately 35 percent of CBD cases will not result in premature mortality. However, the Agency acknowledges that it is possible there have been new developments in medicine and industrial hygiene related to the benefits of early detection, medical intervention, and greater control of exposure achieved within the past decade. For that reason, as elsewhere, the Agency requests comment on these issues.

Also not clear are the negative effects of the illness in terms of lost productivity, medical costs, and potential side-effects of a lifetime of immunosuppressive medication. Nonetheless, the Agency is assigning a valuation of $62,000 per case, to reflect the WTP value of a prevented injury not estimated to precede premature mortality. The Agency believes this is conservative, in part because, with any given case of CBD, the outcome is not known in advance, certainly not at the point of discovery; indeed much of the psychic value of preventing the cases may come from removing the threat of premature mortality. In addition, as previously noted, some of these cases could involve relatively severe forms of CBD where the worker died of other causes; however, in those cases, the duration of the disease would be shortened. While beryllium sensitization is a critical precursor of CBD, this preliminary analysis does not attempt to assign a separate value to sensitization itself.

Particularly given the uncertainties in valuation on these questions, the Agency is interested in public input on the issue of valuing the cost to society of morbidity associated with CBD, both in cases preceding mortality, and those that may not result in premature mortality. The Agency is also interested in comments on whether it is appropriate to assign a separate valuation to prevented sensitization cases in their own right, and if so, how such cases should be valued.

a. Summary of Monetized Benefits

Table IX-12 presents the estimated annualized (over 60 years, using a 0 percent discount rate) benefits from each of these components of the valuation, and the range of estimates, based on uncertainty of the prevention factor (i.e., the estimated range of prevented cases, depending on how large an impact the rule has on cases beyond an airborne-only effect), and the range of uncertainty regarding valuation of morbidity. (Mid-point estimates of the undiscounted benefits for each of the first 60 years are provided in the middle columns of Table VII-A-1 in Appendix VII-A at the end of Chapter VII in the PEA. The estimates by year reach a peak of $3.5 billion in the 60th year. Note that, by using a 60-year time-period, OSHA is not including any monetized fatality benefits associated with reduced worker CBD cases originating after year 40 because the 20-year lag takes these CBD fatalities beyond the 60-year time horizon. To this extent, OSHA will have underestimated benefits.)

As shown in Table IX-12, the full range of monetized benefits, undiscounted, for the proposed PEL of 0.2 µg/m³ runs from $291 million annually, in the case of the lowest estimate of prevented cases of CBD, and the lowest valuation for morbidity, up to $2.1 billion annually, for the highest of both. Note that the value of total benefits is more sensitive to the prevention factor used (ranging from $430 million to $1.6 billion, given estimates at the midpoint of the morbidity valuation) than to the valuation of morbidity (ranging from $666 million to $1.3 billion, given estimates at the midpoint of prevention factor).

Also, the analysis illustrates that most of the morbidity benefits are related to CBD and lung cancer cases that are ultimately fatal. At the valuation and case frequency midpoint, $663 million in benefits are related to mortality, $226 million are related to morbidity preceding mortality, and $4.3 million are related to morbidity not preceding mortality.

b. Adjustment of WTP Estimates to Reflect Rising Real Income Over Time

OSHA's estimates of the monetized benefits of the proposed rule are based on the imputed value of each avoided fatality and each avoided beryllium-related disease. As previously discussed, these, in turn, are derived from a worker's willingness to pay to avoid a fatality (with an imputed value per fatality avoided of $8.7 million in 2010 dollars) and to avoid a beryllium-related disease (with an imputed value per disease avoided of between $62,000 and $5 million in 2010 dollars). To this point, these imputed values have been assumed to remain constant over time. However, two related factors suggest that these values will tend to increase over time.

First, economic theory indicates that the value of reducing life-threatening and health-threatening risks—and correspondingly the willingness of individuals to pay to reduce these risks—will increase as real per capita income increases. With increased income, an individual's health and life becomes more valuable relative to other goods because, unlike other goods, they are without close substitutes and in relatively fixed or limited supply. Expressed differently, as income increases, consumption will increase but the marginal utility of consumption will decrease. In contrast, added years of life (in good health) is not subject to the same type of diminishing returns—implying that an effective way to increase lifetime utility is by extending one's life and maintaining one's good health (Hall and Jones, 2007).

Second, real per capita income has broadly been increasing throughout U.S. history, including recent periods. For example, for the period 1950 through 2000, real per capita income grew at an average rate of 2.31 percent a year (Hall and Jones, 2007), although real per capita income for the recent 25-year period 1983 through 2008 grew at an average rate of only 1.3 percent a year (U.S. Census Bureau, 2010). More important is the fact that real U.S. per capita income is projected to grow significantly in future years. For example, the Annual Energy Outlook (AEO) projections, prepared by the Energy Information Administration (EIA) in the Department of Energy (DOE), show an average annual growth rate of per capita income in the United States of 2.7 percent for the period 2011-2035. The U.S. Environmental Protection Agency prepared its economic analysis of the Clean Air Act using the AEO projections. OSHA believes that it is reasonable to use the same AEO projections employed by DOE and EPA, and correspondingly projects that per capita income in the United States will increase by 2.7 percent a year.

On the basis of the predicted increase in real per capita income in the United States over time and the expected resulting increase in the value of avoided fatalities and diseases, OSHA has adjusted its estimates of the benefits of the proposed rule to reflect the anticipated increase in their value over time. This type of adjustment has been recognized by OMB (2003), supported by EPA's Science Advisory Board (EPA, 2000), and applied by EPA . OSHA proposes to accomplish this adjustment by modifying benefits in year i from [B_i] to [B_i * (1 + k)ⁱ], where “k” is the estimated annual increase in the magnitude of the benefits of the proposed rule.

What remains is to estimate a value for “k” with which to increase benefits annually in response to annual increases in real per capita income, where “k” is equal to “(1+g) * (η)”, “g” is the expected annual percentage increase in real per capita income, and “η” is the income elasticity of the value of a statistical life. Probably the most direct evidence of the value of “k” comes from the work of Costa and Kahn (2003, 2004). They estimate repeated labor market compensating wage differentials from cross-sectional hedonic regressions using census and fatality data from the Bureau of Labor Statistics for 1940, 1950, 1960, 1970, and 1980. In addition, with the imputed income elasticity of the value of life on per capita GNP of 1.7 derived from the 1940-1980 data, they then predict the value of an avoided fatality in 1900, 1920, and 2000. Given the change in the value of an avoided fatality over time, it is possible to estimate a value of “k” of 3.4 percent a year from 1900-2000; of 4.3 percent a year from 1940-1980; and of 2.5 percent a year from 1980-2000.

Other, more indirect evidence comes from estimates in the economics literature of “η”, the income elasticity of the value of a statistical life. Viscusi and Aldy (2003) performed a meta-analysis on 0.2 wage-risk studies and concluded that the confidence interval upper bound on the income elasticity did not exceed 1.0 and that the point estimates across a variety of model specifications ranged between 0.5 and 0.6. Applied to a long-term increase in per capita income of about 2.7 percent a year, this would suggest a value of “k” of about 1.5 percent a year.

More recently, Kniesner, Viscusi, and Ziliak (2010), using panel data quintile regressions, developed an estimate of the overall income elasticity of the value of a statistical life of 1.44. Applied to a long-term increase in per capita income of about 2.7 percent a year, this would suggest a value of “k” of about 3.9 percent a year.

Based on the preceding discussion of these three approaches for estimating the annual increase in the value of the benefits of the proposed rule and the fact that the projected increase in real per capita income in the United States has flattened in recent years and could flatten in the long run, OSHA suggests a conservative value for “k” of approximately two percent a year. The Agency invites comment on this estimate and on estimates of the income elasticity of the value of a statistical life.

The Agency believes that the rising value, over time, of health benefits is a real phenomenon that should be taken into account in estimating the annualized benefits of the proposed rule. Table IX-13, in the following section on discounting benefits, shows estimates of the monetized benefits of the proposed rule (under alternative discount rates) with this estimated increase in monetized benefits over time. The Agency invites comment on this adjustment to monetized benefits.

c. The Discounting of Monetized Benefits

As previously noted, the estimated stream of benefits arising from the proposed beryllium rule is not constant from year to year, both because of the 45-year delay after the rule takes effect until all active workers obtain reduced beryllium exposure over their entire working lives and because of, in the case of lung cancer, a 10-year latency period between reduced exposure and a reduction in the probability of disease. An appropriate discount rate is needed to reflect the timing of benefits over the 60-year period after the rule takes effect and to allow conversion to an equivalent steady stream of annualized benefits.

1. Alternative Discount Rates for Annualizing Benefits

Following OMB (2003) guidelines, OSHA has estimated the annualized benefits of the proposed rule using separate discount rates of 3 percent and 7 percent. Consistent with the Agency's own practices in recent rulemakings, OSHA has also estimated, for benchmarking purposes, undiscounted benefits—that is, benefits using a zero percent discount rate.

The question remains, what is the “appropriate” or “preferred” discount rate to use to monetize health benefits? The choice of discount rate is a controversial topic, one that has been the source of scholarly economic debate for several decades. However, in simplest terms, the basic choices involve a social opportunity cost of capital approach or social rate of time preference approach.

The social opportunity cost of capital approach reflects the fact that private funds spent to comply with government regulations have an opportunity cost in terms of foregone private investments that could otherwise have been made. The relevant discount rate in this case is the pre-tax rate of return on the foregone investments (Lind, 1982, pp. 24-32).

The rate of time preference approach is intended to measure the tradeoff between current consumption and future consumption, or in the context of the proposed rule, between current benefits and future benefits. The individual rate of time preference is influenced by uncertainty about the availability of the benefits at a future date and whether the individual will be alive to enjoy the delayed benefits. By comparison, the social rate of time preference takes a broader view over a longer time horizon—ignoring individual mortality and the riskiness of individual investments (which can be accounted for separately).

The usual method for estimating the social rate of time preference is to calculate the post-tax real rate of return on long-term, risk-free assets, such as U.S. Treasury securities (OMB, 2003, p. 33). A variety of studies have estimated these rates of return over time and reported them to be in the range of approximately 1-4 percent.

In accordance with OMB Circular A-4 (2003), OSHA presents benefits and net benefits estimates using discount rates of 3 percent (representing the social rate of time preference) and 7 percent (a rate estimated using the social cost of capital approach). The Agency is interested in any evidence, theoretical or applied, that would inform the application of discount rates to the costs and benefits of a regulation.

2. Summary of Annualized Benefits under Alternative Discount Rates

Table IX-13 presents OSHA's estimates of the sum of the annualized benefits of the proposed rule, using alternative discount rates of 0, 3, and 7 percent, with the suggested adjustment for increasing monetized benefits in response to annual increases in per capita income over time.

Given that the stream of benefits extends out 60 years, the value of future benefits is sensitive to the choice of discount rate. The undiscounted benefits in Table IX-13 range from $291 million to $2.1 billion annually. Using a 7 percent discount rate, the annualized benefits range from $60 million to $591 million. As can be seen, going from undiscounted benefits to a 7 percent discount rate has the effect of cutting the annualized benefits of the proposed rule by about 74 percent.

Taken as a whole, the Agency's best preliminary estimate of the total annualized benefits of the proposed rule—using a 3 percent discount rate with an adjustment for the increasing value of health benefits over time—is between $158 million and $1.2 billion, with a mid-point value of $576 million.

Step 4: Net Benefits of the Proposed Rule

OSHA has estimated, in Table IX-14, the monetized and annualized net benefits of the proposed rule (with a PEL of 0.2 μg/m³), based on the benefits and costs previously presented. Table IX-14 also provides estimates of annualized net benefits for alternative PELs of 0.1 and 0.5 μg/m³. Both the proposed rule and the alternatives PEL options have the same ancillary provisions and an action level equal to half of the PEL in both cases.

Table IX-14 is being provided for informational purposes only. As previously noted, the OSH Act requires the Agency to set standards based on eliminating significant risk to the extent feasible. An alternative criterion of maximizing net (monetized) benefits may result in very different regulatory outcomes. Thus, this analysis of net benefits has not been used by OSHA as the basis for its decision concerning the choice of a PEL or of other ancillary requirements for the proposed beryllium rule.

Table IX-14 shows net benefits using alternative discount rates of 0, 3, and 7 percent for benefits and costs, having previously included an adjustment to monetized benefits to reflect increases in real per capita income over time. OSHA has relied on a uniform discount rate applied to both costs and benefits. The Agency is interested in any evidence, theoretical or applied, that would support or refute the application of differential discount rates to the costs and benefits of a regulation.

As previously noted in this section, the choice of discount rate for annualizing benefits has a significant effect on annualized benefits. The same is true for net benefits. For example, the net benefits using a 7 percent discount rate for benefits are considerably smaller than the net benefits using a 3 percent discount rate, declining by over half under all scenarios. (Conversely, as noted in Chapter V of the PEA, the choice of discount rate for annualizing costs has a relatively minor effect on annualized costs.)

Based on the results presented in Table IX-14, OSHA finds:

• While the net benefits of the proposed rule vary considerably—depending on the choice of discount rate used to annualize benefits and on whether the benefits being used are in the high, midpoint, or low range—benefits exceed costs for the proposed 0.2 μg/m³ PEL in all cases that OSHA considered.
• The Agency's best estimate of the net annualized benefits of the proposed rule—using a uniform discount rate for both benefits and costs of 3 percent—is between $120 million and $1.2 billion, with a midpoint value of $538 million.
• The alternative of a 0.5 μg/m³ PEL has lower net benefits under all assumptions, whereas the effect on net benefits of the 0.1 μg/m³ PEL is mixed, relative to the proposed 0.2 μg/m³ PEL. However, for these alternative PELs, benefits were also found to exceed costs in all cases that OSHA considered.

Incremental Benefits of the Proposed Rule

Incremental costs and benefits are those that are associated with increasing the stringency of the standard. A comparison of incremental benefits and costs provides an indication of the relative efficiency of the proposed PEL and the alternative PELs. Again, OSHA has conducted these calculations for informational purposes only and has not used these results as the basis for selecting the PEL for the proposed rule.

OSHA provides, in Table IX-15, estimates of the net benefits of the alternative 0.1 and 0.5 μg/m³ PELs. The incremental costs, benefits, and net benefits of meeting a 0.5μg/m³ PEL and then going to a 0.2 μg/m³ PEL (as well as meeting a 0.2 μg/m³ PEL and then going to a 0.1 μg/m³ PEL—which the Agency has not yet determined is feasible), for alternative discount rates of 3 and 7 percent, are presented in Table IX-15. Table IX-15 breaks out costs by provision and benefits by type of disease and by morbidity/mortality. As Table IX-15 shows, at a discount rate of 3 percent, a PEL of 0.2 μg/m³, relative to a PEL of 0.5 μg/m³, imposes additional costs of $4.4 million per year; additional benefits of $172.7 million per year; and additional net benefits of $168.2 million per year. The proposed PEL of 0.2 μg/m³ also has higher net benefits, relative to a PEL of 0.5 μg/m³, using a 7 percent discount rate.

Table IX-15 demonstrates that, regardless of discount rate, there are net benefits to be achieved by lowering exposures from the current PEL of 2.0 μg/m³ to 0.5 μg/m³ and then, in turn, lowering them further to 0.2 μg/m³. However, the majority of the benefits and costs attributable to the proposed rule are from the initial effort to lower exposures to 0.5 μg/m³. Consistent with the previous analysis, net benefits decline across all increments as the discount rate for annualizing benefits increases. As also shown in Table IX-15, there is a slight positive net incremental benefit from going from a PEL of 0.2 μg/m³ to 0.1 μg/m³ for a discount rate of 3 percent, and a slight negative net increment for a discount rate of 7 percent. (Note that these results are for OSHA's midpoint estimate of benefits, although as indicated in Table IX-14, this is not universal across all estimation parameters.)

In addition to examining alternative PELs, OSHA also examined alternatives to other provisions of the standard. These regulatory alternatives are discussed Section IX.H of this preamble.

Step 5: Sensitivity Analysis

In this section, OSHA presents the results of two different types of sensitivity analysis to demonstrate how robust the estimates of net benefits are to changes in various cost and benefit parameters. In the first type of sensitivity analysis, OSHA made a series of isolated changes to individual cost and benefit input parameters in order to determine their effects on the Agency's estimates of annualized costs, annualized benefits, and annualized net benefits. In the second type of sensitivity analysis—a so-called “break-even” analysis—OSHA also investigated isolated changes to individual cost and benefit input parameters, but with the objective of determining how much they would have to change for annualized costs to equal annualized benefits. For both types of sensitivity analyses, OSHA used the annualized costs and benefits obtained from a three-percent discount rate as the reference point.

Again, the Agency has conducted these calculations for informational purposes only and has not used these results as the basis for selecting the PEL for the proposed rule.

a. Analysis of Isolated Changes to Inputs

The methodology and calculations underlying the estimation of the costs and benefits associated with this rulemaking are generally linear and additive in nature. Thus, the sensitivity of the results and conclusions of the analysis will generally be proportional to isolated variations in a particular input parameter. For example, if the estimated time that employees need to travel to (and from) medical screenings were doubled, the corresponding labor costs would double as well.

OSHA evaluated a series of such changes in input parameters to test whether and to what extent the general conclusions of the economic analysis held up. OSHA first considered changes to input parameters that affected only costs and then changes to input parameters that affected only benefits. Each of the sensitivity tests on cost parameters had only a very minor effect on total costs or net costs. Much larger effects were observed when the benefits parameters were modified; however, in all cases, net benefits remained significantly positive. On the whole, OSHA found that the conclusions of the analysis are reasonably robust, as changes in any of the cost or benefit input parameters still show significant net benefits for the proposed rule. The results of the individual sensitivity tests are summarized in Table IX-16 and are described in more detail below.

In the first of these sensitivity tests, where OSHA doubled the estimated portion of employees in need of protective clothing and equipment (PPE), essentially doubling the estimated baseline non-compliance rate (e.g., from 10 to 20 percent), and estimates of other input parameters remained unchanged, Table IX-16 shows that the estimated total costs of compliance would increase by $1.4 million annually, or by about 3.7 percent, while net benefits would also decline by $1.4 million annually, from $538.2 million to $536.8 million annually.

In a second sensitivity test, OSHA increased the estimated unit cost of ventilation from $13.18 per cfm for most sectors to $25 per cfm for most sectors. As shown in Table IX-16, if OSHA's estimates of other input parameters remained unchanged, the total estimated costs of compliance would increase by $2.0 million annually, or by about 5.3 percent, while net benefits would also decline by $2.0 million annually, from $538.2 million to $536.2 million annually.

In a third sensitivity test, OSHA increased the estimated share of workers showing signs and symptoms of CBD from 15 to 25 percent, thereby adding these workers to the group eligible for medical surveillance and assuming that they would not be otherwise eligible for another reason (working in a regulated area, exposed during an emergency, etc.). As shown in Table IX-16, if OSHA's estimates of other input parameters remained unchanged, the total estimated costs of compliance would increase by $1.5 million annually, or by about 4.1 percent, while net benefits would also decline by $1.5 million annually, from $538.2 million to $536.7 million annually.

In a fourth sensitivity test, OSHA increased its estimated incremental time per workers for housekeeping by 50 percent. As shown in Table IX-16, if OSHA's estimates of other input parameters remained unchanged, the total estimated costs of compliance would increase by $5.4 million annually, or by about 14.4 percent, while net benefits would also decline by $5.4 million annually, from $538.2 million to $532.8 million annually.

In a fifth sensitivity test, OSHA increased the estimated number of establishments needing engineering controls. For this sensitivity test, if less than 50 percent of the establishments in an industry needed engineering controls, OSHA doubled the percentage of establishments needing engineering controls. If more than 50 percent of establishments in an industry needed engineering controls, then OSHA increased the percentage of establishment needing engineering control to 100 percent. The purpose of this sensitivity analysis was to check the importance of using a methodology that treated 50 percent of workers in a given occupation exposed above the PEL as equivalent to 50 percent of facilities lacking adequate exposure controls. As shown in Table IX-16, if OSHA's estimates of other input parameters remained unchanged, the total estimated costs of compliance would increase by $4.5 million, or by about 11.9 percent, while net benefits would also decline by $4.5 million, from $538.2 million to $533.7 million annually.

The Agency also performed sensitivity tests on several input parameters used to estimate the benefits of the proposed rule. In the first two tests, in an extension of results previously presented in Table IX-12, the Agency examined the effect on annualized net benefits of employing the high-end estimate of the benefits, as well as the low-end estimate, specifically examining the effect on undiscounted benefits of varying the valuation of individual morbidity cases. Table IX-16 presents the effect on annualized net benefits of using the extreme values of these ranges: the high morbidity valuation case and the low morbidity valuation case. For the low estimate of valuation, the benefits decline by 37.7 percent, to $359 million annually, yielding net benefits of $321 million annually. As shown, using the high estimate of morbidity valuation, the benefits rise by 77.0 percent to $1.0 billion annually, yielding net benefits of $982 million annually.

In a third sensitivity test of benefits, the Agency examined the effect of removing the component for the estimated rising value of health and safety over time. This would reduce the benefits by 54.6 percent, or $314 million annually, lowering the net benefits to $224 million annually.

In Chapter VII of the PEA the Agency examined the effect of raising the discount rate for costs and benefits to 7 percent. Raising the discount rate to 7 percent would increase costs by $1.5 million annually and lower benefits by $320.5 million annually, yielding annualized net benefits of $216.2 million.

Also in Chapter VII of the PEA the Agency performed a sensitivity analysis of dental lab substitution. In the PEA, OSHA estimates that 75 percent of the dental laboratory industry will react to a new standard on beryllium by substituting away from using beryllium to the use of other materials. Substitution is not costless, and Chapter V of the PEA estimates the increased cost due to the higher costs of using non-beryllium alloys. These costs are smaller than the avoided costs of the ancillary provisions and engineering controls. Thus, as indicated in Table VII-8 of the PEA, the benefits of the proposal would be lower and the costs higher if there were less substitution out of beryllium in dental labs. The lowest net benefits would occur if labs were unable to substitute out beryllium-containing materials at all, and had to use ventilation to control exposures. In this case, the proposal would yield only $420 million in net benefits. The highest net benefits, larger than assumed for OSHA's primary estimate, would be if all dental labs substituted out of beryllium-containing materials as a result of the proposal; as a result, the proposal would yield $573 million in net benefits. Another possibility is a scenario is which technology and the market move along rapidly away from using beryllium-containing materials, independently of an OSHA rule, and the proposal itself would therefore produce neither costs nor benefits in this sector. If dental labs are removed from the PEA, the net benefits for the proposal—for the remaining industry sectors—decline to $284 million. This analysis demonstrates, however, that regardless of any assumption regarding substitution in dental labs, the proposal would generate substantially more monetized benefits than costs.

Finally, the Agency examined in Chapter VII of the PEA the effects of changes in two important inputs to the benefits analysis: the factor that transforms CBD prevalence rates into incidence rates, needed for the equilibrium lifetime risk model, and the percentage of CBD cases that eventually lead to a fatality.

From the Cullman dataset, the Agency has estimated the prevalence of CBD cases at any point in time as a function of cumulative beryllium exposure. In order to utilize the lifetime risk model, which tracks workers over their working life in a job, OSHA has turned these prevalence rates into an incidence rate, which is the rate of contracting CBD at a point in time. OSHA's baseline estimate of the turnover rate in the model is 10 percent. In Table VII-10 in the PEA, OSHA also presented alternative turnover rates of 5 percent and 20 percent. A higher turnover rate translates into a higher incidence rate, and the table shows that, from a baseline midpoint estimate with 10 percent turnover the number of CBD cases prevented is 6,367, while raising the turnover rate to 20 percent causes this midpoint estimate to rise to 11,751. Conversely, a rate of 5 percent lowers the number of CBD cases prevented to 3,321. Translated into monetary benefits, the table shows that the baseline midpoint estimate of $575.8 million now ranges from $314.4 million to $1,038 million.

Also in TableVII-10 of the PEA, the Agency looked at the effects of varying the percentage of CBD cases that eventuate in fatality. The Agency's baseline estimate of this outcome is 65 percent, with half of this occurring relatively soon, and the other half after an extended debilitating condition. The Agency judged that a reasonable range to investigate was a low of 50 percent and a high of 80 percent, while maintaining the shares of short-term and long-term endpoint fatality. At a baseline of 65 percent, the midpoint estimate of total CBD cases prevented is 4,139. At the low end of 50 percent mortality this estimate lowers to 3,183 while at the high end of 80 percent mortality this estimate rises to 5,094. Translated into monetary benefits, the table shows that the baseline midpoint estimate of $575.8 million now ranges from $500.1 million to $651.5 million.

b. “Break-Even” Analysis

OSHA also performed sensitivity tests on several other parameters used to estimate the net costs and benefits of the proposed rule. However, for these, the Agency performed a “break-even” analysis, asking how much the various cost and benefits inputs would have to vary in order for the costs to equal, or break even with, the benefits. The results are shown in Table IX-17.

In one break-even test on cost estimates, OSHA examined how much total costs would have to increase in order for costs to equal benefits. As shown in Table IX-17, this point would be reached if costs increased by $538.2 million, or by 1,431 percent.

In a second test, looking specifically at the estimated engineering control costs, the Agency found that these costs would need to increase by $566.7 million, or 6,240 percent, for costs to equal benefits.

In a third sensitivity test, on benefits, OSHA examined how much its estimated monetary valuation of an avoided illness or an avoided fatality would need to be reduced in order for the costs to equal the benefits. Since the total valuation of prevented mortality and morbidity are each estimated to exceed the estimated costs of $38 million, an independent break-even point for each is impossible. In other words, for example, if no value is attached to an avoided illness associated with the rule, but the estimated value of an avoided fatality is held constant, the rule still has substantial net benefits. Only through a reduction in the estimated net value of both components is a break-even point possible.

The Agency, therefore, examined how large an across-the-board reduction in the monetized value of all avoided illnesses and fatalities would be necessary for the benefits to equal the costs. As shown in Table IX-17, a 94 percent reduction in the monetized value of all avoided illnesses and fatalities would be necessary for costs to equal benefits, reducing the estimated value to $733,303 per fatality prevented, and an equivalent percentage reduction to about $4,048 per illness prevented.

In a fourth break-even sensitivity test, OSHA estimated how many fewer beryllium-related fatalities and illnesses would be required for benefits to equal costs. Paralleling the previous discussion, eliminating either the prevented mortality or morbidity cases alone would be insufficient to lower benefits to the break-even point. The Agency therefore examined them as a group. As shown in Table IX-17, a reduction of 96 percent, for both simultaneously, is required to reach the break-even point—90 fewer fatalities prevented annually, and 46 fewer beryllium-related illnesses-only cases prevented annually.

Taking into account both types of sensitivity analysis the Agency performed on its point estimates of the annualized costs and annualized benefits of the proposed rule, the results demonstrate that net benefits would be positive in all plausible cases tested. In particular, this finding would hold even with relatively large variations in individual input parameters. Alternately, one would have to imagine extremely large changes in costs or benefits for the rule to fail to produce net benefits. OSHA concludes that its finding of significant net benefits resulting from the proposed rule is a robust one.

OSHA welcomes input from the public regarding all aspects of this sensitivity analysis, including any data or information regarding the accuracy of the preliminary estimates of compliance costs and benefits and how the estimates of costs and benefits may be affected by varying assumptions and methodological approaches. OSHA also invites comment on the risk analysis and risk estimates from which the benefits estimates were derived.

H. Regulatory Alternatives

This section discusses various regulatory alternatives to the proposed OSHA beryllium standard. Executive Order 12866 instructs agencies to “select those approaches that maximize net benefits (including potential economic, environmental, public health and safety, and other advantages; distributive impacts; and equity), unless a statute requires another regulatory approach.” The OSH Act, as interpreted by the courts, requires health regulations to reduce significant risk to the extent feasible. Nevertheless OSHA has examined possible regulatory alternatives that may not meet its statutory requirements.

Each regulatory alternative presented here is described and analyzed relative to the proposed rule. Where appropriate, the Agency notes whether the regulatory alternative, to be a legitimate candidate for OSHA consideration, requires evidence contrary to the Agency's preliminary findings of significant risk and feasibility. To facilitate comment, OSHA has organized some two dozen specific regulatory alternatives into five categories: (1) Scope; (2) exposure limits; (3) methods of compliance; (4) ancillary provisions; and (5) timing.

1. Scope Alternatives

The first set of regulatory alternatives would alter scope of the proposed standard—that is, the groups of employees and employers covered by the proposed standard. The scope of the current beryllium proposal applies only to general industry work, and does not apply to employers when engaged in construction or maritime activities. In addition, the proposed rule provides an exemption for those working with materials that contain beryllium only as a trace contaminant (less than 0.1percent composition by weight).

As discussed in the explanation of paragraph (a) in Section XVIII of this preamble, Summary and Explanation of the Proposed Standard, OSHA is considering alternatives to the proposed scope that would increase the range of employers and employees covered by the standard. OSHA's review of several industries indicates that employees in some construction and maritime industries, as well as some employees who deal with materials containing less than 0.1 percent beryllium, may be at significant risk of CBD and lung cancer as a result of their occupational exposures. Regulatory Alternatives #1a, #1b, #2a, and #2b would increase the scope of the proposed standard to provide additional protection to these workers.

Regulatory Alternative #1a would expand the scope of the proposed standard to also include all operations in general industry where beryllium exists only as a trace contaminant; that is, where the materials used contain less than 0.1 percent beryllium by weight. Regulatory Alternative #1b is similar to Regulatory Alternative #1a, but exempts operations where beryllium exists only as a trace contaminant and the employer can show that employees' exposures will not meet or exceed the action level or exceed the STEL. Where the employer has objective data demonstrating that a material containing beryllium or a specific process, operation, or activity involving beryllium cannot release beryllium in concentrations at or above the proposed action level or above the proposed STEL under any expected conditions of use, that employer would be exempt from the proposed standard except for recordkeeping requirements pertaining to the objective data. Alternative #1a and Alternative #1b, like the proposed rule, would not cover employers or employees in construction or shipyards.

OSHA has identified two industries with workers engaged in general industry work that would be excluded under the proposed rule but would fall within the scope of the standard under Regulatory Alternatives #1a and #1b: Primary aluminum production and coal-fired power generation. Beryllium exists as a trace contaminant in aluminum ore and may result in exposures above the proposed permissible exposure limits (PELs) during aluminum refining and production. Coal fly ash in coal-powered power plants is also known to contain trace amounts of beryllium, which may become airborne during furnace and baghouse operations and might also result in worker exposures. See Appendices VIII-A and VIII-B at the end of Chapter VIII in the PEA for a discussion of beryllium exposures and available controls in these two industries.

As discussed in Appendix IV-B of the PEA, beryllium exposures from fly ash high enough to exceed the proposed PEL would usually be coupled with arsenic exposures exceeding the arsenic PEL. Employers would in that case be required to implement all feasible engineering controls, work practices, and necessary PPE (including respirators) to comply with the OSHA Inorganic Arsenic standard (29 CFR 1910.1018)—which would be sufficient to comply with those aspects of the proposed beryllium standard as well. The degree of overlap between the applicability of the two standards and, hence, the increment of costs attributable to this alternative are difficult to gauge. To account for this uncertainty, the Agency at this time is presenting a range of costs for Regulatory Alternative #1a: From no costs being taken for ancillary provisions under Regulatory Alternative #1a to all such costs being included. At the low end, the only additional costs under Regulatory Alternative #1a are due to the engineering control costs incurred by the aluminum smelters (see Appendix VIII-A).

Similarly, the proposed beryllium standard would not result in additional benefits from a reduction in the beryllium PEL or from ancillary provisions similar to those already in place for the arsenic standard, but OSHA does anticipate some benefits will flow from ancillary provisions unique to the proposed beryllium standard. To account for significant uncertainty in the benefits that would result from the proposed beryllium standard for workers in primary aluminum production and coal-fired power generation, OSHA estimated a range of benefits for Regulatory Alternative #1a. The Agency estimated that the proposed ancillary provisions would avert between 0 and 45 percent of those baseline CBD cases not averted by the proposed PEL. Though the Agency is presenting a range for both costs and benefits for this alternative, the Agency judges the degree of overlap with the arsenic standard is likely to be substantial, so that the actual costs and benefits are more likely to be found at the low end of this range. The Agency invites comment on all these issues.

Table IX-18 presents, for informational purposes, the estimated costs, benefits, and net benefits of Regulatory Alternative #1a using alternative discount rates of 3 percent and 7 percent. In addition, this table presents the incremental costs, incremental benefits, and incremental net benefits of this alternative relative to the proposed rule. Table IX-18 also breaks out costs by provision, and benefits by type of disease and by morbidity/mortality.

As shown in Table IX-18, Regulatory Alternative #1a would increase the annualized cost of the rule from $37.6 million to between $39.6 and $56.0 million using a 3 percent discount rate and from $39.1 million to between $41.3 and $58.1 million using a 7 percent discount rate. OSHA estimates that regulatory Alternative #1a would prevent as few as an additional 0.3 (i.e., almost one fatality every 3 years) or as many as an additional 31.8 beryllium-related fatalities annually, relative to the proposed rule. OSHA also estimates that Regulatory Alternative #1a would prevent as few as an additional 0.002 or as many as an additional 9 beryllium-related non-fatal illnesses annually, relative to the proposed rule. As a result, annualized benefits in monetized terms would increase from $575.8 million to between $578.0 and $765.2 million, using a 3 percent discount rate, and from $255.3 million to between $256.3 and $339.3 million using a 7 percent discount rate. Net benefits would increase from $538.2 million to between $538.4 and $709.2 million using a 3 percent discount rate and from $216.2 million to somewhere between $215.1 to $281.2 million using a 7 percent discount rate. As noted in Appendix VIII-B of Chapter VIII in the PEA, the Agency emphasizes that these estimates of benefits are subject to a significant degree of uncertainty, and the benefits associated with Regulatory Alternative #1a arguably could be a small fraction of OSHA's best estimate presented here.

OSHA estimates that the costs and the benefits of Regulatory Alternative #1b will be somewhat lower than the costs of Regulatory Alternative #1a, because most—but not all—of the provisions of the proposed standard are triggered by exposures at the action level, 8-hour time-weighted average (TWA) PEL, or STEL. For example, where exposures exist but are below the action level and at or below the STEL, Alternative #1a would require employers to establish work areas; develop, maintain, and implement a written exposure control plan; provide medical surveillance to employees who show signs or symptoms of CBD; and provide PPE in some instances. Regulatory Alternative #1b would not require employers to take these measures in operations where they can produce objective data demonstrating that exposures are below the action level and at or below the STEL. OSHA only analyzed costs, not benefits, for this alternative, consistent with the Agency's treatment of Regulatory Alternatives in the past. Total costs for Regulatory Alternative #1b versus #1a, assuming full ancillary costs, drop from to $56.0 million to $49.9 million using a 3 percent discount rate, and from $58.1 million to $51.8 million using a 7 percent discount rate.

Regulatory Alternative #2a would expand the scope of the proposed standard to include employers in construction and maritime. For example, this alternative would cover abrasive blasters, pot tenders, and cleanup staff working in construction and shipyards who have the potential for airborne beryllium exposure during blasting operations and during cleanup of spent media. Regulatory Alternative #2b would update 29 CFR 1910.1000 Tables Z-1 and Z-2, 1915.1000 Table Z, and 1926.55 Appendix A so that the proposed TWA PEL and STEL would apply to all employers and employees in general industry, shipyards, and construction, including occupations where beryllium exists only as a trace contaminant. For example, this alternative would cover abrasive blasters, pot tenders, and cleanup staff working in construction and shipyards who have the potential for significant airborne exposure during blasting operations and during cleanup of spent media. The changes to the Z tables would also apply to workers exposed to beryllium during aluminum refining and production, and workers engaged in maintenance operations at coal-powered utility facilities. All provisions of the standard other than the PELs, such as exposure monitoring, medical removal, and PPE, would be in effect only for employers and employees that fall within the scope of the proposed rule. Alternative #2b would not be as protective as Alternative #1a or Alternative #1b for employees in aluminum refining and production or coal-powered utility facilities because the other provisions of the proposed standard would not apply.

As discussed in the explanation of proposed paragraph (a) in this preamble at Section XVIII, Summary and Explanation of the Proposed Standard, abrasive blasting is the primary application group in construction and maritime industries where workers may be exposed to beryllium. OSHA has judged that abrasive blasters and their helpers in construction and maritime industries have the potential for significant airborne exposure during blasting operations and during cleanup of spent media. Airborne concentrations of beryllium have been measured above the current TWA PEL of 2 μg/m³ when blast media containing beryllium are used as intended (see Appendix IV-C in the PEA for details).

To address high concentrations of various hazardous chemicals in abrasive blasting material, employers must already be using engineering and work practice controls to limit workers' exposures and must be supplementing these controls with respiratory protection when necessary. For example, abrasive blasters in the construction industry fall under the protection of the Ventilation standard (29 CFR 1926.57). The Ventilation standard includes an abrasive blasting subsection (29 CFR 1926.57(f)), which requires that abrasive blasting respirators be worn by all abrasive blasting operators when working inside blast-cleaning rooms (29 CFR 1926.57(f)(5)(ii)(A)), or when using silica sand in manual blasting operations where the nozzle and blast are not physically separated from the operator in an exhaust-ventilated enclosure (29 CFR 1926.57(f)(5)(ii)(B)), or when needed to protect workers from exposures to hazardous substances in excess of the limits set in § 1926.55 (29 CFR 1926.57(f)(5)(ii)(C); ACGIH, 1971). For maritime, standard 29 CFR 1915.34(c) covers similar requirements for respiratory protection needed in blasting operations. Due to these requirements, OSHA believes that abrasive blasters already have controls in place and wear respiratory protection during blasting operations. Thus, in estimating costs for Regulatory Alternatives #2a and #2b, OSHA judged that the reduction of the TWA PEL would not impose costs for additional engineering controls or respiratory protection in abrasive blasting (see Appendix VIII-C of Chapter VIII in the PEA for details). OSHA requests comment on this issue—in particular, whether abrasive blasters using blast material that may contain beryllium as a trace contaminant are already using all feasible engineering and work practice controls, respiratory protection, and PPE that would be required by Regulatory Alternatives #2a and #2b.

In the estimation of benefits for Regulatory Alternative #2a, OSHA has estimated a range to account for significant uncertainty in the benefits to this population from some of the ancillary provisions of the proposed beryllium standard. It is unclear how many of the workers associated with abrasive blasting work would benefit from dermal protection, as comprehensive dermal protection may already be used by most blasting operators. It is also unclear whether the housekeeping requirements of the proposed standard would be feasible to implement in the context of abrasive blasting work, and to what extent they would benefit blasting helpers, who are themselves exposed while performing cleanup activities. OSHA estimated that the proposed ancillary provisions would avert between 0 and 45 percent of those baseline CBD cases not averted by the proposed PEL.

These considerations also lead the Agency to present a range for the costs of this alternative: From no costs being estimated for ancillary provisions under Regulatory Alternative #2a to including all such costs. Based on the considerations discussed above, the Agency judges that costs and benefits at the low end of this range are more likely to be correct. The Agency invites comment on these issues.

In addition, OSHA believes that a small number of welders in the maritime industry may be exposed to beryllium via arc and gas welding (and none through resistance welding). The number of maritime welders was estimated using the same methodology as was used to estimate the number of general industry welders. Brush Wellman's customer survey estimated 2,000 total welders on beryllium-containing products (Kolanz, 2001). Based on ERG's assumption of 4 welders per establishment, ERG estimated that a total of 500 establishments would be affected. These affected establishments were then distributed among the 26 NAICS industries with the highest number of IMIS samples for welders that were positive for beryllium. To do this, ERG first consulted the BLS OES survey to determine what share of establishments in each of the 26 NAICS employed welders and estimated the total number of establishments that perform welding regardless of beryllium exposure (BLS, 2010a). Then ERG distributed the 500 affected beryllium welding facilities among the 26 NAICS based on the relative share of the total number of establishments performing welding. Finally, to estimate the number of welders, ERG used the assumption of four welders per establishment. Based on the information from ERG, OSHA estimated that 30 welders would be covered in the maritime industry under this regulatory alternative. For these welders, OSHA used the same controls and exposure profile that were used to estimate costs for arc and gas welders in Chapter V of the PEA. ERG judged there to be no construction welders exposed to beryllium due to a lack of any evidence that the construction sector uses beryllium-containing products or electrodes in resistance welding. OSHA solicits comment and any relevant data on beryllium exposures for welders in construction and maritime employment.

Estimated costs and benefits for Regulatory Alternative #2a are shown in Table IX-18a. Regulatory Alternative #2a would increase costs from $37.6 million to between $37.7 and $55.3 million, using a 3 percent discount rate, and from $39.1 million to between $39.2 and $57.3 million using a 7 percent discount rate. Annualized benefits would increase from $575.8 million to between $575.9 and $675.3 million using a 3 percent discount rate, and from $255.3 million to between $255.4 and $299.4 million using a 7 percent discount rate. Net benefits would change from $538.2 million to between $538.2 and $620.0 million using a 3 percent discount rate, and from $216.2 million to between $216.1 and $242.1 million using a 7 percent discount rate.

Table IX-18b presents, for informational purposes, the estimated costs, benefits, and net benefits, of Regulatory Alternative #2b using alternative discount rates of 3 percent and 7 percent. In addition, this table presents the incremental costs, incremental benefits, and incremental net benefits of this alternative relative to the proposed rule. Table IX-18b also breaks out costs by provision and benefits by type of disease and by morbidity/mortality.

As shown in Table IX-18b, this regulatory alternative would increase the annualized cost of the rule from $37.6 million to $39.6 million, using a 3 percent discount rate, and from $39.1 million to $41.1 million using a 7 percent discount rate. Regulatory Alternative #2b would prevent less than one additional beryllium-related fatalities and less than one beryllium-related illness annually relative to the proposed rule. As a result, annualized benefits would increase from $575.8 million to $578.1 million, using a 3 percent discount rate, and from $255.3 million to $256.3 million using a 7 percent discount rate. Net benefits would increase from $538.2 million to $538.5 million using a 3 percent discount rate and slightly decrease from $216.2 million to $215.2 million using a 7 percent discount rate.

2. Exposure Limit (TWA PEL, STEL, and ACTION LEVEL) Alternatives

OSHA is proposing a new TWA PEL for beryllium of 0.2 μg/m³ and a STEL of 2.0 μg/m³ for all application groups covered by the rule. OSHA's proposal is based on the requirements of the Occupational Safety and Health Act (OSH Act) and court interpretations of the Act. For health standards issued under section 6(b)(5) of the OSH Act, OSHA is required to promulgate a standard that reduces significant risk to the extent that it is technologically and economically feasible to do so. See Section II of this preamble, Pertinent Legal Authority, for a full discussion of OSHA legal requirements.

Paragraph (c) of the proposed standard establishes two PELs for beryllium in all forms, compounds, and mixtures: An 8-hour TWA PEL of 0.2 μg/m³ (proposed paragraph (c)(1)), and a 15-minute short-term exposure limit (STEL) of 2.0 μg/m³ (proposed paragraph (c)(2)). OSHA has defined the action level for the proposed standard as an airborne concentration of beryllium of 0.1 μg/m³ calculated as an eight-hour TWA (proposed paragraph (b)). In this proposal, as in other standards, the action level has been set at one-half of the TWA PEL.

As discussed in this preamble explanation of paragraph (c) in Section XVIII, Summary and Explanation of the Proposed Standard, OSHA is considering three regulatory alternatives that would modify the PELs for the proposed standard.

Regulatory Alternative #3 would modify the proposed STEL to be five times the TWA PEL, as is typical for OSHA standards that have STELs. A STEL five times the TWA PEL has more practical effect because a STEL ten times the TWA PEL will rarely be exceeded without also driving exposures above the TWA PEL. For example, assuming a background exposure level of 0.1 μg/m³, a STEL ten times the TWA PEL could only be exceeded once in a work shift for 15 minutes without driving exposures above the TWA PEL, whereas a STEL five times the TWA PEL could be exceeded three times before driving exposures above the TWA PEL. OSHA's standards for methylene chloride (29 CFR 1910.1052), acrylonitrile (29 CFR 1910.1045), benzene (29 CFR 1910.1028), ethylene oxide (29 CFR 1910.1047), and 1,3-Butadiene (29 CFR 1910.1051) all set STELs at five times the TWA PEL. Thus, if OSHA promulgates the proposed TWA PEL of 0.2 μg/m³, the accompanying STEL under this regulatory alternative would be set at 1 μg/m³.

As discussed in this preamble at Section V, Health Effects, immunological sensitization can be triggered by short-term exposures. OSHA believes a STEL for beryllium will help reduce the risk of sensitization and CBD in beryllium-exposed employees. For instance, without a STEL, workers' exposures could be as high as 6.4 μg/m³ (32 × 0.2 μg/m³) for 15 minutes under the proposed TWA PEL, if exposures during the remainder of the 8-hour work shift are non-detectable. A STEL serves to minimize high task-based exposures by requiring feasible controls in these situations, and has the added effect of further reducing the TWA exposure.

OSHA requests comment on the range of short-term exposures in covered industries, the types of operations where these are occurring, and on the proposed and alternative STELs, including any data or information that may help OSHA choose between them.

OSHA identified two job categories where workers would be expected to have short-term exposures in the range between the proposed STEL and the STEL under Regulatory Alternative #3 (that is, between 2.0 and 1.0 μg/m³): Furnace operators in nonferrous foundries and material preparation operators in the beryllium oxide ceramics application group. To estimate the costs for this alternative, OSHA judged, conservatively, that all workers in these job categories would need to wear respirators to meet a STEL of 1.0. OSHA also estimated costs for additional regulated areas and medical surveillance for workers in these two job categories. The costs for this alternative are presented in Table IX-19. Total costs rise from $37.6 million to $37.7 million using a 3 percent discount rate and from $39.1 million to $39.3 million using a 7 percent discount rate.

Under Regulatory Alternative #4, the TWA PEL would be 0.1 μ g/m³ with an action level of 0.05 μ g/m³. The Agency's preliminary risk assessment indicates that the risks remaining at the proposed TWA PEL of 0.2 μg/m³—while lower than risks at the current TWA PEL—are still significant (see this preamble at Section VIII, Significance of Risk). A TWA PEL of 0.1 μg/m³ would reduce some of the remaining risks to workers at the proposed PEL. The OSH Act requires the Agency to set its standards to address significant risks of harm to the extent economically and technologically feasible, so OSHA would have very limited flexibility to adopt a higher PEL if a lower PEL is technologically and economically feasible.

While OSHA's preliminary analysis indicates that the proposed TWA PEL of 0.2 μg/m³ is economically and technologically feasible, OSHA has less confidence in the feasibility of a TWA PEL of 0.1 μg/m³. In some industry sectors it is difficult to determine whether a TWA PEL of 0.1 μg/m³ could be achieved in most operations most of the time (see Section IX.D of this preamble, Technological Feasibility). OSHA believes that one way this uncertainty could be resolved would be with additional information on exposure control technologies and the exposure levels that are currently being achieved in these industry sectors. OSHA requests additional data and information to inform its final determinations on feasibility (see Section IX.D of this preamble, Technological Feasibility) and the alternative PELs under consideration.

Regulatory Alternative #5, which would set a TWA PEL at 0.5 μg/m³ and an action level at 0.25 μg/m³, both higher than in the proposal, responds to an issue raised during the Small Business Advocacy Review (SBAR) process conducted in 2007 to consider a draft OSHA beryllium proposed rule that culminated in an SBAR Panel report (SBAR, 2008). That report included a recommendation that OSHA consider both the economic impact of a low TWA PEL and regulatory alternatives that would ease cost burden for small entities. OSHA has provided a full analysis of the economic impact of its proposed PELs (see Chapter VI of the PEA), and Regulatory Alternative #5 addresses the second half of that recommendation. However, the higher 0.5 μ g/m³ TWA PEL does not appear to be consistent with the Agency's mandate under the OSH Act to promulgate a lower PEL if it is feasible and could prevent additional fatalities and non-fatal illnesses. The data presented in Table IX-20 below indicate that the lower TWA PEL would prevent additional fatalities and non-fatal illnesses, but nevertheless the Agency solicits comments on this alternative and OSHA's analysis of the costs and benefits associated with it.

Table IX-20 below presents, for informational purposes, the estimated costs, benefits, and net benefits of the proposed rule under the proposed TWA PEL of 0.2 μg/m³ and for the regulatory alternatives of a TWA PEL of 0.1 μg/m³ and a TWA PEL of 0.5 μg/m³ (Regulatory Alternatives #4 and #5, respectively), using alternative discount rates of 3 percent and 7 percent. In addition, the table presents the incremental costs, the incremental benefits, and the incremental net benefits, of going from a TWA PEL of 0.5 μg/m³ to the proposed TWA PEL of 0.2 μg/m³ and then of going from the proposed TWA PEL of 0.2 μg/m³ to a TWA PEL of 0.1 μg/m³. Table IX-20 also breaks out costs by provision and benefits by type of disease and by morbidity/mortality.

OSHA has not made a determination that a TWA PEL of 0.1 μg/m³ would be feasible for all application groups (that is, engineering and work practices would be sufficient to reduce and maintain beryllium exposures to a TWA PEL of 0.1 μg/m³ or below in most operations most of the time in the affected industries). For Regulatory Alternative #4, the Agency attempted to identify engineering controls and their costs for those affected application groups where the technology feasibility analysis in Chapter IV of the PEA indicated that a TWA PEL of 0.1 μg/m³ could be achieved. For those application groups, OSHA costed out the set of feasible controls necessary to meet this alternative PEL. For the rest of the affected application groups, OSHA assumed that all workers exposed between 0.2 μg/m³ and 0.1 μg/m³ would have to wear respirators to achieve compliance with the 0.1 μg/m³ TWA PEL and estimated the associated additional costs for respiratory protection. For all affected industries, OSHA also estimated the costs to satisfy the ancillary requirements specified in the proposed rule for all affected workers under the alternative TWA PEL of 0.1 μg/m³. For both controls and respirators, the unit costs were the same as presented in Chapter V of the PEA.

The estimated benefits for Regulatory Alternative #4 were calculated based on the number of workers identified with exposures between 0.1 and 0.2 μg/m³, using the methods and unit benefit values developed in Chapter VII of the PEA.

As Table IX-20 shows, going from a TWA PEL of 0.5 μg/m³ to a TWA PEL of 0.2 μg/m³ would prevent, annually, an additional 29 beryllium-related fatalities and an additional 15 non-fatal illnesses. This is consistent with OSHA's preliminary risk assessment, which indicates significant risk to workers exposed at a TWA PEL of 0.5 μg/m³; furthermore, OSHA's preliminary feasibility analysis indicates that a lower TWA PEL than 0.5 μg/m³ is feasible. Net benefits of this regulatory alternative versus the proposed TWA PEL of 0.2 μg/m³ would decrease from $538.2 million to $370.0 million using a 3 percent discount rate and from $216.2 million to $144.4 million using 7 percent discount rate.

Table IX-20 also shows the costs and benefits of going from the proposed TWA PEL of 0.2 μg/m³ to a TWA PEL of 0.1 μg/m³. As shown there, going from a TWA PEL of 0.2 μg/m³ to a TWA PEL of 0.1 μg/m³ would prevent an additional 2 beryllium-related fatalities and 1 additional non-fatal illness. Net benefits of this regulatory alternative versus the proposed TWA PEL of 0.2 μg/m³ would increase from $538.2 million to $543.5 million using a 3 percent discount rate and decrease from $216.2 million to $214.9 million using a 7 percent discount rate.

Informational Alternative Featuring Unchanged PEL but Full Ancillary Provisions

An Informational Analysis: This proposed regulation has the somewhat unusual feature for an OSHA substance-specific health standard that most of the quantified benefits would come from the ancillary provisions rather than from meeting the PEL with engineering controls. OSHA decided to analyze for informational purposes the effect of retaining the existing PEL but applying all of the ancillary provisions, including respiratory protection. Under this approach, the TWA PEL would remain at 2.0 micrograms per cubic meter, but all of the other proposed provisions (including respiratory protection, which OSHA does not consider an ancillary provision) would be required with their triggers remaining the same as in the proposed rule—either the presence of airborne beryllium at any level (e.g., initial monitoring, written exposure control plan), at certain kinds of dermal exposure (PPE), at the action level of 0.1 μg/m³ (e.g., periodic monitoring, medical removal), or at 0.2 μg/m³ (e.g., regulated areas, respiratory protection, medical surveillance).

Given the record regarding beryllium exposures, this approach is not one OSHA could legally adopt because the absence of a more protective requirement for engineering controls would not be consistent with section 6(b)(5) of the OSH Act, which requires OSHA to “set the standard which most adequately assures, to the extent feasible, on the basis of the best available evidence, that no employee will suffer material impairment of health or functional capacity even if such employee has regular exposure to the hazard dealt with by such standard for the period of his working life.” For that reason, this additional analysis is provided strictly for informational purposes. E.O. 12866 and E.O. 13563 direct agencies to identify approaches that maximize net benefits, and this analysis is purely for the purpose of exploring whether this approach would hold any real promise to maximize net benefits if it was permissible under the OSH Act. It does not appear to hold such promise because an ancillary-provisions-only approach would not be as protective and thus offers fewer benefits than one that includes a lower PEL and engineering controls, and OSHA estimates the costs would be about the same (or slightly lower, depending on certain assumptions) under that approach as under the traditional proposed approach.

On an industry by industry basis, OSHA found that some industries would have lower costs if they could adopt the ancillary-provisions-only approach. Some employers would use engineering controls where they are cheaper, even if they are not mandatory. OSHA does not have sufficient information to do an analysis of the employer-by-employer situations in which there exist some employers for whom the ancillary-provisions-only approach might be cheaper. In the majority of affected industries, the Agency estimates there are no costs saving to the ancillary-provisions-only approach. However, OSHA estimates a total of $2,675,828 per year in costs saving for entire industries where the ancillary-provisions-only approach would be less expensive.

The above discussion does not account for the possibility that the lack of engineering controls would result in higher beryllium exposures for workers in adjacent (non-production) work areas due to the increased level of beryllium in the air. Because of a lack of data, and because the issue did not arise in the other regulatory alternatives OSHA considered (all of which have a PEL of less than 2.0 μg/m³), OSHA did not carefully examine exposure levels in non-production areas for either cost or benefit purposes. To the extent such exposure levels would be above the action level, there would be additional costs for respiratory protection.

The ancillary-provisions-only approach adds uncertainty to the benefits analysis such that the benefits of the rule as proposed may exceed, and perhaps greatly exceed, the benefits of this ancillary-provisions-only approach:

(1) Most exposed individuals would be in respirators, which OSHA considers less effective than engineering controls in preventing employee exposure to beryllium. OSHA last did an extensive review of the evidence on effectiveness of respirators for its APFs rulemaking in 2006 (71 FR 50128-45, August 24, 2006). OSHA has not in the past tried to quantify the size of this effect, but it could partially negate the estimated benefits of 92 CBD deaths prevented per year and 4 lung cancer cases prevented per year by the proposed standard.

(2) As noted above, in the proposal OSHA did not consider benefits caused by reductions in exposure in non-production areas. Unless employers act to reduce exposures in the production areas, the absence of a requirement for such controls would largely negate such benefits from reductions in exposure in the non-productions areas.

(3) OSHA believes that there is a strong possibility that the benefits of the ancillary provisions (a midpoint estimate of eliminating 45 percent of all remaining cases of CBD) would be partially or wholly negated in the absence of engineering controls that would reduce both airborne and surface dust levels. The measured reduction in benefits from ancillary provision was in a facility with average exposure levels of less than 0.2 μg/m³.

Based on these considerations, OSHA believes that the ancillary-provisions-only approach is not one that is likely to maximize net benefits. The costs saving, if any, are estimated to be small, and the difficult-to-measure declines in benefits could be substantial.

3. A Method-of-Compliance Alternative

Paragraph (f)(2) of the proposed rule contains requirements for the implementation of engineering and work practice controls to minimize beryllium exposures in beryllium work areas. For each operation in a beryllium work area, employers must ensure that at least one of the following engineering and work practice controls is in place to minimize employee exposure: Material and/or process substitution; ventilated enclosures; local exhaust ventilation; or process controls, such as wet methods and automation. Employers are exempt from using engineering and work practice controls only when they can show that such controls are not feasible or where exposures are below the action level based on two exposure samples taken seven days apart.

These requirements, which are based on the stakeholders' recommended beryllium standard that beryllium industry and union stakeholders submitted to OSHA in 2012 (Materion and United Steelworkers, 2012), address a concern associated with the proposed TWA PEL. OSHA expects that day-to-day changes in workplace conditions, such as workers' positioning or patterns of airflow, may cause frequent exposures above the TWA PEL in workplaces where periodic sampling indicates exposures are between the action level and the TWA PEL. As a result, the default under the standard is that the controls are required until the employer can demonstrate that exposures have not exceeded the action level from at least two separate measurements taken seven days apart.

OSHA believes that substitution or engineering controls such as those outlined in paragraph (f)(2)(i) provide the most reliable means to control variability in exposure levels. However, OSHA also recognizes that the requirements of paragraph (f)(2)(i) are not typical of OSHA standards, which usually require engineering controls only where exposures exceed the TWA PEL or STEL. The Agency is therefore considering Regulatory Alternative #6, which would drop the provisions of (f)(2)(i) from the proposed standard and make conforming edits to paragraphs (f)(2)(ii) and (iii). This regulatory alternative does not eliminate the need for engineering controls to comply with the proposed TWA PEL and STEL, but does eliminate the requirement to use one or more of the specified engineering or work practice controls where exposures equal or exceed the action level. As shown in Table IX-21, Regulatory Alternative #6 would decrease the annualized cost of the proposed rule by about $457,000 using a discount rate of 3 percent and by about $480,000 using a discount rate of 7 percent. OSHA has not been able to estimate the change in benefits resulting from Regulatory Alternative #6 at this time and invites public comment on this issue.

4. Regulatory Alternatives That Affect Ancillary Provisions

The proposed standard contains several ancillary provisions (provisions other than the exposure limits), including requirements for exposure assessment, medical surveillance, medical removal, training, and regulated areas or access control. As reported in Chapter V of the PEA, these ancillary provisions account for $27.8 million (about 72 percent) of the total annualized costs of the rule ($37.6 million) using a 3 percent discount rate, or $28.6 million (about 73 percent) of the total annualized costs of the rule ($39.1 million) using a 7 percent discount rate. The most expensive of the ancillary provisions are the requirements for housekeeping and training, with annualized costs of $12.6 million and $5.8 million, respectively, at a 3 percent discount rate ($12.9 million and $5.8 million, respectively, at a 7 percent discount rate).

OSHA's reasons for including each of the proposed ancillary provisions are explained in Section XVIII of this preamble, Summary and Explanation of the Standards.

In particular, OSHA is proposing the requirements for exposure assessment to provide a basis for ensuring that appropriate measures are in place to limit worker exposures. Medical surveillance is especially important because workers exposed above the proposed TWA PEL, as well as many workers exposed below the proposed TWA PEL, are at significant risk of death and illness. Medical surveillance would allow for identification of beryllium-related adverse health effects at an early stage so that appropriate intervention measures can be taken. OSHA is proposing regulated areas and access control because they serve to limit exposure to beryllium to as few employees as possible. OSHA is proposing worker training to ensure that employers inform employees of the hazards to which they are exposed, along with associated protective measures, so that employees understand how they can minimize their exposure to beryllium. Worker training on beryllium-related work practices is particularly important in controlling beryllium exposures because engineering controls frequently require action on the part of workers to function effectively.

OSHA has examined a variety of regulatory alternatives involving changes to one or more of the proposed ancillary provisions. The incremental cost of each of these regulatory alternatives and its impact on the total costs of the proposed rule is summarized in Table IX-22 at the end of this section. OSHA has preliminarily determined that several of these ancillary provisions will increase the benefits of the proposed rule, for example, by helping to ensure the TWA PEL is not exceeded or by lowering the risks to workers given the significant risk remaining at the proposed TWA PEL. However, except for Regulatory Alternative #7 (involving the elimination of all ancillary provisions), OSHA did not estimate changes in monetized benefits for the regulatory alternatives that affect ancillary provisions. Two regulatory alternatives that involve all ancillary provisions are presented below (#7 and #8), followed by regulatory alternatives for exposure monitoring (#9, #10, and #11), for regulated areas (#12), for personal protective clothing and equipment (#13), for medical surveillance (#14 through #21), and for medical removal (#22).

a. All Ancillary Provisions

The SBAR Panel recommended that OSHA analyze a PEL-only standard as a regulatory alternative. The Panel also recommended that OSHA consider not applying ancillary provisions of the standard where exposure levels are low so as to minimize costs for small businesses (SBAR, 2008). In response to these recommendations, OSHA analyzed Regulatory Alternative #7, a PEL-only standard, and Regulatory Alternative #8, which would apply ancillary provisions of the beryllium standard only where exposures exceed the proposed TWA PEL of 0.2 μg/m³ or the proposed STEL of 2 μg/m³.

Regulatory Alternative #7 would solely update 1910.1000 Tables Z-1 and Z-2, so that the proposed TWA PEL and STEL would apply to all workers in general industry. This alternative would eliminate all of the ancillary provisions of the proposed rule, including exposure assessment, medical surveillance, medical removal, PPE, housekeeping, training, and regulated areas or access control. Under this regulatory alternative, OSHA estimates that the costs for the proposed ancillary provisions of the rule (estimated at $27.8 million annually at a 3 percent discount rate) would be eliminated. In order to meet the PELs, employers would still commonly need to do monitoring, train workers on the use of controls, and set up some kind of regulated areas to indicate where respirator use would be required. It is also likely that, under this alternative, many employers would follow the recommendations of Materion and the United Steelworkers to provide medical surveillance, PPE, and other protective measures for their workers (Materion and USW, 2012). OSHA has not attempted to estimate the extent to which these ancillary-provision costs would be incurred if they were not formally required or whether any of these costs under Regulatory Alternative #7 would reasonably be attributable to the proposed rule. OSHA welcomes comment on the issue.

OSHA has also estimated the effect of this regulatory alternative on the benefits of the rule. As a result of eliminating all of the ancillary provisions, annualized benefits are estimated to decrease 57 percent, relative to the proposed rule, from $575.8 million to $249.1 million, using a 3 percent discount rate, and from $255.3 million to $110.4 million using a 7 percent discount rate. This estimate follows from OSHA's analysis of benefits in Chapter VII of the PEA, which found that about 57 percent of the benefits of the proposed rule, evaluated at their mid-point value, were attributable to the combination of the ancillary provisions. As these estimates show, OSHA expects that the benefits estimated under the proposed rule will not be fully achieved if employers do not implement the ancillary provisions of the proposed rule.

Both industry and worker groups have recognized that a comprehensive standard is needed to protect workers exposed to beryllium. The stakeholders' recommended standard that representatives of the primary beryllium manufacturing industry and the United Steelworkers union provided to OSHA confirms the importance of ancillary provisions in protecting workers from the harmful effects of beryllium exposure (Materion and USW, 2012). Ancillary provisions such as personal protective clothing and equipment, regulated areas, medical surveillance, hygiene areas, housekeeping requirements, and hazard communication all serve to reduce the risks to beryllium-exposed workers beyond that which the proposed TWA PEL alone could achieve.

Moreover, where there is continuing significant risk at the TWA PEL, the decision in the Asbestos case (Bldg. and Constr. Trades Dep't, AFL-CIO v. Brock, 838 F.2d 1258, 1274 (D.C. Cir. 1988)) indicated that OSHA should use its legal authority to impose additional requirements on employers to further reduce risk when those requirements will result in a greater than de minimis incremental benefit to workers' health. Nevertheless, OSHA requests comment on this alternative.

Under Regulatory Alternative #8, several ancillary provisions that the current proposal would require under a variety of exposure conditions (e.g., dermal contact, any airborne exposure, exposure at or above the action level) would instead only apply where exposure levels exceed the TWA PEL or STEL. Regulatory Alternative #8 affects the following provisions of the proposed standard:

—Exposure monitoring: Whereas the proposed standard requires annual monitoring when exposure levels are at or above the action level and at or below the TWA PEL, Regulatory Alternative #8 would require annual exposure monitoring only where exposure levels exceed the TWA PEL or STEL;

—Written exposure control plan: Whereas the proposed standard requires written exposure control plans to be maintained in any facility covered by the standard, Regulatory Alternative #8 would require only facilities with exposures above the TWA PEL or STEL to maintain a plan;

—Housekeeping: Whereas the proposed standard's housekeeping requirements apply across a wide variety of beryllium exposure conditions, Alternative #8 would limit housekeeping requirements to areas and employees with exposures above the TWA PEL or STEL;

—PPE: Whereas the proposed standard requires PPE for employees under a variety of conditions, such as exposure to soluble beryllium or visible contamination with beryllium, Alternative #8 would require PPE only for employees exposed above the TWA PEL or STEL;