fb tracking

Comments on OSHA’s Proposed Rule on Hexavalent Chromium

Docket Office
Docket H054A
Room N-2625
Occupational Safety and Health Administration
U.S. Department of Labor
200 Constitution Ave, NW
Washington, DC 20210

To whom it may concern:

Public Citizen’s Health Research Group submits the following comments on the Occupational Safety and Health Administration’s (OSHA’s) Proposed Rule on hexavalent chromium.[1] This rulemaking would not be taking place at all were it not for a successful lawsuit filed in March 2002 by Public Citizen and the Paper, Allied-Industrial, Chemical and Energy Workers International Union (PACE).   We look forward to seeing this rulemaking completed, its content improved in the ways set forth below, and the regulation speedily enacted, lest thousands more workers suffer preventable death from lung cancer due to exposure to hexavalent chromium.

OSHA’s Risk Assessment

The essential basis for the risk assessment put forth by OSHA is the decision to place heavy and equal emphasis on just two studies: the Gibb[2] and Luippold[3] studies.  Both studies find markedly elevated risks of lung cancer among hexavalent chromium-exposed workers and, as discussed below, find dose-response relationships well characterized by linear models.  Both followed their subjects for an average of 30 years.  However, their results are quite different.  The Gibb study generates maximum likelihood estimates of excess lung cancer risk that are three to five times higher than those produced by the Luippold study at equivalent levels of cumulative hexavalent chromium exposure.  The confidence intervals for the two studies do not overlap, which OSHA characterizes as “statistical inconsistency.”[4]  Nonetheless, by using the maximum likelihood estimates from the two studies as the range for its risk assessment, the agency has effectively chosen to weigh the studies equally.

Yet the two studies do not merit equal weighting.  As the table below illustrates, in almost every respect the Gibb study is superior to the Luippold study.


Gibb Study

Luippold Study




Person-years of follow-up



Loss to follow-up



Lung cancer deaths



Exposure data collection


Industrial hygiene surveys

Exposure measurements



Includes low exposures?



Smoking assessment

93% of cohort

35% of cohort

The Gibb study has five times more workers and person-years of follow-up, 2.4 times as many lung cancer deaths, much better smoking data and over 80 times as many exposure measurements, and these were more randomly obtained than in the Luippold study. 

Moreover, the Luippold study has inconsistent inclusion criteria, excluding certain workers who later worked at other chromate plants, but including an unknown number who worked at another chromate plant, because exposure data were available.  Although the Gibb study had many short-term workers, a fact played up repeatedly by the industry, when the industry consultant Exponent reanalyzed the Gibb data to exclude the short-term workers, it found little impact on the dose-response relationship described by the full cohort.  The remaining studies cited by OSHA are still weaker, with little or no smoking data, limited follow-up and exposure assessments that are frequently not linked to individuals.

The use of the maximum likelihood estimate from the Luippold study as the lower bound of OSHA’s risk estimates (instead of relying on the Gibb study alone or some weighting of the studies) has the effect of making a higher Permissible Exposure Limit (PEL) appear acceptable.

A Pigment of Their Imagination

A consistent refrain in the more than a decade since Public Citizen and one of PACE’s predecessor unions, the Oil, Chemical and Atomic Workers Union (OCAW) initially petitioned OSHA for a reduced PEL for hexavalent chromium has been the claim from the Color Pigment Manufacturers Association that its products were somehow exempt from the cancer-causing propensities of other hexavalent chromium chemicals.  As discussed below, the record compiled by OSHA makes a convincing case that hexavalent chromium-containing pigments are indeed carcinogens, and may even be more carcinogenic than the compounds used in the studies upon which OSHA’s risk assessment is based.

Even as the industry has sought to sow controversy, scientific experts have increasingly reaped consensus.  As long ago as 1984, the Environmental Protection Agency (EPA) concluded that there was “decisive evidence” that all hexavalent chromium compounds are carcinogenic to humans.[5]  In 1988, the National Institute for Occupational Safety and Health (NIOSH) advised OSHA to consider all hexavalent chromium compounds to be occupational carcinogens.[6]  In 1990, the International Agency for Research on Cancer (IARC) concluded that there was “sufficient evidence” (its highest level of evidence) in humans for the carcinogenicity of hexavalent chromium, including chromium pigments.[7]

In Table VI-2 of its Proposed Rule, OSHA summarizes the major cohort studies of workers exposed to hexavalent chromium in pigment production.  While these studies have their limitations, and are in general not as convincing individually as the Gibb or, to a lesser extent, the Luippold study, a consistent pattern emerges: in adequately-powered studies, the standardized mortality ratios for exposed workers are significantly elevated (range 1.5-4.4) and a relationship between extent of exposure (whether measured by duration of exposure or factory) generally emerges.  These studies must be placed in the context of the large number of cohort studies in chromate production workers and chrome platers which, in general, have similar findings.  They must also be placed in the context of the animal carcinogenicity studies (Tables VI-7, VI-8 and VI-9) and the mechanistic studies reviewed by OSHA which suggest that the carcinogenicity of relatively insoluble chromates (a category including many pigments) may well be higher than the estimates from the Gibb and Luippold studies, which are based on soluble chromates.  For these reasons, the American Conference of Governmental Industrial Hygienists (ACGIH) has actually recommended a lower exposure limit for insoluble hexavalent chromium compounds than soluble ones.

The Elusive Threshold Effect

OSHA’S attempts to regulate occupational carcinogens frequently elicit industry claims that there is a threshold effect (i.e., that chemicals lose their ability to cause cancer below some exposure level).  A related concept is that the dose-response data are non-linear — that the rate of disease does not increase with exposure in a straight line.  Yet both the risk assessment conducted on contract for OSHA by Environ and another by NIOSH find that non-linear models fail to characterize the data in the Gibb study more accurately than simpler linear models.  As the OSHA analysis indicates, the implication by the industry-funded consultant Exponent that the Gibb study’s failure to find a statistically significant elevated risk for lung cancer at lower hexavalent chromium exposures represents a threshold is “misleading, and not considered a valid analysis.”[8]  Elsewhere in the Proposed Rule, the agency dismisses a similar industry claim of a threshold effect with respect to the Luippold study as “not supported by the data.”[9]  Indeed, even the authors of the Luippold study acknowledge that when the data were tested for linearity, “this test showed no departure from linearity, indicating that a linear model may be compatible with the data.” [3]Risk assessments conducted by both Environ for OSHA and the industry-funded Crump, et al.[10] using the Luippold data confirm the linear nature of the dose-response relationship in that study.

But the industry finds the holy grail of the threshold effect scattered throughout the hexavalent chromium literature.  The threshold effect is said to characterize dermal exposures; OSHA says the study cited “does not mimic the occupational experience”[11] because workers with abraded skin were excluded.  It is said to characterize the Glaser rat study; OSHA dismisses this claim for lack of statistical power.  It is said to characterize the Steinhoff and Snyder rat studies; OSHA believes that the dosing schedules in these studies in part explain these findings.  The industry even puts forth convoluted mechanistic arguments based on the De Flora study to assert the existence of a threshold; OSHA rejects these, in part, because the study does not reflect real conditions in the body.  OSHA concludes that it “does not believe that there is sufficient scientific evidence”[12] to support assertions of non-linearity based on these mechanistic data.

We are encouraged that, even in the face of this barrage of misleading information from industry, OSHA has stuck to its principles and declared that “[I]n accordance with the Agency’s long standing cancer policy, OSHA believes it is inappropriate to establish a threshold or ‘no effect’ level of exposure to a carcinogen.”[13]

OSHA’s Proposed Permissible Exposure Limit

If nothing else, the risk assessments make clear that OSHA has been guilty of a massive dereliction of duty in not promulgating this rule sooner.  As early as 1995, the agency had a risk assessment that provided estimates of risk generally similar to those presented in the Federal Register.[14]  Yet nearly a decade has passed and we are still at least a year from a Final Rule regulating a compound about which OSHA acknowledged in 1994 “there is clear evidence that exposure … at the current PEL … can result in an excess risk of lung cancer” and other related illnesses.[15]  Based on the Gibb study, exposure to hexavalent chromium at the current PEL of 52 mg/m3 (measured as Cr) for a working lifetime (the required assessment under the OSHAct) would result in 351 excess lung cancer deaths per 1,000 workers (95% confidence interval [CI]: 181-493).[16]  Stated differently, over one-third of lifetime workers exposed at currently legal concentrations would die from lung cancer due to hexavalent chromium.   Because in our study of hexavalent chromium exposures measured by OSHA between 1990 and 2000, 21.3% of exposures exceeded the current PEL,[17] the preventable damage to human health due to the lack of a standard has been very substantial.

Even under the proposed new PEL of 1 mg/m3, significant risk would remain.  The Gibb study yields an estimate of 9.1 excess lung cancer deaths per 1,000 workers exposed for a working lifetime at that level (95% CI: 4.0-16).[18]  OSHA itself acknowledges that “even at the proposed PEL, the risk posed to workers with a lifetime of regular exposure is still clearly significant.”[19]  The estimate only dips below the 1/1,000 excess lifetime risk used as a yardstick in the 1980 Supreme Court decision in the benzene case[20]when the PEL is lowered to 0.25 mg/m3, the PEL Public Citizen and OCAW petitioned for in 1993.[21]  In the Gibb study, the risk of lung cancer was elevated 1.6-fold (95% CI: 1.1-2.2) in the exposure quartile that included our suggested PEL.  We therefore stand by our assessment that adequate protection for workers can only assured at a PEL of 0.25 mg/m3.

The economic assessment prepared by OSHA indicates that, in general, reaching the proposed PEL is within the reach of the affected industries.   The annualized incremental cost to comply with the new PEL is $223 million for industries with annual revenues of $721 billion (0.03%).  (Incidentally, we reject OSHA’s method of including in the costs of compliance the costs of bringing industries into compliance with existing standards presented in Table IV-4.)  Although the costs can fall disproportionately on particular industries and smaller business entities, in only one industry (chromium catalyst production with three business entities employing 313 people, or 0.08% of the 381,000 workers affected by the standard) did the incremental costs exceed one percent of revenues (1.07%).  In only four industries representing no more than 19% of workers affected by the standard do these costs exceed 10% of profits.  We therefore agree with OSHA that “the proposed standard is economically feasible.”[22]

With the proposed PEL still resulting in significant residual health risk, yet being economically feasible, the only possible rationale for not reducing the PEL below the current proposal is the lack of technological feasibility.  OSHA’s only justification in the Federal Register in claiming this lack of feasibility is literally as follows:

PELs lower than 1 mg/m3 could not be achieved by means of engineering controls and work practices alone for some types of welding (particularly [gas metal arc welding] and [shielded metal arc welding]) and in hard chromium plating.[23]

This hardly qualifies as adequate justification for denying all workers the additional health benefits otherwise required by law.   Given that the claim of technological infeasibility for these particular applications is the only explanation OSHA has offered for concluding that a lower PEL would not be feasible across the board (the agency concedes the economic feasibility of its proposed PEL and that the proposed PEL would still leave “clearly significant” health risks), it is evident that OSHA has failed to offer an adequate justification for not reducing risk further with a lower PEL. 

Even if OSHA is correct about the lack of technical feasibility in two industrial subcategories (the sizes of which are not disclosed in the Federal Register), why can’t the standard be tailored to take these processes into account or to allow longer phase-in for selected industries or processes?   If the agency can adequately justify the technological infeasibility of this lower PEL for welding and hard chromium plating, this would seem like the ideal situation for a Separate Engineering Control Air Limit (SECAL).

Separate Standards for General Industry, Construction and Shipyards

Despite a recommendation from the Advisory Committee for Construction Safety and Health (ACCSH) that all of the construction industries be included with general industry under a single standard, the agency has opted instead to develop separate standards for general industry, construction and shipyards.   (The Maritime Advisory Committee for Occupational Safety and Health recommended a separate standard for shipyards.)   Unlike the more stringent requirements of the general industry standard, the standards for construction and shipyards will require no exposure monitoring (and hence no action level, see below), will require medical surveillance only for those with signs or symptoms consistent with hexavalent chromium exposure (in general industry, exposures over the PEL often trigger the requirement for such surveillance), and will not be subject to the requirement for maintaining “regulated areas” (restricted-acces areas known or expected to have exposures exceeding the PEL).  Somehow employers in the construction and shipyards industries are expected to meet the PEL even without being required to conduct initial, let alone intermittent, exposure monitoring.  The standard allows companies to claim compliance by the use of historical data or objective data, whatever these are.  To our knowledge, this is the first time that OSHA has employed so irrational an approach to exposure reduction.  The lack of a monitoring requirement effectively guts the construction and shipyard standards.

Exposure Assessment

OSHA’s approach to exposure assessment is to require companies covered by the general industry standard to conduct baseline exposure assessment unless they have monitored their workplaces in the previous 12 months and found exposures to be under half the PEL (the action level) or they can provide evidence that their work procedures will not produce exposures exceeding the action level.   If baseline monitoring yields exposure levels under the action level and these are confirmed, monitoring can cease entirely unless work procedures are changed.  Exposures between the action level and the PEL trigger a requirement for six-monthly exposure monitoring and those above the PEL require quarterly testing.  Viewed from another perspective, testing continues at three- to six-monthly intervals until the action level is reached and confirmed with a repeat test a week later, at which point monitoring stops forever unless work practices change.

The problem with this approach, as OSHA readily allows, is that exposures can fluctuate.    The agency has failed to document the likely extent of this fluctuation, either in magnitude or over time.   If fluctuations are substantial, the exposure monitoring requirement is degraded into a form of roulette in which one keeps keeps gambling (testing) until the action level is reached and then carries off the winnings (the termination of the requirement to test). 

Exclusion of Portland Cement from Construction Standard

The agency has also ignored ACCSH’s recommendation that portland cement be included in the construction standard.  It claims that the current OSHA PEL for Particulates Not Otherwise Regulated will be at least as protective as the proposed hexavalent chromium PEL.  It also claims that no workers using portland cement in construction are exposed to hexavalent chromium levels exceeding 0.25 mg/m3, a finding that is not well documented in the Federal Register notice.  We would prefer the approach taken for general industry: that employers in the construction industry using portland cement be required to undertake baseline exposure assessment.  If exposures are as low as OSHA predicts, the general industry standard provides for the termination of further exposure monitoring.

Exemptions to Requirement for Engineering and Work Practice Controls

While OSHA makes much of its belief that engineering and work practice controls are, in general, the preferred method for reducing worker exposures, the agency has carved out a major exception to this principle.   If workers are not exposed to hexavalent chromium at levels exceeding the PEL for more than 30 days per year, the companies can rely upon the less-desirable personal protective equipment to reduce exposure.   As under the rule allowing evasion of the initial exposure assessment (see above), companies in the construction and shipyard sectors can claim this exemption using historical or objective data, including production orders.  This approach is far too vague.

Medical Surveillance

The important requirement that workers with exposures to hexavalent chromium levels exceeding the PEL for more than 30 days per year receive medical monitoring is also undermined by its exceptions.  In particular, construction and shipyards are exempt from this requirement entirely (they must provide medical surveillance only when workers have symptoms or are exposed in an emergency), another casualty of the separate standards for these industries.  We are mystified as to why exposures at a certain level would trigger medical surveillance in one industry but not in another.  Moreover, this departs from previous OSHA practice in which the requirement for medical surveillance was not triggered unless exposures were more than the action level (not the PEL) for more than 30 days.


In sum, the agency has proposed a PEL that it concedes leaves “clearly significant” residual health risks, but is economically feasible.   It has justified not lowering its proposed PEL still further by making the ill-documented claim that a lower PEL is technologically infeasible in just two industrial processes.  The real question left by this Proposed Rule is: What is the basis for not reducing the PEL to the 0.25 mg/m3 level requested by Public Citizen and the OCAW a dozen years ago, since that is the level at which the risk of lung cancer is reduced below the 1 in 1,000 lifetime risk benchmark that defines significant risk under the statute?

Assuming that the agency complies with the court-ordered completion rate for this rulemaking of January 18, 2006, it will be about a decade and a half since OSHA last proposed and then promulgated a regulation concerning an occupational chemical.  As a consequence, the agency’s credibility as a protector of worker health has eroded seriously.  The expeditious conclusion of this rulemaking, with a standard more protective of worker health than the one currently proposed, could be the first step on the way back to new-found legitimacy for the agency.

Yours sincerely,

Peter Lurie, M.D., M.P.H.
Deputy Director

Sidney M. Wolfe, M.D.
Public Citizen’s Health Research Group

[1] 69 Fed Reg 59306-474, October 4, 2004.

[2] Gibb HJ, Lees PSJ, Pinsky PF, Rooney BC. Lung cancer among workers in chromium chemical production. American Journal of Industrial Medicine 2000;38:115-26.

[3] Luippold RS, Mundt KA, Austin RP, et al. Lung cancer mortality among chromate production workers. Occupational and Environmental Medicine 2003;60:451-7.

[4] 69 Fed Reg 59378, October 4, 2004.

[5] US Environmental Protection Agency. Health assessment document for chromium. Environmental Criteria and Assessment Office, Research Triangle Park, NC, 1984 (US EPA-600/8-83-014).

[6] Millar JD, Schulte PA. Testimony of The National Institute for Occupational Safety and Health on the Occupational Safety and Health Administration’s Proposed Rule on Air Contaminants. August 1, 1988.

[7] International Agency for Research on Cancer. IARC monographs on the evaluation of carcinogenic risks to humans, chromium, nickel and welding. 1990;49:49-256. Lyons, France, World Health Organization.

[8] 69 Fed Reg 59370, October 4, 2004.

[9] 69 Fed Reg 59374, October 4, 2004.

[10] Crump C, Crump K, Hack E, at al. Dose-response and risk assessment of airborne hexavalent chromium and lung cancer mortality. Risk Analysis 2003;23:1147-63.

[11] 69 Fed Reg 59359, October 4, 2004.

[12] 69 Fed Reg 59382, October 4, 2004.

[13] 69 Fed Reg 59344, October 4, 2004.

[14] K.S. Crump Division. Evaluation of epidemiological data and risk assessment for hexavalent chromium. ICF Kaiser, Ruston, LA, May 1995.

[15] Dear JA, Asistant Secretary of Labor for Occupational Safety and Health. Letter to Sidney M. Wolfe, Director, Public Citizen’s Health Research Group, March 8, 1994.

[16] The Luippold data produce an estimate of 101 excess lung cancer deaths (95% CI: 62-147) per 1,000 lifetime workers under these conditions.

[17] Lurie P, Wolfe SM. Continuing exposure to hexavalent chromium, a known lung carcinogen: an analysis of OSHA compliance inspections, 1990-2000. American Journal of Industrial Medicine 2002;42:378-83.

[18] The Luippold data produce an estimate of 2.1 excess lung cancer deaths (95% CI: 1.2-3.1) per 1,000 lifetime workers under these conditions.

[19] 69 Fed Reg 59391, October 4, 2004.

[20] Industrial Union Department, AFL-CIO v. American Petroleum Institute, 448 U.S. 607 (1980).

[21]The Luippold data produce an estimate of 1.0 excess lung cancer deaths (95% CI: 0.62-1.6) per 1,000 lifetime workers at a PEL of 0.5 mg/m3.

[22] 69 Fed Reg 59420, October 4, 2004.

[23] 69 Fed Reg 59407, October 4, 2004.