Hexavalent Chromium Case Study

Abstract

Hexavalent chromium (Cr(VI)) is a natural metal used in a wide variety of industrial activities, including the manufacture of stainless steel, welding, painting and pigment application, electroplating, and other surface coating processes. According to the Occupational Safety and Health Administration (OSHA), there are over 558,000 workers exposed to Cr(VI).

Exposure to Cr(VI) has been associated with increased lung cancer risk for more than 50 years. OSHA was petitioned in 1993 and sued in 1997 and 2002 to lower the workplace Cr(VI) exposure limit, resulting in a court order to issue a final standard by February 2006. Faced with the threat of stronger regulation, the chromium industry initiated an effort to challenge the scientific evidence supporting a more protective standard. This effort included the use of “product defense” consultants to conduct post hoc analyses of a publicly funded study to challenge results viewed unfavorable by the industry.

The industry also commissioned a study of the mortality experience of workers at four low-exposure chromium plants, but did not make the results available to OSHA in a timely manner, despite multiple agency requests for precisely these sorts of data. The commissioned study found a statistically significant elevation in lung cancer risk among Cr(VI)-exposed workers at levels far below the 52 µg/m³ standard. This finding changed when the multi-plant cohort was divided into two statistically underpowered components and then published separately. The findings of the first paper published have been used by the chromium industry to attempt to slow OSHA’s standard-setting process. The second paper was withheld from OSHA until it was accepted for publication in a scientific journal, after the rulemaking record had closed.

Studies funded by private sponsors that seek to influence public regulatory proceedings should be subject to the same access and reporting provisions as those applied to publicly funded science. Parties in regulatory proceedings should be required to disclose whether the studies were performed by researchers who had the right to present their findings without the sponsor's consent or influence, and to certify that all relevant data have been submitted to the public record, whether published or not.

Background

The American National Standards Institute in 1943 recommended an exposure standard of 52 micrograms of Cr(VI) per cubic meter of air (52 µg/m³) as a level adequate to prevent nasal perforations in chromium-exposed workers. This limit was adopted by the newly created OSHA in 1971 as the Permissible Exposure Limit (PEL), without any formal review.

In 1993, Public Citizen and the Oil, Chemical and Atomic Workers International Union (OCAW) (now part of the United Steelworkers) petitioned OSHA to reduce its PEL from 52 µg/m³ to 0.25 µg/m³, measured as an 8-hour time-weighted average. Although the OSHA administrator acknowledged in 1994 that “there is clear evidence that exposure… at the current PEL… can result in an excess of lung cancer,” OSHA failed to act. Two lawsuits ensued, challenging OSHA’s “unreasonably delay” in promulgating a stronger standard. Although the chromium industry, through its trade association the Chrome Coalition, had intervened in the lawsuits on OSHA’s behalf opposing a change in the PEL, on April 2, 2003, the US Court of Appeals for the Third Circuit ordered the agency to issue a final rule reducing occupational exposure to Cr(VI) by January 18, 2006, later extended to February 28, 2006. In the words of Judge Edward Becker, OSHA’s decade-long delay in issuing a Cr(VI) standard “exceeded the bounds of reasonableness.” 

On October 4, 2004, OSHA published a proposed PEL of 1 µg/m³ of Cr(VI) per cubic meter of air. The final rule was issued on February 28, 2006. The new standard is significantly weaker than the standard OSHA proposed in October of 2004; once the standard goes into effect (beginning in November 2006 for businesses with 20 or more employees; engineering controls need not be implemented until May 2010), the new PEL will be 5 µg/m³ rather than the level of 1 µg/m³ in the proposed rule. The PEL in the aerospace industry will be 25 µg/m³. According to OSHA’s own estimates, exposure at 5 µg/m³ will result in 10 to 45 lung cancer deaths per 1,000 workers over a lifetime of exposure. In the final rule, OSHA justifies the weaker standard by arguing that a PEL of 5 µg/m³ reduces the significant risk posed to workers by occupational exposure to Cr(VI) “to the maximum extent that is technologically and economically feasible.”  

Industry Strategy to Forestall OSHA Rulemaking

Long before the 2003 court ruling, the chromium industry had initiated an effort to challenge the scientific evidence supporting any stronger OSHA standard, engaging the services of ChemRisk and Exponent, Inc., two consulting firms that specialize in “litigation support” and “product defense.” One industry document noted that “this route [hiring the consultants] is expensive and success is not guaranteed, [but] the longer we wait the more difficult the task becomes.”

In a meeting with chromium industry representatives in 1996, ChemRisk scientists outlined a strategy that included obtaining and analyzing the raw data from a not-yet-published study of Cr(VI) exposure funded by the Environmental Protection Agency (EPA) in order "to forestall the [OSHA] rulemaking." Simultaneously, the industry commissioned new publications that questioned the health effects of low levels of exposure to Cr(VI) (see here, here, and here), a central issue in any OSHA regulatory initiative. The industry paid for services provided by ChemRisk and Exponent, Inc. through its trade association's attorneys. This arrangement was selected to "...preserve the confidentiality of information, opinion, and data to the extent provided for under the attorney-client privilege and attorney work product privilege", ensuring that material developed through the process could be sequestered from public view. [See agreement between Collier, Shannon, Rill & Scott, PLLC, ChemRisk and the Industrial Health Foundation.]

The industry also contracted with a third consulting firm, ENVIRON, to study workers who had only been employed in facilities that were either designed with or converted to production processes that resulted in lower levels of Cr(VI) exposure. ENVIRON was hired through a contract with the Industrial Health Foundation (IHF). (In 2004, ENVIRON purchased Applied Epidemiology, Inc., which had signed the original contract with the industry. In this case study we use the name ENVIRON for both that company and its predecessor.)

The study protocol entailed combining workers from four plants using newer, lower-exposure processes – two in the U.S. (Castle Hayne, NC, and Corpus Christi, TX) and two in Germany – into a single cohort. ENVIRON's proposal noted that "the relatively small study sizes and short follow-up periods resulted in a limited ability of [previous] studies to clarify the relationship between modern [low-level] occupational chromate exposures and cancer in general, and respiratory cancers in particular." According to the draft proposal, creating a single cohort with workers from multiple plants was crucial "to improve statistical power and the inferential value of the results."

In August 2000, the EPA study was published. It is the largest, most comprehensive study ever conducted on the effects of workplace Cr(VI) exposure. Using OSHA's standard assumption of a 45-year working lifetime, the study reported a significantly elevated lung cancer risk of 1.57 among workers whose mean exposure was at levels just above the PEL requested in the Public Citizen-OCAW petition.

No sooner had the EPA study been published than the industry-sponsored critiques began. Scientists with the product defense firm Exponent, Inc. created and analyzed the mortality experience of a "simulated cohort" derived by computer from the EPA study's summary data, standard deviations and ranges. In another report, the consultants obtained the raw data from the EPA study through a Freedom of Information Act request and re-analyzed them. Each of these reports challenged the validity of the EPA study's conclusions and was either entered into the record in litigation or submitted to OSHA by the chromium industry, although not published in the peer-reviewed literature. After extensive analysis, most of the issues raised in these critiques were rejected by OSHA.

OSHA Publishes its Proposed Rule

On October 4, 2004, OSHA published its court-mandated proposed rule for Cr(VI), including a PEL of 1 µg/m³. The agency issued a general request for additional scientific evidence, along with a specific appeal for epidemiological data about the aforementioned cohort in Castle Hayne, where exposure levels were more representative of the concentrations of airborne Cr(VI) found in workplaces today. A mortality study of this group had been published in 1994 and OSHA asked directly, "Are there updated analyses available for [this cohort]?" In addition, in the proposed rule OSHA asked, "Are there other cohorts available to look at low exposures?"

Following a three-month comment period, OSHA held 11 days of public hearings [Docket H054A, Exhibit 45-1 thru 45-11], at which OSHA reiterated its request for more data and industry repeatedly criticized OSHA for relying on data from high-exposure cohorts [see comments from Deborah Proctor, on behalf of the Aerospace Industries Association (Exhibit 45-9, pages 1829-1833 and 1850-1851), Jack Shilling, Chairman, Specialty Steel Industry of North America (Exhibit 45-4, pages 498-499), Joan Fessler, Carpenter Technology Corporation (Exhibit 45-4, pages 659-660), Charles Kelly, Edison Electric Institute (Exhibit 45-3, pages 423-433), Peter Marr, Dominion Colour Corporation (Exhibit 45-9, pages 1745-1751), and Frank White, Organization Resource Counselors Inc (Exhibit 45-6, pages 1060-1065)]. In reviewing the hearing transcript, we found no mention by industry representatives or anyone else of any imminent new epidemiological evidence. The public was given until April 20, 2005, to submit additional data and post-hearing comments.

Selected Science

Just weeks before the close of the comment period, a study reporting on the mortality experience of workers employed at the Castle Hayne and Corpus Christi facilities appeared in the Journal of Occupational and Environmental Medicine (JOEM). The article had been submitted to JOEM in July 2004 and was accepted for publication that October [1], the same month OSHA proposed its rule and specifically asked for information about the Castle Hayne or other cohorts. The analysis has little statistical power and suffers from short follow-up. Even though they collected data on Cr(VI) exposure, none is presented in the paper and the small sample size precludes logistic regression. Nonetheless, the authors offer the "preliminary conclusion" that "the absence of an elevated lung cancer risk at this time may be a favorable reflection of the post-change [i.e., lower exposure] environment."

Three trade associations made reference to the study in their post-hearing comments. For example, the Specialty Steel Industry of North America stated it had "recently" learned of the study:

[W]hile we have not had any opportunity to examine this study...[it] contains potentially incredibly significant data which would allow the development of a dose response relationship based on actual, experienced exposures, as opposed to the modeled exposures upon which OSHA currently relies to set the PEL. Indisputably, this would be much more relevant and appropriate data upon which to establish a risk-based regulatory limit [Post-hearing brief submitted on behalf of Specialty Steel Industry of North America, submitted by Collier Shannon Scott, PLLC, April 20, 2005. Exhibit 47-27-1].

The Specialty Steel Industry warned that OSHA's failure to consider these results would be "arbitrary and capricious," a legal term, signaling that failure to address these "new" findings would be grounds for a legal challenge. The Society of the Plastics Industry, Inc. (SPI) remarked on the "potentially great significance" [Susan Howe, SPI: Letter to OSHA Docket Office; April 20, 2005. Exhibit 47-24-113] of the new ENVIRON study.

Moreover, the comments by these trade associations confirm that they were privy to unpublished details of the ENVIRON analysis regarding a German study, and that the details were discussed among the trade associations. For example, in their submission to OSHA, the Surface Finishing Industry Council wrote:

Recent risk studies on chromate production workers demonstrate further uncertainties associated with the risk assessment science on the relationship between exposure to hexavalent chromium and lung cancer. For example, a study just published in the Journal of Occupational and Environmental Medicine found that at low exposure levels, workers at two chromate production facilities showed no excess lung cancers. Similarly, a German study of chromate production workers that has not yet been published apparently indicates that no excess lung cancers are expected at lower exposure levels. This study relies on very frequent biological monitoring of workers in conjunction with workplace exposure levels.  [Post-Hearing Comments submitted on behalf of the Surface Finishing Industry council, submitted by The Policy Group, April 20, 2005. Exhibit 47-35-1]

The Society of the Plastics Industry, Inc. (SPI) noted that:

[I]n the German plants, excess lung cancer mortality was demonstrated only in the highest exposure group, using chromium exposure estimates based on urinary chromium results. It is possible that the data obtained from the German facilities demonstrates (sic) that no increase in risk at any but the highest exposure levels to CrVI [Letter to OSHA Docket Office].

Intrigued by these developments, we conducted an Internet search, using the terms "Industrial Health Foundation" and "Chrome Coalition." To our surprise, we located a notice for a hearing related to the bankruptcy of IHF. In this proceeding, two chromium industry trade associations asserted that files in the possession of IHF actually belonged to the industry, because the IHF was, according to the petitioner, simply a "third-party administrator of the trade association." [See affidavit.] Using the Public Access to Court Electronic Records system, we obtained documents filed with the court, some of which have been quoted in this case study. These materials also led us to parties in the bankruptcy proceedings who provided additional documents, including ENVIRON's study protocol and the final report of the combined study of the U.S. and German plants.

That report, submitted by ENVIRON to IHF in September 2002 but never by the industry to OSHA and never published in its entirety, provides strong support for the inadequacy of the 52 µg/m³ standard, and raises questions about whether even the PEL proposed in 2004, 1 µg/m³, would be adequately protective. The investigators developed a series of job exposure matrices and utilized air monitoring data from the U.S. plants and urine monitoring from the German plants to estimate the exposure history of each worker. The ENVIRON authors found a significantly elevated risk of lung cancer mortality associated with exposure to Cr(VI) in these newer low-exposure facilities (SMR = 1.66, 95% CI = 1.08–2.46, using a combination of German national rates and U.S. state rates for comparison; SMR = 1.37, 95% CI = 0.89–2.03, using German and U.S. state rates).

Logistic regression analyses of the four-plant cohort found increased risk associated with increased cumulative (or lifetime) exposure to Cr(VI). In one analysis, the lung cancer mortality odds ratio among workers with highest annual exposure (≥ 5.8 ug/m3) was 20.2 (95% CI = 6.2–65.4), compared to the lowest exposure group (< 1.2 µg/m³) [ENVIRON report]. For the intermediate exposure group (1.2 µg/m³ – < 5.8 µg/m³), the odds ratio was 4.9 (95% CI = 1.5–16.0), also in comparison to the lowest exposure group [ENVIRON report]. Thus, the intermediate group includes exposure at levels only slightly higher than the 1 µg/m³ PEL proposed by OSHA in 2004, and showed elevated lung cancer risk at that level.

The final unpublished four-plant report reiterated the strength of the study design: "This study benefited from the multi-site design that provided a reasonably large cohort of post-change [lower exposure] chromium chemical workers, along with the corresponding increase in statistical power generally lacking in previous studies of post-change cohorts."

The published JOEM article, however, reports the mortality experience only of workers at the two U.S. plants studied by the ENVIRON researchers. After submitting the results to their sponsors in 2002, the authors evidently separated the German and U.S. results, despite their repeated emphasis in the protocol on the strength of the combined cohort. Instead of a positive result based on four plants, a negative two-plant study was published. In a response to a letter in the JOEM, the authors stated that the German component of the study had not been published because it was rejected by a journal to which it had been submitted, and defended the exclusion of the German data on the ground that different exposure measurements (air vs. urine) were used. This claim is not consistent with the need for large sample size to increase statistical power, as stated in the protocol and the final report (in which authors were comfortable combining air and urinary exposure measurements). In June 2005, we provided the study protocol and final report of the four-plant study  to OSHA.

On October 17, 2005, the ENVIRON researchers submitted the German component of the study to OSHA, accompanied by a letter stating that the paper had been accepted for publication in JOEM. In this manuscript, the ENVIRON researchers report that "lung cancer risk was elevated only in the highest exposure group (SMR = 2.09 95% CI = 1.08–3.65)".

The authors conducted another logistic regression analysis, but in this new version the estimate of relative risk for workers with high exposure is derived by comparing them to workers in the low and intermediate exposure groups combined. The result of this change is the disappearance of the statistically significant increase in lung cancer mortality risk among the intermediate group that was found in the unpublished final report. Tables 1 and 2 (see below), adapted from the unpublished final report (submitted by Applied Epidemiology, Inc. to the Industrial Health Foundation, September 27, 2002) and the pre-publication manuscript of the German component of the study submitted by the authors to OSHA, respectively, compare the results of the two regression analyses. In addition, while the elevation of the lung cancer SMR in the unpublished final report of the four-plant study was statistically significant, when the cohort was divided into two components, the lung cancer SMR was not statistically significant in either the German or U.S. components.

Table 1. Elevated lung cancer mortality risk in intermediate and high exposure groups in original unpublished study (adapted from Table 17 in final report)

Mean exposure to Cr(VI)*

Odds Ratio

95% CI

 Low (< 1.2 µg/m³)

Ref

--

 Intermediate (1.2- <5.8 µg/m³)

4.9

1.5-16.0

 High (> 5.8 µg/m³)

20.2

6.2-65.4

* Mean Exposure to Cr(VI) derived by dividing cumulative urinary chromium exposure
by 0.77 (conversion factor for air concentration), and then dividing by 45 years
(OSHA's working life assumption)

Table 2. Lung cancer mortality in intermediate group disappears after German component of study published separately, and low and intermediate exposure groups are combined (adapted from Birk et al. 2006)

Mean exposure to Cr(VI)*

Odds Ratio

95% CI

 Low and Intermediate (<5.8 µg/m³)

Ref

--

 High (> 5.8 µg/m³)

6.9

2.6-18.2

* Mean Exposure to Cr(VI) derived by dividing cumulative urinary chromium exposure
by 0.77
(conversion factor for air concentration), and then dividing by 45 years
(OSHA's working life assumption)

Current Status

On February 28, 2006 OSHA issued a final rule addressing occupational exposure to Cr(VI). The new standard is significantly weaker than the standard OSHA proposed in October of 2004; once the standard goes into effect (beginning in November 2006 for businesses with 20 or more employees; engineering controls need not be implemented until May 2010), the new PEL will be 5 µg/m³ rather than the level of 1 µg/m³ in the proposed rule.  (But the PEL for certain painting operations in the aerospace industry will be 25 µg/m³.) According to OSHA’s own estimates, exposure at 5 µg/m³ will result in 10 to 45 lung cancer deaths per 1,000 workers over a lifetime of exposure. In the final rule, OSHA justifies the weaker standard by arguing that a PEL of 5 µg/m³ reduces the significant risk posed to workers by occupational exposure to Cr(VI) “to the maximum extent that is technologically and economically feasible.” Many have criticized the new rule for a variety of other reasons, including: it eliminates training requirements and limits some employee notification of exposure monitoring results. Public Citizen has again sued OSHA over the standard.

On November 21, 2008 a three-judge panel of the U.S. Court of Appeals for the Third Circuit (including retired U.S. Supreme Court Justice Sandra Day O'Connor) heard oral arguments in two legal challenges to OSHA's rule.  One petition for review came from the  Edison Electric Institute.  They argued that OSHA did provide substantial evidence of risk of harm for workers employed in the electric power generation industry and therefore the health standard should not apply in their industry.  The second petition for review came from Public Citizen Health Research Group and the United Steelworkers Union who argued that the health standard OSHA issued failed to adequately reduce workers' risk of lung cancer and that a lower permissible exposure limit is economically and technologically feasible for the affected industries.  [52-page transcript of the oral arguments.]

On February 23, 2009 the U.S. Court of Appeals issued its decision, denying all of the petitioners' objections, except for Public Citizen's claim on the inadequacy of the notification requirement for exposed workers.

Conclusion

Faced with the threat of stronger OSHA regulation of workplace exposure of Cr(VI), a powerful carcinogen, the chromium industry initiated an effort to challenge the scientific evidence that the agency would likely use to justify a new standard. While criticizing OSHA for relying upon data from high-exposure cohorts, the chromium industry also commissioned a study of the mortality experience of workers at four plants with lower exposures, the results of which confirmed the elevated lung cancer risk in such workers. The consultants presented a final report to their chromium industry sponsors in 2002, but industry never provided OSHA a copy of the full four-plant study. Even when the agency specifically asked for precisely these sorts of data during its 2004–2005 rulemaking proceedings, the chromium industry and the authors remained silent.

For publication, industry-funded scientists divided this study into two components and published them separately. The first paper to be published was a statistically underpowered, negative study, the findings of which are being used by industry to attempt to reduce its regulatory burden. The second paper combined two exposure strata from the final report, resulting in the disappearance of the lower exposure stratum of particular regulatory interest in which a statistically significant finding was apparent in the unpublished final report. This allowed the industry trade associations to make the misleading assertion that elevated lung cancer mortality risk was only seen among workers with the highest exposure histories.

OSHA's statute instructs its decision makers to use the "best available evidence" [see Section 6(b)5 of P.L. 91-596] in the rulemaking process. The circumstances regarding these studies raise troubling questions about the ability of government to effectively issue rules protecting public health when studies are conducted, controlled and selectively published or provided to the rulemaking agency by the regulated industry [see Michaels 2005Michaels and Monforton 2005]. The entry of the German study into the OSHA record only after it was accepted for publication, months after the regulatory docket closed and years after data collection was complete, raises an important question for public health research: when regulatory proceedings are underway, should potentially important data be sequestered until the peer review process is complete? Many U.S. regulatory agencies, including the EPA and the Food and Drug Administration (FDA) rely heavily on unpublished studies, submitted by study sponsors, in reaching regulatory decisions. In this case, sponsors withheld data that OSHA requested during an active rulemaking process.

Public health is not well served by the unequal treatment of public and private science [Wagner and Michaels 2003]. Raw data from private studies should be as accessible to the public and regulators as that from publicly funded studies. Parties submitting scientific analyses and reports to the record should be required to disclose the true sponsorship of the study, including the original source of the sponsor's funding. Parties involved in the rulemaking process should also be required to certify that they have submitted all relevant data to the public record, whether or not those data have undergone peer review. Medical journals are increasingly willing to publish findings even if they have already been made available in another form. Regardless, public health rulemakings should not be based on partial records or limited by scientists' career concerns, particularly when lives hang in the balance.

About the Case Study

This case study was largely based upon a February 2006 paper published in Environmental Health, "Selected science: an industry campaign to undermine an OSHA hexavalent chromium standard," written by David Michaels, Celeste Monforton, and Peter Lurie.  They also submitted a Letter to the Editor of JOEM in response to the Birk et al paper published in 2006 in JOEM.

Reference (not linked in text)

[1] Spraycar M: Managing Editor, Journal of Occupational and Environmental Medicine: Email to David Michaels, May 16, 2005.

Additional Links and Resources

ATSDR, 2000. Toxicological Profile for Chromium.

CDC, National Institute for Occupational Safety and Health (NIOSH) Safety and Health Topic: Hexavalent Chromium.

Environmental Working Group report involving Cr(VI) in drinking water.

IARC, 1990, Chromium, Nickel and Welding, Volume 49, IARC Monographs on the Evaluation of Carcinogenic Risks to Humans.

Office of Environmental Health Hazard Assessment, California EPA, 1998. Public Health Goal for Chromium in Drinking Water.

Report on Carcinogens, Eleventh Edition; US Department of Health and Human Services, Public Health Service, National Toxicology Program. Chromium Hexavalent Compounds.

US EPA, 1998. Integrated Risk Information System (IRIS) Health Assessment for Chromium (VI).

US EPA, 1998. Toxicological Review of Hexavalent Chromium.

Update: November 2014: The U.S. Environmental Protection Agency issued a ruling and assessed a $2.57 million penalty against Elementis Chromium for violating Section 8(e) of TSCA.