Letter in JOEM: Lung Cancer Mortality in the German Chromate Industry, 1958 to 1998
This letter to the editor appeared in the Journal of Occupational & Environmental Medicine.
Michaels, David Ph.D., M.P.H.; Lurie, Peter M.D., M.P.H.; Monforton, Celeste M.P.H.. Lung Cancer Mortality in the German Chromate Industry, 1958 to 1998. Journal of Occupational & Environmental Medicine. 48(10):995-997, October 2006.
To the Editor: To put in accurate context the article by Birk et al that concludes that the study’s findings suggest “a possible threshold effect of occupational hexavalent chromium exposure on lung cancer,”[1] three points should be considered: 1) these data were withheld from a federal rulemaking proceeding, 2) these data were actually part of a larger study that refutes the authors’ conclusion, and 3) after examining the study, the U.S. Occupational Safety and Health Administration (OSHA) rejected the authors’ conclusion.
The issue of the carcinogenicity of hexavalent chromium (Cr[VI]) is one of great public health and regulatory interest. In October 2004, OSHA issued a proposed standard that would have greatly reduced workplace exposures to Cr(VI).[2] The proposed rule included a risk assessment that was based on data from workers with generally higher exposure levels than those seen in the Birk et al article. Throughout the rulemaking proceeding, OSHA asked for epidemiologic data on workers with lower exposure to Cr(VI), but neither the chromium industry nor the epidemiologists who conducted this study provided it to the agency at the public hearings or during the public comment period.[3]
In fact, the data presented by Birk et al were part of a larger multiplant study done for the chromium industry, completed more than 2 years before OSHA issued its proposed standard, but never published or given to OSHA by its authors.[3]That study, in which workers from four low-exposure facilities (two German facilities and two U.S. ones) were combined into one cohort reported significantly elevated lung cancer risk at both high and intermediate exposure levels, along with a clear dose–effect relationship[4](see Table 1). Despite repeated emphasis in the study protocol and the final report on the need to maximize statistical power by combining data from the four facilities,[4],[5] the investigators subsequently divided this study into two components and published two statistically underpowered studies.[1],[6] In their logistic regression in the second of these papers, Birk et al combined the low and intermediate exposure categories from the final report into a single referent category. The result was the disappearance of the finding of greatest regulatory importance: the increased risk of lung cancer among those in the intermediate exposure group whose exposures were well below the OSHA standard in effect at the time and close to the exposure limit OSHA was contemplating. The dose–effect relationship vanished as well, because only a single nonreferent exposure category remained (see Table 2).[3] In comments to OSHA, two industry groups cited the supposed lack of a positive finding at low exposures in the German data as evidence the proposed OSHA standard was unnecessarily stringent.[7],[8]
TABLE 1
Elevated Lung Cancer Mortality Risk in Intermediate and High Exposure Groups in Original Unpublished Study*
|
Cumulative Exposure to Cr(VI) |
OR‡ |
95% CI |
|
Low (<40 μg/L) |
Reference |
|
|
Intermediate (40 μg/L to <200/L) |
4.9 |
1.5–16.0 |
|
High (≥200 μg/L) |
20.2 |
6.2–65.4 |
* Adapted from Table 17 in Final report: collaborative cohort mortality study of four chromate production facilities, 1958-1998.[4]
‡ Odds ratio from logistic regression analysis.
CI indicates confidence interval.
TABLE 2
Increased Risk of Lung Cancer Mortality Risk Among Workers With Intermediate Exposure Disappears in JOEM Publication of German Component of Study*
|
Cumulative Exposure to Cr(VI) |
OR‡ |
95% CI |
|
Low and intermediate (<200 g/L) |
Reference |
|
|
High (≥200 μg/L) |
6.9 |
2.6–18.2 |
* Adapted from Birk et al.[1]
‡ Odds ratio from logistic regression analysis.
CI indicates confidence interval.
The authors justify their straying from their own protocol and dividing the study into two smaller, underpowered studies by noting that exposure history was estimated using airborne Cr(VI) levels in the U.S. facilities and urinary chromium levels in the German plants. In the final report, provided to the chromium manufacturing companies that paid for the study, these data were combined, using, ironically, a conversion factor[9] that is cited in the present paper when comparing the measured urinary Cr(VI) levels to OSHA’s permissible exposure limit.
It is important to note that OSHA has rejected Birk et al’s conclusion that “these data suggest a possible threshold effect” of Cr(VI) exposure on lung cancer. The agency’s scientists explained in the final standard that the Birk et al study’s “small cohort size, few lung cancer cases (eg, 10 deaths in the three lowest exposure groups combined) and limited follow up (average 17 years) severely limit the power to detect small increases in risk that may be present with low cumulative exposures.”[10]
The Occupational Safety and Health Act instructs the agency to use the “best available evidence.”[11] We believe that failure to provide the original study to OSHA, as well as the decision to bifurcate the study into two separate, underpowered publications, made after the final report was complete and the results were known, are inconsistent with the obligation of scientists to fully and promptly report findings of public health importance, even those that may trouble the sponsors of their study.
David Michaels, Ph.D., M.P.H.
Project on Scientific Knowledge and Public Policy
George Washington University School of Pubic Health & Health Services
Department of Environmental and Occupational Health
Washington, DC
Peter Lurie, M.D., M.P.H.
Public Citizen Health Research Group
Washington, DC
Celeste Monforton, M.P.H.
Project on Scientific Knowledge and Public Policy
George Washington University School of Pubic Health & Health Services
Department of Environmental and Occupational Health
Washington, DC
REFERENCES:
1 Birk T, Mundt KA, Dell LD, et al: Lung cancer mortality in the German chromate industry, 1958–1998. J Occup Environ Med 48. 426-433.2006.
2 Occupational Safety and Health Administration: Occupational exposure to hexavalent chromium, proposed rule. 69 Federal Register 59307 . October 4, 2004.
3 Michaels D, Monforton D, Lurie P: Selected science: an industry campaign to undermine an OSHA hexavalent chromium standard. Environ Health 5. 5.2006.
4 Applied Epidemiology, Inc. to the Industrial Health Foundation: Final report: collaborative cohort mortality study of four chromate production facilities, 1958–1998. September 27, 2002. Available at: https://www.citizen.org/sites/default/files/31.pdf Accessed June 5, 2006.
5 Epidemiological study of six modern chromate production facilities: a unified strategy for updating mortality experience through 1998: a draft proposal. Accessed June 5, 2006. Available at: https://www.citizen.org/sites/default/files/30.pdf
6 Luippold RS, Mundt KA, Dell LD, Birk T: Low-level hexavalent chromium exposure and rate of mortality among US chromate production employees. J Occup Environ Med 47. 381-385.2005.
7 Howe SR: The Society of the Plastics Industry Inc, letter to OSHA docket office. Accessed June 5, 2006. Available at: https://www.citizen.org/sites/default/files/36.pdf April 20, 2005
8 Richter CM, Hannapel JS: The Policy Group, post-hearing comments of the Surface Finishing Industry Council. Accessed June 5, 2006. Available at: https://www.citizen.org/sites/default/files/28.pdf April 20, 2005
9 Deutsche Forschungsgemeinschaft. Alkali chromates (Cr [VI]). In: Henschler D, Lehnert G, ed. Biological Exposure Values for Occupational Toxicants and Carcinogens: Critical Data Evaluation for BAT and EKA Values, VCH Verlagsgesellschaft Weinhein, Germany 1994: 187-203.
10 Occupational Safety and Health Administration : Occupational exposure to hexavalent chromium, final rule. 71 Federal Register 10180. February 28, 2006.
11 Section 6(b) 5, Occupational Safety and Health Act of 1970, Public Law 91-596.