Readers are invited to submit letters for publication in this department. Submit letters online at http://joem.edmgr.com. Choose “Submit New Manuscript.” A signed copyright assignment and financial disclosure form must be submitted with the letter. Form available at http://www.joem.org under Author and Reviewer information.
To the Editor:
Fordyce et al. attempted to update the Vermont cohort study published by the National Institute for Occupational Safety and Health (NIOSH) in 1979.1,2 The evaluation is limited due to Fordyce's et al mischaracterization of the asbestos exposure, choice of inappropriate expected mesothelioma rates, and inadequate power.2 When properly evaluated, the data provide evidence that talc miners exposed to asbestos had an elevated risk of mesothelioma.
PRESENCE OF ASBESTOS IN VERMONT TALC MINES
Fordyce et al claimed the studied talc mines were asbestos-free. This is not the case.2 The original draft of the NIOSH report stated that multiple air samples from at least one mine contained chrysotile fibers. Johnson & Johnson (J&J) was aware of the fact that NIOSH deleted this asbestos finding from the report.3 As a witness in J&J talc litigation, Dr Moolgavkar also had access to these test results, which J&J posted on the internet in October 2018.4
Neither NIOSH nor Fordyce et al identified the mines or mills studied.1,2 From the mid-1960s through 1993, Engelhard (now BASF) and J&J (1966 to 1968) owned the Johnson mine and Mill, and J&J owned Hammondsville, Argonaut, Chester, Hamm, Rainbow, a few smaller mines and the Windsor Mill, all located in Vermont.5,6 J&J and Engelhard contemporaneous testing of these mines revealed chrysolite, anthophyllite, tremolite, and actinolite in various tests.7 J&J also found chrysotile and tremolite in its commercial “cosmetic talc” products sourced from Vermont.7 Dr Moolgavkar was also aware of J&J's subsidiary Windsor Minerals’ finding of asbestos in Vermont talc ore.8
The funders of this study, Kirkland and Ellis, represent Johnson & Johnson and BASF; thus they had access to Engelhard testing and are aware of the fact that Cahill and Gordon, BASF previous law firm, destroyed tests that had reported asbestos in Johnson mine talc.9 Kirkland and Ellis are also aware that J&J tests found asbestos and fibrous talc in Vermont talc ores used in cosmetic products.10 It is important to note that the asbestos content of the ores is below 0.1%, and thus any cohort study of exposed workers or consumers would need to be large to find any cancer effect.
APPLYING THE PROPER BACKGROUND RATE OF MESOTHELIOMA REVEALS AN INCREASED RATE IN TALC WORKERS
Fordyce et al reported finding two mesothelioma cases in the cohort.2 One case was documented by their study, and another was previously published but not discovered in their study.11
In conducting their statistical analysis, Fordyce et al use a study of Norwegian talc miners to generate an expected rate of 0.6 mesotheliomas or 3.8 per 100,000 person-years to reflect the Norwegian cohort of 15,687.2 Fordyce's discovery of two mesotheliomas in 17,170 person-years would result in a rate of 11.65 per 100,000 person-years, in a cohort they allege had no occupational asbestos exposure.2 In contrast, the “unexposed” background rate of mesothelioma in the United States, which never accounted for talc or asbestos exposure, is estimated to be 1 to 2 per million.12 The national rate for both “exposed” and “unexposed” is only two per 100,000.13 Fordyce's et al expected rate is near twice the general population's expected rate.2 This is further evidenced by calculating the new expected rates, standard mortality ratio (SMR)'s, confidence intervals, and probability values; except for the expected rate, these were not provided by Fordyce et al.2 The general population rate of two per 100,000, means the expected rate of mesothelioma in the Vermont cohort is 0.3434, with a 5% chance of getting two or more cases. At this rate, Fordyce's et al2 observation of two mesothelioma cases gives an SMR of 5.824, 95% CI (0.976, 19.24), P = 0.052. Using the unexposed background rate of two per million, the chance of observing two cases of mesothelioma is 0.0006 and gives an SMR of 58.24, 95% CI (9.76, 192.4), P < 0.001. Both these expected rates are smaller than the 0.6 expected cases provided by Fordyce et al.2
The effect of Fordyce's et al use of an inaccurate expected mesothelioma rate is evident in the analysis of eye cancer. The national rate for eye cancer, 1.02 per 100,000, is half the mesothelioma rate (two per 100,000).14 Fordyce et al only observed one case of eye cancer, but they still concluded, “Cancer of the eye (based on a single case) …[was] also statistically significantly elevated for millers.”2 If Fordyce et al. had utilized an appropriate expected rate for mesothelioma, the rate of mesothelioma was also statistically significantly elevated.2
Fordyce et al. provided another justification for not finding a statistically significant increase in rates of mesothelioma: “Because of the small size of this study, there could be concern that it lacked adequate power to detect an increased risk of rare diseases, such as malignant mesothelioma, despite more than seven decades of follow-up.”2 Fordyce et al admitted that a post-hoc analysis showed that their study did not have adequate power but then commented that “post-study power calculations are considered irrelevant,” citing Greenland et al.2,15 However, Greenland et al. and other statisticians have not questioned post-hoc power calculations nor have they discounted their importance.15–19 Fordyce et al failed to explain why they did not conduct a pre-study power analysis to determine statistical significance.2
UNHEALTHY WORKPLACE EFFECT
Fordyce et al report that talc worker mortality was 30% higher than the US population rate.2 Most of this increase was due to non-malignant respiratory disease that appears to be related to silica exposure. As early as 1918, silicate mining deaths were linked to an increased incidence of respiratory disease and tuberculosis.20 Mesothelioma has a long latent period, and rates will not increase if workers die prematurely from other work-related causes, including injuries (accidental death) and/or lung disease.21 Since the workers would be deceased before the latent period is achieved, the actual rate is lower than the expected rates. We term this the “unhealthy workplace effect.” Due to this effect, the excess mesothelioma rate for cohort studies can be underestimated in epidemiology studies.
When trying to determine health effects, claims of insignificant increased risk can result in an undisclosed health hazard. Greenland et al. stated, “Too often we weaken our capacity to interpret data and to take reasonable decisions whatever the value of P15. And far too often we deduce ‘no difference’ from ‘no significant difference.’”10 Hoenig et al. warned that “In matters of public health and regulation, it is often more important to be protected against erroneously concluding no difference exists when one does.”22 Fordyce's et al study report two mesothelioma cases in a cohort of 427 workers.2 We conducted a statistical analysis of these findings based on national exposed and the unexposed rate; the results contradict their assertion that talc ore exposure does not increase the risk of mesothelioma. This study provides evidence that exposure to talc ore, which contains asbestos and fibrous talc, increases the risk of contracting mesothelioma.
1. Selevan SG, et al. Mortality patterns among miners and millers of non-asbestiform talc: preliminary report. J Environ Pathol Toxicol
2. Fordyce TA, et al. A 37-year update on mortality patterns in an expanded cohort of Vermont Talc Miners and Millers. J Occup Environ Med
3. Lee G. NIOSH-Vermont Talc Environmental Study; 1977. Available at: Available at: https://repository.library.brown.edu/studio/item/bdr:958146/
. Accessed November 18, 2019.
4. (J&J). J.J. Facts About Talc. 2018 October 11; 2019. Available at: https://www.factsabouttalc.com/litigation
. Accessed November 18, 2019.
5. Johnson & Johnson (J&J). Material Specification Luzenac America, Inc. Windsor Grade 66 Talc. Royston, Kolmar, Silliker (NJ); 1997. Available at: https://repository.library.brown.edu/studio/item/bdr:958145/
. Accessed November 18, 2019
6. Miller R. Plaintiff's Videotaped Deposition of Roger Miller, in Consuelo Lopez vs. ACandS, Inc. 2007: Superior Court of the State of California, County of San Francisco. Available at: https://repository.library.brown.edu/studio/item/bdr:841467/PDF/
. Accessed November 18, 2019.
7. Steffen JE, et al. Presence of asbestos in consumer talc products: evaluating a “zero tolerance” policy. APHA
8. Schmitz P, Johnson and Johnson V. Reporter's Transcripts of Proceedings; 2019. Available at: http://www.asbestosandtalc.com/
. Accessed October 17, 2019.
10. Ellis K. J&J Staves Off Asbestos Baby Powder Suit In Calif. Jury Win; 2018. Available at: https://www.kirkland.com/news/in-the-news/2018/11/jj-staves-off-asbestos-baby-powder-suit
. Accessed October 17, 2019.
11. Lamm SH, Starr JA. Similarities in Lung Cancer and Respiratory Disease Mortality of Vermont and New York State Talc Workers. Proceedings of the VIIth: International Pneumoconioses Conference Part I, 1990. VII: pp. 90–108.
12. Teta MJ, Mink EL, Lau E, Sceurman BK, Foster ED. US mesothelioma patterns 1973-2002: indicators of change and insights into background rates. Eur J Cancer Prev
13. Surveillance E, End Results (SEER). Mesothelioma Age-Adjusted SEER Cancer Incidence Rates, 1980-2008 By Race, Sex and Year of Diagnosis, N.C. Institute, Editor; 2008.
14. American Cancer Society (ACS), Key Statistics for Eye Cancer; 2019.
15. Greenland S, Senn SJ, Rothman KJ, Carlin PC, Goodman SN, Altman DG. Statistical tests, P values, confidence intervals, and power: a guide to misinterpretations. Eur J Epidemiol
16. Rosner B. Fundamentals of Biostatistics; 2010.
17. Richard D, De Veaux PV, David E. Bock, Stats: Data and Models; 2012.
18. Winer BJ, Donald RB, Michaels KM. Statistical Principles in Experimental Design. Vol. 3; 1991, New York: McGraw-Hill.
19. Thomas Chalmers HS, Blackburn B, Silverman B, Schroeder B, Reitman D, Ambroz A. A method for assessing the quality of a randomized control trial. Controlled Clinical Trials
20. Hoffman F. Mortality from Respiratory Disease in Dusty Trades. Industrial Accidents and Hygiene Series. Vol. 17; 1918. U.S Department of Labor, Bureau of Labor Statistics.
21. Mazureck J. Malignant Mesothelioma Mortality — United States, 1999-2015, in Morbidity and Mortality Weekly Report. U.D.o.H.a.H.S.C.f.D.C.a. Prevention, Editor; 2017.
22. Hoenig JM, Dennis MH. The abuse of power: the pervasive fallacy of power calculations for data analysis. Am Stat