Secondary Logo

Journal Logo

LETTERS TO THE EDITOR

Letter to the Editor

Misrepresentation by Egilman et al. of the Fordyce et al. (2019) Vermont Talc Miners and Millers Cohort Study Update

Fordyce, Tiffani A. PhD, MPH, MS; Mowat, Fionna S. PhD; Leonhard, Megan J. MPH; Moolgavkar, Suresh H. MBBS, PhD

Author Information
Journal of Occupational and Environmental Medicine: January 2020 - Volume 62 - Issue 1 - p e19–e21
doi: 10.1097/JOM.0000000000001784
  • Free

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.

Reply:

We deplore the tone of the Egilman et al1 letter questioning our (Fordyce et al2) scientific integrity. We also take strong exception to what appear to be intentional misrepresentations of our paper. For example, Egilman et al1 state, “Fordyce's et al2 study report [sic] 2 mesothelioma cases in a cohort of 427 workers.” We went to great lengths to find every case of mesothelioma in this cohort of workers. We found a single case over the entire period of observation, from 1930 through 2012. Egilman et al1 count a second case based on an article by Lamm and Starr,3 which was not published in a peer-reviewed journal and provided no information on the single case of mesothelioma that they claimed to have found. We were unable to verify this claim after an exhaustive search of the death certificates of workers in the expanded cohort of Vermont talc workers, as described in our paper. It is simply incorrect to say that we reported two cases of mesothelioma in this cohort, a false claim that Egilman et al1 repeat at least three times in their letter. All of the calculations in the Egilman et al1 letter based on two mesothelioma deaths are, therefore, incorrect. It is also incorrect to allege that we stated that the workers in the cohort had no occupational exposure to asbestos. We did not state this, and it is not true: the workers in the cohort, just like the general populations used for comparison, could have had occupational asbestos exposure elsewhere in their history. In fact, we explicitly noted that the death certificate for the worker with mesothelioma indicated that this individual was reported as having been exposed to asbestos.2

In a second misrepresentation, Egilman et al1 say, “Fordyce et al2 claimed that the studied talc mines were asbestos free.” This is a deliberate mischaracterization. What we actually said was that the talc was reported by the original authors (ie, Selevan et al4) as being free of both asbestiform minerals and significant quantities of free silica. This description is supported by other studies.5,6 The authors’ statement is currently disputed and we have no contribution to make to this dispute, one way or the other. Whether the talc contains asbestiform minerals or not is a moot point when considering the risk of mesothelioma in the cohort. We specifically noted, “some talc deposits are reported to naturally co-occur with deposits of asbestiform minerals, however, such that possible asbestos ‘contamination’ could complicate evaluation of the association between exposure to talc and development of mesothelioma. Due to the current heightened interest in a possible association between exposure to cosmetic talc and mesothelioma, we made a special effort to find any deaths due to mesothelioma in this cohort”.2 The results of our analyses show that there is no increased risk of mesothelioma in this expanded cohort of talc workers exposed to high levels of cosmetic talc with whatever contaminants the talc might or might not have contained.

Many of the criticisms leveled at us by Egilman et al1 make it evident that they do not understand the basic principles of occupational cohort standardized mortality ratio (SMR) studies. Their letter provides little evidence that any of the authors has either conducted or analyzed an occupational cohort study. As an occupational cohort is followed forward in time, the observed number of deaths from specific causes (as classified by the International Classification of Diseases [ICD] code in use at the time of death) is compared with an expected number of deaths derived from an appropriate comparison group. Many SMR studies span several decades, during which different ICD coding systems could be operational over the period spanned by the study. For example, our update of the Vermont talc workers’ study spanned the period 1930 through 2012. The ICD coding system was revised many times over that period, beginning with the fifth revision (ICD-5) and ending with the tenth (ICD-10), which was first implemented in 1999. The final observed counts of each cause of death and the expected numbers are based on mapping the distinct ICD codes onto the cause of death.

Egilman et al1 appear to have great difficulty with understanding the requirements for an appropriate comparison group for SMR occupational cohort studies. In SMR studies, it is customary to choose the general underlying population from which the cohort members are drawn as the comparison group. When local population mortality rates are not available, national mortality rates are often used. In our paper, we used both local (Vermont) rates and national (US) rates and found that both yielded similar results for our estimates of SMR. We presented results based only on national rates because confidentiality protections stipulated by the National Center for Health Statistics preclude the use of sub-national rates when either the observed or the expected number is less than 10. Because there can be strong temporal trends in both disease incidence and population structure over the course of an occupational cohort study, the rates in the comparison group must be contemporaneously adjusted to the distribution of age (and sex, if applicable) in the occupational cohort to derive the expected numbers. The reason for choosing local and national general populations as the comparison groups is that the members of the occupational cohort are drawn from the general population and would be expected to have the same employment profiles as the general population before and after membership of the occupational cohort under study. For example, the Vermont talc miners and millers would be expected potentially to work in occupations involving asbestos exposure elsewhere before and after membership in the occupational cohort. It is absurd to suggest, as Egilman et al1 do, that rates of asbestos-related diseases in the Vermont talc miners and millers should be compared with rates in populations that were never exposed to asbestos.

The design of occupational cohort SMR studies requires the existence of ICD codes for specific causes of death over the period of the study. Because ICD codes specific to mesothelioma were not available prior to ICD-10, expected numbers of mesothelioma deaths for comparison with observed numbers of deaths cannot be obtained from local or national death statistics before 1999. Surrogates for mesothelioma can be used, for example, most mesotheliomas are either pleural or peritoneal cancers, for which ICD codes were available throughout the period of our study. Therefore, observed numbers of deaths from pleural and peritoneal cancers can be compared with the expected numbers of deaths from such cancers. As we reported in our study,2 there were no deaths from pleural or peritoneal cancer throughout the entire follow-up period in the expanded Vermont cohort. Therefore, following this standard approach would have led to our reporting of zero cases of mesothelioma in the Vermont cohort.

We recognized, however, that under usual nosology practice, mesothelioma could be assigned to certain ICD codes other than pleural or peritoneal cancer prior to the implementation of ICD-10. Therefore, as described in our paper, we sent all of the death certificates within an expanded group of ICD codes for a second review by the nosologist. We compared the observed number of deaths in the broader group of ICD codes with expected number, and found no increased risk for these ICD codes. The nosologist discovered one case of mesothelioma not assigned to either the pleura or the peritoneum, but without ICD codes for mesothelioma during the majority of the study period, an expected number of mesothelioma deaths could not be generated. Therefore, we could not estimate the SMR for mesothelioma based on the usual procedure of using ICD codes in a SMR study, but we could estimate the SMR for the expanded set of ICD codes, where mesothelioma could have been included. Egilman et al1 suggest that we were up to some chicanery by pointing to the completely irrelevant example of the ocular tumor for which, with a single observation, we estimated a large SMR. In this case, unlike for mesothelioma, both the observed number and the expected number of ocular tumors were derived directly from certain ICD codes that were available for that particular tumor over the period of the study.

Given that we reported one case of mesothelioma in the expanded Vermont cohort, how can we determine whether or not this finding represents an increased risk? One approach is to use the expected number reported in a talc worker cohort of a similar size and era, recognizing the limitations of doing so. The Norwegian talc miners and millers cohort7 offers one possible comparison group because it is approximately the same size, includes talc workers employed in the mid-20th century, and has approximately the same number of years of follow-up as the expanded Vermont cohort. Another option is to use reliable national mortality rates for mesothelioma in the United States over the period of our study, that is, 1930 through 2012.

For mesothelioma, the expected number of deaths can be generated from incidence-based mortality data for mesothelioma in the Surveillance, Epidemiology, and End Results (SEER) data.8 SEER 9 is the most reliable database for population-based cancer incidence and mortality in the United States, covering the period 1975 to 2012, which is the relevant period for the single mesothelioma death reported by Fordyce et al.2 Over this period (1975 to 2012), the minimum incidence-based age-adjusted mesothelioma death rate among men in SEER 9, excluding the initial few years required for incidence-based mortality rates to become comparable to mortality rates,9 was approximately one per 100,000. This rate translates to an expected number of 0.17 mesothelioma deaths in the Vermont cohort, given that the number of person-years at risk in the Fordyce et al2 study is more than 17,000. The resulting SMR is 5.9, with a Fisher exact 95% confidence interval (CI) of 0.15 to 32.78 and, therefore, statistically insignificant. Over the same period, the maximum age-adjusted incidence-based mortality rate among men in SEER 9 was approximately two per 100,000. This translates to an expected number of 0.34 mesothelioma deaths, which leads to an SMR of 2.9 (Fisher exact 95% CI = 0.07 to 16.39), which is also statistically insignificant. One limitation of this approach is that the age-adjustment for mesothelioma death rates is not based on the age distribution of the Vermont talc workers cohort.

Even if an SMR analysis demonstrated an increased risk of mesothelioma in the Vermont talc workers cohort—counterfactually, as we do not acknowledge that there is any evidence of increased risk—further investigation would need to be performed before it could be concluded that such increased risk could be attributed to any exposure sustained while working as a member of the cohort. This is a fundamental tenet of occupational cohort studies, of which Egilman et al1 appear to be ignorant. For example, with other non-malignant respiratory disease in the expanded Vermont workers cohort, we found not only an increased risk, but also a positive exposure–response relationship with longer duration of employment, which indicates that the increased risk was likely attributable to exposures encountered in the Vermont talc mines and mills.

Finally, Egilman et al1 ask why we did not perform a power analysis before we undertook an update of the Vermont talc workers cohort. Their attention appears to be focused entirely on increased mesothelioma risk because that issue is front and center in current talc-related litigation. There are many other health issues that constitute a much greater disease burden to workers in the dusty trades than mesothelioma and respiratory cancer. The original study by Selevan et al4 already had more than adequate power to detect increased risks for these diseases. Therefore, our updated study most definitely had the power to detect these risks.

Collectively, the occupational cohort studies of talc miners and millers show quite convincingly that there is absolutely no increased risk of mesothelioma among these cohorts of workers. The Italian workers at Val Chisone have been followed up continuously from 1922 through 2013. Not a single case of mesothelioma has been reported.10–13 Similarly, no cases of mesothelioma have been reported in the other European studies of talc miners and millers.7,14–17

There is an old adage in toxicology that the dose makes the poison. Even if the risk of mesothelioma were marginally increased (and based on the data available to us, it is not) in the expanded Vermont cohort of talc miners and millers exposed to high levels of cosmetic talc (as demonstrated by the high rate of dust pneumoconiosis in the cohort), or in any other cohort of talc workers, what relevance would this have for end-users of cosmetic talc exposed to levels that are orders of magnitude lower? The exposure of an end-user is substantially lower in intensity and frequency, and clearly is not comparable to exposures experienced by talc miners and millers who worked intimately with raw talc. Regulatory agencies regard ionizing radiation as the quintessential non-threshold multisite carcinogen. There is convincing epidemiologic evidence that ionizing radiation increases the risk of mesothelioma.18–31 If every fiber of asbestos above background increases the risk of mesothelioma, then, theoretically, every quantum of radiation over background should as well. Therefore, diagnostic x-rays, ski vacations, and air travel should all increase the risk of mesothelioma.

Given the science, it is time for Egilman et al1 to question their cherished beliefs. It is probably too much to expect that they will abandon them.

REFERENCES

1. Egilman DE, Madigan D, Yiman M, Tran T. Letter to the Editor: Response to Vermont talc miners cohort study update. J Occup Environ Med 2019; Publish Ahead of Print.
2. Fordyce TA, Leonhard MJ, Mowat FS, Moolgavkar SH. A 37-year update on mortality patterns in an expanded cohort of Vermont talc miners and millers. J Occup Environ Med 2019; 61:916–923.
3. Lamm SH, Starr JA. Similarities in lung cancer and respiratory disease mortality of Vermont and New York State talc workers. Epidemiol 1988. 1576–1581.
4. Selevan SG, Dement JM, Wagoner JK, Froines JR. Mortality patterns among miners and millers of non-asbestiform talc: preliminary report. J Environ Pathol Toxicol 1979; 2:273–284.
5. Boundy MG, Gold K, Martin KP Jr. Lemen R, Dement JM, et al. Occupational exposures to non-asbestiform talc in Vermont. Dusts and Disease. Park Forest South, IL: Pathotox Publishers, Inc; 1979. 365–378.
6. Gamble JF, Gibbs GW. An evaluation of the risks of lung cancer and mesothelioma from exposure to amphibole cleavage fragments. Regul Toxicol Pharmacol 2008; 52: (1 suppl): S154–S186.
7. Wergeland E, Gjertsen F, Vos L, Grimsrud TK. Cause-specific mortality and cancer morbidity in 390 male workers exposed to high purity talc, a six decade follow-up. Am J Ind Med 2017; 60:821–830.
8. Surveillance, Epidemiology, and End Results (SEER) Program (www.seer.cancer.gov); 2019. SEER*Stat Database: Incidence-Based Mortality – SEER 9 Regs Research Data, Nov 2018 Sub (1975-2016) <Katrina/Rita Population Adjustment> – Linked To County Attributes – Total U.S., 1969-2017 Counties, National Cancer Institute, Division of Cancer Control and Population Sciences, Surveillance Research Program, released April 2019, based on the November 2018 submission.
9. Chu KC, Miller BA, Feuer EJ, Hankey BF. A method for partitioning cancer mortality trends by factors associated with diagnosis: an application to female breast cancer. J Clin Epidemiol 1994; 47:1451–1461.
10. Coggiola M, Bosio D, Pira E, et al. An update of a mortality study of talc miners and millers in Italy. Am J Ind Med 2003; 44:63–69.
11. Rubino GF, Scansetti G, Piolatto G, Romano C. Mortality study of talc miners and millers. J Occup Med 1976; 18:186–193.
12. Rubino GF, Scansetti G, Piolatto G, Gay G. Lemen RA, Dement JM. Mortality and morbidity among talc miners and millers in Italy. Dusts and Disease. Park Forest South, Illinois: Pathotox Publishers, Inc; 1979. 357–363.
13. Pira E, Coggiola M, Ciocan C, et al. Mortality of talc miners and millers from Val Chisone, northern Italy: an updated cohort study. J Occup Environ Med 2017; 59:659–664.
14. Léophonte P, Didier A. Bignon J. French talc pneumoconiosis. Health Related Effects of Phyllosilicates NATO ASI Series. Berlin: Springer-Verlag; 1990. 203–209. Vol G 21.
15. Léophonte P, Basset MF, Pincemin J, Louis A, Pernet P, Delaude A. Mortality of talc workers in France. Rev Fr Mal Resp 1983; 11:489–490.
16. Wild P. An epidemiologic study of mortality in an industry producing talc. INRS/EE Rapport TMT Version 6 (final) 2000. 1–73.
17. Wild P, Leodolter K, Refregier M, Schmidt H, Zidek T, Haidinger G. A cohort mortality and nested case-control study of French and Austrian talc workers. Occup Environ Med 2002; 59:98–105.
18. Travis LB, Hauptmann M, Gaul LK, et al. Site-specific cancer incidence and mortality after cerebral angiography with radioactive Thorotrast. Radiat Res 2003; 160:691–706.
19. Travis LB, Fossa SD, Schonfeld SJ, et al. Second cancers among 40 576 testicular cancer patients: focus on long-term survivors. J Natl Cancer Inst 2005; 97:1354–1365.
20. Tward JD, Wendland MMM, Shrieve DC, Szabo A, Gaffney DK. The risk of secondary malignancies over 30 years after the treatment of non-Hodgkin lymphoma. Cancer 2006; 107:108–115.
21. Deutsch M, Land SR, Begovic M, Cecchini R, Wolmark N. An association between postoperative radiotherapy for primary breast cancer in 11 National Surgical Adjuvant Breast and Bowel Project (NSABP) studies and the subsequent appearance of pleural mesothelioma. Am J Clin Oncol 2007; 30:294–296.
22. Teta MJ, Lau E, Sceurman BK, Wagner ME. Therapeutic radiation for lymphoma. Risk of malignant mesothelioma. Cancer 2007; 109:1432–1438.
23. Hodgson DC, Gilbert ES, Dores GM, et al. Long-term solid cancer risk among 5-year survivors of Hodgkin's lymphoma. J Clin Oncol 2007; 25:1489–1497.
24. De Bruin ML, Burgers JA, Baas P, et al. Malignant mesothelioma following radiation treatment for Hodgkin's lymphomas. Blood 2009; 113:3679–3681.
25. Goodman JE, Nascarella MA, Valberg PA. Ionizing radiation: a risk factor for mesothelioma. Cancer Causes Control 2009; 20:1237–1254.
26. Berrington de Gonzalez A, Curtis RE, Gilbert E, et al. Second solid cancers after radiotherapy for breast cancer in SEER cancer registries. Br J Cancer 2010; 102:220–226.
27. Farioli A, Violante FS, Mattioli S, Curti S, Kriebel D. Risk of mesothelioma following external beam radiotherapy for prostate cancer: a cohort analysis of SEER database. Cancer Causes Control 2013; 24:1535–1545.
28. Farioli A, Ottone M, Morganti AG, et al. Radiation-induced mesothelioma among long-term solid cancer survivors: a longitudinal analysis of SEER database. Cancer Med 2016; 5:950–959.
29. Schaapveld M, Aleman B, van Eggermond A, et al. Second cancer risk up to 40 years after treatment for Hodgkin's Lymphoma. N Engl J Med 2015; 373:2499–2511.
30. Schubauer-Berigan MK, Daniels RD, Bertke SJ, Tseng C-Y, Richardson DB. Cancer mortality through 2005 among a pooled cohort of U.S. nuclear workers exposed to external ionizing radiation. Radiat Res 2015; 183:620–631.
31. Chang ET, Lau EC, Mowat FS, Teta MJ. Therapeutic radiation for lymphoma and risk of second primary malignant mesothelioma. Cancer Causes Control 2017; 28:971–979.
Copyright © 2020 by the American College of Occupational and Environmental Medicine