Dutheil Et Al. Prostate Cancer and Asbestos: a Systematic Review and Meta-Analysis

n LETTER TO THE EDITOR Re: Dutheil et al. Prostate Cancer and Asbestos: A Systematic Review and Meta-Analysis

Paolo Boffetta, MD, MPH1,2; Nicolo` Franco, MD3; Attilia Gullino, MD, PHD4; Enrico Pira, MD5 Perm J 2020;25:20.215 E-pub: 12/30/2020 https://doi.org/10.7812/TPP/20.215

e letter has not been published or simultaneously sub- be based on stronger methodology than the one used by mitted for publication in whole or in part elsewhere. Dutheil et al.1 Sir, We read with interest the systematic review and meta- Disclosure Statement analysis by Dutheil et al.1 on occupational and residential Paolo Boffetta was involved in litigation on asbestos exposure and prostate exposure to asbestos and prostate cancer risk. We believe cancer risk. Enrico Pira has acted as Court-appointed expert witness and as ff consultant to parties (judge, prosecutor, and defendant attorney) in asbestos this study su ers from several methodological errors. litigations. First, the search of the literature appears to be deficient. We conducted a quick search, which we do not pretend Authors’ Contributions was systematic, and identified 10 studies that reported All authors contributed to the design of the study. Paolo Boffetta and Nicolo` results on risk of prostate cancer among asbestos exposed Franco drafted the manuscript; all authors reviewed and approved the manuscript. workers and were not included in the review by Dutheil et al.1 (Table 1; outdated results were included for 4 Funding No funds were obtained for the study. studies). Notably, the 2 studies that included the largest number of prostate cancer deaths (Lin et al., 2015 and How to Cite this Article Ferrante et al., 2017) were not included in the meta- Boffetta P, Franco N, Gullino A, Pira E. Re: Dutheil et al. Prostate cancer and analysis by Dutheil et al.1 asbestos: A systematic review and meta-analysis [Letter]. Perm J 2020;25:20.215. DOI: https://doi.org/10.7812/TPP/20.215 Second, Dutheil et al.1 double-counted some observations. In particular, the papers by Reid et al.2 and Armstrong et al.3 refer to the same cohort of Australian crocidolite References 1. Dutheil F, Zaragoza-Civale L, Pereira B, et al. Prostate cancer and asbestos: A systematic workers. review and meta-analysis. Perm J 2020 Feb;24:19.086. DOI: https://doi.org/10.7812/TPP/ ird, it is debatable whether cohorts of oil refinery 19.086 4 2. Reid A, Franklin P, Olsen N, et al. All-cause mortality and cancer incidence among adults workers, such as that studied by Tsai et al., should be exposed to blue asbestos during childhood. Am J Ind Med 2013 Feb;56(2):133-45. DOI: included in reviews on asbestos, since these workers are https://doi.org/10.1002/ajim.22103, PMID:22886909 3. Armstrong BK, de Klerk NH, Musk AW, Hobbs MS. Mortality in miners and millers of exposed to other potential carcinogens. In any case, if one crocidolite in Western Australia. Br J Ind Med 1988 Jan;45(1):5-13. DOI: https://doi.org/10. particular cohort of oil refinery workers is included, it is not 1136/oem.45.1.5, PMID:2829959 4. Tsai SP, Waddell LC, Gilstrap EL, Ransdell JD, Ross CE. Mortality among maintenance clear why the many other cohorts from the same industry are employees potentially exposed to asbestos in a refinery and petrochemical plant. Am J excluded. Ind Med 1996 Jan;29(1):89-98. DOI: https://doi.org/10.1002/(sici)1097-0274(199601)29: 1<89::aid-ajim11>3.0.co;2-w, PMID:8808046 Fourth, we think that it is a mistake to combine results of 5. Thompson IM, Ankerst DP. Prostate-specific antigen in the early detection of prostate prostate cancer incidence and mortality. On the one hand, cancer. Can Med Assoc J 2007 Jun;176(13):1853-8 DOI: https://doi.org/10.1503/cmaj. 060955, PMID:17576986 prostate cancer incidence is highly sensitive to the imple- 6. Acheson ED, Gardner MJ, Winter PD, Bennett C. Cancer in a factory using amosite mentation of screening with prostate specific antigen.5 On asbestos. Int J Epidemiol 1984;13(1):3-10. 7. Selikoff IJ, Seidman H. Asbestos-associated deaths among insulation workers in the the other hand, mortality from prostate cancer depends on United States and Canada, 1967-1987. Ann N Y Acad Sci 1991;643:1-14. access to effective treatment. ese 2 factors might vary 8. Sanden´ A, Jarvholm¨ B, Larsson S, Thiringer G. The risk of lung cancer and mesothelioma after cessation of asbestos exposure: A prospective cohort study of shipyard workers. Eur between asbestos-exposed workers and the national or re- Respir J 1992;5(3):281-5. gional populations that are used as reference in the studies. ese sources of bias require careful consideration, and Author Affiliations combining the 2 measures of occurrence, whose biases 1Stony Brook Cancer Center, Stony Brook University, Stony Brook, NY might act in opposite directions, generates confusion in the 2Department of Medical and Surgical Sciences, University of Bologna, Bologna, Italy 3Department of Public Health Sciences and Pediatrics, University of Turin, Turin, Italy results. 4University Hospital City of Health and Science of Turin, Turin, Italy For these reasons, we think that the results of the meta- 5Department of Public Health Sciences and Pediatrics, University of Turin, Turin, Italy analysis by Dutheil et al.1 are not interpretable and their conclusion that asbestos exposure is associated with prostate Corresponding Author Paolo Boffetta, MD, MPH (paolo.boff[email protected]) cancer is not justified. A careful review of this issue should Keywords: asbestos, bias, epidemiology, prostate cancer The Permanente Journal·https://doi.org/10.7812/TPP/20.215 The Permanente Journal·For personal use only. No other uses without permission. Copyright © 2020 The Permanente Press. All rights reserved. 1 LETTER TO THE EDITOR Letters to the Editor

Table 1. Selected studies reporting results on risk of prostate cancer among asbestos exposed workers that were not included in the review by Dutheil et al.1 Study Industry Asbestos type Country Outcome Sex N Obs SMR 95% CI Acheson et al., 1984a,6 Insulation product mft Am UK Mo M 4820 2 0.94 0.11-3.40 Selikoff and Seidman, 1991a,7 Insulation workers Mix US, Canada Mo M 17800 59 1.12 0.85-1.45 Sanden et al., 19928 Shipyard workers P Ch Sweden Mo M 3893 22 0.80 0.50-1.21 Ulvestad et al., 20029 Cement workers P Ch Norway In M 541 8 0.6 0.3-1.1 Lin et al., 201510 Mixed Mix Taiwan In M 121883 307 0.67 0.59-0.75 Levin et al., 201611 Insulation product mft Am USA Mo PM 1130 6 0.78 0.29-1.70 Pira et al., 2016a,b,12 Textile product mft Mix Italy Mo M 894 7 0.74 0.38-1.94 Ferrante et al., 201713 Mixed Mix Italy Mo M 46060 352 0.97 0.87-1.08 Pira et al., 2017a,b,14 Miners Ch Italy Mo M 1056 9 0.91 0.42-1.74 Rusiecki et al., 2018a,b,15 Shipyard workers NA USA Mo PM 4702 281 0.85 0.57-1.23 Results in italics are derived from raw data reported in the original publications. a An earlier report of this cohort was included in the review by Dutheil et al.1 b Workers at high asbestos exposure. Am = amosite; Ch = chrysotile; CI = conﬁdence interval; In = incidence; M = male; mft = manufacture; Mix = mixed ﬁbers; Mo = mortality; N = number of workers; NA = not available; Obs = observed cases or deaths; P Ch = predominantly chrysotile; PM = predominantly (> 90%) male; SMR = standardized mortality ratio (standardized incidence ratio for studies on prostate cancer incidence).

9. Ulvestad B, Kjaerheim K, Martinsen JI, et al. Cancer incidence among workers in the asbestos- 13. Ferrante D, Chellini E, Merler E, et al. Italian pool of asbestos workers cohorts: Mortality cement producing industry in Norway. Scand J Work Environ Health 2002;28(6):411-7. trends of asbestos-related neoplasms after long time since ﬁrst exposure. Occup Environ 10. Lin CK, Chang YY, Wang JD, Lee LJ. Increased standardised incidence ratio of malignant Med 2017;74(12):887-98. pleural mesothelioma in Taiwanese asbestos workers: A 29-year retrospective Cohort 14. Pira E, Romano C, Donato F, Pelucchi C, Vecchia C, Boffetta P. Mortality from cancer and study. BioMed Res In 2015. other causes among Italian chrysotile asbestos miners. Occup Environ Med 2017;74(8): 11. Levin JL, Rouk A, Shepherd S, Hurst GA, McLarty JW. Tyler asbestos workers: A mortality update 558-563. in a cohort exposed to amosite. J Toxicol Environ Health B Crit Rev 2016;19(5-6):190-200. 15. Rusiecki J, Stewart P, Lee D et al. Mortality among coast guard shipyard workers: A 12. Pira E, Romano C, Violante FS, et al. Updated mortality study of a cohort of asbestos retrospective cohort study of speciﬁc exposures. Arch Environ Occup Health 2018;73:1, textile workers. Cancer Med 2016;5(9):2623-8. 4-18.

In Response

To the Editor: We thank Boffetta et al.1 as one line in a table in those articles. The Lastly, we also want to point out that we for their relevant comments concerning only way to systematically retrieve the 10 included twice more articles than the our recent systematic review and meta- suggested articles would have been to preceding metaanalysis of Peng et al.3 analysis on asbestos exposure and prostate broaden the keywords by removing the published quite simultaneously with our cancer.2 “prostate” keyword. Using our algorithm metaanalyses. First, Boffetta et al.1 identified putatively (ie, derivates from “prostate cancer” and Second, we did not double-count the 10 missing articles in our meta-analyses “asbestos”), we had a putative number of same observations. Armstrong et al.4 studied (comprising 30 articles). However, we 2547 articles to include in our meta- occupational asbestos exposure in miners, want to acknowledge that all the suggested analysis. Broadening our algorithm without whereas Reid et al.5 described environ- articles are not retrieved in any databases “prostate” (ie, derivates from “cancer” and mental exposure in children living around using our search strategy. Our algorithm “asbestos”), the number of putative arti- the mine. Moreover, the studies were was (prostate cancer OR prostatic neo- cles to include would rise to more than published in 19884 and 20135 (ie, with a plasm) AND (asbestos* OR crocidolite* 25,000 (10 times more), which is clearly 25-year gap). OR chrysotile* OR amphibole* OR amo- nonmanageable. Boffetta et al.1 are experts Third, concerning the inclusion of sev- site*). All the 10 suggested articles did not on asbestos exposure, and we are thankful eral type of workers (eg, miners, shipyard describe prostate cancer within their title from their comments. However, articles workers, firefighters, textile industry) and or their abstract. They were impossible to were probably retrieved from their knowl- particularly oil refinery workers (ie, main- detect using our algorithm, as those arti- edge without using our algorithm, but this is tenance workers exposed to asbestos con- cles did not focus at all on prostate not the way to compute a metaanalysis tained in insulation materials), all of these cancer—prostate cancer is only described using a reproducible search strategy. occupations were exposed to several

2 The Permanente Journal·For personal use only. No other uses without permission. Copyright © 2020 The Permanente Press. All rights reserved. The Permanente Journal·https://doi.org/10.7812/TPP/20.215 LETTER TO THE EDITOR Letters to the Editor

pollutants and toxics (benzenes, metal Clermont–Ferrand, CHU Clermont– Fares Moustafa, MD, PhD9 fumes, petroleum derived, radiations, Ferrand, Occupational and 9Universite´ Clermont Auvergne, smokes, and carbon residues), rendering Environmental Medicine, WittyFit, University Hospital of Clermont- difficult robust conclusions on the specific Clermont–Ferrand, France Ferrand (CHU), Emergency link between the asbestos exposure and 2Faculty of Health, School of Exercise department, Clermont-Ferrand, France prostate cancer. To avoid exposition bias, Science, Australian Catholic University, we chose to include all workers or residents Melbourne, Victoria, Australia Valentin Navel, MD10 exposed to asbestos particles. There is a 10Universite´ Clermont Auvergne, whole section discussing this limitation in Laetitia Zaragoza-Civale, MD3 CNRS, INSERM, GReD, Translational the dedicated section of our article. 3University Hospital of Clermont Approach to Epithelial Injury and Repair, Fourth, we totally agree that it might be Ferrand, CHU Clermont–Ferrand, CHU Clermont-Ferrand, University a bias to mix incidence data and mortality. Occupational and Environmental Hospital of Clermont-Ferrand, This is why we computed a sensitivity Medicine, Clermont-Ferrand, France Ophthalmology, Clermont-Ferrand, analyses by type of risk (distinguishing France mortality and incidence data): “Stratified Bruno Pereira, PhD4 results by type of risk demonstrated an 4University Hospital of Clermont increased risk of prostate cancer with stan- Ferrand, CHU Clermont–Ferrand, Disclosure Statement The author(s) have no conflicts of interest to dardized incidence ratio (effect size = 1.16, Clinical Research and Innovation disclose. 95% CI = 1.04-1.27) and with standardized Direction, Clermont-Ferrand, rate ratio (1.06, 1.04-1.27), whereas there France Funding was no evidence of an increased risk with This study was funded by the University Hospital of standard mortality ratio and hazard ratio Martial Mermillod, PhD5,6 Clermont-Ferrand, France. (1.09, 0.98-1.19; and 0.79, 0.45-1.13, 5Univ. Grenoble Alpes, Univ. Savoie respectively).” Mont Blanc, CNRS, LPNC, Grenoble, References In total, we understand the comments France 1. Boffetta P, Franco N, Gullino A, Pira E. Re: Dutheil et al. 1 6 Prostate cancer and Asbestos: a systematic review and from Boffetta et al., even if answers were Institut Universitaire de France, Paris, meta-analysis. Perm J 2020 Jul. already within the articles. Metaanalysis France 2. Dutheil F, Zaragoza-Civale L, Pereira B, et al. Prostate follows a very logical methodology and Cancer and asbestos: A systematic review and meta- 7 analysis. Perm J Feb 2020 Feb;24:19.086. DOI: https:// has its own rules that some less- Julien S Baker, PhD doi.org/10.7812/TPP/19.086 7 experienced readers might not be ac- University of the West of Scotland, 3. Peng R, Fang F, Chen Z, et al. Does exposure to customed to. Our conclusions seem Institute for Clinical Exercise and Health asbestos cause prostate cancer? A systematic literature review and meta-analysis. Medicine (Baltim) 2019 Jan; perfectly salient and careful; that is, Science, Hamilton, UK 98:e14108. DOI: https://doi.org/10.1097/MD. “Asbestos exposure “seems to” increase 0000000000014108 the risk of prostate cancer.” v Jeannot Schmidt, MD8 4. Armstrong BK, de Klerk NH, Musk AW, Hobbs MS. 8 Mortality in miners and millers of crocidolite in Western Universite´ Clermont Auvergne, CNRS, Australia. Br J Ind Med 1988 Jan;45(1):5-13. DOI: https:// Fred´ eric´ Dutheil, MD, PhD1,2 LaPSCo, Physiological and Psychosocial doi.org/10.1136/oem.45.1.5 1Universite´ Clermont Auvergne, CNRS, Stress, University Hospital of Clermont- 5. Reid A, Franklin P, Olsen N, et al. All-cause mortality and cancer incidence among adults exposed to blue asbestos LaPSCo, Physiological and Psychosocial Ferrand (CHU), Emergency during childhood. Am J Ind Med 2013 Feb;56(2):133-45. Stress, University Hospital of department, Clermont-Ferrand, France DOI: https://doi.org/10.1002/ajim.22103