Styrene is an important chemical used in making polymers for a wide range of applications, including plastics, latex paints, coatings, insulation, synthetic rubbers, and polyesters. In 2007, 20 million tons of styrene were produced worldwide. The U.S. National Institute for Occupational Safety and Health estimates that >1 million workers in the United States have potential occupational exposure.1 Some epidemiologic studies have reported increased risk of lymphohematopoietic cancers, lung cancer, or pancreatic cancer among workers exposed to styrene, although findings across studies are not consistent.2–4 The International Agency for Research on Cancer classifies styrene as possibly carcinogenic to humans based on limited evidence in both humans and animals.5
Studies of styrene-exposed workers can be classified into three major manufacturing segments, each with its own exposure pattern: styrene-butadiene rubber products,6–11 reinforced plastics from styrene,12–19 and styrene manufacture.2,20–22 The studies of workers in the reinforced plastic industry are considered the most informative because styrene exposure is higher, confounding from other potentially carcinogenic exposures is uncommon, and the study populations are large.3–5,23,24 Four studies have examined cancer risk among workers in glass-fiber reinforced plastic industry: the U.S. industry-wide study,12,13 the European industry-wide study,15 the Danish industry-wide study,16,17 and a study of two sites in Washington State.19 The U.S. industry-wide study is one of the larger studies with well-documented styrene exposures. We have updated this study with 19 additional years of follow-up to determine if styrene exposure is related to increased cancer risk, especially for cancers of the lymphatic and hematopoietic systems, lung, and pancreas.
The original study examined 15,908 workers at 30 U.S. reinforced plastic facilities in the states of Alabama, California, Florida, Illinois, Indiana, Massachusetts, Michigan, North Carolina, Nebraska, New Jersey, New York, Ohio, Oklahoma, Pennsylvania, Texas, and Washington.12 None of these plants has been previously included in another styrene study. A previous update of this study followed vital status to 31 December 1989 and reduced the number of study participants to 15,826 workers by eliminating 30 duplicate records and removing 52 workers exposed to styrene for <6 months.13 This study uses the same 15,826 workers (Table 1). Review and oversight were provided by a human subjects review board in Midland, MI.
This study extended vital-status follow-up to the end of 2008, using death certificates from the previous studies, the National Death Index (NDI) including NDI-plus, the Social Security Death Index, a credit bureau (Trans Union), and a web site (www.ancestry.com). We used death-certificate cause-of-death codes from the NDI when an exact match was identified for more recent decedents. For less certain NDI matches, we examined death-certificate information obtained from the state in which the death occurred to determine matching status. Causes of death on these death certificates were coded by a nosologist according to the revision of the International Classification of Diseases (ICD) in effect when the death occurred. Person-years at risk were accumulated from 1 January 1948 or from the date of first exposure to styrene, whichever was later, and ended at the study end date, the date lost to follow-up, or date of death, whichever was earliest.
There were 4280 workers employed at the end of the original study in 1977. We did not add more recent work-history information or corresponding exposure estimates for these workers. The small amount of exposure information to be gained from updating these work histories did not justify the cost of collection; furthermore, many of the plants included in the study are now closed, making complete collection of additional work-history information impossible. In our study, the average exposure in 1977 was 25 parts per million (ppm) vs. 35 ppm a decade earlier. After 1977, the exposure levels in the reinforced plastics industries declined further.25 Thus, the additional cumulative exposure for the workers employed at the end of the study would have been relatively small.
Construction of the exposure estimates has been discussed previously.12,13 Briefly, a survey team of industrial hygienists and chemical engineers visited each plant to examine engineering controls, work practices, personal protective equipment, and past industrial hygiene monitoring, as well as to collect historical information regarding process change which may have impacted styrene exposures.5,26 In addition to the routine exposure monitoring, a detailed industrial hygiene survey was conducted at each site to provide a basis for exposure estimation. Employment records were used to reconstruct work history for each worker. Area/job titles across the plants were grouped when the jobs had similar tasks and exposure potential. This information was used by survey team to construct styrene exposure estimates across plants and over time.26 Some details of the exposure estimation, such as the number and range of exposure samples, are no longer available, and so the accuracy of the exposure estimates is uncertain. Nevertheless, the historical exposure estimates in this study are consistent with other styrene exposures in the reinforced plastics industry in the United States and Europe.5
We examined four measures of dose: cumulative exposure, peak exposure, duration of exposure, and average exposure.27 For cumulative exposure, we used the time-weighted average for an 8-hour work day. The average time-weighted average for all exposures was 28 ppm. It has been argued that peak exposure may be more relevant measure of risk for styrene.28 For peak exposure, we used a 100-ppm level for 15 minutes of the working day and calculated cumulative days worked above that peak. We chose 100 ppm because it is the lowest level of styrene exposure at which irritation from styrene occurs.5 The average number of peaks across workers was 113. More than half of study members experienced no peak exposures, and only 6% had >5 years of cumulative peak exposures. We calculated cumulative duration as styrene-exposed work time. The mean duration of exposure to styrene was 4.3 years. The arithmetic mean of past exposures (average exposure) was obtained by dividing total cumulative exposure by total cumulative duration. The mean average exposure to styrene was 26.7 ppm. Latency was calculated as the time since first exposure to styrene.
Several concomitant exposures could have been present in some jobs and areas, but typically only occasionally and at low levels.26 Asbestos was once used in some reinforced plastics to provide durability and heat resistance. We had no data on potential asbestos exposure levels or area-specific usage patterns, but we were able to compare plants with and without asbestos use as a crude indicator of potential workplace exposure. Nineteen of the 30 plants reported using asbestos.
The previous studies of these reinforced plastic workers reported an increased risk of lung cancer.12,13 A nested case-control study examined some concomitant workplace exposures and cigarette smoking as potential causes of this excess.12 Although a nested case-control study was beyond the scope of the current effort, we examined details of lung cancer deaths and other deaths commonly related to cigarette smoking including bladder cancer; kidney cancer; bronchitis, emphysema, and asthma; and heart disease.29
The Occupational Cohort Mortality Analysis Program was used to calculate standardized mortality ratios (SMRs) and 95% confidence intervals (CIs).30 We used the United States for comparison, adjusting for differences in sex, age distribution, and time interval. Among 39 cause-of-death categories, we present SMRs for lung cancer; mesothelioma; and bronchitis, emphysema, and asthma for sites reporting use or nonuse of asbestos in their products. For several diseases such as leukemia subtypes, non-Hodgkin lymphoma, and Hodgkin lymphoma, which were not recognized in early ICD revisions, risk was calculated for only those years when the relevant cause of death seems as a distinguishable code in the ICD. For causes of death potentially related to styrene exposure (cancers of the lymphatic and hematopoietic system, lung, and pancreas), cigarette smoking (heart disease; bronchitis, emphysema, and asthma; and cancers of lung, kidney, and bladder), or causes of death with significant excesses, we stratified the study population using cumulative exposure categories of 0.0–149.9, 150–399.9, 400–1199.9, and 1200+ ppm-months and cumulative duration of exposure categories of 0.0–11.9, 12–35.9, 36–95.9, and 96+ months to examine trends in SMRs by exposure. The exposure categories were chosen to approximate quartiles of the number of deaths from all causes of deaths combined. To examine exposure-response for cumulative time-weighted average and cumulative duration, we used the tests for heterogeneity and trend in SMRs based on the levels estimated using the medians of the exposure or duration categories.31
Proportional hazards models were also used to examine exposure-response trends within the study population for cumulative time-weighted average.32 The time scale for the analyses was workers’ age, and all models included sex, year of hire, and year of birth. Exposure was treated as a continuous time-dependent variable expressed in units of 100 ppm-months. Analyses were performed using SAS, version 9.2 (SAS Institute, Inc., Cary, NC).
Table 1 compares the current vital-status follow-up with the two previous studies of these workers. A slight difference from the previous update in the number of workers in duration-of-employment groups was a result of a slightly more precise method of calculating service time in this study. We identified 5026 deaths among the study population and obtained 4998 (99%) causes of death from either the death certificates or NDI-plus. The percentage of workers with censored follow-up has improved with each subsequent study. In this study, only 40 workers (0.3%) were censored, including 4 known foreign deaths, 7 deaths from military actions, and 29 subjects lost to follow-up. The total number of person-years increased from 122,078 in the original study to 307,932 in the first update and to 561,530 in this study.
Table 2 presents cause-specific SMRs for the entire study population and for workers after 15 or more years since first exposure to styrene. For the entire population, all causes of deaths combined (SMR = 1.08 [95% CI = 1.05–1.11]) and all cancers combined (1.12 [1.05–1.18]) are greater than expected. The increase in all cancers combined seems to be attributable mostly to the higher levels of lung cancer (1.34 [1.23–1.46]). Deaths from other causes related to smoking including kidney cancer (1.18 [0.83–1.62]), bladder cancer (1.25 [0.87–1.74]); all heart disease (1.05 [1.00–1.11]); bronchitis, emphysema, and asthma (1.35 [1.17–1.56]); and emphysema (1.30 [0.96–1.71]). Deaths from other cancers of interest occurred at or below expected levels including cancers of the pancreas (0.96 [0.73–1.22]), non-Hodgkin lymphoma (0.72 [0.50–1.00]), and leukemia (0.84 [0.60–1.14]). Other major causes of death with excesses include diabetes (1.29 [1.09–1.51]) and all external causes of death (1.10 [1.00–1.20]). Comparing the total study population with latency of 15 or more years, there are only minor changes to the SMRs, including cancers of the pancreas (from 0.96 to 0.90), lung (from 1.34 to 1.35), non-Hodgkin lymphoma (from 0.72 to 0.75), and leukemia (from 0.84 to 0.88).
We examined the potential impact of asbestos on SMRs for the association of styrene exposures with several causes of death (data not shown). For workers at plants that use asbestos, the SMRs for lung cancer and for bronchitis, emphysema, and asthma are 1.35 (1.23–1.48) and 1.42 (1.21–1.65), respectively; SMRs for workers at plants that reported no asbestos use are 1.30 (1.05–1.58) and 1.04 (0.69–1.51) for same causes of death. A single mesothelioma death occurred in plants with asbestos use, which was less than expected (0.39 [0.01–2.17]).
We also examined exposure-response trends for cancers of interest, causes of death related to cigarette smoking, and diabetes for the four cumulative styrene exposure categories (Table 3). There were no trends with exposure except for lung cancer and breast cancer, which both showed inverse trends. The SMR for kidney cancer is increased only in the highest cumulative exposure category (SMR = 1.79 [95% CI = 1.02–2.91]). We also examined exposure-response trends by sex and found no noteworthy differences (data not shown).
The results from the proportional hazards analysis for these causes of death are also presented in Table 3. Model fit was poor for Hodgkin lymphoma, non-Hodgkin lymphoma, all of the leukemia classifications, and pancreatic cancer. Of the remaining outcomes, kidney cancer shows a weak positive association with cumulative exposure (hazards ratio = 1.009 [95% CI = 1.000–1.017]). As in the SMR analysis, the proportional hazards trend was driven by an increased risk estimate for the highest exposure category (data not shown). Although the fit of the proportional hazard model is poor for pancreatic cancer, there was a weak positive linear association with cumulative styrene exposure (1.008 [1.002–1.015]), but subsequent examination of individual exposure categories revealed an erratic pattern with no important differences for any exposure category compared with the lowest exposure (data not shown). The inverse trends for lung cancer and heart disease in the SMR analyses were not reflected in the internal analyses. No evidence of an association with cumulative styrene exposure was found for Hodgkin lymphoma, non-Hodgkin lymphoma, leukemia, lymphoid or myeloid leukemia, or multiple myeloma, using either external or internal comparisons.
We used the same proportional hazards model used above and added total duration of time worked at the site as a continuous time-independent predictor variable to examine the inverse association of lung cancer with cumulative exposure. The model fits the data well, and the parameter estimates for sex, year of hire, year of birth, and cumulative styrene exposure changed little. The total duration of time worked at the site was a strong predictor, indicating that shorter duration of time at the site was associated with increased lung cancer risk considering cumulative exposure.
The analyses of peak exposures are presented in Table 4. With the exception of Hodgkin lymphoma, there are no major differences among the risk estimates of the four exposure categories. No trends with peak exposures are seen. The highest risk estimates for lymphoid leukemia, other lymphatic cancer, pancreatic cancer, and bladder cancer occurred among workers exposed 60 months or more.
We also examined cumulative duration and average exposure (data not shown). There is no indication of increasing monotonic trends for any of the causes of death examined for cumulative duration. There was an increasing risk with increasing average exposure for pancreatic cancer, with SMRs of 0.75, 0.83, 1.46, and 1.52, and for diabetes, with SMRs of 1.26, 1.27, 1.32, and 1.53.
Our study of a population with relatively high daily styrene exposures, frequent high peaks of styrene exposures and little or no exposure to other potential leukemogens find no evidence of increased risk from styrene exposures for non-Hodgkin lymphoma or any of the other lymphohematopoietic cancers. The levels of these cancers are at or below expected levels in almost all cases. Exposure to styrene has been associated with increased risk of the leukemias and lymphomas in some previous studies but not others. For example, studies of workers in the styrene-butadiene rubber industry reported increased risk of some lymphohematopoietic cancers including non-Hodgkin lymphoma.28 However, more recent studies of this industry indicate that when exposures to butadiene and dimethyldithiocarbamate (a catalyst used for production of synthetic rubber) are taken into account, there is no demonstrated risk of lymphohematopoietic cancers, including non-Hodgkin lymphoma.33 Because butadiene, styrene, and dimethyldithiocarbamate exposures are so highly correlated in this industry, it is difficult to separate the effects of styrene from the other two exposures. The European industry-wide study15 and the Washington State studies19 found no increased risk overall and no trend with cumulative exposure for all cancers of lymphatic or hematopoietic tissue, leukemia, or malignant lymphomas.15,19 However, the European industry-wide study observed significant trends for time since first exposure for cancers of all lymphatic or hematopoietic tissue, and leukemia and for average exposure for all lymphatic or hematopoietic tissue cancers. Given the internal inconsistencies in these findings, the authors concluded that excess risk of all lymphatic or hematopoietic tissue cancers remained an open question.15 The inconsistent findings across previous studies for these cancers, together with the persuasively negative findings in this study, support the notion that risk of cancers of lymphatic and hematopoietic tissue are not increased because of styrene exposures.
More lung cancers were observed than expected. However, we found no increase in risk with increasing cumulative, peak, duration, or average styrene exposure and no increase in risk with increasing latency. In fact, there was an inverse association with cumulative exposure and duration of exposure to styrene with the U.S. comparisons. Other studies of the reinforced plastics industries, the styrene-butadiene rubber industry, and styrene manufacturing have reported relative risk estimates for lung cancer close to expected levels.2,6,11,17,19–21,34–36 The study of the European reinforced plastics industry34 finds, as we do, that the highest risk of lung cancer occurs in the lowest exposure groups. A similar finding is reported among women in the styrene-butadiene rubber industry.35 Our results and the above studies do not support an attribution of excess lung cancer risk to exposure to styrene. Other possible explanations for excess risk include confounding by another workplace exposure or smoking. The examination of potential asbestos exposure in our study showed little difference in lung cancer risk. A previous nested case-control study of this group of workers, which collected data on smoking habits, concluded that smoking was the cause for the excess of lung cancer.12 In that study, 83% of the control group reported that they had ever smoked in 1977, compared with 33% of current or occasional smokers in a national survey in 1978–1980.37 Although these two surveys are not directly comparable, they do seem to indicate a higher smoking prevalence in the styrene workers.
We observed more deaths than expected from other diseases related to cigarette smoking, including bladder cancer; kidney cancer; bronchitis, emphysema, and asthma; and heart disease. The inverse association with duration of employment indicates that the excess lung cancer deaths occur mostly among shorter-term workers. We see similar excesses among the shorter-term workers for bladder cancer; bronchitis, emphysema, and asthma; and heart disease. We used an adjustment for unmeasured confounding by smoking assuming the difference is approximated by the ratio of lung cancer rates to bronchitis, emphysema, and asthma rates.38 This adjustment produced an SMR of 1.0 for the styrene workers overall and SMRs of 1.1, 1.0, 1.3, and 0.7 across increasing cumulative exposure categories for styrene. A gradual reduction over time in confounding from smoking, consistent with a general reduction of smoking over time in the workforce and the general population, is the most likely explanation for the inverse relationship of lung cancer with cumulative exposure to styrene. The increased risk of lung cancer in our study is most likely related to smoking.
Increased pancreatic cancer risk has not been seen consistently across studies.15,17,19,20,36,39 In our study, there were fewer deaths than expected from pancreatic cancer overall. We found some evidence of a weak association with cumulative exposure level in the proportional hazards model, although there was no obvious pattern of risk across four exposure categories in either external or internal analyses. The results from the proportional hazards model indicate that the impact on increased risk, if valid, is relatively small. It is possible that the increased hazard ratio in the proportional hazards analysis is a chance finding of an ill-fitting model.
An unexpected finding was an association of cumulative and peak styrene exposures with kidney cancer. However, other studies of styrene-exposed workers have not found increased kidney cancer risk,15,17,40 and the excess risk of smoking-related cancers in our study indicated that part of the excess in this cancer is probably because of smoking. This finding is likely due to chance because the exposure-response pattern is weak, no increased risk is observed in the other studies of styrene workers, and there is potential confounding with smoking.
This study has several strengths for assessing occupational health effects. First, the size of the study population and the extended follow-up period provide power comparable to the two largest studies of the reinforced plastic industry. This power allows for more precise evaluation of the risk of rare cancers and cancers with long latencies.15,16 Second, compared with the styrene-butadiene rubber industry and styrene manufacturing, routinely high average and peak exposure levels in reinforced plastics are more likely to uncover potential styrene-related health effects. Daily exposures in the reinforced plastics industry averaged 30–50 ppm, with peaks >200 ppm. Exposures in the styrene-butadiene rubber and styrene manufacturing industries averaged <10 ppm, with few peak exposures.5 Third, work histories for each worker included in the study were available and exposure estimates are based on an in-depth job evaluation supplemented by personal styrene monitoring for each of the 30 plants in the study. This approach reduced potential exposure misclassification. Although other studies of the reinforced plastics industry have evaluated exposure response,15,19 our study examines both cumulative exposures and peak exposures.
Among the limitations of this study is the reliance on data from death certificates for investigating some cancer types, especially cancers of the lymphatic and hematopoietic tissues, where misclassification sometimes occurs.41 In addition, exposure estimates were truncated in 1977, leaving 27% of the study group with incomplete exposure history. This missing exposure information may result in some exposure underestimation. However, because the misclassification affects only a small portion of the study population at a time of generally lower styrene exposure, we think it unlikely to appreciably affect the major findings of our study.
In conclusion, we found no increased mortality from lymphatic and hematopoietic tissue cancers overall or in any subdiagnoses including non-Hodgkin lymphoma, Hodgkin lymphoma, multiple myeloma, or leukemia in this highly exposed group of styrene workers. Furthermore, no exposure response was associated with cumulative daily exposure or number of peak exposures. The absence of an increased risk for pancreatic cancer overall plus the inconsistent findings in the trend tests provide little evidence of a causal association. Lung cancer rates, although greater than expected, seemed to be unrelated to styrene exposure and more likely attributable to smoking. In this large study with relatively high styrene exposure, we find no credible evidence that styrene exposure increases risk from cancers of the lymphatic and hematopoietic tissue, pancreas, or lung.
1. . NIOSH. National Occupational Exposure Survey (1981–1983). 1993 Cincinnati, OH National Institute for Occupational Safety and Health
2. Bond GG, Bodner KM, Olsen GW, Cook RR. Mortality among workers engaged in the development or manufacture of styrene-based products—an update. Scand J Work Environ Health. 1992;18:145–154
3. Coggon D. Epidemiological studies of styrene-exposed populations. Crit Rev Toxicol. 1994;24(suppl):S107–S115
4. Boffetta P, Adami HO, Cole P, Trichopoulos D, Mandel JS. Epidemiologic studies of styrene and cancer: a review of the literature. J Occup Environ Med. 2009;51:1275–1287
5. . IARC. Some Traditional Herbal Medicines, Some Mycotoxins, Naphthalene, and Styrene. 2002;Vol. 82 Lyon, France IARC Press
6. Nicholson WJ, Selikoff IJ, Seidman H. Mortality experience of styrene-polystyrene polymerization workers. Initial findings. Scand J Work Environ Health. 1978;4(suppl 2):247–252
7. Meinhardt TJ, Lemen RA, Crandall MS, Young RJ. Environmental epidemiologic investigation of the styrene-butadiene rubber industry. Mortality patterns with discussion of the hematopoietic and lymphatic malignancies. Scand J Work Environ Health. 1982;8:250–259
8. Matanoski GM, Schwartz L. Mortality of workers in styrene-butadiene polymer production. J Occup Med. 1987;29:675–680
9. Matanoski GM, Santos-Burgoa C, Schwartz L. Mortality of a cohort of workers in the styrene-butadiene polymer manufacturing industry (1943–1982). Environ Health Perspect. 1990;86:107–117
10. Santos-Burgoa C, Matanoski GM, Zeger S, Schwartz L. Lymphohematopoietic cancer in styrene-butadiene polymerization workers. Am J Epidemiol. 1992;136:843–854
11. Delzell E, Sathiakumar N, Graff J, Macaluso M, Maldonado G, Matthews R. An updated study of mortality among North American synthetic rubber industry workers. Res Rep Health Eff Inst. 2006:1–63 discussion 65–74
12. Wong O. A cohort mortality study and a case-control study of workers potentially exposed to styrene in the reinforced plastics and composites industry. Br J Ind Med. 1990;47:753–762
13. Wong O, Trent LS, Whorton MD. An updated cohort mortality study of workers exposed to styrene in the reinforced plastics and composites industry. Occup Environ Med. 1994;51:386–396
14. Coggon D, Osmond C, Pannett B, Simmonds S, Winter PD, Acheson ED. Mortality of workers exposed to styrene in the manufacture of glass-reinforced plastics. Scand J Work Environ Health. 1987;13:94–99
15. Kogevinas M, Ferro G, Andersen A, et al. Cancer mortality in a historical cohort study of workers exposed to styrene. Scand J Work Environ Health. 1994;20:251–261
16. Kolstad HA, Lynge E, Olsen J, Breum N. Incidence of lymphohematopoietic malignancies among styrene-exposed workers of the reinforced plastics industry. Scand J Work Environ Health. 1994;20:272–278
17. Kolstad HA, Juel K, Olsen J, Lynge E. Exposure to styrene and chronic health effects: mortality and incidence of solid cancers in the Danish reinforced plastics industry. Occup Environ Med. 1995;52:320–327
18. Okun AH, Beaumont JJ, Meinhardt TJ, Crandall MS. Mortality patterns among styrene-exposed boatbuilders. Am J Ind Med. 1985;8:193–205
19. Ruder AM, Ward EM, Dong M, Okun AH, Davis-King K. Mortality patterns among workers exposed to styrene in the reinforced plastic boatbuilding industry: an update. Am J Ind Med. 2004;45:165–176
20. Frentzel-Beyme R, Thiess AM, Wieland R. Survey of mortality among employees engaged in the manufacture of styrene and polystyrene at the BASF Ludwigshafen works. Scand J Work Environ Health. 1978;4(suppl 2):231–239
21. Hodgson JT, Jones RD. Mortality of styrene production, polymerization and processing workers at a site in northwest England. Scand J Work Environ Health. 1985;11:347–352
22. Ott MG, Kolesar RC, Scharnweber HC, Schneider EJ, Venable JR. A mortality survey of employees engaged in the development or manufacture of styrene-based products. J Occup Med. 1980;22:445–460
23. Tossavainen A. Styrene use and occupational exposure in the plastics industry. Scand J Work Environ Health. 1978;4(suppl 2):7–13
24. Miller RR, Newhook R, Poole A. Styrene production, use, and human exposure. Crit Rev Toxicol. 1994;24(suppl):S1–S10
25. Van Rooij JG, Kasper A, Triebig G, Werner P, Jongeneelen FJ, Kromhout H. Trends in occupational exposure to styrene in the European glass fibre-reinforced plastics industry. Ann Occup Hyg. 2008;52:337–349
26. AD Little Inc. Industrial Hygiene Evaluation of Retrospective Mortality Study Plants.. 1981 Washington, DC
27. Checkoway H, Rice CH. Time-weighted averages, peaks, and other indices of exposure in occupational epidemiology. Am J Ind Med. 1992;21:25–33
28. Matanoski G, Elliott E, Tao X, Francis M, Correa-Villasenor A, Santos-Burgoa C. Lymphohematopoietic cancers and butadiene and styrene exposure in synthetic rubber manufacture. Ann N Y Acad Sci. 1997;837:157–169
29. Doll R, Peto R The Causes of Cancer: Quantitative Estimates of Avoidable Risk of Cancer in the United States Today. 1981 New York Oxford University Press
30. Marsh GM, Youk AO, Stone RA, Sefcik S, Alcorn C. OCMAP-PLUS: a program for the comprehensive analysis of occupational cohort data. J Occup Environ Med. 1998;40:351–362
31. Breslow NE, Day NE Statistical Methods in Cancer Research: Volume II—The Design and Analysis of Cohort Studies. 1987 Lyon, France International Agency for Research on Cancer
32. Cox DR. Regression models and life-tables. J R Statis Soc. 1972;34:187–220
33. Graff JJ, Sathiakumar N, Macaluso M, Maldonado G, Matthews R, Delzell E. Chemical exposures in the synthetic rubber industry and lymphohematopoietic cancer mortality. J Occup Environ Med. 2005;47:916–932
34. Kogevinas M, Boffetta P, Vainio H. Determining the hazards of work place chemicals. Epidemiology. 1993;4:561–562
35. Sathiakumar N, Brill I, Delzell E. 1,3-butadiene, styrene and lung cancer among synthetic rubber industry workers. J Occup Environ Med. 2009;51:1326–1332
36. Anttila A, Pukkala E, Riala R, Sallmén M, Hemminki K. Cancer incidence among Finnish workers exposed to aromatic hydrocarbons. Int Arch Occup Environ Health. 1998;71:187–193
37. Nelson DE, Emont SL, Brackbill RM, Cameron LL, Peddicord J, Fiore MC. Cigarette smoking prevalence by occupation in the United States. A comparison between 1978 to 1980 and 1987 to 1990. J Occup Med. 1994;36:516–525
38. Richardson DB. Occupational exposures and lung cancer: adjustment for unmeasured confounding by smoking. Epidemiology. 2010;21:181–186
39. Sathiakumar N, Delzell E, Hovinga M, et al. Mortality from cancer and other causes of death among synthetic rubber workers. Occup Environ Med. 1998;55:230–235
40. Delzell E, Sathiakumar N, Hovinga M, et al. A follow-up study of synthetic rubber workers. Toxicology. 1996;113:182–189
41. Percy CL, Miller BA, Gloeckler Ries LA. Effect of changes in cancer classification and the accuracy of cancer death certificates on trends in cancer mortality. Ann N Y Acad Sci. 1990;609:87–97 discussion 97