Secondary Logo

Journal Logo

REVIEW ARTICLE

A Meta-Analysis of Cohort Studies Describing Mortality and Cancer Incidence among Chemical Workers in the United States and Western Europe

Greenberg, Raymond S.1; Mandel, Jack S.2; Pastides, Harris3; Britton, Nicole L.2; Rudenko, Larisa4; Starr, Thomas B.5

Author Information
  • Free

Abstract

Considerable attention has been given to the potential health risks to employees in chemical manufacturing and related industries. During the past several decades, hundreds of morbidity or mortality studies of chemical workers have been published, with most focused on the potential cancer risks associated with chemical exposure.

Although many chemicals in general use have exhibited positive responses in cancer or noncancer rodent bioassays when tested at high doses, relatively few workplace exposures have been established as causes of cancer or other diseases in humans. Most epidemiologic studies of the workplace have documented a deficit in mortality caused by cardiovascular disease, accidents, and several other disease categories. These favorable outcomes typically are attributed to the “healthy worker effect.” Notable exceptions are the relations of bladder cancer with aromatic amines and certain dyes, liver angiosarcoma with vinyl chloride, and lung cancer with hexavalent chromium. Each of these causal associations was established because exposure to the specific chemicals was substantial and well documented.

Few comprehensive studies, however, have evaluated the overall health of individuals employed in chemical manufacturing. The meta-analyses that we report herein evaluate the cause-specific mortality and cancer incidence of chemical workers in the United States and western Europe as they have been reported in the peer-reviewed literature between 1966 and 1997. These analyses do not identify specific chemicals, nor rule out their effects. Rather, they provide a general overview of the mortality and cancer incidence experience of persons employed in the chemical industry as reported in cohort studies and indicate several specific areas for future research.

Subject and Methods

Our starting point was the database developed by M.D. Whorton et al (1998; available from the American Chemistry Council), a comprehensive inventory of published peer-reviewed mortality and cancer incidence studies of workers engaged in the manufacture or application of chemical products. They identified articles by searching online databases (for example, Medline, Cancerlit, NIOSHTIC, Toxline, Embase, and Current Contents), reviewing yearly indices of selected journals, and reviewing references of identified studies and relevant review articles. Studies were included if they reported on employees working in companies that would qualify for membership in the U.S. American Chemistry Council. Eligible companies included those involved in the production of chemicals and the manufacture of products involving a substantial chemical operation such as chemical synthesis, formulation, or extraction. Studies also had to be published in English in peer-reviewed journals between 1966 and 1997, to have reported on cohorts in the U.S. or western Europe, and to have included information on mortality or cancer incidence.

We excluded from our review nested case-control studies because they are nearly always initiated after some preliminary observation of an increased risk of disease or mortality that is potentially work-related; therefore, we believe that they are less likely to be useful indicators of the general health and mortality of workers throughout the large and diverse chemical industry. Furthermore, we did not include population-based case-control studies because our goal was to evaluate the general health of chemical-industry workers and not risks related to specific chemical exposures.

We independently assessed the completeness and accuracy of the Whorton et al database by comparing it with the literature citations in recent reviews of chemical-specific epidemiologic investigations, such as International Agency on Cancer Research Monographs, as well as with results from additional online literature searches. Few discrepancies were identified, and they were all satisfactorily resolved.

To prevent multiple appearances of the same cohort in our meta-analysis, studies were grouped so that specific populations could be traced from the earliest to the most recently published report. Generally we chose the most recent publication for data abstraction because it provided the longest follow-up and the most up-to-date information. Exceptions included earlier studies that reported on specific subgroups (for example, women, salary vs hourly) not included in the most recent publication.

For studies reporting multiple standardized mortality ratios (SMRs) based on different comparison populations, such as state and national rates, “local” comparison groups were selected because local populations more directly reflect the source populations from which the workers arose and, therefore, should provide a less biased comparison. Although a few of the studies in the Whorton et al database reported risk ratios resulting from internal comparisons, the only effect estimates selected for use in this analysis were SMRs based on external comparison. We found little consistency in the type of internal group used in these studies (for example, workers from a different plant or workers from the same plant assumed to be unexposed), and the number of studies reporting these results was extremely small relative to the studies using general population comparison groups. Only one study reported risk ratios for employee populations that were not included in SMR results elsewhere in the database.

One of the primary reviewers (R. S. G., J. S. M., or H. P.) reviewed each study and, using a priori quality criteria, assigned a quality score of 1 to 3 for each study, as follows:

1. Satisfactory: The study addressed the methodologic issues in a manner consistent with generally accepted epidemiologic standards.

2. Limited adequacy: The study addressed a methodologic attribute incompletely, or in a manner that might have been appropriate, but provided too few details to allow a definitive assessment.

3. Unsatisfactory: The study either failed to address the methodologic attributes or addressed them in a manner that was inappropriate or inadequate.

We deliberately set the threshold for inclusion low so that only studies with serious deficiencies in methodology, design, or analysis would be excluded.

If one of us judged a study “unsatisfactory,” it was subsequently assessed by all three reviewers. Unanimous accord was required to exclude a study. As a check of reviewer reliability, a sample of one-third of the studies was reassessed by a different reviewer. Greater than 95% concordance was reached initially and, after discussion, unanimous accord was reached on all those studies for which initial quality scores differed.

Of the 461 studies in the database, 69 (15.0%) were ineligible for one of the following reasons: noncohort study design (40 studies, 8.7%), geographical location (27 studies, 5.9%), or employers ineligible for membership in the American Chemistry Council (2 studies, 0.4%). Only seven (1.5%) studies were excluded as unsatisfactory by virtue of inadequacies in design, description, or analysis. For the remaining 385 studies, data were abstracted from the 181 studies representing the most recent report for a cohort, or the earlier study of the same cohort that provided relevant data according to the selection criteria. 1–50, 51–100, 101–150, 151–181 Of these, 156 reported only mortality data, 2–6,8,10–19,21–37,39–47, 51–62,64,67–77,79–81,82–87,89–90,92–102, 104–112,114,116–118,120,123–128,130–138,140–144, 146,148–168,170–181 10 reported only incidence data, 7,20,50,78,88,103,121,129,139,169 and 15 reported both mortality and incidence data. 1,9,38,48,49,63,65,66,82,91,113,115,119,145,147 Sixty-nine per cent of the studies reporting mortality data 1,8–19,21–24,26–29,32–35,37,39,40,42,43,46–49, 51,52,55–61,62–65,67,68,72,74–76,79–87,90,91,94,95,97,99, 101,106,108,113–117,119,120,122–125,128,131,134–138,141–146,148–151, 154,156,158–160,162,164,165,168,170,172–174,176,179–181 and 64% of the studies reporting incidence data 1,9,20,48–50,63,65,82,92,103,113,115,119,139,145 received “satisfactory” quality scores. This selection process is illustrated in Figure 1.

FIGURE 1
FIGURE 1:
Selection and scoring of studies. ACC = American Chemical Council.

We abstracted and entered into an electronic database the observed and expected numbers of cases of each endpoint reported in the selected studies. 1–50, 51–100, 101–150, 151–181 These data were then used to generate SMRs or standardized incidence ratios (SIRs) and corresponding 95% confidence intervals as appropriate. For each disease endpoint, meta-analyses of the data were undertaken from all studies reporting that endpoint (meta command, Stata Corp). We also performed additional meta-analyses for selected subcohorts on the basis of gender, latency (time since first employment to death or diagnosis), or duration of exposure (employment).

Pooled effect estimates were generated using both fixed- and random-effects models. In fixed-effects models, study-specific contributions to the pooled estimates were weighted inversely in proportion to their estimated “within-study” variance estimates. In the random-effects models, we permitted the study-specific effects estimates to have a zero mean normal random component of which the additional variance was determined by a “between-studies” variance estimate. We also routinely conducted tests for heterogeneity of the study-specific estimates. Because the large number of studies included in the meta-analyses led to extremely high precision in the effects estimates, heterogeneity among results was routinely observed, and thus we elected to present only the random-effects model estimates. There were, however, only four causes of death (that is, bronchitis, esophageal cancer, laryngeal cancer, and Hodgkin’s lymphoma) exhibiting sufficient heterogeneity to lead to qualitatively different interpretations based on the model used. We therefore re-examined data evaluated in the meta-analyses of these outcomes to identify potential outliers and/or possible errors in data abstraction or entry.

Results

The population evaluated in these mortality meta-analyses includes more than 1 million chemical workers (>15 million person-years) (Table 1), of whom approximately 85% were male. More than 35,000 workers (>3 million person-years) were included in the meta-analyses of cancer incidence; approximately 62% of these were male.

Table 1
Table 1:
Population Summary

Results of the meta-analyses are presented in Tables 2–6;Table 7 presents the incidence meta-analyses. The rows of each table represent the categories of health outcomes (deaths or incident cases). The first column indicates the health outcome; the second column provides the corresponding codes from the International Classification of Diseases, 8th revision; the third column represents the number of studies included in each meta-analysis; the fourth column provides the total number of observed events; and the last two columns provide the point estimate and 95% confidence interval for either the meta-SMR (Tables 2–6) or the meta-SIR (Table 7). Because the fixed-effects and random-effects models produced results that were generally in close agreement, we have only presented the random-effects model estimates here.

Table 2
Table 2:
Meta-SMR Estimates from Random Effects Model: Full Cohort Analysis by Cause of Death
Table 3
Table 3:
Meta-SMR Estimates from Random-Effects Model: Male Worker Analysis by Cause of Death
Table 4
Table 4:
Meta-SMR Estimates from Random-Effects Model: Female Worker Analysis by Cause of Death
Table 7
Table 7:
Meta-SIR Estimates from Random Effects Model: Incidence Full Cohort Analysis by Cancer Site

As shown in Table 2, the number of studies reporting quantitative data varied markedly according to the mortality outcome considered. Of the 168 full-cohort mortality studies eligible for inclusion, 1–6,8–19,21–49,51–77,79–87,89–102,104–120,122–129,130–138,140–165,167,168,170–179,181 the most commonly reported outcomes were all causes of death (154 studies, 90.6%) 1–6,8–19,21–25,27–31,33–49,53–60,62,63,65–68,70–77,79,80,82–87,90–94,96–102,104–112,114,116–120,122–128,130,131,133–138,140–143,145–148,150–165,167,168,170–179,181 and all malignancies (150 studies, 88.2%). 1–4,8–19,21–25,27–31,33,35–47,49,51–53,56–60,62,63,65–77,79–81,83–87,91–94,96–102,104–111,113–120,122–127,130–138,140–143,145–154,156–165,167,168,170–172,174–179,181 Some mortality outcomes were much less commonly reported, such as thyroid cancer (11 studies, 6.5%) 12,18,29,46,58,83,84,141,145,153 and homicide (5 studies, 2.9%). 24,54,76,138,154

For each mortality outcome, the number of studies and the corresponding total number of events for all persons combined were very similar to those for men (see Tables 2 and 3). The number of studies reporting mortality for women was small, 4,6,15,17,39,46,59–61,76,77,81,84,113,114,118,136–138,140,145,159,179 with all types of cancer combined (18 studies), 4,15,17,39,46,59,60,76,77,81,84,113,114,136,137,145,159,179 all causes of mortality, and breast cancer (17 studies each) 4,6,15,17,39,46,59,60,76,77,84,113,114,118,136–138,140,145,159,179 the outcomes most likely to be reported (Table 4). Information on latency (10 years or more) was available in a modest number of studies for some outcomes, such as cancer of the bronchus, trachea, and lung (49 studies), 2,10,13,14,17,18,22,24,33,34,41,42,43,46,47,51,63–65,68–70,75,84,85,100,106,107,109,114,117,122,127,131,134,136,138,141,149,153,162,164,167,172,174,176,177,179,180 and all malignant neoplasms (46 studies) 2,4,9,10,13,14,17–19,22,38,43,47,51,63–66,68–70,75,83,84,94,100,106,109,114,117,127,130,136,139,141,142,146,152,162,164,172,174,176,177,179,180 (Table 5). This information was infrequently available for other outcomes, such as breast cancer (three studies), 17,138,173 multiple myeloma (two studies), 51,83 thyroid cancer (one study), 18 and homicide (no studies). Information related to workers employed 10 or more years was available most frequently for all malignancies combined (27 studies) 4,12–14,18,23,38,43,51,72,73,80,81,84,100,120,130,132,136,138,156,164,165,170,176,177,180; bronchus, trachea, and lung cancer (23 studies) 13,14,18,24,33,37,41,42,51,84,92,107,108,136,138,149,163,170,176,177,180; and all causes of mortality (23 studies) 4,12,14,38,43,48,54,63,72,73,80,92,100,120,130,136,156,164,170,171,176,177,180 (Table 6). It was available in two or fewer studies for laryngeal cancer, 14,18 breast cancer, 138,173 thyroid cancer, 18 multiple myeloma, bronchitis, emphysema and asthma, 173 suicides, 136,180 and homicides.

Table 5
Table 5:
Meta-SMR Estimates from Random-Effects Model: Longest Latency (≥10 Years) by Cause of Death
Table 6
Table 6:
Meta-SMR Estimates from Random-Effects Model: Longest Duration of Employment (≥10 Years) by Cause of Death

Of the 25 eligible studies with incidence data, 1,7,9,20,38,48–50,63,65,66,78,82,88,91,103,113,115,119,121,129,139,145,147,169 19 provided information on all types of cancer combined, 1,7,9,20,38,48–50,63,65,66,82,91,103,113,115,119,121,145 and an equal number presented data on bladder cancer 7,9,20,38,48–50,63,65,82,103,113,115,121,129,139,145,147,169 (Table 7). For cancers of the larynx, 7,63,113,115,145 breast, 1,7,66,113,145 thyroid, 7,113,145 and multiple myeloma, 7,48,66,91,145 five or fewer studies provided incidence data.

All-Cause Mortality

A total of 154 studies provided data on 143,927 deaths from all causes (Table 2). Overall, there were 10% fewer deaths observed than expected (meta-SMR = 0.90, 95% CI = 0.87–0.92). When only the 119 mortality studies judged to be of the highest quality were evaluated, the meta-SMR was unchanged (data not shown); we therefore evaluated the larger group to maximize the opportunity to detect potential effects. Risk was not elevated after accounting for latency (meta-SMR = 0.92, 95% CI = 0.88–0.96) or duration of employment (meta-SMR = 0.85, 95% CI = 0.79–0.91).

Cancer Mortality

There were 41,914 deaths from cancer reported in 150 studies (Table 2). The observed number of cancer deaths among chemical workers was virtually identical to the expected number (meta-SMR = 0.99, 95% CI = 0.94–1.04). The results for men paralleled those for the overall worker population (Table 3). For women, however, there was a 9% reduction from expected levels in observed cancer mortality (meta-SMR = 0.91, 95% CI = 0.85–0.97) (Table 4).

A slight increase in cancer mortality was observed for all workers after 10 or more years of induction time (meta-SMR = 1.04, 95% CI = 0.99–1.09) (Table 5). In contrast, fewer cancer deaths than expected were observed for employees who worked 10 or more years (meta-SMR = 0.97, 95% CI = 0.91–1.05) (Table 6).

We found no notable difference between observed and expected numbers of deaths in the full cohort for cancers of the buccal cavity and pharynx, esophagus, stomach, large intestine, pancreas, prostate, kidney, and multiple myeloma (Table 2). There were more deaths observed than expected for cancers of the trachea, bronchus and lung (meta-SMR = 1.12, 95% CI = 1.05–1.19), bladder (meta-SMR = 1.29, 95% CI = 1.01–1.65), thyroid (meta-SMR = 1.62, 95% CI = 1.05–2.51), lymphatic and hematopoietic systems (meta-SMR = 1.09, 95% CI = 1.02–1.15), leukemia (meta-SMR = 1.07, 95% CI = 1.01–1.13), larynx (meta-SMR = 1.17, 95% CI = 0.99–1.40), skin (meta-SMR = 1.13, 95% CI = 0.98 = 1.30), lymphomas (meta-SMR = 1.09, 95% CI = 0.97–1.23), liver and biliary passages (meta-SMR = 1.08, 95% CI = 0.92–1.28), respiratory system (meta-SMR = 1.06, 95% CI = 0.97–1.15), Hodgkin’s disease (meta-SMR = 1.06, 95% CI = 0.90–1.23), and brain and other areas of the central nervous system (meta-SMR = 1.05, 95% CI = 0.96–1.16). There were fewer deaths observed than expected for cancers of the digestive organs and peritoneum (meta-SMR = 0.93, 95% CI = 0.89–0.98) and of the breast (meta-SMR = 0.90, 95% CI = 0.83–0.98).

When we considered the analyses by latency (Table 5), length of employment (Table 6), gender (Tables 3 and 4), and study quality (data not presented), some consistency in elevated findings was observed for cancers of the bronchus, trachea and lung, the bladder, lymphatic/hematopoietic systems, and the brain and central nervous system. A list of all studies with SMRs of 2.0 or greater for these cancers is provided in Tables 8–11, respectively. For cancers of the bronchus, trachea, and lung, 5 of the 10 studies with SMRs of 2.0 or greater related to chromium or chromate exposure. 42,55,67,85,95 Three of the studies 97,128,134 on bladder cancer involved aromatic amines and reported SMRs of about 30. There were seven studies with SMRs of 2.0 or greater for lymphatic and hematopoietic cancers. 11,14,36,72,91,125,144 Two of these investigations related to benzene exposure, 36,125 and two related to ethylene oxide exposure. 14,72 The studies reporting brain and central nervous system cancer excesses of 2.0 or more included two each in which the workers were exposed to acrylonitrile 100,174 and vinyl chloride 154,171; these were not, however, the studies reporting the highest SMRs. 3,147

Table 8
Table 8:
Worker Exposures for Studies with Excess Mortality from Bronchus, Trachea, and Lung Cancers (SMR ≥2.0)
Table 9
Table 9:
Worker Exposures for Studies with Excess Mortality from Cancers of the Bladder (SMR ≥2.0)
Table 10
Table 10:
Worker Exposures for Studies with Excess Mortality from Cancers of the Lymphatic and Hematopoietic Systems (SMR ≥2.0)
Table 11
Table 11:
Worker Exposures for Studies with Excess Mortality from Cancers of the Brain and Other Central Nervous System (SMR ≥2.0)

Cancer Incidence

Cancer incidence data were available from only 25 studies 1,7,9,20,38,48,49,50,63,65,66,78,82,88,91,103,113,115,119,121,129,139,145,147,169 (Table 7). For all cancers combined, the results were essentially the same as those obtained with the mortality data (meta-SIR = 1.0, 95% CI = 0.92–1.08). Fewer cases were observed than expected for cancers of the digestive organs and peritoneum (meta-SIR = 0.86, 95% CI = 0.82–0.91) and for esophageal cancer (meta-SIR = 0.67, 95% CI = 0.49–0.92); excesses were found for bladder cancer (meta-SIR = 2.21, 95% CI = 1.18–4.15), multiple myeloma (meta-SIR = 2.19, 95% CI = 1.27–3.78), respiratory cancers (meta-SIR = 1.16, 95% CI = 0.98–1.37), and cancers of the lymphatic and hematopoietic systems (meta-SIR = 1.24, 95% CI = 0.96–1.61).

Noncancer Mortality

As shown in Table 2, there were 33 studies in which mortality from diabetes mellitus was reported, 5,6,17,18,21–25,38,43,46,54,60,63,98,102,114,136–138,149,163–165,168,172,174,175,177–179,181 with more than 750 deaths and a small reduction from the expected level (meta-SMR = 0.96, 95% CI = 0.84–1.10). The 36,421 observed deaths from cardiovascular diseases reported in 99 studies 9,10,12–14,17–19,21–23,25,27,30,32,33,35,36,38,39,42,44,46–49,52,53,56,60,62,63,65–68,70,71,73,74,76,77,79,80,85,87,90–94,96,98,100,104,105,107,110,114,116,117–120,122,124,127,133–138,140,141,145,146,149,150,152–154,157,160,161,163,164,167,168,170,172,174–179,181 represented 14% fewer deaths than expected (meta-SMR = 0.86, 95% CI = 0.83–0.89). Results were similar for deaths from coronary heart disease and cerebrovascular disease in which the meta-SMRs for the full cohort were 0.87 (95% CI = 0.83–0.91) and 0.85 (95% CI = 0.80–0.90), respectively. When we pooled results over 94 studies, 2,4,9,10,13,14,17–19,21–25,27,28,33–39,42–44,46–49,52,56,58,60,65–68,70,71,76,83,85,86,93,98,99,104,105,107–110,114,116,119,120,123,126,127,130,134–138,141,145,146,149–154, 157,158,159–165,167,168,170,172–179,181 the 5,974 observed deaths from noncancer respiratory disease were 23% lower than expected (meta-SMR = 0.77, 95% CI = 0.72–0.83). There was no increase in deaths from bronchitis, emphysema, and asthma. When we confined our analysis to the studies judged to be of quality criterion 1, however, there was a 22% risk increase (meta-SMR = 1.22, 95% CI = 1.01–1.47). The risk was also elevated in the studies of women, but that increase was based on only 25 observed deaths reported in two studies 6,46 (Table 4). The long latency subcohort, with 44 observed deaths in seven stud-ies, 47,63–66,174,177 revealed a 35% increase in deaths (meta-SMR = 1.35, 95% CI = 0.97–1.86) (Table 5). We observed fewer deaths than expected from cirrhosis of the liver (meta-SMR = 0.74, 95% CI = 0.62–0.88) (Table 2) and noted similar results when we confined our analyses to only the studies judged to meet quality criterion 1 (data not presented), as well as those limited to men 4,5,15,18,21,22,25,43,44,46,47,62,63,76,77,91,97–99,102,107,114,134,136,137,141,158,163–165,172,175,178,179,181 (Table 5) or women 6,15,17,46,76,77,114,136,137,179 (Table 4), or the long-latency 17,18,22,47,63,64,114,136,141,164,173,180 (Table 5) and long-duration 43,136,177,180 (Table 6) subcohorts.

When we pooled results of 83 studies, 2,4–6,12–14,19,21,23–25,27,28,30,32,36,38,39,42–44,46–49,53,54,62,65–68,74,76,77,79,80,83,85,87,91–94,98,102,104,105,107,118–120,124,125,127,130,133–137,140,141,149–153,159–163,167,170,172,175,176–179,181 6,773 observed deaths were reported from external causes, which was 16% lower than expected (meta-SMR = 0.84, 95% CI = 0.78–0.91) (Table 2), with similar results in the subgroup analyses (Tables 3–6). The observed numbers of deaths from accidents, suicides, and homicides were all lower than expected (Table 2).

Discussion

This is the first comprehensive analysis of the mortality and cancer incidence of workers in the chemical industry. Based on an exhaustive review of the published literature, it has included an assessment of the methodologic quality of the individual studies. Furthermore, the extremely large size of the aggregated cohort study samples provided sufficient numbers to detect even small differences in health outcomes. We undertook meta-analyses of 35 mortality endpoints and 23 cancer incidence endpoints using data from the full cohort and, in addition, examined the mortality of several subgroups (males; females; workers first employed at least 10 years before their death, incident cancer, or the end of follow-up; and workers with at least 10 years of employment).

The observed cause-specific mortality and site-specific cancer incidence of the workers included in this study were, with few exceptions, largely reassuring, as we observed similar outcomes among the vast majority of chemical workers and comparison populations. Fewer than expected deaths were observed among chemical workers from all causes, cardiovascular disease, noncancer respiratory disease, cirrhosis of the liver, and external causes including accidents, suicide, and homicide. Although some, and perhaps most, of these deficits may be attributed to the healthy worker effect, there nevertheless appears to be no generalized increased risk for mortality or cancer incidence among workers employed in chemical manufacturing.

The only consistently positive findings across the full cohort and the long-latency, long-employment duration, and incidence subgroups were those for cancers of the bronchus, trachea, and lung; the bladder; and the lymphatic/hematopoietic systems. The excesses for these three endpoints appear to be attributable, at least in part, to previously reported associations with specific chemical exposures: lung cancer with chromates, 182 bladder cancer with aromatic amines, 183 and lymphatic/hematopoietic system cancers with ethylene oxide 184 or benzene. 185 From a methodologic standpoint, the detection of these positive associations implies that the analyses had sufficient power to detect increases in risk, despite the fact that small proportions of the chemical workers included in this study were exposed to the specific implicated chemicals.

Limitations of the Meta-analyses

Several issues influence the interpretation of these results. Some issues are common to all occupational cohort studies; others relate to combining studies conducted for a variety of purposes and with varying methods and reporting practices. These are discussed below.

The “Healthy Worker Effect”

The “healthy worker effect” results from generally low relative mortality from selected diseases in an employed population 186 and is caused by the selection process whereby persons with evidence of disease or at increased risk of disease are less likely to gain and to keep jobs. In general, the “healthy worker effect” is more prominent for cardiovascular diseases and certain musculoskeletal conditions than for cancer, and is also more prominent in settings that have rigorous pre-employment testing or fitness requirements. It is difficult, however, to estimate quantitatively the extent of this bias among the pooled cohorts analyzed in this study. The reduced total mortality and cardiovascular disease mortality relative to the general population observed in this analysis is likely attributable, at least in part, to the healthy worker effect.

Confounding Variables

A concern common to virtually all occupational cohort studies is the lack of information about potentially important confounding variables. Age, gender, and time period of observation were accounted for by the authors in nearly all of the original studies; however, race or salaried vs hourly pay status were available in too few studies to be useful in constructing meaningful subgroup meta-analyses (≤15 studies reported data for nonwhite workers and ≤12 studies for salaried workers).

Another confounder of potential importance was smoking. Differential smoking patterns between study and comparison populations might have affected the SMRs for lung and other smoking-related diseases and could account for part of the significant excess of respiratory tract cancers. It is unlikely, however, that smoking prevalence varies enough between chemical workers and the general population to be able to account for the observed 29% excess of bladder cancer (Tsai SP, Wendt JK, Hunter RB. Trends in cigarette smoking among refinery and petrochemical plant employees. In preparation).

The Assumption of Exposure

An important assumption that underlies our analyses is that all members of the included cohorts were exposed to one or more chemicals in their workplaces. This level of information on exposure is adequate for a comparison of the mortality and cancer incidence experience of persons employed in the chemical industry to those not employed in this industry. On the other hand, the relative paucity of information on worker exposures makes it difficult to infer whether any observed associations with mortality or cancer incidence were attributable to chemical exposures per se. For causal inference, it is important to be able to examine the risk of disease according to the amount of exposure. Only a weak surrogate measure of dose of exposure, that is, length of employment, was possible in the present analysis.

“Publication” Bias

These meta-analyses were conducted only on studies that were published in the peer-reviewed literature. Not all of the studies ever conducted have been published, nor have all results obtained been reported in the published studies. There are a number of reasons that some completed occupational cohort studies have not been published. For example, mortality surveillance programs performed either by individual industries, trade associations, regulatory agencies, or the academic community can be performed for specific purposes (for example, corporate stewardship, academic research, or legislatively mandated population surveillance) without publishing the results. Within the private sector, even when elevated SMRs are observed, companies can disclose data, either voluntarily or by mandate, to the appropriate government agencies without publication of the results in peer-reviewed journals. Furthermore, studies with negative or null results may have difficulty gaining acceptance in peer-reviewed journals. Thus, the number of unpublished studies that would be relevant to these meta-analyses is not known.

For studies that were included, the authors generally gave two reasons for undertaking the studies: (1) there was evidence for, or suspicion of, an adverse health risk, or (2) the study was conducted as part of a general health surveillance program. Studies conducted for the former reason might be more likely to produce positive findings; this is especially true for case-control studies of industrial populations. When available, we abstracted the authors’ rationale for initiating the studies, but the explanation was often incomplete or insufficient, making it impractical to conduct meta-analyses stratified by “reason for study.” Nevertheless, occupational groups for which there is little suspicion of increased risk of disease are not as likely to be studied or reported as those with suspected health concerns. Thus, in general the published literature, and hence our meta-analyses, are weighted with cohorts at higher suspicion for elevated morbidity or mortality rates. This weighting would have been even higher had case-control studies been included.

Selection Bias

By limiting this review to studies published in the English language, conducted in North America and western Europe, and published in peer-reviewed journals since 1966, the population base was restricted. Although these restrictions may limit the ability to extrapolate results to other time periods and countries, they should not affect the validity for the populations considered.

The studies included in this review had to meet certain eligibility criteria and therefore do not include all the studies on specific chemicals. Therefore, conclusions regarding specific disease-chemical relationships cannot be made without considering all of the available data.

Variability in Reporting

Individual studies exhibited considerable variability in the amount of information reported in tables and texts, with many authors choosing to report, selectively, analyses that resulted in elevated SMRs and not to report many SMRs that were presumably not elevated. For example, of the 168 full-cohort mortality studies evaluated in this meta-analysis, 1–6,8–19,21–49,51–77,79–87,89–102,104–120,122–138,140–165,167,168,170–179,181 150 studies presented SMRs for all malignant neoplasms, 1–4,8–19,21–25,27–31,33,35–47,49,51–53,56–60,62,63,65–77,79–81,83–87,91–94,96–102,104–111,113–120,122–127,130–138,140–143,145–154,156–165,167,168,170–172,174–179,181 99 for cardiovascular diseases, 9,10,12–14,17–19,21–23,25,27,30,32,33,35,36,38,39,42,44,46–49,52,53,56,60,62,63,65–68,70,71,73,74,76,77,79,80,85,87,90–94,96,98,100,104,105,107,110,114,116,117–120,122,124,127,133–138,140,141,145,146,149,150,152–154,157,160,161,163,164,167,168,170,172,174–179,181 83 for external causes of death, 2,4–6,12–14,19,21,23–25,27,28,30,32,36,38,39,42–44,46–49,53,54,62,65–68,74,76,77,79,80,83,85,87,91–94,98,102,104,105,107,118–120,124,125,127,130,133–137,140,141,149–153,159–163,167,170,172,175,176–179,181 120 for lung cancer, 1–6,8,10,12–15,17–19,21–24,28–31,33–39,41–47,51–55,57–60,62,63,65,67–71,74–77,83–87,89,91,92,95,96,98–101,104,106–110,112,113–117,122,124,127,130,131,134–138,141,143,145,146,149–151,153,154,161–165,167,168,170,172,176–179,181 and 77 for brain cancer. 3–6,10–12,15,17–19,21,23–25,27–29,33,37,39,41–44,46,47,51,52,56,58,68,76,83,84,87,99,100,102,104,105,107,108,112–114,116,117,120,126,127,131,135–137,141,142,145–147,151,153,154,158,159,162–165,170–172,174,175,177,178,181 This selective reporting also occurred in the cancer incidence studies, in which, of the 25 studies analyzed, 1,7,9,20,38,48,49,50,63,65,66,78,82,88,91,103,113,115,119,121,129,139,145,147,169 19 studies provided data on all malignant neoplasms, 1,7,9,20,38,48–50,63,65,66,82,91,103,113,115,119,121,145 17 on lung cancer, 1,7,9,20,38,48,40,63,65,66,82,88,103,113,115,119,145 14 on prostate cancer, 7,9,20,38,48,63,66,82,103,113,115,119,121,145 and only 5 on multiple myeloma. 7,48,66,91,145

For the reasons just discussed, it is likely that the selective underreporting of subgroup analyses in the original studies resulted in a reporting bias toward elevated or statistically significant results.

Independence of Cohorts

We made every effort to avoid duplicate observations of individual cohort members by abstracting results primarily from the most recently updated publication regarding a cohort. Still, the possibility remains that we inadvertently included cohorts containing overlapping populations reported in multiple publications in the meta-analyses, and thus, the mortality and morbidity experience of such populations would be inappropriately overrepresented in the results.

Summary

In summary, in this investigation we reviewed and analyzed a large number of published studies of workers in the chemical industry and found that, with a few exceptions, their mortality and cancer morbidity experience is similar to or better than that of the general population. We observed excesses of bronchus, trachea, and lung cancers and bladder cancer that were consistent with exposure to known chemical carcinogens. A few unanswered questions remain with respect to some of the apparent excesses, such as the somewhat consistently elevated findings for brain and central nervous system cancers, and especially the observation of elevated SMRs and an elevated SIR for lymphatic and hematopoietic cancers. This category is difficult to evaluate for a number of reasons. Approximately half the studies provided data on lymphatic and hematopoietic cancers, with the classifications used across studies varying considerably. We found it was often difficult to group subcategories on the basis of the limited information provided in the study, and virtually impossible to distinguish among specific types of leukemias or lymphomas. As the diseases that constitute lymphatic and hematopoietic cancers probably have different etiologies, it is important to have a more precise breakdown of these data to properly assess the results. Important difficulties with disease nomenclature and with coding, as well as the limited reporting in these published studies, hamper a clear assessment of which specific cancers within this general category are in excess among chemical workers.

References

1. Acquavella JF, Riordan SG, Anne M, Lynch CF, Collins JJ, Ireland BK, Heydens WE. Evaluation of mortality and cancer incidence among alachlor manufacturing workers. Environ Health Perspect 1996; 104: 728–733.
2. Al-Dabbagh S, Forman D, Bryson D, Stratton I, Doll R. Mortality of nitrate fertiliser workers. Br J Ind Med 1986; 43: 507–515.
3. Alderson MR, Rattan NS. Mortality of workers on an isopropyl alcohol plant and two MEK dewaxing plants. Br J Ind Med 1980; 37: 85–89.
4. Amoateng-Adjepong Y, Sathiakumar N, Delzell E, Cole P. Mortality among workers at a pesticide manufacturing plant. J Occup Environ Med 1995; 37: 471–478.
5. Andjelkovich DA, Taulbee J, Symons M. Mortality experience of a cohort of rubber workers, 1964–1973. J Occup Med 1976; 18: 387–394.
6. Andjelkovich DA, Taulbee J, Blum S. Mortality of female workers in a rubber manufacturing plant. J Occup Med 1978; 20: 409–413.
7. Anttila A, Pukkala E, Sallmen M, Hernberg S, Hemminki K. Cancer incidence among Finnish workers exposed to halogenated hydrocarbons. J Occup Environ Med 1995; 37: 797–806.
8. Armstrong BG, Kazantzis G. The mortality of cadmium workers. Lancet 1983; 1: 1425–1427.
9. Barregard L, Sallsten G, Jarvholm B. Mortality and cancer incidence in chloralkali workers exposed to inorganic mercury. Br J Ind Med 1990; 47: 99–104.
10. Becher H, Flesch-Janys D, Kauppinen T, Kogevinas M, Steindorf K, Manz A, Wahrendorf J. Cancer mortality in German male workers exposed to phenoxy herbicides and dioxins. Cancer Causes Control 1996; 7: 312–321.
11. Benson LO, Teta MJ. Mortality due to pancreatic and lymphopoietic cancers in chlorohydrin production workers. Br J Ind Med 1993; 50: 710–716.
12. Bernardinelli L, De Marco R, Tinelli C. Cancer mortality in an Italian rubber factory. Br J Ind Med 1987; 44: 187–191.
13. Bertazzi PA, Pesatori AC, Radice L, Zocchetti C, Vai T. Exposure to formaldehyde and cancer mortality in a cohort of workers producing resins. Scand J Work Environ Health 1986; 12: 461–468.
14. Bisanti L, Maggini M, Raschetti R, Alegiani SS, Ippolito FM, Caffari B, Segnan N, Ponti A. Cancer mortality in ethylene oxide workers. Br J Ind Med 1993; 50: 317–324.
15. Blair A, Stewart PA, O’Berg MT, Gaffey WR, Walrath J, Ward J, Bales R, Kaplan SD, Cubit D. Mortality among industrial workers exposed to formaldehyde. J Natl Cancer Inst 1986; 76: 1071–1084.
16. Bloemen U, Mandel JS, Bond GG, Pollock AF, Vitek RP, Cook RR. An update of mortality among chemical workers potentially exposed to the herbicide 2,4-dichlorophenoxyacetic acid and its derivatives. J Occup Med 1993; 3 5: 1208–1212.
17. Bond GG, McLaren EA, Cartmill JB, Wymer KT, Lipps TE, Cook RR. Mortality among female employees of a chemical company. Am J Ind Med 1987; 12: 563–578.
18. Bond GG, McLaren EA, Cartmill JB, Wymer KT, Sobel W, Lipps TE, Cook RR. Cause-specific mortality among male chemical workers. Am J Ind Med 1987; 12: 353–383.
19. Bond GG, Reeve GR, Ott MG, Waxweiler RJ. Mortality among a sample of chemical company employees. Am J Ind Med 1985; 7: 109–121.
20. Bond GG, Austin DF, Gondek MR, Chiang M, Cook RR. Use of a population-based tumor registry to estimate cancer incidence among a cohort of chemical workers. J Occup Med 1988; 30: 443–448.
21. 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.
22. Burchfiel CM, Cartmill JB, Axe FD, Bond GG. General mortality and respiratory cancer among a cohort of male chemical workers in California. Am J Ind Med 1992; 22: 69–83.
23. Carlo GL, Jablinske MR, Lee NL, Sund KG, Corn M. Reduced mortality among workers at a rubber plant. J Occup Med 1993; 35: 611–616.
24. Checkoway H, Heyer NJ, Demers PA. An updated mortality follow-up study of Florida phosphate industry workers. Am J Ind Med 1996; 30: 452–460.
25. Chiazze L, Wolf P, Ference LD. An historical cohort study of mortality among salaried research and development workers of the Allied Corporation. J Occup Med 1986; 28: 1185–1188.
26. Collingwood KW, Pasternack BS, Shore RE. An industry-wide study of respiratory cancer in chemical workers exposed to chloromethyl ethers. J Natl Cancer Inst 1987; 78: 1127–1136.
27. Collins JJ, Page LC, Caporossi JC, Utidjian HM, Saipher IN. Mortality patterns among men exposed to methyl methacrylate. J Occup Med 1989; 31: 41–46.
28. Collins JJ, Page LC, Caporossi JC, Utidjian HM, Lucas LJ. Mortality patterns among employees exposed to acrylonitrile. J Occup Med 1989; 31: 368–371.
29. Collins JJ, Swaen GMH, Marsh GM, Utidjian HM, Caporossi JC, Lucas LJ. Mortality patterns among workers exposed to acrylamide. J Occup Med 1989; 31: 614–617.
30. Cowles SR, Tsai SP, Gilstrap EL, Ross CE. Mortality among employees at a plastics and resins research and development facility. Occup Environ Med 1994; 51: 799–803.
31. Craig R, Gillis CR, Hole DJ, Paddle GM. Sixteen year follow up of workers in an explosives factory. J Soc Occup Med 1985; 35: 107–110.
32. Davies JM. Long term mortality study of chromate pigment workers who suffered lead poisoning. Br J Ind Med 1984; 4: 170–178.
33. Davies JM, Easton DE, Bidstrup PL. Mortality from respiratory cancer and other causes in United Kingdom chromate production workers. Br J Ind Med 1991; 48: 299–313.
34. Davies JM. Lung cancer mortality among workers making lead chromate and zinc chromate pigments at three English factories. Br J Ind Med 1984; 4: 158–169.
35. De Jong G, Swaen GMH, Slangen JJM. Mortality of workers exposed to dieldrin and aldrin: a retrospective cohort study. Occup Environ Med 1997; 54: 702–707.
36. Decoufle P, Blattuer WA, Blair A. Mortality among chemical workers exposed to benzene and other agents. Environ Res 1983; 30: 16–25.
37. Dell L, Teta MJ. Mortality among workers at a plastics manufacturing and research and development facility: 1946–1988. Am J Ind Med 1995; 28: 373–384.
38. Delzell E, Louik C, Lewis RJ, Monson RR. Mortality and cancer morbidity among workers in the rubber tire industry. Am J Ind Med 1981; 2: 209–216.
39. Delzell E, Monson RR. Mortality among rubber workers. III. Cause-specific mortality, 1940–1978. J Occup Med 1981; 23: 677–684.
40. Delzell E, Sathiakumar N, Hovinga M, Macaluso M, Julian J, Larson R, Cole P, Muir D. A follow-up study of synthetic rubber workers. Toxicology 1996; 113: 182–189.
41. Delzell E, Macaluso M, Cole P. A follow-up study of workers at a dye and resin manufacturing plant. J Occup Med 1989; 31: 273–278.
42. Deschamps F, Moulin JJ, Wild P, Labriffe H, Haguenoer JM. Mortality study among workers producing chromate pigments in France. Int Arch Occup Environ Health 1995; 67: 147–152.
43. Divine BJ, Hartman CM. Mortality update of butadiene production workers. Toxicology 1996; 113: 169–181.
44. Divine BJ, Barron V, Kaplan SD. Texaco mortality study. I. Mortality among refinery, petrochemical, and research workers. J Occup Med 1985; 27: 445–447.
45. Duck BW, Carter JT, Coombes EJ. Mortality study of workers in a polyvinyl-chloride production plant. Lancet 1975; 2: 1197–1199.
46. Easton DF, Peto J, Doll R. Cancers of the respiratory tract in mustard gas workers. Br J Ind Med 1988; 45: 652–659.
47. Egedahl RD, Olsen GW, Coppock E, Young ML, Arnold IMF. An historical prospective mortality study of the Sarnia Division of Dow Chemical Canada Inc., Sarnia, Ontario (1950–1984). Can J Public Health 1989; 80: 441–446.
48. Ellingsen DG, Andersen A, Nordhagen HP, Efskind J, Kjuus H. Incidence of cancer and mortality among workers exposed to mercury vapour in the Norwegian chloralkali industry. Br J Ind Med 1993; 50: 875–880.
49. Englander V, Sjoberg A, Hagmar L, Attewell R, Schutz A, Moller T, Skerfving S. Mortality and cancer morbidity in workers exposed to sulphur dioxide in a sulphuric acid plant. Int Arch Occup Environ Health 1988; 61: 157–162.
50. Fandrem SI, Kjuus H, Andersen A, Amlie E. Incidence of cancer among workers in a Norwegian nitrate fertiliser plant. Br J Ind Med 1993; 50: 647–652.
51. Fingerhut MA, Halperin W, Marlow DA, Piacitelli LA, Honchar PA, Sweeney MH, Greife AL, Dill PA, Steenland K, Suruda AJ. Cancer mortality in workers exposed to 2,3,7,8-tetrachlorodibenzo-p-dioxin. N Engl J Med 1991; 324: 212–218.
52. Fox AJ, Collier PF. A survey of occupational cancer in the rubber and cablemaking industries: analysis of deaths occurring in 1972–74. Br J Ind Med 1976; 33: 249–264.
53. Fraser P, Chilvers C, Day M, Goldblatt P. Further results from a census based mortality study of fertiliser manufacturers. Br J Ind Med 1989; 46: 38–42.
54. 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.
55. Frentzel-Beyme R. Lung cancer mortality of workers employed in chromate pigment factories: a multicentric European epidemiological study. J Cancer Res Clin Oncol 1983; 105: 183–188.
56. Gardiner JS, Walker SA, MacLean AJ. A retrospective mortality study of substituted anthraquinone dyestuffs workers. Br J Ind Med 1982; 39: 355–360.
57. Gardner MJ, Coggon D, Pannett B, Harris EC. Workers exposed to ethylene oxide: a follow-up study. Br J Ind Med 1989; 46: 860–865.
58. Gardner MJ, Pannett B, Winter PD, Cruddas AM. A cohort study of workers exposed to formaldehyde in the British chemical industry: an update. Br J Ind Med 1993; 50: 827–834.
59. Gibbs GW, Amsel J, Soden KJ. A cohort mortality study of cellulose triacetate-fiber workers exposed to methylene chloride. J Occup Environ Med 1996; 38: 693–697.
60. Gilliland FD, Mandel JS. Mortality among employees of a perfluorooctanoic acid production plant. J Occup Med 1993; 35: 950–954.
61. Goldberg MS, Theriault G. Retrospective cohort study of workers of a synthetic textiles plant in Quebec. II. Colorectal cancer mortality and incidence. Am J Ind Med 1994; 25: 909–922.
62. Guberan E, Usel M. Unusual mortality pattern among short term workers in the perfumery industry in Geneva. Br J Ind Med 1987; 44: 595–601.
63. Guberan E, Raymond LW. Mortality and cancer incidence in the perfumery and flavour industry of Geneva. Br J Ind Med 1985; 42: 240–245.
64. Gustavsson A, Hogstedt C, Holmberg B. Mortality and incidence of cancer among Swedish rubber workers, 1952–1981. Scand J Work Environ Health 1986; 12: 538–544.
65. Hagmar L, Bellander T, Eng M, Englander V, Ranstam J, Attewell R, Skerfving S. Mortality and cancer morbidity among workers in a chemical factory. Scand J Work Environ Health 1986; 12: 545–551.
66. Hagmar L, Welinder H, Mikoczy Z. Cancer incidence and mortality in the Swedish polyurethane foam manufacturing industry. Br J Ind Med 1993; 50: 537–543.
67. Hayes RB, Lilienfeld AM, Snell LM. Mortality in chromium chemical production workers: a prospective study. Int J Epidemiol 1979; 8: 365–374.
68. Hearne FT, Pifer JW, Grose F. Absence of adverse mortality effects in workers exposed to methylene chloride: an update. J Occup Med 1990; 32: 234–240.
69. Heldaas SS, Andersen A, Langard S. Incidence of cancer among vinyl chloride and polyvinyl chloride workers: further evidence for an association with malignant melanoma. Br J Ind Med 1987; 44: 278–280.
70. Hodgson JT, Jones RD. A mortality study of carbon black workers employed at five United Kingdom factories between 1947and 1980. Arch Environ Health 1985; 40: 261–268.
71. 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.
72. Hogstedt C. Epidemiological studies on ethylene oxide and cancer: an updating. IARC Sci Publ 1988; 89: 265–270.
73. Hogstedt C, Andersson K. A cohort study of mortality among dynamite workers. J Occup Med 1979; 21: 553–556.
74. Holmes TM, Buffler PA, Holguin AH, Hsi BP. A mortality study of employees at a synthetic rubber manufacturing plant. Am J Ind Med 1986; 9: 355–362.
75. Hours M, Cardis E, Marciniak A, Quelin P, Fabry J. Mortality of a cohort in a polyamide-polyester factory in Lyon: a further follow up. Br J Ind Med 1989; 46: 665–670.
76. Huebner WW, Schnatter AR, Nicolich MJ, Jorgensen G. Mortality experience of a young petrochemical industry cohort: 1979–1992 follow-up study of US-based employees. J Occup Environ Med 1997; 39: 970–982.
77. Ietri E, Belli S, Comba P, Gerosa A, Raffi GB, Pirastu R. Cohort mortality study of rubber and plastics product makers in Italy. Occup Med 1997; 47: 417–422.
78. Kaleja R, Horbach L, Amsel J. Polypropylene production workers and colorectal cancer in Germany: a brief report. J Occup Environ Med 1994; 51: 784–785.
79. Kiesselbach N, Ulm K, Lange HJ, Korallus U. A multicentre mortality study of workers exposed to ethylene oxide. Br J Ind Med 1990; 47: 182–188.
80. Kilpikari I. Mortality among male rubber workers in Finland. Arch Environ Health 1982; 37: 295–299.
81. Kilpikari I, Pukiala E, Lehtonen M, Hakama M. Cancer incidence among Finnish rubber workers. Int Arch Occup Environ Health 1982; 51: 65–71.
82. Kjuus H, Andersen A, Langard S. Incidence of cancer among workers producing calcium carbide. Br J Ind Med 1986; 43: 237–242.
83. Kogevinas M, Ferro G, Andersen A, Bellander T, Biocca M, Coggon D, Gennaro V, Hutchings S, Kolstad HA, Lundberg I, Lynge E, Partanen T, Saracci R. Cancer mortality in a historical cohort study of workers exposed to styrene. Scand J Work Environ Health 1994; 20: 251–261.
84. Kogevinas M, Becher H, Benn T, Bertazzi PA, Boffetta P, Bueno de Mesquita BH, Coggon D, Cohn D, Flesch-Janys D, Fingerhut MA, Green LM, Kauppinen T, Littorin M, Lynge E, Mathews JD, Neuberger M, Pearce N, Saracci R. Cancer mortality in workers exposed to phenoxy herbicides, chlorophenols, and dioxins: an expanded and updated international cohort study. Am J Epidemiol 1997; 145: 1061–1075.
85. Korallus U, Ulm K, Steinmann-Steiner-Haldenstaett W. Bronchial carcinoma mortality in the German chromate-producing industry: the effects of process modification. Int Arch Occup Environ Health 1993; 65: 171–178.
86. Lanes SF, Rothman KJ, Dreyer NA, Soden KJ. Mortality update of cellulose fiber production workers. Scand J Work Environ Health 1993; 19: 426–428.
87. Lanes SF, Rothman KJ, Soden KJ, Amsel J, Dreyer NA. Mortality among synthetic fiber workers exposed to glycerol polyglycidyl ether. Am J Ind Med 1994; 25: 689–696.
88. Langard S, Vigander T. Occurrence of lung cancer in workers producing chromium pigments. Br J Ind Med 1983; 40: 71–74.
89. Lemen PA, Johnson WM, Wagoner JK, Archer VE, Saccomanno G. Cytologic observations and cancer incidence following exposure to BCME. Ann NY Acad Sci 1976; 71: 71–80.
90. Levine RJ, Andjelkovich DA, Kersteter SL, Arp EW, Balogh SA, Blunden PB, Stanley JM. Heart disease in workers exposed to dinitrotoluene. J Occup Med 1986; 28: 811–816.
91. Lundberg I. Mortality and cancer incidence among Swedish paint industry workers with long-term exposure to organic solvents. Scand J Work Environ Health 1986; 12: 108–113.
92. Mabuchi K, Lilienfeld AM, Snell LM. Lnng cancer among pesticide workers exposed to inorganic arsenicals. Arch Environ Health 1979; 34: 312–320.
93. MacMahon B, Monson RR. Mortality in the US rayon industry. J Occup Med 1988; 30: 698–705.
94. Maher KV, DeFonso LR. An historical cohort study of mortality among chemical researchers. Arch Environ Health 1986; 41: 109–116.
95. Mancuso TF. Chromium as an industrial carcinogen. Part I. Am J Ind Med 1997; 31: 129–139.
96. Mancuso TF, El-Attar AA. Cohort study of workers exposed to betanaphthylamine and benzidine. J Occup Med 1967; 9: 277–285.
97. Marsh GM, Leviton LC, Talbott EO, Callahan C, Pavlock D, Hemstreet GP, Logue JN, Fox AJ, Schulte PA. Drake Chemical Workers’ Health Registry Study. I. Notification and medical surveillance of a group of workers at high risk of developing bladder cancer. Am J Ind Med 1991; 19: 291–301.
98. Marsh GM. Mortality among workers from a plastics producing plant: a matched case-control study nested in a retrospective cohort study. J Occup Med 1983; 25: 219–230.
99. Marsh GM, Stone RA, Esmen NA, Henderson VL, Lee KY. Mortality among chemical workers in a factory where formaldehyde was used. Occup Environ Med 1996; 53: 613–627.
100. Mastrangelo G, Serena R, Marzia V. Mortality from tumours in workers in an acrylic fibre factory. Occup Med 1993; 43: 155–158.
101. McCallum RI, Woolley V, Petrie A. Lung cancer associated with chloromethyl methyl ether manufacture: an investigation at two factories in the United Kingdom. Br J Ind Med 1983; 40: 384–389.
102. McMichael AJ, Spirtas R, Kupper LL. An epidemiologic study of mortality within a cohort of rubber workers, 1964–72. J Occup Med 1974; 16: 458–464.
103. Meigs JW, Marrett LD, Ulrich FU, Flannery JT. Bladder tumor incidence among workers exposed to benzidine: a thirty-year follow-up. J Natl Cancer Inst 1986; 76: 1–8.
104. Morgan RW, Hyg SM, Kaplan SD, Gaffey WR. A general mortality study of production workers in the paint and coatings manufacturing industry: a preliminary report. J Occup Med 1981; 23: 13–21.
105. Morgan RW, Claxton KW, Divine BJ, Kaplan SD, Harris VB. Mortality among ethylene oxide workers. J Occup Med 1981; 23: 767–770.
106. Morris SN, Densem JW, Wald NJ, Doll R. Occupational exposure to hydrazine and subsequent risk of cancer. Occup Environ Med 1995; 52: 43–45.
107. Moulin JJ, Wild P, Mur JM, Fournier-Betz M, Mercier-Gallay M. A mortality study of cobalt production workers: an extension of the follow up. Am J Ind Med 1993; 23: 281–288.
108. Negri E, Piolatto G, Pira E, Decarli A, Kaldor J, La Vecchia C. Cancer mortality in a northern Italian cohort of rubber workers. Br J Ind Med 1989; 46: 624–628.
109. Newhouse ML, Matthews G, Sheikh K, Knight KL, Oakes D, Sullivan KR. Mortality of workers at acetylene production plants. Br J Ind Med 1988; 45: 63–69.
110. Nicholson WJ, Selikoff IJ, Seidman H. Mortality experience of styrene-polystyrene polymerization workers: initial findings. Scand J Work Environ Health 1978; 4: 247–252.
111. Nicholson WJ, Hammond EC, Seidman H, Selikoff IJ. Mortality experience of a cohort of vinyl chloride-polyvinyl chloride workers. Ann NY Acad Sci 1975; 246: 225–230.
112. Norseth T, Andersen A, Giltvedt J. Cancer incidence in the rubber industry in Norway. Scand J Work Environ Health 1983; 9: 69–71.
113. O’Berg MT, Burke CA, Chen JL, Wairath J, Pell S, Gallie CR. Cancer incidence and mortality in the Du Pont Company: an update. J Occup Med 1987; 29: 245–252.
114. Olsen GW, Lacy SE, Cartmill JB, Kravat BA, Chamberlin SR, Spadacene NW, Lipps TE. Half-century of cause-specific mortality experience of chemical manufacturing employees. Am J Ind Med 1994; 26: 203–219.
115. Olsen GW, Gondek MR, Flannery JT, Cartmill JB, Bodner KM. A thirty-eight year comparison of cancer incidence and mortality among employees at a Connecticut chemical plant. Conn Med 1997; 61: 83–89.
116. Olsen GW, Bodner KM, Stafford BA, Cartmill JB, Gondek MR. Update of the mortality experience of employees with occupational exposure to 1,2-dibromo-3-chloropropane (DBCP). Am J Ind Med 1995; 28: 399–410.
117. Olsen GW, Lacy SE, Bodner KIM, Chau M, Arceneaux TG, Cartmill JB, Ramlow JM, Boswell JM. Mortality from pancreatic and lymphopoietic cancer mortality among workers in ethylene and propylene chlorohydrin production. Occup Environ Med 1997; 54: 592–598.
118. Ott MG, Skory LK, Holder BB, Bronson JM, Williams PR. Health evaluation of employees occupationally exposed to methylene chloride. Scand J Work Environ Health 1983; 9: 8–16.
119. Ott MG, Zober A. Cause specific mortality and cancer incidence among employees exposed to 2,3,7,8-TCDD after a 1953 reactor accident. Occup Environ Med 1996; 53: 606–612.
120. Ott MG. Mortality experience among Louisiana chemical manufacturing employees, 1957-1992. J La State Med Soc 1996; 148: 260–266.
121. Ouellet-Hellstrom R, Rench JD. Bladder cancer incidence in arylamine workers. J Occup Environ Med 1996; 38: 1239–1247.
122. Paddle GM, Osborn AJ, Parker GDJ. Mortality of employees in plants manufacturing 4,4′-bipyridyl. Scand J Work Environ Health 1991; 17: 175–178.
123. Pasternack BS, Shore RE, Albert RE. Occupational exposure to chloromethyl ethers: a retrospective cohort mortality study (1948–1972). J Occup Med 1977: 19: 741–746.
124. Pastides H, Austin R, Lemeshow S, Klar J, Mundt KA. A retrospective-cohort study of occupational exposure to hexavalent chromium. Am J Ind Med 1994; 25: 663–675.
125. Paxton MB, Chinchilli VM, Brett SM, Rodricks JV. Leukemia risk associated with benzene exposure in the Pliofilm cohort. I. Mortality update and exposure distribution. Risk Anal 1994; 14: 147–154.
126. Pifer JW, Hearne FT, Friedlander BR, McDonough JR. Mortality study of men employed at a large chemical plant, 1972 through 1982. J Occup Med 1986; 28: 438–444.
127. Pifer JW, Hearne FT, Swanson FA, O’Donoghue JL. Mortality study of employees engaged in the manufacture and use of hydroquinone. Int Arch Occup Environ Health 1995; 67: 267–280.
128. Piolatto G, Negri E, La Vecchia C, Pira E, Decarli A, Peto J. Bladder cancer mortality of workers exposed to aromatic amines: an updated analysis. Br J Cancer 1991; 63: 457–459.
129. Popp W, Schmieding W, Speck M, Vahrenholz C, Norpoth K. Incidence of bladder cancer in a cohort of workers exposed to 4-chloro-toluidine while synthesising chlordimeform. Br J Ind Med 1992; 49: 529–531.
130. Rafnsson V, Gunnarsdottir H. Mortality study of fertiliser manufacturers in Iceland. Br J Ind Med 1990; 47: 721–725.
131. Rapiti E, Fantini F, Dell_Orco V, Fano V, Blasetti F, Bracci C, Forastiere F, Comba P. Cancer mortality among chemical workers in an Italian plant. Eur J Epidemiol 1997; 13: 281–285.
132. Robertson JM, Ingalls TH. A mortality study of carbon black workers in the United States from 1935 to 1974. Arch Environ Health 1980; 35: 181–186.
133. Robinson TR. Twenty-year mortality of tetraethyl lead workers. J Occup Med 1974; 16: 601–605.
134. Rubino GF, Scansetti G, Piolatto G, Pira E. The carcinogenic effect of aromatic amines: an epidemiological study on the role of o-toluidine and 4,4′-methylene bis (2-methylaniline) in inducing bladder cancer in man. Environ Res 1982; 27: 241–254.
135. Sathiakumar N, Delzell E, Cole P. Mortality among workers at two triazine herbicide manufacturing plants. Am J Ind Med 1996; 29: 143–151.
136. Satin KY, Wong O, Yuan LA, Bailey WJ, Newton KL, Wen CP, Swencicki RE. A 50 year mortality follow up of a large cohort of oil refinery workers in Texas. J Occup Environ Med 1996; 38: 492–506.
137. Schnatter AR, Theriault G, Katz AM, Thompson ES, Donaleski DL, Murray N. A retrospective mortality study within operating segments of a petroleum company. Am J Ind Med 1992; 22: 209–229.
138. Schnorr TM, Steenland K, Egeland GM, Boeniger M, Egilman D. Mortality of workers exposed to toluene diisocyanate in the polyurethane foam industry. Occup Environ Med 1996; 53: 703–707.
139. Schulte PA, Ringen K, Hemstreet GP, Altekruse EB, Gullen MTH, Patton MG, Allsbrook WC, Crosby JH, West SS, Witherington R, Koss LG, Bales CE, Tillet S, Rooks SCF, Stern FB, Stringer W, Schmidt VA, Brubaker MM. Risk assessment of a cohort exposed to aromatic amines. J Occup Med 1985; 27: 115–121.
140. Shindell S, Ulrich S. Mortality of workers employed in the manufacture of chlordane: an update. J Occup Med 1986; 28: 497–501.
141. Simonato L, L’Abbe KA, Andersen A, Belli S, Comba P, Engholm G, Ferro G, Hagmar L, Langard S, Lundberg I, Pirastu R, Thomas PG, Winkelmann R, Saracci R. A collaborative study of cancer incidence and mortality among vinyl chloride workers. Scand J Work Environ Health 1991; 17: 159–169.
142. Sorahan T, Parkes HG, Veys CA, Waterhouse JAH, Straughan JK, Nutt A. Mortality in the British rubber industry 1946–85. Br J Ind Med 1989; 46: 1–10.
143. Sorahan T, Cathcart M. Lung cancer mortality among workers in a factory manufacturing chlorinated toluenes: 1961–84. Br J Ind Med 1989; 46: 425–427.
144. Sorahan T, Waterhouse JAH, Cooke MA, Smith EMB, Jackson JR, Temkin L. A mortality study of workers in a factory manufacturing chlorinated toluenes. Ann Occup Hyg 1983; 27: 173–182.
145. Sorahan T, Pope D. Mortality and cancer morbidity of production workers in the United Kingdom flexible polyurethane foam industry. Br J Ind Med 1993; 50: 528–536.
146. Sorahan T, Pope D. Mortality study of workers employed at a plant manufacturing chemicals for the rubber industry: 1955–86. Br J Ind Med 1993; 50: 998–1002.
147. Stasik MJ. Carcinomas of the urinary bladder in a 4-chloro-o-toluidine cohort. Int Arch Occup Environ Health 1988; 60: 21–24.
148. Stayner LT, Dannenberg AL, Bloom TF, Thun M. Excess hepatobiliary cancer mortality among munitions workers exposed to dinitrotoluene. J Occup Med 1993; 35: 291–296.
149. Stayner LT, Meinhardt TJ, Lemen RA, Bayliss D, Herrick R, Reeve GR, Smith AB, Halperin W. A retrospective cohort mortality study of a phosphate fertilizer production facility. Arch Environ Health 1985; 40: 133–138.
150. Stern FB, Murthy LI, Beaumont JJ, Schulte PA, Halperin W. Notification and risk assessment for bladder cancer of a cohort exposed to aromatic amines. J Occup Med 1985; 27: 495–500.
151. Strauss ME, Barrick ED, Bannister RM. Mortality experience of employees exposed to 2-mercaptobenzothiazole at a chemical plant in Nitro, West Virginia. Br J Ind Med 1993; 50: 888–893.
152. Swaen GMH, Braun C, Slangen JJM. Mortality of Dutch workers exposed to carbon disulfide. Int Arch Occup Environ Health 1994; 66: 103–110.
153. Swaen GMH, Bloemen U, Twisk J, Scheffers T, Slangen JJM, Sturmans F. Mortality of workers exposed to acrylonitrile. J Occup Med 1992; 34: 801–809.
154. Sweeney MH, Beaumont JJ, Waxweiler RJ, Halperin W. An investigation of mortality from cancer and other causes of death among workers employed at an east Texas chemical plant. Arch Environ Health 1986; 41: 23–28.
155. Sweetnam PM, Taylor SW, Elwood PC. Exposure to carbon disulphide and ischaemic heart disease in a viscose rayon factory. Br J Ind Med 1987; 44: 220–227.
156. Tassignon JP, Bos GD, Craigen AA, Jacquet B, Kueng HL, Lanouziere-Simon C, Pierre C. Mortality in an European cohort occupationally exposed to epichlorohydrin (ECH). Int Arch Occup Environ Health 1983; 51: 325–336.
157. Taylor FH. The relationship of mortality and duration of employment as reflected by a cohort of chromate workers. Am J Public Health 1966; 56: 218–229.
158. Teta MJ, Ott MG, Schnatter AR. An update of mortality due to brain neoplasms and other causes among employees of a petrochemical facility. J Occup Med 1991; 33: 45–51.
159. Teta MJ, Schnatter AR, Ott MG, Pell S. Mortality surveillance in a large chemical company: the Union Carbide Corporation experience, 1974–1983. Am J Ind Med 1990; 17: 435–447.
160. Theriault G, Allard P. Cancer mortality of a group of Canadian workers exposed to vinyl chloride monomer. J Occup Med 1981; 23: 671–676.
161. Thomas TL. Mortality among flavour and fragrance chemical plant workers in the United States. Br J Ind Med 1987; 44: 733–737.
162. Tomenson JA, Bonner SM, Heijne CG, Farrar DG, Cummings TF. Mortality of workers exposed to methylene chloride employed at a plant producing cellulose triacetate film base. Occup Environ Med 1997; 54: 470–476.
163. Tsai SP, Gilstrap EL, Cowles SR, Snyder PJ, Ross CE. A cohort mortality study of two California refinery and petrochemical plants. J Occup Med 1993; 35: 415–421.
164. Tsai SP, Gilstrap EL, Colangelo TA, Menard AK, Ross CE. A mortality study of oil refinery and petrochemical employees. J Occup Environ Med 1997; 39: 448–454.
165. Tsai SP, Gilstrap EL, Cowles SR, Snyder PJ, Ross CE. Long-term follow-up mortality study of petroleum refinery and chemical plant employees. Am J Ind Med 1996; 29: 75–87.
166. Walker AM, Cohen AJ, Loughlin JE, Rothman KJ, DeFonso LR. Mortality from cancer of the colon or rectum among workers exposed to ethyl acrylate and methyl methacrylate. Scand J Work Environ Health 1991; 17: 7–19.
167. Wang HH, MacMahon B. Mortality of workers employed in the manufacture of chlordane and heptachlor. J Occup Med 1979; 21: 745–748.
168. Ward EM, Fajen JM, Ruder AM, Riusky RA, Halperin W, Fessler-Flesch CA. Mortality study of workers in 1,3-butadiene production units identified from a chemical workers cohort. Environ Health Perspect 1995; 103: 598–603.
169. Ward EM, Carpenter A, Markowitz S, Roberts D, Halperin W. Excess number of bladder cancers in workers exposed to ortho-toluidine and aniline. J Natl Cancer Inst 1991; 83: 501–506.
170. Waxweiler RJ, Alexander V, Leffingwell SS, Haring M, Lloyd JW. Mortality from brain tumor and other causes in a cohort of petrochemical workers. J Natl Cancer Inst 1983; 70: 75–81.
171. Weber H, Reinl W, Greiser E. German investigations on morbidity and mortality of workers exposed to vinyl chloride. Environ Health Perspect 1981; 41: 95–99.
172. Weiland SK, Mundt KA, Keil U, Kraemer B, Birk T, Person M, Bucher AIVI, Straif K, Schumann J, Chambless L. Cancer mortality among workers in the German rubber industry: 1981–91. Occup Environ Med 1996; 53: 289–298.
173. Welp E, Partanen T, Kogevinas M, Andersen A, Bellander T, Biocca M, Coggon D, Gennaro V, Kolstad HA, Lundberg I, Lynge E, Spence A, Ferro G, Saracci R, Boffetta P. Exposure to styrene and mortality from non-malignant respiratory diseases. Occup Environ Med 1996; 53: 499–501.
174. Werner JB, Carter JT. Mortality of United Kingdom acrylonitrile polymerisation workers. Br J Ind Med 1981; 38: 247–253.
175. Wolf P, Chiazze L, Ference LD. An historical cohort study of mortality among salaried research and development pensioners of the Allied Corporation. J Occup Med 1987; 29: 613–615.
176. Wong O. A cohort mortality study of employees exposed to chlorinated chemicals. Am J Ind Med 1988; 14: 417–431.
177. 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.
178. Wong O, Brocker W, Davis HV, Nagle GS. Mortality of workers potentially exposed to organic and inorganic brominated chemicals, DBCP, TRIS, PBB, and DDT. Br J Ind Med 1984; 41: 15–24.
179. 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.
180. Wong O. An industry wide mortality study of chemical workers occupationally exposed to benzene. II. Dose response analyses. Br J Ind Med 1987; 44: 382–395.
181. Wong O. An industry wide mortality study of chemical workers occupationally exposed to benzene. I. General results. Br J Ind Med 1987; 44: 65–81.
182. Barceloux DG. Chromium. J Toxicol Clin Toxicol 1999; 37: 173–194.
183. Cohen SM. Urinary bladder carcinogenesis. Toxicol Pathol 1988; 26: 121–127.
184. Hogstedt LC. Epidemiological studies on ethylene oxide and cancer: an updating. IARC Sci Publ 1988; 89: 265–270.
185. Rinsky RA, Smith AB, Hornung R, Filloon TG, Young RJ, Okun AH, Landrigan PJ. Benzene and leukemia: an epidemiologic risk assessment. N Engl J Med 1987; 316: 1044–1050.
186. McMichael AJ. Standardized mortality ratios and the “healthy worker effect”: scratching beneath the surface. J Occup Med 1976; 18: 165–168.
Keywords:

chemical industry; worker health; cancer incidence; mortality; occupational exposures; bladder neoplasms; lower respiratory tract neoplasms; lymphohematopoietic neoplasms

© 2001 Lippincott Williams & Wilkins, Inc.