Abstract
Rationale
Benzene has been classified as carcinogenic to humans, but there is limited evidence linking benzene exposure to lung cancer.
Objectives
We aimed to examine the relationship between occupational benzene exposure and lung cancer.
Methods
Subjects from 14 case-control studies across Europe and Canada were pooled. We used a quantitative job-exposure matrix to estimate benzene exposure. Logistic regression models assessed lung cancer risk across different exposure indices. We adjusted for smoking and five main occupational lung carcinogens and stratified analyses by smoking status and lung cancer subtypes.
Measurements and Main Results
Analyses included 28,048 subjects (12,329 cases, 15,719 control subjects). Lung cancer odds ratios ranged from 1.12 (95% confidence interval, 1.03–1.22) to 1.32 (95% confidence interval, 1.18–1.48) (Ptrend = 0.002) for groups with the lowest and highest cumulative occupational exposures, respectively, compared with unexposed subjects. We observed an increasing trend of lung cancer with longer duration of exposure (Ptrend < 0.001) and a decreasing trend with longer time since last exposure (Ptrend = 0.02). These effects were seen for all lung cancer subtypes, regardless of smoking status, and were not influenced by specific occupational groups, exposures, or studies.
Conclusions
We found consistent and robust associations between different dimensions of occupational benzene exposure and lung cancer after adjusting for smoking and main occupational lung carcinogens. These associations were observed across different subgroups, including nonsmokers. Our findings support the hypothesis that occupational benzene exposure increases the risk of developing lung cancer. Consequently, there is a need to revisit published epidemiological and molecular data on the pulmonary carcinogenicity of benzene.
Keywords: lung cancer, benzene, occupational exposure
At a Glance Commentary
Scientific Knowledge on the Subject
Benzene has been classified as carcinogenic to humans, but there is limited evidence linking benzene exposure to lung cancer.
What This Study Adds to the Field
In this large, pooled analysis of 14 case-control studies, we provide evidence to support the hypothesis based on robust and consistent associations between occupational benzene exposure and increased lung cancer risk.
Benzene is a volatile and ubiquitous air pollutant that is produced mainly from anthropogenic sources. It was a common solvent and ingredient in paint, printing inks, and glues, but the benzene content in these products has been either replaced or reduced since the 1980s after regulations and other mitigation measures (1, 2). Nevertheless, it remains a high-production-volume chemical and is still widely present in low- and middle-income countries (3). Occupational exposure to benzene occurs in various industries, including petroleum, chemical, painting, rubber, coke making, and manufacturing. Benzene has been classified as carcinogenic to humans (International Agency for Research on Cancer [IARC] group 1) on the basis of its causal link with acute myeloid leukemia (4).
Lung cancer is one of the leading causes of death worldwide (5). However, the association between benzene exposure and lung cancer has been less understood because previous studies have reported inconsistent results. For example, some studies indicated an increased risk of lung cancer among exposed subjects (6–11), whereas other studies showed no evidence of such association (12–19). In 2018, an IARC monographs working group on benzene concluded that the evidence of carcinogenicity for lung cancer was limited (4). The main concerns were the lack of sufficient adjustments for smoking and exposure to other occupational lung carcinogens (4). Shortly after the publication of the IARC monograph on benzene, a Canadian case-control study (which is also included in the SYNERGY project), including 733 cases and 894 control subjects, provided support for the association between ever exposure to benzene and lung cancer risk (odds ratio [OR], 1.35; 95% confidence interval [CI], 0.99–1.84) after adjusting for both smoking and several lung carcinogens (10). The association, however, was not present among nonsmokers and low-level smokers (OR, 0.94; 95% CI, 0.49–1.81), possibly because of the limited sample size in this study, leaving the possibility that the observed association was driven by residual confounding. Therefore, studies with larger sample sizes and stricter control for confounding factors (e.g., smoking and coexposures) are needed to further elucidate the possible association between occupational benzene exposure and lung cancer.
This study aimed to examine the association between occupational benzene exposure and lung cancer using a large-scale, pooled case-control study. We investigated lung cancer risk in relation to various benzene exposure metrics (ever/never, cumulative exposure, duration, and time since last exposure), stratified by smoking status and histologic types of lung cancer. Some of the results have been reported in the form of a conference abstract (20).
Methods
Study Population
Fourteen population- and hospital-based case-control studies on lung cancer were pooled from 13 European countries and Canada in the SYNERGY project (see Table E1 in the online supplement). A detailed description of the study population is presented elsewhere (21) and online (http://synergy.iarc.fr). Briefly, all studies have provided lifetime occupational and smoking histories (except for MORGEN [Monitoring Project on Risk Factors and Chronic Diseases in the Netherlands]). In most studies, cases and control subjects were frequency matched by sex and age. Most interviews (84%) were conducted face-to-face with the subjects. Lung cancer subtypes were classified on the basis of World Health Organization guidelines (22) after histological or cytological confirmation by the pathologists associated with the participating hospitals. Ethical approvals for the SYNERGY project were obtained according to the legislation in each participating country and the IARC institutional review board.
Exposure Assessment
We used a benzene-specific job-exposure matrix (BEN-JEM) to assess occupational exposure on the basis of participants’ lifetime occupational histories. One of the authors (R.V.) previously developed BEN-JEM (23) by combining expert assessments of exposure levels and probability factors and incorporating trends from various industries/job titles over time. BEN-JEM assesses occupational exposure to benzene encoded to the International Classification of Occupations 1988 (ISCO-88). For each ISCO-88 job code, BEN-JEM assigns the proportion of exposed workers (P, also as probability) and the mean level of exposure (L, in parts per million [ppm]) by eight periods (1945–1959, 1960–1984, 1985–1994, 1995–1997, 1998–2000, 2001–2003, 2004–2006, and 2007–2009), accounting for the downward trend of workplace exposure in Europe and North America (2). Exposure was calculated as a product of the probability and level (P × L). To improve exposure assessment quality, we included only subjects whose job records were within the BEN-JEM assessment period (1945–2009). Details of the exposure assessment are shown in the online supplement.
For other occupational exposures, we used SYN-JEM to assess the cumulative exposure levels to five main lung carcinogens (asbestos, hexavalent chromium [Cr(VI)], nickel, polycyclic aromatic hydrocarbons [PAHs], and respirable crystalline silica [RCS]) and diesel engine exhaust (DEE)-JEM to assess exposure to DEE. Detailed descriptions of SYN-JEM and DEE-JEM can be found in (24) and in (25), respectively, and the positive associations of the individual agents with lung cancer risk within the SYNERGY study were previously published (lung cancer ORs for the ever-exposed ranged from 1.09 to 1.27 [25–29]). In sensitivity analyses, we also accounted for chlorinated and other types of solvents that may co-occur with benzene using the semiquantitative ALOHA + JEM (30, 31).
Statistical Analyses
Unconditional logistic regression models were used to calculate ORs and 95% CIs of lung cancer associated with various indices of occupational benzene exposure. The exposure indices include ever/never exposed, cumulative exposure (>0–1, >1–5, and >5 ppm-yr), exposure duration (1–9, 10–19, 20–29, and >29 yr), and time since last exposure (<5, 5–9, 10–19, 20–29, 30–39, and >39 yr). Different histological types of lung cancer have different morphological features, molecular characterization, and etiology (32). To examine the possibly heterogeneous associations, we stratified the analyses by main lung cancer subtypes (adenocarcinoma, squamous cell carcinoma, small cell carcinoma, and large cell carcinoma). We also stratified the analyses by smoking status (never, former, and current smokers), in which smokers were defined as participants who had smoked more than one cigarette per day for >1 year and former smokers were defined as those who had stopped smoking at least 2 years before diagnosis/interview. P values for linear trend were obtained by treating the exposure metrics (e.g., cumulative exposure, duration) as continuous variables in the logistic regression models for both all subjects and the exposed only.
Assumptions made in the main analyses are shown in a directed acyclic graph (Figure E1). On the basis of the directed acyclic graph, we adjusted for age group (<45, 45–49, 50–54, 55–59, 60–69, 70–74, >74 yr), study, sex, pack-years (log [cigarette pack-years + 1]) and time since quitting smoking (current smokers; quitting 2–7, 8–15, 16–25, >25 yr before diagnosis/interview; never smokers), five known lung carcinogens (cumulative exposure to asbestos, PAHs, RCS, Cr[VI], and DEE), and ever-employed in a list A job. “List A” is a list of occupations and industries with known excess risk to lung cancer (33, 34), and it is used as an adjustment for other occupational lung carcinogens in this study. We did not adjust for nickel because of its high correlation with Cr(VI) (Pearson r = 0.81) (Figure E2).
We evaluated the robustness of associations with the following sensitivity analyses:
-
1.
One-by-one omitting groups of subjects who were ever employed in various specific industries/jobs (e.g., construction, mining, printers, painters, shoemakers) or exposed to known or suspected occupational lung carcinogens (e.g., asbestos, silica, Cr[VI], PAHs, chlorinated solvents) to explore if excluding any specific industry/job/exposure would alter the associations.
-
2.
Restricting the analyses to blue collar workers to limit possible residual confounding from socioeconomic factors.
-
3.
Applying different exposure lag times (0, 5, 10, 20 yr), in which the exposure for the specified years before diagnosis/interview was disregarded.
-
4.
Omitting the probability factor P in calculating cumulative benzene exposure levels to evaluate the associations on the basis of the concentration (L) estimate only.
-
5.
Assessing the association of benzene exposure with the risk of the four major lung cancer subtypes among nonsmokers only.
-
6.
Examining the associations without adjusting for the other known lung carcinogens (asbestos, Cr[VI], nickel, PAHs, silica, DEE) and/or list A jobs.
We applied spline analyses (thin-plate regression) to assess the shapes of exposure–response relationships for cumulative benzene exposure and exposure duration. The smoothing function “tp” was used with R package mgcv without specifying the smooth term k. The spline analyses were also performed for the effect of cumulative exposure on the four lung cancer subtypes.
We evaluated the interaction between smoking and benzene exposure by estimating the relative excess risks due to interaction (RERI) from models on the basis of stratified population according to their status of benzene exposure and smoking (ever/never) (35). The CI of the RERI was estimated on the basis of bootstrapping (36). Meta-analyses were performed using the R package meta (37). For the pooled studies, we stratified the analyses on the basis of sources of control subjects (hospital vs. population) and imposed both the fixed and random effects on the pooled results. All statistical analyses were conducted with R version 4.2.1 (38).
Results
From the SYNERGY population, we omitted 1,653 subjects (804 cases, 849 control subjects) because of incomplete covariate data and 335 subjects (145 cases, 191 control subjects) because of incomplete occupational records. We further excluded 5,055 control subjects and 4,427 cases because at least part of their job history was before 1945. The final study sample included 28,048 subjects (12,329 cases and 15,719 control subjects).
Table 1 shows the characteristics of the study subjects by lung cancer status. Cases were more frequently current and heavier smokers. The distribution between the exposed and unexposed is comparable for most demographic features and smoking indicators, except for ever employment in list A jobs, in which, for both cases and control subjects, more subjects can be found in the exposed than the unexposed group. Around 77% of study subjects were males.
Table 1.
Descriptive Characteristics of Study Participants, Stratified by Occupational Benzene Exposure Status
| Characteristics | Control Subjects (n = 15,719) |
Cases (n = 12,329) |
||
|---|---|---|---|---|
| Exposed | Unexposed | Exposed | Unexposed | |
| (n = 6,253) | (n = 9,466) | (n = 5,838) | (n = 6,491) | |
| Sex | ||||
| Female | 910 (14.6%) | 2,801 (29.6%) | 797 (13.7%) | 1,791 (27.6%) |
| Male | 5,343 (85.4%) | 6,665 (70.4%) | 5,041 (86.3%) | 4,700 (72.4%) |
| Age group, yr | ||||
| <45 | 484 (7.7%) | 870 (9.2%) | 330 (5.7%) | 376 (5.8%) |
| 45–49 | 547 (8.7%) | 760 (8.0%) | 529 (9.1%) | 534 (8.2%) |
| 50–54 | 909 (14.5%) | 1,209 (12.8%) | 897 (15.4%) | 901 (13.9%) |
| 55–59 | 1,340 (21.4%) | 1,724 (18.2%) | 1,380 (23.6%) | 1,246 (19.2%) |
| 60–64 | 1,259 (20.1%) | 1,812 (19.1%) | 1,205 (20.6%) | 1,315 (20.3%) |
| 65–69 | 1,052 (16.8%) | 1,697 (17.9%) | 913 (15.6%) | 1,111 (17.1%) |
| 70–74 | 590 (9.4%) | 1,160 (12.3%) | 504 (8.6%) | 810 (12.5%) |
| >75 | 72 (1.2%) | 234 (2.5%) | 80 (1.4%) | 198 (3.1%) |
| List A jobs | ||||
| Never | 5,478 (87.6%) | 9,324 (98.5%) | 4,770 (81.7%) | 6,310 (97.2%) |
| Ever | 775 (12.4%) | 142 (1.5%) | 1,068 (18.3%) | 181 (2.8%) |
| Smoking status | ||||
| Never smoker | 1,860 (29.7%) | 3,655 (38.6%) | 321 (5.5%) | 663 (10.2%) |
| Former smoker | 2,437 (39.0%) | 3,345 (35.3%) | 1,651 (28.3%) | 1,861 (28.7%) |
| Current smoker | 1,956 (31.3%) | 2,466 (26.1%) | 3,866 (66.2%) | 3,967 (61.1%) |
| Pack-years | ||||
| >0–10 | 1,052 (16.8%) | 1,604 (16.9%) | 275 (4.7%) | 377 (5.8%) |
| >10–19 | 795 (12.7%) | 1,072 (11.3%) | 516 (8.8%) | 517 (8.0%) |
| >19 | 2,546 (40.7%) | 3,135 (33.1%) | 4,726 (81.0%) | 4,934 (76.0%) |
| Time since quitting smoking, yr | ||||
| >0–7 | 428 (6.8%) | 661 (7.0%) | 661 (11.3%) | 757 (11.7%) |
| 8–15 | 599 (9.6%) | 758 (8.0%) | 457 (7.8%) | 521 (8.0%) |
| 16–25 | 695 (11.1%) | 996 (10.5%) | 356 (6.1%) | 374 (5.8%) |
| >25 | 715 (11.4%) | 930 (9.8%) | 177 (3.0%) | 209 (3.2%) |
| Histological type | ||||
| Squamous cell carcinoma | 0 (0%) | 0 (0%) | 2,254 (38.6%) | 2,163 (33.3%) |
| Small cell lung cancer | 0 (0%) | 0 (0%) | 975 (16.7%) | 1,034 (15.9%) |
| Adenocarcinoma | 0 (0%) | 0 (0%) | 1,604 (27.5%) | 2,117 (32.6%) |
| Large cell lung carcinoma | 0 (0%) | 0 (0%) | 294 (5.0%) | 317 (4.9%) |
| Unavailable | 0 (0%) | 0 (0%) | 29 (0.5%) | 30 (0.5%) |
| Others/unspecified | 0 (0%) | 0 (0%) | 682 (11.7%) | 830 (12.8%) |
On the basis of subjects’ lifetime occupation records, 47.4% of cases and 39.8% of control subjects were ever exposed to benzene (Table 2). The prevalence of ever benzene exposure ranged between 29.5% and 63.6% for cases and between 17.1% and 61.3% for control subjects among pooled studies (Table E1). The prevalence of the six known lung carcinogens by study is presented in Table E2. Benzene exposure level gradually declined from 1950 to almost none in 2009 among the included population (Figure E3). The job title “painters and related workers” had the highest average level of exposure (1.11 ppm for 544 ever-employed subjects), followed by “varnishes and related painters” and “shoemakers and related workers” (Table E3).
Table 2.
Lung Cancer Odds Ratios and 95% Confidence Intervals Based on Various Indices of Occupational Benzene Exposure
| Occupational Benzene Exposure | Cases, n (%) | Control Subjects, n (%) | OR* | 95% CI |
|---|---|---|---|---|
| Never | 6,491 (52.6) | 9,466 (60.2) | 1 | Referent |
| Ever exposure | 5,838 (47.4) | 6,253 (39.8) | 1.17 | 1.10–1.24 |
| Cumulative exposure, ppm-yr† | ||||
| >0–1 | 2,190 (17.8) | 2,549 (16.2) | 1.12 | 1.03–1.21 |
| >1–5 | 2,553 (20.7) | 2,698 (17.2) | 1.17 | 1.08–1.27 |
| >5 | 1,095 (8.9) | 1,006 (6.4) | 1.32 | 1.18–1.48 |
| Test for trend (exposed only), P Value | 0.05 | |||
| Test for trend (all subjects), P Value | 0.002 | |||
| Duration, yr | ||||
| 1–9 | 2,128 (17.3) | 2,414 (15.4) | 1.10 | 1.02–1.19 |
| 10–19 | 1,181 (9.6) | 1,401 (8.9) | 1.10 | 1.00–1.22 |
| 20–29 | 1,022 (8.3) | 1,017 (6.5) | 1.23 | 1.10–1.37 |
| >29 | 1,507 (12.2) | 1,421 (9.0) | 1.34 | 1.21–1.48 |
| Test for trend (exposed only), P Value | <0.001 | |||
| Test for trend (all subjects), P Value | <0.001 | |||
| Time since last exposure‡, yr | ||||
| <5 | 1,165 (9.4) | 1,083 (6.9) | 1.43 | 1.20–1.70 |
| 5–9 | 738 (6.0) | 783 (5.0) | 1.12 | 0.94–1.33 |
| 10–19 | 1,485 (12.0) | 1,571 (10.0) | 1.17 | 1.02–1.34 |
| 20–29 | 858 (7.0) | 983 (6.3) | 1.06 | 0.94–1.21 |
| 30–39 | 953 (7.7) | 1,081 (6.9) | 1.07 | 0.95–1.20 |
| >39 | 639 (5.2) | 752 (4.8) | 1.02 | 0.89–1.16 |
| Test for trend (exposed only), P Value | 0.020 | |||
Definition of abbreviations: CI = confidence interval; OR = odds ratio.
P values for trend test were obtained by taking the continuous variables (e.g., duration of exposure in years) in the logistic regression models (same for all subsequent analyses).
OR adjusted for study, age group, sex, smoking (pack-years, time since quitting smoking), list A jobs, cumulative exposures of asbestos, hexavalent chromium, polycyclic aromatic hydrocarbons, silica, and diesel engine exhaust.
Cumulative exposure was calculated by taking the product of intensity (L) and probability (P).
OR in time since last exposure is additionally adjusted for duration (continuous) of benzene exposure.
In Table 2, subjects ever exposed to benzene showed higher lung cancer risk than the unexposed (OR, 1.17; 95% CI, 1.10–1.24). Increased ORs were found for subjects with higher cumulative benzene exposure (e.g., >5 ppm-yr; OR, 1.32; 95% CI, 1.18–1.48), longer duration (e.g., >29 yr; OR, 1.34; 95% CI, 1.21–1.48), and more recent exposure, in which the highest lung cancer risk was observed (<5 yr since last exposure; OR, 1.43; 95% CI, 1.20–1.70). Analyses based on continuous exposure metrics showed evidence for linear increasing trends of cumulative benzene exposure (Ptrend = 0.002 for all subjects) and exposure duration (Ptrend < 0.001) and a declining trend in time since last exposure (Ptrend = 0.02). These trends in cumulative exposure categories were similar in men and women (Table E4).
Table 3 shows ORs associated with cumulative benzene exposure by major lung cancer subtypes. Increased risks for all four subtypes were associated with ever benzene exposure (ORs ranged from 1.13 to 1.26). The three cumulative exposure groups had mostly positive associations with all subtypes: ORs of the four subtypes ranged from 1.04–1.19 to 1.15–1.55, among the lowest (>0–1 ppm-yr) and highest (>5 ppm-yr) benzene exposure groups, respectively. For exposure duration, we found evidence for the increasing trends in ORs among all lung cancer subtypes (all Ptrend < 0.05). For time since last exposure, we observed the decreasing trend in the risk of squamous cell carcinoma (Ptrend = 0.01).
Table 3.
Odds Ratios of Different Lung Cancer Subtypes Associated with Various Occupational Benzene Exposure Indices
| Occupational Benzene Exposure | Adenocarcinoma |
Squamous Cell Carcinoma |
Small Cell Carcinoma |
Large Cell Carcinoma |
||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Cases, n (%) | OR | 95% CI | Cases, n (%) | OR | 95% CI | Cases, n (%) | OR | 95% CI | Cases, n (%) | OR | 95% CI | |
| Unexposed | 2,117 (56.9) | 1 | Referent | 2,163 (49.0) | 1 | Referent | 1,034 (51.5) | 1 | Referent | 317 (51.9) | 1 | Referent |
| Ever exposed | 1,604 (43.1) | 1.13 | 1.04–1.23 | 2,254 (51.0) | 1.20 | 1.10–1.30 | 975 (48.5) | 1.18 | 1.05–1.32 | 294 (48.1) | 1.26 | 1.04–1.51 |
| Cumulative exposure, ppm-yr | ||||||||||||
| >0–1 | 691 (18.6) | 1.10 | 0.98–1.22 | 751 (17.0) | 1.19 | 1.07–1.33 | 355 (17.7) | 1.11 | 0.96–1.28 | 107 (17.5) | 1.04 | 0.81–1.32 |
| >1–5 | 633 (17.0) | 1.11 | 0.99–1.25 | 1,055 (23.9) | 1.18 | 1.06–1.32 | 452 (22.5) | 1.27 | 1.09–1.46 | 137 (22.4) | 1.48 | 1.16–1.88 |
| >5 | 280 (7.5) | 1.31 | 1.10–1.55 | 448 (10.1) | 1.26 | 1.08–1.46 | 168 (8.4) | 1.15 | 0.93–1.42 | 50 (8.2) | 1.55 | 1.08–2.20 |
| Test for trend (exposed only), P Value | 0.08 | 0.36 | 0.97 | 0.08 | ||||||||
| Test for trend (all subjects), P Value | 0.03 | 0.06 | 0.35 | 0.08 | ||||||||
| Duration of benzene exposure, yr | ||||||||||||
| 1–9 | 643 (17.3) | 1.12 | 1.00–1.25 | 760 (17.2) | 1.11 | 0.99–1.24 | 354 (17.6) | 1.09 | 0.94–1.27 | 109 (17.8) | 1.13 | 0.88–1.43 |
| 10–19 | 319 (8.6) | 1.01 | 0.87–1.17 | 466 (10.6) | 1.18 | 1.03–1.36 | 204 (10.2) | 1.12 | 0.93–1.35 | 57 (9.3) | 1.04 | 0.76–1.41 |
| 20–29 | 271 (7.3) | 1.15 | 0.98–1.35 | 406 (9.2) | 1.31 | 1.13–1.52 | 173 (8.6) | 1.25 | 1.02–1.53 | 65 (10.6) | 1.74 | 1.27–2.36 |
| >29 | 371 (10.0) | 1.29 | 1.11–1.50 | 622 (14.1) | 1.31 | 1.15–1.50 | 244 (12.1) | 1.38 | 1.14–1.66 | 63 (10.3) | 1.52 | 1.10–2.09 |
| Test for trend (exposed only), P Value | 0.05 | 0.002 | 0.007 | 0.002 | ||||||||
| Test for trend (all subjects), P Value | 0.003 | <0.001 | <0.001 | 0.001 | ||||||||
| Time since last exposure*, yr | ||||||||||||
| <5 | 226 (6.1) | 1.21 | 0.93–1.57 | 504 (11.4) | 1.53 | 1.21–1.95 | 238 (11.8) | 1.45 | 1.06–1.98 | 70 (11.5) | 1.55 | 0.94–2.54 |
| 5–9 | 210 (5.6) | 1.09 | 0.85–1.40 | 271 (6.1) | 1.18 | 0.92–1.51 | 120 (6.0) | 1.16 | 0.84–1.59 | 23 (3.8) | 0.83 | 0.46–1.44 |
| 10–19 | 430 (11.6) | 1.00 | 0.82–1.22 | 560 (12.7) | 1.36 | 1.12–1.65 | 225 (11.2) | 1.17 | 0.91–1.51 | 85 (13.9) | 1.26 | 0.84–1.87 |
| 20–29 | 261 (7.0) | 1.08 | 0.90–1.29 | 302 (6.8) | 1.07 | 0.89–1.28 | 150 (7.5) | 1.06 | 0.84–1.34 | 48 (7.9) | 1.10 | 0.75–1.58 |
| 30–39 | 278 (7.5) | 1.12 | 0.95–1.32 | 373 (8.4) | 1.10 | 0.94–1.28 | 154 (7.7) | 1.06 | 0.86–1.32 | 38 (6.2) | 0.91 | 0.62–1.31 |
| >39 | 199 (5.3) | 1.02 | 0.85–1.23 | 244 (5.5) | 1.05 | 0.87–1.25 | 88 (4.4) | 1.00 | 0.77–1.29 | 30 (4.9) | 1.10 | 0.71–1.64 |
| Test for trend (exposed only), P Value | 0.80 | 0.01 | 0.18 | 0.57 | ||||||||
Definition of abbreviations: CI = confidence interval; OR = odds ratio.
OR adjusted for study, age group, sex, smoking (pack-years, time since quitting smoking), list A jobs, and cumulative exposures of asbestos, hexavalent chromium, polycyclic aromatic hydrocarbons, silica, and diesel engine exhaust.
OR in time since last exposure is additionally adjusted for duration (continuous) of benzene exposure.
Nonparametric spline analyses showed a monotonic increase in lung cancer risks with higher cumulative benzene exposure and longer exposure duration (Figure 1). We also observed similar exposure–response relationships for most lung cancer subtypes (except for large cell carcinoma, for which a nonlinear relationship was observed) (Figure E4).
Figure 1.
Odds ratios adjusted for study, age group, sex, smoking (pack-years, time since quitting smoking), list A jobs, and cumulative exposures of asbestos, hexavalent chromium, polycyclic aromatic hydrocarbons, silica, and diesel engine exhaust. Note the histograms above two plots only included exposed subjects to better reflect the distribution of exposure metrics, and the x-axis range of cumulative exposure (0–22 ppm-year) covered the exposure level of 99% included population.
We found associations between occupational benzene exposure and lung cancer risk, regardless of smoking status (Table 4). For nonsmokers, we observed increased ORs with both ever benzene exposure (OR, 1.18; 95% CI, 1.00–1.38) and cumulative benzene exposure (Ptrend = 0.005 for both exposed and all subjects). The OR for nonsmokers in the highest exposure group was 1.80 (95% CI, 1.26–2.53) versus an OR of 1.09 (95% CI, 0.89–1.32) in the lowest exposure group. Benzene exposure among nonsmokers that lasted for 20–29 years and occurred most recently (<5 yr) was associated with higher lung cancer risk (ORs, 1.46 [95% CI, 1.07–1.98] and 1.79 [95% CI, 1.10–2.88], respectively). For both former and current smokers, higher lung cancer ORs were found in subjects ever exposed to benzene and with longer exposure duration. Time since last exposure showed no clear trends among the three smoking strata (Ptrend > 0.1). Interaction analysis suggested the joint effect of ever benzene exposure and ever smoking was on an additive scale (RERI, 2.66; 95% CI, 1.89–3.23).
Table 4.
Lung Cancer Odds Ratios Associated with Different Degrees of Benzene Exposure, by Smoking Status
| Occupational Benzene Exposure | Never Smokers |
Former Smokers |
Current Smokers |
||||||
|---|---|---|---|---|---|---|---|---|---|
| Cases, n (%)/Control Subjects, n (%) | OR* | 95% CI | Cases, n (%)/Control Subjects, n (%) | OR† | 95% CI | Cases, n (%)/Control Subjects, n (%) | OR‡ | 95% CI | |
| Unexposed | 663 (67.4)/3,655 (66.3) | 1 | Referent | 1,861 (53.0)/3,345 (57.9) | 1 | Referent | 3,967 (50.6)/2,466 (55.8) | 1 | Referent |
| Ever exposed | 321 (32.6)/1,860 (33.7) | 1.18 | 1.00–1.38 | 1,651 (47.0)/2,437 (42.1) | 1.20 | 1.08–1.33 | 3,866 (49.4)/1,956 (44.2) | 1.14 | 1.04–1.24 |
| Cumulative benzene exposure, ppm-yr | |||||||||
| >0–1 | 159 (16.2)/862 (15.6) | 1.09 | 0.89–1.32 | 506 (14.4)/875 (15.1) | 1.09 | 0.95–1.25 | 1,525 (19.5)/812 (18.4) | 1.12 | 1.00–1.25 |
| >1–5 | 107 (10.9)/749 (13.6) | 1.17 | 0.92–1.50 | 775 (22.1)/1,132 (19.6) | 1.21 | 1.06–1.37 | 1,671 (21.3)/817 (18.5) | 1.15 | 1.03–1.29 |
| >5 | 55 (5.6)/249 (4.5) | 1.80 | 1.26–2.53 | 370 (10.5)/430 (7.4) | 1.46 | 1.22–1.75 | 670 (8.6)/327 (7.4) | 1.15 | 0.98–1.36 |
| Test for trend (exposed only), P Value | 0.005 | 0.94 | 0.07 | ||||||
| Test for trend (all subjects), P Value | 0.005 | 0.36 | 0.02 | ||||||
| Duration of benzene exposure, yr | |||||||||
| 1–9 | 121 (12.3)/742 (13.5) | 1.05 | 0.84–1.31 | 562 (16.0)/914 (15.8) | 1.14 | 0.99–1.30 | 1,445 (18.4)/758 (17.1) | 1.07 | 0.96–1.20 |
| 10–19 | 81 (8.2)/427 (7.7) | 1.20 | 0.91–1.56 | 328 (9.3)/533 (9.2) | 1.12 | 0.94–1.32 | 772 (9.9)/441 (10.0) | 1.04 | 0.91–1.20 |
| 20–29 | 60 (6.1)/299 (5.4) | 1.46 | 1.07–1.98 | 273 (7.8)/394 (6.8) | 1.15 | 0.95–1.38 | 689 (8.8)/324 (7.3) | 1.22 | 1.05–1.43 |
| >29 | 59 (6.0)/392 (7.1) | 1.25 | 0.90–1.71 | 488 (13.9)/596 (10.3) | 1.43 | 1.22–1.68 | 960 (12.3)/433 (9.8) | 1.32 | 1.15–1.53 |
| Test for trend (exposed only), P Value | 0.15 | 0.006 | 0.001 | ||||||
| Test for trend (all subjects), P Value | 0.02 | <0.001 | <0.001 | ||||||
| Time since last exposure§, yr | |||||||||
| <5 | 64 (6.5)/295 (5.3) | 1.79 | 1.10–2.88 | 228 (6.5)/375 (6.5) | 1.32 | 0.95–1.82 | 873 (11.1)/413 (9.3) | 1.38 | 1.09–1.75 |
| 5–9 | 40 (4.1)/221 (4.0) | 1.28 | 0.77–2.08 | 171 (4.9)/273 (4.7) | 1.11 | 0.80–1.53 | 527 (6.7)/289 (6.5) | 1.07 | 0.85–1.35 |
| 10–19 | 84 (8.5)/514 (9.3) | 1.15 | 0.79–1.65 | 495 (14.1)/608 (10.5) | 1.23 | 0.96–1.58 | 906 (11.6)/449 (10.2) | 1.08 | 0.90–1.31 |
| 20–29 | 42 (4.3)/288 (5.2) | 0.97 | 0.66–1.41 | 235 (6.7)/364 (6.3) | 1.17 | 0.94–1.47 | 581 (7.4)/331 (7.5) | 0.98 | 0.82–1.16 |
| 30–39 | 49 (5.0)/316 (5.7) | 0.99 | 0.70–1.38 | 264 (7.5)/459 (7.9) | 1.00 | 0.83–1.22 | 640 (8.2)/306 (6.9) | 1.11 | 0.95–1.31 |
| >39 | 42 (4.3)/226 (4.1) | 1.22 | 0.84–1.73 | 258 (7.3)/358 (6.2) | 1.07 | 0.88–1.29 | 339 (4.3)/168 (3.8) | 0.93 | 0.75–1.14 |
| Test for trend (exposed only), P Value | 0.71 | 0.11 | 0.21 | ||||||
Definition of abbreviations: CI = confidence interval; OR = odds ratio.
ORs are adjusted for study, age group, sex, list A jobs, and cumulative exposures of asbestos, hexavalent chromium, polycyclic aromatic hydrocarbons, silica, and diesel engine exhaust.
ORs are adjusted for study, age group, sex, smoking (pack-years, time since quitting smoking), list A jobs, and cumulative exposures of asbestos, hexavalent chromium, polycyclic aromatic hydrocarbons, silica, and diesel engine exhaust.
ORs are adjusted for study, age group, sex, smoking (pack-years), list A jobs and, and cumulative exposures of asbestos, hexavalent chromium, polycyclic aromatic hydrocarbons, silica, and diesel engine exhaust.
OR in time since last exposure is additionally adjusted for duration (continuous) of benzene exposure.
After omitting subjects in each of the benzene-related jobs and industries and the subjects ever exposed to known lung carcinogens and other solvents, increased ORs remained for the various benzene exposure metrics and lung cancer risk (Table 5). The meta-analysis, stratified by control types, showed consistent ORs (for ever benzene exposure) across most studies (heterogeneity statistic I2 = 19%). No obvious difference was observed in the pooled ORs between hospital- and population-based studies (Figure 2). We also showed stable associations after setting different lag years (Table E5), omitting the probability factor from the cumulative exposure calculation (Tables E6–E8) and excluding ever-smokers (Table E9). Similar exposure–response relationships were found for analyses in which the probability factor was omitted (Figure E5). Applying models without adjusting for the five main occupational lung carcinogens resulted in higher risk estimates for benzene and lung cancer risk (Table E10).
Table 5.
Sensitivity Analyses for Overall Lung Cancer Risk with Various Occupational Benzene Exposure Indices, by Various Subgroups
| Occupational Benzene Exposure | OR1 | 95% CI | OR2 | 95% CI | OR3 | 95% CI | OR4 | 95% CI |
|---|---|---|---|---|---|---|---|---|
| All subjects | 1.17 | 1.10–1.24 | 1.32 | 1.18–1.48 | 1.34 | 1.21–1.48 | 1.43 | 1.20–1.70 |
| Blue collar worker only | 1.08 | 1.00–1.16 | 1.24 | 1.09–1.40 | 1.26 | 1.13–1.41 | 1.32 | 1.09–1.59 |
| Omit specific industries/jobs* | ||||||||
| Construction | 1.14 | 1.06–1.22 | 1.33 | 1.15–1.53 | 1.28 | 1.13–1.44 | 1.44 | 1.17–1.78 |
| Mining | 1.16 | 1.09–1.23 | 1.32 | 1.18–1.49 | 1.35 | 1.22–1.50 | 1.41 | 1.18–1.68 |
| Metal-related industry | 1.17 | 1.10–1.24 | 1.31 | 1.16–1.48 | 1.37 | 1.24–1.52 | 1.40 | 1.17–1.67 |
| Transport | 1.20 | 1.12–1.28 | 1.37 | 1.21–1.55 | 1.39 | 1.25–1.54 | 1.44 | 1.20–1.74 |
| Farmer | 1.18 | 1.11–1.26 | 1.36 | 1.21–1.54 | 1.36 | 1.22–1.50 | 1.51 | 1.26–1.82 |
| Vehicle mechanic | 1.17 | 1.10–1.25 | 1.33 | 1.18–1.50 | 1.34 | 1.21–1.49 | 1.40 | 1.17–1.68 |
| Shoemakers and related workers | 1.16 | 1.09–1.23 | 1.31 | 1.16–1.47 | 1.33 | 1.21–1.48 | 1.42 | 1.19–1.69 |
| Printers and related workers† | 1.17 | 1.10–1.24 | 1.31 | 1.16–1.49 | 1.35 | 1.22–1.50 | 1.47 | 1.23–1.75 |
| Painter and related workers‡ | 1.17 | 1.10–1.24 | 1.34 | 1.18–1.51 | 1.32 | 1.19–1.46 | 1.46 | 1.22–1.75 |
| Omit subjects with specific exposures§ | ||||||||
| Asbestos | 1.16 | 1.07–1.26 | 1.29 | 1.09–1.52 | 1.32 | 1.14–1.53 | 1.43 | 1.10–1.86 |
| PAHs | 1.21 | 1.13–1.30 | 1.39 | 1.21–1.60 | 1.43 | 1.27–1.61 | 1.41 | 1.13–1.75 |
| Nickel | 1.17 | 1.10–1.25 | 1.29 | 1.13–1.48 | 1.39 | 1.23–1.56 | 1.39 | 1.14–1.70 |
| Cr(VI) | 1.17 | 1.09–1.25 | 1.29 | 1.13–1.48 | 1.38 | 1.23–1.56 | 1.35 | 1.10–1.67 |
| Silica | 1.16 | 1.09–1.25 | 1.39 | 1.23–1.58 | 1.36 | 1.22–1.52 | 1.41 | 1.15–1.73 |
| Diesel engine exhaust | 1.18 | 1.09–1.28 | 1.38 | 1.18–1.60 | 1.37 | 1.21–1.56 | 1.32 | 1.03–1.67 |
| Chlorinated solvents | 1.19 | 1.08–1.30 | 1.83 | 1.29–2.59 | 1.43 | 1.17–1.77 | 1.38 | 1.01–1.89 |
| Other types of solvents | 1.17 | 1.05–1.31 | 1.58 | 1.10–2.27 | 1.49 | 1.17–1.90 | 1.68 | 1.17–2.41 |
Definition of abbreviations: CI = confidence interval; Cr(VI) = hexavalent chromium; OR = odds ratio; PAH = polycyclic aromatic hydrocarbon.
OR1 represents the comparisons of ever exposed to benzene versus never exposed; OR2 represents the comparisons of exposed to high benzene exposure group (cumulative exposure > 5 ppm-yr) versus never exposed; OR3 represents the comparisons of exposed to longest duration group (duration >29 yr) versus never exposed; and OR4 represents the comparisons of exposed within 5 years before enrollment (time since last exposure <5 yr) versus never exposed. Note that all OR4 are additionally adjusted for benzene exposure duration (continuous).
For specific industries and jobs, all ORs (OR1–OR4) were adjusted for study, age group, sex, smoking (pack-years, time since quitting smoking), list A, cumulative exposures of asbestos, Cr(VI), PAHs, silica, and diesel engine exhaust.
Printers were classified on the basis of corresponding ISCO-68 codes 92XXX.
Painters were classified on the basis of corresponding ISCO-68 codes 93XXX.
For specific exposures, all ORs (OR1–OR4) were adjusted for study, age group, sex, smoking (pack-years, time since quitting smoking), and list A. Exposures were assessed with SYN-job-exposure matrix (JEM) and ALOHA (arial locations of hazards atmosphere) + JEM.
Figure 2.
Comparisons were made between ever exposed to occupational benzene versus never exposure. Odds ratios are adjusted for study, age group, sex, smoking (pack-years, time since quitting smoking), list A jobs, and cumulative exposures of asbestos, hexavalent chromium, polycyclic aromatic hydrocarbons, silica, and diesel engine exhaust. For LUCA and LUCAS, odds ratios were not adjusted for sex, because they only included male participants. Controls from TORONTO and INCO_Poland were from both population and hospitals. We consider the control subjects to be hospital based here for the meta-analyses. CI = confidence interval; IV = inverse-variance weighting; SE = standard error; TE = treatment effect (log-transformed).
Discussion
We investigated the association between occupational benzene exposure and lung cancer within a large, pooled international case-control study. Our main analyses suggested that increased lung cancer risk was associated with higher cumulative benzene exposure, longer exposure duration, and time since exposure cessation. Positive associations were present for the main lung cancer subtypes and among all smoking subgroups, including nonsmokers.
Previous studies have reported mixed results regarding the effect of benzene on lung cancer (6–11, 13–19). On the basis of available studies, it appears the evidence for the association is stronger in case-control studies than in cohort studies. Specifically, in case-control studies (6, 10, 11, 16, 18), the relative risk estimates for ever benzene exposure ranged from 1.10 to 1.84, whereas in cohort studies (7–9, 13–15, 17, 19), the estimates ranged from 0.22 to 1.50. However, restricting the evidence base on benzene and lung cancer to the cohort studies that reported a positive association for the known causal benzene–acute myeloid leukemia association (7–9), as previously shown to be informative in the systematic evaluation of the risk of benzene on lymphoma (4, 39, 40), showed that of the remaining three cohort studies, all reported a positive association for ever benzene exposure and lung cancer risk. The more informative cohort studies are more in line with the case-control evidence, together lending support for a possible association between benzene exposure and lung cancer.
Smoking is the main contributor to lung cancer risk, as shown also in previous analysis in SYNERGY (OR for lung cancer, 23.6 [95% CI, 20.4–27.2] for current smokers) (41). Confounding from smoking is thus the main concern in occupational epidemiologic analyses involving lung cancer and was also highlighted by the IARC monographs as one of the main factors for calling the evidence on benzene-related lung cancer limited (4). In our analyses, we not only controlled for pack-years and time since quitting smoking but also performed various stratified analyses by smoking status. We observed clear exposure–response relationships for benzene-related lung cancer risk among nonsmokers, suggesting that confounding by smoking is not the explanation for the observed associations in this study.
The second reservation expressed by the IARC working group was the potential confounding by other occupational exposures (4). We adjusted for five main lung carcinogens in the regression models, in which four were from SYN-JEM, a JEM that was developed specifically for the SYNERGY population. SYN-JEM estimates time-, job-, and region-specific exposure levels from statistical modeling based on large amounts of personal measurement data (24). This quantitative assessment of exposure to the major lung carcinogens, in addition to the adjustment of list A jobs in regression models, enabled us to further reduce residual confounding by other occupational exposures. By performing various sensitivity analyses to study the potential effect of confounding by other exposures, we showed robust associations for all exposure metrics (increased ORs for ever exposure, highest cumulative level, longest duration group, and most recent exposure) after excluding subjects ever employed in various occupations, indicating that the effect was not limited to a certain industry or occupation. In addition, we excluded subjects ever exposed to any of the six known lung carcinogens that had shown effects with lung cancer in previous analyses in the SYNERGY study (asbestos [26], silica [27], nickel [28], Cr[VI] [28], PAHs [29], and diesel engine exhaust [25]). These analyses clearly showed that the observed associations between benzene and lung cancer were not driven by any coexposure. In addition, we accounted for exposure to other solvents (e.g., chlorinated solvents) that have weaker evidence for an association with lung cancer (42). Also excluding these exposed subjects did not lead to an obvious change in results. Altogether these analyses suggest that confounding by known or suspected occupational lung carcinogens or jobs with known excess risk to lung cancer is unlikely to explain the observed results.
The observed benzene–lung cancer association is further strengthened by the coherent associations observed between increased lung cancer risk with longer exposure duration and the decrease in risk after exposure cessation (Table 2) and the strong mechanistic evidence showing that benzene exhibits 7 of the 10 key characteristics of carcinogens (43), including being metabolically activated to electrophilic metabolites, inducing oxidative stress, being genotoxic, being immunosuppressive, altering DNA repair, and causing genomic instability (4). In addition, benzene is absorbed in humans mainly by inhalation, making the lungs the first organ to be involved in benzene’s metabolism (44) and thus a potential target for its carcinogenic effects.
The observed risk patterns for all lung cancer combined were largely reflected by the positive associations found for four major lung cancer subtypes. The effect of ever exposure to benzene was stronger for large cell carcinoma, followed by squamous cell carcinoma and small cell carcinoma. Evidence to support an exposure–response relationship for each subtype was weaker than that for all lung cancer combined, possibly because of the limited statistical power in individual subtype analyses.
Limitations
There are several potential limitations of this study. Given that the individual benzene exposure was assigned by a JEM based on job titles, exposure may be misclassified because of the exposure variability within a job title (45). However, such misclassification is unlikely to be differential for cases and control subjects because job coding and exposure assignments were done blinded for case-control status. Moreover, misclassification of group-based exposure assessment (e.g., using a JEM) usually has a Berkson-like error structure, in which obtained risk estimates are (in most scenarios) unbiased but with less precision (46). Unlike many other JEMs, the high time-resolved temporal variation of benzene exposure was incorporated in BEN-JEM, and we included only subjects whose job history fell within the assessment timespan (1945–2009) to reduce uncertainties in exposure assessment.
We acknowledge that detailed harmonized data on socioeconomic status (SES) were not available in the SYNERGY project. The effect of unmeasured confounding from SES could be reflected by the attenuated effect estimates when limiting the analyses to blue collar workers only. Nevertheless, by restricting the analyses to blue collar workers, to minimize confounding by SES (but at the expense of losing some informative exposure contrast), we still observed an increased risk of lung cancer (Table 5), further supporting the consistent effect of benzene on lung cancer. From the sensitivity analyses using different covariate adjustments, we also observed lower risk estimates when list A job and/or other lung carcinogens were included in the models compared with the models without (Table E10). Although we rigorously adjusted for main lung carcinogens and performed analyses in which subjects working in certain industries or exposures to lung carcinogens were excluded, residual confounding from other unmeasured coexposures cannot be fully excluded.
We also acknowledge the limited statistical power to examine the exposure–response relationship in females. The analyses nevertheless demonstrated clear evidence of increased lung cancer risk among females with ever benzene exposure and low cumulative exposure (Table E4).
Conclusions
Benzene has been regulated extensively over the past several decades, and, as a result, occupational exposure to benzene has declined to <1 ppm among most occupational groups in North America and Europe (2). However, benzene exposure at unregulated workplaces and in low- and middle-income countries remains of great concern because higher exposure levels are still often observed (47, 48). In addition to its occupational occurrence, benzene is widely present in the general environment via the emission of motor vehicle exhaust, burning of coal and oil, and fuel evaporation (4, 49), leading to a far greater population being potentially exposed.
In conclusion, we found consistent and robust associations between different dimensions of occupational benzene exposure and lung cancer after adjusting for smoking and main occupational lung carcinogens. These associations are coherent over different strata of the study population, including nonsmokers. Our findings support the hypothesis of an effect of occupational benzene exposure on lung cancer risk and warrant revisiting the published epidemiological and molecular data addressing the pulmonary carcinogenicity of benzene.
Acknowledgments
Acknowledgment
The authors acknowledge the contributions of the late Dr. Bas Bueno-de-Mesquita and the contribution by Dr. Neil E. Caporaso and Prof. Vladimir Janout in data acquisition and data analysis design of the project. The authors also thank Dr. Johan Ohlander for his contribution to the data preparation, Max Oosterwegel for his suggestions on the analysis, and Prof. Loredana Radoï for her comments on the manuscript.
Footnotes
Supported by Deutsche Gesetzliche Unfallversicherung grant FP 271. Grant sponsors of the individual studies were the Canadian Institutes of Health Research and Guzzo-SRC Chair in Environment and Cancer, the Fondation de France, the German Federal Ministry of Education, Science, Research, and Technology (grants 01 HK 173/0 and 01 HK 546/8) and the Ministerstvo Práce a Sociálních Věcí České Republiky (grant IIIb7-27/13), EC’s INCO-COPERNICUS Program, Polish State Committee for Science Research, Roy Castle Foundation, the National Institutes of Health, National Cancer Institute, Division of Cancer Epidemiology and Genetics, Intramural Research Program, Lombardy Region, Istituto nazionale per l’assicurazione contro gli infortuni sul lavoro and the European Union Nuclear Fission Safety Program, Italian Association for Cancer Research, Region Piedmont, Compagnia di San Paolo, Europe Against Cancer Program, the Swedish Council for Work Life Research, and the Swedish EPA, the University of Oviedo, the Ministry of Health of the Czech Republic—MH CZ—DRO (MMCI, 00209805), CIBERESP, and FISS-PI060604.
Author Contributions: W.W., S.P., L.P., R.V., A.O., J. Schüz, and K.S. contributed significantly to data analysis, results interpretation, and original drafting of the work. T. Brüning, K.S., H.K., and R.V. contributed to the original conception of the project and secured project funding. W.A., M.S., P.B., T. Behrens, T. Brüning, B.K., D.C., P.A.D., E.F., G.F.-T., J.K.F., F.F., L.F., P. Guénel, P. Gustavsson, K.-H.J., S.K., M.T.L., J.L., C.B., D.M., E.M., J.R.M., F.M., L.R., T.P., H.P., J. Siemiatycki, B.Ś., H.-E.W., D.Z., C.G., K.S., H.K., and R.V. participated in data acquisition and data analysis design of the project. All authors participated in critical revision of the manuscript and provided approval of the finalized submitted version.
The views expressed in this paper are those of the authors and not necessarily those of the sponsor. Where authors are identified as personnel of the International Agency for Research on Cancer/World Health Organization, the authors alone are responsible for the views expressed in this article, and they do not necessarily represent the decisions, policy, or views of the International Agency for Research on Cancer/World Health Organization.
This article has an online supplement, which is accessible from this issue’s table of contents at www.atsjournals.org.
Originally Published in Press as DOI: 10.1164/rccm.202306-0942OC on October 9, 2023
Author disclosures are available with the text of this article at www.atsjournals.org.
References
- 1. Feitshans IL. Law and regulation of benzene. Environ Health Perspect . 1989;82:299–307. doi: 10.1289/ehp.8982299. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Capleton AC, Levy LS. An overview of occupational benzene exposures and occupational exposure limits in Europe and North America. Chem Biol Interact . 2005;153-154:43–53. doi: 10.1016/j.cbi.2005.03.007. [DOI] [PubMed] [Google Scholar]
- 3.Markit I.2017. https://ihsmarkit.com/products/benzene-chemical-economics-handbook.html
- 4.International Agency for Research on Cancer. 2018. https://publications.iarc.fr/Book-And-Report-Series/Iarc-Monographs-On-The-Identification-Of-Carcinogenic-Hazards-To-Humans/Benzene-2018
- 5.Wild CP, Weiderpass E, Stewart BW. World cancer report: cancer research for cancer prevention. Lyon, France: International Agency for Research on Cancer; 2020. [Google Scholar]
- 6. Villeneuve PJ, Jerrett M, Brenner D, Su J, Chen H, McLaughlin JR. A case-control study of long-term exposure to ambient volatile organic compounds and lung cancer in Toronto, Ontario, Canada. Am J Epidemiol . 2014;179:443–451. doi: 10.1093/aje/kwt289. [DOI] [PubMed] [Google Scholar]
- 7. Collins JJ, Ireland B, Buckley CF, Shepperly D. Lymphohaematopoeitic cancer mortality among workers with benzene exposure. Occup Environ Med . 2003;60:676–679. doi: 10.1136/oem.60.9.676. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Sorahan T, Kinlen LJ, Doll R. Cancer risks in a historical UK cohort of benzene exposed workers. Occup Environ Med . 2005;62:231–236. doi: 10.1136/oem.2004.015628. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Linet MS, Yin SN, Gilbert ES, Dores GM, Hayes RB, Vermeulen R, et al. Chinese Center for Disease Control and Prevention-U.S. National Cancer Institute Benzene Study Group A retrospective cohort study of cause-specific mortality and incidence of hematopoietic malignancies in Chinese benzene-exposed workers. Int J Cancer . 2015;137:2184–2197. doi: 10.1002/ijc.29591. [DOI] [PubMed] [Google Scholar]
- 10. Warden H, Richardson H, Richardson L, Siemiatycki J, Ho V. Associations between occupational exposure to benzene, toluene and xylene and risk of lung cancer in Montréal. Occup Environ Med . 2018;75:696–702. doi: 10.1136/oemed-2017-104987. [DOI] [PubMed] [Google Scholar]
- 11. Cheng I, Yang J, Tseng C, Wu J, Shariff-Marco S, Park SL, et al. Traffic-related air pollution and lung cancer incidence: the California Multiethnic Cohort Study. Am J Respir Crit Care Med . 2022;206:1008–1018. doi: 10.1164/rccm.202107-1770OC. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Koh DH, Chung EK, Jang JK, Lee HE, Ryu HW, Yoo KM, et al. Cancer incidence and mortality among temporary maintenance workers in a refinery/petrochemical complex in Korea. Int J Occup Environ Health . 2014;20:141–145. doi: 10.1179/2049396714Y.0000000059. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Koh DH, Kim TW, Yoon YH, Shin KS, Yoo SW. Lymphohematopoietic cancer mortality and morbidity of workers in a refinery/petrochemical complex in Korea. Saf Health Work . 2011;2:26–33. doi: 10.5491/SHAW.2011.2.1.26. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. 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. doi: 10.1136/oem.44.6.382. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Wong O, Harris F, Smith TJ. Health effects of gasoline exposure. II. Mortality patterns of distribution workers in the United States. Environ Health Perspect . 1993;101(Suppl 6):63–76. doi: 10.1289/ehp.93101s663. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Gérin M, Siemiatycki J, Désy M, Krewski D. Associations between several sites of cancer and occupational exposure to benzene, toluene, xylene, and styrene: results of a case-control study in Montreal. Am J Ind Med . 1998;34:144–156. doi: 10.1002/(sici)1097-0274(199808)34:2<144::aid-ajim7>3.0.co;2-x. [DOI] [PubMed] [Google Scholar]
- 17. Bove FJ, Ruckart PZ, Maslia M, Larson TC. Evaluation of mortality among marines and navy personnel exposed to contaminated drinking water at USMC base Camp Lejeune: a retrospective cohort study. Environ Health . 2014;13:10. doi: 10.1186/1476-069X-13-10. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Yuan JM, Butler LM, Gao YT, Murphy SE, Carmella SG, Wang R, et al. Urinary metabolites of a polycyclic aromatic hydrocarbon and volatile organic compounds in relation to lung cancer development in lifelong never smokers in the Shanghai Cohort Study. Carcinogenesis . 2014;35:339–345. doi: 10.1093/carcin/bgt352. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Collins JJ, Anteau SE, Swaen GM, Bodner KM, Bodnar CM. Lymphatic and hematopoietic cancers among benzene-exposed workers. J Occup Environ Med . 2015;57:159–163. doi: 10.1097/JOM.0000000000000324. [DOI] [PubMed] [Google Scholar]
- 20. Wan W, Peters S, Portengen L, Olsson A, Schüz J, Straif K, et al. O-59 Increased lung cancer risk and occupational benzene exposure: results from a pooled case-control study [abstract] Occup Environ Med . 2023;80:A82–A83. [Google Scholar]
- 21. Olsson AC, Gustavsson P, Kromhout H, Peters S, Vermeulen R, Brüske I, et al. Exposure to diesel motor exhaust and lung cancer risk in a pooled analysis from case-control studies in Europe and Canada. Am J Respir Crit Care Med . 2011;183:941–948. doi: 10.1164/rccm.201006-0940OC. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Travis WD, Colby TV, Corrin B, Shimosato Y, Brambilla E. Histological typing of lung and pleural tumors. 3rd ed. Berlin: Springer Verlag; 1999. [Google Scholar]
- 23. Spycher BD, Lupatsch JE, Huss A, Rischewski J, Schindera C, Spoerri A, et al. Swiss Paediatric Oncology Group; Swiss National Cohort Study Group Parental occupational exposure to benzene and the risk of childhood cancer: a census-based cohort study. Environ Int . 2017;108:84–91. doi: 10.1016/j.envint.2017.07.022. [DOI] [PubMed] [Google Scholar]
- 24. Peters S, Vermeulen R, Portengen L, Olsson A, Kendzia B, Vincent R, et al. SYN-JEM: a quantitative job-exposure matrix for five lung carcinogens. Ann Occup Hyg . 2016;60:795–811. doi: 10.1093/annhyg/mew034. [DOI] [PubMed] [Google Scholar]
- 25. Ge C, Peters S, Olsson A, Portengen L, Schüz J, Almansa J, et al. Diesel engine exhaust exposure, smoking, and lung cancer subtype risks. A pooled exposure-response analysis of 14 case-control studies. Am J Respir Crit Care Med . 2020;202:402–411. doi: 10.1164/rccm.201911-2101OC. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Olsson AC, Vermeulen R, Schüz J, Kromhout H, Pesch B, Peters S, et al. Exposure-response analyses of asbestos and lung cancer subtypes in a pooled analysis of case-control studies. Epidemiology . 2017;28:288–299. doi: 10.1097/EDE.0000000000000604. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Ge C, Peters S, Olsson A, Portengen L, Schüz J, Almansa J, et al. Respirable crystalline silica exposure, smoking, and lung cancer subtype risks. A pooled analysis of case-control studies. Am J Respir Crit Care Med . 2020;202:412–421. doi: 10.1164/rccm.201910-1926OC. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Behrens T, Ge C, Vermeulen R, Kendzia B, Olsson A, Schüz J, et al. Occupational exposure to nickel and hexavalent chromium and the risk of lung cancer in a pooled analysis of case-control studies (SYNERGY) Int J Cancer . 2023;152:645–660. doi: 10.1002/ijc.34272. [DOI] [PubMed] [Google Scholar]
- 29. Olsson A, Guha N, Bouaoun L, Kromhout H, Peters S, Siemiatycki J, et al. Occupational exposure to polycyclic aromatic hydrocarbons and lung cancer risk: results from a pooled analysis of case-control studies (SYNERGY) Cancer Epidemiol Biomarkers Prev . 2022;31:1433–1441. doi: 10.1158/1055-9965.EPI-21-1428. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Matheson MC, Benke G, Raven J, Sim MR, Kromhout H, Vermeulen R, et al. Biological dust exposure in the workplace is a risk factor for chronic obstructive pulmonary disease. Thorax . 2005;60:645–651. doi: 10.1136/thx.2004.035170. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Sunyer J, Kogevinas M, Kromhout H, Antó JM, Roca J, Tobias A, et al. Pulmonary ventilatory defects and occupational exposures in a population-based study in Spain. Spanish Group of the European Community Respiratory Health Survey. Am J Respir Crit Care Med . 1998;157:512–517. doi: 10.1164/ajrccm.157.2.9705029. [DOI] [PubMed] [Google Scholar]
- 32. Travis WD, Brambilla E, Nicholson AG, Yatabe Y, Austin JHM, Beasley MB, et al. WHO Panel The 2015 World Health Organization Classification of Lung Tumors: impact of genetic, clinical and radiologic advances since the 2004 classification. J Thorac Oncol . 2015;10:1243–1260. doi: 10.1097/JTO.0000000000000630. [DOI] [PubMed] [Google Scholar]
- 33. Ahrens W, Merletti F. A standard tool for the analysis of occupational lung cancer in epidemiologic studies. Int J Occup Environ Health . 1998;4:236–240. doi: 10.1179/oeh.1998.4.4.236. [DOI] [PubMed] [Google Scholar]
- 34. Mirabelli D, Chiusolo M, Calisti R, Massacesi S, Richiardi L, Nesti M, et al. Database of occupations and industrial activities that involve the risk of pulmonary tumors [in Italian] Epidemiol Prev . 2001;25:215–221. [PubMed] [Google Scholar]
- 35. Buckley JP, Doherty BT, Keil AP, Engel SM. Statistical approaches for estimating sex-specific effects in endocrine disruptors research. Environ Health Perspect . 2017;125:067013. doi: 10.1289/EHP334. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Assmann SF, Hosmer DW, Lemeshow S, Mundt KA. Confidence intervals for measures of interaction. Epidemiology . 1996;7:286–290. doi: 10.1097/00001648-199605000-00012. [DOI] [PubMed] [Google Scholar]
- 37. Balduzzi S, Rücker G, Schwarzer G. How to perform a meta-analysis with R: a practical tutorial. Evid Based Ment Health . 2019;22:153–160. doi: 10.1136/ebmental-2019-300117. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.R Core Team. A language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing; 2022. [Google Scholar]
- 39. Vlaanderen J, Lan Q, Kromhout H, Rothman N, Vermeulen R. Occupational benzene exposure and the risk of lymphoma subtypes: a meta-analysis of cohort studies incorporating three study quality dimensions. Environ Health Perspect . 2011;119:159–167. doi: 10.1289/ehp.1002318. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40. Rana I, Dahlberg S, Steinmaus C, Zhang L. Benzene exposure and non-Hodgkin lymphoma: a systematic review and meta-analysis of human studies. Lancet Planet Health . 2021;5:e633–e643. doi: 10.1016/S2542-5196(21)00149-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41. Pesch B, Kendzia B, Gustavsson P, Jöckel KH, Johnen G, Pohlabeln H, et al. Cigarette smoking and lung cancer—relative risk estimates for the major histological types from a pooled analysis of case-control studies. Int J Cancer . 2012;131:1210–1219. doi: 10.1002/ijc.27339. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42. Vizcaya D, Christensen KY, Lavoué J, Siemiatycki J. Risk of lung cancer associated with six types of chlorinated solvents: results from two case-control studies in Montreal, Canada. Occup Environ Med . 2013;70:81–85. doi: 10.1136/oemed-2012-101155. [DOI] [PubMed] [Google Scholar]
- 43. Samet JM, Chiu WA, Cogliano V, Jinot J, Kriebel D, Lunn RM, et al. The IARC monographs: updated procedures for modern and transparent evidence synthesis in cancer hazard identification. J Natl Cancer Inst . 2020;112:30–37. doi: 10.1093/jnci/djz169. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44. Powley MW, Carlson GP. Cytochromes P450 involved with benzene metabolism in hepatic and pulmonary microsomes. J Biochem Mol Toxicol . 2000;14:303–309. doi: 10.1002/1099-0461(2000)14:6<303::AID-JBT2>3.0.CO;2-8. [DOI] [PubMed] [Google Scholar]
- 45. Peters S. Although a valuable method in occupational epidemiology, job-exposure -matrices are no magic fix. Scand J Work Environ Health . 2020;46:231–234. doi: 10.5271/sjweh.3894. [DOI] [PubMed] [Google Scholar]
- 46. Seixas NS, Sheppard L. Maximizing accuracy and precision using individual and grouped exposure assessments. Scand J Work Environ Health . 1996;22:94–101. doi: 10.5271/sjweh.116. [DOI] [PubMed] [Google Scholar]
- 47. Wang L, Zhou Y, Liang Y, Wong O, Armstrong T, Schnatter AR, et al. Benzene exposure in the shoemaking industry in China, a literature survey, 1978–2004. Regul Toxicol Pharmacol . 2006;46:149–156. doi: 10.1016/j.yrtph.2006.06.009. [DOI] [PubMed] [Google Scholar]
- 48. Wong O. Regulation of occupational exposures in China. Regul Toxicol Pharmacol . 2003;38:109–111. doi: 10.1016/s0273-2300(03)00078-3. [DOI] [PubMed] [Google Scholar]
- 49. Weisel CP. Benzene exposure: an overview of monitoring methods and their findings. Chem Biol Interact . 2010;184:58–66. doi: 10.1016/j.cbi.2009.12.030. [DOI] [PMC free article] [PubMed] [Google Scholar]