Document Z8nb7xQG0QNXdQwLEqwz5vR9L

Review Occupational Exposure to Asbestos and Ovarian Cancer: A Meta-analysis M. Constanza Camargo,1 Leslie T. Stayner,1 Kurt Straif,2 Margarita Reina,1 Umaima Al-Alem,1 Paul A. Demers,3 and Philip J. Landrigan4 ''Division of Epidemiology and Biostatistics, University of Illinois, Chicago, Illinois, USA; international Agency for Research on Cancer, Lyon, France; Occupational Cancer Research Centre, Cancer Care Ontario, Toronto, Ontario, Canada; 4Department of Preventive Medicine, Mount Sinai School of Medicine, New York, New York, USA OBJECTIVE: A recent Monographs Working Group of the International Agency for Research on Cancer (IARC) concluded that there is sufficient evidence for a causal association between exposure to asbestos and ovarian cancer. We performed a meta-analysis to quantitatively evaluate this association. Data SOURCES: Searches of PubMed and unpublished data yielded a total of 18 cohort studies of women occupationally exposed to asbestos. Data extraction: Two authors independently abstracted data; any disagreement was resolved by consulting a third reviewer. Data synthesis: All but one study reported standardized mortality ratios (SMRs) comparing observed numbers of deaths with expected numbers for the general population; the exception was a study that reported standardized incidence ratios. For simplicity, we refer to all effect estimates as SMRs. The overall pooled SMR estimate for ovarian cancer was 1.77 (95% confidence interval, 1.37--2.28), with a moderate degree of heterogeneity among the studies (12 = 35.3%, p = 0.061). Effect estimates were stronger for cohorts compensated for asbestosis, cohorts with estimated lung cancer SMRs > 2.0, and studies conducted in Europe compared with other geographic regions. Effect estimates were similar for studies with and without pathologic confirmation, and we found no evidence of publication bias (Egger's test p-value = 0.162). CONCLUSIONS: Our study supports the IARC conclusion that exposure to asbestos is associated with increased risk of ovarian cancer. Key WORDS: asbestos, chrysotile, crocidolite, meta-analysis, ovarian cancer, SMR. Environ Health Perspect 119:1211-1217 (2011). http://dx.doi.org/10.1289/ehp.1003283 [Online 3 June 2011] In 2008, cancer of the ovary represented the second leading cause of gynecologic cancer death worldwide (Ferlay et al. 2010). The geographical distribution of ovarian cancer is characterized by wide international variation. Highest rates are observed in North America and Northern Europe. In the United States, white women have higher incidence and mor tality rates than do other racial and ethnic groups (Horner et al. 2009). Although the etiology of ovarian cancer is not well under stood, multiparity, lactation, oral contracep tive use, and tubal ligation or hysterectomy are inversely associated with risk (Permuth- Wey and Sellers 2009; Sueblinvong and Carney 2009), whereas estrogen-only meno pausal therapy, tobacco smoking, and other environmental, occupational, and genetic fac tors are positively associated with ovarian can cer (Antoniou et al. 2000; Grosse et al. 2009; Secretan et al. 2009; Shen et al. 1998). Approximately 125 million people around the world work in environments in which they are exposed to asbestos, and at least 90,000 people die from asbestos-related lung cancer, mesothelioma, or asbestosis every year (Burki 2009). Asbestos exposure has been identified in some previous reviews as a possi ble risk factor for ovarian cancer (Hankinson and Danforth 2006; Ness and Cottreau 1999; Shoham 1994). However, this association has not been widely recognized. Perineal use of talc, which may in some formulations contain asbestiform or talc mineral fibers, has also been associated with ovarian cancer in a num ber of studies (Baan et al. 2006; Langseth et al. 2008). The association between ovarian cancer risk and asbestos exposure was addressed by a Monographs Working Group that was con vened in March 2009 by the International Agency for Research on Cancer (IARC). After considering the potential role of chance, confounding, and other forms of bias, the working group concluded that the evidence is sufficient for a causal association between occupational exposure to asbestos and ovar ian cancer (Straif et al. 2009). To more fully evaluate and characterize this association, we performed a meta-analysis. Materials and Methods We searched for studies of workers exposed to asbestos published in any language before March 2010 using PubMed software to search Medline (U.S. National Library of Medicine, Bethesda, MD). Combinations of the following keywords were used: "ovarian cancer," "cancer of the ovary," "asbestos," "chrysotile," "crocidolite," "mortality," "stan dardized mortality ratio" (SMR), "incidence," "standardized incidence ratio" (SIR), "can cer," "mesothelioma," "cohort," "female," and "women." In addition, we searched major cohorts of asbestos-exposed workers for data on ovarian cancer. References cited in the selected articles were also considered. Two investigators in our team independently reviewed the articles and extracted the data; any disagreement was resolved by consulting a third reviewer. We incorporated into the meta-analysis all studies of women who were occupationally exposed to asbestos meeting the following two criteria: a) an estimate of relative risk (i.e., SMRs or SIRs) for ovarian cancer or data allowing such estimates to be derived were presented, and b) the study was of a population with clear and unequivocal evidence of occupational exposure to asbestos such as asbestos cement and textile workers; asbestos miners and millers; friction material, insulator, and insulation board manufactur ers; and workers compensated for asbestosis. Population- or hospital-based case-control studies that were based on jobs and industries with only limited documentation of asbes tos exposures were excluded (Langseth and Kjaerheim 2004; Rosenblatt et al. 1992; Shu et al. 1989). The following information was recorded for each study: first author, journal, geo graphic region of the cohort, year of publica tion, outcome (mortality or incidence), overall number of women, duration of follow-up, total person-years of observation, period of employment, industry sector, type of asbes tos, SMR or SIR and 95% confidence inter val (CI) for ovarian cancer (for simplicity, we refer to all effect estimates as SMRs), observed ovarian cancer cases or deaths, expected ovar ian cancer cases or deaths, whether pathologic confirmation of the tumors was conducted, potential confounding variables adjusted for, total number of deaths, total number of cancer cases, total number of peritoneal mesothelioma cases, SMRs for lung cancer, and whether workers received compensation for asbestosis. In addition, data on national Address correspondence to L.T. Stayner, Division of Epidemiology and Biostatistics, School of Public Health, University of Illinois at Chicago MC923, 1603 West Taylor St., Chicago, IL 60612-4392 USA. Telephone: (312) 355-3693. Fax: (312) 996 0064. E-mail: lstayner@uic.edu We are grateful to the following colleagues who provided unpublished information: M. Hein, D. Loomis, and B. Clin. We thank C. Mamo for pro viding us a copy of the report of the study conducted by his research group in Italy. We also thank J.M. Samet for his helpful comments on an earlier version of this manuscript. This review was performed without external funding. The authors declare they have no actual or poten tial competing financial interests. Received 1 December 2010; accepted 3 June 2011. Environmental Health Perspectives volume 119 I number 9 I September 2011 1211 Camargo et al Figure 1. Flow chart of the meta-analysis. incidence rates for ovarian cancer were obtained from GLOBOCAN 2008 estimates for individual countries (Ferlay et al. 2010). Statistical analysis. Based on the reported CIs, we estimated the standard errors (SEs) for the ln(SMR) or the ln(SIR) given by the for mula SE = [ln(upper limit) - ln(lower limit)] (2 x Zi-a/2), where for a 95% CI, Zi-a/2 equals 1.96 (Bradburn 2004). For the studies for which the 95% CI was not reported, we calculated them by the Fischer's exact method using the observed deaths and expected deaths reported in the articles (Dean et al. 2010). Overall pooled SMR estimates and their corresponding 95% CIs were obtained using fixed-effects (Mantel-Haenszel method) and random-effects (DerSimonian and Laird method) methods (Harris 2008). Given the significant amount of heterogeneity, only the random-effects estimates are presented. Meta-regression techniques were used to examine the extent to which one or more of the following variables might explain hetero geneity: outcome (mortality or incidence), cohort size (< 500, 500-1,000, or > 1,000 women), follow-up period (< 25 or > 25 years), geographic region (Europe vs. United States and Australia), national ovarian cancer incidence rate (< 12 and > 12 cases/100,000 women), type of industry (mining, textiles, cement, gas mask manufacturing, mixed, or other), type of asbestos (chrysotile, crocidolite, or mixed), compensation for asbestosis (yes or no), magnitude of the SMR for lung cancer (< 2.0 or > 2.0), and pathological confirma tion (yes or no). Between-study heterogeneity was assessed using the Q and I2 statistics, with Pq < 0.10 or I2 > 25% indicating significant heterogeneity (Higgins and Thompson 2002; Higgins et al. 2003). A log-likelihood test was used to measure the improvement in fit when explanatory variables were included com pared with the null model. The proportion of between-study variance explained by a specific Table 1. Study characteristics. Reference Acheson et al. 1982 Gardner et al. 1986 Newhouse and Sullivan 1989 Country United Kingdom United Kingdom United Kingdom National incidence rate for ovarian cancer3 12.8 12.8 Outcome studied Mortality Mortality Mortality Industry type Cohort Asbestos type size Gas mask assemblers (in Crocidolite Leyland and Preston) Gas mask assemblers Chrysotile (in Blackburn) Cement Chrysotile 757 570 657 Period of employment 1927-1939 1927-1945 1941-1954 12.8 Mortality Production of friction Chrysotile materials 4,345 1941-1979 Follow-up period 1951-1980 1951-1980 1941-1984 1941-1986 Ovarian cancer results Observed/ Lung expected Person- Total Total cancer deaths or SMR or SIR years deaths cancers SMR cases (95% CI) 18,781 219 66 2.41* 12/4.4 2.75 (1.42-4.81) 14,324 177 44 1.45* 5/3.4 1.48 (0.48-3.44) -- 102 26 1.42 3/2.7 1.11 (0.23-3.25) -- 522 148 0.66* 11/10.1 1.08 (0.61-1.79) Rosler et al. 1994 Germany 10.0 Mortality Mixed (mainly textile) Mixed (mainly 616 -- 1977-1988 6,236 64 32 3.39 2/1.8 1.09 chrysotile) (0.13-3.95) Tarchi et al. 1994 Italy 8.7 Mortality Mining Chrysotile 120 -- 1965-1989 -- 28 8 4.14c 2/0.42 4.76 (0.58-17.2) Germani et al. 1999 Italy 8.7 Mortality Textile (compensated Chrysotile 276 -- 1980-1997 3,761 123 40 6.82 4/0.76 5.26 for asbestosis) (1.43-13.47) Cement (compensated Mixed (mainly 278 -- 1980-1997 3,932 129 54 2.39 5/0.93 5.40 for asbestosis) crocidolite) (1.75-12.61) Berry et al. 2000 United 12.8 Mortality Textile and prefabricated Mixed 700 1936-1942 Up to June, 17,146 -- 129 7.46 9/3.56 2.53 Kingdom cement pipes 1980 (1.16-4.80) Szeszenia-Dabrowska Poland 12.6 Mortality Mixed (compensated Mixed 490 1970-1997 Up to Dec. 31, -- 121 34 6.21 1/1.27 0.79 et al. 2002 for asbestosis, mainly (diagnosis 1999 (0.02-4.39) asbestos processing period) plants) Mamo 2004 Italy 8.7 Mortality Textile Chrysotile 645 1951-1978 1981-1995 7,450 84 36 5.23 1/0.78 1.28 (0.02-7.12) Wilczynska et al. Poland 12.6 Mortality Mixed Mixed 1,201 1945-1980 Up to Dec. 31, -- 414 124 2.09 8/4.5 1.76 2005 1999 (0.76-3.47) McDonald et al. United 12.8 Mortality Gas mask assemblers Crocidolite 1,073 1940-1944 1963-2003 -- -- -- 2.73d 10/5.6 1.80 2006 Kingdom (0.9-3.3) Hein et al. 2007 United States 8.8 Mortality Textile Chrysotile 1,265 1940-1965 1979-2001 49,922 709 169 2.22c 6/9.68 0.62 (0.23-1.35) Pira et al. 2007 Italy 8.7 Mortality Textile Mixed 1,077 1946-1984 Up to Dec. 31, 36,886 254 130 6.5 8/2.8 2.83 2004 (1.22-5.57) Magnani et al. 2008 Italy 8.7 Mortality Cement Mixed 777 1912-1986 1965-2003 22,367 371 169 2.21 9/4 2.27 (1.04-4.32) Loomis et al. 2009 United States 8.8 Mortality Textile Chrysotile 1,795 1950-1973 Up to Dec. 31, 59,949 608 160 1.73c 9/7.34 1.23 2003 (0.56-2.33) Reid et al. 2009 Australia 7.7 Incidence Mining and milling Crocidolite 416 1943-1966 1960-2006 -- -- -- -- 1/1.54 0.65 (0.02-3.64) Harding et al. 2009 United 12.8 Mortality Mixed Mixed 4,495 -- 1971-2005 103,394 -- -- -- 17/15.2 1.12 Kingdom (0.66-1.80) Clin et al. 2009 France 7.7 Mortality Mixed Mixed 420 -- 1978-2004 -- -- 11 -- 3/1.88 1.60 (0.33-4.67) --, Not available. "Cases per 100,000 women, GLOBOCAN 2008 (Ferlay et al. 2010). *Lung and pleura are included. "Trachea, bronchus, and lung are included. rfFrom Wignall and Fox (1982). 1212 volume 119 I number 9 I September 2011 Environmental Health Perspectives Ovarian cancer and asbestos exposure variable was described using adjusted R2 esti mates that could be negative for variables that explain less of the heterogeneity than would be expected by chance (Sterne 2009). Publication bias was investigated by visual inspection of Begg's funnel plots and formally tested using Egger's regression asymmetry method. The influence of individual studies was assessed by sequentially dropping each one before pooling study-specific estimates. A visual impression of the amount of heterogeneity was explored using Galbraith plots to identify the studies contributing to heterogeneity (the z-statistic O/SEo was plotted against the reciprocal SE 1/SE$, where O was the effect estimated from the individual study and SEo was its SE). Tumor misclassification could lead to both false-positive and false-negative diag noses. Nielsen et al. (1994), based on a large series of peritoneal mesotheliomas, estimated that 16% of cases were misdiagnosed as ovar ian cancer cases. Assuming that a similar proportion of peritoneal mesotheliomas are misclassified as ovarian cancer cases, we evalu ated the effect of misclassification on the over all pooled SMR estimate by removing 20% of ovarian cancer cases from every study and repeating the meta-analysis. As a proxy for a dose-response analysis, we performed a meta-analysis that combined the results of women with the highest occupa tional asbestos exposure from studies reporting either duration or cumulative exposure. The meta-analyses were performed with Stata software (version 10; StataCorp LP, College Station, TX, USA) using a combina tion of available macros (Sterne 2009). Meta regression analyses were performed using the Proc Mixed in SAS (version 9.1; SAS Institute Inc., Cary, NC, USA). A Rvalue < 0.05 was considered statistically significant for all tests except for the heterogeneity. Results Literature search. We identified 15 references that met the criteria for inclusion in the meta analysis (Acheson et al. 1982; Berry et al. 2000; Gardner et al. 1986; Germani et al. 1999; Harding et al. 2009; Magnani et al. 2008; Mamo 2004; McDonald et al. 2006; Newhouse and Sullivan 1989; Pira et al. 2007; Reid et al. 2009; Rosler et al. 1994; Szeszenia-Dabrowska et al. 2002; Tarchi et al. 1994; Wilczynska et al. 2005) (Figure 1). In addition, we were able to obtain results from the investigators of three cohort studies that had not previously reported findings for ovar ian cancer (Clin et al. 2009; Hein et al. 2007; Loomis et al. 2009). Two of the identified 15 articles (Acheson et al. 1982; Germani et al. 1999) reported findings from two dis tinct cohorts, and thus we analyzed data on 20 distinct populations. The data were from the most recently published reference for each cohort, except when results for highexposure groups were not reported in the latest publication (Pira et al. 2005; Wignall and Fox 1982). We excluded studies conducted among workers who were predominantly exposed to other known or suspected carcinogens but also had some potential for exposure to asbes tos (Atkinson et al. 2004; Beall et al. 2005; Boice et al. 1999; Bulbulyan et al. 1999; Coggon et al. 1997; Costantini et al. 1994; Langseth and Andersen 1999; Lewis et al. 2003; Richardson et al. 2007; Settimi et al. 1999; Vasama-Neuvonen et al. 1999; Ward et al. 1994). Table 1 summarizes the main characteris tics of the selected studies. Only one study reported findings for ovarian cancer incidence (Reid et al. 2009); the remaining studies were based on mortality. We included a total of 125 ovarian cancer deaths and one inci dent cancer case in our main analysis. SMR estimates reported by the individual studies ranged from 0.62 to 5.40 (Figure 2). Most of the studies had been carried out in Europe (n = 15). Two studies were conducted in the United States, and one in Australia. Although some cohorts included only females, the majority included both males and females (n = 14). The industries involved included the manufacture of textiles, mining, cement production, manufacture of friction mate rial, and manufacture of gas masks. Some industries included manufacturing of a wide range of goods containing asbestos. Only Author Acheson et al. (Blackburn) Acheson et al. (Leyland/Preston) Gardner et al. Newhouse and Sullivan Rosler et al. Tarchi et al. Germani et al. (cement) Germani et al. (textile) Berry et al. Szeszenia-Dabrowska et al. Mamo et al. Wilczynska et al. McDonald et al. Hein et al. Pira et al. Magnani et al. Clin et al. Harding et al. Loomis et al. Reid et al. Overall (12 = 35.3%, p = 0.061) Publication year 1982 1982 1986 1989 1994 1994 1999 1999 2000 2002 2004 2005 2006 2007 2007 2008 2009 2009 2009 2009 SMR estimates (95% CI) 1.48 (0.48-3.44) 2.75 (1.42-4.81) 1.11 (0.23-3.25) 1.08 (0.61-1.79) 1.09 (0.13-3.95) 4.76 (0.58-17.20) 5.40 (1.75-12.61) 5.26 (1.43-13.47) 2.53 (1.16-4.80) 0.79 (0.02-4.39) 1.28 (0.02-7.12) 1.76 (0.76-3.47) 1.80 (0.90-3.30) 0.62 (0.23-1.35) 2.83 (1.22-5.57) 2.27 (1.04-4.32) 1.60 (0.33-4.67) 1.12 (0.66-1.80) 1.23 (0.56-2.33) 0.65 (0.02-3.64) 1.77 (1.37-2.28) Percent weight 4.66 8.26 2.96 9.25 1.92 1.95 4.64 3.85 7.06 0.83 0.71 6.53 7.76 5.39 6.53 7.04 2.96 9.80 7.02 0.89 100.00 Figure 2. SMR estimates and 95% CIs of ovarian cancer associated with occupational exposure to asbestos. Weights are from random-effects analysis. Studyspecific SMRs are shown as triangles, with the size of the boxes being inversely proportional to the study-specific SMR variance. Horizontal lines represent 95% CIs for the study-specific SMRs. The pooled SMR is shown as a diamond. The middle of the diamond corresponds to the SMR, and the width of the diamond rep resents the 95% CI. The vertical dashed red line provides a visual comparison of the pooled SMR with the corresponding study-specific SMRs. Environmental Health Perspectives volume 119 I number 9 I September 2011 1213 Camargo et al two studies conducted pathologic review of cases (Magnani et al. 2008; Reid et al. 2009). Peritoneal mesothelioma cases were reported in five of the six studies that reported a sig nificant excess mortality from ovarian can cer (Acheson et al. 1982; Berry et al. 2000; Germani et al. 1999; Magnani et al. 2008; Pira et al. 2007), and we did not include these when deriving SMRs for ovarian cancer. Pooled SMR estimate. Figure 2 shows the SMR estimates and 95% CIs from the indi vidual studies and the pooled SMR estimate based on a random-effects model. The average pooled estimate for ovarian cancer among asbestos exposed women was 1.77 (95% CI, 1.37-2.28), with a moderate degree of heterogeneity among the studies (I2 = 35.3%, p = 0.061). Table 2. Pooled random-effects model-based SMR estimates and 95% CIs of ovarian cancer associated with asbestos exposure by study characteristics. Covariable No covariables Type of outcome Incidence Mortality Type of asbestos Chrysotile Crocidolite Mixed Compensation for asbestosis Yes No Geographic region Europe United States and Australia Pathology confirmation Yes No Follow-up period in years < 25 > 25 Sample size < 500 500-1,000 > 1,000 SMR for lung cancer < 2.0 > 2.0 No data Type of industry Mining Textile Cement Gas mask manufacturing Mixed Others Ovarian cancer incidence ratec < 12 > 12 na Pooled SMR (95% CI) 20 1.77 (1.37-2.28) 1-- 19 1.79 (1.38-2.31) 8 1.40 (0.88-2.21) 3 2.18 (1.40-3.37) 9 2.00 (1.41-2.84) 3 4.67 (2.28-9.54) 17 1.60 (1.28-2.00) 17 1.95 (1.51-2.51) 3 0.92 (0.54-1.59) 2 2.08 (1.05-4.14) 18 1.76 (1.34-2.31) 6 1.83 (0.81-4.16) 14 1.73 (1.38-2.16) 6 3.37 (1.82-6.25) 7 2.16 (1.54-3.03) 7 1.35 (0.99-1.84) 4 1.18 (0.81-1.72) 13 2.25 (1.64-3.07) 3 1.15 (0.73-1.82) 2 2.27 (0.34-14.97) 5 1.73 (0.81-3.70) 3 2.56 (1.17-5.58) 3 2.10 (1.40-3.15) 6 1.50 (1.07-2.10) 1-- 11 2.02 (1.27-3.21) 9 1.59 (1.22-2.06) /2 (%) 35.3 Pq 0.06 ---- 37.6 0.05 39.2 0.12 0.0 0.42 29.9 0.18 0.0 0.41 17.6 0.25 28.2 0.13 0.0 0.48 0.0 0.36 39.7 0.04 67.2 0.01 7.9 0.37 9.0 0.36 0.0 0.80 34.0 0.17 0.0 0.96 30.5 0.14 0.0 0.81 36.7 0.21 65.0 0.02 47.8 0.15 0.0 0.48 0.0 0.56 ---- 46.2 0.05 14.5 0.31 Pllr -- 0.48 0.26 0.01 0.03 1.0 1.0 0.01 < 0.001 0.55 0.58 Adjusted R2 (%)b -1.8 17.8 52.0 26.2 -14.0 -15.2 100.0 89.7 -20.0 -5.9 Abbreviations: --, not applicable; Pq, p-value for the heterogeneity test; Pllr, p-value of the log-likelihood ratio test (when compared with the model containing no covariables). SMR includes data from one cohort that reported an SIR. "Number of cohorts included. ^Estimates can be negative if the covariable explains less of the heterogeneity than would be expected by chance (Sterne 2009). "National rate as cases per 100,000 women, GLOBOCAN 2008 (Ferlay et al. 2010). Exploring between-study heterogeneity. The evidence for heterogeneity was strong enough to warrant investigation of potential explana tory factors. Table 2 presents the findings from the meta-regression models for all covariates. Compensation for asbestosis, magnitude of the SMR for lung cancer, geographic region, and sample size were each statistically signifi cant predictors of the pooled ovarian cancer SMR based on the -2 log-likelihood test. The simultaneous inclusion of these predictors in a regression model virtually eliminated any degree of heterogeneity (I2 = 0%, R2 = 100%). Pooled SMR estimates were increased for cohorts that were compensated for asbesto sis (SMR = 4.67; 95% CI, 2.28-9.54) com pared with cohorts that were not compensated (SMR = 1.60; 95% CI, 1.28-2.00); cohorts that reported an SMR for lung cancer > 2.0 (SMR = 2.25; 95% CI, 1.64-3.07) compared with other cohorts (SMR = 1.18; 95% CI, 0.81-1.72); European cohorts (SMR = 1.95; 95% CI, 1.51-2.51; 15 studies representing 17 cohorts) compared with cohorts from the United States and Australia, for which there appeared to be no increase in ovarian cancer mortality (SMR = 0.92; 95% CI, 0.54-1.59; three studies); and for smaller cohorts com pared with larger cohorts. We found sugges tive albeit nonsignificant evidence for variation by type of asbestos. Pooled SMRs were larger for cohorts exposed predominantly to crocidolite (SMR = 2.18; 95% CI, 1.40-3.37) or mixed asbestos (SMR = 2.00; 95% CI, 1.41-2.84) than for cohorts exposed to chrysotile (SMR = 1.40; 95% CI, 0.88-2.21). Geographic region was no longer a signifi cant predictor of the pooled SMR when we excluded six European studies of gas mask assemblers (n = 3) and cohorts compensated for asbestosis (n = 3; data not shown). Influence ofindividualstudies. The pooled SMR estimates were relatively robust to the exclusion of any one study from the overall meta-analysis and did not change by > 10% (data not shown). Based on a visual impression ofthe amount of heterogeneity across all studies (data not shown), a study by Germani et al. (1999), con ducted in cement workers, and a study by Hein et al. (2007) were the greatest contributors to Table 3. SMR estimates of studies included in the analysis of highly exposed groups. Reference Pira et al. 2005, 2007 Berry et al. 2000 Wignall and Fox 1982 McDonald et al. 2006 Loomis et al. 2009 Hein et al. 2007 Country Italy United Kingdom United Kingdom United States United States Industry type Textile Textile and prefabricated cement pipe Gas mask assemblers Ovarian cancer results Observed/expected deaths SMR (95% CI) 3/0.53 5/0.9 5.74 (1.18-16.7) 5.56 (2.04-12.31) 3/0.95 3.16 (0.65-9.23) Textile Textile 6a/5.45 1/1.99 1.10 (0.37-2.21) 0.50 (0.01-2.80) Magnani et al. 2008 Italy Cement "Three ovarian cancer cases from the fourth plant were omitted. 2/0.7 2.97 (0.35-10.32) Variable (highest category) Duration of employment (> 10 years) Exposure and duration (severe exposure with > 2 years of duration) Duration of employment (> 1 year) Cumulative exposure (> 120 fiber-days/ml) Cumulative exposure and duration (> 30 years of employment and > 5,479 fiber-days/mL) Duration of exposure (> 30 years) 1214 volume 119 I number 9 I September 2011 Environmental Health Perspectives Ovarian cancer and asbestos exposure between-study variation. SMRs for both stud ies were outside the 95% CI of the regression line, in opposite directions. The pooled SMR estimates for ovarian cancer excluding the cement worker study by Germani et al. (1999), the Hein et al. (2007) study, or both were 1.67 (95% CI, 1.32-2.11), 1.86 (95% CI, 1.47-2.36), and 1.74 (95% CI, 1.41-2.16), respectively. Heterogeneity was reduced when we removed both of these studies (I2 = 11%, p = 0.323). Analysis of highly exposed groups. Estimates of cumulative or duration of exposure among asbestos-exposed workers were reported for only six studies (Berry et al. 2000; Hein et al. 2007; Loomis et al. 2009; Magnani et al. 2008; Pira et al. 2005; Wignall and Fox 1982; Table 3). The pooled SMR estimate of ovarian cancer based on these six high-exposure groups was 2.78 (95% CI, 1.36- 5.66; Figure 3). We found a moder ate degree of heterogeneity across all studies (I2 = 45%, p = 0.108). Influence of tumor misclassification. Results of the sensitivity analysis assuming that 20% of the cases were misclassified as ovarian cancers suggested some attenuation of the pooled effect estimate (SMR = 1.42; 95% CI, 1.11-1.82), with a moderate degree of heterogeneity among the studies (p = 0.104; I2 = 29.7%). Assessing publication bias. Figure 4 pres ents the Begg's funnel plot including all 20 populations. Egger's test produced a p-value of 0.162, which provides little evidence of publication bias. Discussion The association between asbestos and ovarian cancer has been assessed here among studies of workers in which a major portion of the cohort is presumed to have been exposed to asbestos. Our results demonstrate an increase in the pooled estimate (SMR = 1.77; 95% CI, 1.37- 2.28) for ovarian cancer in relation to exposure to asbestos. The magnitude of the pooled estimate is similar to that reported by Edelman (1992), who included six studies conducted in the United Kingdom published before 1989 (pooled SMR = 1.65; 95% CI, 1.27-2.16). They concluded, however, that despite the positive and significant association, there was insufficient information to infer that ovarian cancers were caused by occupational exposure to asbestos because of concerns about tumor misclassification, inappropriate compari son populations, and the failure to take into account for known risk factors. A more recent meta-analysis by Li et al. (2004) of three stud ies published before February 2003 of workers exposed only to chrysotile found a nonsig nificant association (pooled SMR = 1.81; 95% CI, 0.61-5.36; Pq < 0.05). These previous meta-analyses by Edelman (1992) and Li et al. (2004) did not evaluate sources of betweenstudy variability. Our analysis addressed heterogeneity and was based on studies included in the published meta-analyses, other available data, and more recent publications. Our search for sources of heterogene ity revealed that SMRs based on European cohorts suggested stronger effects of occupa tional asbestos exposure than did estimates based on cohorts from other geographic regions. This geographic variation seems to have been driven by data from studies of Italian and Polish women compensated for asbestosis and of United Kingdom women who manufactured gas masks, who most likely had been exposed to high levels of asbestos fibers. Indeed, the estimated effect of geographic region was no longer significant when we excluded both subsets of studies. Our analysis of heterogeneity also suggested that stratification according to sample size reduced heterogeneity. The observation that the smaller the cohort size, the larger the SMR was related to limited cohort size (< 500) in the three studies of women compensated for asbestosis. Sample size was no longer an important predictor once we dropped the studies of women with asbestosis and gas mask production. The results from the analysis of highly exposed workers indicate a stronger effect than among all workers combined. Similarly, occupational exposure was more strongly associated with ovarian cancer among cohorts with a lung cancer SMR > 2.0. We found a suggestive but nonsignifi cant association between asbestos type and the pooled ovarian cancer SMR. Cohorts predominantly exposed to crocidolite or mixed asbestos showed larger SMRs than did those exposed only to chrysotile asbestos. This finding is similar to what Stayner et al. (1996) found for mesothelioma. In addition, the nonsignificant SMR based on the eight cohorts with exposure to chrysotile asbestos Author Publication year SMR estimates (95% CI) Wignall and Fox Berry et al. 1982 2000 -- ---------- A----------i --1-----A--------- Pira et al. 2005 --------- 1------A------------ Hein et al. Magnani et al. 2007 2008 <---------------------------------------- A-- 1 ----------- ----------A------------- Loomis et al. 2009 Overall (I2 = 44.7%, p = 0.108) 4------ i 3.16 (0.65-9.23) 5.56 (2.04-12.31) 5.74 (1.18-16.70) 0.50 (0.01-2.80) 2.97 (0.35-10.32) 1.10 (0.37-2.21) 2.78 (1.36-5.66) Percent weight 16.71 24.34 16.74 5.51 12.26 24.44 100.00 0.5 1 2 3 5 10 Figure 3. SMR estimates and 95% CIs of ovarian cancer associated with high occupational exposure to asbestos. Weights were from random-effects analysis. Study-specific SMRs are shown as triangles, with the size of the boxes being inversely proportional to the study-specific SMR variance. Horizontal lines represent 95% CIs for the study-specific SMRs. The pooled SMR is shown as a diamond. The middle of the diamond corresponds to the SMR, and the width of the diamond represents the 95% CI. The vertical dashed red line provides a visual comparison of the pooled SMR with the corresponding study-specific SMRs. 0 0.5 1.0 1.5 SE ln(SMR) Figure 4. Begg's funnel plot with pseudo-95% CIs for ovarian cancer SMRs associated with occupational exposure [natural log (In)] of the pooled SMR of the 20 cohorts = 0.57]. Environmental Health Perspectives volume 119 I number 9 I September 2011 1215 Camargo et al only seems to confirm the results by Li et al. (2004) based on three studies. The observed overall heterogeneity among studies seemed to be explained by two cohorts (Germani et al. 1999; Hein et al. 2007). The study by Germani et al. (1999) of 278 Italian cement industry workers compensated for asbestosis (with mixed exposure, mainly crocidolite) reported a very large increase in mor tality for ovarian cancer (SMR = 5.4; 95% CI, 1.75-12.61), possibly because the study was limited to subjects with asbestosis who were likely highly exposed. The study by Hein et al. (2007) of 1,265 U.S. women exposed to chrysotile in a textile plant reported a non significant decrease in mortality for ovarian cancer (SMR = 0.62; 95% CI, 0.23-1.35). Pathophysiologic mechanisms by which asbestos may confer susceptibility to ovar ian cancer have been proposed. They relate mainly to the hypothesis that the persistent presence of asbestos fibers in ovarian tissue causes chronic inflammation. This hypoth esis is supported by reports of asbestos fibers in the ovaries of women occupationally and non-occupationally exposed to asbestos (Heller et al. 1996; Langseth et al. 2007). The mechanism of transportation of asbestos fibers to the ovary is not clearly understood. Retrograde movement of particles through the reproductive tract to the ovaries has been sug gested (Baan et al. 2006; Heller et al. 1996). Alternatively, blood-borne or lymph-borne fibers could penetrate to the ovary through the mesothelium. This mechanism is sup ported by the findings of in vivo studies in animal models demonstrating changes in the ovaries of guinea pigs and rabbits after perito neal injection of asbestos fibers (Graham and Graham 1967). In addition, perineal exposure to talc, which may in the past have contained asbestos or talc fibers, has also been associated in a number of studies with an increased risk of ovarian cancer (Baan et al. 2006; Langseth et al. 2008). A major concern in interpreting our find ings is that until recently it has been very dif ficult to distinguish pathologically between peritoneal mesothelioma and ovarian can cer (Kannerstein et al. 1977). In fact, mis diagnosis of cases of peritoneal mesotheliomas as ovarian cancer was previously identified in two studies that included pathologic review (Newhouse et al. 1972; Wignall and Fox 1982). Reid et al. (2009) examined the potential for misclassification by reviewing pathologic material on ovarian, colon, and peritoneal cancer and reported that none of the cancer specimens had been misclassified in their study. We did not observe a difference in pooled SMRs between studies with and with out pathologic confirmation, but the power of this test was limited because there were only two studies with pathologic confirmation (Magnani et al. 2008; Reid et al. 2009). We also did not observe a large attenuation of the association when we assumed that 20% of the ovarian cancer cases in each study were misclassified. Given our findings from this sen sitivity analysis, it would seem unlikely that the association between occupational asbestos exposure and ovarian cancer could be fully explained by tumor misdiagnosis. Our meta-analysis mainly represents stud ies conducted in developed areas, particularly among European populations. It is possible that studies conducted in other geographic regions (e.g., developing countries) may be available through other biomedical litera ture databases. However, the previous meta analysis by Li et al. (2004), which searched Chinese literature, found no articles on ovar ian cancer published in that language. A further limitation of our analysis was its inability to account for nonoccupational risk factors for ovarian cancer other than age. Differences in the definitions of duration or latency of asbestos exposure measures prevented a proper evaluation of a dose-response relation ship. Although imperfect, our meta-analysis restricted to highly exposed women is compat ible with an underlying dose-response effect. Finally, of even greater potential concern was the fact that some of the published studies failed to include findings for ovarian cancer or only reported results for cancers of the female genital organs. We identified 20 cohort stud ies of asbestos-exposed women that failed to report findings for ovarian cancer (Armstrong et al. 1988; Cheng and Kong 1992; Clin et al. 2009; Finkelstein 1989; Hein et al. 2007; Karjalainen et al. 1999; Knox et al. 1968; Loomis et al. 2009; Luberto et al. 2004; McDonald et al. 1980; Morinaga et al. 1991; Pang et al. 1997; Peto et al. 1977; Raffaelli et al. 2007; Raffn et al. 1989; Sichletidis et al. 2009; Smailyte et al. 2004; Sun et al. 2003; Woitowitz et al. 1986; Zhu and Wang 1993). Because of our familiarity with the authors, we were able to obtain unpublished results from three of these studies (Clin et al. 2009; Hein et al. 2007; Loomis et al. 2009). The remaining studies had, in general, a relatively small number of women or included short follow-up periods. Conclusion The findings from this analysis are consistent with the hypothesis that exposure to asbestos is associated with an increased risk of ovar ian cancer. Based on our sensitivity analy sis, it appears unlikely that our results can be fully explained by misclassification of ovarian cancer and peritoneal mesothelioma or other sources of bias and confounding. Our results therefore support the conclusion by IARC that exposure to asbestos is causally associated with an increased risk of ovarian cancer. References Acheson ED, Gardner MJ, Pippard EC, Grime LP. 1982. Mortality of two groups of women who manufactured gas masks from chrysotile and crocidolite asbestos: a 40-year follow-up. Br J Ind Med 39:344-348. Antoniou AC, Gayther SA, Stratton JF, Ponder BA, Easton DF. 2000. Risk models for familial ovarian and breast cancer. Genet Epidemiol 18:173-190. Armstrong BK, de Klerk NH, Musk AW, Hobbs MS. 1988. Mortality in miners and millers of crocidolite in Western Australia. Br J Ind Med 45:5-13. Atkinson WD, Law DV, Bromley KJ, Inskip HM. 2004. Mortality of employees of the United Kingdom Atomic Energy Authority, 1946-1997. Occup Environ Med 61:577-585. Baan R, Straif K, Grosse Y, Secretan B, El Ghissassi F, Cogliano V et al. 2006. Carcinogenicity of carbon black, titanium dioxide, and talc. Lancet Oncol 7:295-296. Beall C, Bender TJ, Cheng H, Herrick R, Kahn A, Matthews R, et al. 2005. Mortality among semiconductor and storage device manufacturing workers. J Occup Environ Med 47:996-1014. Berry G, Newhouse ML, Wagner JC. 2000. Mortality from all cancers of asbestos factory workers in East London 1933-1980. Occup Environ Med 57:782-785. Boice JD Jr, Marano DE, Fryzek JP, Sadler CJ, McLaughlin JK. 1999. Mortality among aircraft manufacturing workers. Occup Environ Med 56:581-597. Bradburn MJ. 2004. Updated and New Commands for Meta analysis in STATA. Cancer Research UK Medical Statistics Group. Oxford:Centre for Statistics in Medicine. Bulbulyan MA, Ilychova SA, Zahm SH, Astashevsky SV, Zaridze DG. 1999. Cancer mortality among women in the Russian printing industry. Am J Ind Med 36:166-171. Burki T. 2009. Asbestos production increases despite WHO opposition. Lancet Oncol 10:846. Cheng WN, Kong J. 1992. A retrospective mortality cohort study of chrysotile asbestos products workers in Tianjin 1972-1987. Environ Res 59:271-278. Clin B, Morlais F, Dubois B, Guizard AV, Desoubeaux N, Marquignon MF, et al. 2009. Occupational asbestos expo sure and digestive cancers--a cohort study. Aliment Pharmacol Ther 30:364-374. Coggon D, Wield G, Pannett B, Campbell L, Boffetta P. 1997. Mortality in employees of a Scottish paper mill. Am J Ind Med 32:535-539. Costantini AS, Pirastu R, Lagorio S, Miligi L, Costa G. 1994. Studying cancer among female workers: methods and preliminary results from a record-linkage system in Italy. J Occup Med 36:1180-1186. Dean AG, Sullivan KM, Soe MM. 2010. OpenEpi: Open Source Epidemiologic Statistics for Public Health. Version 2.3.1. Available: http://www.OpenEpi.com [accessed 20 October 2010]. Edelman DA. 1992. Does asbestos exposure increase the risk of urogenital cancer? Int Arch Occup Environ Health 63:469-475. Ferlay J, Shin HR, Bray F, Forman D, Mathers C, Parkin DM. 2010. GLOBOCAN 2008, v1.2, Cancer Incidence and Mortality Worldwide: IARC CancerBase No. 10 [Internet]. Available: http://globocan.iarc.fr [accessed 20 October 2010]. Finkelstein MM. 1989. Mortality rates among employees poten tially exposed to chrysotile asbestos at two automotive parts factories. CMAJ 141:125-130. Gardner MJ, Winter PD, Pannett B, Powell CA. 1986. Follow up study of workers manufacturing chrysotile asbestos cement products. Br J Ind Med 43:726-732. Germani D, Belli S, Bruno C, Grignoli M, Nesti M, Pirastu R, et al. 1999. Cohort mortality study of women compensated for asbestosis in Italy. Am J Ind Med 36:129-134. Graham J, Graham R. 1967. Ovarian cancer and asbestos. Environ Res 1:115-128. Grosse Y, Baan R, Straif K, Secretan B, El Ghissassi F, Bouvard V, et al. 2009. A review of human carcinogens-- part A: pharmaceuticals. Lancet Oncol 10:13-14. Hankinson SE, Danforth KN. 2006. Ovarian cancer. In: Cancer Epidemiology and Prevention (Schottenfeld D, Fraumeini JF Jr, eds). New York:Oxford University Press, 1013-1026. Harding AH, Darnton A, Wegerdt J, McElvenny D. 2009. Mortality among British asbestos workers undergoing regular medical examinations (1971-2005). Occup Environ Med 66:487-495. Harris R. 2008. metan: fixed- and random-effects meta-analysis. Stata J 8:3-28. Hein MJ, Stayner LT, Lehman E, Dement JM. 2007. Follow-up 1216 volume 119 I number 9 I September 2011 Environmental Health Perspectives Ovarian cancer and asbestos exposure study of chrysotile textile workers: cohort mortality and exposure-response. Occup Environ Med 64:616-625. Heller DS, Gordon RE, Westhoff C, Gerber S. 1996. Asbestos expo sure and ovarian fiber burden. Am J Ind Med 29:435-439. Higgins JP, Thompson SG. 2002. Quantifying heterogeneity in a meta-analysis. Stat Med 21:1539-1558. Higgins JP, Thompson SG, Deeks JJ, Altman DG. 2003. Measuring inconsistency in meta-analyses. BMJ 327:557-560. Horner MJ, Ries LAG, Krapcho M, Neyman N, Aminou R, Howlader N, et al. (eds). 2009. SEER Cancer Statistics Review, 1975-2006. Bethesda, MD:National Cancer Institute. Bethesda, MD, Available: http://seer.cancer.gov/ csr/1975_2006/ [accessed 30 September 2010]. Kannerstein M, Churg J, McCaughey WT, Hill DP. 1977. Papillary tumors of the peritoneum in women: mesothelioma or papil lary carcinoma. Am J Obstet Gynecol 127:306-314. Karjalainen A, Pukkala E, Kauppinen T, Partanen T. 1999. Incidence of cancer among Finnish patients with asbestos-related pul monary or pleural fibrosis. Cancer Causes Control 10:51-57. Knox JF, Holmes S, Doll R, Hill ID. 1968. Mortality from lung cancer and other causes among workers in an asbestos textile factory. Br J Ind Med 25:293-303. Langseth H, Andersen A. 1999. Cancer incidence among women in the Norwegian pulp and paper industry. Am J Ind Med 36:108-113. Langseth H, Hankinson SE, Siemiatycki J, Weiderpass E. 2008. Perineal use of talc and risk of ovarian cancer. J Epidemiol Community Health 62:358-360. Langseth H, Johansen BV, Nesland JM, Kjaerheim K. 2007. Asbestos fibers in ovarian tissue from Norwegian pulp and paper workers. Int J Gynecol Cancer 17:44-49. Langseth H, Kjaerheim K. 2004. Ovarian cancer and occu pational exposure among pulp and paper employees in Norway. Scand J Work Environ Health 30:356-361. Lewis RJ, Schnatter AR, Drummond I, Murray N, Thompson FS, Katz AM, et al. 2003. Mortality and cancer morbidity in a cohort of Canadian petroleum workers. Occup Environ Med 60:918-928. Li L, Sun TD, Zhang X, Lai RN, Li XY, Fan XJ, et al. 2004. Cohort studies on cancer mortality among workers exposed only to chrysotile asbestos: a meta-analysis. Biomed Environ Sci 17:459-468. Loomis D, Dement JM, Wolf SH, Richardson DB. 2009. Lung cancer mortality and fibre exposures among North Carolina asbestos textile workers. Occup Environ Med 66:535-542. Luberto F, Amendola P, Belli S, Bruno C, Candela S, Grignoli M, et al. 2004. Mortality study of asbestos cement workers in Emilia-Romagna [in Italian]. Epidemiol Prev 28:239-246. M a g na n i C, Ferra n te D , Baron e- Ad esi F, B ertol otti M , Todesco A, Mirabelli D, et al. 2008. Cancer risk after ces sation of asbestos exposure: a cohort study of Italian asbestos cement workers. Occup Environ Med 65:164-170. Mamo C. 2004. Mortality experience in an historical cohort of chrysotile asbestos textile workers. In: Proceedings from the Global Asbestos Congress, 19-21 November 2004, Waseda University, Tokyo, Japan. McDonald JC, Harris JM, Berry G. 2006. Sixty years on: the price of assembling military gas masks in 1940. Occup Environ Med 63:852-855. McDonald JC, Liddell FD, Gibbs GW, Eyssen GE, McDonald AD. 1980. Dust exposure and mortality in chrysotile mining, 1910-1975. Br J Ind Med 37:11-24. Morinaga K, Hanai A, Fujimoto I, Ohtsuka J, Matsumura T, Sakato J, et al. 1991. A retrospective cohort study of workers in small asbestos industries in south Osaka [in Japanese]. Nippon Koshu Eisei Zasshi 38:267-271. Ness RB, Cottreau C. 1999. Possible role of ovarian epithe lial inflammation in ovarian cancer. J Natl Cancer Inst 91:1459-1467. Newhouse ML, Berry G, Wagner JC, Turok ME. 1972. A study of the mortality of female asbestos workers. Br J Ind Med 29:134-141. Newhouse ML, Sullivan KR. 1989. A mortality study of workers manufacturing friction materials: 1941-1986. Br J Ind Med 46:176-179. Nielsen AM, Olsen JH, Madsen PM, Francis D, Almind M. 1994. Peritoneal mesotheliomas in Danish women: review of histopathologic slides and history of abdominal surgery. Acta Obstet Gynecol Scand 73:581-585. Pang ZC, Zhang Z, Wang Y, Zhang H. 1997. Mortality from a Chinese asbestos plant: overall cancer mortality. Am J Ind Med 32:442-444. Permuth-Wey J, Sellers TA. 2009. Epidemiology of ovarian can cer. Methods Mol Biol 472:413-437. Peto J, Doll R, Howard SV, Kinlen LJ, Lewinsohn HC. 1977. A mortality study among workers in an English asbestos fac tory. Br J Ind Med 34:169-173. Pira E, Pelucchi C, Buffoni L, Palmas A, Turbiglio M, Negri E, et al. 2005. Cancer mortality in a cohort of asbestos textile workers. Br J Cancer 92:580-586. Pira E, Pelucchi C, Piolatto PG, Negri E, Discalzi G, La Vecchia C. 2007. First and subsequent asbestos exposures in relation to mesothelioma and lung cancer mortality. Br J Cancer 97:1300-1304. Raffaelli I, Festa G, Costantini AS, Leva G, Gorini G. 2007. Mortality in a cohort of asbestos cement workers in Carrara, Italy [in Italian]. Med Lav 98:156-163. Raffn E, Lynge E, Juel K, Korsgaard B. 1989. Incidence of cancer and mortality among employees in the asbestos cement industry in Denmark. Br J Ind Med 46:90-96. Reid A, Segal A, Heyworth JS, de Klerk NH, Musk AW. 2009. Gynecologic and breast cancers in women after expo sure to blue asbestos at Wittenoom. Cancer Epidemiol Biomarkers Prev 18:140-147. Richardson DB, Wing S, Wolf S. 2007. Mortality among workers at the Savannah River Site. Am J Ind Med 50:881-891. Rosenblatt KA, Szklo M, Rosenshein NB. 1992. Mineral fiber exposure and the development of ovarian cancer. Gynecol Oncol 45:20-25. Rosier JA, Woitowitz HJ, Lange HJ, Woitowitz RH, Ulm K, Rodelsperger K. 1994. Mortality rates in a female cohort following asbestos exposure in Germany. J Occup Med 36:889-893. Secretan B, Straif K, Baan R, Grosse Y, El Ghissassi F, Bouvard V, et al. 2009. A review of human carcinogens--part E: tobacco, areca nut, alcohol, coal smoke, and salted fish. Lancet Oncol 10:1033-1034. Settimi L, Costellati L, Naldi M, Bersani G, Olanda S, Maiozzi P. 1999. Mortality among workers in an Italian cigarette fac tory. Occup Med (Lond) 49:361-364. Shen N, Weiderpass E, Antilla A, Goldberg MS, VasamaNeuvonen KM, Boffetta P, et al. 1998. Epidemiology of occupational and environmental risk factors related to ovarian cancer. Scand J Work Environ Health 24:175-182. Shoham Z. 1994. Epidemiology, etiology, and fertility drugs in ovarian epithelial carcinoma: where are we today? Fertil Steril 62:433-448. Shu XO, Brinton LA, Gao YT, Yuan JM. 1989. Population-based case-control study of ovarian cancer in Shanghai. Cancer Res 49:3670-3674. Sichletidis L, Chloros D, Spyratos D, Haidich AB, Fourkiotou I, Kakoura M, et al. 2009. Mortality from occupational exposure to relatively pure chrysotile: a 39-year study. Respiration 78:63-68. Smailyte G, Kurtinaitis J, Andersen A. 2004. Mortality and can cer incidence among Lithuanian cement producing work ers. Occup Environ Med 61:529-534. Stayner LT, Dankovic DA, Lemen RA. 1996. Occupational expo sure to chrysotile asbestos and cancer risk: a review of the amphibole hypothesis. Am J Public Health 86:179-186. Sterne JAC. 2009. Meta-Analysis in Stata: An Updated Collection from the Stata Journal. College Station, TX:Stata Press. Straif K, Benbrahim-Tallaa L, Baan R, Grosse Y, Secretan B, El Ghissassi F, et al. 2009. A review of human carcinogens-- part C: metals, arsenic, dusts, and fibres. Lancet Oncology 10:453-454. Sueblinvong T, Carney ME. 2009. Current understanding of risk factors for ovarian cancer. Curr Treat Options Oncol 10:67-81. Sun T, Li L, Shi N, Zhang X. 2003. A 40-year cohort study on cancer mortality among female workers with manual spin ning of chrysotile asbestos [in Chinese]. Wei Sheng Yan Jiu 32:511-513. Szeszenia-Dabrowska N, Urszula W, Szymczak W, Strzelecka A. 2002. Mortality study of workers compensated for asbestosis in Poland, 1970-1997. Int J Occup Med Environ Health 15:267-278. Tarchi M, Orsi D, Comba P, De Santis M, Pirastu R, Battista G, et al. 1994. Cohort mortality study of rock salt workers in Italy. Am J Ind Med 25:251-256. Vasama-Neuvonen K, Pukkala E, Paakkulainen H, Mutanen P, Weiderpass E, Boffetta P, et al. 1999. Ovarian cancer and occupational exposures in Finland. Am J Ind Med 36:83-89. War d E M , Rud e r AM, Sur u d a A, Sm ith AB, H a lper in W, Fessler CA, et al. 1994. Cancer mortality patterns among female and male workers employed in a cable manufac turing plant during World War II. J Occup Med 36:860-866. Wignall BK, Fox AJ. 1982. Mortality of female gas mask assem blers. Br J Ind Med 39:34-38. Wilczyn'ska U, Szymczak W, Szeszenia-Dabrowska N. 2005. Mortality from malignant neoplasms among workers of an asbestos processing plant in Poland: results of prolonged observation. Int J Occup Med Environ Health 18:313-326. Woitowitz HJ, Lange HJ, Beierl L, Rathgeb M, Schmidt K, Ulm K, et al. 1986. Mortality rates in the Federal Republic of Germany following previous occupational exposure to asbestos dust. Int Arch Occup Environ Health 57:161-171. Zhu H, Wang Z. 1993. Study of occupational lung cancer in asbestos factories in China. Br J Ind Med 50:1039-1042. Environmental Health Perspectives volume 119 I number 9 I September 2011 1217