TSpace Research Repository tspace.library.utoronto.ca Pesticide exposures and the risk of multiple myeloma in men: An analysis of the North American Pooled Project Presutti R., Harris S.A., Kachuri L., Spinelli J.J., Pahwa M., Blair A., Zahm S.H., Cantor K.P., Weisenburger D.D., Pahwa P., McLaughlin J.R., Dosman J.A., Beane Freeman L. Version Post-Print/Accepted Manuscript Citation (published version) Presutti R, Harris SA, Kachuri L, Spinelli JJ, Pahwa M, Blair A, Zahm SH, Cantor KP, Weisenburger DD, Pahwa P, McLaughlin JR, Dosman JA, Beane Freeman L. Pesticide exposures and the risk of multiple myeloma in men: An analysis of the North American Pooled Project (NAPP). International Journal of Cancer 2016 Jun. http://www.ncbi.nlm.nih.gov/pubmed/27261772. Publisher s Statement This is the peer reviewed version of the following article: Presutti R, Harris SA, Kachuri L, Spinelli JJ, Pahwa M, Blair A, Zahm SH, Cantor KP, Weisenburger DD, Pahwa P, McLaughlin JR, Dosman JA, Beane Freeman L. Pesticide exposures and the risk of multiple myeloma in men: An analysis of the North American Pooled Project (NAPP). International Journal of Cancer 2016 Jun, which has been published in final form at https://dx.doi.org/10.1002/ijc.30218. This article may be used for non-commercial purposes in accordance with Wiley Terms and Conditions for Self-Archiving. How to cite TSpace items Always cite the published version, so the author(s) will receive recognition through services that track citation counts, e.g. Scopus. If you need to cite the page number of the TSpace version (original manuscript or accepted manuscript) because you cannot access the published version, then cite the TSpace version in addition to the published version using the permanent URI (handle) found on the record page.
Pesticide exposures and the risk of multiple myeloma in men: an analysis of the North American Pooled Project (NAPP) Roseanna Presutti 1,2, Shelley A. Harris* 1,2,3, Linda Kachuri 1,2,3,4, John J. Spinelli 5, Manisha Pahwa 1,2, Aaron Blair 6, Shelia Hoar Zahm 6, Kenneth P. Cantor 6, Dennis D. Weisenburger 7, Punam Pahwa 8, John R. McLaughlin 4,9, James A. Dosman 8, Laura Beane Freeman 6 1 Occupational Cancer Research Centre, Cancer Care Ontario, Toronto, ON, Canada 2 Prevention and Cancer Control, Cancer Care Ontario, Toronto, ON, Canada 3 Dalla Lana School of Public Health, University of Toronto, Toronto, ON, Canada 4 Lunenfeld-Tanenbaum Research Institute, Toronto, ON, Canada 5 British Columbia Cancer Agency Research Centre, University of British Columbia, Vancouver, BC, Canada 6 U.S. National Cancer Institute, Bethesda, MD, USA 7 Department of Pathology, City of Hope Medical Center, Duarte, CA, USA 8 University of Saskatchewan, Saskatoon, SK, Canada 9 Public Health Ontario, Toronto, ON, Canada *Corresponding Author: Shelley A. Harris, PhD Prevention and Cancer Control, Cancer Care Ontario 620 University Avenue, Toronto, ON, M5G 2L7 Fax: 416-971-6888 E-mail: Shelley.Harris@cancercare.on.ca Tel: 416-971-9800 x3234 Keywords: multiple myeloma; pesticides; pooled case-control study; carbamates; insecticides Article category: Epidemiology Abbreviations used: 2,4-D: 2,4-dichlorophenoxyacetic acid CCSPH: Cross Canada Study of Pesticides and Health CI: Confidence Interval DDT: Dichlorodiphenyltrichloroethane HL: Hodgkin Lymphoma IARC: International Agency for Research on Cancer LD: Lifetime-days MGUS: Monoclonal gammopathy of undetermined significance MM: Multiple Myeloma NAPP: North American Pooled Project NHL: Non-Hodgkin Lymphoma OR: Odds Ratio
Novelty and Impact: Case-control studies from the United States and Canada were pooled to form the North American Pooled Project, the largest study of its kind focused on agricultural and pesticide exposures and haematological cancers, including multiple myeloma. We observed significant increases in multiple myeloma risk for self-reported use of carbaryl, captan and DDT. This study provides further insight into the relationship between pesticide use and the risk of MM, while offering more refined risk estimates using different metrics of exposure. 1
Abstract: Multiple myeloma (MM) has been consistently linked with agricultural activities, including farming and pesticide exposures. Three case-control studies in the United States and Canada were pooled to create the North American Pooled Project (NAPP) to investigate associations between pesticide use and haematological cancer risk. This analysis used data from 547 MM cases and 2700 controls. Pesticide use was evaluated as follows: ever/never use; duration of use (years); and cumulative lifetime-days (LD) (days/year handled years of use). Odds ratios (OR) and 95% confidence intervals (CI) were estimated using logistic regression adjusted for age, province/state of residence, use of proxy respondents, and selected medical conditions. Increased MM risk was observed for ever use of carbaryl (OR=2.02, 95% CI=1.28-3.21), captan (OR=1.98, 95% CI=1.04-3.77), and DDT (OR=1.44, 95% CI=1.05-1.97). Using the Canadian subset of NAPP data, we observed a more than 3-fold increase in MM risk (OR=3.18, 95% CI=1.40-7.23) for 10 cumulative LD of carbaryl use. The association was attenuated but remained significant for >10 LD of carbaryl use (OR=2.44; 95% CI=1.05-5.64; p trend=0.01). For captan, 17.5 LD of exposure was also associated with a more than 3-fold increase in risk (OR=3.52, 95% CI=1.32-9.34), but this association was attenuated in the highest exposure category of >17.5 LD (OR=2.29, 95% CI=0.81-6.43; p trend=0.01). An increasing trend (p trend=0.04) was observed for LD of DDT use (LD>22; OR=1.92, 95% CI=0.95-3.88). In this large North American study of MM and pesticide use, we observed significant increases in MM risk for use of carbaryl, captan and DDT. 2
Introduction Multiple myeloma (MM) is a cancer of terminally-differentiated B-cells (plasma cells) that originates in the bone marrow and is nearly always preceded by a benign condition termed monoclonal gammopathy of undetermined significance (MGUS) 1, 2. MM accounts for approximately 1% of all cancers and up to 10% of all haematological malignancies in the United States 3. Worldwide, MM incidence in men shows substantial geographical variation, with the highest rates observed in developed regions, ranging from 4.3 per 100,000 in Western Europe and North America to 4.6 per 100,000 in Australia/New Zealand 4. The prognosis for MM is less favourable than for many common cancers, with a relative 5-year survival rate of 43% in Canada, and median survival of 3 years following diagnosis 1, 5. Established risk factors for MM include older age (> 65 years), male gender, African American ethnicity, and a family history of MGUS and/or MM 1, 6. Suspected risk factors include autoimmune conditions such as rheumatoid arthritis and lupus, certain viral infections, and occupational and environmental exposures, including having lived or worked on a farm and specific pesticide exposures 6-8. Elevated rates of MM have been observed among farmers compared to the general population, suggesting that exposures in the farming environment contribute to the increased incidence of MM 8. Several case-control studies were conducted during the 1980s and 1990s in Canada and the United States (U.S.) to investigate associations between occupational and environmental exposures, including pesticides, and the risk of MM 9-13. In the U.S., there were positive associations observed between MM and exposure to several insecticides, including carbaryl, chlordane, malathion and dichlorodiphenyltrichloroethane (DDT), and the herbicide glyphosate 12. Similarly, in Canada, significant associations were observed for captan, carbaryl and mecoprop while positive, non-significant associations were observed for chlordane, lindane, DDT, methoxychlor and 2,4-dichlorophenoxyacetic acid (2,4-D) 9, 13. These studies used dichotomous (ever/never) exposure classifications and duration and frequency measures of exposure. For most pesticides, statistical power was limited due to the small numbers of exposed cases and controls 9, 13. 3
To refine risk estimates for previously reported associations and to increase study power to evaluate possible risks from specific pesticides, the Canadian and U.S. case-control studies were pooled to form the North American Pooled Project (NAPP). The objective of this analysis was to evaluate possible associations between individual pesticide exposures and the risk of MM using data from the NAPP. Materials and Methods Study Population The NAPP is comprised of three population-based incident case-control studies conducted by the U.S. National Cancer Institute in Kansas, Iowa/Minnesota, and Nebraska in the 1980s 12, 14, 15, and the Cross Canada Study of Pesticides and Health (CCSPH) 9-11, 13, 16, a population-based incident case-control study that was conducted in Quebec, Ontario, Manitoba, Saskatchewan, Alberta, and British Columbia in the early 1990s. All four studies aimed to investigate the effects of pesticides and other agricultural exposures on the risk of lymphatic and hematopoietic cancers. The present analysis is restricted to a subset of NAPP studies conducted in Iowa, Nebraska, and Canada (CCSPH) where MM cases were recruited. The study design and data collection in the CCSPH were modeled after the U.S. studies, making the data amenable to pooling. Detailed methodologies of the individual studies included in the NAPP have been previously published 12, 14-16. Briefly, eligible participants included white men aged 30 years or older in Iowa/Minnesota, white men and women aged 21 years or older in Nebraska, and men aged 19 years or older in the CCSPH. Deceased participants were considered eligible in Iowa and Nebraska, but not in the CCSPH. In addition to U.S. studies, proxy respondents were also permitted for CCSPH participants requiring assistance due to illness or disability. Incident MM cases were identified using state and provincial cancer registry records, with the exception of Nebraska and Quebec, where cases were recruited from hospitals. Population-based controls were identified using random digit dialing (all studies), Medicare records, and state mortality files (Iowa and Nebraska), health insurance records (Alberta, Saskatchewan, Manitoba, and Quebec), 4
telephone listings (Ontario), and voter lists (British Columbia). Cases and controls were frequencymatched to the overall case distribution by age (±2 years in Nebraska and CCSPH, ±5 years in Iowa/Minnesota), vital status and year of death (if applicable), sex (Nebraska), and province of residence (Canada). The participation rates in the MM subset of NAPP studies were modest for Canadian controls (48%) 16, 17, and higher in Iowa (78%) 12 and Nebraska (85%) 15, 18. Participation rates were higher for cases in Canada (58%) 16, 17, Iowa (84%) 12 and Nebraska (88%) 15, 18. Exposure Assessment A set of a priori pesticides to be investigated in this analysis included agents that showed positive (significant or non-significant) associations in the earlier U.S. and Canadian studies. In addition, exposure information had to be available in the Canadian study and at least one U.S. study. Pesticides that met these criteria were 2,4-D, captan, carbaryl, chlordane, DDT, glyphosate, lindane, malathion, methoxychlor, permethrin, and the pyrethrins. However, since exposure to permethrin and pyrethrin was low among MM cases in the NAPP (n=0 and n=2, respectively), these pesticides were excluded from the analysis. The details of the exposure assessment procedures for each study have been described elsewhere 9-16. Briefly, self-reported information on pesticide use, farming activities, and demographic characteristics was obtained through standardized interviews with participants. The Canadian and U.S. studies involved a sequential ascertainment of pesticide exposures. Individuals who provided an affirmative answer to general questions about pesticide use or exposure to substances within broad groups (i.e.: insecticides, herbicides, fungicides) were subsequently asked more detailed follow-up questions regarding specific agents, including the frequency and duration of exposure. Participants who did not report any pesticide use were excluded from these follow-up questions and were classified as unexposed. Among individuals reporting pesticide use, missing information for duration or frequency of exposure was treated as missing or unknown. Information on duration of pesticide use (years) was collected in all studies, whereas frequency information (days per year) was only collected in the Canada (CCSPH) and Nebraska. However, data 5
from Nebraska were excluded from the analysis since the number of exposed cases for pesticides of interest was low and the proportion of missing data was prohibitively high (>40%). Exposure Metrics To facilitate comparison with previous publications, associations were examined for dichotomous exposure (ever/never pesticide use) and by major chemical classes: phenoxy herbicides, and organochlorine, organophosphate, and carbamate insecticides (listed in Supplementary Table S1). Duration of exposure was evaluated for each individual pesticide using years of self-reported use. Cumulative exposure was investigated using a composite lifetime days (LD) metric, defined as: LD = years of pesticide use days/year of pesticide use. Analyses of cumulative exposure were restricted to the Canadian subset of the NAPP data, where both years and days/year information was available. For subjects with missing information for duration of pesticide use, simple conditional imputation was carried out. Age- and state/province-specific median values for years and days/year were assigned to participants classified as exposed based on the ever/never metric. Imputed values were only assigned if <35% of exposure duration data were missing among cases, and if the proportions of missing data differed by less than 20% between cases and controls. Statistical Analysis Descriptive analyses were performed on potential confounders identified from the literature including age, province/state of residence, use of a proxy respondent, farming history (ever lived or worked on a farm), and personal medical history. Covariates that were significantly (p < 0.05) associated with MM or those that produced meaningful changes ( 5%) in the OR estimates were retained in the final models. Unconditional logistic regression was used to calculate odds ratios (OR) and 95% confidence intervals (CI) for pesticide exposure variables with adjustment for age, province/state of residence, use of proxy respondent, ever being diagnosed with any allergy, hay fever, or rheumatoid arthritis. 6
For all analyses of individual pesticides, the referent population consisted of subjects who did not report any pesticide use, and those who indicated that they had not used that specific agent. Duration of pesticide exposure and cumulative lifetime days (LD) were modeled as ordinal variables and linear trends were examined (p-trend). Cut-offs for categories were based on the median duration and LD among cases and controls for each pesticide. The use of proxy respondents was also considered as an effect modifier and sensitivity analyses were conducted excluding information provided by proxy respondents. In addition, analyses were conducted and results are presented in Supplementary Tables S2-S3 for duration and cumulative lifetime exposure using the dataset without imputed values. All analyses were performed using SAS version 9.3 (Cary, North Carolina, USA) 19. The original studies were approved by the local Research Ethics or Institutional Review Boards. This analysis was approved by the University of Toronto Health Sciences Research Ethics Board and exempted by the U.S. National Institutes of Health Office of Human Subjects Research Protection. Results The analysis included a total of 587 MM cases and 3588 controls from Iowa, Nebraska and six Canadian provinces. Female cases (N=40) and female matched controls (N=707) contributed by the Nebraska study were excluded due to the very low prevalence of pesticide use among females. The youngest MM case in the NAPP dataset was 31 years of age and, therefore, controls age 30 years and younger were excluded (n=181) in order to maintain a comparable distribution of age between cases and controls. This approach was used by Kachuri et al. 13 in a previous analysis of the CCSPH data. The final analysis included 547 male cases and 2700 male controls. Participant demographic characteristics are presented in Table 1. Among the participants, cases were older than controls. This was expected since a common age-matched control group was used for all cancer sites in the NAPP (NHL, HL, STS, and MM), and the majority of MM cases are typically diagnosed at slightly older ages (>65 years) than NHL and HL. Proxy respondents were used for 35% of the cases and 28% of controls overall. Associations of demographic characteristics and medical history covariates with MM were modeled using logistic regression, with adjustment for age (in years) and province or state of residence. 7
A history of any type of cancer among first-degree relatives was significantly associated with MM risk (OR=1.34, 95% CI=1.10-1.62). However, a positive family history of a first-degree relative with any lymphatic or haematopoietic cancer, including MM, was not associated with MM. A number of conditions associated with stimulation of the immune system showed statistically significant inverse associations with MM risk (Table 1). This pattern was observed for any allergy (OR=0.48, 95% CI=0.36-0.63), including specific allergies to food (OR=0.58, 95% CI=0.39-0.87) and drugs (OR=0.47, 95% CI=0.33-0.67), as well as hayfever (OR=0.58, 95% CI=0.37-0.89) and rheumatoid arthritis (OR=0.23, 95% CI=0.14-0.40). Pesticide Exposure (Ever/Never) The associations between MM and self-reported exposure to pesticides, by major chemical class, are presented in Table 2. Individuals were considered exposed to a chemical class of pesticides if they reported use of one or more pesticides within that class. Non-significant positive associations were observed for phenoxy herbicides (OR=1.22, 95% CI=0.97-1.54), organochlorine insecticides (OR=1.24, 95% CI=0.97-1.57), organophosphate insecticides (OR=1.23, 95% CI=0.92-1.64) and carbamate insecticides (OR=1.37, 95% CI=0.97-1.93). The exclusion of proxy respondents did not have an appreciable impact on OR estimates, except for organochorines (OR=1.33, 95% CI=1.01-1.75). Measures of association by individual active agent are presented in Table 3. The most notable finding was the 2-fold increase in risk observed for captan and carbaryl. Compared to the unexposed, men who reported ever using captan had 1.98-times higher odds of MM (95% CI=1.04-3.77). The estimates for captan increased when proxy respondents were excluded (OR=2.37, 95% CI=1.13-4.97). Similar positive associations were observed for carbaryl. For men who reported ever using carbaryl, compared to those who were unexposed, an OR of 2.02 (95% CI=1.28-3.21) was observed; without proxies, an OR= 2.14 (95% CI=1.28-3.58) was observed. An increased risk of MM associated with DDT exposure was also observed. Although smaller in magnitude, these risks were statistically significant (OR=1.44, 95% CI: 1.05-1.97) and increased with the exclusion of proxy respondents (OR=1.62, 95% CI=1.13-2.34). Several other agents exhibited positive associations with MM risk that were not statistically significant including lindane, 2,4-D, and malathion (Table 3). 8
In general, the exclusion of exposure data provided by proxy respondents increased the magnitude of the observed associations for most agents. However, the exclusion of data provided by proxy respondents attenuated the observed associations for glyphosate (OR=1.29, 95% CI=0.90-1.85, compared to OR=1.07 95% CI=0.69-1.66 without proxies), and to a lesser extent, lindane (OR=1.38, 95% CI= 0.90-2.10 compared to OR=1.32, 95% CI=0.82-2.14 without proxies). Years of Pesticide Exposure Several positive associations with MM risk were observed when examining years of selfreported pesticide exposure (Table 4). Use of captan for 7 years was associated with a significantly increased risk of MM (OR=3.14, 95% CI=1.30-7.58); however, the magnitude of increased risk was reduced for >7 years of exposure (OR=1.27, 95% CI=0.49-3.31; p trend=0.13). Despite the statistically significant trend observed for carbaryl (p trend=0.01), the risk increase among men with 6 years of exposure (OR=2.69, 95% CI=1.41-5.13) was larger than for >6 years of exposure (OR=1.76, 95% CI=0.87-3.58). A statistically significant (p trend=0.02) positive trend in risk was also observed for duration of DDT use. Men reporting 9 years of use had modestly elevated risks of MM (OR=1.48, 95% CI=0.93-2.35), and for >9 years of DDT use (OR=1.56, 05% CI=1.01-2.41). Similar to analyses of dichotomous exposure, the observed effects for duration of captan, carbaryl and DDT use were slightly increased after the exclusion of proxy respondents. Cumulative Pesticide Exposure Lifetime Days The general pattern of associations observed for cumulative lifetime days (Table 5) was similar to the findings observed for years of self-reported use. Statistically significant trends in risk were observed for carbaryl (p trend=0.01), DDT (p trend=0.04), and captan (p trend=0.01). Men reporting 10 lifetime days (LD) of carbaryl use had a three-fold increase in MM risk (OR=3.18; 95% CI=1.40-7.23) compared those unexposed to carbaryl. The association was attenuated but remained statistically significant for >10 LD of carbaryl exposure, (OR=2.44; 95% CI=1.05-5.64). Use of DDT was more prevalent compared to carbaryl, resulting in a larger number of cumulative days of exposure. Men reporting 22 LD of DDT use had a modestly elevated risk (OR=1.51, 95% CI=0.74-3.09) compared to unexposed men. The magnitude of this association increased at higher 9
levels of cumulative exposure with an OR of 1.92 (95% CI: 0.95-3.88) for >22 LD of DDT use. The exclusion of information provided by proxy respondents strengthened the latter association. Individuals in the highest category of DDT exposure, (>22 LD) had a statistically significant two-fold increase in MM risk (OR=2.23, 95% CI=1.10-4.55; p trend=0.02). Risks of similar magnitude were also observed for lindane, although the effect estimates for each exposure category did not reach statistical significance: 9 LD of lindane exposure (OR=1.97; 95% CI=0.70-5.58) and >9 LD (OR=2.42, 95% CI=0.84-7.02; p trend=0.05). Examining cumulative captan exposure uncovered a statistically significant positive trend in risk (p trend=0.01), with the risk in the highest exposure level attenuated. Men with 17.5 LD of captan exposure had a greater than three-fold increase in MM risk (OR=3.52, 95% CI=1.32-9.34), and this risk estimate increased upon exclusion of proxy respondents (OR=3.82, 95% CI=1.26-11.56). Among men with >17.5 cumulative days of captan exposure, an increased risk was observed in the full dataset (OR=2.29, 95% CI=0.81-6.43), and after the exclusion of proxy respondents (OR=2.66, 95% CI=0.79-9.02). None of the other pesticides showed a relationship with cumulative exposure. Discussion The role of agricultural exposures, including pesticides, in the etiology of multiple myeloma (MM) remains unclear. Although there is literature focusing on leukemia and lymphoma 20-22, fewer studies have investigated MM. There are several hypothesized mechanisms of pesticide-induced carcinogenesis including oxidative stress and receptor-mediated toxicity 23 that could be relevant to MM. Several pesticides have also been associated with monoclonal gammopathy of undetermined significance (MGUS), an established precursor for MM 24. Pesticide exposure may increase the overall production of reactive oxygen species contributing to DNA damage 23. Additionally, endocrine disruption due to pesticide exposure may alter gene expression networks involved in carcinogenesis 25, 26. MM is characterized by numerous chromosomal aberrations and oncogene and tumor-suppressor gene mutations 1, therefore genotoxicity resulting from pesticide exposure could also contribute to the pathogenesis of the disease. Animal models have also shown that pesticide exposure, particularly carbaryl, may modulate cancer risk via immunotoxic effects and unbalancing the helper T-cell (Th1/Th2) 10
immune response 27. Other emerging pathways include inflammatory and aberrant epigenetic mechanisms; however, the literature in these areas is limited 23. Recently, the International Agency for Research on Cancer (IARC) classified malathion, glyphosate, and DDT as probably carcinogenic to humans (Group 2A) and lindane as carcinogenic to humans (Group 1) 21, 22, based on associations with non-hodkgin lymphoma. In this first analysis of NAPP data, we assessed the relationship between MM risk and several high priority agents, and identified significant positive associations for carbaryl and DDT exposures. We did not observe any statistically significant increases in risk associated with self-reported exposure to glyphosate. Use of malathion was generally not associated with statistically significant excess risks, but there was a consistent increase in point estimates for cumulative exposure, with the largest risks observed for the highest exposure category. For lindane there was some evidence of a positive relationship from the relationship between cumulative lifetime days and MM based on the analysis of CCSPH data alone. In the NAPP dataset, having ever used carbaryl was associated with a two-fold increase in MM risk (OR=2.14). This finding was consistent with previous analyses of CCSPH data, which demonstrated positive and significant associations with carbaryl (OR=1.89; OR=2.71) 9, 13. However, it should also be noted that the CCSPH contributed 63% (n=342) of the MM cases to this NAPP dataset. In addition to ever exposure, the analysis of cumulative exposure based on CCSPH data revealed pronounced effects for self-reported carbaryl use, where ORs of 3.18 and 2.44 were observed for 10 and >10 lifetime exposure days, respectively. Although previous CCSPH analyses of MM have examined associations for days/year of pesticide use, none have investigated cumulative exposure using lifetime days. The carcinogenic potential of carbaryl, a carbamate insecticide, was formally assessed by IARC in 1987, and it was deemed non-classifiable due to inadequate animal studies and human evidence 28. However, more recent assessments by the U.S. EPA consider carbaryl likely to be carcinogenic in humans as a result of increased tumor incidence in mice 29. Modest increases in MM risk were observed for several metrics of DDT exposure. Moreover, the significant trends in risk were based on monotonically increasing OR estimates. The increase in MM risk among men exposed to DDT in this analysis is similar to that observed by Nanni et al. among Italian 11
agricultural workers, although their findings were not significant 30. IARC classifies DDT as probably carcinogenic to humans 21. DDT use has been banned in the U.S. since 1972 and in Canada since 1985 31. Despite not being in use for about a decade at the time of the U.S. and Canadian studies, DDT was reported as the second most commonly used pesticide within the NAPP dataset (10.3% of MM cases and controls). This indicates historical DDT exposure of at least 5 to 10 years prior to the recruitment of cases and controls in the studies comprising the NAPP. DDT and its metabolites are still found in human biological samples and the environment in Canada and the U.S. 31-33 and DDT continues to be used in other parts of the world for agricultural and vector-control purposes. The fungicide captan was also associated with notable increases in MM risk. Odds ratios above two were observed for any exposure to captan, consistent with previous analyses of the CCSPH (OR=2.35; OR=2.96) 9, 13. Although cumulative captan exposure did not result in a monotonic exposureresponse pattern, the trend in risk was statistically significant, as were the effect estimates in each exposure category. In addition, use of captan for 7 years was associated with a three-fold increase in MM risk; however, risk estimates were markedly attenuated for greater than 7 years of self-reported use. This observation is similar to findings in the Agricultural Health Study, which found no association between the highest level of captan exposure and risk of all cancers combined or any specific cancer site 34. This was also cited as one of the studies supporting the recent revision of the carcinogenic classification of captan in 2004 from a probable human carcinogen to not likely carcinogenic to humans 35. Evidence from animal toxicity studies is limited and inconsistent. Captan shares a mode of action with folpet, with both pesticides producing tumors in the gastrointestinal tract in mice, however, captan does not appear to be carcinogenic to rats 36, 37. In interpreting the findings of this analysis, several limitations should be acknowledged. We observed several attenuated (and inversed) point estimates for the highest pesticide exposure category, which may be attributed to exposure measurement error, uncontrolled confounding, chance findings, small numbers of cases and controls. Our analyses could not account for the intensity of exposure, as these data were not collected, and this may have contributed to the attenuation observed for the highest duration or cumulative exposure categories. Furthermore, ordinal categorization of a non- 12
normal continuous variable in the presence of non-differential measurement error can result in attenuation of risk estimates in the highest exposure category. However, simulation studies suggest that this attenuation will not result in risk estimates that are below the true risk observed in the middle category or below the null 38-40. Additionally, recall bias arising from differential reporting of exposure by cases and controls is a concern in case-control studies, and may lead to inflated risk estimates. This pattern of results was not observed in our data, and investigations of self-reported pesticide use among farmers in studies with comparable methods found that cases and controls reported similar numbers of specific pesticides used when this information was volunteered or when it was reported after probing 41. Despite the increase in the overall sample size resulting from pooling data from the CCSPH and U.S. NCI studies, the numbers of exposed participants were still low for some pesticides, and information for duration or frequency was sparse and not collected in all MM studies. Our ability to investigate the effects of high levels of exposure was further limited since few participants reported frequent and long-term pesticide use. Exposure misclassification due to the use of proxy respondents may also influence results. This is supported by the results of sensitivity analyses typically showing an increase in the effect estimates after the exclusion of data provided by proxy respondents. This pattern was observed for most, but not all, pesticides (i.e. glyphosate and lindane) and could be due to differences in the accuracy of information provided by different types of proxy respondents. Unlike the CCSPH, which covered urban and rural areas across six Canadian provinces, the U.S. studies were primarily conducted in areas with large rural farming populations 12. Studies have shown that farmers may be able to recall pesticide use better compared to non-farmers 41, 42, and certain types of proxy respondents, such as friends and family members who also work in agriculture, may be able to provide reasonably accurate and detailed reports of a participant s pesticide exposures compared to proxy respondents recruited from urban centers, or those who had limited knowledge of agricultural practices. Lastly, despite uncovering several statistically significant observations, it should be recognized that a large number of comparisons were made and some of the effect estimates were based on small 13
numbers of exposed cases and controls. Therefore, we cannot exclude the possibility that some of the observed associations may represent chance findings. Despite these limitations, our analysis has several important strengths. The NAPP is one of the largest pooled case-control studies of agricultural exposures and haematopoietic cancers. To our knowledge, this analysis is the first to investigate the association between pesticide exposure and MM risk in a pooled sample of Canadian and the American participants. Since similar pesticides were used in both Canada and the United States, it was possible to investigate the effects of these exposures in the NAPP overall. Furthermore, similarities in the design of the case-control studies facilitated successful pooling of these datasets, which afforded a larger sample size for more comprehensive and powerful analyses. Specifically, the investigation of different aspects of pesticide exposure, such as duration of use and cumulative lifetime exposure, provides an informative and novel addition to this analysis of pesticide use alone. A further advantage of this study lies in the extensive medical history information that was collected by the Canadian and U.S. studies. This allowed us to take into account the influence of several conditions that result in sustained stimulation of the immune system, such as rheumatoid arthritis, systemic lupus erythematosus, certain viral infections, and allergies, which were inversely associated with MM in our data 6. Although some of the pesticides we investigated in the NAPP have been banned in developed countries since the 1970 s and 1980 s, they continue to be used and exported to other developing regions. For instance, while DDT was banned in the United States in 1972, China maintained production until 2007; and currently India remains the largest consumer and sole manufacturer of DDT 43, 44. Use of DDT may also occur under public health exemptions that prioritize vector control for containing the spread of mosquito-born diseases, such as malaria 43, 44. Thus, it is important to characterize the spectrum of pesticide effects on health (including beneficial) to help guide public health decision-making and develop fair and actionable risk management strategies. In summary, this analysis provides further insight on the relationship between pesticide exposures and the risk of MM in a large population that included both rural and urban areas in North America. Significant increases in the risk of MM were observed for several pesticides. Although the 14
pattern of risk was complex, these results are in agreement with some previous findings and offer more refined estimates of MM risk using data on exposure duration and additional metrics of cumulative exposure. This study demonstrates a resourceful use of previously collected data, but also further underscores the need for future studies to collect detailed information on multiple dimensions of pesticide exposure and other factors such as use of protective clothing, that modify those exposures. In addition, further pooling efforts remain an attractive avenue to increase statistical power to detect significant risks and risk patterns with greater precision than possible in the NAPP. 15
Acknowledgements: The authors wish to thank Joe Barker of IMS Inc. for his programming services to pool data from the CCSPH and U.S. case-control studies. This work was supported by the Canadian Cancer Society Research Institute Grant # 703055 and the U.S. National Institutes of Health Intramural Research Program of the National Cancer Institute. The contributions of the late Dr. Helen McDuffie to the CCSPH, and the late Dr. Leon Burmeister to the U.S. studies, are recognized. 16
References 1. Raab MS, Podar K, Breitkreutz I, Richardson PG, Anderson KC. Multiple myeloma. Lancet 2009;374: 324-39. 2. Landgren O, Kyle RA, Pfeiffer RM, Katzmann JA, Caporaso NE, Hayes RB, Dispenzieri A, Kumar S, Clark RJ, Baris D, Hoover R, Rajkumar SV. Monoclonal gammopathy of undetermined significance (MGUS) consistently precedes multiple myeloma: a prospective study. Blood 2009;113: 5412-7. 3. Rajkumar SV. Multiple myeloma: 2013 update on diagnosis, risk-stratification, and management. Am J Hematol 2013;88: 226-35. 4. Ferlay J, Soerjomataram I, Dikshit R, Eser S, Mathers C, Rebelo M, Parkin DM, Forman D, Bray F. Cancer incidence and mortality worldwide: sources, methods and major patterns in GLOBOCAN 2012. Int J Cancer 2015;136: E359-86. 5. Canadian Cancer Society's Steering Committee on Cancer Statistics, Canadian Cancer Statistics 2014. Canadian Cancer Society, 2014. 6. Alexander DD, Mink PJ, Adami HO, Cole P, Mandel JS, Oken MM, Trichopoulos D. Multiple myeloma: a review of the epidemiologic literature. Int J Cancer 2007;120 Suppl 12: 40-61. 7. Lope V, Perez-Gomez B, Aragones N, Lopez-Abente G, Gustavsson P, Plato N, Zock JP, Pollan M. Occupation, exposure to chemicals, sensitizing agents, and risk of multiple myeloma in Sweden. Cancer Epidemiol Biomarkers Prev 2008;17: 3123-7. 8. Perrotta C, Staines A, Cocco P. Multiple myeloma and farming. A systematic review of 30 years of research. Where next? J Occup Med Toxicol 2008;3: 27. 9. Pahwa P, Karunanayake CP, Dosman JA, Spinelli JJ, McDuffie HH, McLaughlin JR. Multiple myeloma and exposure to pesticides: a Canadian case-control study. J Agromedicine 2012;17: 40-50. 10. Pahwa P, McDuffie HH, Dosman JA, McLaughlin JR, Spinelli JJ, Robson D, Fincham S. Hodgkin lymphoma, multiple myeloma, soft tissue sarcomas, insect repellents, and phenoxyherbicides. J Occup Environ Med 2006;48: 264-74. 11. Pahwa P, McDuffie HH, Dosman JA, Robson D, McLaughlin JR, Spinelli JJ, Fincham S. Exposure to animals and selected risk factors among Canadian farm residents with Hodgkin's disease, multiple myeloma, or soft tissue sarcoma. J Occup Environ Med 2003;45: 857-68. 12. Brown LM, Burmeister LF, Everett GD, Blair A. Pesticide exposures and multiple myeloma in Iowa men. Cancer Causes Control 1993;4: 153-6. 13. Kachuri L, Demers PA, Blair A, Spinelli JJ, Pahwa M, McLaughlin JR, Pahwa P, Dosman JA, Harris SA. Multiple pesticide exposures and the risk of multiple myeloma in Canadian men. Int J Cancer 2013;133: 1846-58. 17
14. Blair A, Cantor KP, Zahm SH. Non-hodgkin's lymphoma and agricultural use of the insecticide lindane. Am J Ind Med 1998;33: 82-7. 15. Zahm SH, Weisenburger DD, Babbitt PA, Saal RC, Vaught JB, Cantor KP, Blair A. A casecontrol study of non-hodgkin's lymphoma and the herbicide 2,4-dichlorophenoxyacetic acid (2,4-D) in eastern Nebraska. Epidemiology 1990;1: 349-56. 16. McDuffie HH, Pahwa P, McLaughlin JR, Spinelli JJ, Fincham S, Dosman JA, Robson D, Skinnider LF, Choi NW. Non-Hodgkin's lymphoma and specific pesticide exposures in men: cross-canada study of pesticides and health. Cancer Epidemiol Biomarkers Prev 2001;10: 1155-63. 17. McDuffie HH, Pahwa P, Karunanayake CP, Spinelli JJ, Dosman JA. Clustering of cancer among families of cases with Hodgkin Lymphoma (HL), Multiple Myeloma (MM), Non-Hodgkin's Lymphoma (NHL), Soft Tissue Sarcoma (STS) and control subjects. BMC Cancer 2009;9: 70. 18. Zahm SH, Weisenburger DD, Babbitt PA, Saal RC, Vaught JB, Blair A. Use of hair coloring products and the risk of lymphoma, multiple myeloma, and chronic lymphocytic leukemia. Am J Public Health 1992;82: 990-7. 19. 9.3 SV. Cary, NC, USA: Institute Inc., 2011. 20. Schinasi L, Leon ME. Non-Hodgkin lymphoma and occupational exposure to agricultural pesticide chemical groups and active ingredients: a systematic review and meta-analysis. Int J Environ Res Public Health 2014;11: 4449-527. 21. Loomis D, Guyton K, Grosse Y, El Ghissasi F, Bouvard V, Benbrahim-Tallaa L, Guha N, Mattock H, Straif K, International Agency for Research on Cancer Monograph Working Group ILF. Carcinogenicity of lindane, DDT, and 2,4-dichlorophenoxyacetic acid. Lancet Oncol 2015;16: 891-2. 22. Guyton KZ, Loomis D, Grosse Y, El Ghissassi F, Benbrahim-Tallaa L, Guha N, Scoccianti C, Mattock H, Straif K, International Agency for Research on Cancer Monograph Working Group ILF. Carcinogenicity of tetrachlorvinphos, parathion, malathion, diazinon, and glyphosate. Lancet Oncol 2015;16: 490-1. 23. Alavanja MC, Ross MK, Bonner MR. Increased cancer burden among pesticide applicators and others due to pesticide exposure. CA Cancer J Clin 2013;63: 120-42. 24. Landgren O, Kyle RA, Hoppin JA, Beane Freeman LE, Cerhan JR, Katzmann JA, Rajkumar SV, Alavanja MC. Pesticide exposure and risk of monoclonal gammopathy of undetermined significance in the Agricultural Health Study. Blood 2009;113: 6386-91. 25. Schug TT, Janesick A, Blumberg B, Heindel JJ. Endocrine disrupting chemicals and disease susceptibility. J Steroid Biochem Mol Biol 2011;127: 204-15. 26. Kretschmer XC, Baldwin WS. CAR and PXR: xenosensors of endocrine disrupters? Chem Biol Interact 2005;155: 111-28. 27. Jorsaraei SG, Maliji G, Azadmehr A, Moghadamnia AA, Faraji AA. Immunotoxicity effects of carbaryl in vivo and in vitro. Environ Toxicol Pharmacol 2014;38: 838-44. 18
28. International Agency for Research on Cancer, IARC Monographs on the Evaluation of Carcinogenic Risks to Humans. World Health Organization, 1987. 29. US Environmental Protection Agency, Amended reregistration eligibility decision (RED) for Carbaryl. Office of Prevention, Pesticides, and Toxic Substances, 2008. 30. Nanni O, Falcini F, Buiatti E, Bucchi L, Naldoni M, Serra P, Scarpi E, Saragoni L, Amadori D. Multiple myeloma and work in agriculture: results of a case-control study in Forli, Italy. Cancer Causes Control 1998;9: 277-83. 31. US Environmental Protection Agency, DDT. General Fact Sheet. National Pesticide Information Center, 1999. 32. Spinelli JJ, Ng CH, Weber JP, Connors JM, Gascoyne RD, Lai AS, Brooks-Wilson AR, Le ND, Berry BR, Gallagher RP. Organochlorines and risk of non-hodgkin lymphoma. Int J Cancer 2007;121: 2767-75. 33. Canadian Council of Ministers of the Environment, Canadian tissue residue guildines for the protection of wildlife consumers of aquatic biota: DDT (total). 1999. 34. Greenburg DL, Rusiecki J, Koutros S, Dosemeci M, Patel R, Hines CJ, Hoppin JA, Alavanja MC. Cancer incidence among pesticide applicators exposed to captan in the Agricultural Health Study. Cancer Causes Control 2008;19: 1401-7. 35. Gordon E. Captan: transition from 'B2' to 'not likely'. How pesticide registrants affected the EPA Cancer Classification Update. J Appl Toxicol 2007;27: 519-26. 36. Cohen SM, Gordon EB, Singh P, Arce GT, Nyska A. Carcinogenic mode of action of folpet in mice and evaluation of its relevance to humans. Crit Rev Toxicol 2010;40: 531-45. 37. US Environmental Protection Agency, Captan: Cancer Reclassification; Amendment of Reregistration Eligibility Decision; Notice of Availability., 2004. 38. Dosemeci M, Wacholder S, Lubin JH. Does nondifferential misclassification of exposure always bias a true effect toward the null value? Am J Epidemiol 1990;132: 746-8. 39. Wacholder S. When measurement errors correlate with truth: surprising effects of nondifferential misclassification. Epidemiology 1995;6: 157-61. 40. Wang D, Gustafson P. On the Impact of Misclassification in an Ordinal Exposure Variable. Epidemiologic Methods 2014;3: 97-106. 41. Blair A, Zahm SH. Patterns of pesticide use among farmers: implications for epidemiologic research. Epidemiology 1993;4: 55-62. 42. Blair A, Tarone R, Sandler D, Lynch CF, Rowland A, Wintersteen W, Steen WC, Samanic C, Dosemeci M, Alavanja MC. Reliability of reporting on life-style and agricultural factors by a sample of participants in the Agricultural Health Study from Iowa. Epidemiology 2002;13: 94-9. 19
43. Stockholm Convention on Persistent Organic Pollutants, Report of the Third Meeting of the Expert Group on the assessment of the production and use of DDT and its alternatives for disease vector control, 2010. 44. van der Berg H, Global status of DDT and its alternatives for use in vector control to prevent disease. Stockholm Convention on Persistent Organic Pollutants, 2008. 20
Table&1:&Selected"characteristics"of"multiple"myeloma"cases"and"controls"in"the"North"American"Pooled" Project"(NAPP)&& & & Cases&& (N=547)& Controls& (N=2700)& Age"(years;"mean"+"SD)" 67.2"±"11.4" 62.7"±"15.0" OR 1 && (95%&CI)& " " N&(%)& N&(%)& & Age"group"(years)" " " 31H39" 11"(2.0)" 249"(9.2)" 40H49" 32"(5.9)" 330"(12.2)" " 50H59" 80"(14.6)" 431"(16.0)" " 60H69" 172"(31.4)" 721"(26.7)" " 70H79" 182"(33.3)" 627"(23.2)" " 80" 70"(12.8)" 342"(12.7)" " Residence" " " Iowa" " 173"(31.6)" 650"(24.1)" Nebraska" 32"(5.9)" 701"(26.0)" Quebec"" 37"(6.8)" 255"(9.4)" " Ontario" 103"(18.8)" 516"(19.1)" " Manitoba" 25"(4.6)" 100"(3.7)" " Saskatchewan"" 28"(5.1)" 83"(3.1)" " Alberta"" 58"(10.6)" 177"(6.6)" " British"Columbia" 91"(16.6)" 218"(8.1)" " Vital"status"" " " " Living" 354"(64.7)" 1790"(66.3)" " Deceased"" 92"(16.8)" 458"(17.0)" " Unknown/Missing"" 101"(18.5)" 452"(16.7)" " Education"level" " " <8"years" 58"(10.6)" 256"(9.5)" 1.00" 8H11"years" 208"(38.0)" 823"(30.5)" 1.15"(0.82H1.61)" High"school"(12"years)" 99"(18.1)" 535"(19.8)" 1.24"(0.84H1.82)" College,"junior"college"(1H3"years)" 83"(15.2)" 359"(13.3)" 1.05"(0.70H1.56)" College"graduate" 40"(7.3)" 288"(10.7)" 0.92"(0.58H1.46)" PostHgraduate" 18"(3.3)" 166"(6.2)" 0.63"(0.35H1.14)" Other 2 " 41"(7.5)" 273"(10.1)" 0.92"(0.58H1.47)" Respondent"Type" " " " Subject"" 356"(65.1)" 1945"(72.0)" 1.00" Proxy"" 191"(34.9)" 755"(28.0)" 1.61"(1.30H2.01)" Ever"lived/worked"on"farm" " " " No" 207"(37.8)" 1066"(39.5)" 1.00" Yes" 340"(62.2)" 1634"(60.5)" 1.08"(0.87H1.33)" " 1"
Ever"smoked"cigarettes" " " " No" " 158"(28.8)" 872"(32.3)" 1.00" Yes" 389"(71.1)" 1828"(67.7)" 1.19"(0.97H1.47)" Any"cancer"in"first"degree"relative" " " " No" 278"(50.8)" 1661"(61.5)" 1.00" Yes" 269"(49.2)" 1039"(38.5)" 1.34"(1.10H1.62)" Lymphatic"or"hematopoietic"cancer"in" first"degree"relative" " " " No" 525"(96.0)" 2576"(95.4)" 1.00" Yes" 22"(4.0)" 124"(4.6)" 0.81"(0.50H1.31)" Ever"diagnosed"with"any"allergy" " " " No" 483"(88.3)" 2158"(79.9)" 1.00" Yes" 64"(11.7)" 542"(20.1)" 0.48"(0.36H0.63)" Ever"diagnosed"with"food"allergy" " " " No" 515"(94.2)" 2494"(92.4)" 1.00" Yes" 32"(5.9)" 206"(7.6)" 0.58"(0.39H0.87)" Ever"diagnosed"with"drug"allergy" " " " No" 504"(92.1)" 2373"(87.9)" 1.00" Yes" 43"(7.9)" 327"(12.1)" 0.47"(0.33H0.67)" Ever"diagnosed"with"asthma" " " " No" 521"(95.3)" 2525"(93.5)" 1.00" Yes" 26"(4.8)" 175"(6.5)" 0.69"(0.45H1.07)" Ever"diagnosed"with"hay"fever" " " " No" 522"(95.4)" 2467"(91.4)" 1.00" Yes" 25"(4.6)" 233"(8.6)" 0.58"(0.37H0.89)" Ever"diagnosed"with"mononucleosis" " " " No" 540"(98.7)" 2656"(98.4)" 1.00" Yes" 7"(1.3)" 44"(1.6)" 1.01"(0.44H2.31)" Ever"diagnosed"with"rheumatoid"arthritis" " " " No" 531"(97.1)" 2497"(92.5)" 1.00" Yes" 16"(2.9)" 203"(7.5)" 0.23"(0.14H0.40)" Ever"diagnosed"with"tuberculosis" " " " No" 539"(98.5)" 2683"(99.4)" 1.00" Yes" 8"(1.5)" 17"(0.6)" 1.61"(0.67H3.84)" 1" Adjusted"for"age,"and"province/state"of"residence" 2" Educational"level"that"was"not"reported"as"one"of"the"aforementioned"categories& " 2"
Table&2:&Adjusted"odds"ratios"(OR)"and"corresponding"95%"confidence"intervals"(CI)"for"multiple" myeloma"in"relation"to"any"pesticide"exposure"(ever/never"use)"in"the"napp,"with"individual"pesticides" grouped"by"major"chemical"class"" Chemical&Class& & Cases&(%)& (n=547)" & Controls&(%)& (n=2700)" & Adjusted&OR 1 && (95%&CI)& Cases&(%)& (n=356)& Proxy&Respondents&Excluded& Controls&(%)& (n=1945)& Adjusted&OR 2 & (95%&CI)& Phenoxy"herbicides" " " " " " No" 412"(75.3)" 2115"(78.3)" 1.00" 258"(72.5)" 1471"(75.6)" 1.00" Yes" 135"(24.7)" 585"(21.7)" 1.22"(0.97H1.54)" 98"(27.5)" 474"(24.4)" 1.26"(0.96H1.65)" Organochlorine"insecticides" " " " " " No" 425"(77.7)" 2153"(79.7)" 1.00" 264"(74.2)" 1503"(77.3)" 1.00" Yes" 122"(22.3)" 547"(20.3)" 1.24"(0.97H1.57)" 92"(25.8)" 442"(22.7)" 1.33"(1.01H1.75)" Organophosphate"insecticides" " " " " " No" 474"(86.6)" 2355"(87.2)" 1.00" 300"(84.3)" 1657"(85.2)" 1.00" Yes" 73"(13.4)" 345"(12.8)" 1.23"(0.92H1.64)" 56"(15.7)" 288"(14.8)" 1.20"(0.86H1.67)" Carbamate"insecticides" " " " " " No" 496"(90.7)" 2471"(91.5)" 1.00" 317"(89.0)" 1761"(90.5)" 1.00" Yes" 51"(9.3)" 229"(8.5)" 1.37"(0.97H1.93)" 39"(11.0)" 184"(9.5)" 1.41"(0.96H2.09)" & & & 1 "Adjusted"for"age,"province/state"of"residence,"use"of"proxy"respondent,"ever"diagnosed"with"any"allergy,"rheumatoid"arthritis," or"hayfever"" 2 "Adjusted"for"age,"province/state"of"residence,"ever"diagnosed"with"any"allergy,"rheumatoid"arthritis,"or"hayfever"""" & &
" Pesticide& Table&3:&Adjusted"odds"ratios"(OR)"and"corresponding"95%"confidence"intervals"(CI)"for"multiple" myeloma"in"relation"to"any"selfhreported"exposure"(ever/never"use)"to"selected"individual"pesticides"in"" the"napp&& & & Cases&(%)& (n=547)" & Controls&(%)& (n=2700)" Adjusted&OR 1 & (95%&CI)& Cases&(%)& (n=356)& Proxy&Respondents&Excluded& Controls&(%)& (n=1945)& 2,4HD" " " " " " " Adjusted&OR 2 & (95%&CI)& No" 420"(76.8)" 2147"(79.5)" 1.00" 264"(74.2)" 1496"(76.9)" 1.00" Yes" 127"(23.2)" 553"(20.5)" 1.20"(0.95H1.52)" 92"(25.8)" 449"(23.1)" 1.22"(0.93H1.61)" Captan" " " " " " " No" 532"(97.3)" 2660"(98.5)" 1.00" 345"(96.9)" 1913"(98.4)" 1.00" Yes" 15"(2.7)" 40"(1.5)" 1.98"(1.04H3.77)" 11"(3.1)" 32"(1.7)" 2.37"(1.13H"4.97)" Carbaryl" " " " " " " No" 518"(94.7)" 2604"(96.4)" 1.00" 333"(93.5)" 1869"(96.1)" 1.00" Yes" 29"(5.30)" 96"(3.6)" 2.02"(1.28H3.21)" 23"(6.5)" 76"(3.9)" 2.14"(1.28H"3.58)" Chlordane" " " " " " " No" 509"(93.0)" 2510"(93.0)" 1.00" 323"(90.7)" 1788"(91.9)" 1.00" Yes" 38"(7.0)" 190"(7.0)" 1.15"(0.78H1.68)" 33"(9.3)" 157"(8.1)" 1.32"(0.87H"1.99)" DDT" " " " " " " No" 481"(87.9)" 2430"(90.0)" 1.00" 306"(86.0)" 1739"(89.4)" 1.00" Yes" 66"(12.1)" 270"(10.0)" 1.44"(1.05H1.97)" 50"(14.0)" 206"(10.6)" 1.62"(1.13H2.34)" Glyphosate" " " " " " " No" 502"(91.8)" 2504"(92.7)" 1.00" 327"(91.9)" 1771"(91.0)" 1.00" Yes" 45"(8.2)" 196"(7.3)" 1.29"(0.90H1.85)" 29"(8.1)" 174"(9.0)" 1.07"(0.69H1.66)" Lindane" " " " " " " No" 514"(94.0)" 2546"(94.3)" 1.00" 331"(93.0)" 1818"(93.5)" 1.00" Yes" 33"(6.0)" 154"(5.7)" 1.38"(0.90H2.10)" 25"(7.0)" 127"(6.5)" 1.32"(0.82H2.14)" Malathion" " " " " " " No" 499"(91.2)" 2474"(91.6)" 1.00" 314"(88.2)" 1751"(90.0)" 1.00" Yes" 48"(8.8)" 226"(8.4)" 1.19"(0.84H1.69)" 42"(11.8)" 194"(10.0)" 1.33"(0.91H1.94)" Methoxychlor" " " " " " " No" 503"(92.0)" 2498"(92.5)" 1.00" 320"(89.9)" 1767"(90.9)" 1.00" Yes" 44"(8.0)" 202"(7.5)" 1.12"(0.77H1.61)" 36"(10.1)" 178"(9.1)" 1.23"(0.82H1.85)" " 1 "Adjusted"for"age,"province/state"of"residence,"use"of"proxy"respondent,"ever"diagnosed"with"any"allergy,"rheumatoid"arthritis," or"hayfever" 2" Adjusted"for"age,"province/state"of"residence,"ever"diagnosed"with"any"allergy,"rheumatoid"arthritis,"or"hayfever"" " 4"