Skip to main content
PMC home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2021 May 1.
Published in final edited form as: Int J Hyg Environ Health. 2020 Feb 19;226:113486. doi: 10.1016/j.ijheh.2020.113486

Prenatal Pesticide Exposure and Childhood Leukemia – a California statewide case-control study

Andrew S Park a, Beate Ritz a, Fei Yu b, Myles Cockburn c, Julia E Heck a,d
PMCID: PMC7174091  NIHMSID: NIHMS1564216  PMID: 32087503

Abstract

Background

A number of epidemiologic studies with a variety of exposure assessment approaches have implicated pesticides as risk factors for childhood cancers. Here we explore the association of pesticide exposure in pregnancy and early childhood with childhood acute lymphoblastic leukemia (ALL) and acute myeloid leukemia (AML) utilizing land use and pesticide use data in a sophisticated GIS tool.

Methods

We identified cancer cases less than 6 years of age from the California Cancer Registry and cancer-free controls from birth certificates. Analyses were restricted to those living in rural areas and born 1998-2011, resulting in 162 cases of childhood leukemia and 9,805 controls. Possible carcinogens were selected from the Environmental Protection Agency’s classifications and pesticide use was collected from the California Department of Pesticide Regulation’s (CDPR) Pesticide Use Reporting (PUR) system and linked to land-use surveys. Exposures for subjects were assessed using a 4000m buffer around the geocoded residential addresses at birth. Unconditional logistic and hierarchical regression models were used to assess individual pesticide and pesticide class associations.

Results

We observed elevated risks for ALL with exposure to any carcinogenic pesticide (adjusted Odds Ratio (aOR): 2.83, 95% CI: 1.67-4.82), diuron (Single-pesticide model, adjusted (OR): 2.38, 95% CI: 1.57-3.60), phosmet (OR: 2.10, 95% CI: 1.46-3.02), kresoxim-methyl (OR: 1.77, 95% CI: 1.14-2.75), and propanil (OR: 2.58, 95% CI: 1.44-4.63). Analyses based on chemical classes showed elevated risks for the group of 2,6-dinitroanilines (OR: 2.50, 95% CI: 1.56-3.99), anilides (OR: 2.16, 95% CI: 1.38-3.36), and ureas (OR: 2.18, 95% CI: 1.42-3.34).

Conclusion

Our findings suggest that in rural areas of California exposure to certain pesticides or pesticide classes during pregnancy due to residential proximity to agricultural applications may increase the risk of childhood ALL and AML. Future studies into the mechanisms of carcinogenicity of these pesticides may be beneficial.

Keywords: Childhood leukemia, pesticide, agriculture, ambient exposure, pregnancy

Introduction

Leukemia, which makes up 40% of all childhood cancers, is the most frequent childhood malignancy. The most common subtype, acute lymphocytic leukemia (ALL) alone accounts for 25% of all childhood malignancies in children younger than 15 years of age (Howlader, 2013b), and 78% of all childhood leukemia cases. Nevertheless, it is a rare disease with an incidence of 39 cases per million, with diagnoses peaking between ages 2-5 and the highest rates estimated for Hispanics (CDC, 2009). Acute myeloid leukemia accounts for most of the remaining cases of childhood leukemia (16%). Few risk factors for childhood leukemia are considered established apart from ionizing radiation and specific genetic abnormalities.

In 2010, 173 million pounds of pesticide active ingredients were used in California (CDPR, 2011). Higher concentrations of pesticides have been measured in communities abutting agricultural fields, as measured by urine and hand wipe samples as well as in house dust (Loewenherz et al., 1997; Lu et al., 2000; Simcox et al., 1995; Ward et al., 2006). Previous studies have shown that pesticides and their metabolites (organophosphates, pyrethroids) persist in media such as house dust and clothing in families of farmworkers living near agricultural fields (Arcury et al., 2007; Bradman et al., 2005; Bradman et al., 2007).These families show a higher body burden of pesticides compared to those living further away from fields (Curwin et al., 2007; Fenske et al., 2005). This is of great concern as pesticide residues may build up in residences located in agricultural communities and expose pregnant women and infants (Arcury et al., 2007).

A number of previous studies suggested that pesticide exposures are risk factors for childhood ALL (Bailey et al., 2015; Infante-Rivard and Weichenthal, 2007; Jurewicz and Hanke, 2006; Nasterlack, 2007; Vinson et al., 2011). A meta-analysis (Vinson et al., 2011) concluded that exposure during pregnancy to any pesticide exposure increases the risk of leukemias (48%) and lymphomas (53%) in offspring. However, the contributing studies mainly relied on parental interviews to assess in-home pesticide use or professional applications of residences, and the retrospective assessment of exposures made these studies vulnerable to recall bias since parents of children with cancer may recall exposures in more detail and may want to explain their child’s disease. In addition, interviews fail to assess levels of ambient exposure. In 2007, a study (Infante-Rivard and Weichenthal, 2007) critiqued the literature as lacking clarity on the most important time period of exposure, the type of leukemia most impacted by pesticide exposures, and which parent’s exposure may be more important. This study aims to shed light on the subtypes of leukemia and to determine any differences in risks whether children were exposed during pregnancy or early childhood (<1 year).

Importantly, most prior studies did not report on risk related to individual pesticides, instead they assessed pesticides by type (insecticides, fungicides, herbicides, fumigants, etc.), chemical class (organophosphates, organochlorines, triazines, azoles, etc.), classifications according to level of carcinogenicity (Group A, B, probable carcinogen, etc.) (Alderton et al., 2006; Infante-Rivard and Sinnett, 1999; Lowengart et al., 1987), or they targeted specific pesticides but investigated all childhood leukemias together instead of separating ALL and AML (Reynolds et al., 2005). Our study expands upon earlier publications by identifying specific pesticide exposures and by assessing risk of ALL and AML separately. Since agricultural pesticide use changes frequently due to improvements in formulations and to address pest adaptation and evolution (Tomasetto et al., 2017), updates with more recent exposure data are needed. We are focusing on children born in more recent years (1998-2011) and we are employing a very sophisticated GIS tool we developed that combines land use and pesticide use records for those years; we applied this tool to all rural areas in California that use pesticides agriculturally because urban communities apply pesticides for non-agricultural purposes such as home and building fumigations or right-of way and other public pesticide applications, which are not reported to the PUR on a fine spatial level (reported only at the county-level). While agricultural pesticide use is generally rarer in urban areas, urban dwellers nevertheless experience other types of ambient pesticide exposures that may impact risk (Heck et al., 2013; Park et al., 2017). Hence we decided to focus on agricultural applications and rural populations with spatially well captured pesticide exposures. This exposure assessment method does not rely on recall and does not suffer the same limitations as many prior studies.

Methods

Study Population

Cases were collected from records of incident ALL [International Classification of Childhood Cancer, Third Edition (ICCC) code 011] and AML (ICCC code 012) diagnosed in 1986-2012 (born 1983-2011) in the California Cancer Registry, younger than age 6, and born in California. Using a probabilistic linkage program (LinkPlus, CDC), we linked cases to birth certificates using first and last names, date of birth, and social security number when available. With this method, 89% of cases were matched to a California birth certificate. The remaining 11% were likely born out of state and moved to California prior to diagnosis (Urayama et al., 2009).

Controls were selected at random from California birth records for the same time period, frequency-matched by year of birth to all childhood cancers in the study. Controls had no record of a cancer diagnosis in California before age 6. Results from cases and controls were restricted to birth years 1998-2011 in this study. As this was a record-based study, we did not seek informed consent from individual subjects. The study received approvals from the human subject protection boards of the University of California, Los Angeles, the University of Southern California, and the California Committee for the Protection of Human Subjects.

Exposure Assessment

Starting in 1974, California’s Department of Pesticide Regulation required farmers to report the use of restricted pesticides, defined as pesticides whose application was restricted by federal law due to potential harm to people, crops, or the environment, to the state via California Pesticide-Use Reports (PUR). In 1990, this California PUR program was expanded to require full reporting of all agricultural pesticide use. In the PUR, detailed information is given regarding the active ingredients, the acreage treated, the amounts used, the crops that are treated, and the location and date of the application. Our group combined data from the PUR with data from land-use surveys from the California’s Public Land Survey System (PLSS). The PLSS is a countywide survey conducted every 7-10 years by the California Department of Water Resources recording the extent of land use in terms of crop cover by location. The combination of the PUR and land-use data allowed us to more precisely locate pesticide applications by linking them with their respective crops. PUR data tracks the target crop(s) for each pesticide application, which is then mapped using the land-use data to the field containing the target crop(s). We calculated monthly and annual application rates (total pounds applied per acre in a PLSS section, within a given time period) for each pesticide as detailed elsewhere (Cockburn et al., 2011; Goldberg et al., 2008).

Previously, we successfully characterized chronic organophosphate (OP) pesticide exposure via differential DNA methylation (Paul et al., 2018). We used our geographic information system-based pesticide exposure assessment tools to generate long-term OP pesticide exposure measures in healthy control individuals and identified epigenetic profiles related to chronic OP exposures. Using systems analysis and marker set enrichment, we found that the OP-associated CpGs were enriched for nicotinic and muscarinic acetylcholine esterase receptors as a response to OP exposure.

Case data were obtained from California Cancer Registry records. Parental demographics such as age, education, race/ethnicity, socioeconomic status, as well as child’s information including gestational age, and child’s sex were obtained from birth certificates and additionally from the year 2000 census. Neighborhood socioeconomic status was measured with a multi-factorial index that used principal components analysis to develop a single, 5-level measure from seven census-level indicators (education, median household income, percent living 200% below poverty, percent blue-collar workers, percent older than 16 years without employment, median rent, and median house value) (Yost et al., 2001). Based on the date of last reported menses we excluded children with implausibly long (>45 weeks, n=2729) gestations, likely non-viable births (gestational age <20 weeks or birth weight <500g, n=32), and those with missing data for gestation length (n=12,786). In addition, those with a residential address outside of California (n=632), and controls who had died before age 6 (n=1,202) were excluded from the dataset resulting in 5,112 cases of leukemia and 270,776 controls prior to restriction. We used an urban/rural designation based on rural-urban commuting area codes (RUCA) created by a collaboration of government organizations based on Census Tract information (Morrill, 2013) to restrict our sample to rural residents only (RUCA levels 4-10). Our rural study population contained 132 cases of ALL, 30 cases of AML, and 9,805 controls.

Residential addresses as listed on the birth certificate were mapped using Geographic Information System (GIS) tools based on street address (54%), street intersection (38%), city centroid, or ZIP code centroid (7%). We examined several buffer sizes (500m, 2000m, 4000m) around the home . Using the residential address, for each buffer size, the buffer radius is drawn around the residence to determine pesticide use within the respective buffer. Due to the distances that pesticides can reach from where they are initially applied (pesticide drift) (Wofford et al., 2014), and to improve sample size, we present results at the 4000m distance. We also examined smaller buffers (500m, 2000m) in sensitivity analyses.

We pre-selected 133 pesticides considered possible, likely, or probable carcinogens by the EPA (Vinson et al., 2011), (See Supplement 1 for more detail on EPA carcinogen classifications) (EPA, 2017). We further categorized pesticides according to physiochemical type and class, based on information from the Pesticide Action Network (PAN) Pesticide Database (Kegley, 2014) and we also considered 6 additional pesticides (methyl bromide, diazinon, paraquat, chlorpyrifos, glyphosate, and simazine (Metayer et al., 2013)) because these chemicals are widely used in California and little is known about potential adverse effects from co-exposures. Of these, we report results for 65 pesticides for which at least 10 cases were exposed during the study years.

Approximately 85% of PUR reports have an exact date of application, while the remaining 15% only have the year. We conducted sensitivity analyses to elucidate whether exposure in a specific trimester was most relevant for ALL risk.

Statistical Analyses

Unconditional logistic regression was utilized to derive ORs and 95% CIs adjusting for the matching factor (birth year) in single and multiple-pesticides models for the 65 selected pesticides. We examined ‘ever’ exposure during pregnancy for each pesticide as well as for chemical class for ALL and AML outcomes. The single-pesticide models adjusted for any exposure to another carcinogenic pesticide as a binary variable. Multiple-pesticide models, co-adjusted for ever/never exposure to all other pesticides selected via semi-Bayesian hierarchical logistic regression, i.e. assuming that the estimated effects are either 1) drawn from the same distribution for all carcinogenic pesticides or 2) are specific to a pesticide class (2,6-dinitroanilines, azoles, chloroacetanilides, dithiocarbamates, n-methyl carbamates, organochlorines, organophosphates, pyrethroids, substituted benzenes, and triazines). Both assumptions yielded similar results, thus here we present the hierarchical logistic regression model results using the first assumption (same distribution of effects for all carcinogenic pesticides) only. Due to the highly correlated nature of agricultural pesticide use and exposure, we employed hierarchical logistic modelling (HLM) in an attempt to identify the pesticides with the most consistently estimated effects after adjusting for other carcinogenic pesticide exposures. This, however, assumes that all carcinogenic pesticides have a similar (or mean) effect on childhood leukemia for exposure in pregnancy or early childhood. This assumption is likely not valid. We also conducted sensitivity analyses examining the entire state among those born 1998-2011 and restricted to those exposed to at least one pesticide in the PUR database.

Selection of variables for adjustment was based upon literature review as well as the 10% change in estimate criterion (Chow et al., 2010; Howlader, 2013a; Johnson et al., 2009; Oksuzyan et al., 2015). We adjusted for birth year (matching factor), mother’s race (White non-Hispanic, Black, and Other race/refused to report), neighborhood socioeconomic index, and the binary indicator for any other carcinogenic pesticide exposure as possible confounders. We considered for inclusion in models neighborhood and individual level socioeconomic indicators: maternal education (years), paternal education (years), and the source of payment for prenatal care, which was defined as private insurance (including Health Maintenance Organizations, Blue Cross-Blue Shield, or any other private insurance) versus other payment methods (including self-pay and government aid programs, such as Medicare, Medi-Cal, worker’s compensation, Title V, and CHAMPUS/TRICARE), which we have previously observed to be a reasonable proxy for family income (Ritz et al., 2007). However, these socioeconomic factors were not included in final models due to not fulfilling the 10% change in estimate criterion.

Results

In terms of demographic characteristics (Table 1) fathers of ALL cases were more likely to be White non-Hispanics and Hispanics than any other race. Mothers of ALL cases were older and more likely to be Hispanic. Families of ALL cases were more likely to have had their prenatal care paid by private insurance. ALL cases were more often male children.

Table 1.

Demographic characteristics of children in California born in 1998-2011 living in rural areas.

Characteristic ALL N = 132 AML N = 30 Controls N = 9805
Sex of Child, n (%)
 Male 77 (58.3) 23 (76.7) 5062 (51.6)
 Female 55 (41.7) 7 (23.3) 4743 (48.4)
Paternal Race/Ethnicity, n (%)
 White non-Hispanic 38 (28.8) 9 (30.0) 2916 (29.7)
 Hispanic 69 (52.3) 15 (50.0) 4657 (47.5)
 Other 25 (18.9) 6 (20.0) 2232 (22.8)
Maternal Race/Ethnicity, n (%)
 White non-Hispanic 44 (33.3) 10 (33.3) 4033 (41.1)
 Hispanic 80 (60.6) 18 (60.0) 5036 (51.4)
 Other 8 (6.1) 2 (6.7) 736 (7.5)
Maternal Age, n (%)
 19 or less 19 (14.4) 4 (13.3) 1251 (12.8)
 20-24 37 (28.0) 7 (23.3) 2831 (28.9)
 25-29 32 (24.2) 8 (26.7) 2672 (27.3)
 30-34 26 (19.7) 4 (13.3) 1856 (18.9)
 35 and older 18 (13.6) 7 (23.3) 1186 (12.1)
 Missing 9
Maternal Education (years), n (%)
 8 or less 14 (10.7) 4 (13.3) 1134 (11.8)
 9-11 28 (21.4) 6 (20.0) 1931 (20.0)
 12 45 (34.4) 8 (26.7) 3213 (33.3)
 13 to 15 27 (20.6) 8 (26.7) 2179 (22.6)
 16 more 17 (13.0) 4 (13.3) 1194 (12.4)
 Missing 154
Principal Payment of Prenatal Care, n (%)
 Private/HMO/BCBS 59 (45.4) 12 (40.0) 3435 (35.7)
 MediCal/Govt/Self-pay 71 (54.6) 18 (60.0) 6198 (64.3)
 Missing 12 172
Census-based SES index level, n (%)
 1 (reference) 33 (25) 13 (43.3) 2924 (29.8)
 2 56 (42.4) 8 (26.7) 3158 (32.2)
 3 34 (25.8) 6 (20.0) 2980 (30.4)
 4 9 (6.8) 3 (10.0) 743 (7.6)
Open in a new tab

Exposure to any of the 59 carcinogenic pesticides during pregnancy resulted in elevated odds for ALL (OR: 2.83, 95% CI: 1.67-4.82) and AML (OR: 3.75, 95% CI: 0.97-11.57). Table 2 and 3 provide the exposure distributions of cases and controls based on specific pesticide exposures during pregnancy, presented by pesticide class. We estimated elevated odds for ALL with exposures to the following classes in our single-pesticide models: 2,6-dinitroanilines (OR: 2.50, 95% CI: 1.56-3.99), anilides (OR: 2.16, 95% CI: 1.38-3.36), and ureas (OR: 2.18, 95% CI: 1.42-3.34). Point estimates for AML were elevated for organophosphates (OR: 1.82, 95% CI: 0.70-4.74) and ureas (OR: 3.38, 95% CI 1.22-9.38) but small sample sizes resulted in wide confidence intervals.

Table 2.

Associations between ALL and exposure to agricultural pesticide applications within 4000m of the residential address, by pesticide class

Chemical Class Exposed Cases N=132 Exposed Controls N=9,805 Crude ORa Single-pesticide class modelb HLM OR (95% CI)c
2,6-Dinitroaniline 108 6,343 2.51 2.50 (1.56, 3.99) 1.76 (0.91, 3.38)
Amide 36 2,134 1.37 1.23 (0.83, 1.83) 1.09 (0.72, 1.66)
Anilide 52 3,204 2.23 2.16 (1.38, 3.36) 1.62 (1.02, 2.56)
Azole 85 5,037 1.73 1.69 (1.16, 2.45) 1.25 (0.76, 2.05)
Chloroacetanilide 55 3,119 1.54 1.47 (1.03, 2.11) 1.27 (0.82, 1.99)
Dicarboximide 77 5,127 1.25 1.15 (0.79, 1.67) 0.54 (0.33, 0.89)
Dithiocarbamate 97 5,799 1.91 1.83 (1.21, 2.75) 1.10 (0.61, 1.97)
N-Methyl Carbamate 69 4,166 1.40 1.30 (0.90, 1.87) 0.93 (0.61, 1.42)
Organochlorine 60 3,435 1.40 1.35 (0.93, 1.97) 0.98 (0.63, 1.52)
Organophosphorus 110 6,753 2.26 2.38 (1.48, 3.83) 1.54 (0.81, 2.93)
Pyrethroid 85 5,228 1.62 1.51 (1.03, 2.23) 0.89 (0.51, 1.54)
Substituted-Benzene 72 4,623 1.35 1.21 (0.84, 1.75) 0.81 (0.51, 1.28)
Sulfonylurea 13 794 1.21 1.07 (0.59, 1.94) 0.99 (0.55, 1.77)
Triazine 83 5,018 1.61 0.96 (0.60, 1.54) 0.77 (0.47, 1.25)
Urea 100 5,741 2.22 2.18 (1.42, 3.34) 1.43 (0.79, 2.58)
Open in a new tab
a

Adjusted for matching variable, birth year

b

Adjusted for birth year, SES-index variable, and mother’s race

c

Adjusted for birth year, mother’s race, SES-index variable, and using an overall pesticide effect in the hierarchical model

Table 3.

Associations between AML and exposure to agricultural pesticide applications within 4000m of the residential address, by pesticide class

Chemical Class Exposed Cases N=30 Exposed Controls N=9,805 Crude ORa Single-pesticide class modelb HLM OR (95% CI)c
2,6-Dinitroaniline 26 6,343 3.54 3.31 (1.10, 10.03) 1.63 (0.58, 4.58)
Anilide 14 3,204 2.18 2.00 (0.80, 4.98) 1.15 (0.49, 2.69)
Azole 21 5,037 2.21 1.99 (0.88, 4.53) 1.00 (0.43, 2.33)
Chloroacetanilide 13 3,119 1.64 1.50 (0.71, 3.17) 1.11 (0.52, 2.38)
Dicarboximide 22 5,127 2.52 2.31 (0.97, 5.51) 1.30 (0.51, 3.28)
Dithiocarbamate 23 5,799 2.27 2.08 (0.85, 5.10) 1.03 (0.40, 2.64)
N-Methyl Carbamate 17 4,166 1.82 1.62 (0.75, 3.51) 1.04 (0.49, 2.19)
Organochlorine 10 3,435 0.94 0.76 (0.33, 1.75) 0.50 (0.23, 1.08)
Organophosphorus 28 6,753 6.33 1.82 (0.70, 4.74) 0.76 (0.28, 2.05)
Pyrethroid 20 5,228 1.75 1.50 (0.65, 3.48) 0.75 (0.31, 1.79)
Substituted-Benzene 19 4,623 1.94 1.76 (0.79, 3.92) 1.16 (0.52, 2.61)
Triazine 22 5,018 2.63 1.24 (0.50, 3.10) 1.05 (0.46, 2.39)
Urea 25 5,741 3.54 3.38 (1.22, 9.38) 1.89 (0.70, 5.11)
Open in a new tab
a

Adjusted for matching variable, birth year

b

Adjusted for birth year, SES-index variable, mother’s race, and exposure to any other pesticide class

c

Adjusted for birth year, mother’s race, SES-index variable, and using an overall pesticide effect in the hierarchical model

In Table 4 we present ORs for individual pesticides. We estimated elevated odds ratios for ALL with exposure to diuron (OR: 2.38, 95% CI: 1.57-3.60), phosmet (OR: 2.10, 95% CI: 1.46-3.02), kresoxim-methyl (OR: 1.77, 95% CI: 1.14-2.75), and propanil (OR: 2.58, 95% CI: 1.44-4.63) in the single-pesticide model. We also observed unexpectedly increased odds ratios for glyphosate (OR: 2.20, 95% CI: 1.33, 3.63) and paraquat dichloride (OR: 1.74, 95% CI: 1.16, 2.62). Using hierarchical regression models instead these effects estimates attenuated slightly except for phosmet (OR: 2.10, 95% CI: 1.3-3.39). The addition of all other pesticides in the hierarchical logistic models attenuated the effects of most pesticides, likely due in part to co-adjustment for the other pesticides in the list, along with drawing each pesticide to the overall mean of the carcinogenic pesticides.

Table 4.

Associations between ALL and exposure to agricultural pesticide applications within 4000m of the residential address, by individual pesticide

Pesticide Class Pesticide Exposed Cases N=132 Exposed Controls N=9,805 Crude ORa Single-pesticide modelb HLM OR (95% CI)c
2,6-Dinitroaniline
Pendimethalin 89 5,223 1.87 1.82 (1.22, 2.71) 1.32 (0.74, 2.33)
Trifluralin 81 4,931 1.57 1.44 (0.97, 2.13) 1.03 (0.58, 1.83)
Oryzalin 76 4,602 1.57 1.54 (1.08, 2.21) 0.98 (0.58, 1.65)
Ethalfluralin 24 1,123 1.66 1.51 (0.96, 2.38) 1.15 (0.66, 1.99)
Benefin 16 933 1.22 1.05 (0.61, 1.81) 0.99 (0.54, 1.83)
Amide
Propyzamide 22 1,177 1.45 1.25 (0.78, 2.01) 2.12 (1.06, 4.23)
Isoxaben 16 1,044 1.20 1.11 (0.65, 1.89) 1.02 (0.58, 1.78)
Anilide
Boscalid 43 2,947 1.89 1.81 (1.10, 2.97) 1.38 (0.78, 2.43)
Propanil 13 405 2.58 2.58 (1.44, 4.63) 2.21 (1.16, 4.22)
Azole
Propiconazole 66 3,818 1.60 1.56 (1.09, 2.24) 1.57 (0.92, 2.66)
Tebuconazole 55 3,484 1.34 1.24 (0.86, 1.80) 0.80 (0.48, 1.32)
Fenbuconazole 21 1,273 1.32 1.21 (0.75, 1.96) 1.01 (0.55, 1.84)
Triadimefon 11 753 0.90 0.86 (0.45, 1.63) 0.79 (0.41, 1.51)
Chloroacetanilide
S-Metolachlor 38 2,462 1.37 1.30 (0.87, 1.94) 1.19 (0.70, 2.00)
Metolachlor 23 1,061 1.49 1.45 (0.89, 2.35) 1.09 (0.62, 1.91)
Dicarboximide
Iprodione 75 5,104 1.19 1.08 (0.74, 1.56) 0.50 (0.29, 0.86)
Dithiocarbamate
Ziram 60 3,200 1.70 1.68 (1.17, 2.40) 1.09 (0.65, 1.83)
Maneb 52 3,490 1.16 1.03 (0.71, 1.48) 0.62 (0.38, 1.00)
Mancozeb 70 4,324 1.41 1.28 (0.89, 1.83) 0.90 (0.58, 1.40)
Metam-Sodium 39 2,208 1.37 1.21 (0.82, 1.79) 1.07 (0.67, 1.73)
Halogenated organic
Methyl Bromide† 54 3,106 1.42 1.33 (0.93, 1.92) N/A
1,3-Dichloropropene 48 2,905 1.39 1.26 (0.86, 1.82) 1.03 (0.62, 1.69)
N-Methyl Carbamate
Carbaryl 66 4,044 1.34 1.25 (0.86, 1.79) 0.91 (0.58, 1.42)
Thiodicarb 19 744 1.76 1.56 (0.93, 2.63) 1.65 (0.88, 3.09)
Organochlorine
Dicofol 60 3,345 1.47 1.41 (0.97, 2.06) 1.11 (0.68, 1.82)
Organophosphorus
Chlorpyrifos† 97 5,934 1.81 1.71 (1.12, 2.60) N/A
Dimethoate 82 4,982 1.56 1.44 (0.98, 2.12) 1.02 (0.61, 1.71)
Malathion 78 4,515 1.67 1.54 (1.06, 2.22) 1.16 (0.73, 1.85)
Diazinon† 76 4,731 1.38 1.27 (0.87, 1.83) N/A
Phosmet 76 3,757 2.09 2.10 (1.46, 3.02) 2.10 (1.30, 3.39)
Acephate 48 3,146 1.18 1.06 (0.73, 1.54) 0.82 (0.50, 1.33)
Methidathion 29 2,053 0.97 0.91 (0.58, 1.41) 0.62 (0.36, 1.07)
S,S,S-Tributyl Phosphorotrithioate 27 1,282 1.57 1.47 (0.94, 2.30) 1.59 (0.81, 3.12)
Pyrethroid
Permethrin 73 4,765 1.31 1.18 (0.81, 1.72) 0.65 (0.37, 1.14)
Bifenthrin 56 3,760 1.27 1.18 (0.82, 1.70) 0.82 (0.51, 1.34)
(S)-Cypermethrin 40 2,768 1.25 1.13 (0.76, 1.69) 1.05 (0.64, 1.74)
Substituted Benzene
Chlorothalonil 67 4,262 1.34 1.19 (0.83, 1.72) 0.99 (0.59, 1.66)
Dicloran 26 1,632 1.17 1.05 (0.67, 1.63) 0.96 (0.54, 1.68)
Triazine
Simazine† 72 4,340 1.50 1.49 (1.04, 2.14) N/A
Pymetrozine 14 980 1.17 1.08 (0.61, 1.91) 1.00 (0.52, 1.89)
Urea
Diuron 98 5,364 2.38 2.38 (1.57, 3.60) 1.79 (1.02, 3.15)
Linuron 22 1,209 1.42 1.25 (0.78, 2.00) 1.63 (0.85, 3.12)
Other classes
Glyphosate‡ 113 7,125 2.28 2.20 (1.33, 3.63) N/A
Oxyfluorfen 104 6,247 2.15 2.10 (1.35, 3.27) 1.43 (0.72, 2.81)
Paraquat Dichloride† 95 5,711 1.83 1.74 (1.16, 2.62) N/A
Propargite 68 4,219 1.39 1.32 (0.91, 1.90) 0.76 (0.41, 1.42)
Norflurazon 58 3,466 1.35 1.27 (0.88, 1.84) 0.84 (0.51, 1.38)
Thiophanate-Methyl 56 3,127 1.61 1.52 (1.06, 2.17) 1.34 (0.87, 2.05)
Captan 55 2,893 1.61 1.58 (1.10, 2.28) 1.52 (0.91, 2.56)
Clofentezine 38 2,173 1.40 1.33 (0.90, 1.96) 1.09 (0.65, 1.83)
Buprofezin 32 2,181 1.30 1.20 (0.78, 1.85) 1.01 (0.61, 1.67)
Kresoxim-Methyl 31 1,673 1.86 1.77 (1.14, 2.75) 1.47 (0.83, 2.60)
Hexythiazox 30 2,219 1.17 1.09 (0.71, 1.69) 0.74 (0.44, 1.26)
Pyrithiobac-Sodium 24 1,666 1.01 0.94 (0.59, 1.50) 0.61 (0.31, 1.18)
Metaldehyde 17 984 1.30 1.22 (0.73, 2.05) 1.39 (0.75, 2.56)
Pyraflufen-Ethyl 16 1,310 1.18 1.11 (0.62, 1.99) 0.97 (0.50, 1.87)
Chlorthal-Dimethyl 15 1,351 0.80 0.67 (0.39, 1.16) 0.41 (0.18, 0.93)
Pyrimethanil 15 1,290 1.18 1.11 (0.60, 2.06) 0.85 (0.43, 1.68)
Piperonyl Butoxide 15 911 1.15 1.08 (0.62, 1.87) 0.94 (0.51, 1.72)
Cacodylic Acid 14 647 1.35 1.30 (0.71, 2.36) 1.00 (0.48, 2.08)
Bromacil 13 884 1.09 1.02 (0.57, 1.83) 0.87 (0.44, 1.72)
Thiazopyr 12 546 1.58 1.47 (0.80, 2.70) 1.26 (0.68, 2.34)
Triflusulfuron-Methyl 12 772 1.13 0.97 (0.52, 1.78) 1.05 (0.55, 2.02)
Hydrogen Cyanamide 11 1,192 0.70 0.65 (0.35, 1.22) 0.53 (0.27, 1.02)
Spirodiclofen 10 787 1.36 1.28 (0.62, 2.62) 1.35 (0.63, 2.89)
Open in a new tab
†

Added based on previous literature or high usage in California

‡

Glyphosate was defined as the sum of the following chemicals: glyphosate, glyphosate (salt), glyphosate (diammonium salt), glyphosate (isopropylamine salt), glyphosate (potassium salt), and glyphosate (trimesium)

a

Adjusted for matching variable, birth year

b

Adjusted for birth year, mother’s race, and SES-index variable

c

Adjusted for birth year, mother’s race, SES-index variable, and using an overall pesticide effect in the hierarchical model

In relation to AML (Table 5), we saw elevated odds for metam-sodium (OR: 2.56, 95% CI: 1.19, 5.49) and paraquat dichloride (OR: 3.38, 95% CI: 1.23, 9.27) in single-pesticide models.

Table 5.

Associations between AML and exposure to agricultural pesticide applications within 4000m of the residential address, by individual pesticide

Pesticide Class Pesticide Exposed Cases N=30 Exposed Controls N=9,805 Crude ORa Single-pesticide modelb HLM OR (95% CI)c
2,6-Dinitroaniline
Pendimethalin 22 5,223 2.41 2.19 (0.90, 5.30) 1.31 (0.74, 2.33)
Oryzalin 21 4,602 2.63 2.39 (1.06, 5.38) 0.99 (0.55, 1.78)
Trifluralin 19 4,931 1.71 1.46 (0.63, 3.36) 2.23 (1.38, 3.60)
Anilide
Boscalid 13 2,947 2.27 2.00 (0.76, 5.30) 0.82 (0.5, 1.34)
Azole
Propiconazole 14 3,818 1.37 1.18 (0.56, 2.52) 1.66 (0.97, 2.84)
Tebuconazole 14 3,484 1.58 1.37 (0.64, 2.94) 1.09 (0.65, 1.83)
Chloroacetanilide
S-Metolachlor 11 2,462 1.74 1.58 (0.72, 3.46) 0.78 (0.48, 1.28)
Dicarboximide
Iprodione 22 5,104 2.55 2.32 (0.97, 5.52) 1.02 (0.61, 1.71)
Dithiocarbamate
Mancozeb 18 4,324 1.91 1.75 (0.82, 3.77) 0.49 (0.28, 0.86)
Maneb 13 3,490 1.39 1.22 (0.57, 2.61) 1.32 (0.86, 2.04)
Metam-Sodium 13 2,208 2.69 2.56 (1.19, 5.49) 1.53 (0.91, 2.58)
Ziram 12 3,200 1.38 1.17 (0.55, 2.52) 0.80 (0.48, 1.33)
N-Methyl Carbamate
Carbaryl 17 4,044 1.91 1.71 (0.79, 3.69) 0.91 (0.58, 1.42)
Organochlorine
Dicofol 10 3,345 0.98 0.79 (0.34, 1.81) 1.44 (0.81, 2.56)
Organophosphorus
Chlorpyrifos† 25 5,934 3.26 3.05 (1.09, 8.52) N/A
Dimethoate 20 4,982 1.95 1.72 (0.75, 3.94) 1.12 (0.7, 1.79)
Diazinon† 18 4,731 1.64 1.43 (0.66, 3.13) N/A
Acephate 15 3,146 2.13 1.94 (0.92, 4.12) 0.92 (0.58, 1.44)
Malathion 15 4,515 1.17 0.99 (0.46, 2.12) 0.79 (0.42, 1.48)
Phosmet 11 3,757 0.94 0.79 (0.36, 1.71) 1.06 (0.64, 1.76)
Pyrethroid
Permethrin 18 4,765 1.59 1.33 (0.60, 2.96) 0.95 (0.56, 1.60)
Bifenthrin 14 3,760 1.40 1.20 (0.56, 2.58) 1.12 (0.68, 1.84)
(S)-Cypermethrin 14 2,768 2.29 2.09 (0.95, 4.59) 0.85 (0.51, 1.40)
Substituted Benzene
Chlorothalonil 18 4,262 1.95 1.79 (0.82, 3.89) 0.65 (0.37, 1.14)
Triazine
Simazine† 19 4,340 2.18 1.94 (0.89, 4.22) N/A
Urea
Diuron 22 5,364 2.28 2.02 (0.85, 4.84) 1.70 (0.96, 2.99)
Other classes
Glyphosate‡ 27 7,125 3.38 2.99 (0.88, 10.19) N/A
Paraquat Dichloride† 25 5,711 3.59 3.38 (1.23, 9.27) N/A
Oxyfluorfen 24 6,247 2.27 1.99 (0.76, 5.17) 1.39 (0.71, 2.75)
Propargite 15 4,219 1.33 1.07 (0.49, 2.34) 0.97 (0.58, 1.64)
Norflurazon 12 3,466 1.24 1.02 (0.47, 2.24) 0.60 (0.37, 0.98)
Thiophanate-Methyl 10 3,127 1.07 0.92 (0.42, 2.01) 1.09 (0.66, 1.80)
Captan 10 2,893 1.21 1.04 (0.47, 2.31) 1.09 (0.68, 1.77)
Open in a new tab
†

Added based on previous literature or high usage in California

‡

Glyphosate was defined as the following chemicals: glyphosate, glyphosate (salt), glyphosate (diammonium salt), glyphosate (isopropylamine salt), glyphosate (potassium salt), and glyphosate (trimesium)

a

Adjusted for matching variable, birth year

b

Adjusted for birth year, mother’s race, SES-index variable, and exposure to any other pesticide

c

Adjusted for birth year, mother’s race, SES-index variable, and using an overall pesticide effect in the hierarchical model

Effect estimates were similar across trimesters and comparable in size to the estimates for entire pregnancy exposures (Supplemental table 1). Exposure correlations between the first and second trimesters, and second and third trimesters were around 0.40-0.55 while the first and third trimesters were not correlated (r2 = −0.01-0.30). In the sensitivity analysis in which we relied on 2000m buffers to assess exposure, associations with the four pesticides mentioned above were slightly weaker for ALL but remained elevated with CIs including the null (ORs: 1.41-2.28) except for diuron, which was higher (OR: 2.28, 95% CI: 1.56-3.33). With regards to AML, paraquat dichloride remained elevated (OR: 2.59, 95% CI: 1.08-6.20), and metam-sodium had only 7 exposed cases at 2000m but remained elevated (OR: 2.17, 95% CI: 0.91-5.21). Results of sensitivity analyses statewide and among those exposed to at least one pesticide were similar among the pesticides with elevated effect estimates, although the magnitude of effect was generally closer to the null.

Discussion

In this rural population-based study of ALL and AML which was not subject to recall or participation biases, we observed an over two-fold increased odds for ALL and an over four-fold increased odds for AML with exposure to any carcinogenic pesticide. This risk estimate is higher than the estimates from meta-analyses for ALL (meta-OR: 1.74, 95% CI:1.37-2.21) (Van Maele-Fabry et al., 2011) and for AML, from a meta-analysis (meta-OR: 1.4) (Turner et al., 2010). The analyses for AML, however, were generally under-powered due to the small number of cases in less densely populated rural areas, which prevented us from estimating risks for one-third of the potentially carcinogenic pesticides. Many pesticides exhibited effect estimates above the null in single pollutant models. When comparing our results from the single-pesticide models to the multiple-pesticide adjusted model (with HLM), several specific pesticide associations remained elevated: diuron, phosmet, kresoxim-methyl, and propanil. The effects estimated for the classes of anilides and ureas can be explained by the presence of propanil and diuron in these classes, and these classes only contain two pesticides each for ALL. Additionally, propyzamide had no effect in the single-pesticide model, but showed an association in the hierarchical regression model.

While pesticides have previously been associated with increased risk of ALL, those results were almost exclusively based on studies that utilized ecological or case-control designs fraught with a potential for biased parental recall of pesticide exposure. Other studies that avoided retrospective self-reporting of exposures assessed exposures according to census block groups or acreage devoted to farming, exposure assessment methods that are relatively crude (Carozza et al., 2008; Reynolds et al., 2005; Reynolds et al., 2002; Ward et al., 2006). Our study is unique in that it utilizes specific address information and time of pregnancy to assess exposures of subjects in the state of California.

A similar case-control study in California assessing residential proximity to agricultural fields and childhood cancers in 1990-1997 found elevated risks in children of mothers exposed to metam-sodium (≥50th percentile) and dicofol (≥50th percentile) for all leukemias grouped together (Reynolds et al., 2005). Our results suggest the elevated risk these authors observed with metam-sodium may have been driven by the association we also observed for AML. We were not able to confirm this study’s finding of an increased risk of AML for dicofol as our point estimate was OR: 0.79, however, the ALL estimate was slightly elevated (OR: 1.41). Also, paraquat was found to increase risk of AML. However, while paraquat is currently not considered a carcinogen, the increased risk of AML might be linked to paraquat being an agent that causes oxidative stress and mitochondrial DNA damage (Sanders, 2017). These mechanisms may explain the increase in risk for AML seen here.

Of the pesticides we identified as being associated with ALL, diuron, a substituted phenylurea, is classified as a “known/likely” carcinogen according to the EPA. It is currently approved for use as an herbicide on terrestrial crops, for terrestrial non-crops (highways, pipelines, storage areas, etc.), and aquatic non-crops (irrigation, drainage ditches). Diuron has previously been linked to reproductive toxicity in rats (Fernandes et al., 2007), and shown to harm placental choriocarcinoma cells (Huovinen et al., 2015). Studies of human cell lines (breast adenocarcinoma and placental choriocarcinoma cells) showed that diuron is cytotoxic and potentially genotoxic; generation of reactive oxygen species (ROS) is a likely mechanism for its toxicity. To the best of our knowledge, no studies have examined the effects of diuron exposure on human cancers.

The second pesticide with strong and consistent associations to ALL, phosmet is an organophosphate (OP) insecticide, which have been shown to be able to cross the placental barrier and OP metabolites have been discovered in meconium and cord blood samples (Ostrea et al., 2009; Whyatt et al., 2003). A study testing phosmet exposures on a human choriocarcinoma cell line reported decreased cell viability, reduced DNA synthesis, and induction of inflammatory cytokines (Guinazu et al., 2012).

Propanil is an anilide classified as a suggestive carcinogen by EPA based upon studies in rats that found an increase in testicular interstitial cell adenomas after exposure (EPA, 2017). However, there is relatively little previous research in humans. Lastly, the strobilurin class fungicide kresoxim-methyl has been classified as a likely carcinogen by the EPA based on findings that it is highly toxic to several aquatic species such as daphnia magna and grass carp (Cui et al., 2017; Liu et al., 2013). It has also been seen as having neurotoxic effects in a study of cultured cortical neurons (Regueiro et al., 2015). This is the first human study to find a link between kresoxim-methyl and ALL.

Although we observed increases in risk of ALL and AML for those exposed to glyphosate and paraquat dichloride and these chemicals were chosen because they are widely used in California they were not included in the hierarchical models because they had not been classified in the 2011 document we originally used to select carcinogenic pesticides. However, more recently glyphosate has been classified as a type 2A carcinogen by IARC (IARC Monographs Program., 2017) and been implicated in an increased risk of AML in farmers who apply pesticides (Andreotti et al., 2018).

Pesticides become airborne during and after applications with ‘drift’ determined by factors such as application method (aircraft dispersion vs. hydraulic spraying vs. controlled droplet application), wind direction, wind speed and pesticide volatility (Richards et al., 2001; Richter et al., 1986; Weppner et al., 2006). Validation studies of PUR data, which utilized biomarkers or employed air monitoring to compare pesticide use reported in the PUR system with chemical concentrations, have reported that pesticides have been measured at up to 8,000m from the location where they are applied (Curwin et al., 2005; Harnly et al., 2005; Rull and Ritz, 2003; Wofford et al., 2014). One validation study comparing air samples collected by the California EPA Toxic Air Contaminant program and California Department of Pesticide Regulation records showed strong correlations between reported agricultural applications within a 3-mile radius and pesticide air concentrations measured up to 4 days later (Harnly et al., 2005). In addition to these validations of our exposure assessment, another strength of our study was that we were able to examine individual-level exposures at the residential address at birth.

Our study attempts to identify true confounders and adjust for them accordingly. However, with correlated exposures, it is difficult to identify and adjust for true confounders while avoiding over-adjustment (Momoli et al., 2010). In order to address this, we presented single-pesticide models which only adjust for likely confounders while excluding other pesticide exposures, while the semi-Bayesian approach accounts for co-exposures among carcinogens. Our results show most pesticides having effects greater than the null in the single-pesticide models while only a few survive the HLM. While HLM allowed us to adjust for other pesticide exposures, the HLM makes the unrealistic assumption that all pesticides examined are truly affecting leukemia risk in the same manner. At the current state of knowledge, it is however impossible to know which pesticide exposures are truly confounders and risk factors for the outcome or simply indicators of exposure to carcinogenic pesticides we should not be adjusting for. Thus, we prefer to present and interpret the ordinal logistic regression results as our main outcomes.

This study has some limitations common to studies of pregnancy exposures. Since fetuses would have had to survive to birth and into early childhood, it is possible that those exposed who would have later developed cancers may at the same time have been less likely to survive if high exposures to a pesticide results in miscarriage or spontaneous abortion, which would generate a live birth bias. That is, selection of healthier fetuses with less exposure would result in an attenuation of effects. Another limitation is that residential information was only available at birth and not throughout pregnancy or early childhood. Previous studies have estimated that between 11%-32% of pregnant women move at least once during pregnancy with median move distances ranging from 4.2 to 10 km (Lupo et al., 2010; Pennington et al., 2016; Pereira et al., 2016). With an expected low percentage of women moving, and those who do moving relatively small distances on average, our 4km exposure radius might have captured most residential ambient pesticide exposures from agricultural applications quite well. However, information on maternal occupational addresses and jobs that may also contribute to pesticide exposures and increased leukemia risk in these rural communities were not available (Bailey et al., 2014). Importantly, we did not have information whether parents were employed in agriculture and incurred occupational exposures possibly leading to a severe underestimation of total exposures. In addition, exposure to in-home or active occupational pesticide use was not accounted for and may result in an underestimation of total pesticide exposure. Nevertheless, living on a farm not only generates ambient exposures, which we captured with our GIS tool, but is also associated with occupational exposures, thus our residential exposure estimates for these rural communities may reflect an appropriate exposure ranking for the most heavily exposed subjects.

Although this study focuses on a highly exposed population, these cases represent only a fraction of childhood leukemia cases in California. Nevertheless, we hope to shed light on the possible link between ambient carcinogenic pesticide exposures and the risk of childhood leukemia.

Conclusion

In conclusion, our study results suggest that exposure to any carcinogenic pesticide exposure, or 2,6-dinitroanilines, anilides, and ureas classes of pesticides, specifically diuron, phosmet, kresoxim-methyl, and propanil increase the odds of ALL among those children exposed during pregnancy. Furthermore, exposure to metam-sodium and paraquat dichloride may increase the odds of AML. This study adds to the growing body of knowledge regarding prenatal pesticide exposures and childhood leukemias.

Supplementary Material

Supplementary material

Acknowledgements

This study was supported by grants from the US National Institutes of Health (R21ES019986, R21ES018960).

Dr. Ritz currently is an expert witness and consultant for law suits against Monsanto that pertain to glyphosate and NHL and also has been retained as a consultant for a law suits against Syngenta on paraquat exposure and Parkinson’s disease. Dr. Heck was retained to conduct a literature review on behalf of a lawsuit against the Monsanto Company.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

  1. Alderton LE, Spector LG, Blair CK, Roesler M, Olshan AF, Robison LL, Ross JA, 2006. Child and maternal household chemical exposure and the risk of acute leukemia in children with Down’s syndrome: a report from the Children’s Oncology Group. American journal of epidemiology 164, 212–221. [DOI] [PubMed] [Google Scholar]
  2. Andreotti G, Koutros S, Hofmann JN, Sandler DP, Lubin JH, Lynch CF, Lerro CC, De Roos AJ, Parks CG, Alavanja MC, Silverman DT, Beane Freeman LE, 2018. Glyphosate Use and Cancer Incidence in the Agricultural Health Study. J. Natl. Cancer Inst 110, 509–516. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Arcury TA, Grzywacz JG, Barr DB, Tapia J, Chen H, Quandt SA, 2007. Pesticide urinary metabolite levels of children in eastern North Carolina farmworker households. Environ. Health Perspect 115, 1254–1260. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Bailey HD, Fritschi L, Infante-Rivard C, Glass DC, Miligi L, Dockerty JD, Lightfoot T, Clavel J, Roman E, Spector LG, Kaatsch P, Metayer C, Magnani C, Milne E, Polychronopoulou S, Simpson J, Rudant J, Sidi V, Rondelli R, Orsi L, Kang AY, Petridou E, Schuz J, 2014. Parental occupational pesticide exposure and the risk of childhood leukemia in the offspring: findings from the Childhood Leukemia International Consortium. Int J Cancer 135, 2157–2172. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Bailey HD, Infante-Rivard C, Metayer C, Clavel J, Lightfoot T, Kaatsch P, Roman E, Magnani C, Spector LG, Th Petridou E, Milne E, Dockerty JD, Miligi L, Armstrong BK, Rudant J, Fritschi L, Simpson J, Zhang L, Rondelli R, Baka M, Orsi L, Moschovi M, Kang AY, Schuz J, 2015. Home pesticide exposures and risk of childhood leukemia: Findings from the Childhood Leukemia International Consortium. Int J Cancer 137, 2644–2663. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Bradman A, Eskenazi B, Barr DB, Bravo R, Castorina R, Chevrier J, Kogut K, Harnly ME, McKone TE, 2005. Organophosphate urinary metabolite levels during pregnancy and after delivery in women living in an agricultural community. Environ. Health Perspect 113, 1802–1807. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Bradman A, Whitaker D, Quiros L, Castorina R, Claus Henn B, Nishioka M, Morgan J, Barr DB, Harnly M, Brisbin JA, Sheldon LS, McKone TE, Eskenazi B, 2007. Pesticides and their metabolites in the homes and urine of farmworker children living in the Salinas Valley, CA. J Expo Sci Environ Epidemiol 17, 331–349. [DOI] [PubMed] [Google Scholar]
  8. Carozza SE, Li B, Elgethun K, Whitworth R, 2008. Risk of childhood cancers associated with residence in agriculturally intense areas in the United States. Environmental Health Perspectives 116, 559–565. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. CDC, 2009. United States Cancer Statistics (USCS). Centers for Disease Control and Prevention, Atlanta, GA. [Google Scholar]
  10. California Department of Pesticide Regulation, 2011. Summary of Pesticide Use Report Data - 2010. California Environmental Protection Agency, Sacramento, CA. [Google Scholar]
  11. Chow EJ, Puumala SE, Mueller BA, Carozza SE, Fox EE, Horel S, Johnson KJ, McLaughlin CC, Reynolds P, Von Behren J, Spector LG, 2010. Childhood cancer in relation to parental race and ethnicity: a 5-state pooled analysis. Cancer 116, 3045–3053. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Cockburn M, Mills P, Zhang X, Zadnick J, Goldberg D, Ritz B, 2011. Prostate cancer and ambient pesticide exposure in agriculturally intensive areas in California. Am. J. Epidemiol 173, 1280–1288. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Cui F, Chai T, Liu X, Wang C, 2017. Toxicity of three strobilurins (kresoxim-methyl, pyraclostrobin, and trifloxystrobin) on Daphnia magna. Environmental toxicology and chemistry 36, 182–189. [DOI] [PubMed] [Google Scholar]
  14. Curwin BD, Hein MJ, Sanderson WT, Barr DB, Heederik D, Reynolds SJ, Ward EM, Alavanja MC, 2005. Urinary and hand wipe pesticide levels among farmers and nonfarmers in Iowa. J Expo Anal Environ Epidemiol 15, 500–508. [DOI] [PubMed] [Google Scholar]
  15. Curwin BD, Hein MJ, Sanderson WT, Striley C, Heederik D, Kromhout H, Reynolds SJ, Alavanja MC, 2007. Urinary pesticide concentrations among children, mothers and fathers living in farm and non-farm households in iowa. Ann. Occup. Hyg 51, 53–65. [DOI] [PubMed] [Google Scholar]
  16. EPA, 2017. Evaluating Pesticides for Carcinogenic Potential. US Environmental Protection Agency, https://www.epa.gov/pesticide-science-and-assessing-pesticide-risks/evaluating-pesticides-carcinogenic-potential. Accessed March 28, 2017. [Google Scholar]
  17. Fenske RA, Lu C, Curl CL, Shirai JH, Kissel JC, 2005. Biologic monitoring to characterize organophosphorus pesticide exposure among children and workers: an analysis of recent studies in Washington State. Environ. Health Perspect 113, 1651–1657. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Fernandes GS, Arena AC, Fernandez CD, Mercadante A, Barbisan LF, Kempinas WG, 2007. Reproductive effects in male rats exposed to diuron. Reprod. Toxicol 23, 106–112. [DOI] [PubMed] [Google Scholar]
  19. Goldberg DW, Wilson JP, Knoblock CA, Ritz B, Cockburn MG, 2008. An effective and efficient approach for manually improving geocoded data. Int J Health Geogr 7, 60. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Guinazu N, Rena V, Genti-Raimondi S, Rivero V, Magnarelli G, 2012. Effects of the organophosphate insecticides phosmet and chlorpyrifos on trophoblast JEG-3 cell death, proliferation and inflammatory molecule production. Toxicology in vitro : an international journal published in association with BIBRA 26, 406–413. [DOI] [PubMed] [Google Scholar]
  21. Harnly M, McLaughlin R, Bradman A, Anderson M, Gunier R, 2005. Correlating agricultural use of organophosphates with outdoor air concentrations: a particular concern for children. Environmental Health Perspectives 113, 1184–1189. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Heck JE, Wu J, Lombardi C, Qiu J, Meyers TJ, Wilhelm M, Cockburn M, Ritz B, 2013. Childhood cancer and traffic-related air pollution exposure in pregnancy and early life. Environ. Health Perspect 121, 1385–1391. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Howlader N, Noone AM, Krapcho M, Garshell J, Neyman N, Altekruse SF, Kosary CL, Yu M, Ruhl J, Tatalovich Z, Cho H, Mariotto A, Lewis DR, Chen HS, Feuer EJ, Cronin KA, 2013a. SEER Cancer Statistics Review, 1975-2010, Section 29. National Cancer Institute, Bethesda, MD. [Google Scholar]
  24. Huovinen M, Loikkanen J, Naarala J, Vahakangas K, 2015. Toxicity of diuron in human cancer cells. Toxicology in vitro : an international journal published in association with BIBRA 29, 1577–1586. [DOI] [PubMed] [Google Scholar]
  25. IARC Monographs Program., 2017. Some organophosphate insecticides and herbicides Monographs Vol 112 International Agency for Research on Cancer (IARC) Press: Lyon, France. [PubMed] [Google Scholar]
  26. Infante-Rivard C, Sinnett D, 1999. Preconceptional paternal exposure to pesticides and increased risk of childhood leukaemia. Lancet 354, 1819. [DOI] [PubMed] [Google Scholar]
  27. Infante-Rivard C, Weichenthal S, 2007. Pesticides and childhood cancer: an update of Zahm and Ward’s 1998 review. Journal of Toxicology and Environmental Health. Part B, Critical Reviews 10, 81–99. [DOI] [PubMed] [Google Scholar]
  28. Johnson KJ, Carozza SE, Chow EJ, Fox EE, Horel S, McLaughlin CC, Mueller BA, Puumala SE, Reynolds P, Von Behren J, Spector LG, 2009. Parental age and risk of childhood cancer: a pooled analysis. Epidemiology 20, 475–483. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Jurewicz J, Hanke W, 2006. Exposure to pesticides and childhood cancer risk: has there been any progress in epidemiological studies? Int J Occup Med Environ Health 19, 152–169. [DOI] [PubMed] [Google Scholar]
  30. Kegley SE, Hill BR, Orme S, Choi AH, 2014. PAN Pesticide Database. Pesticide Action Network, Oakland. [Google Scholar]
  31. Liu L, Jiang C, Wu ZQ, Gong YX, Wang GX, 2013. Toxic effects of three strobilurins (trifloxystrobin, azoxystrobin and kresoxim-methyl) on mRNA expression and antioxidant enzymes in grass carp (Ctenopharyngodon idella) juveniles. Ecotoxicol. Environ. Saf 98, 297–302. [DOI] [PubMed] [Google Scholar]
  32. Loewenherz C, Fenske RA, Simcox NJ, Bellamy G, Kalman D, 1997. Biological monitoring of organophosphorus pesticide exposure among children of agricultural workers in central Washington State. Environ. Health Perspect 105, 1344–1353. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Lowengart RA, Peters JM, Cicioni C, Buckley J, Bernstein L, Preston-Martin S, Rappaport E, 1987. Childhood leukemia and parents’ occupational and home exposures. Journal of the National Cancer Institute 79, 39–46. [PubMed] [Google Scholar]
  34. Lu C, Fenske RA, Simcox NJ, Kalman D, 2000. Pesticide exposure of children in an agricultural community: evidence of household proximity to farmland and take home exposure pathways. Environmental Research 84, 290–302. [DOI] [PubMed] [Google Scholar]
  35. Lupo PJ, Symanski E, Chan W, Mitchell LE, Waller DK, Canfield MA, Langlois PH, 2010. Differences in exposure assignment between conception and delivery: the impact of maternal mobility. Paediatr Perinat Epidemiol 24, 200–208. [DOI] [PubMed] [Google Scholar]
  36. Metayer C, Colt JS, Buffler PA, Reed HD, Selvin S, Crouse V, Ward MH, 2013. Exposure to herbicides in house dust and risk of childhood acute lymphoblastic leukemia. J Expo Sci Environ Epidemiol 23, 363–370. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Momoli F, Abrahamowicz M, Parent ME, Krewski D, Siemiatycki J, 2010. Analysis of multiple exposures: an empirical comparison of results from conventional and semi-bayes modeling strategies. Epidemiology 21, 144–151. [DOI] [PubMed] [Google Scholar]
  38. Morrill R, Cromartie J, Hart G, 2013. Metropolitan, Urban, and Rural Commuting Areas: Toward a Better Depiction of the United States Settlement System. Urban Geography 20, 727–748. [Google Scholar]
  39. Nasterlack M, 2007. Pesticides and childhood cancer: an update. International journal of hygiene and environmental health 210, 645–657. [DOI] [PubMed] [Google Scholar]
  40. Oksuzyan S, Crespi CM, Cockburn M, Mezei G, Vergara X, Kheifets L, 2015. Race/ethnicity and the risk of childhood leukaemia: a case-control study in California. J. Epidemiol. Community Health 69, 795–802. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Ostrea EM Jr., Bielawski DM, Posecion NC Jr., Corrion M, Villanueva-Uy E, Bernardo RC, Jin Y, Janisse JJ, Ager JW, 2009. Combined analysis of prenatal (maternal hair and blood) and neonatal (infant hair, cord blood and meconium) matrices to detect fetal exposure to environmental pesticides. Environ. Res 109, 116–122. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Park AS, Ritz B, Ling C, Cockburn M, Heck JE, 2017. Exposure to ambient dichloromethane in pregnancy and infancy from industrial sources and childhood cancers in California. Int J Hyg Environ Health 220, 1133–1140. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Paul KC, Chuang YH, Cockburn M, Bronstein JM, Horvath S, Ritz B, 2018. Organophosphate pesticide exposure and differential genome-wide DNA methylation. Sci. Total Environ 645, 1135–1143. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Pennington AF, Strickland MJ, Klein M, Zhai X, Russell AG, Hansen C, Darrow LA, 2016. Measurement error in mobile source air pollution exposure estimates due to residential mobility during pregnancy. J Expo Sci Environ Epidemiol. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Pereira G, Bracken MB, Bell ML, 2016. Particulate air pollution, fetal growth and gestational length: The influence of residential mobility in pregnancy. Environ. Res 147, 269–274. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Regueiro J, Olguin N, Simal-Gandara J, Sunol C, 2015. Toxicity evaluation of new agricultural fungicides in primary cultured cortical neurons. Environ. Res 140, 37–44. [DOI] [PubMed] [Google Scholar]
  47. Reynolds P, Von Behren J, Gunier RB, Goldberg DE, Harnly M, Hertz A, 2005. Agricultural pesticide use and childhood cancer in California. Epidemiology 16, 93–100. [DOI] [PubMed] [Google Scholar]
  48. Reynolds P, Von Behren J, Gunier RB, Goldberg DE, Hertz A, Harnly ME, 2002. Childhood cancer and agricultural pesticide use: an ecologic study in California. Environmental Health Perspectives 110, 319–324. [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Richards SM, McClure GY, Lavy TL, Mattice JD, Keller RJ, Gandy J, 2001. Propanil (3,4-dichloropropionanilide) particulate concentrations within and near the residences of families living adjacent to aerially sprayed rice fields. Arch. Environ. Contam. Toxicol 41, 112–116. [DOI] [PubMed] [Google Scholar]
  50. Richter ED, Rosenvald Z, Kaspi L, Levy S, Gruener N, 1986. Sequential cholinesterase tests and symptoms for monitoring organophosphate absorption in field workers and in persons exposed to pesticide spray drift. Toxicol. Lett 33, 25–35. [DOI] [PubMed] [Google Scholar]
  51. Ritz B, Wilhelm M, Hoggatt KJ, Ghosh JK, 2007. Ambient air pollution and preterm birth in the environment and pregnancy outcomes study at the University of California, Los Angeles. American journal of epidemiology 166, 1045–1052. [DOI] [PubMed] [Google Scholar]
  52. Rull RP, Ritz B, 2003. Historical pesticide exposure in California using pesticide use reports and land-use surveys: an assessment of misclassification error and bias. Environmental Health Perspectives 111, 1582–1589. [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Sanders LH, Paul KC, Howlett EH, Lawal H, Boppana S, Bronstein JM, Ritz B, Greenamyre JT. Base excision repair variants and pesticide exposure increase Parkinson’s disease risk. Toxicological Sci 158, 188–198. [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. Simcox NJ, Fenske RA, Wolz SA, Lee IC, Kalman DA, 1995. Pesticides in household dust and soil: exposure pathways for children of agricultural families. Environmental Health Perspectives 103, 1126–1134. [DOI] [PMC free article] [PubMed] [Google Scholar]
  55. Tomasetto F, Tylianakis JM, Reale M, Wratten S, Goldson SL, 2017. Intensified agriculture favors evolved resistance to biological control. Proc Natl Acad Sci U S A. [DOI] [PMC free article] [PubMed] [Google Scholar]
  56. Turner MC, Wigle DT, Krewski D, 2010. Residential pesticides and childhood leukemia: a systematic review and meta-analysis. Environmental Health Perspectives 118, 33–41. [DOI] [PMC free article] [PubMed] [Google Scholar]
  57. Urayama KY, Von Behren J, Reynolds P, Hertz A, Does M, Buffler PA, 2009. Factors associated with residential mobility in children with leukemia: implications for assigning exposures. Annals of epidemiology 19, 834–840. [DOI] [PMC free article] [PubMed] [Google Scholar]
  58. Van Maele-Fabry G, Lantin AC, Hoet P, Lison D, 2011. Residential exposure to pesticides and childhood leukaemia: a systematic review and meta-analysis. Environ Int 37, 280–291. [DOI] [PubMed] [Google Scholar]
  59. Vinson F, Merhi M, Baldi I, Raynal H, Gamet-Payrastre L, 2011. Exposure to pesticides and risk of childhood cancer: a meta-analysis of recent epidemiological studies. Occupational and Environmental Medicine 68, 694–702. [DOI] [PubMed] [Google Scholar]
  60. Ward MH, Lubin J, Giglierano J, Colt JS, Wolter C, Bekiroglu N, Camann D, Hartge P, Nuckols JR, 2006. Proximity to crops and residential exposure to agricultural herbicides in iowa. Environmental Health Perspectives 114, 893–897. [DOI] [PMC free article] [PubMed] [Google Scholar]
  61. Weppner S, Elgethun K, Lu C, Hebert V, Yost MG, Fenske RA, 2006. The Washington aerial spray drift study: children’s exposure to methamidophos in an agricultural community following fixed-wing aircraft applications. J Expo Sci Environ Epidemiol 16, 387–396. [DOI] [PubMed] [Google Scholar]
  62. Whyatt RM, Barr DB, Camann DE, Kinney PL, Barr JR, Andrews HF, Hoepner LA, Garfinkel R, Hazi Y, Reyes A, Ramirez J, Cosme Y, Perera FP, 2003. Contemporary-use pesticides in personal air samples during pregnancy and blood samples at delivery among urban minority mothers and newborns. Environ. Health Perspect 111, 749–756. [DOI] [PMC free article] [PubMed] [Google Scholar]
  63. Wofford P, Segawa R, Schreider J, Federighi V, Neal R, Brattesani M, 2014. Community air monitoring for pesticides. Part 3: using health-based screening levels to evaluate results collected for a year. Environmental monitoring and assessment 186, 1355–1370. [DOI] [PubMed] [Google Scholar]
  64. Yost K, Perkins C, Cohen R, Morris C, Wright W, 2001. Socioeconomic status and breast cancer incidence in California for different race/ethnic groups. Cancer Causes Control 12, 703–711. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary material
NIHMS1564216-supplement-Supplementary_material.pdf (98.3KB, pdf)

RESOURCES