Comparison of Current-Use Pesticide and Other Toxicant Urinary Metabolite Levels among Pregnant Women in the CHAMACOS Cohort and NHANES

Background We measured 34 metabolites of current-use pesticides and other precursor compounds in urine samples collected twice during pregnancy from 538 women living in the Salinas Valley of California, a highly agricultural area (1999–2001). Precursors of these metabolites included fungicides, carbamate, organochlorine, organophosphorus (OP), and pyrethroid insecticides, and triazine and chloroacetanilide herbicides. We also measured ethylenethiourea, a metabolite of the ethylene-bisdithiocarbamate fungicides. Repeat measurements of the compounds presented here have not been reported in pregnant women previously. To understand the impact of the women’s regional environment on these findings, we compared metabolite concentrations from the CHAMACOS (Center for the Health Assessment of Mothers and Children of Salinas) cohort with U.S. national reference data for 342 pregnant women sampled by the National Health and Nutrition Examination Survey (1999–2002). Results The eight metabolites detected in > 50% of samples [2,4-dichlorophenol (2,4-DCP); 2,5-dichlorophenol (2,5-DCP); 1- and 2-naphthol; ortho-phenylphenol (ORTH); para-nitrophenol (PNP); 2,4,6-trichlorophenol (2,4,6-TCP); and 3,4,6-trichloro-2-pyridinol (TCPy)] may be related to home or agricultural pesticide use in the Salinas Valley, household products, and other sources of chlorinated phenols. More than 78% of women in this study had detectable levels of at least one of the OP pesticide-specific metabolites that we measured, and > 30% had two or more. The 95th percentile values of six of the most commonly detected (> 50%) compounds were significantly higher among the CHAMACOS women after controlling for age, race, socioeconomic status, and smoking [(2,4-DCP; 2,5-DCP; ORTH; PNP; 2,4,6-TCP; and TCPy); quantile regression p < 0.05]. Conclusions Findings suggest that the CHAMACOS cohort has an additional burden of precursor pesticide exposure compared with the national sample, possibly from living and/or working in an agricultural area.

volume 118 | number 6 | June 2010 • Environmental Health Perspectives Research Prenatal exposure to pesticides, including organophosphorus (OP) and organochlorine (OC) compounds, has been shown to increase the risk of adverse developmental outcomes in children (Eskenazi et al. 2008;Jurewicz and Hanke 2008;Perera et al. 2005). However, few studies have comprehensively investigated maternal exposures to currently used pesticides during pregnancy. Biomonitoring results from the National Health and Nutrition Examination Survey (NHANES) (1999)(2000)(2001)(2002) [Centers for Disease Control and Prevention (CDC) 2005] and other studies suggest that many pregnant women are exposed to a mixture of insecticidal and fungicidal compounds (Berkowitz et al. 2003;Bradman et al. 2003Bradman et al. , 2005Takser et al. 2004;Ye et al. 2008).
There are multiple pesticide exposure pathways, but recent research (Lu et al. 2006;McKone et al. 2007) suggests that diet is the dominant route of exposure for the general population. Para-occupational or take-home exposure from farmworkers and ambient exposure from living in proximity to agricultural activity can add to total pesticide exposure in agricultural communities (Goldman et al. 2004;McKone et al. 2007). Biomonitoring can help to quantify the additional exposure for pregnant women in agricultural communities by allowing for the direct comparison of detection frequencies and metabolite levels between populations.
The Center for the Health Assessment of Mothers and Children of Salinas (CHAMACOS) is a longitudinal birth cohort study investigating in utero and postnatal environmental exposures and their effects on the health of children residing in the largely agricultural Salinas Valley of Monterey County, California . The growing season in this productive region is nearly yearround. Row crops are grown with a predominance of strawberries, lettuce, and broccoli in the north and warmer-weather crops such as grapes in the south. Previously we presented CHAMACOS prenatal measure ments of three specific metabolites of OP pesticides ) and ethylenethiourea (ETU), a toxic metabolite of the ethylene bisdithiocarbamate (EBDC) fungicides (Montesano et al. 2007). In this study, we present exposure data on an additional 30 pesticide metabolites in repeat samples collected from pregnant women late in the first and second trimesters. To better understand the impact of the regional agricultural environment on these findings, we also compare levels in the CHAMACOS cohort with those measured in a representative national sample of pregnant women participating in the NHANES (1999)(2000)(2001)(2002) (CDC 2005).

Study location characteristics. The Salinas
Valley includes approximately 91,054 irrigated hectares devoted to row, vineyard, and orchard crops. In 2000, approximately 570,000 kg herbicide, EBDC fungicide, and carbamate, OP, OC, and pyrethroid pesticide active ingredients were applied for agricultural and landscaping purposes in this region, a level typical of recent years (10% herbicides; 27% EBDC fungicides; 44% OP and 2% OC pesticides; and 4% pyrethroid pesticides) [California Department of Pesticide Regulation (CDPR) 2000a, 2006]. Of the total kilograms of nonfumigant pesticides applied in the Salinas Valley (excluding sulfur and petroleum products), 65% percent devolve into one of the 34 urinary metabolites we measured. Applications of maneb, diazinon, malathion, acephate, and chlorpyrifos account for 78% of the kilograms of pesticides applied whose urinary metabolites were detectable using the analytic methods available in this study (Table 1).
Study population. CHAMACOS. Between September 1999 and November 2000, 601 pregnant women were enrolled in the CHAMACOS birth cohort study from six local prenatal clinics. Women were eligible to participate in this study if they were ≤ 20 weeks gestation at the time of enrollment, were ≥ 18 years of age, qualified to receive poverty-based government health insurance, and planned to continue receiving prenatal care at a participating health center . Written informed consent was obtained from all participants in accordance with procedures approved by the University of California Berkeley Committee for the Protection of Human Subjects.
NHANES 1999-2002 data set. As part of the ongoing NHANES, the CDC reported levels of current-use pesticide metabolites measured in spot urine collected from representative samples of the U.S. population stratified by age, sex, and racial/ethnic group CDC 2004). Metabolite concentrations were measured in 3,048 urine samples collected from U.S. residents 6-59 years of age during 1999 and 2002 and included 342 pregnant women 15-50 years of age (CDC 2004). The public release versions of the NHANES data sets, including demographic information and metabolite data, were used for these analyses (CDC 2007a(CDC , 2007b. We applied no sample weights to the NHANES data. Pesticide use reporting data. In California, all agricultural, right-of-way, and structural pesticides applied by farmers or licensed applicators must be reported to the state (CDPR 2000b). Agricultural pesticide applications, representing 90% of all reported use, are geocoded to 1-square-mile units based on the Public Land Survey System (PLSS). To summarize agricultural pesticide use in the Salinas Valley, we first defined the region as the area bounded by the Pacific Ocean and an isopleth at 200 feet in elevation above the Salinas River. We then identified the PLSS sections within the Salinas Valley and abstracted pesticide use from the California pesticide use reporting (PUR) data set.
a Kilograms active ingredient applied includes agricultural, landscape maintenance, structural pest control, and roadside pesticide usage (CDPR 2000a).
volume 118 | number 6 | June 2010 • Environmental Health Perspectives and occupation of CHAMACOS participants was collected through personal interviews. Interviews were conducted in English or Spanish by bilingual, bicultural study staff. The first prenatal interview occurred shortly after enrollment in the study (mean ± SD = 13 ± 5.2 weeks gestation) and a follow-up interview occurred late in the second trimester (mean = 26 ± 2.6 weeks gestation). Urine collection and analysis. Two spot urine samples were collected from CHAMACOS participants at each prenatal interview and analyzed by the Division of Laboratory Sciences, CDC. Women voided into a sterile urine cup in bathroom facilities at our field office or in a mobile clinic. Specimens were placed into precleaned glass containers with Teflon-lined caps, bar coded, and stored at -80°C until shipment. Samples were shipped on dry ice to the CDC and stored at -70°C until analysis.
Laboratory quality control (QC) included repeat analysis of three in-house urine pools enriched with known amounts of pesticide residues whose target values and confidence limits were previously determined (Westgard 2003). Detection limits ranged from 0.03 to 1.5 µg/L. We assigned an imputed value of the limit of detection (LOD)/√2 to levels below the detection limit (Barr et al. 1999;Hornung and Reed 1990). A total of 153 (blind) field QC samples were analyzed, representing 15% of total samples. Mean relative recoveries for the metabolites in field QC samples ranged from 78 to 120%. The mean levels of metabolites in 51 blank field samples were ≤ 1 µg/L. The creatinine concentration in each urine sample was reported in milligrams of creatinine per deciliter of urine. Of the 601 women who enrolled in the study, adequate urine samples with valid creatinine levels were collected from 538 (90%) women at the first prenatal sampling point and 481 (80%) women at the second.
Ratio of 1-naphthol to 2-naphthol. Both naphthalene and carbaryl can devolve to urinary 1-naphthol. Because naphthalene, but not carbaryl, also metabolizes to 2-naphthol, the ratio of urinary 1-naphthol to 2-naphthol within an individual can provide information about the source of 1-naphthol. Thus, we calculated the ratio of 1-naphthol to 2-naphthol to identify subgroups within the CHAMACOS cohort and pregnant women in NHANES in which carbaryl was expected to be the primary source of 1-naphthol. We assumed that a ratio > 2 indicates carbaryl as the source of 1-naphthol (Meeker et al. 2007). Data analysis. We calculated descriptive statistics for metabolite levels from the two CHAMACOS prenatal urine samples. For the eight CHAMACOS metabolites with detection frequencies > 50% in both samples, we computed the Spearman's rank correlation between the first and second prenatal samples. In addition, we computed metabolite cor relations within each of the two CHAMACOS samples and within samples collected from pregnant women in NHANES. To broadly compare the NHANES and CHAMACOS metabolite levels, we analyzed the 11 compounds with detection frequencies > 40% in either study. We used the Wilcoxon rank-sum test and quantile regression (95th percentile) to compare differences in metabolite concentrations between the CHAMACOS cohort and NHANES. We compared differences in detection frequencies using analysis of variance (ANOVA). For compounds with LODs that differed between the two populations (e.g., 1-and 2-naphthol), the higher LODs were applied to the metabolite concentrations of both populations for analyses of differences in detection frequencies. When comparing NHANES and CHAMACOS women's metabolite levels using quantile regression, we used demographic variables to adjust our models for age (continuous), current smoking (yes/no), ethnicity (Mexican or Mexican American; other Hispanic; non-Hispanic white; non-Hispanic black; or other race, including multiracial), and socioeconomic status (at or below poverty threshold; within 200% of poverty threshold; above 200% of poverty threshold) (Table 2). Finally, we calculated the total number of OP pesticide metabolites detected in the first and second prenatal urine samples of each CHAMACOS woman. We did not count PNP, because compounds other than OP pesticides (e.g., nitrobenzene) are additional sources of urinary PNP.
All analyses were conducted using Stata software, version 10 (StataCorp LP, College Station, TX). Table 2 presents the demographic characteristics of pregnant women in the CHAMACOS cohort and NHANES. Compared with pregnant women in NHANES, the CHAMACOS women were less educated (80% vs. 23% attained less than a 12th grade education), lived with more people in their households (74% vs. 29% live with ≥ 4 household members), and were poorer (62% vs. 25% living at or below the federal poverty threshold). Ninetyfive percent of CHAMACOS women were Mexican or Mexican American compared with 26% in NHANES. Forty-four percent of CHAMACOS women were employed as farm workers at some point during their pregnancy, and 81% percent shared a home with at least one agricultural worker ). In addition, the CHAMACOS women tended to have higher prepregnancy body mass indexes (BMIs) compared with NHANES (23% vs. 15% with BMI > 30, respectively), but fewer CHAMACOS women smoked (1.5%) compared with pregnant women in NHANES (7.4%).

) (CDC 2005).
Among compounds with > 40% detection, the CHAMACOS and NHANES detection frequencies were statistically different Table 3. Summary of eight CHAMACOS first and second prenatal and NHANES urinary metabolite concentrations (µg/L) with detection frequencies > 50%. Abbreviations: DF, detection frequency; Max, maximum. CHAMACOS prenatal sample 1 was collected at approximately 13 weeks gestation, and prenatal sample 2 was collected at approximately 26 weeks gestation; the NHANES prenatal samples were collected at approximately 24 weeks gestation (mean = 24.4 ± 9.5 weeks). a NHANES detection frequencies for 1-and 2-naphthol become 99% and 93%, respectively, when recalculated using the CHAMACOS LODs.
In the CHAMACOS cohort, the detection frequency of ETU was higher at the first (23.5%) compared with the second (7.7%) sampling time. In addition, the 95th percentile value of ETU was significantly higher in the first compared with the second prenatal sample (1.5 µg/L vs. 0.4 µg/L, respectively; p < 0.01).
In addition, there were statistically significant differences between the distributions of eight (2,5-DCP; MDA; 2-naphthol; PNP; 3PBA2; 2,4,5-TCP; 2,4,6-TCP; and TCPy) of the 11 frequently detected metabolites in the CHAMACOS second sample versus the NHANES metabolite levels (p < 0.05). Among these eight compounds, the distributions of six (2,5-DCP; MDA; PNP; 2,4,5-TCP; 2,4,6-TCP; and TCPy) of the second prenatal CHAMACOS metabolites were higher compared with those in NHANES, and the distributions of two NHANES metabolite levels (2-naphthol and 3PBA2) were higher compared with those in CHAMACOS. The NHANES prenatal samples were collected at approximately 24 weeks gestation, and the CHAMACOS first and second prenatal samples were collected at approximately 13 and 26 weeks gestation, respectively. Thus, the CHAMACOS second prenatal metabolite levels may be more comparable with those from pregnant women in NHANES. We compared median concentrations of the six compounds with detection frequencies > 50% in both populations using quantile regression and adjusting for age, race, socio economic status, and smoking. The CHAMACOS first prenatal median 2,4-DCP and 2,5-DCP concentrations were higher than NHANES (p < 0.05). Further, the CHAMACOS second prenatal median 2,5-DCP, 2,4,6-TCP, and TCPy concentrations were also higher than NHANES (p < 0.05). By contrast, the NHANES 2-naphthol median levels were significantly higher than the CHAMACOS first and second prenatal levels (p < 0.05) ( Table 3). Figure 1 presents a comparison of the 95th percentile metabolite values for the CHAMACOS cohort (both first and second prenatal samples) and NHANES for the 11 metabolites detected at > 40% in either study. Using quantile regression, the CHAMACOS first and/or second pre natal 95th percentile values were statistically higher compared with NHANES for seven of the 11 metabolites. These seven included six of the most commonly detected metabo lites (2,4-DCP; 2,5-DCP; ORTH; PNP; 2,4,6-TCP; and TCPy). [Probability plots examining the z-scores of CHAMACOS vs. NHANES TCPy, 2,4,5-TCP, and 2,5-DCP levels are presented in the Supplemental Material (doi:10.1289/ ehp.0901568).] We found no significant differences between the CHAMACOS first and second prenatal 95th percentile metabolite levels presented in Figure 1.
A metabolite of several commonly used pyrethroids (3PBA2), including permethrin, cypermethrin, cyhalothrin, deltamethrin, and fenvalerate, was detected in urine samples of 67% of NHANES pregnant women compared with 21% and 27% in the CHAMACOS first and second prenatal samples, respectively. In addition, the CHAMACOS 3PBA2 95th-percentile metabolite levels were lower than the NHANES level, albeit not significantly (Figure 1). We could not compare the CHAMACOS and NHANES metabolite levels of several high-use precursor pesticides (dimethoate, acephate, methamidophos, and maneb) because they have not yet been reported in NHANES. Figure 2 presents the cumulative distributions of urinary 1-naphthol to 2-naphthol ratios in the CHAMACOS cohort and NHANES. Ratios > 2 occurred in 22% and 24% of CHAMACOS first and second prenatal samples compared with 13% among pregnant women in NHANES. Conversely, the frequency of 1-naphthol to 2-naphthol ratios in the region near unity (i.e., ratios ranging from 0.75 to 1.25) were higher among women in NHANES (32%) compared with the CHAMACOS women at the first and second prenatal sampling times (19% and 20%, respectively) (Figure 2 inset).

Discussion
We measured 34 specific metabolites in two urine samples collected at approximately 13 and 26 weeks gestation from 538 pregnant