Data for this analysis were collected as part of Casa y Campo, a 4-year community-based participatory environmental justice project in which environmental health scientists, health care providers, and farmworker advocates collaborated to reduce pesticide exposure among farmworkers and their families. Casa y Campo was implemented in a six-county area of eastern North Carolina, including Duplin, Harnett, Johnston, Sampson, Wake, and Wayne counties. For 2004, the North Carolina Employment Security Commission (Raleigh, NC) estimated that 21,614 migrant and seasonal farmworkers (not counting dependents) worked in these counties during peak harvest, accounting for one-quarter of the 86,040 migrant and seasonal farmworkers in the state. Agriculture in these six counties is diverse (Table 1). Most of the farms in these counties producing tobacco (1,329 farms), sweet potatoes (188 farms), and vegetables (435 farms) would employ migrant and seasonal farmworkers. All of these farms, as well as the large number of farms producing grains, soybeans, and cotton, would use pesticides to which farmworkers and their families could be exposed. Results describing the dialkylphosphate urinary metabolites of OP pesticides for participants in this analysis have been published previously (Arcury et al. 2006).

Urine samples from most study children living in farmworker homes have a variety of pesticide metabolites. At least one specific pesticide metabolite was found in the urine samples for 59 of these 60 children; the urine samples for 88.3% of these children had three or more specific pesticide metabolites present. Analysis of the general dialkylphosphate OP pesticide metabolites for these same children found at least one of these six metabolites in every sample (Arcury et al. 2006).

The specific pesticide metabolites present indicate multiple sources and pathways of potential exposure among children in farm-worker homes, and illustrate the length of time these pesticides remain in the environments of farmworker children. The pesticide metabolites present in these children’s urine indicate at least four pathways of exposure: a) a paraoccupational take-home pathway in which workers bring pesticides into their homes on their person or on their clothing; b) an environmental pathway in which pesticides applied to nearby fields drift into the residential environment; c) a residential pathway of pesticides applied in the home; and d) a residual pathway in which pesticides deposited inside the home from any of the pathways at an earlier time remain active. Earlier research in North Carolina showed that farmworkers had little knowledge of the pesticides used where they work, and that they did not have information about pesticides applied to dwellings that were rented or grower-provided (Arcury et al. 2001; Early et al. 2006). Drift has been shown to result in the dispersal of agricultural pesticides to dwellings on surrounding land (Fenske et al. 2000b; Ward et al. 2006; Weppner et al. 2006). Finally, agricultural communities have generally higher levels of environmental pesticides compared with nonagricultural communities, with residues from discontinued pesticides found in contemporary rural environments and homes (Lu et al. 2004; Quandt et al. 2004; Wolz et al. 2003). However, the possible sources and pathways of exposure for several pesticide metabolites found in the samples from these children remain unclear.

In discussing the sources and pathways of potential exposure, it is important to remember that the urine samples used in this analysis were collected in June and July 2004. Exposure of these children to pesticides in Mexico should be discounted as the source of the metabolites; these are the children of seasonal rather than migrant farmworkers, who have been settled in North Carolina for several years. Many of these children were born in the United States, and have limited direct or indirect contact with sources of exposure from Mexico.

Chlorpyrifos and diazinon have been used for residential as well as agricultural applications. However, chlorpyrifos was banned for residential use by the end of 2001, and diazinon was banned for residential use at the end of 2004. Fifty of the 60 children had TCPy, the chlorpyrifos metabolite, in their urine. This would indicate that chlorpyrifos exposure for these children did not result from recent residential application, but could result from the take-home and drift pathways as well as from residual deposition. The diazinon metabolite IMPY was present in 33 of the 60 samples. Because it was still available for residential use at the time these data were collected, it is possible that diazinon exposure could result from any of the pathways.

Malathion has outdoor residential use, whereas parathion has no residential use. Yet 54 of the 60 children had the parathion metabolite PNP and 17 had the malathion metabolite MDA. Farmworkers may be exposed to malathion at work (e.g., picking peppers), but it is unlikely that they would work in fields to which parathion has been applied (e.g., cotton). Therefore, exposure resulting in the presence of PNP is most likely the result of the drift and residual deposition pathways, whereas MDA could result from the take-home, drift, or residual pathways.

Pyrethroid insecticides are widely used for residential and agricultural applications. The 24 children with 3PBA, the pyrethorid metabolite, in their urine sample could have been exposed through any of the pathways. The herbicides 2,4-D and acetochor are widely used for residential and agricultural applications, and their metabolites could result from exposure through any of the pathways.

Atrazine is used on corn (85% of all use), sorghum (10%), and sugar cane (3%). Three-quarters of all corn has atrazine applied. Corn and sorghum are grown in North Carolina. However, field corn and sorghum are machine cultivated and harvested. Sweet corn is hand picked, but few farmworkers are employed for this task. The most plausible explanation for four farmworker children having the metabolite for atrazine in their urine can be limited to drift or contaminated water supplies.

The sources and pathways of exposure among farmworker children to the OP pesticides coumaphas, pirimiphos methyl, and isazophos, and to the herbicide 2,4,5-T are not apparent. Coumaphas is used to control insects on livestock, and farmworkers in North Carolina seldom work with livestock. However, CMHC, the metabolite of coumaphas, was detected in samples from seven of the children. Pirimiphos methyl is used postharvest to treat stored corn and sorghum grain, activities with limited farmworker participation. Yet DEAMPY, the metabolite of pirimiphos methyl, was detected in samples from three children. Isazophos was used to treat lawns and turf; use of isazophos was cancelled in North Carolina at the end of 1998, and cancelled by the U.S. Environmental Protection Agency (EPA) in mid-1999. Yet CIT, the metabolite of isazophos, was found in one child. Finally, the herbicide 2,4,5-T has been banned for use in the United States since 1986, years before most Latino farm-workers had immigrated to North Carolina. The metabolite for 2,4,5-T was detected in the sample of one child. The metabolites for isazophos and 2,4,5-T were found in samples from different children.

Few data exist with which our results can be compared. The 1999–2000 National Health and Nutrition Examination Survey (NHANES) data provide results for the proportion of detects and levels detected for the OP pesticide metabolites TCPy, IMPY, MDA, and PNP (Barr et al. 2005). Caution must be taken in the comparison of our results with those reported for the NHANES sample. Development is a major factor affecting pesticide metabolism; and our participants were 1–6 years of age, whereas the youngest age group included in the NHANES sample was 6–11 years of age (Table 5). For TCPy, the percent of children with pesticide metabolite detected in the entire 1999–2000 NHANES sample (91%), among the Mexican-American sample (87%), and among children 6 to 11 years of age (97%) was somewhat greater than for the children in our sample (83%), with the GMs, 50th percentile, and 95th percentile (creatinine unadjusted and adjusted) being lower for the entire and Mexican-American samples, but higher among children 6 to 11 years of age. The percent detects for IMPY and PNP was higher for our sample (55% and 90%, respectively) compared with the NHANES samples (total sample, 29%, 22%; Mexican-American sample, 24%, 34%; children 6 to 11 years of age sample, 26%, 26%). The 95th percentile of IMPY for our sample (3.94 μg/L unadjusted for creatinine, 3.01 μg/g adjusted for creatinine) approximates those for the NHANES samples. The 95th percentile for PNP is greater among the children in our sample (6.32 μg/L unadjusted for creatinine, 7.99 μg/g adjusted for creatinine) than for the entire NHANES sample (5.0, 4.2) and among children 6 to 11 years of age (4.2, 4.2). However, the 95th percentile is much higher in the NHANES Mexican-American sample (21, 17). The percent detects for MDA for the children in our sample (28%) is lower than the percent detects in the entire 1999–2000 NHANES sample (52%), among the Mexican-American sample (61%), and among children 6 to 11 years of age (56%). However, the 95th percentile (6.87 μg/L unadjusted for creatinine, 8.64 μg/g adjusted for creatinine) for the children in our sample is greater than that for the entire NHANES sample (1.6, 1.8), the NHANES Mexican-American sample (1.6, 1.7), and among children 6 to 11 years of age (2.8, 3.7). No comparison data are available for the pyrethroid insecticide metabolite 3PBA, DEET repellent, or any of the herbicide metabolites.

Fenske and colleagues (2002) reported on the presence and levels of two OP pesticide metabolites for children residing in Washington State agricultural households. Not all of the children lived in farmworker households. They found TCPy in urine samples from 24% (18 of 75) children, with level of 6.0 μg/L (SD = 17) among children residing within 200 feet of an agricultural field and 1.3 μg/L (SD = 4.9) for those living more than 200 feet from a field. This metabolite was present in 83.3% of our samples, and had a GM of 1.92 μg/L. They reported PNP in urine samples for 7% (5 of 75) of the children. The comparable mean PNP levels were 33 μg/L (SD = 210) and 0. This metabolite was present in 90.0% of our samples, and had a GM of 1.00 μg/L. These differences reflect regional differences in agricultural uses.

Earlier studies have not considered the number of different pesticide metabolites present in individual children. Combining doses from several pesticides in a single child may result in interactions beyond the effects of a single toxicant (Eskenazi et al. 1999). Among these 60 children, 54 had two or more pesticide metabolites, with nine children having six pesticide metabolites and three children having seven pesticide metabolites. There was no discernable pattern in the combinations of pesticide metabolites present in the samples. Boys versus girls, and those living in rented housing versus owned or other housing had a greater number of pesticide metabolites in bivariate analysis. This may reflect sex differences in behavior. More important, it reflects how control of a dwelling may decrease exposure. Children whose mothers worked part-time versus full-time or who did not work had a greater number of pesticide metabolites in the bivariate and multivariate analysis. The interpretation of this association is not clear. We conducted further analysis to examine whether mothers working part-time were more likely to be employed doing farm work; this was not the case. It is possible that mothers working part-time are less able then unemployed women to attend to household hygiene, whereas mothers working full-time could count on more help in home care. Research is needed that has greater precision in measuring of pesticide exposure risk factors (Arcury et al. 2006; Quandt et al. 2006).

The results of this study should be considered in light of its limitations. The cross-sectional design limits analysis to association rather than causation. The lack of environmental pesticide samples does not allow for consideration of the location of exposure. Data on residential pesticide application were not collected. The metabolites present in the urine may also include exposure to the less toxic environmental degradates of the target pesticides. The determination of whether a metabolite was detected in a child’s urine sample, and therefore the total number of metabolites detected in a child, may be limited by the LOD of current analytic procedures; therefore, the number of metabolites detected is a conservative estimate. Finally, the measurement of risk factors may lack precision. This study also has several strengths. It is based on a relatively large sample of children. It is one of the few studies of farmworker child exposure on the East Coast. Finally, it is one of only a few studies to examine a large number of specific pesticide metabolites for farmworker children.

Several implications for environmental and occupational health practice and research can be drawn from our results. Pesticide exposure pathways for farmworkers and their children are multiple and complex. Pathways include take-home, drift, residential application, and residues. The sources of exposure to several of the pesticide metabolites found in the urine samples are not clear. Farmworkers must be educated about the take-home pathway. Current U.S. EPA Worker Protection Standard training (U.S. EPA 1992), when workers receive it (Arcury et al. 2001; U.S. Government Accountability Office 2000), provides little information about pesticides that can be taken home on the clothes, boots, and skin of workers, and about how to reduce or eliminate the amount taken home. Regulations on pesticide application to reduce drift must be reviewed. Farmworker and all rural families must be educated about drift and how to reduce exposure. Farmworker housing regulations must be improved to reduce the need for residential pesticide application (Early et al. 2006; Housing Assistance Council 2001). Farmworkers need to be educated about alternatives to using pesticides, such as residential integrated pest management. Farmworker and other rural homes need to be deep cleaned to remove pesticides from indoor environments. Procedures for this deep cleaning need to be developed and tested (McCauley et al. 2006a).

Research is needed that includes more precise measures of exposure pathways. Laboratory techniques measuring pesticide metabolites in urine and other matrices as well as environmental monitoring have improved substantially (Barr et al. 2006; Hoppin et al. 2006). However, measurement of exposure pathways in epidemiologic research has not improved greatly. For example, questions about recent employment as a farmworker must be changed to questions about the amount of farm work conducted in the 3 days before samples for biomonitoring are collected.

Research on the health effects of concurrent exposure to multiple individual pesticides and classes of pesticides is also needed. Current knowledge of health effects of pesticides in general is limited, but documentation is beginning (McCauley et al. 2006b). Examining health effects in light of the interactions of several different pesticides will provide greater insight into the actual risks to those exposed.