The objective of this study was to determine reliable measures of fetal exposure to environmental pesticides. A few studies have reported on the analysis of cord blood, maternal blood or meconium for pesticides: cord blood and maternal blood for chlorpyrifos, diazinon and propoxur (Whyatt et al., 2003
) and meconium for organophosphates (Whyatt and Barr, 2001
), DDE (Hong et al., 2002
), organochlorines (Ortega Garcia et al., 2006
) and other pesticides (Ostrea et al., 2002
; Bielawski et al., 2005
). However, this is the first study to simultaneously analyze and compare five matrices (maternal hair and blood, cord blood, infant hair and meconium) to determine the optimum matrix or combination of matrices to detect antenatal pesticide exposure.
For the infant matrices (cord blood, infant hair or meconium), meconium was the best matrix for this purpose. Of eleven pesticides analyzed, eight were detected in meconium with a high prevalence rate for propoxur (23.2%). In contrast, cord blood and infant hair were each only positive for a single pesticide, propoxur and chlorpyrifos, respectively. Furthermore, the concentrations of the pesticides detected were also significantly higher in meconium than cord blood and infant hair. Pesticides that partition and accumulate in adipose tissues such as organochlorines were poorly found in cord blood, infant hair or meconium. Although DDT was found in meconium, the frequency of detection was low (0.7%). The use therefore of cord blood, infant hair and meconium as matrices for the detection of fetal exposure to these compounds is a recognized limitation of the study. However, access to fetal or infant adipose tissue is normally not feasible; thus its diagnostic use in clinical settings may not be practical. On the other hand, failure to detect lipophilic pesticides from the analysis of non-adipose tissue matrices should not imply non-exposure to these types of pesticides. We did not include the analysis of infant’s urine due to inherent problems and difficulty associated with urine collection in infants. Besides, there are added limitations associated with the interpretation of urine results for pesticides particularly if only spot samples are collected (Barr and Needham, 2002
). Cord blood also had an advantage over urine since parent pesticides are more readily detectable in blood compared to urine (Barr and Needham, 2002
The high rate of detection of pesticides in meconium is consistent with the reported high rates of detection of most xenobiotics in meconium which include illicit drugs, licit drugs, nicotine metabolites and alcohol metabolites (Ostrea, 1999
). This is attributed to the repository nature of meconium thus providing a wide window of exposure to xenobiotics. Meconium is formed at around the third or fourth month of gestation and most xenobiotics that the fetus is exposed to during gestation are deposited in meconium, through fetal swallowing and/or bile secretion from that period up to the time of birth (Ostrea et al., 1989
). Since meconium, unlike fetal urine, is not normally excreted in utero, compounds that deposit in meconium accu mulate and increase in concentration thus enhancing their detection. In contrast, pesticides in cord blood represent acute exposure and may not be readily detected due to their low concentrations in the blood as a result of the metabolism, excretion and deposition in tissues of the pesticides. A highly sensitive technique to detect pesticides in blood has recently been published (Barr et al., 2002
) with LOD’s of three orders of magnitude lower than the LOD in this report. However, the specificity of the method was compromised since in many instances, only a single mass ion, often not the molecular ion, was used for compound identity. Thus, the methodology was associated with an imprecision that was abot double that of methods using higher detection limits (Barr et al., 2002
). In our study, we used stringent criteria for the identity of any compound, including (i) appropriate retention time in the chromatogram based on positive controls, (ii) the presence in the mass spectra of specific mass and qualifier ions and (iii) appropriate mass/qualifier ion ratios. Our adherence to these standard GC/MS criteria may have decreased the sensitivity of methodology, but retained the high specificity inherent to GC/MS analysis.
We did not detect any pesticides in infant hair except in one sample that was positive for chlorpyrifos. It appears that the deposition of pesticides in infant hair does not occur as readily compared to other compounds such as illicit drugs, nicotine, and most recently, fatty acid ethyl esters (Ostrea, 1999
; Koren et al., 2002
; Berkowitz et al., 2003
). The pharmacokinetics and tissue distribution of pesticides in the fetus is largely unknown. Fetal metabolism of pesticides is low due to the poor detoxification mechanisms (Waliszewski et al., 1998
). Furthermore, fetal hair starts to grow at approximately 6–7 months of gestation (Koren et al., 2002
) so that the timing of maternal exposure during pregnancy could also influence incorporation of pesticides into the growing hair shaft. It is also likely that due to the small amount of hair that could be collected from the newborn infant, the limited sample size for analysis prevented the detection of minute quantities of pesticides in infant’s hair. In contrast, our results with infant hair analysis differed markedly from the results with maternal hair due, in part, to more hair sample that could be obtained from the mother compared to newborn hair. Overall, newborn hair is not ideal for the analysis of pesticides because of low concentration of pesticides in infant hair and the limited amount of hair that could be collected for analysis. Furthermore, it has been reported that pesticide metabolites tend to partition predominantly towards blood rather than hair (Altshul et al., 2004
With regards to maternal matrices, the analysis of maternal hair showed significantly higher detection rates for pesticides compared to maternal blood. When combined with meconium analysis for pesticides, the analysis of maternal hair at midgestation and at birth increased the detection rate of prenatal exposure to pesticides by almost two fold. For propoxur, the increase was from 23.25% with meconium alone to 38.5% with meconium plus maternal hair A and B and for pyrethroids, from 2.8% for meconium alone to 16.7% with meconium plus maternal hair A and B. However, unlike fetal matrices (meconium, cord blood or infant hair) which when positive for pesticides are indicative of active fetal exposure to these compounds, this relationship is not necessarily valid for maternal hair. Inherent in maternal hair analysis is the difficulty in distinguishing between active and passive exposure to pesticides. In this study, we purposely did not wash the maternal hair prior to analysis for pesticides because we were interested in maternal exposure to the pesticide, regardless of whether it was active or passive. We have conducted some preliminary study on hair washing before analysis of the hair for propoxur and bioallethrin. There was no difference in the concentration of propoxur in the paired hair samples before and after washing (p = 0.175, Wilcoxon signed ranks test), but for bioallethrin, the concentration of the pesticide was significantly higher in the pre-washed compared to the post-washed hair samples (p = 0.001, Wilcoxon signed ranks test). However, the post wash concentration was undetectable in only one sample whereas all other samples were still positive after washing. The low concordance between the prevalence of propoxur and pyrethroids in maternal hair and meconium strongly suggests passive rather than active exposure in maternal hair. Nonetheless, there is ample justification and clinical use for the prenatal analysis of maternal hair for pesticides since important information on exposure can be provided (whether active or passive), which therefore provides an opportunity to initiate intervention measures during pregnancy that can reduce further exposure of the fetus to the pesticides.
Pesticide metabolites were rarely detected in the present study in any of the matrices analyzed despite a few publications that have reported on detecting pesticide metabolites in meconium, e.g., DDE (Hong et al., 2002
) and organophosphate metabolites (Whyatt and Barr, 2001
). Hong et al. (2002)
randomly sampled 60 meconium samples in Germany and detected DDE in 3 of them. However, the pesticide metabolite concentration they detected was 11.1 ng g−1
, which is lower than our LOD for DDE. Our method had a higher LOD since it was optimized to detect many classes of metabolites, especially the pyrethroids, whereas Hong and colleagues were selectively searching for DDE. However, we did find one meconium sample positive for DDE. Whyatt and Barr (2001)
found diethylthiophosphate (DETP), an organophosphate metabolite, in 100% of meconium samples studied in New York. We attempted to analyze for this compound using our current meconium liquid-liquid extraction method. However, it was discontinued due to difficulty in the chromatographic separation of DETP from TCP. The survey had reported higher use of malathion and chlorpyrifos, for which we had specific metabolites that we could accurately measure, than for diazinon, for which DETP would be a potential metabolite.
Overall, this study has shown that exposure to home, rather than farm, pesticides was the major source of pesticide exposure in the pregnant woman and her fetus even in an agricultural environment. This observation parallels reports of high exposure rate to home pesticides among pregnant women and their infants residing in urban areas (Whyatt et al., 2003
; Ostrea et al., 2002
). Thus, whether in the urban or rural areas, home pesticides constitute a high health risk in pregnant women and are likely related to the widespread and inappropriate use of pesticides at home. In our study, due to the widespread problems of pests at home, including flies, mosquitoes and roaches, spray pesticides were commonly used (38.2%), principally Baygon™ (91.5%) which contains propoxur and cyfluthrin. Inappropriate use of these home pesticides was evident since 39.9% of the spraying was done by the pregnant woman and reentry time to the sprayed area was ≤60min in 73.2% of the cases. Poor education and inadequate labeling on the safe use of the pesticide are major reasons for its improper use. The pesticide labels do not warn that the product should not be used by the pregnant woman, nor explicitly instruct on the appropriate reentry time of the sprayed area. Corrective measures to minimize further pesticide exposure in our study site have been instituted as a result of our findings. Assessment of clinical outcomes in the child in relation to prenatal and ongoing exposure to pesticides, are also under way.
In conclusion, our study has demonstrated that compared to cord blood or infant hair, meconium is the most sensitive matrix to analyze for fetal exposure to pesticides. The accumulation of pesticides in meconium, the ease of meconium collection and the large amount of sample that could be obtained for analysis are all factors that contribute to the increased sensitivity of this matrix. The prenatal analysis of maternal hair significantly adds to the detection rate of fetal exposure to pesticides and also provides the important advantage of initiating intervention measures during pregnancy that will reduce further fetal exposure to these potent neurotoxicants.