However, the mechanisms underlying environmental impacts on preterm delivery are still insufficiently understood and further experimental research is warranted. Pesticide exposures affected preterm birth in our study mostly in female children and to a lesser extent if at all males, similar to a Chinese study that found high levels of non-specific metabolites of organophosphate pesticides in maternal urine to have adversely affected duration of gestation only in girls . It has been suggested that exposures to pesticides in early pregnancy trigger more spontaneous abortions of male fetuses , or stillbirth in late pregnancy , outcomes not captured in our study. It is well known that the male fetus is more vulnerable in utero and is at greater risk of fetal death with the male-to-female ratio falling from around 120 male conceptions to 105 boys per 100 girls at birth . Some pesticides are endocrine disruptors such as those in the organophosphate family that mimic sex steroidal action and resemble estrogenic more than androgenic action in fish models . In general, we observed stronger ORs among infants born to Hispanic mothers partly because Hispanic mothers had higher exposure prevalence during pregnancy. According to a recent agricultural survey, about 90% of female farm workers in California were Mexico-born Hispanics ; thus, the foreign-born Hispanic mothers may live near fields where they work, making them more likely to be exposed to ambient pesticides when at home. Unfortunately,plastic flower pots information of specific occupations and occupational addresses of the mothers was not collected on birth certificates and therefore we could not determine exposures at workplaces.
Fetal growth restriction, the main reason for low birthweight other than preterm birth, has been associated with transplacental oxygen and nutrient transport, hypoxia, oxidative stress, placental inflammation, and inhibition of placental growth hormone ; these possible mechanisms may be influenced by toxic exposure to organophosphate and carbamate pesticides . We did not find much evidence for associations between term low birthweight and many specific pesticide exposures, in line with some previous studies. Others however, reported associations for low birthweight or a decrease in birthweight for some pesticides, including chlorpyrifos and/or diazinon, carbaryl, methyl bromide, as well as with organophosphate and pyrethroid metabolites measured in maternal urine . However, it also has been reported that when adjusting for gestational age associations with low birthweight were attenuated . Our results for pyrethroids are consistent with these previous observations for both preterm birth and low birthweight. Our term low birthweight results may have been under powered but still seem to corroborate a previous report that found residential proximity to methyl bromide use to reduce birthweight overall . Most previous pesticide and birth outcome studies examined exposures from home/garden or professional use of pesticides and relied on parental interviews after birth; these studies have been criticized for their potential selection or recall bias . Other studies using job exposure matrices may have been prone to non-differential exposure measurement errors, and often could not distinguish between types of chemicals. Smaller studies were able to employ biomarkers such as maternal blood or urine collected in pregnancy, or umbilical cord blood samples to measure prenatal chemical concentrations . The necessarily small size of such pregnancy cohorts limits the number of outcomes and hence study power considerably, and they also have to assume that chemical concentrations measure in bio-samples reflect exposures during multiple gestational windows accurately when many pesticides have relatively short half-lives, e.g., hours to a few days for organophosphates , and few studies have multiple bio-samples available throughout pregnancy.
Recently, several studies have examined the associations of ambient pesticide exposures and adverse birth outcomes in large populations based on proximity to applications modeling . These GIS-PUR based approaches applied to birth records avoids selection bias due to non-response and recall bias that threatens studies relying on interviews after births, in which mothers who had babies with adverse outcomes may be more likely to participate or recall their pesticide exposures. Similar to ours, these Californian studies were exclusively based on California’s PUR records and land use surveys. Particularly, two studies focused on the agriculturally dominated San Joaquin Valley; one assessed exposures by comparing high exposure to low exposure to pesticides with acute toxicity based on EPA signal word, while the other that reported negative associations with spontaneous preterm birth focused on exposures to frequently used chemicals and physicochemical groupings yet they mainly reported results by month in late pregnancy. Our sensitivity analysis stratified by season of conception supported our hypothesis that early but not late pregnancy is the critical period. Besides, different from previous studies, our method employed a four-tier mechanism to improve the match rate of PUR and land use maps , successfully reducing potential non-differential misclassification of exposure. We also expanded the study area to the California statewide to include more agricultural regions that might also be populous outside of the Central Valley while providing us with a large sample size and thus high statistical power in this record linkage-based design. Our study has some limitations. The ambiguous location, only at county level, of those non-agricultural pesticide applications in pesticide reporting, made it difficult to properly assess exposures for mothers living in urban areas at birth, whom were excluded to avoid substantial underestimation of exposure. However, our restriction to women living within 2km of fields might partially and indirectly ‘matched on’ location and generated a more homogeneous population in terms of potential geographically-specific confounders such as air pollution.
Similarly, since our study question is whether proximity to fields with agricultural pesticide applications increases risks of adverse birth outcomes, despite that other unassessed sources of pesticide exposure including occupational, home and garden use, or dietary exposures to pesticides could potentially confound our results, our ‘matching’ through restriction to women living within 2km of fields might have accounted for such factors. For example, the SUPERB study in northern California suggested that those who live near fields are more similar in their use of pesticides for other purposes, than residents in urban areas . Residents in the San Francisco metropolitan area, had a lower percentage of using outdoor pesticides than two inland areas in northern California they studied . Another assumption was that addresses at birth reflected the location of mothers over the entirety of pregnancy. A review of research on residential mobility during pregnancy showed that on average 24% of mothers move during pregnancy in the US ; although most moving distances were short , it may result in exposure misclassification in our GIS-based estimates based on a 2km buffer. Particularly, Hispanic mothers are more mobile than White mothers , increasing their chances of living close to the fields and receiving pesticide exposures during pregnancy. Similar to all other studies of live birth outcomes ours may also be subject to live birth bias, i.e., the fact that early exposures could lead to fetal loss. While data on the potential confounders maternal smoking and pre-pregnancy BMI, was only available for 4 out of 13 years of our study period, additional adjustment for these variables did not change our results more than minimally and suggests that they may not be confounders. In summary,plastic garden container this study found that first and second trimester exposures to most selected pesticides known or suspected to be reproductive toxicants were associated with preterm delivery but only one pesticide and perhaps pyrethroids as a class and were related to term low birthweight in California among women living near agricultural fields in California. These associations seemed stronger for female infants suggesting possible sex specificity for some of these agents.Epidemiologic studies of environmental exposures in early childhood, often assign exposures based upon the child’s or mother’s residence. While large-scale record-linkage based studies can avoid selection and recall bias that often impacts smaller studies with active subject recruitment, previous record-based studies often relied solely on maternal residential address at birth, which is readily available on many birth certificates and/or residential address at diagnosis, as done in some childhood cancer studies . The reliance on one address implicitly makes the assumption that a child’s residence remains the same throughout early childhood, or if they moved, that the exposure levels remained the same. Consequently, this may lead to exposure mis-classification for those who move in early childhood especially for exposures with high spatial heterogeneity. In a 2003-2007 California statewide representative survey, only 14% of all women moved in the 2-7 months post-partum , but with increasing age the frequency of residential moves also increased. For more than 50% of childhood cancer cases under age 5 diagnosed in California between 1988 and 2005, address at birth differed from the address at cancer diagnosis , which raises concerns about using residence at birth to assess exposures in early childhood. Exposure misclassification due to moving is a ubiquitous problem encountered by nearly all record-based studies that lack a complete residential history for each child. Previous studies suggested that residential mobility may be associated with certain risk factors for childhood cancers such as maternal age, marital status, parity, family income, and other socioeconomic status metrics , resulting in differential misclassification of exposures. While previous studies that examined residential proximity to exposure have mentioned the potential bias resulting from residential mobility during pregnancy , they rarely investigated the impact of residential mobility in early childhood on exposure measures or effect estimates.
While it is not feasible to acquire complete residential histories from interviews for subjects in large record-based studies as a gold standard to compare against the recorded birth or diagnosis address, databases containing public records of individuals collected by commercial companies have become available in recent years, allowing us to trace individuals without a self-reported residential history. For example, LexisNexis® Public Records , a commercial credit reporting company, provides all known addresses for a set of individuals upon request. Earlier validation studies have proven addresses acquired from LexisNexis to be useful for reconstructing residential histories for subjects in epidemiological studies with an overall match rate of ~70-85% with detailed address history obtained from interviews ; however, these subjects mostly consisted of mid-aged or older individuals, whose residential mobility may differ from that of women at child-bearing age. Such information, if of high quality, could potentially augment existing address information and help us to reconstruct residential histories for subjects in large record based studies and provide more accurate exposure estimates. The degree of exposure misclassification due to mobility depends on the distance moved, the spatial heterogeneity of the exposure , and the method of exposure assessment each study employed. For example, one study used ecological measures of agricultural activity at the county level , thus moving within a county would not alter exposure estimates. Other studies have assessed agricultural land use and crop coverage within a 1-km buffer of a child’s residence as proxies of pesticide applications or exposures to pesticides within a ½ mile buffer of child’s residence in relation to childhood cancers; such individual-level measures might be more sensitive to changes in location. Compared with these methods to estimate agricultural pesticide exposures near residences, our GIS-based system that integrates California’s unique Pesticide Use Reporting database and land use maps in California estimates children’s early life exposures at a finer resolution, but may be subject to more misclassification due to residential mobility. For the purpose of this study, we identify individuals’ exposures in early childhood using a 2-km buffer. The objectives of the present study are to assess patterns of mobility and identify maternal and child characteristics that may predict residential mobility in early childhood, and examine the impact of mobility on early childhood exposure measures for agriculturally applied pesticides and childhood cancers in California.To examine the associations between maternal and child characteristics and the cases’ likelihood of moving between birth and diagnosis, we conducted univariate and multivariate logistic regression analysis and estimated odds ratios and 95% confidence intervals . Based on previous literature , we considered factors that potentially influence mobility in pregnancy or early childhood including age at diagnosis , year of birth , maternal age at delivery , maternal race/ethnicity , maternal birthplace , maternal education , parity , rural/urban classification of residence at birth , and several socioeconomic variables including payment source for prenatal care as a proxy for family income and neighborhood level SES.