Maternal Occupational Exposure to PAHs
We observed an association between estimated maternal occupational exposure to PAHs and gastroschisis in offspring. However, although case and control mothers < 20 years of age were more likely to be classified as exposed (13.2% and 12.5%, respectively) than were older case and control mothers (6.7% and 2.8%, respectively), the association was limited to women ≥ 20 years of age. Other factors have also been reported to be associated with gastroschisis in the children of older women but not younger women. For instance, results from the NBDPS and a case–control study in Utah both suggest that the association between direct maternal smoking is stronger in women ≥ 20 years of age than in younger women (Feldkamp et al. 2008; Werler et al. 2009). Assuming older smokers have smoked for more years, it has been suggested that longer duration of smoking might contribute to uterine vascular damage (Suzuki et al. 1980), which in turn can lead to the development of gastroschisis (Werler et al. 2009). This could also be the case for prolonged exposure to occupational PAHs, although long-term information on maternal occupation before conception was not available for this analysis. Alternatively, the underlying mechanisms that produce gastroschisis among young mothers may differ from those that produce gastroschisis in the children of older mothers. Lastly, the difference in the magnitude of the effect measure estimates between younger and older mothers could be attributable to between-job exposure variability because exposed case mothers ≥ 20 years of age were more likely to be cooks, whereas exposed case mothers < 20 years of age were more likely to be restaurant cashiers. Cooks are likely to have a greater intensity of exposure to PAHs because of close proximity to high-temperature cooking compared with cashiers in the same restaurant.
Because PAHs are lipophilic, they readily penetrate cellular membranes (including the placenta) (ATSDR 1995). During PAH metabolism, enzymatic activity can result in the formation of reactive intermediates that covalently bind to DNA, forming adducts. DNA adducts have been shown to result in a spectrum of cellular mutations that may be teratogenic (Wells et al. 2010). PAH–DNA adducts have been isolated not only in adult tissues but also in placental tissues, amniotic fluid, and umbilical cord blood (Arnould et al. 1997; Madhavan and Naidu 1995; Ravindra et al. 2001). Furthermore, there is some evidence that occupational PAH exposure is associated with PAH–DNA adduct formation. For instance, a study by Perera et al. (1994) demonstrated that foundry workers with low-level exposure to PAHs had detectable levels of PAH–DNA adducts; however, a review by Brandt and Watson (2003) indicated that associations between measured PAH exposure and PAH–DNA adducts is equivocal. PAHs have been shown to be developmental toxicants in animal models, causing a range of birth defects (Anwer and Mehrotra 1988; Barbieri et al. 1986; Farwell et al. 2006; Incardona et al. 2004; Shum et al. 1979; Wassenberg and Di Giulio 2004; Wassenberg et al. 2005). To our knowledge, there have been only two other human studies of PAHs and birth defects (both case–control studies in China assessing neural tube defects). In a study by Naufal et al. (2010), PAH concentrations measured in venous blood samples were significantly (p < 0.05) higher in case mothers compared with control mothers. In a study by Ren et al. (2011), which included part of the same population as the Naufal et al. (2010) study, placental concentrations of PAHs were significantly higher (p < 0.001) in case placentas than in controls.
Our findings must be considered in light of certain limitations. The main limitation is related to the occupational exposure assessment. Although our approach relied on expert industrial hygienist consensus, there is still a potential for misclassification when assigning exposure based on questionnaire responses about jobs held. In an attempt to limit bias due to exposure misclassification, we restricted our analysis to those mothers with jobs that were rated with the highest confidence in the exposure assessment and found our results were similar. Furthermore, our approach is superior to a strategy that relies solely on maternal self-report of PAH exposure, where knowledge of PAH exposure is likely to be limited (Olsson et al. 2010). Although the use of personal monitoring or biomarkers of exposure would be preferred, these data are typically unavailable in population-based studies of birth defects, because these outcomes, although clinically significant, are relatively rare (e.g., the prevalence of gastroschisis is ~ 5 per 10,000 births) (Benjamin et al. 2010) and often not assessed in the context of prospective cohort studies (Yoon et al. 2001). Another limitation related to the occupational exposure assessment is the lack of information on intensity and frequency of exposure, which limits inferences about between- and within-job exposure variability and precludes exposure–response analyses.
A limitation with this and other case–control studies is the potential for recall bias. Because occupational PAH exposure was based on expert assessment rather than self-report, this may be less of a problem for our study (Jackson et al. 2004; Rocheleau et al. 2011). Furthermore, the impact of recall bias appears to be minimal in the NBDPS for many important risk factors, such as maternal smoking (MacLehose et al. 2009). The absence of information on environmental sources of PAHs is also a potential limitation, but occupational exposures are generally higher than those found in the environment (Brandt and Watson 2003). Additionally, we evaluated potential confounding by direct and secondhand smoke and meat consumption, which are important sources of environmental PAHs (Boers et al. 2005; Hansen et al. 2008). Finally, although we controlled for many measured maternal factors, there is still potential confounding by unidentified factors (i.e., unmeasured factors that have not been established as risk factors for gastroschisis). For instance, because most exposed women worked in food preparation or restaurant-related occupations, there may be some factor related to these occupations that is associated with both gastroschisis and PAH exposure. However, we attempted to adjust for several factors that may be associated with employment in these occupations (e.g., maternal age, education).
Strengths of this study include the use of data from the NBDPS, the largest population-based case–control study exploring risk factors for birth defects, which has an extensive occupational PAH exposure assessment available for study participants from 1997 through 2002. As part of the NBDPS, we also had information on potentially important confounding factors such as maternal nutrition, prepregnancy BMI, and smoking. Additionally, the case classification undertaken by NBDPS clinical geneticists to exclude cases due to single gene disorders or chromosomal abnormalities or those that are part of a limb–body wall complex or amniotic band sequence, resulted in a more homogeneous gastroschisis case group. Specifically, the exclusion of cases with known causes (e.g., single gene disorders) reduces the potential for etiologic heterogeneity in studies of birth defects (Khoury et al. 1982a, 1982b).
Discussion
We observed an association between estimated maternal occupational exposure to PAHs and gastroschisis in offspring. However, although case and control mothers < 20 years of age were more likely to be classified as exposed (13.2% and 12.5%, respectively) than were older case and control mothers (6.7% and 2.8%, respectively), the association was limited to women ≥ 20 years of age. Other factors have also been reported to be associated with gastroschisis in the children of older women but not younger women. For instance, results from the NBDPS and a case–control study in Utah both suggest that the association between direct maternal smoking is stronger in women ≥ 20 years of age than in younger women (Feldkamp et al. 2008; Werler et al. 2009). Assuming older smokers have smoked for more years, it has been suggested that longer duration of smoking might contribute to uterine vascular damage (Suzuki et al. 1980), which in turn can lead to the development of gastroschisis (Werler et al. 2009). This could also be the case for prolonged exposure to occupational PAHs, although long-term information on maternal occupation before conception was not available for this analysis. Alternatively, the underlying mechanisms that produce gastroschisis among young mothers may differ from those that produce gastroschisis in the children of older mothers. Lastly, the difference in the magnitude of the effect measure estimates between younger and older mothers could be attributable to between-job exposure variability because exposed case mothers ≥ 20 years of age were more likely to be cooks, whereas exposed case mothers < 20 years of age were more likely to be restaurant cashiers. Cooks are likely to have a greater intensity of exposure to PAHs because of close proximity to high-temperature cooking compared with cashiers in the same restaurant.
Because PAHs are lipophilic, they readily penetrate cellular membranes (including the placenta) (ATSDR 1995). During PAH metabolism, enzymatic activity can result in the formation of reactive intermediates that covalently bind to DNA, forming adducts. DNA adducts have been shown to result in a spectrum of cellular mutations that may be teratogenic (Wells et al. 2010). PAH–DNA adducts have been isolated not only in adult tissues but also in placental tissues, amniotic fluid, and umbilical cord blood (Arnould et al. 1997; Madhavan and Naidu 1995; Ravindra et al. 2001). Furthermore, there is some evidence that occupational PAH exposure is associated with PAH–DNA adduct formation. For instance, a study by Perera et al. (1994) demonstrated that foundry workers with low-level exposure to PAHs had detectable levels of PAH–DNA adducts; however, a review by Brandt and Watson (2003) indicated that associations between measured PAH exposure and PAH–DNA adducts is equivocal. PAHs have been shown to be developmental toxicants in animal models, causing a range of birth defects (Anwer and Mehrotra 1988; Barbieri et al. 1986; Farwell et al. 2006; Incardona et al. 2004; Shum et al. 1979; Wassenberg and Di Giulio 2004; Wassenberg et al. 2005). To our knowledge, there have been only two other human studies of PAHs and birth defects (both case–control studies in China assessing neural tube defects). In a study by Naufal et al. (2010), PAH concentrations measured in venous blood samples were significantly (p < 0.05) higher in case mothers compared with control mothers. In a study by Ren et al. (2011), which included part of the same population as the Naufal et al. (2010) study, placental concentrations of PAHs were significantly higher (p < 0.001) in case placentas than in controls.
Our findings must be considered in light of certain limitations. The main limitation is related to the occupational exposure assessment. Although our approach relied on expert industrial hygienist consensus, there is still a potential for misclassification when assigning exposure based on questionnaire responses about jobs held. In an attempt to limit bias due to exposure misclassification, we restricted our analysis to those mothers with jobs that were rated with the highest confidence in the exposure assessment and found our results were similar. Furthermore, our approach is superior to a strategy that relies solely on maternal self-report of PAH exposure, where knowledge of PAH exposure is likely to be limited (Olsson et al. 2010). Although the use of personal monitoring or biomarkers of exposure would be preferred, these data are typically unavailable in population-based studies of birth defects, because these outcomes, although clinically significant, are relatively rare (e.g., the prevalence of gastroschisis is ~ 5 per 10,000 births) (Benjamin et al. 2010) and often not assessed in the context of prospective cohort studies (Yoon et al. 2001). Another limitation related to the occupational exposure assessment is the lack of information on intensity and frequency of exposure, which limits inferences about between- and within-job exposure variability and precludes exposure–response analyses.
A limitation with this and other case–control studies is the potential for recall bias. Because occupational PAH exposure was based on expert assessment rather than self-report, this may be less of a problem for our study (Jackson et al. 2004; Rocheleau et al. 2011). Furthermore, the impact of recall bias appears to be minimal in the NBDPS for many important risk factors, such as maternal smoking (MacLehose et al. 2009). The absence of information on environmental sources of PAHs is also a potential limitation, but occupational exposures are generally higher than those found in the environment (Brandt and Watson 2003). Additionally, we evaluated potential confounding by direct and secondhand smoke and meat consumption, which are important sources of environmental PAHs (Boers et al. 2005; Hansen et al. 2008). Finally, although we controlled for many measured maternal factors, there is still potential confounding by unidentified factors (i.e., unmeasured factors that have not been established as risk factors for gastroschisis). For instance, because most exposed women worked in food preparation or restaurant-related occupations, there may be some factor related to these occupations that is associated with both gastroschisis and PAH exposure. However, we attempted to adjust for several factors that may be associated with employment in these occupations (e.g., maternal age, education).
Strengths of this study include the use of data from the NBDPS, the largest population-based case–control study exploring risk factors for birth defects, which has an extensive occupational PAH exposure assessment available for study participants from 1997 through 2002. As part of the NBDPS, we also had information on potentially important confounding factors such as maternal nutrition, prepregnancy BMI, and smoking. Additionally, the case classification undertaken by NBDPS clinical geneticists to exclude cases due to single gene disorders or chromosomal abnormalities or those that are part of a limb–body wall complex or amniotic band sequence, resulted in a more homogeneous gastroschisis case group. Specifically, the exclusion of cases with known causes (e.g., single gene disorders) reduces the potential for etiologic heterogeneity in studies of birth defects (Khoury et al. 1982a, 1982b).