In this study we found that DDE had a small association with BMI SDS in 3-year-old children of nonsmoking mothers. However, this association was enhanced in children of smoking mothers. This relation seemed to be mediated by an influence on weight gain during the preschool period. Second, PCBs were also associated with higher BMI SDS values. This association was not modified by smoking of the mother, nor did it change with the age of the child. The association between PCBs and BMI SDS seemed to be mediated by a negative influence on birth weight and length and on length between 1 and 3 years.
Our results should be considered in light of some important study limitations. We found significant associations and interactions with maternal smoking ever, which includes mothers who smoked before pregnancy or during pregnancy, but not with smoking during pregnancy alone. This result could be attributable to insufficient power. Second, although it has been suggested that smoking could increase susceptibility to pollutants by interfering with immune function and by promoting oxidative stress, which could sensitize cells to pollutants, it is also known that mothers who quit smoking before pregnancy gain more weight during pregnancy (Favaretto et al. 2007
), which could modulate both birth weight/length and cord blood pollutant levels. This could suggest that pharmacokinetic factors are responsible for these associations. Unfortunately, we did not record maternal weight gain during pregnancy in our study, which should be considered as a major study limitation. This could also suggest that maternal smoking has long-term health effects. These limitations require elucidation in future studies.
A second limitation of this study is that we related reported results to BMI SDS until a child reached 3 years of age. Although there is a correlation between BMI during pre-school age and adult BMI (Nader et al. 2006
; Sachdev et al. 2005
), we do not know whether the correlation with BMI will persist into later periods and whether these children will actually become overweight or obese. Third, although we found no differences in confounders between this cohort and the larger cohort from which it originates, there are significant regional differences in exposure to these pollutants in Flanders (Koppen G, Den Hond E, Nelen V, Van De Mieroop E, Bruckers L, Bilau M, et al., unpublished data). Therefore, our findings cannot be generalized to the entire Flemish population. Another limitation is that the exact duration of breast-feeding was not available, so we had to create two crude variables to assess postnatal exposure. Finally, as in every longitudinal study, missing data are a hindrance. However, we were able to confirm our results by multiple imputation.
The preschool period is a critical time in the development of obesity, because it coincides with the so-called adiposity rebound phenomenon. Indeed, BMI rises in infancy but declines during the preschool years, increasing again during the primary school years (Rolland-Cachera et al. 1984
). Even though we did not assess adiposity rebound in this study, it is accepted that higher BMI SDS during this period is associated with an increased risk of obesity in adult life (Whitaker et al. 1998
). Furthermore, there is a known correlation between BMI during the preschool years and adult BMI (Nader et al. 2006
; Sachdev et al. 2005
). This is the first study demonstrating that environmental pollution may influence BMI during these critical years. DDE and PCBs may influence BMI by interfering with several endocrine mediators. For instance, PCB and DDE levels are related to lower thyroid hormone levels in pregnant women (Takser et al. 2005
). DDT may also have estrogenic, antiandrogenic, and antiprogestin effects (Jurewicz et al. 2006
; Kelce and Wilson 1997
). PCBs may exert estrogenic, antiandrogenic, or antiestrogenic effects (Koppe et al. 2006
). Other studies provide evidence that PCBs are associated with increased insulin resistance in adults (Lee et al. 2007
; Vasiliu et al. 2006
). Although endocrine disruption caused by these pollutants has been well documented, few studies have investigated the influence of pre-natal exposure to pollutants and BMI later in life. One follow-up study found that increased transplacental exposure to DDE is associated with a significantly higher weight adjusted for height in adolescent boys (Gladen et al. 2000
). However, a study of the same group in adolescent boys with higher exposure to DDT and DDE failed to replicate this finding (Gladen et al. 2004
). A study by Karmaus and Eneli (2004)
demonstrates that intrauterine exposure to DDE, but not to PCBs, is associated with a higher weight and BMI in adult females.
The association between PCBs and BMI in our cohort seemed to begin early, demonstrated by its association with birth weight and length. Furthermore, maternal smoking modified this effect. This finding is in agreement with a recent study conducted by Sagiv et al. (2007)
, who also found effect modification by smoking on the association between PCBs and birth weight. After 1 year of age, PCBs were associated with smaller heights. Hertz-Picciotto et al. (2005)
found that prenatal PCBs were associated with increased growth in 5-year-old girls but not in boys. Lamb et al. (2006)
found that the association between PCBs and childhood growth among boys depended on the degree of ortho
On the other hand, estimated effects of DDE appeared to manifest later in life, because we found no association with birth size. We did find an association with weight that depended on the child’s age and whether the mother smoked. It is important to emphasize that there was a trend for DDE effect modification by sex (data not shown), but the present study did not have sufficient power to render this effect significant. Another prospective study found that increased prenatal DDE concentrations were associated with decreased height at 1, 4, and 7 years of age (Ribas-Fito et al. 2006
). Karmaus et al. (2002)
demonstrated that growth during childhood was significantly reduced in girls with high DDE concentrations measured at 8 years of age. They observed no growth effect of DDE in boys. Contradictory results may be attributable to different degrees of exposure, the inclusion of premature children, the age of assessment of anthropometric measurements (prepubertal or pubertal children), and the number of assessed confounding variables.
Finally, familial influences have been confirmed as risk factors increasing BMI in children. The finding that increased birth weight is associated with higher BMI in early childhood has been previously documented (Li et al. 2007
; Whitaker and Dietz 1998
), as has the finding that exposure to prenatal smoking increases incidence of overweight children (Oken et al. 2008
In conclusion, this study documents that intrauterine exposure to DDE and PCBs is associated with increased BMI during early childhood. The association with PCBs starts from birth, with a negative effect on birth weight and birth length, whereas the association with DDE is measurable between 1 and 3 years of age, with no effect on birth weight or length. Future prospective studies are needed to confirm these findings, including studies assessing possible mechanisms by which these pollutants could alter energy metabolism. Finally, we recommend more studies investigating the mechanisms of effect modification by smoking, especially on the influence of maternal weight gain during pregnancy.