shows the geometric mean levels of PAH exposure overall and by categories of obesity risk factors for the entire cohort and for children at ages 5 and 7 years. For the cohort overall, mothers receiving public assistance had significantly higher exposure levels, and for children followed up to age 5 years, mothers of African-American children had higher exposure levels than mothers of Dominican children. In the indoor air monitoring substudy, the intraclass correlation coefficient for the consecutive 2-week indoor air sampling periods was 0.65 during the last 6–8 weeks of pregnancy. Estimates of prenatal PAH exposure based on individual 48-hour personal air monitoring and based on the mean of the consecutive 2-week integrated indoor air monitoring periods were significantly correlated (r = 0.58, P < 0.001).
Geometric Mean Levels (ng/m3) of Polycyclic Aromatic Hydrocarbons in Prenatal Personal Air Monitoring Samples in the Mothers and Children Study in Northern Manhattan and the South Bronx, 1998–2011
Complete data for outcomes and key covariates (sex, ethnicity, receipt of public assistance during pregnancy, birth weight, and maternal prepregnancy obesity) were available from 422 of the 453 children followed to age 5 years for anthropometric outcomes and from 341 of the 371 children followed to age 7 years. Data on body composition and key covariates were available for 297 children. The primary missing data element was maternal prepregnancy weight. presents descriptive statistics for the sociodemographic characteristics and risk factors for all cohort children at baseline and for those followed up. At age 5 years, 21% of the children were obese, and at age 7 years, 25% were obese; the mean percentage of body fat at age 7 years was 24.1% (7.2 kg of fat mass). As expected, for the 331 children from whom anthropometric data were available at ages 5 and 7 years, obesity at age 5 was strongly predictive of obesity at age 7 (κ = 0.67, P < 0.001; relative risk = 8.39, 95% confidence interval (CI): 5.63, 12.50).
Sociodemographic and Anthropometric Characteristics of the Cohort at Baseline and During Follow-up, Mothers and Children Study in Northern Manhattan and the South Bronx, 1998–2011
In unadjusted analyses, compared with children whose mothers were in the first tertile of prenatal PAH exposure, at age 5 years the BMI z score was 0.33 units higher (95% CI: 0.02, 0.65) for children of mothers in the second exposure tertile, and it was 0.43 units higher (95% CI: 0.12, 0.75) for children of mothers in the third exposure tertile. Similar effects were seen at age 7 years; compared with the first tertile, the second tertile of exposure was associated with a 0.17-unit higher BMI z score (95% CI: −0.14, 0.48), and the third tertile of exposure was associated with a 0.32-unit higher BMI z score (95% CI: 0.01, 0.62).
shows the multivariate associations between model covariates and categories of maternal PAH exposure and BMI z score and body composition measures. After control for the model covariates, the highest tertile of prenatal PAH exposure was significantly associated with higher BMI z score at both age 5 years and age 7 years. In addition, at age 7 years, higher prenatal PAH exposure was significantly associated with increased percentage of body fat and fat mass but not with variation in lean mass (e.g., organs, bone, and muscle). shows the covariate-adjusted mean BMI z scores (and 95% confidence intervals) at ages 5 and 7 years by tertile of prenatal PAH exposure. To place the results in context, for a 5-year-old boy with a weight of 21 kg and a height of 115 cm (BMI z score = 0.37, BMI percentile = 64.5—median values for the study population), an increase of 0.39 z-score units equates to a 0.7-kg increase in weight and a 13-unit increase in BMI percentile. Similarly, for a 7-year-old boy with a weight of 27 kg and a height of 126 cm (BMI z score = 0.86, BMI percentile = 80.5), an increase of 0.30 z-score units equates to a 1.11-kg increase in weight and places the child at the 87.7th BMI percentile. Being in the third tertile of prenatal PAH exposure as compared with the first was associated with 1.1-kg higher fat mass. shows that the second and third tertiles of maternal PAH exposure were associated with higher relative risks of obesity at ages 5 and 7 years.
Table 3. Associations of Prenatal Exposure to Polycyclic Aromatic Hydrocarbons and Sociodemographic and Early-Life Characteristics With Body Size and Composition at Ages 5 and 7 Years Among Children in the Mothers and Children Study in Northern Manhattan and the (more ...)
Figure 2. Covariate-adjusted mean body mass index (BMI; weight (kg)/height (m)2) z scores according to tertile of prenatal ambient air polycyclic aromatic hydrocarbon (PAH) exposure for children at ages 5 and 7 years, Mothers and Children Study in Northern Manhattan (more ...)
Table 4. Risks of Obesity at Ages 5 and 7 Years According to Prenatal Exposure to Polycyclic Aromatic Hydrocarbons and Sociodemographic and Early-Life Characteristics Among Children in the Mothers and Children Study in Northern Manhattan and the South Bronx, 1998–2011 (more ...)
We conducted further analyses to assess the sensitivity of the results to additional model specifications and adjustment for other covariates. Neighborhood socioeconomic status represents a potential confounding factor; however, further control for neighborhood poverty rate or median household income did not alter the results. In regards to factors hypothesized to be associated with higher ambient air PAH levels, PAH levels were not associated with the mother’s reporting that a smoker lived in the home or with neighborhood socioeconomic status. Air monitoring samples collected during heating months showed 47% higher PAH levels (95% CI: 29, 67) than samples collected during the nonheating months, and higher PAH levels were associated with higher street density within a 1-km radius of the home (2.94% higher per linear kilometer of street; 95% CI: 0.20, 5.65). However, season of sampling and neighborhood street density did not predict BMI z score at age 5 or 7 years or body composition at age 7 years, and adjustment for these variables did not alter the analytical results.
Weighting the data by the inverse probability of follow-up and complete data collection at age 5 years only slightly reduced the estimated effect of the third tertile of PAH exposure on BMI z score (weighted estimate = 0.34 BMI z-score units; 95% CI: 0.02, 0.66) and did not alter the effect estimate for the second tertile of exposure. However, inverse probability-weighted analyses for BMI z score at age 7 years suggested that the initial analyses were biased toward the null value. Compared with the first tertile of PAH exposure, the covariate-adjusted weighted β coefficient for the second tertile of exposure was 0.28 BMI z-score units (95% CI: −0.01, 0.58), while in the unweighted analyses the covariate-adjusted coefficient was 0.17 units (95% CI: −0.12, 0.46). For comparisons between the third and first tertiles of exposure, the covariate-adjusted weighted β coefficient was 0.40 BMI z-score units (95% CI: 0.11, 0.69), while the corresponding covariate-adjusted coefficient in the unweighted analyses was 0.30 BMI z-score units (95% CI: 0.01, 0.58).