Of the 4,704 eligible children 4–15 years of age, 344 (8.2%) had only parent-reported ADHD and 154 (4.3%) reported stimulant medication use, equivalent to 3.8 million and 2.0 million U.S. children and adolescents, respectively. Of the 4,704 children, 135 (4.2% weighted percent) had parent report of both ADHD and stimulant medication use, equivalent to 1.8 million children in the United States. (). In bivariate analyses, we found a significant association between parent-reported ADHD and stimulant medication use with prenatal ETS exposure (p = 0.023), preschool attendance (p = 0.003), male sex (p < 0.001), increasing age in years (p < 0.001), and health insurance coverage (p < 0.001) (). Non-Hispanic white children were more likely than other racial groups to report ADHD (p = 0.001).
Prevalence of stimulant medication use and parent-reported ADHD among children 4–15 years of age in NHANES 1999–2002 according to demographic and medical factors.
In multivariable analysis, prenatal ETS exposure and blood lead concentration were significant predictors of ADHD (). The adjusted odds ratio (AOR) for prenatal ETS exposure was 2.5 [95% confidence interval (CI), 1.2–5.2]. We also found a significant dose–response relationship between lead exposure and ADHD (). Compared with children in the lowest quintile of blood lead concentration, children with blood lead levels in the fifth quintile (AOR = 4.1; 95% CI, 1.2–14.0) were at significantly higher risk for ADHD. The risk of ADHD was also significantly associated with male sex (AOR = 3.7; 95% CI, 2.1–6.6). Mexican-American and non-Hispanic black children had lower risks for reported ADHD diagnosis and stimulant medication use (AOR = 0.3; 95% CI, 0.1–0.7 and AOR = 0.5; 95% CI, 0.3–0.8, respectively) compared with non-Hispanic white children. Postnatal ETS exposure, as measured by the presence of a smoker in the home, was not a significant predictor of ADHD status in adjusted models (AOR = 0.6; 95% CI, 0.3–1.3; p = 0.224). The risk for ADHD was significantly associated with preschool attendance (AOR = 2.4; 95% CI, 1.1–5.1).
Logistic regression analysis for parent-reported attention deficit disorder among children 4–15 years of age, NHANES 1999–2002.a
Figure 1 AOR for ADHD among U.S. children, NHANES 1999–2002, by blood lead concentration (μg/dL). The model was adjusted for child’s age, sex, race/ethnicity, preschool attendance, serum ferritin, prenatal ETS exposure, smoker in the household, (more ...)
Next, we tested for interactions between sex and prenatal exposure to ETS, sex and blood lead concentration, and prenatal exposure to ETS and blood lead concentration. We did not find a significant interaction between prenatal ETS exposure and sex using a formal interaction term (p = 0.141). Compared with unexposed females, females who were pre-natally exposed to ETS were at a 4.6-fold higher risk for ADHD compared with unexposed females (OR = 4.6; 95% CI, 1.7–12.4), whereas exposed males were at an almost significant 2-fold higher risk for ADHD than unexposed males (OR = 2.1; 95% CI, 0.9–4.7; p = 0.073) (). There was not a significant interaction between blood lead levels by sex (p = 0.242) or blood lead levels by maternal smoking (p = 0.837).
Figure 2 AOR for ADHD among U.S. children by prenatal ETS exposure and sex. The risk for ADHD among ETS-exposed children was greater in females; females who were prenatally exposed to tobacco were at 4.6-fold higher risk for ADHD compared with unexposed females (more ...)
We conducted secondary analyses to examine the effects of lead exposure at blood lead levels < 5 μg/dL and to test the stability of our results. When the sample was restricted to children with concurrent blood lead concentrations ≤ 5 μg/dL, there was still a significant association between higher blood lead levels and ADHD. Compared with children in the lowest quintile (nondetectable to 0.7 μg/dL), children with blood lead levels in the highest quintile (2.0–5 μg/dL) had a 4.5-fold (95% CI, 1.3–15.3) higher risk for ADHD.
When birth weight and NICU were added to the primary model, the adjusted OR for prenatal ETS exposure and the fifth quintile of blood lead levels did not change, remaining at 2.2 (95% CI, 1.0–5.1; p = 0.055) and 4.5 (95% CI, 1.3–15.6; p = 0.019), respectively. Consistent with postnatal ETS exposure measured by parent report, postnatal ETS exposure using serum cotinine was not associated with ADHD (AOR = 0.99; 95% CI, 0.97–1.00; p = 0.092). Using the same multivariable model, we found that when the ADHD outcome was defined simply by parent report or by stimulant medication use, rather than the combination of the two, the AOR of maternal smoking and blood lead level did not differ appreciably. Finally, there was no substantive change in the relationship of either prenatal ETS exposure or blood lead concentration when we included children without routine access to health care in the model.
The PAF for prenatal ETS exposure for both males and females was 18.4% (95% CI, 5.1–24.8%), corresponding to 270,000 cases of ADHD in children 4–15 years of age (). Although there was a significant dose–response relationship between lead exposure and ADHD, we estimated the population attributable fraction only for children who had blood lead levels in the fifth quintiles of blood lead concentration. Our estimates indicate that 21.1% (95% CI, 4.6–25.9%) of ADHD cases among children 4–15 years of age were attributable to having a blood lead > 2.0 μg/dL. This corresponds to 290,000 excess cases of ADHD among U.S. children 4–15 years of age (). Finally, we calculated the PAF for having either prenatal ETS exposure or blood lead concentration > 2.0 μg/dL to account for children who had both exposures. Our estimates indicate that 32.2% (95% CI, 4.2–41.3%) of ADHD cases among children 4–15 years of age were attributable to having either prenatal ETS exposure or blood lead > 2.0 μg/dL, which corresponds to 480,000 excess cases of ADHD among U.S. children 4–15 years of age.
Population-attributable fraction of prenatal tobacco exposure and environmental lead exposure for parent-reported ADHD and stimulant medication use in children 4–15 years of age, NHANES 1999–2002.a