Table 5.
Estimated coefficients (95% CIs) for regression models with repeated measures: birth and 4, 7, and 17 years of age.
| Outcomea |
||||
|---|---|---|---|---|
| Exposureb | Girls’ weight | Girls’ height | Boys’ weight | Boys’ height |
| PCB-15 (coplanar) | 1.1 (–5.9 to 8.1) | 1.9 (–0.0 to 3.8) | 0.7 (–1.1 to 2.4) | 0.2 (–0.4 to 0.7) |
| ∑PCBmono | –13.7 (–20.8 to –6.6) | –1.8 (–4.2 to 0.6) | 0.6 (–4.9 to 6.0) | 0.0 (–1.7 to 1.8) |
| ∑PCBdi | –10.6 (–18.2 to –3.0) | –1.1 (–3.5 to 1.4) | 3.7 (–1.6 to 9.0) | 0.6 (–0.9 to 2.1) |
| ∑PCBtri | –9.2 (–15.0 to –3.3) | –1.0 (–3.1 to 1.2) | 1.8 (–2.0 to 5.6) | 1.2 (0.1 to 2.2) |
| ∑PCBall | –13.2 (–20.7 to –5.7) | –0.6 (–3.2 to 2.0) | 1.7 (–4.7 to 8.1) | 0.5 (–1.4 to 2.3) |
Abbreviations: ∑PCBall, sum of all PCBs; ∑PCBdi, sum of di-ortho-substituted PCBs; ∑PCBmono, sum of mono-ortho-substituted PCBs; ∑PCBtri, sum of tri-ortho-substituted PCBs. Data are 100 × the parameter estimates for the log-transformed PCB measure, adjusted for maternal prepregnancy weight, preterm status, and serum triglyceride and cholesterol. Models with height as the outcome were also adjusted for maternal height.
a
Height and weight measures were log-transformed.
b
All PCB exposure measures were summed by structural categorization and then log-transformed.