Table 3.
Adjusted parameter estimates (95% CI) for the change in birth outcomes with each unit increase in natural log–transformed maternal 2nd-trimester urinary arsenic by maternal prepregnancy weight statusa and infant sex strata.b
| Maternal prepregnancy weight status/infant sex | Arsenic exposure | β (95% CI) | ||
|---|---|---|---|---|
| Birth length (cm) | Birth weight (g) | Ponderal index (kg/m3) | ||
| Normal weight | ||||
| Female (n = 204) | Ln(iAs) | –0.04 (–0.36, 0.27) | –7.1 (–59.3, 45.2) | 0.00 (–0.40, 0.40) |
| Ln(MMA) | 0.02 (–0.24, 0.29) | 14.9 (–29.2, 58.9) | 0.05 (–0.29, 0.39) | |
| Ln(DMA) | 0.03 (–0.36, 0.42) | 34.5 (–29.2, 98.3) | 0.17 (–0.32, 0.66) | |
| Ln(tAs) | 0.00 (–0.40, 0.40) | 23.5 (–42.4, 89.4) | 0.13 (–0.38, 0.63) | |
| Male (n = 202) | Ln(iAs) | 0.04 (–0.24, 0.32) | –0.7 (–50.9, 49.4) | –0.04 (–0.45, 0.36) |
| Ln(MMA) | 0.04 (–0.20, 0.27) | 20.5 (–22.0, 63.0) | 0.12 (–0.23, 0.46) | |
| Ln(DMA) | 0.17 (–0.16, 0.49) | 11.6 (–45.8, 69.0) | –0.12 (–0.58, 0.34) | |
| Ln(tAs) | 0.13 (–0.20, 0.47) | 3.7 (–56.1, 63.4) | –0.14 (–0.62, 0.34) | |
| Overweight/obese | ||||
| Female (n = 153) | Ln(iAs) | 0.00 (–0.32, 0.32) | –58.6 (–115.2, –1.9) | –0.50 (–0.97, –0.03) |
| Ln(MMA) | –0.07 (–0.38, 0.25) | –85.0 (–139.2, –30.8)** | –0.58 (–1.04, –0.12) | |
| Ln(DMA) | 0.00 (–0.41, 0.40) | –90.7 (–160.6, –20.8)* | –0.69 (–1.27, –0.10) | |
| Ln(tAs) | 0.01 (–0.40, 0.41) | –90.7 (–161.0, –20.5)* | –0.73 (–1.32, –0.14) | |
| Male (n = 147) | Ln(iAs) | 0.34 (–0.09, 0.77) | 16.9 (–65.2, 98.9) | –0.41 (–0.96, 0.15) |
| Ln(MMA) | 0.45 (0.14, 0.76)** | 28.6 (–31.2, 88.4) | –0.50 (–0.90, –0.09) | |
| Ln(DMA) | 0.75 (0.32, 1.17)** | 39.2 (–44.5, 122.9) | –0.95 (–1.50, –0.39) | |
| Ln(tAs) | 0.74 (0.30, 1.19)** | 38.7 (–47.7, 125.0) | –0.94 (–1.51, –0.36) | |
| p for 3-way interaction between infant sex, maternal weight status and Ln(tAs)c | 0.13 | 0.03 | 0.89 | |
| aNormal weight defined as prepregnancy BMI 18.5–24.9 kg/m2 and overweight/obese defined as prepregnancy BMI ≥ 25 kg/m2; 25 underweight mothers (BMI < 18.5 kg/m2) excluded from analysis. bResults from a series of regression models fitting each birth outcome on each arsenic exposure measure [natural log–transformed inorganic arsenic (iAs), monomethylarsonic acid (MMA), dimethylarsinic acid (DMA), total arsenic (iAs + MMA + DMA) (tAs)]. All models adjusted for maternal age (continuous), education level (ordinal with the five categories listed in Table 1 assigned values of 1–5, respectively), parity (continuous), and gestational age (continuous). Models adjusted for maternal age (continuous), education level (ordinal with the five categories listed in Table 1 assigned values of 1–5, respectively), parity (continuous), gestational age (continuous) and infant sex. cTo test for a 3-way interaction between maternal weight status, infant sex, and Ln(tAs), regressions were fit as described in footnote b with the addition of six 2-way multiplicative interaction terms between maternal weight status, infant sex, and arsenic exposure, and one 3-way multiplicative interaction term between all 3 variables; p-value from a two-sided Wald t-test on the coefficient for the 3-way multiplicative term. Results for test of interaction with Ln(MMA) and Ln(DMA) are similar (data not shown). *p < 0.05. **p < 0.01. | ||||