TABLE 3.
Multivariate Cox proportional hazards regression models assessing risk of growth faltering in EU infants by feeding as randomized and as practiced1
|
Underweight (WAZ < −2) |
Wasting (WLZ < −2) |
Stunting (LAZ < −2) |
||||
| aHR (95% CI) | P-value | aHR (95% CI) | P-value | aHR (95% CI) | P-value | |
| Model 1 (as randomized)2 | ||||||
| Infant feeding as randomized | ||||||
| BF arm | 1.00 (ref) | — | 1.00 (ref) | — | 1.00 (ref) | — |
| FF arm | 1.05 (0.63, 1.77) | 0.84 | 1.34 (0.67, 2.68) | 0.40 | 1.01 (0.69, 1.50) | 0.94 |
| Infant WAZ, WLZ, or LAZ at birth | 0.38 (0.27, 0.53) | <0.001 | 0.78 (0.59, 1.02) | 0.07 | 0.48 (0.40, 0.57) | <0.001 |
| Diarrheal episode in past month3 | ||||||
| No | — | — | 1.00 (ref) | — | — | — |
| Yes | — | — | 2.26 (1.11, 4.62) | 0.03 | — | — |
| Maternal HIV RNA viral load, log10 | 1.25 (0.91, 1.72) | 0.18 | 0.87 (0.57, 1.34) | 0.53 | 1.30 (1.01, 1.66) | 0.04 |
| Maternal education, y | 0.90 (0.83, 0.99) | 0.03 | — | — | 0.91 (0.85, 0.97) | 0.004 |
| Model 2 (as practiced)4 | ||||||
| Infant feeding as practiced | ||||||
| Ever breast-fed | 1.00 (ref) | — | 1.00 (ref) | — | 1.00 (ref) | — |
| Never breast-fed | 1.14 (0.67, 1.92) | 0.64 | 1.40 (0.71, 2.73) | 0.33 | 0.94 (0.64, 1.38) | 0.74 |
| Infant WAZ, WLZ, or LAZ at birth | 0.38 (0.28, 0.53) | <0.001 | 0.78 (0.59, 1.02) | 0.07 | 0.48 (0.41, 0.57) | <0.001 |
| Diarrheal episode in past month3 | ||||||
| No | — | — | 1.00 (ref) | — | — | — |
| Yes | — | — | 2.21 (1.09, 4.50) | 0.03 | — | — |
| Maternal HIV RNA viral load, log10 | 1.23 (0.89, 1.71) | 0.20 | 0.86 (0.56, 1.31) | 0.48 | 1.30 (1.02, 1.67) | 0.04 |
| Maternal education, y | 0.90 (0.82, 0.99) | 0.03 | — | — | 0.91 (0.85, 0.97) | 0.003 |
Values are aHR (95% CI); = 285 for underweight, n = 280 for wasting, n = 264 for stunting. aHR, adjusted HR; BF, breastfeeding; EU, HIV-1-exposed, uninfected; FF, formula feeding; LAZ, length-for-age Z-score; ref, reference; WAZ, weight-for-age Z-score; WLZ, weight-for-length Z-score.
Model 1 covariates: infant feeding as randomized, infant birth Z-score, diarrheal episode, maternal HIV viral load, and maternal education. Due to colinearity, the impact of maternal education and maternal height was evaluated in separate models. When all variables in model 1 except for education were adjusted for, greater maternal height was associated with lower risk of underweight and stunting [multivariate model: aHR = 0.92 (95% CI: 0.87, 0.97), = 0.002, and aHR = 0.96 (95% CI: 0.92, 0.99), P = 0.02, respectively].
Episode of diarrhea occurring in the month prior to Z-score < −2.
Model 2 covariates: infant feeding as practiced, infant birth Z-score, diarrheal episode, maternal HIV viral load, and maternal education.