TABLE 4.
Outcome | Level of effect modifier | IFA | MMN | LNS | P-interaction2 | P value3 |
IgA | BMI | 0.008 | ||||
10th percentile | 0.33 (0.30, 0.36)a | 0.27 (0.25, 0.30)b | 0.29 (0.26, 0.32)a,b | 0.003 | ||
90th percentile | 0.30 (0.27, 0.33) | 0.33 (0.30, 0.37) | 0.34 (0.31, 0.37) | 0.121 | ||
Low fucosylated HMO | HHAZ | 0.048 | ||||
10th percentile | 0.26 (0.20, 0.34) | 0.17 (0.12, 0.24) | 0.23 (0.17, 0.31) | 0.174 | ||
90th percentile | 0.26 (0.17, 0.37) | 0.37 (0.27, 0.48) | 0.27 (0.17, 0.38) | 0.270 | ||
Low antitrypsin | Parity | 0.012 | ||||
Primiparous | 0.10 (0.06, 0.17)a | 0.16 (0.10, 0.24)a,b | 0.29 (0.20, 0.41)b | 0.003 | ||
Multiparous | 0.44 (0.31, 0.59) | 0.32 (0.21, 0.45) | 0.29 (0.19, 0.42) | 0.234 | ||
Maternal age | 0.046 | |||||
10th percentile | 0.15 (0.09, 0.26)a | 0.24 (0.15, 0.35)a,b | 0.35 (0.25, 0.48)b | 0.015 | ||
90th percentile | 0.28 (0.18, 0.41) | 0.20 (0.12, 0.32) | 0.22 (0.13, 0.34) | 0.459 |
Values are estimated means (95% CIs) at the 10th and 90th percentiles of the effect modifier for continuous outcome variables as determined by ANCOVA after adjusting for covariates unless otherwise indicate. The estimated probability of low outcomes (95% CIs) at the 10th and 90th percentiles of the effect modifier for binary outcomes was determined by logistic regression after adjusting for covariates. IgA was log transformed before analyses were performed. IgA was adjusted for HHAZ, maternal HIV status, maternal BMI (in kg/m2), and season. Low fucosylated HMO was adjusted for HHAZ and season. Low antitrypsin was adjusted for maternal age, parity, and season. Values with different superscript letters were significantly different (P < 0.05) in pairwise comparisons. HHAZ, household asset z score; HMO, human milk oligosaccharide; IFA, iron and folic acid; LNS, lipid-based nutrient supplement; MMN, multiple micronutrient.
P values are presented for the interaction of group × effect modifier on the outcome variable.
Global P values are presented for the 3-group comparison.