Abundance of pathogenic E. coli at endpoint, by group, in Kenyan infants (n = 115) receiving daily for 4 months an MNP containing either no iron (−FeMNP) or 2.5 mg or 12.5 mg iron (+FeMNP). Univariate general linear models with baseline values as covariates were used to estimate the intervention effect. Adapted from Jaeggi et al.²⁰