Abstract Objectives Routine amoxicillin for children with uncomplicated severe acute malnutrition raises concerns of increasing antibiotic resistance. We performed an ancillary study nested within a double-blind, placebo-controlled trial in Niger testing the role of routine 7 day amoxicillin therapy in nutritional recovery of children 6 to 59 months of age with uncomplicated severe acute malnutrition. Methods We screened 472 children for rectal carriage of ESBL-producing Enterobacteriaceae (ESBL-E) as well as their household siblings under 5 years old, at baseline and Week 1 (W1) and Week 4 (W4) after start of therapy, and characterized strains by WGS. ClinicalTrials.gov: NCT01613547. Results Carriage in index children at baseline was similar in the amoxicillin and the placebo groups (33.8% versus 27.9%, P = 0.17). However, acquisition of ESBL-E in index children at W1 was higher in the amoxicillin group than in the placebo group (53.7% versus 32.2%, adjusted risk ratio = 2.29, P = 0.001). Among 209 index and sibling households possibly exposed to ESBL-E transmission, 16 (7.7%) had paired strains differing by ≤10 SNPs, suggesting a high probability of transmission. This was more frequent in households from the amoxicillin group than from the placebo group [11.5% (12/104) versus 3.8% (4/105), P = 0.04]. Conclusions Among children exposed to amoxicillin, ESBL-E colonization was more frequent and the risk of transmission to siblings higher. Routine amoxicillin should be carefully balanced with the risks associated with ESBL-E colonization.