Résumé
Abstract only This analysis assessed morbidity risks associated with feeding mode in HIV‐1‐exposed uninfected African infants for two age periods (0–3 and 3–6 mo). HIV‐1‐infected women from 5 sites in Burkina Faso, Kenya and South Africa were enrolled during pregnancy and counselled to either breastfeed (BF) exclusively and wean before six mo postpartum or formula‐feed from birth. Infant morbidity was defined in two distinct ways, as ‘trivial’ (i.e. mother‐reported fever, diarrhea, or vomiting since the preceding clinic visit) or as ‘SAE’ (i.e. selected clinical serious adverse events: gastroenteritis, lower respiratory infections, severe malnutrition or death). The analysis (n= 982) was conducted using multiple piecewise exponential regression with adjustment for study site, maternal CD4 cell count at enrollment and cotrimoxazole prophylaxis (CTX). Feeding mode and CTX were time‐dependent variables, defined monthly. Not BF (i.e. never BF or weaned) was associated with increased risks of ‘trivial’ morbidity from 0–3 mo [HR=1.3 (95% CI: 1.1–1.6), P<0.001] but not from 3–6 mo [1.1(0.8–1.5)], and with increased risks of SAE at 0–3 [5.6 (2.9–11.6), P<0.001] and 3–6 mo [2.9 (1.4–5.8), P<0.01]. Partial or predominant BF modes were not associated with significantly increased risks, as compared to exclusive mode. BF prevents serious morbidity and fatal outcomes in African infants up till 6 months. Funding: ANRS, WHO, EDCTP.