RESUMO
OBJECTIVES: HIV-exposed infants, including those who do not become infected, have higher morbidity and mortality rates than HIV unexposed infants. The underlying mechanisms of this difference are largely unknown. The objective of this study was to identify the risk factors for mortality among HIV-exposed (infected as well as uninfected) infants in a prevention of mother-to-child transmission (PMTCT) programme in Cameroon. METHODS: We analysed the data from 319 mother-infant pairs included in a PMTCT programme at a rural and an urban hospital between 2004 and 2012. The programme offered free formula feeding, monthly follow-up visits and antiretroviral therapy (ART) according to national PMTCT guidelines. Mother-infant pairs were divided in three study groups, based on year of recruitment and study site: (I) rural hospital, 2004-07; (II) rural hospital, 2008-12; (III) urban hospital, 2008-12. RESULTS: Two hundred and eighty-five medical records were included in the final analysis. Infant mortality rates were 23.9%, 20.0% and 5.3% in group I, II and III, respectively (P = 0.02). Hazard ratios of infant mortality were 6.4 (P < 0.001) for prematurity, 4.6 (P = 0.04) for no maternal use of ARTs, 5.6 (P = 0.025) for mixed feeding, 2.7 for home deliveries (P = 0.087) and 0.4 (P = 0.138) for urban study group. CONCLUSIONS: In this programme, prematurity, no ART use, and the practice of mixed feeding were independent predictors of infant mortality. Mixed feeding and not using ART increased the hazard of death, probably through its increased risk of HIV infection. Although mortality rates were significantly higher in the rural area, rural setting was not a risk factor for infant mortality. These findings may contribute to the development of tailor-made programmes to reduce infant mortality rates among HIV-exposed infants.