HIV-exposed uninfected infant morbidity and mortality within a nationally representative prospective cohort of mother-infant pairs in Zimbabwe.

OBJECTIVE To examine morbidity and mortality risk among HIV-exposed uninfected (HEU) infants. DESIGN Secondary data analysis of HEU infants in a prospective cohort study of mother-infant pairs. METHODS Infants were recruited from immunization clinics (n = 151) in Zimbabwe from February to August 2013, enrolled at 4-12 weeks age, and followed every 3 months until incident HIV-infection, death, or 18-month follow-up. We estimated cumulative mortality probability and hazard ratios with 95% confidence intervals (CIs) using Kaplan-Meier curves and Cox regression, respectively. We also described reported reasons for infant hospitalization and symptoms preceding death. Median weight-for-age z-scores (WAZ) and median age were calculated and analyzed across study visits. RESULTS Of 1188 HIV-exposed infants, 73 (6.1%) contracted HIV; we analyzed the remaining 1115 HEU infants. In total, 54 (4.8%) infants died, with median time to death of 5.5 months since birth (interquartile range: 3.6-9.8 months). Diarrhea, difficulty breathing, not eating, fever, and cough were commonly reported (range: 7.4-22.2%) as symptoms preceding infant death. Low birth weight was associated with higher mortality (adjusted hazard ratio 2.66, CI: 1.35-5.25), whereas maternal antiretroviral therapy predelivery (adjusted hazard ratio 0.34, CI: 0.18-0.64) and exclusive breastfeeding (adjusted hazard ratio 0.50, CI: 0.28-0.91) were associated with lower mortality. Overall, 9.6% of infants were hospitalized. Infant median WAZ declined after 3 months of age, reaching a minimum at 14.5 months of age, at which 50% of infants were underweight (WAZ below -2.0). CONCLUSION Clinical interventions including maternal antiretroviral therapy; breastfeeding and infant feeding counseling and support; and early prevention, identification, and management of childhood illness; are needed to reduce HEU infant morbidity and mortality.