Background: There are few data on predictors of HIV progression in untreated children in resource-limited settings.
Methods: Children with HIV Antibiotic Prophylaxis (CHAP) was a randomized trial comparing cotrimoxazole prophylaxis with placebo in HIV-infected Zambian children. The prognostic value of baseline characteristics was investigated using Cox models.
Results: Five hundred fourteen children aged 1 to 14 (median 5.5) years contributed 607 years follow-up (maximum 2.6 years). Half were boys, and in 67%, the mother was the primary carer; at baseline, median CD4 percentage was 11% and weight was less than third percentile in 67%. One hundred sixty-five children died (27.2 per 100 years at risk; 95% confidence interval 23.3-31.6). Low weight-for-age, CD4 percentage, hemoglobin, mother as primary carer, current malnutrition, and previous hospital admissions for respiratory tract infections or recurrent severe bacterial infections were independent predictors of poorer survival, whereas oral candidiasis predicted poorer survival only when baseline CD4 percentage was not considered. Mortality rates per 100 child years of 44.5 (37.2-53.2), 14.7 (10.9-19.8), and 2.3 (0.3-16.7) were associated with new World Health Organization stages 4, 3, and 1/2, respectively, applied retrospectively; very low weight-for-age was the only staging feature for 42% of stage 4 children.
Conclusions: Malnutrition and hospitalizations for respiratory/bacterial infections predict mortality independent of immunosuppression, suggesting that they capture HIV- and non-HIV-related mortality, whereas oral candidiasis is a proxy for immunosuppression.