Background: Excluding HIV infection among infants and young children in resource-poor settings where nucleic acid amplification tests (NAAT) are not routinely available remains a considerable challenge. Objectives: To assess the performance of two rapid HIV antibody tests (RT) used alone and in parallel for excluding HIV infection among acutely ill infants and children <18 months in comparison to NAAT in a region where maternal HIV prevalence was approximately 7%.

Study Design: Infants and children ≥2 < 18 months admitted to hospital with an acute febrile illness had two rapid antibody tests in parallel, with single and parallel results subsequently compared against NAAT.

Results: HIV prevalence among 1602 enrolled infants was 3.4%. All 1526 infants with 2 negative RT were HIV negative by NAAT. All 46 infants with 2 positive RT were HIV positive by NAAT. The overall specificity of two rapid tests for excluding HIV infection was 99.5%. Sensitivity and specificity were ≥99% and >98%, respectively, across all age brackets ≥2 < 18 months. Overall sensitivity and specificity for a single RT was 98.2% and 99%, respectively, for Determine, and 85.5% and 99.6%, respectively, for Capillus.

Conclusions: In a setting with a maternal HIV prevalence rate of <10%, a single negative RT had excellent specificity and two negative RT performed in parallel had a perfect negative predictive value for HIV infection among acutely ill patients <18 months of age. In this and similar settings, RT could assist with excluding HIV infection with much lower complexity and cost than NAAT.