Background: Failure to determine the HIV status of all pregnant women impedes progress in preventing and treating paediatric HIV because vertically exposed infants are not identified for prophylaxis, early HIV diagnosis and care.
Objectives: To assess the performance of rapid HIV tests in comparison to a laboratory-based HIV ELISA test for determining HIV-exposure and excluding HIV infection during infancy.
Study design: Seven rapid HIV tests were evaluated on 2266 stored samples from 116 HIV-exposed infants of known HIV status at four ages during infancy. The HIV ELISA for each sample was the standard against which rapid results were assessed to establish HIV-exposure.
Results: Rapid tests did not perform uniformly during infancy. For detecting HIV-exposure the sensitivity of most rapid tests to 3 months of age approached that of an HIV ELISA however only Determine maintained this sensitivity (99.7%) throughout infancy. For excluding HIV infection (i.e. for correctly identifying HIV-uninfected infants) the specificity of all rapid tests except Determine exceeded that of the HIV ELISA from 7 months of age.
Conclusions: The use of rapid tests in infancy could improve identification and care of HIV-exposed infants. Further evaluation under field conditions is required before rapid tests can be incorporated into evidence-based diagnostic algorithms.