Pitfalls with rapid HIV antibody testing in HIV-infected children in the Western Cape, South Africa.

Background: Rapid HIV antibody tests are commonly used for HIV diagnosis in the developing world. These tests are generally reported as sensitive, despite paucity of evaluations in paediatric populations.
Objectives: We tested specimens of paediatric patients, known to be HIV-infected, to detect any false negative tests and determine associations with such an outcome.
Study Design: One hundred and fifty-three specimens, from 109 patients, recorded to be HIV-infected by standard testing, were tested on the Capillustrade mark HIV-1/HIV-2 test (Trinity Biotech, Ireland); 150 specimens also had sufficient volume to be tested on Abbott Determinetrade mark HIV1/2 assay (Abbott GmbH, Wiesbaden, Germany). Treatment information, CD4 counts and HIV-1 viral load measurements were obtained from patient files and laboratory databases.
Results: Twenty-one of 153 specimens tested negative on the Capillus (sensitivity 86.3%). False negative results by Capillus were associated with antiretroviral treatment (ART) (p=0.0018) and lower HIV-1 viral load (p=0.013). Serial dilutions of some of the specimens indicated that both rapid tests, and the Capillus in particular, became negative at lower dilutions than an HIV enzyme immunoassay (EIA).
Conclusions: The Capillus test had an unexpectedly low sensitivity in a South African population of HIV-infected children that had access to antiretroviral treatment, posing a risk of false negative HIV testing.