Objective: To evaluate the World Health Organisation (WHO) antibody testing strategy for the individual patient diagnosis of HIV infection (strategy III).

Design: Evaluation of a combination of enzyme-linked immunosorbent assays (ELISAs) for the detection of antibodies to HIV-1 and HIV-2 infection. The WHO strategy III calls for a combination of three ELISAs, based on different antigens and/or differing test principles, to be used in a sequential fashion. The first part of the study evaluated various kits as part of a selection process. The second part of the study was an assessment of the three-ELISA testing strategy on routine sera submitted to the National Institute for Virology (NIV) for HIV testing. In all instances, the conventional testing protocol, which utilises a screening ELISA followed by a confirmatory Western blot (WB) on positive specimens, was used as the 'gold standard'.

Setting: The HIV-testing laboratory at the NIV, Johannesburg.

Results: In the first part of the study, all of the ELISA kits evaluated showed high sensitivity and specificity, and a selection of the kits for part two of the study was made in accordance with the WHO recommendation. The kits selected, in order of use, were the Biotest anti-HIV 1/2 (test 1), Pasteur Genelavia Mixt HIV-1/2 (test 2) and Murex Wellcozyme HIV-1 competitive assay (test 3). This combination was evaluated using 7,812 sera submitted to the NIV for serology testing. The sensitivities of the tests were highly satisfactory (99.6-100%) as were the specificities (99.2-100%). The positive predictive value of strategy III at various seroprevalence (0.5-25.5%) was > or = 99.6%. The rate of WB usage when compared with the previous HIV testing protocol was low (4.6%).

Conclusions: The results of this study suggest that this testing protocol could be introduced in South Africa with considerable cost-saving and no reduction in specificity.