Performance of an alternative laboratory-based algorithm for diagnosis of HIV infection utilizing a third generation immunoassay, a rapid HIV-1/HIV-2 differentiation test and a DNA or RNA-based nucleic acid amplification test in persons with established HIV-1 infection and blood donors

Abstract Background The HIV-1 Western blot (WB) and immunofluorescence assay used to confirm HIV infections are less sensitive during seroconversion than immunoassays (IAs) used for screening. An alternative diagnostic algorithm has been proposed to detect early HIV-1 infection and differentiate HIV-1 from HIV-2. Objectives We evaluated the performance of an algorithm with a third generation IA that when reactive was followed by a rapid test (Multispot) that differentiates HIV-1 from HIV-2. Multispot-reactive specimens were considered HIV-infected. Multispot-negative specimens were tested with a nucleic acid amplification test (NAAT) for resolution. Study design WB-positive specimens [serum ( n = 2202), plasma ( n = 1109) and peripheral blood mononuclear cells (PBMCs) ( n = 1065)] were obtained from HIV-infected persons not taking antiretrovirals. HIV-uninfected specimens [plasma ( n = 1517) and PBMCs ( n = 1508)] with negative IA and NAAT results were obtained from blood donors. Specimens were tested with third generation IAs (Abbott rDNA, ADVIA Centaur, GS HIV1-2 Plus O, Ortho VITROS) in singlet, Multispot, and NAAT (APTIMA (RNA) and AMPLICOR (DNA)). We calculated algorithm sensitivity and specificity and the proportion of IA-reactive specimens requiring NAAT. Results Algorithm sensitivity was 99.95% with APTIMA and 100% with AMPLICOR. One WB-positive specimen reactive by all IAs and AMPLICOR was negative by Multispot and APTIMA. Algorithm specificity was 100% using APTIMA or AMPLICOR as NAAT. From 0.10% (Abbott) to 2.43% (VITROS) of IA-reactive specimens required NAAT. Conclusions The proposed algorithm performs with high sensitivity and specificity in specimens from persons with established HIV infection and uninfected blood donors and appears to be a good alternative to the current algorithm.