LARGE-SCALE EVALUATION OF AN ALTERNATIVE STRATEGY FOR CONFIRMATION OFHIV ANTIBODIES

Objective: To retrospectively compare the accuracy of combinations of two enzyme-linked immunosorbent assays (ELISAs) with a Western blot ba sed strategy for identification of human immunodeficiency virus type 1 (HIV-1) seropositivity.Materials and methods: 48,977 sera, sent to th e National Bacteriological Laboratory, Stockholm, Sweden, for HIV anti body determinations between October 1988 and June 1993, were investiga ted. All samples were tested in parallel with two different ELISAs, ei ther Abbott Recombinant HIV-1 EIA and Wellcozyme Recombinant Anti-HIV- 1 EIA, or Enzygnost Anti-HIV-1/2 and Wellcozyme Recombinant Anti-HIV-1 EIA, or Enzygnost Anti-HIV-1/2 and Wellcozyme Anti-HIV-1 + 2 EIA. 156 5 sera repeatedly reactive by one or both ELISAs were investigated by Western blot (WB). Furthermore, a total of 2820 referred sera, screen reactive at primary laboratories but negative on our combinations of t wo ELISAs were analysed by WB. Results: Out of 1244 truly HIV antibody positive samples 1203 were WB positive and 41 (3.4%) were WB indeterm inate. A sensitivity of 100% was obtained by all three combinations of two ELISAs on examination of these 1244 sera including repeated testi ng of 5 samples with initially discrepant results. Among 2820 sera fro m HIV-negative individuals 649 (23%) sera were WB indeterminate. The c ombination of Enzygnost (indirect test with synthetic peptides) and We llcozyme (sandwich test with recombinant and synthetic peptides) Anti- HIV 1 + 2 EIAs was 100% specific when used for analysis of 9111 sera. One of 30,323 HIV-1 antibody negative sera tested was initially reacti ve on both Enzygnost Anti-HIV 1 + 2 and Wellcozyme Recombinant Anti-HI V-1 EIA (competitive assay) but was found to be negative by repeated t esting, resulting in a specificity of 100% for that combination of ELI SAs. Abbott Recombinant Anti-HIV-1 EIA. (indirect assay) combined with Wellcozyme Recombinant Anti-HIV-l EIA was initially falsely reactive with 12 of 8272 sera of which 6 were repeatedly reactive. Conclusions: This large-scale evaluation demonstrates that combinations of two ELI SAs based on different test principles and antigens increase the accur acy of the HN antibody determination and could be used as an alternati ve or complement to WB.