The performance of four enzyme immunoassays, manufactured by Abbott, Diagnostics Pasteur, Genetic Systems, and Organon Teknika, for the combined detection of anti-human immunodeficiency virus type 1 (HIV-1) and anti-HIV-2, was examined in a multisite evaluation. The collaborative efforts of 7 Australian Red Cross Blood Transfusion and 12 Australian Public Health Laboratories minimized potential biases in data by providing large numbers of anti-HIV-1-negative and -positive samples. Sensitivity was estimated using samples that were positive for anti-HIV-1 from individuals known to be infected and seroconversion samples. Sensitivity estimates in the four assays were 99.71, 99.94, 99.49, and 99.68%, respectively. Specificity was measured using fresh, sequential blood donations and samples with previous false-positive reactions in other assays. Specificity estimates from blood donations were 99.92, 99.46, 99.67, and 99.85%, respectively. The data were analyzed further using the delta statistic, which distinguishes the performance of assays of similar sensitivity and specificity by providing a measure of how well results in a population of positive or negative samples are removed from the assay's cutoff value.

Australia Comparative Study False Positive Reactions Human HIV Antibodies/*ANALYSIS HIV Infections/*DIAGNOSIS HIV Seronegativity/IMMUNOLOGY HIV Seropositivity/IMMUNOLOGY HIV-1/*IMMUNOLOGY HIV-2/*IMMUNOLOGY *Immunoenzyme Techniques Predictive Value of Tests Reagent Kits, Diagnostic Sensitivity and Specificity Support, Non-U.S. Gov't CLINICAL TRIAL JOURNAL ARTICLE MULTICENTER STUDY

www.aegis.org