False-positive results in the Elecsys HIV combi PT assay: A retrospective analysis of contributing factors and cutoff index values

Fourth-generation assays detecting human immunodeficiency virus (HIV)-1/2 antibodies and the p24 antigen are widely used for HIV screening; however, false-positive (FP) results remain a limitation, particularly in low-prevalence settings. This study aimed to identify factors associated with false positivity in the Elecsys HIV combi PT assay and to evaluate the relationship between cutoff index (COI) values and confirmed HIV infection. A total of 36,046 individuals were screened using the Elecsys HIV combi PT assay. Serum samples were analyzed for anti-HIV antibodies and p24 antigen using a fourth-generation electrochemiluminescence immunoassay. Repeatedly reactive samples were referred to a reference laboratory and evaluated according to the national HIV diagnostic algorithm, including VIDAS HIV-1/2 DUO Ultra, HIV-1/2 antibody differentiation assay, and HIV-1 ribonucleic acid polymerase chain reaction testing when required. Receiver operating characteristic curve analysis was performed to assess the ability of the COI values to differentiate true-positive results from presumed FP results. Presumed FP cases and nonreactive controls were evaluated to identify factors associated with false positivity. Of the screened individuals, 122 (0.33%) had repeatedly reactive screening results. Among these, 24 were confirmed HIV-positive and 98 were classified as presumed FP, corresponding to an overall false-positivity rate of 0.27%. Presumed FP results were moderately associated with concurrent infection or recent vaccination within the preceding 6 months ( P  < .05). The median COI values were significantly higher in the true-positive cases than in the presumed FP cases (458.9 and 1.75, respectively; P  < .05). Receiver operating characteristic analysis suggested that COI values demonstrated discriminatory performance within the study population, with an assay- and population-specific cutoff value of 72.085 identified. Our findings indicate that COI values may serve as a supportive laboratory parameter for interpreting screening reactivity, particularly in low-prevalence settings; however, confirmatory testing should remain essential for diagnosis.