Concordance of PCR and antibody results from HIV testing of injecting drug users

Abstract

Standard HIV-1 testing relies on the enzyme immunoassay (EIA) for detecting antibodies specific to HIV-1. This technique may misclassify persons as HIV-1-negative in instances where testing follows infection but precedes development of antibody to HIV-1. To evaluate the occurrence of HIV infection in the absence of positive antibody, polymerase chain reaction (PCR) for viral DNA in the blood has been applied. Research comparing these two testing techniques has generally focused on populations of homosexual and bisexual men. This study compares PCR and antibody testing of 337 injecting drug users recruited from street settings in San Francisco. Of 286 HIV-1 antibody-negative samples, 3 (1.0%) were PCR-positive. Of 49 HIV-1 antibody-positive samples, 1 (2.0%) was PCR-negative. Two samples were antibody-indeterminate and PCR-negative. This yielded an overall concordance of 331/335 (98.8%), excluding the indeterminate results. These results suggest that current antibody methodology is adequate. However, misclassification among recently infected individuals may occur, which is of concern in high-incidence groups