Specificity of antibody tests for human immunodeficiency virus in alcohol and parenteral drug abusers with chronic liver disease.

Abstract

Parenteral drug abusers are at risk for acquired immunodeficiency syndrome (AIDS), which is caused by human immunodeficiency virus (HIV). We tested stored sera for antibody to HIV (anti-HIV) using two enzyme-linked immunosorbent assay (ELISA) methods and Western blot. The patients were parenteral drug abusers who had undergone percutaneous liver biopsy for chronic liver disease. Current or former alcohol abuse was noted in 88 (80%) of the 110 patients. The sensitivities of the two ELISA tests in comparison with Western blot, the more specific test for HIV, were 100 and 94%, respectively; the specificities were 94 and 99%. Western blot was positive in 36 (33%) of 110 patients. False-positive ELISA reactions for anti-HIV were seen in five (7%) of 70 patients with negative Western blot analyses. Compared to true-negatives, false-positives had significantly more years of alcohol abuse, younger ages of onset of alcohol abuse, greater frequencies of jaundice and edema, higher levels of alkaline phosphatase, total billirubin, total protein, and globulins, and lower levels of serum albumin. In a stepwise logistic regression, only hyperglobulinemia was significantly associated with a false-positive anti-HIV. We conclude that: (a) ELISA tests for anti-HIV are useful for screening abusers of alcohol and parenteral drugs with chronic liver disease for HIV infection, but positive results must be confirmed with more specific tests such as Western blot; (b) false-positive ELISA reactions in this population are associated with hyperglobulinemia; and (c) studies of HIV testing are needed in other populations of patients with alcoholism or liver disease.