May 23, 1994

Risk Factors for Repeatedly Reactive HIV-1 EIA and Indeterminate Western BlotsA Population-Based Case-Control Study

Author Affiliations

From the Departments of Medicine (Drs Celum, Coombs, Corey, Inui, Wener, and Holmes, and Ms Murphy), Laboratory Medicine (Drs Coombs, Corey, and Wener), and Biostatistics (Ms Jones and Dr Fisher), and the Center for AIDS Research (Ms Murphy and Dr Holmes), University of Washington, Seattle; and the Pacific Northwest Research Foundation, Seattle (Dr Grant). Dr Grant is now with Custom Monoclonals, Sacramento, Calif. Dr Inui is now with the Department of Ambulatory Care and Prevention, Harvard Medical School, Boston, Mass.

Arch Intern Med. 1994;154(10):1129-1137. doi:10.1001/archinte.1994.00420100115015

Objective:  Causes of indeterminate results of Western blot testing (IWB) for human immunodeficiency virus (HIV) type 1 include seroconversion, HIV-2 cross-reactivity, and autoimmune disease, but most IWB results remain unexplained. This case-control study assessed risk factors for IWB results, including early HIV infection, other retroviral infection, autoantibodies, and other medical conditions.

Design:  Prospective study to determine HIV seroconversion rate, with a case-control design to assess other risk factors for IWB. Cases (persons with one or more repeatedly reactive HIV-1 enzyme immunoassay with IWB), their current sexual partners, and controls (persons with negative enzyme immunoassay and Western blot results) were recruited from blood banks, health department and prenatal clinics, and private providers in Washington and Oregon.

Results:  Of 244 cases enrolled, 206 were followed up for 6 months or longer, and six (3.0%; 95% confidence interval [CI], 0.7% to 5.3%) with recent HIV risk behaviors seroconverted. The Western blot banding patterns differed among groups; cases usually had p17 or p24 bands, while controls and cases' sexual partners usually had polymerase bands. Conditional logistic regression indicated that independent risk factors for IWB among male cases and controls were a tetanus booster in the past 2 years (odds ratio, 3.2; 95% CI, 1.2 to 8.6) and sexual contact with a prostitute (odds ratio, 3.0; 95% CI, 1.0 to 9.5). Independent risk factors for women were parity (odds ratio, 1.2;95%CI, 1.02 to 1.4) and autoantibodies, either rheumatoid factor or antinuclear antibodies (odds ratio, 2.3; 95% CI, 1.03 to 5.6). No cross-reactivity was detected with HIV-2, human T-lymphotrophic virus type 1, feline immunodeficiency or feline leukemia, or bovine immunodeficiency viruses.

Conclusions:  Evaluation of persons with reactive HIV-1 enzyme immunoassays and IWB should include an assessment of HIV risk and other possible risk factors, such as alloimmunization (ie, parity or recent immunization) or autoantibodies (ie, antinuclear antibodies and rheumatoid factor). The relationship of IWB among men who reported sex with prostitutes is intriguing and warrants further study.(Arch Intern Med. 1994;154:1129-1137)