To ascertain the sensitivity, specificity, predictive value, and clinical use of a human immunodeficiency virus (HIV)-IgA immunoblot assay for diagnosing perinatal HIV infection in infants tested at birth to 1 month, 3 months, and 6 months of age. Prospective, longitudinal cohort study of children born to HIV-infected and noninfected women. The HIV-IgA immunoblot assays were performed at birth to 1 month, 3 months, and 6 months of age and compared with the Centers for Disease Control's classification system of HIV infection in the children. Children were followed up for at least 15 months to ensure accuracy of infection status. Municipal hospital in central Brooklyn, NY, where the prevalence of HIV infection is high. Serum samples from 58 children, 22 with documented HIV infection, 18 noninfected children born to seropositive women, and 18 children born to noninfected women, were studied. Diagnosis of HIV infection using the Centers for Disease Control's classification scheme was compared with diagnosis using the HIV-IgA immunoblot assay for children 6 months of age or younger. The HIV-IgA immunoblot assay yielded negative results at 3 and 6 months of age for all 18 infants born to seronegative women; for the 18 seroreverting, noninfected children born to infected women, the assay yielded negative results at 1 month, 3 months, and 6 months of age. The positive predictive value of the assay was 100%--no false-positive results were identified in the 88 serum samples obtained from noninfected infants. For the HIV-infected children, sensitivity was a function of age: one (5.9%) of 17 infants had an assay that yielded positive results at birth to 1 month of age, 13 (62%) of 21 infants had assays that yielded positive results at 3 months of age, and 17 (77%) of 22 infants had assays that yielded positive results at 6 months of age. The presence or absence of symptoms did not affect the sensitivity. The HIV-IgA immunoblot assay can detect a significant proportion of infected children during an early asymptomatic period of their life. This relatively inexpensive, easily standardized assay may allow for institution of therapy before the onset of clinical symptoms.