OBJECTIVE To determine the sensitivity and specificity of human immunodeficiency virus (HIV)-specific IgA for vertically transmitted HIV infection, particularly during the first month of life. METHODS Prospective cohort study of 140 infants born to HIV-seropositive women in a large urban teaching hospital and of 248 older infants and children referred for diagnosis and treatment of HIV infection. METHODS The HIV-specific IgA immunoblot results were compared with the infection status of patients as determined by Centers for Disease Control and Prevention (Atlanta, Ga) criteria or by sequential early diagnostic assays for HIV. Sensitivity, specificity, and predictive values were calculated for each age range. RESULTS Among infants studied from birth, the rate of vertical transmission of HIV was 21.6% (25/116). The sensitivity of HIV-specific IgA for the first month of life was 8.0% (2/25), and the specificity was 90.1% (82/91). Sensitivity increased progressively during the first year of life, and the negative predictive value was 94.6% by 6 to 8 months of age. The positive predictive value of this assay was 18.2% for neonates but was 96% to 100% after the first month of life. CONCLUSIONS False-positive test results for HIV-specific IgA occurred with diminishing frequency during the first 4 weeks of life, and the frequency of detectable HIV-specific IgA was similar among the HIV-infected and uninfected groups at this age. Beyond 1 month of age, detection of HIV-specific IgA is highly specific and is a useful serum-based assay for early diagnosis of HIV infection. These results suggest that maternal-fetal transfusion is common and support the hypothesis that the majority of maternal-fetal transmission of HIV occurs around the time of parturition.