A comparison of peripheral blood coculture versus 18- or 24-month serology in the diagnosis of human immunodeficiency virus infection in the offspring of infected mothers. Women and Infants Transmission Study.

The Women and Infants Transmission Study (WITS) has established virologic definitions of human immunodeficiency virus (HIV)-infected and uninfected children that have been widely used but never formally compared with serologic definitions of infection. Data from the offspring of HIV-infected women in the WITS with frequent HIV cultures during the first year of life and with HIV serology at 18 and/or 24 months of age were analyzed. Seventy-seven infants were HIV-infected and 430 uninfected by both virologic and serologic criteria. Thirteen were virologically infected (> or = 2 positive cultures) but either seronegative or serologically indeterminate. All but 1 of these had clinical HIV disease at the time of analysis. In this pediatric cohort, children defined as infected by virologic criteria often (13/90) had negative or indeterminate serology despite symptoms of HIV disease. Results suggest that serology at 18-24 months has high specificity but poor sensitivity. It should not be considered the reference standard in identifying HIV infection in perinatally exposed children.