An intravenous immunoglobulin (IVIG) preparation was evaluated prospectively in hospitalized low birth weight infants for the prevention of respiratory virus infection. Premature neonates were evaluated from October 19, 1987, through July 31, 1988. Nasopharyngeal secretions were cultured weekly for viruses and clinical information was obtained daily on each infant. Ninety-one infants with birth weights between 500 and 1750 g were randomized to receive either IVIG, 500 mg/kg (46 infants), or 5% albumin-normal saline (placebo), 10 ml/kg (45 infants), between Days 3 and 7 of life, 7 days later and every 14 days thereafter for a maximum of 5 doses. Demographic and life event data during pregnancy were similar for IVIG and placebo groups. Birth weight, gestational age, gender, age at entry into the study and incidence of respiratory distress syndrome at birth were also similar in both groups of premature infants. Twenty-six viruses were isolated from 25 infants. There were 13 and 12 infections in the IVIG and placebo groups, respectively. Severity of disease, as measured by clinical factors and outcomes of virus-infected infants were no different in IVIG-treated and placebo groups. Adenoviruses and cytomegalovirus accounted for 57.7 and 23.1%, respectively, of the viral isolates. In this study the use of IVIG did not prevent or modify adenovirus and cytomegalovirus infections in premature infants.