Abstract

AIMS--To determine the effectiveness of intravenous immunoglobulin administration to premature infants in the prevention and/or treatment of bacterial infection. METHODS--Computer searches of MEDLINE, EMBASE, SCISEARCH and Oxford Database of Perinatal Trials were made. Two independent researchers applied inclusion criteria of: randomised controlled trial; premature and/or low birthweight infant; use of intravenous immunoglobulin; and infection or mortality. Nineteen of 44 identified studies fulfilled these criteria. Study quality was assessed and information on study population, intervention, and outcomes were collected. RESULTS--Studies were divided into prophylaxis or treatment; results were tabulated for infection, sepsis, and death from all causes. For 17 studies of prophylaxis (n = 5245), the relative risk and confidence interval were, for proved infection 0.81, 0.67-0.97; for sepsis 0.87, 0.66-1.13; for death from all causes 0.85, 0.64-1.14. Some outcome results were heterogeneous. Two treatment studies showed no reduction in mortality when combined. CONCLUSIONS--Routine administration of intravenous immunoglobulin to preterm infants is not recommended.