A survey of infants with neonatal sepsis in U.S. Army hospitals.

Data on 109,312 singleton births at U.S. Army hospitals over 3 years were examined to determine differences in perinatal risk factors between infants with a diagnosis of sepsis confirmed by blood culture and those whose cultures remained negative. The incidence of confirmed sepsis (1.1/1000 live births) and the importance of perinatal complications and prematurity were consistent with other reported findings. No reliable indicators of bacteremia that could be used in conjunction with other clinical data were found. These findings support continued empiric therapy in infants at risk until infection can be verified by culture.