TERM NEWBORNS WHO ARE AT RISK FOR SEPSIS - ARE LUMBAR PUNCTURES NECESSARY

Objectives. To determine: (1) whether a lumbar puncture (LP) is indica ted in asymptomatic full-term newborns delivered by mothers at risk of intrapartum sepsis; and (2) whether gentamicin improves bacterial cov erage for such newborns when used with ampicillin. Design. A retrospec tive chart review from 1987 through 1993 of all newborns with positive blood and/or cerebrospinal fluid cultures in the first 7 days of life . Methods. Pregnant women were screened in the second trimester for gr oup B streptococci and given ampicillin during labor if two or more ri sk factors were present: group B streptococci colonization, maternal f ever or leukocytosis, rupture of membranes at more than 18 hours, foul -smelling amniotic fluid, and fetal tachycardia. After sepsis evaluati on (LP, blood culture, white blood cell count, and differential), asym ptomatic infants received ampicillin and gentamicin for 48 to 72 hours unless cultures grew pathogens. Results. Of approximately 24 452 full -term births in 7 years, 7% (1712) had evaluations for symptoms of sep sis, and 14% (3423) were asymptomatic but had evaluations for maternal risk factors. There were 11 cases of meningitis, all involving sympto matic newborns; 10 of these 11 had positive blood cultures for the sam e organism. In asymptomatic infants, none of the 3423 had meningitis ( 95% confidence interval, 0 to 0.0008), although 35 grew contaminants. Of 73 pathogens isolated from blood or cerebrospinal fluid, 7 (9.5%) w ere resistant to ampicillin. Addition of gentamicin provided coverage for only 2 of these 7 pathogens. Of 5135 infants who received ampicill in and gentamicin, only 2 required gentamicin for improved coverage. C onclusions. (1) LP is unnecessary in asymptomatic full-term newborns. (2) Empiric coverage for asymptomatic newborns with maternal risk fact ors need not include gentamicin at all hospitals, because it only impr oved the coverage of ampicillin alone from 90% to 93% of pathogens, bu t it exposed more than 5000 infants to the side effects of gentamicin. (3) The presence of leukopenia (<5000 white blood cells/mm(3)) is hig hly predictive of bacteremia.