Infusions of 50 mg of sulbactam per kg per day and 400 mg of ampicillin per kg per day in divided doses to infants and children with bacterial meningitis produced levels in cerebrospinal fluid approximately one-third those in serum. Concentrations in cerebrospinal fluid of 5.5 ,ug of sulbactam per ml and 16.0 ,ug of ampicillin per ml declined within a few days of therapy to 1.9 ,g of sulbactam per ml and 5.2 ,ug of ampicillin per ml.Haemophilus influenzae, Neisseria meningitidis, and Streptococcus pneumoniae are the leading pathogens causing bacterial meningitis in children (16,22). For years, ampicillin was the mainstay of therapy (4). However, its effectiveness was compromised by the emergence (9, 11, 21) and increasing incidence (8,16,18,19,21) of strains of H. influenzae which produce 1-lactamase enzymes that hydrolyze ampicillin, destroying its activity.Sulbactam, a ,-lactamase inhibitor which combines with ,-lactamases to destroy their activity (10) and prevent the destruction of P-lactams, is under clinical development for coadministration with ampicillin for treatment of infections. The activity of sulbactam plus ampicillin against ampicillinresistant strains of H. influenzae and S. pneumoniae (1,3,12,13,23) and the enhanced penetration of sulbactam through inflamed meninges (7,17) suggested that this combination might be useful in the treatment of bacterial meningitis in children.Therefore, we examined the efficacy of ampicillin plus sulbactam in the treatment of meningitis in infants and young children and compared this therapy with the standard therapy of ampicillin plus chloramphenicol. The results (reported previously on 41 of 53 patients receiving sulbactam plus ampicillin [14]) showed eradication of bacteria, including four ampicillin-resistant pathogens, from cerebrospinal fluid (CSF) from 34 of 35 evaluable patients during therapy with ampicillin plus sulbactam. As a part of this study, we also examined the penetration of sulbactam and ampicillin into CSF. range, 1 to 13 doses]). Sample set 2 was obtained 1 to 5 (median, 3) days later and sample set 3 was obtained 6 to 11 (median, 8) days after sample set 1. Serum samples were also obtained at the time of the first sampling. Because of the nature of the patient population, sufficient volume to run the desired assays was sometimes not available. The samples were frozen until assays for ampicillin by bioassay (6) and for sulbactam by gas chromatography (5, 6) were done. Each assay was free from interference by the coadministered drug and had a coefficient of variation of less than 10% for quadruplicate assays. The few CSF samples with unusually high drug concentrations were correlated with very high concentrations in serum, suggesting that the high concentrations in CSF were not due to contamination.
MATERIALS AND METHODSComparisons were made by means of paired t tests, whenever possible, to reduce the extensive variability expected from this patient population. Group t tests were used otherwise.
