Fifty children with bacterial meningitis were prospectively randomized to receive cefotaxime (50 mg/kg/dose every 6 hours) or ampicillin and chloramphenicol in standard doses. Twenty-three patients received cefotaxime and 27 received standard therapy. Bacterial isolates included: Haemophilus influenzae (29), Streptococcus pneumoniae (eight), Neisseria meningitidis (eight), group B streptococci (three), and Salmonella enteritidis (two). Ten (34%) of the H. influenzae isolates were resistant to ampicillin, nine on the basis of beta-lactamase production. All strains were susceptible to cefotaxime. Clinical cure rates for the cefotaxime (100%) and standard therapy (96%) groups were similar; survival without detectable sequelae was similar, at 78% and 77%, respectively. The duration of therapy, 11.1 +/- 2.4 days (range 10 to 21 days) vs 11.9 +/- 3.9 days (range 10 to 21 days), and days to defervescence, 4.7 +/- 2.6 days (range 1 to 14 days) vs 5.6 +/- 2.9 days (range 2 to 17 days), were similar in the cefotaxime and standard therapy groups, respectively. No adverse drug reactions or side effects were noted in either group. Cefotaxime was found to be as safe and effective as standard therapy for the treatment of bacterial meningitis in children.