Summary Background New antimicrobials or adjunctive treatments have not substantially reduced mortality from acute childhood bacterial meningitis. Paracetamol seems to have beneficial effects in bacteraemic adults and some experts recommend initial slow β-lactam infusion. We investigated whether these treatments had benefits in children with bacterial meningitis. Methods We did a prospective, double-blind, single-centre study with a two-by-two factorial design in Luanda, Angola. 723 participants aged 2 months to 13 years were randomly assigned two 12 h intravenous infusions, without loading doses, of 125 mg/kg bodyweight cefotaxime (total dose 250 mg/kg) given over 24 h, or 250 mg/kg bodyweight cefotaxime given as four boluses, one every 6 h over 24 h. Patients also received oral paracetamol at an initial dose of 30 mg/kg then 20 mg/kg every 6 h for 48 h or placebo. Two primary endpoints, death or severe neurological sequelae and deafness, were analysed by intention to treat. The study was registered as ISRCTN62824827. Findings 183 patients were assigned cefotaxime infusion plus paracetamol and 180 patients to each of the other three treatment groups. Causative agents were identified in 63% of cases and were mostly Haemophilus influenzae type b, Streptococcus pneumoniae, or Neisseria meningitidis. Death or severe neurological sequelae were seen in 340 (47%) of 723 children and deafness in 45 (12%) of 374 tested, both distributed similarly across treatment groups. In a predefined subgroup analysis of death or any sequelae, by causative agent, a benefit was seen in favour of infusion over bolus in children with pneumococcal meningitis (infusion plus placebo, odds ratio 0·18, 95% CI 0·03–0·90, p=0·04). A similar effect was seen for children receiving cefotaxime infusion plus paracetamol, but the difference was not significant (OR 0·22, 95% CI 0·04–1·09, p=0·06). A post-hoc analysis suggested that cefotaxime infusion plus paracetamol lowered mortality at least during the first 3 days, irrespective of cause. Interpretation Although no tested regimen improved the final outcomes of these very ill children, studies of longer courses of β-lactam infusion plus paracetamol seem warranted. Funding The Päivikki and Sakari Sohlberg, the Sigrid Jusélius, and the Paediatric Research Foundations, and the daily newspaper Helsingin Sanomat.