In this study the clinical and laboratory findings of 48 children with purulent meningitis were examined, prospectively, to determine the prognostic factors in childhood meningitis in a developing country. Patients were examined for the following variables: history of antibiotic use; period between onset of symptoms and hospital admission; age at presentation; sex; fever; convulsion; level of consciousness; malnutrition; anemia; leukocyte and thrombocyte counts; erythrocyte sedimentation rate; serum C-reactive protein (CRP) level; and cerebrospinal fluid (CSF) including white blood cell count; glucose, protein, and CRP concentrations; antibiotic treatment; neurological sequelae; and fatality rate during the hospital stay. Most of these parameters were re-evaluated in all patients 36-48 h after admission. Patients were divided into 3 groups: surviving without sequelae, surviving with sequelae, and not surviving (deceased). A total of 48 children, 19 girls (39.5%) and 29 boys (60.5%), aged 2 months to 13 years, were included in the study. Of the 48 patients, 29 (60.5%) survived without sequelae, 13 (27%) survived with sequelae and 6 (12.5%) died. In a comparison among groups, we found that absence of anemia, low (< 1,000) CSF white blood cell (WBC) count, and high CRP level at admission were the indicative of poor prognosis. Thirty-six to 48 h after admission, the presence of fever, depressed level of consciousness, high (> 1,000) CSF WBC count, and low CRP level were also poor prognostic factors. In addition, we observed that mortality rate was lower in the penicillin G+ chloramphenicol group than in the ampicillin-sulbactam + cefotaxime group (P < 0.05). The mean period between onset of symptoms and hospital admission was longer in the surviving with sequelae and in the not surviving groups than in the surviving without sequelae group (P < 0.05).