Prediction of academic and behavioural limitations in school-age survivors of bacterial meningitis

Aim: To develop a prediction rule to identify postmeningitic children at high risk of academic and behavioural limitations. Methods: 182 children (mean age 10 y; range 5-14) were selected from a cohort of 674 school-age survivors of bacterial meningitis. These children had neither meningitis with "complex onset", nor prior cognitive or behavioural problems, nor severe disease sequelae. On average, 7 y after the meningitis, they were evaluated using an "Academic Achievement Test", and their parents filled in the "Child Behaviour Checklist". By reviewing the medical records, potential risk factors for academic and/or behavioural limitations were collected. Independent predictors were identified using multivariate logistic regression analysis, leading to the formulation of a prediction rule. Results: The cumulative incidence of academic and/or behavioural limitations among children who survived bacterial meningitis without severe disease sequelae was 32%. The prediction rule was based on nine independent risk factors: gender, birthweight, educational level of the father, S. pneumoniae, cerebrospinal fluid leukocyte count, delay between admission and start of antibiotics, dexamethasone use, seizures treated with anticonvulsive therapy, and prolonged fever. When 10 was taken as a cut-off point for the risk score computed using this rule, 76% of the children with limitations could be identified, while 38% of the children in the cohort were selected as at risk for these limitations. Conclusion: With a prediction rule based on nine risk factors, postmeningitic children at high risk of developing academic and/or behavioural limitations could be identified. Additional research is required to further validate this prediction rule. In the future, a careful follow-up of high risk children may enhance early detection and treatment of these limitations