Multicenter study confirms validity of bacterial meningitis prediction rule

March 1, 2007

The widespread introduction of highly effective bacterial conjugate vaccines against Haemophilus influenzae type b and Streptococcus pneumoniae has significantly decreased the incidence of pediatric bacterial meningitis, reducing the probability that a child with cerebrospinal fluid (CSF) pleocytosis has bacterial meningitis. As a result, investigators who several years ago developed a clinical prediction rule, the Bacterial Meningitis Score (BMS), set out to perform a large-scale revalidation study of the rule in the context of reduced incidence of the disease. The BMS classifies a patient as at very low risk of bacterial meningitis if he or she lacks all of the following criteria: positive CSF Gram stain, CSF absolute neutrophil count of at least 1,000 cells/mL, CSF protein of at least 80mg/dL, a peripheral blood absolute neutrophil count of at least 10,000, and a history of seizure before or at the time of presentation.

Investigators analyzed data from a network of 20 US emergency departments at academic medical centers that are associated with the Pediatric Emergency Medicine Collaborative Research Committee of the AAP. The coinvestigators from these institutions reviewed the medical records of all patients aged 29 days to 19 years who were brought to the emergency department with CSF pleocytosis, and had not received antibiotic treatment before lumbar puncture. Of 3,295 such children, 121 (3.7%) had bacterial meningitis and 3,174 (96.3%) had aseptic meningitis. Of the 1,714 children categorized as at very low risk for bacterial meningitis because they met none of the BMS criteria, only two (0.1%) actually had bacterial meningitis, and both were younger than 2 months. Investigators concluded that these results validate the BMS prediction rule as an accurate decision-support tool (Nigrovic LE et al: JAMA 2007;297:52).

Commentary