A Scoring Index for Acute Bacterial Meningitis: Implications for the Population-Based Assessment of Hib, Meningococcal and Pneumococcal Vaccine Impact

A Scoring Index for Acute Bacterial Meningitis: Implications for the Population-Based Assessment of Hib, Meningococcal and Pneumococcal Vaccine Impact

Background. Acute bacterial meningitis (ABM) causes high mortality and severe sequelae. Yet, in many countries, a paucity of microbiology diagnostic facilities limit identification of children with treatable bacterial infections. Prospective surveillance for ABM in Vietnam provided an opportunity to develop a clinical score for bacterial meningitis among children aged 1--59 months.

Methods. We randomly allocated 1,773 children with suspected ABM into two cohorts (derivation and validation). Independent predictors of ABM were identified by logistic regression. The resulting scoring index was evaluated in the validation cohort for its accuracy and ability to rule-in and rule-out acute bacterial meningitis.

Results. A 10-point score for ABM consisting of cerebrospinal fluid (CSF) white blood cell count (score: 4 if >100 cells/mm3), age (score: 3 if <24 months of age), CSF glucose concentration (score: 2 if <40 mg/dl) and CSF protein concentration (score: 1 if >100 mg/dl) was validated and yielded an area under the receiver operating characteristic curve of 0.9050 (95% CI: 0.9044 to 0.9056). Two cut points were identified that resulted in three risk levels for ABM: low (score: 1 to 3), moderate (4 to 7) and high (8 to 10).

Conclusions. The performance of this index suggests it may be useful in the assessment of Hib, meningococcal and pneumococcal conjugate vaccine impact at the population level. However, the score's present accuracy limits its application in clinical settings where near 100% accuracy is required. To improve the score's clinical application, prospective studies using additional data from sensitive and specific rapid diagnostic tests are needed.