The Bacterial Meningitis Score correctly identified acute bacterial meningitis in 884 out of 889 affected children, indicating that the tool, while 99.6% sensitive, still missed cases, the authors of a French study reported.
“As with any rule, clinicians must remain aware that no rule can completely eliminate the possibility of bacterial meningitis and that cases of bacterial meningitis can occur among patients without [cerebrospinal fluid] pleocytosis as well as among rare patients with CSF pleocytosis and a [Bacterial Meningitis Score] indicating low risk,” Dr. Francois Dubos of Saint Vincent de Paul Hospital in Paris and his colleagues warned.
The investigators based their conclusion on a secondary analysis of data from children seen in emergency departments in France over a 4-year period. They undertook their study because of concerns regarding the tool's sensitivity.
The Bacterial Meningitis Score (BMS) was developed by the Pediatric Emergency Medicine Collaborative Research Committee of the American Academy of Pediatrics (AAP), and its specificity has been well validated among more than 3,000 patients (Pediatrics 2002;110:712–9).
The sensitivity of the BMS, however, was evaluated only in 196 patients from the original test population and external populations, mainly in the United States,” they wrote. The current study was undertaken to evaluate the sensitivity of the BMS in a large population of children with bacterial meningitis in a country with a different epidemiology of the disease (J. Pediatr. 2008;152:378–82).
On the basis of the BMS, children with CSF pleocytosis are considered to be at very low risk of bacterial meningitis if they lack all of the following criteria: positive CSF Gram stain, CSF absolute blood neutrophil count (ANC) of at least 1,000 cells/mm3, CSF protein of at least 80 mg/dL, peripheral blood ANC of at least 10,000 cells/mm3, and a history of seizure before or at the time of presentation.
Because implementation of routine immunization for Haemophilus influenzae type b and pneumococcus has further reduced the chances that a febrile child with CSF pleocytosis will have bacterial meningitis rather than aseptic meningitis, the committee validated the BMS prediction rule as an accurate decision support tool in the era of these vaccines (JAMA 2007;297:52–60).
Dr. Dubos and his associates used data obtained from the Bacterial Meningitis French Surveillance Network to identify children who presented to French emergency departments between January 2001 and February 2005.
The study patients were between age 29 days and 18 years and had a diagnosis of acute bacterial meningitis, but no known neurosurgical disease, no known immunosuppression, no CSF culture revealing an organism usually associated with contamination, no clinical sepsis, no purpura fulminans, and no CSF red blood cell count greater than 10,000/mm. The mean age of the 889 patients was 3 years; the death rate was 6%, and the primary bacterial pathogens were Neisseria meningitidis (44%) and Streptococcus pneumoniae (42%).
With respect to four of the five BMS criteria, 14% of the patients had a history of seizure with the illness, 89% had a positive CSF Gram stain, 84% had a CSF protein level of 80 mg/dL or higher, and 55% had a CSF neutrophil count of 1,000 cells/mm3 or higher.
The blood neutrophil count was not collected by use of the standardized network data form, so the authors retrospectively collected data for this variable by searching the medical records of the 23 patients who did not have a bacterial meningitis diagnosis based on the other four variables. For 18 of the 23 patients, the ANC was 10,000 cells/mm3 or higher. “Thus, the sensitivity of the BMS was 99.6%,” they wrote.
“The main limitation of our study is that not all cases of [bacterial meningitis] were included in the database of the Bacterial Meningitis French Surveillance Network,” the researchers reported. “The exhaustiveness of this database for cases of pneumococcal meningitis has been estimated to be 61% and is unknown for cases of infection with other pathogens.”
Additionally, the data for the study were collected when immunization with pneumococcal conjugate vaccine was implemented in France. The estimated vaccination coverage at the end of the study period was 41%, and the vaccination coverage against meningococcus C during that time was very low.
“The rule might have different results in areas with differing vaccination patterns for pneumococcus and meningococcus,” the investigators stated.
Prospective studies in clinical practice are needed to assess whether use of the BMS can safely reduce hospital admissions and antibiotic use for children with aseptic meningitis, they concluded.