Models of predicting the risk of brain herniation in bacterial meningitis

Abstract number: P1225

Meyer C.N., Augustesen S.

Objectives: The purpose of our study was to evaluate models predicting a risk of brain herniation among patients with verified acute bacterial meningitis. These models are used in daily practise to indicate when to do a head CT scanning before lumbar puncture; ideally thus preventing brain herniation caused by accelerated brain shift following the spinal tap.

Methods: The predictive models originated from North-American, British, and the Dutch guidelines (Tunkel-AR 2004, BIS algoritme 2004 (similar to Fitch-MT 2007), van Crevel-H 2002). On a national basis, unselected patients with microbiologically and clinically verified acute bacterial meningitis from 47 hospitals during 2 years (n = 320) were included. Survival data and clinical data from the medical records were evaluated retrospectively. Two-tailed p < 0.05 indicated a significant difference.

Results: In 5 patients, brain herniation occurred. Among the 316 patients with available sufficient clinical data, 85% fulfilled the North-American model for early CT-scanning (before lumbar puncture), but in practice only 19% (52/269) fulfilling the criteria were handled accordingly. The British model was fulfilled by 33% of the patients, but only 32% of these (35/111) were handled accordingly. And very similar numbers were found in the Dutch model (31% and 34% of these, respectively).

Antibiotics were given before lumbar puncture in 30% (6/20) of patients suspected of meningitis, who were sent to early (before lumbar puncture) CT scanning.

Conclusions: According to North-American guidelines, most of the patients with meningitis should have had a CT-scan done before lumbar puncture, though according to Dutch or British guidelines this was recommended in much fewer patients. In a majority of suspected meningitis patients, timely therapy before early CT-scan was not given. Unnecessary CT scanning may still delay relevant treatment and thus have detrimental effect in acute bacterial meningitis.