Nosocomial Meningitis: Are We too Suspicious?


Synopsis: None of 51 nonsurgical patients who underwent lumbar puncture because of suspected nosocomial meningitis had abnormal cerebrospinal fluid.

Source: Metersky ML, et al. Retrospective analysis: Are fever and altered mental status indications for lumbar puncture in a hospitalized patient who has not undergone neurosurgery? Clin Infect Dis 1997;25:285-288.

Metersky and colleagues at the university of Connecticut retrospectively reviewed all patients who had a sample of cerebrospinal fluid (CSF) sent to the microbiology laboratories of two teaching hospitals. The records of all such nonsurgical patients older than 15 years of age whose lumbar puncture (LP) had been performed to rule out acute central nervous system (CNS) infection were then further examined.

During the 20 months examined, 51 patients underwent LP because of suspected nosocomial meningitis. Forty-seven percent were febrile, 78% had had a change in mental status, and only 22% had headache or meningeal signs. The CSF from all 51 patients was normal.

When data were included from a group of 181 patients with suspected community-acquired meningitis, of whom 26 (14%) proved to have this diagnosis, the absence of headache or meningeal signs had a negative predictive value of 0.98. The presence of either had a sensitivity for the presence of meningitis of 0.92.


Only in a teaching hospital would so many non-neurosurgical patients be sufficiently suspected of having nosocomial meningitis as to undergo a LP. As the authors point out, the National Nosocomial Infections Surveillance System reported a rate of such infections on medical services of seven per 100,000 admissions, but even some of these were undoubtedly associated with procedures that may have contaminated the central nervous system.

Although the authors did not provide confidence intervals, one must be concerned about the low yield of LPs in the paper reviewed here vis-à-vis the low, but real risk of this procedure. These risks are discussed both in the paper and in an accompanying editorial by Tauber. Nonetheless, since nosocomial meningitis does, on rare occasion, occur in the medical patient whose CNS has not been invaded, LPs are appropriate and necessary in some cases. This study, although retrospective and small, suggests that the risk of such infection in the absence of headache or meningeal signs is quite low. We are left with the necessity of clinical judgment, which can only be acquired through experience.