P D Friedmann1, M H Samore. 1. Department of Medicine, University of Chicago Pritzker School of Medicine, IL, USA.
Abstract
BACKGROUND: When human immunodeficiency virus (HIV)-infected patients present with neurologic problems, lumbar puncture and cerebrospinal fluid (CSF) analysis are often performed, usually to rule-out the presence of meningitis. However, the test characteristics of CSF analyses in this population are unknown. METHODS: To examine the diagnostic characteristics of CSF analyses for secondary causes of meningitis in HIV-infected adults, we performed a case-control study of 322 HIV-infected adults who received diagnostic lumbar puncture in an urban academic hospital from 1989 to 1992. Using multivariable logistic regression analyses, we determined the independent CSF correlates of cryptococcal and other types of secondary meningitis. RESULTS: Eighty percent were men who had sex with men, 89% were white, and 85% had AIDS. Thirty-seven case-subjects had a secondary cause of meningitis though none had bacterial or tuberculous meningitis. Excluding the india ink as a candidate variable, logistic regression analysis identified 2 independent CSF correlates of secondary meningitis: the leukocyte count (odds ratio [OR] per 10 cells, 1.2; 95% confidence interval [CI], 1.1 to 1.3), and the CSF-to-serum glucose ratio (OR per 0.10 units, 0.61; CI, 0.43 to 0.88). The areas (+/- SD) under the receiver operating characteristic curves were 0.76 +/- 0.05 and 0.72 +/- 0.05, respectively. CONCLUSIONS: In similar clinical settings, these findings may help clinicians assess the probability of secondary meningitis in HIV-infected patients when initial CSF results first become available.
BACKGROUND: When human immunodeficiency virus (HIV)-infectedpatients present with neurologic problems, lumbar puncture and cerebrospinal fluid (CSF) analysis are often performed, usually to rule-out the presence of meningitis. However, the test characteristics of CSF analyses in this population are unknown. METHODS: To examine the diagnostic characteristics of CSF analyses for secondary causes of meningitis in HIV-infected adults, we performed a case-control study of 322 HIV-infected adults who received diagnostic lumbar puncture in an urban academic hospital from 1989 to 1992. Using multivariable logistic regression analyses, we determined the independent CSF correlates of cryptococcal and other types of secondary meningitis. RESULTS: Eighty percent were men who had sex with men, 89% were white, and 85% had AIDS. Thirty-seven case-subjects had a secondary cause of meningitis though none had bacterial or tuberculous meningitis. Excluding the india ink as a candidate variable, logistic regression analysis identified 2 independent CSF correlates of secondary meningitis: the leukocyte count (odds ratio [OR] per 10 cells, 1.2; 95% confidence interval [CI], 1.1 to 1.3), and the CSF-to-serum glucose ratio (OR per 0.10 units, 0.61; CI, 0.43 to 0.88). The areas (+/- SD) under the receiver operating characteristic curves were 0.76 +/- 0.05 and 0.72 +/- 0.05, respectively. CONCLUSIONS: In similar clinical settings, these findings may help clinicians assess the probability of secondary meningitis in HIV-infectedpatients when initial CSF results first become available.