Risk factors for cryptococcal meningitis in HIV-infected patients.

Abstract

To identify the risk factors for cryptococcal meningitis in patients with HIV disease we conducted a nested case-control study of 37 incident cases of cryptococcal meningitis and 74 controls, identified from a cohort of more than 2000 HIV-infected patients. Conditional logistic regression was used to study demographic and AIDS-related variables in addition to fluconazole and steroid use. No difference in demographic variables, HIV risk factors, or stage of AIDS was detected between cases and controls. Exposure to fluconazole for more than 90 days reduced the risk of cryptococcal meningitis by 82% (OR=0.18; 95% CI=0.04-0.85; p=0.03). We did not find a difference in steroid use between cases and controls for either the length or amount of steroid exposure (p=0.41). No difference in survival during follow-up in the clinic was observed by the log-rank test (p=0.74). Among the cases, a cryptococcal antigen was positive in more than 97% of the CSF or blood samples. CSF and blood cultures were positive in 81 and 44% of the samples, respectively. We conclude that demographic factors did not affect the risk of cryptococcal meningitis in an inner city United States population. While fluconazole use has a protective effect, steroid use was not associated with an increased risk of cryptococcal meningitis in HIV-infected patients.