Cryptococcal meningitis carries a high mortality. Further understanding of immune suppression factors associated with neuroinvasive infection will improve risk stratification and enhance early diagnosis and treatment with antifungal therapy. The aim of the study was to corroborate established or find novel clinical predictors for cryptococcal meningitis. We performed a matched case-control study of Cryptococcus infection in immunocompromised patients with or without cryptococcal meningitis. Data of all patients with a diagnosis of cryptococcal disease were collected at University of Colorado Hospital between 2000 and 2015 (n = 51). Thirty patients were diagnosed with cryptococcal meningitis. We built a logistic regression model for risk factors associated with cryptococcal meningitis. The single-predictor univariate model found that a positive blood culture, positive serum cryptococcal antigen, current malignancy, and headaches were significantly associated with cryptococcal meningitis (p = 0.02). In the adjusted multivariate model, central nervous system disease was significantly associated with a diagnosis of HIV infection (OR 24.45, 95 % CI 1.62-350.37; p = 0.022) and a positive serum cryptococcal antigen test (OR 42.92, 95 % CI 3.26-555.55; p = 0.0055). In patients with HIV infection or a positive serum cryptococcal antigen, the pretest probability of neuroinvasive Cryptococcus infection is increased and an aggressive diagnostic evaluation should be conducted to exclude infection and consider empiric therapy.
Keywords: Cryptococcal antigen; Cryptococcal meningitis; Cryptococcus neoformans; Fungemia; HIV; Risk factors.