Abstract
BACKGROUND: High mortality rates among asymptomatic cryptococcal antigen (CrAg)-positive patients identified through CrAg screening, despite preemptive fluconazole treatment, may be due to undiagnosed cryptococcal meningitis. METHODS: Symptoms were reviewed in CrAg-positive patients identified by screening 19233 individuals with human immunodeficiency virus infection and CD4 cell counts <100/µL at 17 clinics and 3 hospitals in Johannesburg from September 2012 until September 2015, and at 2 hospitals until June 2016. Cerebrospinal fluid samples from 90 of 254 asymptomatic patients (35%) and 78 of 173 (45%) with headache only were analyzed for cryptococcal meningitis, considered present if Cryptococcus was identified by means of India ink microscopy, culture, or CrAg test. CrAg titers were determined with stored blood samples from 62 of these patients. The associations between blood CrAg titer, concurrent cryptococcal meningitis, and mortality rate were assessed. RESULTS: Cryptococcal meningitis was confirmed in 34% (95% confidence interval, 25%-43%; 31 of 90) of asymptomatic CrAg-positive patients and 90% (81%-96%; 70 of 78) with headache only. Blood CrAg titer was significantly associated with concurrent cryptococcal meningitis in asymptomatic patients (P < .001) and patients with headache only (P = .003). The optimal titer for predicting cryptococcal meningitis was >160 (sensitivity, 88.2%; specificity, 82.1%); the odds ratio for concurrent cryptococcal meningitis was 34.5 (95% confidence interval, 8.3-143.1; P < .001). CONCLUSIONS: About a third of asymptomatic CrAg-positive patients have concurrent cryptococcal meningitis. More effective clinical assessment strategies and antifungal regimens are required for CrAg-positive patients, including investigation for cryptococcal meningitis irrespective of symptoms. Where it is not possible to perform lumbar punctures in all CrAg-positive patients, blood CrAg titers should be used to target those most at risk of cryptococcal meningitis.