The Role of Cryptococcal Antigen Assay in Diagnosis and Monitoring of Cryptococcal Meningitis

Abstract

In a recent paper evaluating the significance of cryptococcal antigen test results for 29 Chinese human immunodeficiency virus (HIV)-negative patients affected by cryptococcal meningitis, Lu and colleagues (8) showed in all patients a decrease of antigen titer from the baseline following antifungal therapy and suggested that a decrease can be used to monitor antifungal therapy efficacy but cannot be used as an index of cure. We have reviewed our experience with 66 HIV-positive patients out of 118 with cryptococcal meningitis for whom at least three serial determinations (at baseline, day 7, and day 14) of cryptococcal antigen tests on cerebrospinal fluid (CSF) were available (1). A total of 440 determinations (range, 3 to 28 antigen determinations; median, 5 antigen determinations) were available, and for 55 patients the last determination was considered (median, 13 weeks; range, 2 to 84 weeks). In Fig. ​Fig.11 is depicted the kinetics of CSF cryptococcal antigen together with the results of CSF culture. Overall, 53 patients (80%) showed a decrease of CSF antigen titer from the baseline (7 of whom had negative results), as follows: 27 cases of 1 to 3 dilutions, 16 cases of 4 to 6 dilutions, and 10 cases of 7 or more dilutions. However, 13 out of 15 of these patients for whom postmortem examination was available, despite evidence of several intravitam negative CSF cultures, still had cryptococcal meningitis or disseminated disease at autopsy (demonstrated by histopathology). Eight patients had an increase in the CSF antigen titer (four of 1 to 3 dilutions and four of 4 to 8 dilutions), and five showed stable (i.e., the same value) results throughout the follow-up. All the patients but two with an increase of CSF antigen titer had persistent positive CSF culture and died; four underwent autopsy showing disseminated cryptococcosis. FIG. 1. Change over time of CSF cryptococcal antigen titers and correlation with CSF cryptococcal cultures in 66 HIV-positive patients. Data of CSF antigen are median values. Data regarding CSF cultures refers to the total number of patients (n = 66), ... Our experience regarding the role of cryptococcal antigen to monitor antifungal therapy in AIDS patients is in keeping with that previously reported by Powderly et al. (11), who showed the lack of any correlation of changes of CSF or serum cryptococcal antigen and the outcome of cryptococcal meningitis. However, a high CSF antigen level has been identified as a sign of poor prognosis in patients with AIDS (1, 7); interestingly, more recently Thay cohorts of HIV-positive patients with cryptococcal meningitis showed a significant positive correlation between CSF cryptococcal colony-forming units (CFU) and CSF cryptococcal antigen titers at baseline (P < 0.0001), but the rapid rate of decline in CSF CFU was not correlated with that in CSF cryptococcal antigen titers (2). As shown in Table ​Table1,1, regardless of the different hosts in whom cryptococcal meningitis is diagnosed, among all methods employed the cryptococcal CSF antigen had the best overall sensitivity (94.1%) followed by the serum antigen (93.6%). However, some differences were observed in the different categories of hosts, with lower sensitivity in AIDS and immunocompetent patients (92%) and higher sensitivity among the other immunocompromised hosts without HIV infection. Persistently elevated CSF cryptococcal antigen in HIV-infected patients carries a poor prognosis and indicates ongoing production of viable Cryptococcus neoformans in tissue. In conclusion, CSF cryptococcal antigen seems to be the best test for diagnosis of cryptococcal meningitis in terms of sensitivity, but it is an unreliable tool, at least among HIV-positive patients, to drive therapeutic monitoring, particularly in assessing the point of discontinuation of antifungal therapy in HIV-infected patients. TABLE 1. Efficiency of different techniques in the diagnosis of cryptococcal meningitis in different hostsa