Cryptococcal meningitis in HIV uninfected patients in tropical Viet Nam is most commonly due to Cryptococcus neoformans var grubii. Most patients (81%) have no concomitant immunosuppressive disease, although CD4 lymphopenia was detected in 9 patients, 3 of whom had no other underlying conditions.
C. neoformans var grubii, the commonest cause of cryptococcal meningitis since the HIV epidemic, usually occurs in patients with an underlying immunosuppressive condition and has rarely been reported in the absence of underlying disease, in southeast Asia or elsewhere [2, 12, 14, 22–27]. Classical teaching would suggest that, in our tropical location, more cases would have been due to infection with Cryptococcus gattii [28, 29]. Against expectation, in our series C. neoformans var grubii infection outweighed C. gattii infection 2.5-fold. We could only revive isolates for speciation from 35 patients, but since the study has finished we continue to isolate C. neoformans var grubii and C. gattii from HIV uninfected patients in the ratio 3:1. We could not detect differences in the rate of underlying disease according to infecting species, although our study lacked power since there were only 10 cases of C. gattii infection.
All patients in our study had positive CSF India ink examination and culture - a rate higher than generally reported in the literature . This may be a reflection of the large volumes of CSF (5 - 10 mls) routinely taken from patients with suspected meningitis in our hospital in order to exclude tuberculous meningitis.
In the tropical Northern Territories of Australia the incidences of C. gattii and C. neoformans infection are equal. In Queensland the experience is more like our own - in a 3 year period C. neoformans infections outweighed C. gattii 5 fold, although the total number of cases is not clear from the report. However, it is not clear whether these cases were meningoencephalitis or extra-neural disease, and the varietal form of C. neoformans was not described for the majority of isolates.
All the C. neoformans var grubii strains from our series were of URA5- RFLP molecular type VN1, as also recently reported from China . In that series, there was remarkable genetic homogeneity between the VN1 isolates, perhaps representing a strain in Asia with increased ability to cause infection in the immunocompetent.
The significance of CD4 lymphopenia in our patients is uncertain. Only 20 patients had the investigation, and the test was not performed serially. Thus it is not clear whether it represents a transient phenomenon as a consequence of infection, or whether it represents a genuine pre-existing immune deficit, such as idiopathic CD4 lymphopenia, known to be associated with an increased risk of cryptococcosis . HIV infection is unlikely since each patient received at least 2 HIV antibody tests.
While cryptococcal meningitis remains rare in HIV uninfected patients, when it occurs the consequences are severe . The death rate in our patients is 19% at hospital discharge, and blindness affects 20% of survivors. Most patients were admitted to provincial or district hospitals before transfer to our centre, and delay in diagnosis may play a role in the high death rate seen in this disease, although we did not find length of history to be associated with death. The independent variables associated with death were age and convulsions. We did not find the presence of underlying disease to be independently associated with outcome, although analysis was limited by the low number of deaths. CSF opening pressure at baseline was highly associated with development of blindness, but there was no association with papilloedema, which probably reflects the fact that papilloedema is a late indicator of raised intracranial pressure. However, it could be a function of the size of our study. Only 25% of patients made a complete recovery.
Historically, the disease phenotypes due to infection with C. neoformans var grubii and C. gattii are considered to be different. We found no difference in clinical phenotype between the 2 infections, and the pattern of disease seen in our cases mimics that seen in case series of C. gattii infection published from Papua New Guinea[17, 27]. Like us, Seaton et al found presence of convulsions to be a predictor of death in patients with C. gattii meningitis. In addition, the rate of blindness in our patients is similar to that reported for C. gattii infection in immunocompetent adults, and higher than in HIV infected patients. The lack of difference in the clinical phenotypes of disease according to infecting species and the difference in rates of visual loss between HIV positive and negative patients support the hypothesis that host immune status is key in determining clinical phenotype in cryptococcal disease. There are retrospective data suggesting that immune mediating drugs (corticosteroids) may reduce the risk of blindness in C. gattii infection [34, 35]. The similar clinical phenotype seen in our predominantly C. neoformans var grubii infected patients suggests that this may be a worthwhile hypothesis to test in our patients.
There are no established breakpoints for anti-fungal drugs and C. neoformans, which makes the interpretation of anti-fungal MICs difficult [36–39]. The geometric mean MICs of flucytosine and amphotericin B were statistically significantly higher for C. neoformans var grubii. This is consistent with the experience in Malaysia, but differs from reports from Brazil and Taiwan which found C. gattii to be inherently less susceptible to amphotericin B and flucytosine [39–41]. Thompson found no difference in MICs by species, but was comparing strains from multiple geographic regions. Our data are reassuring, suggesting that it is safe to extrapolate treatment doses for our patients with C. gattii from trials in C. neoformans infection. However, the clinical significance of differences in MICs of flucytosine and amphotericin B remain unclear. MIC data were available for only 7 of the isolates from the 11 patients that died - we could not demonstrate any correlation between the MIC of any drug and patient outcome.
gattii has been seen to establish new ecological niches, with subsequent successful human-mediated dispersal . It is possible that disease incidence in the immunocompetent will increase [44, 45]. There has been little change in the treatment of cryptococcal meningitis over the past 10 years. Amphotericin B remains a key component of treatment, but therapy is protracted, expensive and difficult to administer. Combined with flucytosine, it has been associated with more rapid CSF sterilisation in HIV patients, and more recently slower rates of CSF sterilisation have been shown to be associated with an increased risk of death [46, 47]. The identification of a reliable surrogate marker of outcome in cryptococcal meningitis enables anti-fungal therapeutic trials to be undertaken with smaller numbers of patients, and it should now be possible to answer clinical questions in groups where disease is rare, such as HIV uninfected patients. However, differences in complication rates, such as blindness, between HIV infected and uninfected patients may be due to differences in disease pathogenesis (perhaps immune mediated) and clinical endpoint trials, although difficult, may still be necessary.