Ugandan study supports the use of fluconazole to prevent cryptococcal meningitis

Oral fluconazole prophylaxis safely prevents invasive cryptococcal disease in people with advanced HIV, according to a randomised, double-blind, placebo-controlled trial conducted in over 1500 participants in rural Uganda, and presented on Monday at the Sixteenth Conference on Retrovirus and Opportunistic Infections (CROI) in Montreal. Prophylaxis was effective in people who were waiting to receive antiretroviral therapy (ART) and in those who had recently started ART, but who had not yet had a significant improvement in their immune status. It also significantly reduced the incidence of other serious fungal infections like oesophageal candidiasis.

Taking fluconazole did not affect overall mortality (from all causes) in the study but the number of deaths attributable to cryptococcal disease in the study was likely to be lower than would typically be seen in clinical settings in most of sub-Saharan Africa - and access to rapid diagnosis and treatment better (see below). “So in programmatic settings, fluconazole prophylaxis may have potentially even greater impact,” said Dr Rosalind Parkes-Ratanshi, who presented the findings on behalf of the team at the Medical Research Council (MRC) in Uganda and collaborating institutions.

The findings could prompt more programmes to consider the inclusion of fluconazole prophylaxis as part of the basic package of care that should be provided to anyone with advanced HIV disease before, and for the first few months after, starting ART, although the cost of purchasing fluconazole is likely to be a disincentive. The findings may also increase pressure on Pfizer, the drug’s maker, to expand their drug assistance programme supplying fluconazole - currently for treatment only - in resource-limited settings.

People with CD4 cell counts below 100 cells/mm³ are at increased risk of cryptococcal disease, especially cryptococcal meningitis, which can lead to rapid death, and may cause permanent brain damage and other disabilities in survivors, especially when diagnosis and treatment are delayed - as is commonly the case in sub-Saharan Africa.

ART has dramatically reduced the incidence of cryptococcal disease in industrialised settings, but since many people in resource-limited settings are not diagnosed with HIV until they have AIDS (and most who qualify for ART are still not receiving it), cryptococcal disease remains a significant cause of morbidity and mortality in much of the world. In fact, earlier this decade, one study found that cryptococcal disease had been responsible for about 17% of deaths in a cohort of people with HIV from Uganda (French).

Several randomised studies have evaluated the effect of primary fluconazole prophylaxis on the incidence of cryptococcal disease in people with advanced HIV disease. Notably, in the pre-ART era. Powderly et al. found that daily fluconazole significantly reduced the incidence of invasive fungal infections (cryptococcosis, and oesophageal and oropharyngeal candidiasis), compared to clotrimazole troches in 428 people with advanced HIV infection - though survival was similar in the two groups. Conversely, a placebo-controlled study in 90 people with advanced HIV infection in Thailand reported that once-weekly fluconazole was associated with a lower death rate, though its effect on cryptococcal disease was not significant (because of the small number of events) (Chetchotisakd). In 2005, a Cochrane Review meta-analysis of the evidence from all azole-drug prophylaxis studies concluded that their use was associated with a 0.21 reduction in the relative risk of cryptococcal disease but that there was no clear effect on mortality (Chang).

WHO currently recommends that, in areas where cryptococcal disease is common, programmes consider fluconazole prophylaxis for severely immunocompromised people with HIV (WHO clinical stage 4 or CD4 cell counts < 100 cells), regardless of whether they are on ART or not. But very few have put this into practice because there is little evidence of benefit from the developing world. In addition, concerns have been raised about potential drug interactions (for example, with nevirapine), the potential for teratogenicity of fluconazole, and the risk of developing fluconazole-resistant thrush infections.

The study presented this week at CROI is not only the first large trial of fluconazole as primary prophylaxis to be performed in Africa, it is the largest fluconazole prophylaxis study to ever be conducted and its findings lend substantial support to the current WHO recommendation.

The study was a prospective, double-blind, randomised controlled trial to determine whether 200 mg of fluconazole three times a week could reduce invasive cryptococcal disease and all-cause mortality when compared to placebo in adults from rural Uganda. The trial ran between 2004 and 2008, but the sample size had to be increased when ART became available, and the protocol amended to recruit ART-naive adult patients with CD4 cell counts under 200, and to follow them as they started ART.

The study excluded anyone who was pregnant or breastfeeding, severely ill (bed-bound), with liver-function tests (LFTs) more than three times the upper limit of normal and anyone with previous/active cryptococcal disease or who screened positive for cryptococcal antigen (CrAg) at study entry.

A total of 1519 adults were enrolled, with follow up provided every eight weeks, or earlier if they became unwell. LFTs were monitored at each visit and CD4 cell counts every 16 weeks. Once ART became available, participants were actively referred to local providers for ART and 1335 (88%) participants started ART, a median of 11 weeks after enrolment (interquartile range (IQR) 7-17). Fluconazole (or placebo) was continued until CD4 counts rose to 200 (median 197 days, IQR 160 to 339). Survival and time to the study endpoints were assessed in intention-to-treat Kaplan-Meier survival analyses and log rank tests. Cox proportional-hazard regression-model analyses were used to adjust for the effect of ART and CD4 count.

A total of 760 participants received fluconazole and 759 received placebo: baseline characteristics were similar. About 60% of the cohort was female and the median CD4 cell count was 111 cells/mm³.

Over the course of follow-up, 19 participants developed confirmed cryptococcal disease, one on fluconazole, and 18 on placebo (p = 0.0001), and the effect was observed both before and after starting ART - twelve developed cryptococcus before ART (eleven on placebo) and 7 while on ART (all on placebo) (p=0.0001). The overall adjusted hazard ratio was 18.74, 95% CI (2.50; 140.67), p=0.004. In a discussion after the presentation, it was suggested that some of the seven cryptococcal events that occurred on ART in the placebo arm could have been cryptococcal disease that was ‘unmasked’ by immune reconstitution inflammatory syndrome (IRIS). The researchers did not look at IRIS specifically but there is a chance that fluconazole may have prevented such events.

As expected, the majority of cryptococcal events occurred in people with CD4 cell counts below 100 at baseline (13 were below 50), though a couple of events were seen in participants with CD4 cell counts between 100 and 149 at baseline.

Seven participants died of cryptococcal disease - all on placebo. However, there was no difference in all-cause mortality between fluconazole (n = 100) and placebo (n = 98).

Fluconazole also reduced the incidence of oesphageal Candida, with 55 cases on placebo versus eight on fluconazole, p = <0.0001, though this effect was less profound after ART was initiated. The adjusted hazard ratio was 7.1 for fluconazole. Very similar effects were seen with oropharyngeal and vaginal Candida.

The most common adverse event was having an abnormal LFT (transaminases more than five times the upper limit of normal) but the rate of events was similar in both arms (aHR = 1.06, 95% CI 0.74; 1.53 p = <0.74). Contrary to fears, fluconazole did not increase the rate of toxicity among people taking nevirapine in their ART regimen - though it is important to note that this study excluded people with CD4 cell counts over 200 who are more likely to experience hypersensitivity reactions and hepatitis on nevirapine.

The most common reason for stopping the trial drug was having a CD4 cell count that rose above 200 cells/mm³ as a result of ART. This occurred in 751 or 49.4% of the participants. Another 3.5% quit when they became pregnant. Of note, the loss-to-follow-up rate was quite low - at 3.3% - while about 1% requested to leave the study.

Dr Parkes-Ratanshi believes the quality of care provided in this research setting may have resulted in an understatement of the effect of fluconazole prophylaxis on mortality. For instance “overall, in this study, the rate of crytococcal infection was low. One reason was that we excluded those people who were CrAg positive at the beginning of the study,” she said.

However, most settings in sub-Saharan Africa do not have the capacity to screen all patients (without symptoms) for CrAg, so the rates of invasive cryptococcal disease would normally be higher. In addition, the study participants had better access to rapid diagnosis and optimal treatment, including access to ART. So she believes that the effect of fluconazole prophylaxis is potentially greater and that it “is likely to reduce mortality in patients who are not yet on ART. However, the benefit also extends to those in the first few months on ART before significant improvements in immune status,” she said.

The study makes the evidence base for the current WHO fluconazole prophylaxis recommendation considerably stronger, and many of the often cited reasons for not implementing fluconazole prophylaxis as a routine intervention did not appear to be an issue in this low resourced setting. Critically, it was safe in this population and as a prophylactic measure, resistance was not a major issue in this large cohort. “We didn’t see any resistance in our patients with cryptococcal disease,” said Dr Parkes-Ratanshi. “We haven’t yet gotten the drug sensitivities for candida, but in the clinical setting, there are very few cases of recurrent candida, and we have not found very resistant candida though we are waiting for the sensitivities.”

The fact that fluconazole offered benefit beyond ART is important. Some regional guidelines have quite rightly stated that ART is the most effective preventive measure for cryptococcal disease and that access to ART should be prioritised over introducing fluconazole prophylaxis. But this may be a false dichotomy if fluconazole can easily delivered in a programmatic setting while someone is being prepared to take ART: “It is safe and simple to administer. It doesn’t have to be given by a doctor; it can be given by other healthcare workers; and it can be given at peripheral health centres where there may be no access to ART, certainly no CD4 cell count testing and monitoring for ART,” said Dr Parkes-Ratanshi. Furthermore, it is a time-limited intervention in those starting ART, with a median duration of treatment of about five months.

Of note, a late breaker study presented by Dr Azure Makadzange of the University of Washington, Seattle, and the University of Zimbabwe, Harare, during the same session of the conference, reported a very high rate of death (62%) in people with cryptococcal meningitis treated with high-dose fluconazole (which is the only available option in much of Africa) regardless of whether they started ART or not. In fact, the risk of death was more than twice as high in people who received early versus deferred ART (before finishing their course of fluconazole as treatment). There was much debate about the study’s findings, which seem to contradict other reports, but it is important to note that most of these deaths occurred early on ART, most likely due to IRIS - which fluconazole prophylaxis before ART could possibly prevent.

However, many programmes are still having trouble supplying cotrimoxazole, a clearly life-saving intervention, despite its low cost and widespread availability. Fluconazole is a much more expensive drug. Although fluconazole is made available by Pfizer free of charge to governmental and non-governmental organisations to treat severe AIDS-related fungal infections in less developed countries, it is not clear that the company would be willing to expand access to include its use for prophylaxis. So the inclusion of fluconazole as part of the basic package of care may be limited to countries with access to low-cost, generic formulations of the drug.

A clinical review of the management of meningitis – including cryptococcal meningitis – was published in HIV & AIDS Treatment in Practice in December 2007 (http://www.aidsmap.com/cms1256448.asp).