A phase III randomised controlled trial of oral fluconazole plus flucytosine versus amphotericin B-based therapy for o

Cryptococcal meningitis is one of the commonest causes of death in patients with AIDS and
is associated with up to 500,000 deaths each year in Africa alone. A large proportion of
patients die from the infection, in part because the current recommended treatment,
amphotericin B for 2 weeks, is difficult to give in hospitals in the developing world, because
it is relatively expensive and needs to be given intravenously and has side effects, often
starting in the second week, meaning monitoring is needed with frequent blood tests. The
alternative oral tablet treatment, fluconazole, that is available and cheap and currently
commonly used, is much less effective.
Therefore, based on a number of earlier small trials by the study team, we wish to test 2
new treatments, (1) Short, 1-week amphotericin B, and (2) Combination tablet treatment
with high dose fluconazole plus another drug called flucytosine, that are as fast and effective
in killing the infection as 2 weeks of amphotericin B. We will compare these new treatments
with 2 weeks amphotericin B, in a larger, randomised trial that will enable us to see whether
they are as good in preventing deaths from the infection. After 2 weeks of the study
treatments, all patients will receive the usual follow on therapy with fluconazole, and will be
started on drugs for the underlying HIV infection, as currently recommended, and followed
up for 10 weeks. 570 patients (190 given each alternative treatment) will be studied. This is
the minimum number needed to reliably compare the results of the treatments.
The project has been developed with doctors in 3 centres in Malawi and Zambia where there
are many cases and where alternative, affordable and practical treatments are urgently
needed. Each centre has the laboratories needed, and experience in doing such trials. Both
test treatments have been shown to be much more rapidly effective than fluconazole alone,
and would have less side effects and be much more easily given in developing countries than
2 weeks amphotericin B. However, if 2 weeks amphotericin B was found to be the best
treatment, then the costs required for its use could be justified. The costs as well as the
effectiveness of the treatments will be compared to help decide which treatment to
recommend in the future.

Cryptococcal meningitis is one of the commonest causes of death in cohorts of HIV-infected
patients, and of early mortality in antiretroviral (ART) programmes in Africa, being
associated with up to 500,000 deaths per year in Africa alone. The high case fatality rate of
cryptococcal meningitis is driven in substantial part by the lack of effective antifungal
regimens for resource-constrained settings. In many centres the standard initial treatment of
2 weeks intravenous amphotericin B (AmB) is not available or not safely sustainable due to
the drug and administration costs, the close monitoring required, and common dose-related
side effects. Fluconazole, the commonly used but slowly-acting alternative, widely available
through a donation programme, when used alone, is associated with a 10-week mortality of
50-60%.
Therefore, based on a series of phase II studies by the study team, 2 novel regimens, (1)
Short, 1-week AmB-based induction, and (2) Combination oral therapy with high dose
fluconazole plus flucytosine, that are associated with a rate of clearance of infection similar
to that seen with 2 weeks AmB, will be compared with the internationally-accepted standard
of 2 weeks AmB, in a randomised non-inferiority trial with 2-week mortality as primary
endpoint. After 2 weeks, all patients will receive fluconazole consolidation and maintenance
therapy, will be commenced on ART at 3-4 weeks, in line with current guidelines, and
followed to 10 weeks. 570 patients (190 per arm) will be studied in order to achieve 85%
power, using a non-inferiority design with 10% non-inferiority margin, and assuming 85% 2-
week survival in the 2-week AmB arm.
The project has been developed with collaborators from 3 African sites (Blantyre and
Lilongwe, Malawi, and Lusaka, Zambia) with a high, ongoing burden of cryptococcal disease
where alternative, sustainable regimens are urgently needed. Each site has laboratory
infrastructure in place and experience of conducting trials to GCP standard. Both test
regimens have been shown to be much more rapidly active than fluconazole monotherapy,
and would be better tolerated and much more easily implemented in resource-limited
settings than 2 weeks AmB. Alternatively, if 2 weeks AmB proved superior, then the
resources required for its wider implementation could be justified. A cost effectiveness substudy
will be developed to inform policy-making decisions.