Cryptococcal meningitis was first described in 1905. It is an infection of the brain parenchyma and sub‐arachnoid space by the encapsulated saprophytic yeast organism Cryptococcus neoformans, of which there are three subtypes: var. gatti, var. neoformans and var. grubbi (Lee 1996). Before the global human immunodeficiency virus (HIV) pandemic the main pathogenic strain was C. Neoformans var. gatti, which is a rare cause of meningitis in immunocompetent individuals in several tropical countries.

Over the last twenty years, HIV has created a large and severely immunocompromised population to whom cryptococcosis is a dangerous opportunistic infection. The major burden of disease is in South‐East Asia and the African sub‐continent. During the 1990s cryptococcal meningitis was the leading reported cause of adult meningitis in parts of Africa (Maher 1994, Moosa 1997, Mwaba 2001), and 10‐30% of AIDS deaths are attributed to the disease (French 2002, Bicanic 2005, Day 2004). This makes it the second leading cause of death in HIV‐infected individuals in Africa after tuberculosis. The majority of these patients are infected with sub‐type C. neoformans var. grubbi (Casadevall 2001) and infection predominantly occurs in patients with CD4 counts of less than 100x106 cells/l (French 2002, Saag 2000). Cryptococcal meningititis is classified by the World Health Organization as HIV/AIDS stage 4 disease.

Over 75% of patients with cryptococcal meningitis present with headache (Day 2004), usually evolving over 2‐4 weeks. Fever, nausea, vomiting and seizures are common, but neck stiffness and classical signs of meningism are seen in less than 25% of cases; confusion or behaviour change may be the only sign of infection (French 2002, Day 2004). Sub‐acute dementia can develop, which is reversible with treatment. Space occupying lesions such as cryptococcomas, sub‐dural effusions and spinal cord granulomas are occasionally seen and present with focal neurological signs. Visual impairment can occur and blindness is a long‐term complication of infection in some immunocompetent patients (Seaton 1997 a; Seaton 1997 b).

Transmission of C. neoformans is usually by inhalation. A symptomatic respiratory illness may occur but this is rare in HIV‐infected patients (Casadevall 2001). The skin is commonly involved as a manifestation of disseminated disease. Cutaneous lesions often resemble molluscum contagiosum but a wide spectrum of plaques and papules have been reported, especially in the immunocompromised population.

Diagnosing cryptococcal meningitis requires the examination of cerebrospinal fluid (CSF). Indian ink staining is the cheapest technique and is relatively easy to perform in developing countries. It has 75% sensitivity on a centrifuged sample (Day 2004). A more reliable but expensive test is Cryptococcal antigen detection, which has 99% sensitivity in subjects with titres > 1:2048 (Van der Horst 1997). Fungal culture has a sensitivity of 90%. Microscopic and biochemical features of positive CSF analysis may include a moderate mononuclear leucocytosis, mildly elevated protein and slightly low CSF/blood glucose ratio. These parameters may be normal in 17% of HIV patients with known cryptococcal disease (Moosa 1997). Computer tomography brain imaging is of limited supportive diagnostic value as the scan is normal in 50% of proven cases. It can be justified in some cases, however, to exclude mass lesions and other HIV‐related pathology (e.g. toxoplasmosis, CNS lymphoma).

HIV‐associated cryptococcal meningitis carries a high mortality even when treated appropriately, a study from Thailand reported 43% mortality in patients treated in a tertiary facility, with the majority of the deaths occuring in the first two weeks of therapy (Imwidthaya 2000). Three main groups of anti‐fungal drugs are used in treatment, amphotericin B, flucytosine and the azoles. Drug toxicity has been found to be a problem and is reported in up to 60% of patients (Day 2004).

Amphotericin B, the traditional mainstay of therapy, can be given intravenously, or, rarely, intrathecally, but it has poor oral bioavailability. Although actively fungicidal, it is associated with significant nephrotoxicity. This can be reduced by giving intravenous normal saline pre‐medication. Liposomal amphoteracin preparations are associated with fewer side effects but are very expensive and not available in resource poor settings.

Intra‐venous or oral flucytosine displays synergy with amphotericin. It cannot be used as monotherapy because of rapidly developing drug resistance and significant gastro‐intestinal side effects may develop (Saag 2000, Graybill 1997). Drug levels or haematological markers need to be monitored.

The azole compounds (e.g. fluconazole, itraconazole, ketoconazole) were introduced in the 1980s as effective, well tolerated oral and intravenous anti‐fungals. Their action against cryptococcal disease has been well described.

Most modern treatment protocols involve combinations of the above drugs. The Infectious Diseases Society of America recommends a 3 step treatment approach: induction (2 weeks of amphotericin B 0.7‐1mg/kg/day plus flucytosine 100mg/kg/day) followed by consolidation (8 weeks of fluconazole 400mg/day) and then maintenance (fluconazole 200mg/day for life) (Saag 2000), but some commentators have questioned the scientific evidence behind this (Day 2004).Other proposed regimes include prolonged courses of amphotericin and flucytosine, fluconazole and flucytosine dual therapy and the use of amphotericin alone. In much of the developing world fluconazole is the only drug available, via a donation programme and azole based monotherapy is widely used. A number of these regimes are based on anecdotal reports of open‐label phase II clinical trials and the rationale for their use has never been rigorously reviewed.

Raised intracranial pressure (ICP), is common in cryptococcal meningitis occurring in more than 50% of patients (Graybill 1997). Raised ICP during the first 2 weeks of treatment is associated with a poor clinical response and careful management of ICP is thought to reduce mortality and long term sequelae. Proposed mechanisms to reduce ICP include repeated lumbar puncture, drain insertion and ventriculo‐peritoneal shunting (Denning 1991). Mannitol, acetazolamide and corticosteroids have also been tried, but the effects of these interventions are, at best, unproven.

Initial antifungal treatment rarely eradicates cryptococcosis in HIV infected patients but symptoms may improve and control of infection is obtainable with secondary prophylaxis. With the advent of antiretroviral therapy this picture may change with patients that are initially treated with antifungal agents being established on antiretroviral therapy and no longer needing secondary prophylaxis as their immune system improves (Vibhagool 2003). Antiretroviral therapy is becoming increasingly available in the developing world, many patients in these setting still present with a major opportunistic infection such as cryptococcal meningitis having not known they were HIV infected. Defining effective evidence based protocols for treatment of cryptococcal meningitis is important in these patients if they are to benefit from subsequent antiretroviral therapy.

The mainstay of treatment for cryptococcal meningitis in the developing world is fluconazole, which in many cases is the only drug available. The aim of this review will be to look at trials comparing the current gold standard of treatment amphoteracin to fluconazole. Also combinations of treatment will be compared as interventions to manage raised intracranial pressure.

This review will not look at primary or secondary prophylaxis against cryptococcal meningtis.