Survival after cryptococcal meningitis not improving despite ART in Uganda

The 14-day mortality rate for cryptococcal meningitis (CM) in HIV-1-infected Ugandan patients remained high at 20-42 % despite antifungal therapy with amphotericin B and access to highly active antiretroviral therapy (HAART), according to the findings of a longitudinal prospective study published in the June 1^st edition of Clinical Infectious Diseases. The findings highlight the challenges of managing opportunistic infections like CM during the era of HAART.

Meningitis, inflammation of the membrane covering the spinal cord and brain, is a common manifestation of HIV infection (reviewed in December 2007 in HIV & AIDS Treatment and Practice, NAM's newsletter on HIV treatment in resource-limited settings). CM is the commonest and fatal opportunistic infection of the central nervous system (CNS) in HIV-infected patients in many African countries. The diagnosis and management of CM in these resource-poor settings is fraught with problems.

Lumbar puncture must be carried out to obtain cerebrospinal fluid (CSF) for laboratory diagnosis but many African patients will not accept this invasive method. The diagnostic laboratory techniques require trained personnel and expensive reagents which are beyond the reach of resource-poor clinics.

CM treatment in Africa is based on amphotericin B and fluconazole while in North America it is based on a combination of amphotericin B, flucytosine, and aggressive management of intracranial pressure. The fact that the latter treatment modality is not practical in Africa means that the management of CM in both regions is not comparable.

Paradoxically, the mortality rate of CM before and after HAART has remained significantly elevated in HIV-infected African patients by comparison with HIV-infected patients in North America. The mechanisms underlying the differences in the epidemiology of CM in Africa and North America are not known.

A team of US, Canadian, and Ugandan researchers re-examined this issue by investigating the survival of HIV-1-infected Ugandan patients before and after the availability of HAART at Mulago Hospital in Kampala.

The study patients were recruited between November 2001 and March 2002 before HAART was available (cohort 1) and between July and December 2006 when HAART was available (cohort 2). The patients were 18 years old and above, HIV seropositive, not receiving HAART (cohort 1, n= 92), receiving HAART (cohort 2, n= 44), and having a laboratory confirmed clinical diagnosis of CM.

Hospitalised patients received amphotericin B deoxycholate for 14 consecutive days. Patients with renal dysfunction were monitored daily while receiving daily intravenous saline and amphotericin B on alternate days. Acute elevated intracranial pressure was managed by draining CSF after lumbar puncture. Upon hospital discharge, patients took indefinite fluconazole prophylaxis and those in cohort 2 initiated HAART within 1-2 weeks of discharge.

Patients presented with prolonged CM symptoms lasting for a median duration of 14 days. Amphotericin B toxicity consisted of nausea, vomiting, and rigors but renal dysfunction was rare. The 14-day survival rates were 49% in 2001–2002 before HAART and 80% in 2006 after HAART (P

The most remarkable findings were the elevated CM mortality rates which, despite HAART, continued to increase after hospitalization. During 2006–2007, the 6-month survival rate after CM diagnosis was only 41%. Surprisingly, there was no significant association between initial intracranial pressures and death at the time of CM diagnosis.

Managing intracranial pressure is a critical component of CM treatment. Although the majority of Ugandan patients had elevated CSF opening pressures ( ≥ 200 mm H2O), only 5 consented to therapeutic lumbar puncture. This is an important constraint for treating CM.

Immune reconstitution inflammatory syndrome (IRIS) is an emerging public health problem in Africa in the wake of increased access to HAART. CM-related IRIS was diagnosed a median of 12 weeks after initiation of HAART in 42% of patients. Four out of 10 patients with CM-IRIS died.

In conclusion, unabatedly high CM-associated mortality was observed before and after the availability of HAART in HIV-infected Ugandan patients. The mortality in the post-HAART era might be due to IRIS. There are two important policy implications.

The timing of HAART initiation after CM diagnosis needs to be optimized in order to mitigate the risk of CM-related IRIS. There is an urgent need for less invasive procedures for managing intracranial pressure since lumbar puncture is not acceptable to patients.