Timing of antiretroviral therapy after diagnosis of cryptococcal meningitis

Abstract

Background: Cryptococcal meningitis accounts for 20 to 25% of acquired immunodeficiency syndrome-related deaths in Africa. Antiretroviral therapy (ART) is essential for survival; however, the question of when ART should be initiated after diagnosis of cryptococcal meningitis remains unanswered.

Methods: We assessed survival at 26 weeks among 177 human immunodeficiency virus-infected adults in Uganda and South Africa who had cryptococcal meningitis and had not previously received ART. We randomly assigned study participants to undergo either earlier ART initiation (1 to 2 weeks after diagnosis) or deferred ART initiation (5 weeks after diagnosis). Participants received amphotericin B (0.7 to 1.0 mg per kilogram of body weight per day) and fluconazole (800 mg per day) for 14 days, followed by consolidation therapy with fluconazole.

Results: The 26-week mortality with earlier ART initiation was significantly higher than with deferred ART initiation (45% [40 of 88 patients] vs. 30% [27 of 89 patients]; hazard ratio for death, 1.73; 95% confidence interval [CI], 1.06 to 2.82; P=0.03). The excess deaths associated with earlier ART initiation occurred 2 to 5 weeks after diagnosis (P=0.007 for the comparison between groups); mortality was similar in the two groups thereafter. Among patients with few white cells in their cerebrospinal fluid (<5 per cubic millimeter) at randomization, mortality was particularly elevated with earlier ART as compared with deferred ART (hazard ratio, 3.87; 95% CI, 1.41 to 10.58; P=0.008). The incidence of recognized cryptococcal immune reconstitution inflammatory syndrome did not differ significantly between the earlier-ART group and the deferred-ART group (20% and 13%, respectively; P=0.32). All other clinical, immunologic, virologic, and microbiologic outcomes, as well as adverse events, were similar between the groups.

Conclusions: Deferring ART for 5 weeks after the diagnosis of cryptococcal meningitis was associated with significantly improved survival, as compared with initiating ART at 1 to 2 weeks, especially among patients with a paucity of white cells in cerebrospinal fluid. (Funded by the National Institute of Allergy and Infectious Diseases and others; COAT ClinicalTrials.gov number, NCT01075152.).

Figure 1. Screening and Randomization

Of 237 patients with cryptococcal meningitis, 177 were enrolled in the study after 7 to 11 days of antifungal treatment. One participant randomly assigned to receive earlier antiretroviral therapy (ART) withdrew consent 2 days after randomization. All other participants were followed for 46 weeks or until death. No participants were lost to follow-up. Analyses were performed according to the intention-to-treat principle. Randomization was stratified according to site and the presence or absence of altered mental status (Glasgow Coma Scale [GCS] score <15) at the time informed consent was obtained (within 48 hours before randomization).

Figure 2. Cumulative Probability of Survival According to Timing of ART

Overall survival from randomization (time 0) at 7 to 11 days after diagnosis of cryptococcal meningitis to 46 weeks is shown in Panel A. Earlier ART initiation, at 7 to 13 days after diagnosis, was associated with a risk of death within 26 weeks that was 15 percentage points higher than that associated with ART initiation at 5 weeks after diagnosis (P = 0.03). Panels B and C show survival stratified according to the cerebrospinal fluid (CSF) white-cell count at the time of randomization. Among study participants with a paucity of CSF white cells (<5 cells per cubic millimeter), mortality was significantly higher with earlier ART initiation than with deferred ART initiation (P = 0.008). In all panels, the vertical dashed line at 1 month indicates the time of ART initiation in the deferred-ART group. The primary outcome was the survival at 26 weeks (vertical dashed line at 6 months).

Figure 3. Subgroup Analyses of Mortality

Prespecified subgroups were assessed to determine whether differences in survival between treatment groups at 26 weeks were dependent on baseline clinical characteristics. Values for subgroups are 6-month mortality. The CSF white-cell count at randomization was the only characteristic with a significant interaction, which suggested a differential response to the timing of ART initiation.