Abstract & Commentary
Delaying Initiation of Antiretroviral Therapy for Cryptococcal Meningitis Improves Survival
By Richard R. Watkins, MD, MS, FACP
Division of Infectious Diseases, Akron General Medial Center, Akron, OH; Associate Professor of Internal Medicine, Northeast Ohio Medical University, Rootstown, OH
Dr. Watkins reports no financial relationships in this field of study.
SOURCE: Boulware DR, et al. Timing of antiretroviral therapy after diagnosis of cryptococcal meningitis. N Engl J Med 2014;370:2487-2498.
SYNOPSIS: HIV-infected patients diagnosed with cryptococcal meningitis who received antiretroviral therapy (ART) 5 weeks after starting antifungal therapy had improved survival at 26 weeks compared to similar patients who received ART at 1-2 weeks (45% vs 30%, respectively, P=0.03).
Cryptococcal meningitis causes significant mortality in patients with HIV, especially in sub-Saharan Africa. The timing of ART initiation in these patients has been controversial due to the risk of developing immune reconstitution inflammatory syndrome (IRIS), which may be fatal when it involves the central nervous system, vs. the survival benefit from ART. Therefore, Boulware and colleagues sought to determine if early initiation of ART, defined as 1-2 weeks from the diagnosis of cryptococcal meningitis, led to a difference in 26 week survival compared to delaying ART for 5 weeks after diagnosis.
The study was conducted at 2 hospitals in Uganda and 1 in South Africa from November 2010 until enrollment was stopped by the Data and Safety Monitoring Board (DSMB) in April 2012 due to excess mortality in the earlier ART group. Patients entered the trial after 7 to 11 days of amphotericin B and fluconazole combination therapy. They were randomized in a 1:1 ratio to receive ART within 48 hours of randomization or at 4 weeks after randomization. The primary end point of the study was survival at 26 weeks. Secondary endpoints were survival through 46 weeks, cryptococcal IRIS, relapse of cryptococcal meningitis, fungal clearance, HIV viral load suppression at 26 weeks, adverse events and discontinuation of ART for more than 3 days for any reason.
After 389 patients were screened, 177 were randomized to one of the two treatment groups. The baseline and demographic characteristics of the groups were very similar. The number of patients who died by 26 weeks was significantly greater in the earlier-ART group compared to the later-ART group (40 of 88 [45%] vs. 27 of 89 [30%]; hazard ratio for death, 1.73; 95% confidence interval [CI], 1.06 to 2.82; P=0.03). The difference in mortality occurred early in treatment, during study days 8 through 30, when the patients were receiving consolidation fluconazole therapy. During this time frame, 21 of 75 patients (28%) in the earlier-ART group and 8 of 80 (10%) in the later-ART group died (hazard ratio, 3.10; 95% CI, 1.37 to 7.00; P=0.007). Mortality was not significantly different between the two groups after 30 days (P=0.87). By 46 weeks, 1 additional patient in the earlier-ART group and 2 in the later-ART group died (hazard ratio 1.66; 95% CI, 1.03 to 2.68; P=0.04). Furthermore, subgroup analysis showed that a CSF white-cell count of <5 cells per cubic millimeter was associated with higher mortality in the earlier-ART group than with later-ART.
Survival was similar between treatment groups in the subgroup with CSF white-cell counts of ≥5 cells per cubic millimeter. Earlier-ART was not favorable in any subgroup including patients with CD4 counts <50 per cubic millimeter. Participants at lower risk of death, such as those with lower CSF fungal burden at day 7 of antifungal therapy, also did not benefit from earlier ART. Timing of ART did not affect any of the secondary endpoints. Interestingly, the incidence of cryptococcal IRIS did not differ significantly between the earlier-ART vs. later-ART groups (20% [17 of 87] and 13% [9 of 69], respectively; P=0.32). Adverse events were similar between the two groups and were mostly related to amphotericin. Causes of death were also similar except for an excess of cryptococcal meningitis-related deaths in the earlier-ART group vs. the later-ART group (19 vs 10 deaths, respectively).
The only benefit for earlier ART in the study was on the immune response in the CSF. At day 14 of amphotericin therapy, the proportion of patients who had CSF white blood cell counts of 5 per cubic millimeter or higher was greater in the earlier-ART group (58%) than in the later-ART group (40%; P=0.047). However, this finding did not seem to translate into any appreciable clinical benefit.
COMMENTARY
This was a critical study because it helps clarify an important clinical dilemma, when to start ART in cryptococcal meningitis. On the one hand, the benefits of early ART in patients with HIV and AIDS are clear and DHHS guidelines now recommend initiating ART regardless of a patient’s CD4 count. On the other, the risk for IRIS should not be overlooked, especially in those with central nervous system infections where it is particularly hazardous. The investigators found that delaying ART for 5 weeks after initiating antifungal therapy improved survival in patients with cryptococcal meningitis. Moreover, besides a slightly improved immune CSF response, there were no significant benefits from earlier ART. Indeed, earlier ART was actually associated with more deaths from cryptococcal meningitis, leading the DSMB to stop the trial early. Since IRIS with CNS infections causes an intense local inflammatory response within a confined space, the authors hypothesized that earlier-ART was most harmful in high-risk patients with a predisposition to cryptococcal IRIS and it was this mechanism that drove the results of the study. Thus, certain patients who had not yet recovered from the detrimental effects of their cryptococcal meningitis were especially susceptible to a second insult from IRIS, which was insurmountable and led to their deaths.
One of the limitations of the study was the difficulty in determining whether progressive clinical deterioration was due to cryptococcal meningitis or cryptococcal IRIS, which probably resulted in ascertainment bias and an underdetection of early IRIS events. Another limitation was that the trial was stopped early which may have caused an overestimation of the magnitude of benefit or harm. Finally, the relatively small sample size limits the power of the subgroup analysis to provide guidance for individual patients. Despite these limitations, Boulware and colleague’s carefully conducted study shows the benefit of delaying ART in cryptococcal meningitis while amphotericin is being given. Although the trial was conducted in a resource-limited setting, it seems logical that delaying ART will also benefit HIV-infected patients with cryptococcal meningitis in high-income countries like the United States.
