Abstract:
Background. In sub-Saharan Africa, 20%–25% of people starting antiretroviral therapy (ART) have severe immunosuppression;
approximately 10% die within 3 months. In the Reduction of EArly mortaLITY (REALITY) randomized trial, a broad enhanced
anti-infection prophylaxis bundle reduced mortality vs cotrimoxazole. We investigate the contribution and timing of different causes
of mortality/morbidity.
Methods. Participants started ART with a CD4 count <100 cells/μL; enhanced prophylaxis comprised cotrimoxazole plus 12
weeks of isoniazid + fluconazole, single-dose albendazole, and 5 days of azithromycin. A blinded committee adjudicated events and
causes of death as (non–mutually exclusively) tuberculosis, cryptococcosis, severe bacterial infection (SBI), other potentially azith-
romycin-responsive infections, other events, and unknown.
Results. Median pre-ART CD4 count was 37 cells/μL. Among 1805 participants, 225 (12.7%) died by week 48. Fatal/nonfatal
events occurred early (median 4 weeks); rates then declined exponentially. One hundred fifty-four deaths had single and 71 had
multiple causes, including tuberculosis in 4.5% participants, cryptococcosis in 1.1%, SBI in 1.9%, other potentially azithromycin-re-
sponsive infections in 1.3%, other events in 3.6%, and unknown in 5.0%. Enhanced prophylaxis reduced deaths from cryptococcosis
and unknown causes (P < .05) but not tuberculosis, SBI, potentially azithromycin-responsive infections, or other causes (P > .3); and
reduced nonfatal/fatal tuberculosis and cryptococcosis (P < .05), but not SBI, other potentially azithromycin-responsive infections,
or other events (P > .2).
Conclusions. Enhanced prophylaxis reduced mortality from cryptococcosis and unknown causes and nonfatal tuberculosis
and cryptococcosis. High early incidence of fatal/nonfatal events highlights the need for starting enhanced-prophylaxis with ART
in advanced disease.
