Ellis J, Hale G, Nsangi LJ, et al.  Inpatient initiation of tuberculosis preventive therapy with 1 month of isoniazid and rifapentine for adults with advanced HIV disease and cryptococcal meningitis (IMPROVE): a non-inferiority, randomised controlled trial.  Lancet HIV.  2026 Feb 6: online ahead of print.  https://doi.org/10.1016/S2352-3018(25)00246-2.

IMPROVE is a phase 3, open-label trial which recruited hospitalized adults receiving cryptococcal meningitis treatment across three tertiary hospitals in Uganda to compare the safety and feasibility of two 1HP delivery strategies (i.e., inpatient initiation versus standard of care outpatient initiation at week 6).  A 28-day course of 600mg rifapentine plus 300mg isoniazid daily with adjunctive pyridoxine was initiated either before hospital discharge (inpatient group, n = 103) or at 6 weeks after meningitis diagnosis (outpatient group, n = 102), with the first dose given as DOT in a healthcare setting for both groups.  Participants were followed for 18 weeks and ART was initiated, reinitiated, or switched during weeks 4 to 6 in accordance with Ugandan guidelines.  Adjusted intention-to-treat analyses indicated that 70% of participants in the inpatient group met the primary endpoint of tuberculosis disease-free survival and 1HP completion compared with 62% in the outpatient group, meeting prespecified non-inferiority criteria.  18-week survival was similar between groups, and treatment completion was achieved in 76% of the inpatient group versus 66% of the outpatient group (21% of outpatient group participants did not start 1HP as planned, with the most frequent reasons being diagnosis of active tuberculosis disease before week 6 or death).  There was no difference in frequency of grade 3 or above adverse events or serious adverse events, 55% versus 58%.

Findings from this open-label, non-inferiority trial offer important insights into the promise of inpatient 1HP initiation among adults with advanced HIV disease being treated for cryptococcal meningitis, a population with very high morbidity and mortality rates.  Inpatient initiation was feasible, and 18-week tuberculosis disease-free survival and 1HP completion outcomes were comparable to standard of care outpatient initiation.  Additional studies to determine if timely 1HP completion leads to other improved outcomes e.g., reduced risk of TB-IRIS and active tuberculosis disease, could provide additional evidence supporting its broad public health impact.