Why HIV Sepsis and Tuberculosis Collide
HIV-RELATED sepsis remains a major cause of death in sub-Saharan Africa, where people living with HIV are at high risk of severe infections, including tuberculosis. Sepsis is a life-threatening response to infection that can lead to organ failure, and diagnosing tuberculosis in critically ill patients is often delayed or impossible due to limited access to rapid diagnostics. This has raised the question of whether starting antituberculosis therapy immediately, before confirmation, could save lives.
Testing Immediate Antituberculosis Therapy in Critical Care
The ATLAS trial set out to answer this question in real-world hospital settings in Tanzania and Uganda. Researchers enrolled adults living with HIV who were hospitalised with sepsis and randomly assigned them to receive either immediate antituberculosis therapy or treatment only after a tuberculosis diagnosis. Participants were also assigned to either high-dose or standard-dose therapy, creating a 2×2 study design.
All patients received standard antibiotic treatment for bacterial sepsis, reflecting routine clinical care. The main outcome was death within 28 days of starting treatment.
No Overall Survival Benefit for Immediate or High-Dose Treatment
Among nearly 400 participants analysed, overall 28-day mortality was high, at around 25%. The trial found no difference in survival between those given immediate antituberculosis therapy and those treated only after diagnosis. Similarly, increasing drug doses above standard WHO recommendations did not reduce deaths.
These findings suggest that, for unselected patients with HIV-related sepsis, routinely starting antituberculosis therapy immediately or at higher doses does not improve short-term survival.
Clear Benefit When Tuberculosis Is Confirmed
However, results differed in the subgroup of patients who were later confirmed to have tuberculosis, accounting for just over half of participants. In this group, those who received immediate conventional-dose antituberculosis therapy had significantly lower 28-day mortality compared with those whose treatment was delayed until diagnosis.
This pattern supports the idea that, when tuberculosis is truly driving sepsis, earlier active treatment may be life-saving. Importantly, higher drug doses did not confer additional benefit and were associated with more frequent liver injury events.
Implications for Practice in Resource-Limited Settings
The study highlights the challenge of managing HIV-related sepsis where tuberculosis is common but hard to diagnose. While immediate treatment for everyone may not be justified, faster identification of tuberculosis, and prompt initiation of standard-dose therapy when confirmed, could improve outcomes. The findings underline the need for better diagnostics rather than blanket escalation of treatment.
Reference
Heysell SK et al. Immediate or high-dose antituberculosis therapy for HIV-related sepsis in Tanzania and Uganda (ATLAS): a phase 3, open-label, randomised, controlled, 2 × 2 factorial, superiority trial. Lancet Infect Dis. 2026; doi:10.1016/S1473-3099(25)00747-9.
