The recently published ATLAS study of sepsis in patients living with HIV suggests that early, aggressive therapy for TB may reduce mortality. In regions with endemic TB, it is the leading cause of sepsis in patients with HIV. The ATLAS results show immediate TB treatment in patients with sepsis reduced mortality by 23% compared with standard care. Unfortunately, standard rapid tests miss a significant number of TB infections.1
“What was not known prior to our studies and the clinical trial was just how common TB is as a pathogen of sepsis,” said lead author Scott Heysell, MD, MPH, Thomas H. Hunter Professor of International Health at the University of Virginia and Director of the Center for Global Health. “TB was the most common sepsis-related pathogen we found, more than half overall. TB was present in the bloodstream, and it was commonly missed by routine diagnostics.”
Dr. Heysell and fellow lead author Stellah Mpagama, MD, MSc, PhD, Director of Research at Kibong’oto Infectious Diseases Hospital in Tanzania, noted that although ATLAS was conducted in East Africa, the findings may apply to people living with HIV in other regions with endemic TB, including parts of Europe and the Americas.
The phase 3 trial randomized 437 patients living with HIV in semi-urban Tanzania and Uganda who were admitted to the hospital with sepsis to immediate or diagnosis-dependent initiation of either conventional World Health Organization-recommended TB therapy or high-dose TB therapy.1
Participants randomized to conventional TB therapy received daily fixed-dose combination tablets of rifampicin (10 mg/kg) and isoniazid (5 mg/kg), along with weight-based doses of pyrazinamide and ethambutol, plus pyridoxine. Those randomized to high-dose therapy received daily rifampicin (30 mg/kg) and isoniazid (7.5 mg/kg), along with weight-based doses of pyrazinamide and ethambutol, plus pyridoxine. Each group was further randomized to immediate TB therapy or diagnosis-dependent therapy.
The trial design included four treatment groups: immediate conventional dose, diagnosis-dependent conventional dose, immediate high dose, and diagnosis-dependent high dose. By day 28, most participants (77%) had received TB therapy, regardless of randomization.
Diagnostic testing was conducted at enrollment, including blood and urine samples for bacterial culture. TB tests included GeneXpert MTB/RIF Ultra on sputum and urine, Alere Determine LF-LAM on urine, and mycobacterial culture on sputum. Baseline blood samples were tested for 43 pathogens, including Mycobacterium tuberculosis, using quantitative polymerase chain reaction (PCR); however, the PCR results were not available for clinical decision-making. Participant eligibility criteria included concern for infection and two or more modified quick Sequential Organ Failure Assessment score criteria.1
The primary end point was mortality at 28 days. Secondary end points included time to TB therapy, time to temperature normalization, time to ambulation, duration of hospitalization, in-hospital mortality, adverse drug events, six-month survival, and sepsis etiology. Of the participants, 57% were female and the median age was 42 years old.
ATLAS found that TB was the predominant sepsis-related pathogen, identified in 52% of participants. Another 19% of participants had an additional identified pathogen within the spectrum of activity of ceftriaxone—primarily Klebsiella species and E. coli,identified in 9% and 8% of participants, respectively.1–2 A secondary analysis of ATLAS found that urine LF-LAM and sputum GeneXpert MTB/RIF testing missed a combined 32% of TB bloodstream infections and that many of the non-TB bacterial pathogens were ceftriaxone-resistant.2
“Not only was TB the most common cause of sepsis, it was commonly present in the bloodstream,” Dr. Heysell said. “We typically think of TB as a smoldering lung infection, but it was functioning more like an E. coli or Staph aureus that causes an abrupt infection like a urinary tract infection that becomes septicemic. And when it was present in the bloodstream, it was missed by routine diagnostics a third of the time.”
Overall, there was no difference in 28-day survival between the immediate and diagnosis-dependent TB treatment groups, which was driven by an interaction between high-dose and immediate treatment, Dr. Heysell said. However, prespecified subgroup analyses of those ultimately diagnosed with TB found a 23% reduction in the risk of death for the immediate conventional dose treatment group compared with the diagnosis-dependent conventional dose.1 The mortality risk reduction was slightly lower for the high-dose groups.
Dr. Heysell said the mortality risk reduction for immediate treatment was sustained at six months, and there was no significant difference in adverse events between the immediate conventional dose TB group and the diagnosis-dependent conventional dose group, suggesting that immediate TB treatment is safe even in people with sepsis and at high risk of organ toxicity.1 He said detailed six-month data will be reported at a later date.
“When someone presents with sepsis, in the United States or in sub-Saharan Africa, they get ceftriaxone or a similar antibiotic,” Dr. Heysell said. “It kills the bacteria of urinary tract infections [and] pneumonia, the common things, but it doesn’t kill TB. We believe that in any region where TB is endemic, clinicians should consider immediate TB treatment for people with HIV-related sepsis. If, at some point in time, TB testing is negative, or diagnostics were not available, and the person is doing better, there can be a discussion about continuing TB treatment.”
A follow-up study, STARS, is exploring corticosteroids and enhanced spectrum antibiotics to target TB and ceftriaxone-resistant non-TB bacteria vs conventional treatment for refractory sepsis in patients with HIV.3
References
1. Heysell SK, Mpagama SG, Nuwagira E, et al. Immediate or high-dose antituberculosis therapy for HIV-related sepsis in Tanzania and Uganda (ATLAS): a phase 3, open-label, randomized, controlled, 2 × 2 factorial, superiority trial. Lancet Infect Dis. Published online January 28, 2026. doi:10.1016/S1473-3099(25)00747-9
2. Otoupalova E, Ampaire L, Null M, et al. Aetiology of sepsis in adults living with HIV in East Africa: a secondary analysis of an open-label, multicentre, randomised, controlled phase 3 trial. EClinicalMedicine. 2026;92:103719. doi:10.1016/j.eclinm.2025.103719
3. STeroids and Enhanced Spectrum Antibiotics for the Treatment of Patients in Africa With Refractory Sepsis (STARS). ClinicalTrials.gov
