In China trial, a severity-tailored oral regimen matches 18-month drug-resistant TB care
An all-oral short regimen, its length set by chest-imaging severity, was non-inferior to 18-month treatment and caused fewer serious adverse events in a Chinese randomised trial.
| Group | Value (%) |
|---|---|
| Short regimen (STR) | 87.1 |
| 18-month control | 81.5 |
A shorter, all-oral treatment for drug-resistant tuberculosis, in which the length of therapy was set by how severe the disease looked on a chest X-ray, was non-inferior to the standard 18-month course — and safer — in a randomised trial run across multiple centres in China [s1]. The results were published in Clinical Microbiology and Infection on 29 June [s1].
Drug-resistant TB is one of the hardest problems in the disease. Multidrug-resistant TB remains a public health crisis and a health security threat, the World Health Organization says, and only about 2 in 5 people with drug-resistant TB accessed treatment in 2024 [s2]. TB overall killed 1.23 million people in 2024, including 150 000 among people with HIV, and an estimated 10.7 million people fell ill with it worldwide that year [s2]. Shortening and simplifying treatment is central to closing the gap, because longer regimens are harder to finish and harder to deliver.
How the trial was designed
Participants with multidrug- or rifampicin-resistant TB were randomly assigned in a 2:1 ratio to receive either the all-oral short treatment regimen or an 18-month regimen based on the 2019 WHO guidelines, which served as the control [s1]. The novel feature was stratification: the short-regimen group was divided into three subgroups according to radiographic severity scores, so that patients with more extensive disease on imaging received a correspondingly different course [s1].
The primary efficacy endpoint was a favourable treatment outcome up to 6 months after discontinuing treatment, and the primary safety endpoint was grade 3 or higher adverse events during the study period [s1]. The trial was designed to test non-inferiority, with a prespecified margin of less than 12 percentage points for the difference in favourable outcomes between the groups [s1].
What it found
A total of 259 participants were randomised, 173 to the short regimen and 86 to the control [s1]. In the modified intention-to-treat population, the favourable treatment outcome rate was 87.1% (148 of 170) in the short-regimen group and 81.5% (66 of 81) in the control group [s1]. The between-group difference was 5.6 percentage points, with a 95% confidence interval of minus 5.7 to 13.8 and a p-value of 0.03 for non-inferiority [s1].
The per-protocol analysis pointed the same way. There, the favourable outcome rate was 89.0% (146 of 164) in the short-regimen group against 83.1% (64 of 77) in the control group, a difference of 5.9 percentage points (95% CI minus 7.9 to 11.7; p 0.02 for non-inferiority) [s1]. In both analyses the shorter, tailored regimen was not worse than the long one — and its point estimate sat slightly higher.
Safety was where the two courses parted most clearly. Grade 3 or higher adverse events occurred significantly less often in the short-regimen group than in the control group, 27.7% against 43.0%, with a p-value below 0.01 [s1]. The authors concluded that the radiographic severity-stratified short regimen was non-inferior to the traditional longer regimen for favourable outcomes and demonstrated better safety [s1].
Why stratifying by imaging is the interesting part
Short all-oral regimens for drug-resistant TB are no longer novel in themselves; WHO has moved the field toward them. What this trial tests is a way to match regimen length to the individual patient by using a cheap, widely available tool — a chest radiograph — rather than treating every patient for the same fixed duration. If disease severity on imaging can safely sort patients into shorter and longer courses, that offers a route to spare lighter cases from unnecessary months of drugs while keeping protection for those with more extensive disease.
The limits
This was an open-label trial, meaning patients and clinicians knew which regimen was being given, which can influence how outcomes are assessed. It was conducted entirely in China, and drug-resistance patterns, imaging practice and health-system capacity differ across the high-burden countries that carry most of the world's drug-resistant TB. The sample was modest — 259 people — and the confidence intervals around the difference, which extend into negative values, reflect that. A non-inferiority result also establishes that the new regimen is not meaningfully worse, not that it is better.
Still, the direction is consistent across analyses and the safety advantage is real, and it feeds a larger effort: making treatment of a disease that most of its patients still cannot access shorter, better tolerated and easier to complete.
That effort has a track record behind it. Global efforts to combat TB have saved an estimated 83 million lives since the year 2000, and the disease is preventable and curable, WHO says [s2]. It is also a major cause of deaths related to antimicrobial resistance in its own right [s2] — which makes a drug-resistant regimen that patients can finish, without the injectable agents and long durations that drove earlier courses off the rails, more than a convenience. Each patient who completes treatment is one less source of onward, harder-to-treat transmission. The fact that the shorter regimen's advantage held in both the modified intention-to-treat and per-protocol populations, rather than appearing only among the patients who adhered perfectly, is part of why the authors landed on non-inferiority with better safety.
Sources
- Efficacy and safety of a radiographic severity-stratified all-oral short treatment regimen for multidrug/rifampicin-resistant tuberculosis in China — Clinical Microbiology and Infection, 29 June 2026
- Tuberculosis (fact sheet) — World Health Organization
Sources
- Efficacy and safety of a radiographic severity-stratified all-oral short treatment regimen for multidrug/rifampicin-resistant tuberculosis in China: an open-label, multicentre, randomised controlled, noninferiority trial — Clinical Microbiology and Infection , June 29, 2026
- Tuberculosis (fact sheet) — World Health Organization
More on
Two trials cut treatment length in high-burden settings without losing ground
A six-month regimen for drug-resistant tuberculosis matched the standard of care in South Africa. Four days of antibiotics matched eight in African children hospitalised with severe pneumonia.
Drug-resistant TB care met WHO's success target in war-torn South Kivu, Congo
An all-oral, bedaquiline-based programme treated 277 patients through six years of conflict in eastern Congo and reached 77.6% treatment success — above WHO's 75% mark — but mortality stayed at 13.4%.
TB scoring tool caught 97% of cases in children with HIV, four-country trial finds
External validation of the PAANTHER decision algorithm in 277 children across four African countries recorded 93.3% negative predictive value and 97.4% sensitivity against a composite reference standard.
A rifampicin-free four-month TB regimen matched standard care in a small trial
In a 122-patient South African proof-of-concept study, a four-month all-oral combination of quabodepistat, delamanid and bedaquiline cleared TB from sputum as often as the six-month standard.