Treatment shortening and relapse control in drug-susceptible pulmonary tuberculosis: a network meta-analysis of randomized trials.

UNLABELLED: Drug-susceptible pulmonary tuberculosis (DS-PTB) is conventionally treated with a 6-month multidrug regimen, but prolonged treatment may compromise adherence, prompting efforts to develop shorter and effective therapeutic strategies. However, the comparative long-term effectiveness of alternative regimens, particularly their ability to prevent relapse, remains uncertain. We systematically searched multiple databases from inception through August 31, 2025, for randomized controlled trials enrolling adolescents and adults with DS-PTB. Regimens were classified according to treatment duration and architecture, including standard 6-month therapy, rifamycin-intensified strategies, structurally modified 6-month regimens, 4-month treatment-shortening approaches, and intended 8-week multidrug regimens. A random-effects pairwise meta-analysis and a fixed-effect consistency network meta-analysis (NMA) compared treatment success, relapse/recurrence-related outcomes, and mortality. Eight randomized controlled trials involving 10,244 participants were included. Six source trials involving 8,457 randomized participants contributed to the quantitative synthesis and informed a 14-node NMA; outcome data from 6,752 participants contributed to 12 shortened-regimen-versus-standard comparisons in the pairwise meta-analysis. Two trials involving 1,787 participants were summarized narratively because their thrice-weekly control regimens were not considered exchangeable with the daily reference regimen. In the pooled pairwise analysis, shortened-duration regimens were associated with a higher risk of trial-defined unfavorable long-term outcomes than the 6-month standard regimen (RR = 1.68, 95% CI 1.37-2.06; P < 0.001; I² = 39.9%). Treatment success was close to that of the standard regimen for the 4-month daily rifapentine-moxifloxacin regimen evaluated in Study 31/A5349 (RR = 0.98, 95% CI 0.94-1.01) and the 6-month regimen with once-weekly rifapentine-moxifloxacin continuation therapy evaluated in RIFAQUIN (RR = 1.01, 95% CI 0.93-1.09). In contrast, all four intended 8-week TRUNCATE-TB regimens showed lower treatment success and higher relapse/recurrence-related risk estimates than the standard regimen. Mortality was uncommon, and mortality comparisons were imprecise. These findings suggest that long-term outcomes should be interpreted in relation to regimen design, dosing, and treatment duration rather than duration alone. The standard 6-month regimen remains a reliable reference, and conclusions about individual alternatives should be based on effect estimates and uncertainty rather than rankings. Because no closed loop was formed by independent trial evidence, formal inconsistency assessment was not feasible. SYSTEMATIC REVIEW REGISTRATION: https://www.crd.york.ac.uk/PROSPERO/, identifier CRD420251267175.