Cavitary pulmonary disease is associated with higher bacillary burden and has been linked to increased post-treatment recurrence in pulmonary tuberculosis (TB). International guidance has suggested extending therapy for patients with cavitary disease, particularly when cultures remain positive at 2 months, but contemporary real-world evidence is limited. This study evaluated whether treatment extended beyond 200 days was associated with reduced bacteriologically confirmed recurrence among patients with drug-susceptible pulmonary tuberculosis who had a cavity on chest radiograph.
The analysis used a prospective multicenter cohort recruited from 18 centers in South Korea between July 2019 and June 2023. From 1,204 enrolled pulmonary TB patients, those with any drug resistance, those not treated with the standard HRZE regimen, and patients who did not achieve treatment success were excluded. The final analytic sample comprised 159 pan-susceptible patients with cavitary lesions on baseline chest radiograph who completed post-treatment follow-up.
Baseline demographics, comorbidities, clinical labs, chest radiograph findings, and sputum AFB smear and culture results were recorded. Adverse drug reactions were monitored at predefined intervals and analyzed as binary outcomes (presence/absence). The primary outcome was bacteriologically confirmed recurrence after treatment completion among patients classified as cured or having successfully completed treatment.
Patients were categorized by total treatment duration. Standard treatment was defined as ≤200 days; extended treatment was >200 days, reflecting an allowance of ~2 weeks beyond the standard 6-month regimen to distinguish meaningful prolongation from minor variation. This fixed 200-day cutoff was used to avoid immortal-time bias related to treatment duration being determined during therapy.
Of 159 patients, 58 (36.5%) received standard treatment and 101 (63.5%) received extended treatment. The extended group had higher baseline AFB smear positivity (56.4% vs. 29.3%, p = 0.002). Two-month culture positivity rates were low and comparable between groups (6.3% vs. 4.4%, p > 0.999), and mean time to culture conversion did not differ significantly (8.00 ± 3.91 vs. 7.45 ± 3.73 weeks, p = 0.454).
Four bacteriologically confirmed recurrences (2.5% overall) were observed: two in the standard-treatment group and two in the extended-treatment group. Recurrences occurred at varied intervals after treatment completion: 197 days and 414 days in the standard group; 295 days and 378 days in the extended group. Molecular typing to distinguish relapse versus reinfection was not available, so definitive classification was not possible; earlier events were considered more compatible with relapse based on timing alone.
Overall adverse event rates were comparable between groups. Medication interruptions attributed to adverse drug reactions were numerically more frequent in the extended-treatment group, but the difference was not statistically significant. Adverse drug reactions were analyzed as present versus absent due to inconsistent severity grading in the dataset.
Because total treatment duration is determined during therapy and only four recurrence events occurred, the authors used a 200-day fixed-time exposure classification and applied Firth’s penalized logistic regression to mitigate bias from sparse events and separation. Age and sex were included a priori in multivariable models; variables with p < 0.05 in univariable screening were considered for inclusion while limiting model complexity to avoid overfitting.
Unadjusted recurrence proportions did not differ between standard and extended treatment groups (3.4% vs. 2.0%, p = 0.966). In the exploratory adjusted fixed-time analysis, extended treatment was associated with a numerically lower risk of recurrence (adjusted odds ratio 0.07), but this result did not reach conventional statistical significance (95% CI 0.01–1.06; p = 0.056) and is imprecise because only four events occurred. Low body mass index (<18.5 kg/m²) was associated with recurrence in multivariable analysis; however, that estimate was extremely imprecise and should be interpreted with caution.
Key limitations include the small number of recurrence events, absence of molecular genotyping to discriminate relapse from reinfection, and potential residual confounding. Adverse event severity was not uniformly graded. Given these constraints, the authors conclude that the presence of a cavity on chest radiograph alone may not be sufficient justification for routine treatment extension beyond 200 days. The exploratory favorable point estimate for extended treatment is imprecise and not definitive.
These real-world data suggest that routine extension of therapy solely for cavitary disease is not clearly supported. The findings underscore the need for larger prospective studies that systematically capture microbiologic response (including serial cultures and molecular typing) and detailed radiologic burden to identify which patients with cavitary TB might benefit from individualized extension of therapy. Clinicians should balance potential benefits of prolonged therapy against increased cumulative drug exposure, potential adverse events, interruptions, and costs when considering extension.