*3.3. Evaluating the Role of Surgery Through IPD*

A recent IPD based on the large MDR-TB cohort coordinated by McGill University [33–35] utilized a sophisticated analysis (propensity score matching) to evaluate the benefits offered by surgery. Individual patient data from 26 cohort studies were analysed, including clinical features and information on both medical and surgical therapy. Primary analyses compared treatment success (cure and completion) to a combined outcome of failure, relapse, or death. The effects of all forms of resection surgery, pneumonectomy, and partial lung resection were evaluated [35].

The final analysis was conducted on 4,238 patients from 18 surgical studies and 2,193 from 8 non-surgical ones. Pulmonary resection surgery (478 patients) was associated with improved treatment success (adjusted odds ratio (aOR), 3.0; 95% confidence interval (CI), 1.5–5.9), but pneumonectomy was not (aOR, 1.1; 95% CI, 0.6–2.3). Treatment success was achieved in 95.2% of patients undergoing surgery after culture conversion compared with 91.2% of those who had surgery before it (aOR, 2.6; 95% CI, 0.9–7.1).

Patients undergoing partial lung resection achieved better treatment success and lower failure/death rates than patients who had either pneumonectomy or no surgery. The median duration of medical therapy was 20 months (interquartile range [IQR], 13.7–24.0 months) for those who had surgery after culture conversion versus 29 months (IQR, 22–45 months) for those undergoing surgery before conversion. The loss to follow-up was lower among patients who had surgery (11%; 95% CI, 4–17%) than among those who had not (22%; 95% CI, 14–31%).

The authors concluded that, among MDR-TB patients, partial lung resection (but not pneumonectomy) was associated with improved treatment success, although selection bias cannot be excluded [35]. This finding can be explained with the lower rate of mortality among surgical versus non-surgical TB patients. Furthermore, patients undergoing surgery had, overall, more severe drug-resistance profiles and more extensive diseases [35]. Importantly, both surgical and non-surgical patients were rather young with a low probability of confounding co-morbidities [35].

A summary of the available evidence is reported in Table 1


*Appl. Sci.* **2020**, *10*, 2734


*Appl. Sci.* **2020**, *10*, 2734

**Table 1.** *Cont.*
