Comparison of Effectiveness Between Delamanid and Bedaquiline Among Patients with Multidrug-Resistant Tuberculosis: A Markov Model Simulation Study

Background: No clear evidence on the comparative effectiveness of delamanid (DLM) and bedaquiline (BDQ) has been published. Objective: This study aims to estimate the incremental effectiveness of DLM versus BDQ in patients with multidrug-resistant tuberculosis (MDR-TB). Methods: We developed a Markov model based on a cohort with MDR-TB, which consisted of success, failure, loss to follow-up, and death. The cohort simulation was conducted assuming each patient was 36 years old and, lived until age 82, and that the cycle length was 1 year. Patients with an inadequate response to DLM, the background regimen, or BDQ for 2 years were transitioned through the next treatment sequence. We evaluated the incremental effectiveness of the drugs using the quality-adjusted life-year (QALY) resulting from this Markov model over a lifetime. Results: The incremental effectiveness of DLM (13.96 QALYs) was greater by 2.44 QALYs per patient than the background regimen (11.52 QALY), while the incremental effectiveness of BDQ (10.40 QALY) was higher by 1.14 QALY per patient than the background regimen (9.26 QALY). Consequently, the incremental effectiveness of DLM was relatively more positive by 1.30 QALY than those of BDQ per patient over a lifetime. Limitations: This study is a simulation study. Therefore, the treatment sequence for patients may be different in the real world. Conclusions: Our lifetime simulated data found that DLM was relatively more favorable than BDQ. A Markov model can be considered an alternative approach when there is an absence of head-to-head clinical data.