Multi-biomarker integration in predicting bedaquiline treatment response among diabetic MDR-TB patients

Diabetes mellitus significantly complicates multidrug-resistant tuberculosis (MDR-TB) treatment outcomes. This study investigated the predictive value of integrating bedaquiline pharmacokinetics, inflammatory markers, and glycemic parameters for treatment response in diabetic MDR-TB patients.

We conducted a retrospective cohort study of 186 patients with MDR-TB and diabetes mellitus receiving bedaquiline-containing regimens from January 2019 to December 2023. Bedaquiline trough concentrations, inflammatory markers (CRP, IL-6, TNF-α), and glycemic indices (HbA1c, fasting glucose) were measured at baseline and during follow-up. The primary outcome was sputum culture conversion or clinical improvement at 6 months. Multivariate logistic regression identified predictors of treatment response, and a predictive model was developed using 70 % training and 30 % validation cohorts.

Overall treatment response rate was 72.0 %. Responders demonstrated significantly higher bedaquiline trough concentrations and better glycemic control. The integrated predictive model incorporating bedaquiline concentration, baseline CRP, HbA1c, and HbA1c change achieved superior discrimination compared to individual biomarkers. Subgroup analysis revealed a critical interaction between glycemic control and drug exposure, with poor glycemic control patients showing 0 % response in the lowest bedaquiline quartile but 100 % response in higher quartiles.

Multi-biomarker integration effectively predicts bedaquiline treatment response in diabetic MDR-TB patients. The synergistic relationship between adequate drug exposure and glycemic control underscores the necessity for integrated therapeutic drug monitoring and diabetes management strategies in this high-risk population.