Prediction of treatment outcomes for multidrug-resistant tuberculosis by whole genome sequencing.

BACKGROUND: Whole genome sequencing (WGS) has been proposed to be a powerful tool to predict drug resistance for antitubercular drugs. However, the feasibility of WGS in predicting final treatment outcomes for multidrug-resistant tuberculosis (MDR-TB) patients remains unclear. PATIENTS AND METHODS: In the prospective observational study conducted from January 2014 to September 2016, MDR-TB patients were enrolled consecutively. Genotypic drug sensitivity testing was performed via WGS using the culture isolates. Patients were followed for two years to determine the treatment outcomes. Multivariate analysis was used to identify the association between information provided by WGS and the final treatment outcomes. RESULTS: A total of 123 patients with MDR-TB were included in this study. The overall favorable treatment outcome rate was 60.2%. Multivariate analysis showed that independent risk factors associated with unfavorable treatment outcome including high-level moxifloxacin phenotypic resistance (OR, 4.362; 95%CI, 1.364-13.950; p=0.013), cycloserine phenotypic resistance (OR, 7.457; 95%CI, 1.644-33.819; p=0.009), mutations causing high level fluoroquinolones resistance (OR, 3.947; 95%CI, 1.195-13.034; p=0.024), and ethA mutation (OR, 3.817; 95% CI, 1.154-12.823; p=0.028). WGS costs for each patient are yen450 ($63) and the average turnaround time was one week. CONCLUSIONS: In summary, WGS showed a promising feasibility in predicting treatment outcomes for MDR-TB patients within a clinically relevant time frame.