Risk factors for multidrug-resistant tuberculosis: a predictive model study.

Objective: To investigate the risk factors associated with Multidrug-resistant tuberculosis (MDR-TB) in people with drug-resistant tuberculosis (DR-TB) and develop a predictive model. Methods: A total of 893 individuals with DR-TB treated at Wenzhou Central Hospital from January 2018 to December 2022 were included in the study after excluding 178 individuals with incomplete clinical and laboratory data, leaving 715 individuals for analysis. Data on demographic information, baseline clinical characteristics, laboratory and imaging results, and clinical diagnosis were collected to identify the risk factors for MDR-TB and establish a predictive model. Results: Multivariate logistic regression analysis identified residence in rural areas, retreatment of TB, presence of pulmonary cavity, uric acid (UA) ≥ 346 μmol/L and c-reactive protein (CRP) < 37.3 mg/L as independent risk factors for MDR-TB in individuals with DR-TB. A nomogram model was constructed using these five factors to predict the risk of MDR-TB, with an area under the ROC curve (AUC) of 0.758 for the training group and 0.775 for the validation group. Calibration curve analysis showed good agreement between predicted and actual MDR-TB incidence in both groups, and decision curve analysis showed that the nomogram model had a higher rate of clinical net benefit. Conclusion: This study suggests that residence, types of TB treatment, presence of pulmonary cavity, UA and CRP are associated with MDR-TB occurrence in individuals with DR-TB, and the nomogram model developed in this study shows promising predictive value.