Abstract
OBJECTIVE: Acute kidney injury (AKI) is a common and serious adverse effect during tuberculosis (TB) treatment in clinical settings, particularly in patients with drug-resistant TB. AKI may lead to treatment interruption and poor prognosis. Early identification of patients at high risk for AKI is crucial to improve clinical outcomes. METHODS: We retrospectively enrolled 571 TB patients, divided into training and validation cohorts. LASSO and multivariate logistic regression were used to identify risk factors, and the nomogram was evaluated using AUC, calibration, and decision curve analysis (DCA). RESULTS: This study included 571 patients with TB. In this study, five variables (age, hypertension, diabetes, Scr, and ALB) were included to construct a nomogram for predicting AKI caused by drug resistance to TB. The AUC of the training set and validation set were 0.809 (95% CI: 0.7480-0.871, P < 0.001) and 0.841 (95% CI: 0.765-0.918, P < 0.001), respectively, indicating that the prediction model had good discriminative performance. The calibration curve shows that the predicted values of the model are basically consistent with the actual values, indicating good performance. DCA suggests that almost all ranges of TB patients can benefit from this new predictive model, indicating good clinical utility. CONCLUSION: The nomogram model of AKI caused by drug resistance to TB established in this study has good predictive value and helps identify high-risk populations.