Abstract:

Objective: To explore the risk factors of drug-induced liver injury in patients with tuberculosis, and to establish and validate a nomogram prediction model. Methods: The clinical data of 498 patients with anti-tuberculosis drug-induced liver injury (ATB-DILI) in the Third People’s Hospital of Zhenjiang from January 2017 to June 2023 were retrospectively collected. The patients were divided into modeling group and validation group according to the ratio of 7:3. The independent risk factors were screened using LASSO regression and multivariate logistic regression analysis, a nomogram prediction model was established for internal and external verification. Results: After LASSO screening variables, multivariate logistic regression analysis showed that hyperglycemia (OR=1.183, 95%CI: 1.037-1.349), high hemoglobin (OR=1.028, 95%CI: 1.011-1.045), extrapulmonary tuberculosis (OR=2.159, 95%CI: 1.240-3.759), and high international normalized ratio (OR=5.767, 95%CI: 1.259-26.421) were independent risk factors for ATLI, while high uric acid (OR=0.998, 95%CI: 0.996-0.999) and low PLT (OR=0.990, 95%CI: 0.986-0.995) were protective factors. The nomogram model was constructed based on the above related factors. The area under the receiver operating characteristic (ROC) curve was 80.3% and 79.1% in the modeling groups and validation groups, calibration curve P-value were 0.318 and 0.605, the decision curve showed that the nomogram model had certain clinical practicability in the high risk threshold range 9%-85% and 1%-93%. Conclusion: The nomogram model for risk predicting ATB-DILI among inpatients with tuberculosis in this study has good predictability, consistency and clinical practicability, which can provide individualized basis for the prevention and control of anti-tuberculosis drug-induced liver injury.

Key words: Tuberculosis, Drug-induced liver injury, Forecasting, Nomograms