Analysis of influencing factors and predictive model construction of anti-tuberculosis drug-induced liver injury

GU Zhengtai; LIU Weiwei; GU Lingyun; MA Chenbin

doi:10.12173/j.issn.1004-5511.202503224

Abstract：Objective To analyze the influencing factors for anti-tuberculosis drug-induced liver injury(ATB-DILI)and construct a prediction model.Methods Clinical data were retrospectively collected from patients who received anti-tuberculosis(TB)treatment and completed follow-up from January 2019 to October 2022 at Rugao People's Hospital.Patients from January 2019 and December 2019 were included in the training set,and patients from January 2020 to October 2022 were included in the testing set.A nomogram was constructed,and its predictive performance was evaluated using the receiver operating characteristic(ROC)curve and its area under the curve(AUC),calibration curve,decision curve analysis,and clinical impact curve and verified in the testing set.Packages such as DynNom were used to publish nomograms to the network and develop dynamic nomograms that predicted ATB-DILI risk.Results A total of 1,396 TB patients were included,with an ATB-DILI incidence rate of 10.03%(140/1,396),comprising 872 in the training set and 524 in the testing set.The results of the multivariate Logistic regression analysis revealed that older age[OR=1.03,95%CI(1.01,1.05)],diabetes[OR=2.66,95%CI(1.40,5.07)],smoking history[OR=1.98,95%CI(1.09,3.59)],liver-related diseases[OR=2.07,95%CI(1.05,4.10)],high aspartate aminotransferase(AST)levels[OR=1.03,95%CI(1.02,1.05)],high gamma-glutamyl transferase(GGT)levels[OR=1.18,95%CI(1.06,1.30)],and high total bilirubin(TBil)levels[OR=1.22,95%CI(1.16,1.29)]were independent risk factors for ATB-DILI,while high albumin(Alb)levels[OR=0.87,95%CI(0.81,0.93)]were an independent protective factor(P<0.05).The nomogram for both the training set and testing set showed a high degree of fit with the ideal curve,and both demonstrated significant net benefit characteristics,with an AUC of 0.88、0.86 and high clinical impact gains.Conclusion In this study,the dynamic nomogram developed based on age,diabetes,smoking,liver-related diseases,Alb,AST,GGT and TBil can effectively identify patients with high ATB-DILI risk,and provide theories and methods for patients to implement more active prevention strategies.