Development And Validation Of The Risk Model For Predicting Clinical Outcomes After PCI In Patients With Coronary Artery Disease And Diabetes

Objective: We aim to develop and validate a simple risk model for predicting the risk of clinical outcomes in patients with DM and multivessel coronary artery disease(CAD)undergoing PCI.Methods: A total of 458 CAD patients with DM undergoing PCI in Department of Cardiovascular Medicine,Hospital of Xinjiang Medial University were enrolled as the derivation cohort.The validation cohort including 674 CAD patients with DM undergoing PCI in the First Affiliated Hospital of Zhengzhou University,all were hospitalized from January 2008 to December 2017.The primary endpoint was cardiac mortality(CM).the predictors for CM were identificated using Cox multivariate egression with stepwise inclusion of variables,and establish three different COX regression models to compare the prediction performance of different models.The discrimination and calibration of the risk model were assessed by ROC curve and calibration curve,the final model showed by Nomogram.The model was further validated using ROC analysis in the validation cohort,and the derivation group and validation group were compared by using AUC.Results: The average follow-up time was 48 months,the overall incidence of CM was 29,Cox multivariate regression analyses showed that heart rate,NTpro BNP,PLT(platelet),TG(Triglycerides)Were most important predictors for CM,and established stepwise model.The simple risk model had high discriminatory ability for CM(AUC:0.826,P<0.0001)in the derivation cohort,(AUC:0.795,P<0.0001)In the validation cohort,with good calibration in both cohorts.Conclusion: A simple risk model,which included four independent variables(heart rate,PLT,NT-pro BNP,TG),might be a good tool for predicting CM in patients with DM undergoing PCI,and may possibly be employed to complement clinical assessment and guide management based on CV risk.