Introduction: The accurate assessment of individual risk can be of great value to facilitate the prevention of recurrent coronary thrombosis events (CTEs) in patients after percutaneous coronary intervention (PCI) with Drug-Eluting Stents (DES). However, current prediction models in common use were often formulated in white populations and were of limit predicting value in Chinese patients. The current study is aimed at developing and validating a prediction model for recurrent CTEs from 2 contemporary Chinese cohorts of patients after PCI with DES. We formulated a CTEs prediction model from PROMISE trial, a PR ospective O bservational M ulticenter cohort of I schemic and h E morrage risk after PCI. Hypothesis We assessed the hypothesis that the PROMISE model was an effective tool for predicting 2-year CTEs risk in Chinese patients after PCI with DES. Methods 7998 Chinese patients underwent DES-PCI from PROMISE trial with 2-year follow up were used as the derivation cohort to develop the PROMISE model. The external validation was evaluated in one independent Chinese cohort with 3824 participants. Cardiac death, myocardial infarction, definite and probable in-stent thrombosis were defined as CETs. We compared candidate variables between CETs and non-CETs groups. Backward selection with an entry criterion of p<0.05 were used to formulate the model. The model performance was assessed by discrimination C statistics and calibration χ2. We also assessed the internal and external validation. Results Over 2-year follow up in the derivation cohort, 151 patients developed CTEs. Five variables including, previous revascularization, baseline elevated cTNI, left ventricular ejection fraction, residual SYNTAX score after PCI and estimated glomerular filtration rate were finally involved in the PROMISE model. The C statistics of the model was 0.750(95% confidence interval, 0.700-0.799), and the calibration χ2 was 7.2(p=0.612). In the subsequent external validation, the PROMISE model had C statistics of 0.607 (95%CI,0.508-0.706) , and the calibration χ2 was 14.0(p=0.121). Conclusion The PROMISE model had a good internal validation and moderate external validation and it was an effective tool for 2-year CTEs risk prediction in Chinese patients after PCI with DES.
Read full abstract