Usefulness of the RISK-PCI score to predict stent thrombosis in patients treated with primary percutaneous coronary intervention for ST-segment elevation myocardial infarction: a substudy of the RISK-PCI trial

Abstract

Stent thrombosis (ST) is an important cause of death after primary percutaneous coronary intervention (pPCI). This substudy aimed at evaluating the usefulness of the RISK-PCI score, originally developed for the prediction of 30-day major adverse cardiovascular events, to predict the occurrence of ST after pPCI. We analyzed 1972 consecutive patients who underwent pPCI with stent implantation between February 2007 and December 2009. Early ST (EST), late ST (LST), and cumulative 1-year ST (CST) were the predefined end points. Definite, probable, and possible ST were included. Models discrimination and calibration to predict ST was tested using receiver-operating characteristics curves and the goodness-of-fit (GoF) test. Sensitivity analyses and 1000-resample bootstrapping were used to evaluate the model's performance. The rates of EST, LST, and CST were 4.6, 1.4, and 6.0 %, respectively. Compared with controls, the cumulative ST group was associated with much higher rates of adverse clinical outcomes at 30-day follow-up (adjusted odds ratio (OR) for death 6.45, adjusted OR for major bleeding 4.41) and at 12-month follow-up (adjusted OR for death 7.35, adjusted OR for major bleeding 4.56). Internal validation confirmed a reasonably good discrimination and calibration of the RISK-PCI score for the prediction of EST (area under the curve (AUC) 0.71, GoF 0.42), LST (AUC 0.69, GoF 0.36), and CST (AUC 0.70, GoF 0.22) after pPCI. ST after pPCI is associated with adverse 30-day and 1-year clinical outcomes. We conclude that the risk of ST could be accurately assessed using the RISK-PCI score, which might help in deciding upon measures aimed at preventing adverse prognosis.