Survival of patients undergoing rescue percutaneous coronary intervention: development and validation of a predictive tool.

OBJECTIVES: This study sought to develop a tool for predicting an individual's risk of mortality following rescue percutaneous coronary intervention (PCI). BACKGROUND: Although fibrinolytic therapy is appropriate and improves survival for certain ST-segment elevation myocardial infarction patients, a substantial proportion suffer ongoing myocardial ischemia, a class I indication for emergent percutaneous coronary intervention (rescue PCI). METHODS: Using the National Cardiovascular Data Registry (NCDR), rescue PCI was defined as nonelective PCI following failed fibrinolysis in patients with continuing or recurrent myocardial ischemia. Multivariable logistic regression was used to determine mortality predictors and the C-statistic for model discrimination. The NCDR-RESCUE (Real-World Estimator of Survival in Catheterized STEMI Patients Following Unsuccessful Earlier Fibrinolysis) score was developed using a shortened list of 6 pre-angiographic variables and 70% of the cohort; performance was subsequently validated against the remaining 30%. RESULTS: Among 166,516 PCI procedures on patients with an admission diagnosis of ST-segment elevation myocardial infarction, 8,007 (4.8%) represented rescue PCI. In-hospital mortality occurred in 464 (5.8%). Factors in the final model were age, glomerular filtration rate, history of congestive heart failure, insulin-treated diabetes, cardiogenic shock, and salvage status. The NCDR-RESCUE score effectively segregated individuals into 6 clinically meaningful risk categories, with 0.4% (0.0% to 1.3%), 1.6% (0.9% to 2.4%), 7.6% (5.3% to10.4%), 27.5% (20.7% to 35.1%), 64.2% (49.8% to 76.9%), or 100% (59.0% to 100.0%) risk, respectively, of in-hospital mortality (mean ± 95% confidence interval, C-index = 0.88, Hosmer-Lemeshow p = 0.898). CONCLUSIONS: In-hospital mortality risk among individuals undergoing rescue PCI varies from minimal to extreme and can be easily calculated using the NCDR-RESCUE score. This information can be of value in counseling patients, families, and referring caregivers.

Duke Scholars

Author Sunil Vadlakonda Rao Medicine, Cardiology

Author Matthew Todd Roe Medicine, Cardiology