Background: Primary percutaneous coronary intervention (PPCI) is the default treatment for patients with ST elevation myocardial infarction (STEMI) and carries a higher risk of adverse outcomes when compared with elective and urgent PCI. Conventional PCI risk scores tend to be complex and may underestimate the risk associated with PPCI due to under-representation of patients with STEMI in their datasets. This study aimed to develop a simple, practical and contemporary risk model to provide risk stratification in PPCI.
Methods: Demographic, clinical and outcome data were collected for all patients who underwent PPCI between January 2009 and October 2013 at the Northern General Hospital, Sheffield. Multiple regression analysis was used to identify independent predictors of mortality and to construct a risk model. This model was then separately validated on an internal and external dataset.
Results: The derivation cohort included 2870 patients with a 30-day mortality of 5.1% (145 patients). Only four variables were required to predict 30-day mortality: age [OR:1.047, 95% CI:1.031–1.063], call-to-balloon (CTB) time [OR:1.829, 95% CI:1.198–2.791], cardiogenic shock [OR:13.886, 95% CI:8.284–23.275] and congestive heart failure [OR:3.169, 95% CI:1.420–7.072]. Internal validation was performed in 693 patients and external validation in 660 patients undergoing PPCI. Our model showed excellent discrimination on ROC-curve analysis (C-Stat = 0.87 internal and 0.86, external), and excellent calibration on Hosmer-Lemeshow testing (p = 0.37 internal, 0.55 external).
Conclusions: We have developed a bedside risk model which can predict 30-day mortality after PPCI using only four variables: age, CTB time, congestive heart failure and shock.