Validation of Mayo clinic risk adjustment model for in-hospital complications after percutaneous coronary interventions, using the National Heart, Lung, and Blood Institute dynamic registry

Abstract

Objectives

We sought to validate the recently proposed Mayo Clinic risk score model for complications after percutaneous coronary interventions (PCI), using an independent data set.

Background

The Mayo Clinic risk score has eight simple clinical and angiographic variables for the prediction of complications defined as either death, Q-wave myocardial infarction, emergent or urgent coronary artery bypass graft surgery, or cerebrovascular accident after PCI. External validation using an independent data set is lacking.

Methods

A total of 3,264 patients undergoing PCI at each of the 17 sites in the National Heart, Lung, and Blood Institute's Dynamic Registry during two enrollment periods (July 1997 to February 1998 and February to June 1999) were studied. Logistic regression was used to model the calculated risk score and major procedural complications. The expected number of complications, with 95% confidence bounds (CBs), was also calculated.

Results

There were 96 (2.94%) observed procedural complications, and the Mayo Clinic risk score predicted 93.5 events (2.86%; 95% CB 2.32% to 3.41%; p = NS). The Hosmer-Lemeshow goodness-of-fit p value was 0.28, and the area under the receiver operating curve was 0.76, indicating excellent overall discrimination. There were no statistical differences between observed and predicted procedural complications using the Mayo Clinic risk score among the most selected high- and low-risk subgroups.

Conclusions

Eight variables were combined into a convenient risk scoring system that accurately predicts cardiovascular complications after PCI. The Mayo clinic predictive model for procedural complications yielded excellent results when applied to a multi-center external data set.