Predicting Likelihood for Coronary Artery Bypass Grafting After Non–ST-Elevation Myocardial Infarction: Finding the Best Prediction Model

Ali Shafiq, Jae Sik Jang, Faraz Kureshi, Timothy J. Fendler, Kensey Gosch, Phil G. Jones, David J. Cohen, Richard Bach, John A. Spertus

Research output: Contribution to journalArticle

2 Scopus citations

Abstract

Background Up to half of patients with non–ST-elevation myocardial infarction (NSTEMI) do not receive dual antiplatelet therapy before angiography “pretreatment” because of the risk of increased bleeding if coronary artery bypass grafting (CABG) operation is needed. Several models have been published that predict the likelihood of CABG after NSTEMI, but they have not been independently validated. The purpose of this study was to validate these models and improve the best one. Methods We studied patients with NSTEMI who were enrolled in the 24-center Translational Research Investigating Underlying Disparities in Acute Myocardial Infarction Patients’ Health Status (TRIUMPH) registry between 2005 and 2008. Previous CABG prediction models were assessed using c-statistics and calibration assessments to determine the best model. Variables from TRIUMPH likely to be associated with CABG were tested to see whether they could improve the best model's performance. Results Among 2,473 patients with NSTEMI, 11.8% underwent in-hospital CABG. C-statistics for the Modified Thrombolysis in Myocardial Infarction, Treat Angina With Aggrastat and Determine the Cost of Therapy With an Invasive or Conservative Strategy–Thrombolysis in Myocardial Infarction 18, Poppe, and Global Risk of Acute Coronary Events (GRACE) models were 0.54, 0.61, 0.61, and 0.62, respectively. The GRACE model showed the best discrimination and calibration. From the TRIUMPH registry, preselected variables were added to the GRACE model but did not significantly improve model discrimination. A GRACE model risk score of less than 9 had high sensitivity (96%), thus making it useful for predicting patients with NSTEMI who were at low risk for requiring CABG, which included approximately 21% of patients with NSTEMI. Conclusions This study could not improve on the GRACE model, which had the best predictive value for identifying a need for CABG after NSTEMI with a broader range of predicted risk levels and high sensitivity, especially in patients with scores lower than 9.

Original languageEnglish
Pages (from-to)1304-1311
Number of pages8
JournalAnnals of Thoracic Surgery
Volume102
Issue number4
DOIs
StatePublished - Oct 1 2016

Fingerprint Dive into the research topics of 'Predicting Likelihood for Coronary Artery Bypass Grafting After Non–ST-Elevation Myocardial Infarction: Finding the Best Prediction Model'. Together they form a unique fingerprint.

  • Cite this