Abstract 18783: Predictive Model for High-Risk Coronary Artery Disease: An Analysis From PROMISE
Background: Patients with high-risk coronary artery disease (CAD) benefit from revascularization but may be difficult to identify clinically.
Methods: Using the Prospective Multicenter Imaging Study for Evaluation of Chest Pain (PROMISE) cohort of stable patients with suspected CAD randomized to coronary computed tomographic angiography as their initial diagnostic test (n=4520), this study categorized patients with high-risk CAD as obstructive left main, 3-vessel or 2-vessel CAD involving the proximal left anterior descending artery. Demographics, risk factors and risk scores were examined as potential predictors of high-risk CAD using stepwise logistic regression in the original cohort and 500 bootstrap samples. Restricted cubic splines were used to assess the association of continuous variables with high-risk CAD. Where relationships were non-linear, appropriate transformations were used. Discrimination was characterized using c-statistics, for which 95% confidence limits were determined using bootstrapping.
Results: High-risk CAD was identified in 111 patients (2.5%). In univariable analysis, older age, male sex, lower BMI, diabetes mellitus, metabolic syndrome, peripheral arterial disease, history of smoking, lower GFR, lower total cholesterol, lower HDL, typical angina symptoms, Diamond-Forrester pretest likelihood (DF) and Framingham Risk Score (FRS) were associated with high-risk CAD. In a multivariable analysis, DF (OR 1.054, 95% CI 1.037-1.072, for each 1-point increase >50%, but no effect for values <50) and high FRS (OR 1.030, 95% CI 1.019-1.041, for each 1-point increase) were strong independent predictors of high-risk CAD, producing a c-statistic of 0.75 (95% CI 0.70-0.80). Three additional variables (BMI, sedentary lifestyle, family history of premature CAD) were marginally predictive, but were excluded from the final model because they did not substantially improve model performance.
Conclusions: In contemporary practice, approximately 2.5% of patients without prior CAD who have stable symptoms and an indication for noninvasive cardiovascular testing have high-risk CAD. Using a simple combination of the readily obtained FRS and DF, high-risk CAD can be predicted with good discrimination.
Author Disclosures: J. Jang: None. S. Vemulapalli: Research Grant; Significant; American College of Cardiology, Abbott Vascular, Agency for Healthcare Research and Quality. Consultant/Advisory Board; Modest; Novella, Premiere. Other; Modest; Phillips Healthcare. M. Bhapkar: None. C.B. Fordyce: None. K.L. Lee: None. J.E. Udelson: None. U. Hoffmann: Research Grant; Significant; HeartFlow, Siemens Healthcare. J. Tardif: None. W.S. Jones: None. D.B. Mark: Research Grant; Significant; Eli Lilly, Bristol-Myers Squibb, Gilead Sciences, AGA Medical, Merck, Oxygen Biotherapeutics, AstraZeneca. Consultant/Advisory Board; Modest; Medtronic, CardioDx, St. Jude Medical. V.L. Sorrell: None. A. Espinoza: None. P.S. Douglas: Research Grant; Significant; General Electric, HeartFlow. M.R. Patel: Research Grant; Significant; AstraZeneca, CSL, HeartFlow, Janssen Research & Development, Maquet, Medtronic, NHLBI. Consultant/Advisory Board; Modest; AstraZeneca, Bayer, CSL, Genzyme, Janssen Research & Development, Medtronic, Merck & Co.
- © 2016 by American Heart Association, Inc.