Abstract 12072: Predictors of Cardiovascular and Non-Cardiovascular Death: From the Clinic to the Lab
Introduction: Mortality risk estimates are often limited by the availability of individual patient risk factor data. We developed a series of prediction models for cardiovascular (CV) and non-cardiovascular (non-CV) death using increasing levels of information starting with the 1) standard history and physical examination, with incremental addition of 2) ECG and common lab values, 3) catheterization variables, and 4) research lab values.
Methods: BARI 2D enrolled 2368 patients with type 2 diabetes and documented stable coronary artery disease from 6 countries. Patients were treated according to national diabetes, hypertension and cholesterol guidelines. Average follow-up was 5.4 years, and an independent adjudication committee classified 149 deaths as cardiovascular and 167 as non-cardiovascular. A series of four multivariable Cox regression models corresponding to the above levels of information were generated for CV and non-CV death, and Harrell c-statistics were used to evaluate model discrimination.
Results: A history of heart failure, COPD and eGFR were significant independent predictors of both types of death. Classic cardiac risk factors including history of hypertension, ST depression, number of vessels with ≥50% stenosis, and reduced LV function were predictors of death due to cardiovascular causes whereas elevated CRP and total number of lesions were associated with death due to non-cardiovascular causes. Model discrimination for CV death improved with the addition of lab, ECG and angiographic information (c-statistic=0.720 with demographic and clinical predictors only, c-statistics=0.770 with all candidate variables). Improvement in model discrimination across the four models for non-CV death was less marked (c-statistics 0.740 to 0.766).
Conclusion: Incremental prediction models provide clinically useful information as a patient is sequentially evaluated from the clinical to the lab.
Author Disclosures: M.M. Brooks: Research Grant; Significant; Genway Biotech, Inc.. B.R. Chaitman: None. M. Singh: None. H. Vlachos: None. G. Steiner: None. A. Krishnaswami: None. T.J. Orchard: None. F. Feit: Ownership Interest; Significant; Johnson and Johnson, Medtronic, Boston Scientific. C. Rihal: None. O. Marroquin: None.
- © 2014 by American Heart Association, Inc.