Abstract 1925: Prognostic Value of Cardiac MRI in Patients Referred to Evaluate Known or Possible Coronary Artery Disease
Introduction: The assessment of left ventricular (LV) function and myocardial viability is part of a comprehensive evaluation for coronary artery disease (CAD) by cardiac MRI. The specific aim of this large community based study was to determine the predictors of all-cause mortality from cardiac MRI findings in patients referred to evaluate CAD.
Methods: The study included 838 consecutive patients referred to evaluate CAD in a community based program from 2000 to 2004. LV ejection fraction (LVEF) was assessed by cine MRI and myocardial viability by delayed enhancement (DE) imaging in a 1.5 T scanner. The extent of DE was scored on a scale from 0 – 68 representing 17 segments with a transmural extent graded 0 to 4. All-cause mortality was confirmed by the National Death Index. Cox Proportion Hazard was used to define the relative risk of death.
Results: Of 838 subjects studied, the mean age was 60.5 ± 13.9 years, 589 (70%) were men and 362 (43%) had known CAD. Cardiac MRI revealed abnormal DE in 381 (46%) and LVEF < 46% in 187 (22%) subjects. Those with known CAD accounted for 77% and 65% of the abnormal DE and low LVEF, respectively. The extent of DE correlated with lower LVEF (Pearson correlation coefficient 0.60, p < 0.001). During a mean follow up 1.7 ± 1.0 years, 31 individuals died (3.7%). Receiver operator curve analysis was performed to define optimal thresholds for predicting mortality with LVEF and DE. The best cutoff was LVEF < 46% and DE score ≤ 2 out of 68 possible points. In the univariate analyses LVEF < 46% and any abnormal DE were significantly associated with mortality with hazard ratios 12.20 (5.24, 28.57) and 2.99 (1.38, 6.49), respectively. However, history of CAD was not predictive of mortality with hazard ratio 1.59 (0.79, 3.23). In the multivariable analysis only LVEF < 46% remained a significant risk for death in a model that included history of CAD, abnormal DE and LVEF < 46%.
Conclusion: Overall the prognosis of patients in this community based study was good with a 2 year all-cause mortality rate of 3.7%. While abnormal DE and low LVEF each predicted mortality in the univariate analysis the multivariable analysis suggests low LVEF (< 46%) is the most important predictor for death in the first two years after MRI in a community based cohort with known or possible CAD.