A new prognostic index for acute heart failure and non- arrhythmic death in subjects with a cardiac defibrillator and ischemic cardiomyopathy

Subjects suffering from ischemic cardiomyopathy receiving a defibrillator (ICD) are still at high risk of heart failure and non-arrhythmic death. Predictors of increased risk are lacking in these patients. In a preliminary study in patients receiving an ICD for MADIT II criteria we found that age, 24 hour Systolic blood pressure, measured by Ambulatory blood pressure monitoring and creatinine, were independent predictors for acute decompensated heart failure and cardiac non arrhythmic death, while ejection fraction (EF), evaluated by 2D echo, was not predictive. In the present study we combined the same predictor factors in a Prognostic Index (PI) built according to the formula: 120 - age + m24hSBP - (creatinine*10). This PI was prospectively evaluated in 192 patients (all with ICD for MADIT II criteria and ischemic cardiomyopathy) for the combined endpoint of non-arrhythmic death and hospitalization for acute heart failure, in one year follow-up. Other variables assessed included EF, hemoglobin concentration, 24 hour mean heart rate, sodium levels, biventricular pacing and diabetes. We have registered 48 events (25%) in one year follow-up: 7 cardiac deaths and 41 hospitalizations for acute heart failure. The Cox multivariate analysis showed that low values of PI are the only independent predictor of events ( HR= 0.96; CI 95% 0.944-0.976, p < 0.0001). The ROC curve has shown the best cut-off for PI equal to 144, (AUC 0.79, p < 0.0001. Sensitivity 77%, specificity 74%, Positive predictive value 50%, Negative predictive value 90%). The Kaplan Meier curve has confirmed a significant cumulative risk for value < 144 (Log rank test p > 0.0001). Interestingly, in these patients with reduced ventricular function, EF was not predictive of new events, while PI was significantly associated with new events (acute heart failure and non arrhythmic death). PI is easy to calculate and could be applied in clinical practice to stratify this very high risk population.