A prognostic index for risk stratification for acute heart failure and death in subjects with ischemic cardiomyopathy and cardiac defibrillator

To propose a clinical prognostic index for death and heart failure in patients with ischemic cardiomyopathy implanted with an ICD. This prospective study included 192 consecutive patients (age 68 ± 10) recruited from 2004 to 2009 and implanted with an ICD for MADIT II criteria. All patients performed 24-h ambulatory blood pressure monitoring after discharge and common haematological samples. The prognostic index (PI) was built according to the formula: 120 - age + mean 24 h systolic blood pressure - (creatinine * 10). Other variables were assessed: EF, haemoglobin concentration, mean 24 h heart rate and diastolic blood pressure, sodium level, pacing mode and diabetes. Non-arrhythmic cardiac death and new hospitalizations for heart failure during 1-year follow-up were the combined end point. A total of 48 events (25 %) occurred during the follow-up: 7 cardiac deaths and 41 hospitalizations for acute heart failure. Cox proportionalhazards model showed that PI was the only predictor of events (HR = 0.96; CI 95 % 0.944–0.976, p<0.0001). ROC curve showed that PI best cut-off was 144, with AUC 0.79, p<0.0001; sensitivity 77 %, specificity 74 %, positive predictive value 50 %, negative predictive value 90 %. PI was predictive of events in a clinical setting where EF had no predictive value. PI works according to the rule ‘‘the lower the worse’’. The high negative predictive value (90 %) of PI allows to identify subjects at lower risk for death and heart failure. PI can be a practical tool to stratify risk in ischemic cardiomyopathy.