Comparison of Multivariate Risk Estimation Models to Predict Prognosis in Patients With Implantable Cardioverter Defibrillators With or Without Cardiac Resynchronization Therapy

Abstract

Several multivariate risk score models were developed to predict prognosis of patients with heart failure (HF). We compared 3 models with regard to prediction of mortality in patients with HF who received an implantable defibrillator (ICD) or a cardiac resynchronization therapy defibrillator (CRT-D), as primary prevention of sudden death. The study cohort consisted of 823 patients (ICD= 410; CRT-D= 413). The evaluated models were the Seattle Heart Failure Model (SHFM), the Multicenter Automatic Defibrillator Implantation Trial II (MADIT II) score, and an adjusted Charlson Comorbidity Index (aCCI). End point was the performance of the models to predict all-cause mortality at 5years. This was determined by c-statistics, for both subgroups. Multivariate analysis was used to analyze the relations between the risk score models, their individual components and mortality, and its applicability to the entire population. Cumulative mortality was 4.9% at 1year and 21.1% at 5years. Discriminatory power for 5-year mortality was highest for the SHFM (0.73; p<0.001) compared with the MADIT II score and the aCCI for the entire population. SHFM performed better than the MADIT II score for CRT-D group. In the entire population, the SHFM and the aCCI were significant predictors of mortality in multivariate analysis (hazard ratio 1.90, 95% confidence interval 1.49 to 2.43 vs hazard ratio 1.11, 95% confidence interval 1.01 to 1.22). The strongest individual components were age, HF, impaired renal function, and cancer, whereas CRT-D use was no predictor. In conclusion, the SHFM has the best discriminatory power for 5-year mortality in patients with HF with an ICD or CRT-D. The aCCI and MADIT II scores are less powerful but viable alternatives.