Differential multivariable risk prediction of appropriate shock versus competing mortality - A prospective cohort study to estimate benefits from ICD therapy

Background and objective We prospectively investigated combinations of risk stratifiers including multiple EP diagnostics in a cohort study of ICD patients. Methods For 672 enrolled patients, we collected history, LVEF, EP study and T-wave alternans testing, 24-h Holter, NT-proBNP, and the eGFR. All-cause mortality and first appropriate ICD shock were predefined endpoints. Results The 635 patients included in the final analyses were 63 ± 13 years old, 81% were male, LVEF averaged 40 ± 14%, 20% were inducible at EP study, 63% had a primary prophylactic ICD. During follow-up over 4.3 ± 1.5 years, 108 patients died (4.0% per year), and appropriate shock therapy occurred in n = 96 (3.9% per year). In multivariate regression, age (p < 0.001), LVEF (p < 0.001), NYHA functional class (p = 0.007), eGFR (p = 0.024), a history of atrial fibrillation (p = 0.011), and NT-pro-BNP (p = 0.002) were predictors of mortality. LVEF (p = 0.002), inducibility at EP study (p = 0.007), and secondary prophylaxis (p = 0.002) were identified as independent predictors of appropriate shocks. A high annualized risk of shocks of about 10% per year was prevalent in the upper quintile of the shock score. In contrast, a low annual risk of shocks (1.8% per year) was found in the lower two quintiles of the shock score. The lower two quintiles of the mortality score featured an annual mortality <0.6%. Conclusions In a prospective ICD patient cohort, a very good approximation of mortality versus arrhythmic risk was possible using a multivariable diagnostic strategy. EP stimulation is the best test to assess risk of arrhythmias resulting in ICD shocks.