Wolfram Grimm, MD

Department of Internal Medicine and Cardiology, Philipps University of Marburg, Germany.

Supported by The German Science Foundation (Grant # Gr 1130/2-1 and Gr 1130/2-2)

Background: Based on the results of recent controlled trials, implantable cardioverter-defibrillator (ICD) therapy has become the method of first choice to prevent sudden cardiac death in patients at high risk for malignant ventricular arrhythmias. Sudden cardiac death has been observed in about 50% of patients with idiopathic dilated cardiomyopathy (IDC). Arrhythmia risk stratification in patients with IDC with regard to prophylactic ICD implantation, however, is a completely unsolved issue. Therefore, we performed a prospective observational study to investigate prediction of arrhythmic events in IDC by the following 11 predefined variables: age, gender, NYHA class, LV end-diastolic dimension (LVEDD) and ejection fraction (EF) by echocardiography, atrial fibrillation and left bundle branch block on ECG, time domain and spectral turbulance analysis of the signal-averaged ECG, frequent VPDs and nonsustained VT (NSVT) on Holter.

Methods and Results: Between 1992 and 1997, 202 patients with IDC (age: 50±13 years, EF: 30±10%) without previous sustained VT or VF and without antiarrhythmic drugs were prospectively enrolled in this study. During 32±15 months follow-up, arrhythmic events including sustained VT, VF or sudden death occurred in 32 pts (16%). By multivariate Cox regression analysis, only LVEDD ³ 70 mm and NSVT on Holter were identified as significant arrhythmia risk predictors. The relative relative risk for the combination of LVEDD ³ 70 mm and NSVT was 14 (95% CI: 2.3 - 90). When EF £ 30% was forced to remain in the Cox regression model, only EF £ 30% and NSVT on Holter were found to be significant arrhythmia risk predictors. The relative risk of patients with the combination of EF £ 30% and NSVT was also 14 (95% CI: 2.2 - 97). Sensitivity, specifity and positive predictive accuracy for arrhythmic events for the combination of EF £ 30% and NSVT were 59%, 84% and 40%, respectively.

Conclusions: The combination of LVEDD ³ 70 mm and NSVT as well as the combination of EF £ 30% and NSVT identify a subgroup of patients with IDC with a 14-fold increased risk for subsequent arrhythmic events. The clinical utility of these parameters, however, is limited by a positive predictive accuracy of £ 40%. Therefore, the Marburg Cardiomyopathy Study (MACAS) has been initiated in 1996 to determine whether the positive predictive accuracy of "traditional" arrhythmia risk predictors including LVEF and NSVT on Holter can be improved by newer potential arrhythmia risk predictors including heart rate variability, baroreflex sensitivity and T wave alternans.