P2276Sudden cardiac death in non-ischemic systolic heart failure: do we only need LVEF to decide ICD implantation?

Background Implantable cardioverter defibrillator (ICD) is recommended in patients (pts) with non-ischaemic heart failure with left ventricular systolic dysfunction who receive optimal medical therapy (OMT) in order to prevent sudden cardiac death (SCD). However, the results of the DANISH study have recently shown the limits of these recommendations. It is therefore mandatory to reconsider the risk stratification of SCD in this population. Purpose The purpose of our study is to determine independent predictors of severe arrhythmic events (AE in pts with non-ischemic systolic heart failure. Methods Between January 1998 and December 2014, all consecutive outpatients with non-ischemic systolic heart failure, receiving OMT and without a history of significant arrhythmic events, were included. We performed to all the pts a clinical and biological evaluation, an echocardiography, a cardiopulmonary exercise test, a radionuclide angiography and a Holter-ECG. Follow-up was performed either by direct examination, by contact with the general practitioner or the cardiologist and by remote monitoring if available. The composite primary endpoint was the occurrence of SCD, recovered cardiac arrest, sustained ventricular tachycardia, or appropriate therapy by the ICD. Results We included 910 pts with a mean age of 53±12 years, 244 (27%) were women, LVEF was 36±10%. Most of the pts received renin-angiotensin blockers (97%) and betablockers (84%), 77% received diuretics and 41% spironolactone. During a median follow-up period of 6.33 [3.29–10.18] years, 160 (17.6%) pts presented the composite primary endpoint. The median time between the assessment and the occurrence of AE was 4.05 [1.68–7.85] years. The most powerful independent predictor of AE was non-sustained ventricular tachycardia (≥3 ectopic beats) (HR: 2.8 [1.66–4.72], p<0.0001). The other independent factors of AE were left atrial diameter (HR: 1.03 [1.01–1.06], p<0.0001); gender (HR: 0.71 [0.55–0.92], p=0.010); digoxin intake (HR: 1.63 [1.10–2.44], p=0.016); QRS duration (HR: 1.01 [1.00–1.01], p=0.022), sinus rhythm (HR: 0.70 [0.56–0.87], p=0.001). LVEF, as a quantitative parameter, was not an independent predictor of AE. However, LVEF dichotomized with a value of 35% was a modest predictor of AE (HR =1.38 [1.12–1.70], p=0.002). Neither the NYHA classification nor the parameters of the cardiopulmonary exercise test were independent factors of AE occurrence. Conclusion LVEF is not the most powerful predictor of severe arrhythmic events in outpatients with non-ischemic systolic heart failure receiving optimal medical therapy. New risk scores are required. We found that in addition to LVEF, gender, QRS duration, sinus rhythm, left atrial diameter and more particularly non-sustained ventricular tachycardia were independent predictors of AE. This score needs to be validated in an independent population. Acknowledgement/Funding None