Background: Several pre-implant risk-stratification algorithms based on multiple variables routinely available have been proposed to predict mortality after cardiac resynchronization therapy (CRT). Purpose: To compare different models in predicting mortality and assess their association with clinical response in patients with HF who received a CRT device in an unselected, real-world population.
Methods: We collected data of 1066 consecutive patients who underwent CRT implantation from the CRT-MORE registry from 2011 to 2014. Prediction score models derived for the present analysis were: EARN Score, VALID-CRT Risk Score and ScREEN Score. Aggregated risk score (AGR) comprised all common factors used by at least two of these scores (Age=75years; LVEF=25%; GFR=60 mL/min/1.73 mq; persistent/permanent AF at implantation; NYHA class III or IV; gender male). For comparison each model was categorized stratifying patients in five risk groups (S1-S5). The primary endpoint was total mortality after CRT implantation. Clinical Response (CR) at 12-month follow-up was evaluated considering a hierarchical composite criterion which includes alive status, hospitalization for HF, and variations in NYHA functional class.
Results: 745 patients had complete data for risk-scores comparison (mean age 70±10 years, 74% male, 53% ischemic, 60% NYHA III/IV, 20% with persistent/permanent atrial fibrillation at the time of implantation, mean LVEF 29±7%). During a median follow-up of 1012 [616-1371] days 111 patients died, 69 had at least one HF hospitalization and 167 met the combined endpoint of death or HF hospitalization. 69% of pts displayed an improvement in their CR at 12 months. All risk-stratification algorithms were able to predict total mortality after CRT (survival ranging from 7.9% to 10% -S1- and 18.1% to 20.5% -S5-; HR=2, 95%CI: 1.3 to 3, p=0.0013 for ScREEN; HR= 2.2, 95%CI: 1.5 to 3.5, p<0.0003 for VALID-CRT; HR=2.3, 95%CI: 1.5 to 3.4, p<0.0001 for EARN). AGR showed a good discriminatory power with an HR=2.8, 95%CI: 1.8 to 4.4, p<0.0001). All Scores predicted also hospitalizations for HF (HR ranged from 1.8 to 2) and CR (it decreases according to the severity of the risk profile ranging from 88% -S1- to 20% -S5-).
Conclusion: A simple score that includes a limited set of variables appears to be predictive for total mortality in an unselected, real-world population and seem to be useful also for predicting hospitalization for HF and CRT response. At the time of CRT implantation identifying patients in whom the benefit of device therapy is attenuated may improve their management facilitating consideration of alternative therapies and tailored follow-up.