P1736Two-step risk stratification of life-threatening ventricular tachyarrhythmias in patients with dilated cardiomyopathy

Abstract

The absence of clinical risk algorithms to guide decisions and selection of potential ICD candidates lead to discussion of the potential benefit from primary preventive device implantation. Purpose: Design of prognostic algorithm to define personalized risk of life-threatening ventricular tachyarrhythmias in patients with DCM was the aim of this study. Material and methods: The study enrolled and followed up 53,1 612,6 months 268 pts with DCM (74,6% male, aged 47,2 ± 11,7 years, LVEF 28,7 ± 10,1%; NYHA class 2,760,4). We analyzed age, gender, NYHA, 6-MWT, peak VO2, EchoCG, Holter ECG (PVEs, sVT, nsVT, VF), device interrogation and high definition ECG (Intecard-7) with assessment of HRT, mTWA, JTd, QTd, QRS width, and AC/DC. For stepwise discriminate, multivariate regression and ROC analysis, and binary logit-regression model as a primary endpoints were accepted SCD documented, successful resuscitation, sustained VT, and CRT-D\ICD appropriated discharges. Results: A stepwise discriminate analysis of ECG-markers showed a few independent predictor. Discriminators Y1 and Y2 are calculated as follows: Y1=1,92xmQTd+6,19xmHRTO+6,35xmmTWA+2,42xmDC-5,63 Y2=1,86xmQTd+0,92xmHRTO+0,71xmmTWA+1,39xmDC-1,09 The calculated value Y1>Y2 suggests that the risk of life-threatening VTffs is high, whereas Y2 > Y1 it is low. Cox regression analysis including ECG, HM-ECG, and EchoCG showed independent prognostic predictors: non-sustained VT (≥ 5 complexes ≥150 bpm, HR=3.24; h=0.007); mTWA (25 % of abnormal mTWA ≥ 46 mcV, HR=1.99; h=0.011); HRT (TO≥ 0,9 %, HR=1.93; h=0.017); LV GLS (GLSmean ≥- 6,5%, HR=1.98; h=0.02); ventricular ectopy (PVEs ≥1500/24h, HR=1.43; h=0.039); JT disp (JTd≥ 70 ms, HR=1.36; h=0.042). Predictors (and their cut-off values defined by a ROC analysis) were included in the binary logit regression. As a result binary analysis predictive model was constructed (F=31,2; 2=143,2; p=0,0001) in the final. The classification matrix demonstrated a high prognostic significance (F=31.2; χ2=143,2; p<0.0001; sensitivity 80.8 %; specificity 99.1 %). Two-step risk stratification and P-value risk grading are presented in Figure. Conclusion: The proposed two-step algorithm of the individualised risk stratification allows to classify and potentially assist decisions to defer or implant an ICD during the dynamic monitoring. (Figure Presented).