Cardiac resynchronization therapy: towards personalized device selection. Results of a two-year prospective study

Aim . To conduct a comparative analysis of clinical, instrumental, and laboratory diagnostic methods and to identify factors determining the likelihood of sustained paroxysmal ventricular tachyarrhythmias (VT) in patients with indications for cardiac resynchronization therapy (CRT). Methods . The study included 124 patients with chronic heart failure (CHF) and an implanted CRT-D system. The median age was 58 (52-63) years. Patients were followed for 24 months. Clinical and demographic characteristics, electrocardiographic data, speckle-tracking echocardiographic parameters, and blood biomarker levels were assessed. The primary endpoint was the occurrence of sustained VT episodes recorded by the implanted device. A multivariate logistic regression model was developed to predict the two-year probability of VT occurrence. Results . During the follow-up period, 29 patients (23.3%) experienced episodes of sustained VT. Univariate analysis identified seven candidate predictors with the highest potential for reaching the endpoint. These included: clinical factors (presence of coronary artery disease and atrial fibrillation); ECG parameters (modified QRS index >0.6, presence of left bundle branch block (LBBB) according to Strauss criteria); echocardiographic findings (global longitudinal strain ≥ -6%, mitral regurgitation of grade 2 or higher); and laboratory markers (galectin-3 ≥ 12 ng/mL). Based on these variables, a predictive model was developed using binary logistic regression to estimate the two-year risk of VT in patients with CRT indications. The Strauss LBBB criterion, although statistically significant in univariate analysis, was not included in the final model. At a regression function cut-off value of 0.228, the model demonstrated a diagnostic accuracy of 73.6% (sensitivity - 86.2%, specificity - 69.6%). The area under the ROC curve was 0.779, which, according to expert grading, indicates good model performance. Conclusion . The study identified several independent predictors of sudden cardiac death risk in patients with implanted CRT-D devices and enabled the construction of a multifactorial prognostic model. The findings suggest the potential for developing a personalized algorithm for device selection.