Aim. Identification of clinical and instrumental predictors for non-arrhythmic death in patients with heart failure (HF) and implantable cardioverter-defibrillator (ICD).Material and methods. Through a telephone survey and examination of medical records from hospital and polyclinic databases, data were obtained on the alive/dead status and causes of death for 260 patients with heart failure (HF) and ICD included in the Kuzbass Registry of Patients with ICD. The follow-up period was 1.5 years. Clinical and instrumental parameters entered into the registry before the ICD implantation were included in a univariate and multivariate step-by-step analysis using the logistic (for qualitative variables) and linear (for quantitative variables) regression with calculation of regression coefficients and construction of a prognostic regression model. The quality of the created model was assessed using a ROC analysis.Results. During the observation period, 54 (20.8%) patients died. In 21 (38.8%) patients, death occurred in the hospital and was caused by acute decompensated heart failure in 15 (71.4%) patients, myocardial infarction in 3 (14.3%) patients, stroke in 1 (4.7%) patient, and pneumonia in 2 (9.5%) patients. 33 (61.2%) patients died outside the hospital; the cause of death was stated as the underlying disease associated with acute decompensated heart failure: in 9 (27.2%) patients, dilated cardiomyopathy; in 1 (3.0%) patient, rheumatic mitral disease; and in 23 (69.7%) patients, ischemic cardiomyopathy. According to the univariate regression model, the risk of death in the long-term period was increased by the QT interval prolongation (U 2.41, p = 0.0161); elevated pulmonary artery systolic pressure (U 4.30, p=0.0000) and increased left atrial size according to echocardiography (U 2.98, p=0.0029); stage IIB HF (OR 2.41; 95% CI: 1.26-4.6), NYHA III-IV (OR 3.03; 95% CI: 1.58-5.81); chronic obstructive pulmonary disease (OR 5.24; 95% CI: 2.04-13.45); and lack of optimal drug therapy (ODT) for HF before ICD implantation (OR 2.41; 95% CI: 1.29-4.49). The multivariate analysis identified the most significant factors included in the prognostic regression model: pulmonary artery systolic pressure above 45 mm Hg, social status, chronic obstructive pulmonary disease, and lack of ODT for HF.Conclusion. To ensure a maximum benefit from ICD, the factors that increase the likelihood of non-arrhythmic death should be considered before making a decision on ICD implantation. Particular attention should be paid to mandatory ODT for HF as the main modifiable risk factor for unfavorable prognosis.