Aim. To develop a multifactorial model for predicting the development of polymorphic ventricular tachycardia (VT) in patients with drug-induced long QT syndrome (LQTS) induced by class III antiarrhythmic drugs (AADs) by identifying electrocardiographic, laboratory and molecular genetic predictors.Methods. The study included 64 patients (37 (57.9%) women and 27 (42.1%) men, mean age 57.2±9.4 years) with ischemic heart disease and/or arterial hypertension. and cardiac arrhythmias, in which drug-induced prolongation of the QTc interval (Bazett) (over 450 ms in men and over 470 ms in women) was noted with the use of class III AADs (amiodarone or sotalol) in a cardiac hospital. Depending on the presence or absence of non-sustained polymorphic VT according to 24-hour ECG Holter monitoring, patients were further divided into two groups: 17 patients with episodes of non-sustained polymorphic VT and 47 patients without such episodes. All patients underwent clinical and laboratory, instrumental and molecular genetic studies, which included taking an anamnesis, recording ECG in 12 leads, biochemical blood test, determining the levels of neuronal NO-synthase (NOS1) and the adapter protein of neuronal NO-synthase (NOS1AP) in blood serum by enzyme immunoassay, determination of nitric oxide synthase gene polymorphisms by polymerase chain reaction. To assess the relationship of the studied parameters with the achievement of the end point, the method of logistic regression with a binary response and the logit function of the connection was used.Results. To assess the risk of developing non-sustained polymorphic VT in patients with drug-induced LQTS while taking class III AADs, a complex binary logistic regression model was developed, including the following indicators: patient gender (p=0.019), relative variance of the QT interval (p=0.002), duration of the Tpeak-Tend interval, (p=0.034), serum magnesium (p=0.004) and NOS1 (p=0.004) levels, as well as the AA genotype of the G84A polymorphism of the NOS1 gene (p=0.049). With the calculated value of the threshold probability p≥0.48, the developed model makes it possible to identify patients at high risk of developing polymorphic VT in patients with drug-induced LQTS with a sensitivity of 94.12%, a specificity of 89.36%, and an area under the ROC curve of 0.977 (0.95-1.0, p<0.001).Conclusion. The developed complex model will allow predicting the risk of proarrhythmic effects in patients with drug-induced LQTS, which will lead to a decrease in the number of cardiovascular events in this category of patients.