Ventricular tachyarrhythmias prediction in patients with coronary artery disease and left ventricular dysfunction

Type of funding sources: None. Introduction. The cardioverter-defibrillator (ICD) implantation is the most effective method for the sudden cardiac death (SCD) prevention. However, about 25% patients didn"t have an incidence of ICD shocks during first battery life. Most of them are mainly represented by patients who had the ICD implanted for the primary prevention of the SCD. It us known, reduced left ventricular ejection fraction (LVEF) is an independent predictor of the SCD. So, it’s necessary to find out new predictors of the SCD and ventricular arrhythmias (VTA) incidence, which will help to optimize the selection of patients who really need a ICD implantation. Purpose. To identify predicting methods of the VTA in patients with coronary artery disease and LVEF 35% or less. Materials and methods. The study included 40 patients (males – 36, mean age – 63,4 ± 7,8 years) with coronary artery disease (CAD), LVEF 35% or less and ICD implantation indications (primary prevention of the SCD). Patients before ICD implantation underwent 6-minute walk test, echocardiography, heart rate variability analysis and cardiac single-photon emission computed tomography with 123I-meta-iodobenzylguanidine (123I-MIBG). All patients after ICD implantation were followed-up during two years and divided into two groups. Results. The 1-st group consisted of 36 (90,0%) patients with VTA events. The 2-nd group consisted of 4 (10,0%) patients without VTA events. The univariate ROC-analysis showed that the low values of the average NN intervals (AUC = 0,986, p = 0,0001, t ≤ 1211 ms), standard deviation of NN intervals (AUC = 0,986, p = 0,0001, t ≤ 119 ms), standard deviation of the average NN intervals (AUC = 0,861, p = 0,0001, t ≤ 94 ms), average standard deviation of NN intervals (AUC = 0,792, p = 0,004, t ≤ 48 ms), root mean square of successive differences (AUC = 0,847, p = 0,0003, t ≤ 18 ms), very low frequency domain (AUC = 0,792, p = 0,02, t ≤ 2411 ms), low frequency domain (LFD) (AUC = 0,903, p = 0,0001, t ≤ 1046 ms), high frequency domain (AUC = 0,875, p = 0,0001, t ≤ 743 ms), total frequency domains (AUC = 0,847, p = 0,0003, t ≤ 2785 ms), heart/mediastinum ratio on early (AUC = 0,889, p = 0,0001, t ≤ 2,29) and delayed (AUC = 0,806, p = 0,001, t ≤ 1,65) scintigrams, as well as high values of the end-diastolic index (AUC = 0,944, p = 0,0001, t > 65,9 ml/m2), end-systolic index (AUC = 1,000, p = 0,0001, t > 23,6 мл/м2), 123I-MIBG accumulation defect on early (AUC = 0,958, p = 0,0001, t > 15,0%) and delayed (AUC = 0,958, p = 0,0001, t > 18,0%) scintigrams leaded to frequent occurrences of the VTA. The multivariate ROC-analysis demonstrated that the LFD (p = 0,0136) is independent predictor of the VTA. Also, increase of the VTA predictive ratio, calculated according to this predictive model, more than 0,8936 leaded to frequent occurrences of the VTA (AUC = 0,903, p = 0,0001). Conclusion. A decrease in the LFD can be a predictor of the VTA in patients with CAD. An increase in the predictive ratio more than 0,8936 is a prognostic marker of the life-threatening ventricular arrhythmias.