Remote monitoring can predict lethal arrhythmic events through time-domain analysis of heart rate variability in patients with implantable cardioverter defibrillator

Background We prospectively collected device and heart rate data, i.e. heart rate variability (HRV), through remote monitoring (RM) of patients with implantable cardioverter defibrillator (ICD). Several studies have demonstrated usefulness of RM in implanted device patients, however, reports concerning the predictors of lethal ventricular arrhythmias are limited. Purpose The objective was to identify the predictors of lethal arrhythmic events (VT/VF). Methods Thirty-three patients (mean age: 50 years) with ICDs [with functionality of HRV analysis] were divided into 2 groups [VT/VF(+), VT/VF(−)]. Clinical, device (ventricular lead impedance; amplitude of ventricular electrogram), and HRV data were compared between the 2 groups. As the index of time-domain HRV analysis, NN intervals-index (SDNNi) was calculated for every 5 minutes, and the mean, maximum, and minimum SDNNi during the 24-hour period were used. Results During the observation period (median 12 months), 10 patients experienced VT/VF events. In HRV data, the mean, max, and min SDNNi were higher in VT/VF(+) than VT/VF(−) group (132.9±9.3 v.s. 93.5±6.1, p=0.0013; 214.6±10.6 v.s. 167.0±7.0, p=0.0007; 71.2±7.5 v.s. 43.9±4.9, p=0.0047). The other parameters did not exhibit significant difference. On logistic regression analysis, the mean SDNNi of 100.1, max SDNNi of 185.0 and min SDNNi of 52.0 as cut-off values for prediction of VT/VF event demonstrated significant receiver operating characteristics (ROC) curves (AUC=0.86, p=0.0007; AUC=0.84, p=0.0005; AUC=0.78, p=0.0030). Furthermore, in cases of VT/VF(+) group, the max ΔSDNNi, i.e., difference from baseline SDNNi, and min ΔSDNNi in 7 and 28 days preceding VT/VF events exhibited time course changes in comparison with baseline values. They were significant predictors of VT/VF events (max ΔSDNNi cut-off: 46.8, AUC=0.91, p=0.0002; min ΔSDNNi cut-off: −42.4, AUC=0.88, p=0.0014). Conclusion Time-domain analysis of HRV through RM may help identify patients at high risk of lethal arrhythmic events, and predict occurrence of such arrhythmic events. Funding Acknowledgement Type of funding source: None