Value of fragmented atrial electrograms for identification of optimal sites for radio-frequency delivery during cardioneuroablation for functional bradyarrhythmias - the ROMAN 5 study

Introduction Cardioneuroablation (CNA) is a promising method for the treatment of functional bradyarrhythmias resulting in syncope or presyncope, however, the optimal method for identification of sites for effective radio-frequency (RF) current delivery is not known. Some studies suggested that such sites are characterised by increased fragmentation, duration or amplitude of local atrial electrogram (LAE). Purpose To assess the predictive value of LAE for identification of successful sites of vagal denervation during CNA. Methods In this prospective study 40 patients (62.5 % females, mean age 40±2 years) with cardioinhibitory reflex syncope or symptomatic bradycardia resistant to all non-pharmacological therapies underwent anatomically-guided (electro-anatomical system and intra-cardiac echocardiography) CNA. Paraseptal superior and inferior ganglionated plexi (GP) areas from the left (LA) and right atrium (RA) were targeted, and, if ineffective, followed by GPs located close to the left pulmonary veins. Intra-procedural efficacy was assessed using extra-cardiac vagal stimulation (ECVS) which was performed after each consecutive RF delivery. LAE were recorded at an ECG speed of 400 mm/s and filter setting 100-500 Hz. Three parameters were measured – number of deflections, maximal amplitude and total duration. The operator was blinded to LAE while performing CNA (signals from ablation electrode were clipped) in order not to influence the choice of the site for RF delivery. Results CNA was effectively performed in all patients – ECVS was negative in all at the end of procedure. LAE from three paraseptal GP areas were analysed (93 for SPSGP-LA, 122 for IPSGP-LA and 80 for IPSGP-RA). Other sites were not analysed due to low number of RF applications because CNA of paraseptal GPs was enough to achieve full vagal denervation in the majority of patients. Compared with ineffective sites (n=241), effective sites (n=126) were characterised by significantly higher number of LAE deflections (7.00±1.5 vs 6.00±1.33, p=0.002), longer duration (47±6.33 vs 46.7±6.67 ms, p=0.045) and trend towards higher amplitude (0.55±0.25 vs 0.49±0.26 mV, p=0.085). The values of sensitivity, specificity as well as positive, negative and total predictive accuracy for various cut-off values of the number of LAE deflections are shown in Table1. Conclusions LAE at effective sites for CNA are characterised by higher number of deflections and greater amplitude than unsuccessful sites, however, their predictive value is rather low and should not be used as the only parameter to guide CNA. Combination of various anatomical approaches and LAE analysis is probably a better approach until more accurate methods for identification of GP areas become available.