ABC | Volume 114, Nº3, March 2020

Original Article Silva et al. AF ablation with rivaroxaban Arq Bras Cardiol. 2020; 114(3):435-442 performed whenever there was a typical atrial flutter electrocardiographic record or if it occurred (spontaneously or not) during the procedure. The standard protocol consisted of three right femoral punctures, not guided by ultrasound; deflectable decapolar catheter placed in the coronary sinus using a 7F introducer sheath and two transseptal punctures, performed only with the aid of fluoroscopy. Decapolar or duodecapolar circular catheters were used in a conventional SL1 sheath (SwartzTM; St. Jude Medical/Abbott) for mapping of LA/PVs and irrigated catheter for ablation (without or with contact sensor) in an SL1 or deflectable sheath (AgilisTM; St Jude Medical/Abbott). RF applications were limited to the power of 20 to 25 W in the posterior wall and 30 to 35 W in the other walls and monitored by the impedance curve, esophageal temperature and contact force (when available). The criteria for interrupting an RF application were: sudden increase in impedance, esophageal temperature reaching 37.5ºC and contact force greater than 40 g. RF applications were performed continuously to fill the entire circumference of the PV antra (Figure 2). We considered as full isolation of the PVs the complete disappearance of the electrograms in the circular catheter placed at its most proximal portion (inlet block) and also the demonstration of electrical dissociation between the PVs and the LA through programmed stimulation of the same circular catheter (outlet block). The adenosine test (12 mg) was performed 20 minutes after the completion of PV isolation and additional applications were performed if PV-LA reconnection was observed. Anticoagulation in the procedure Prior to transseptal punctures, the sheaths and transseptal needle were washed with saline solution containing 50 IU/mL of heparin, and basal ACT was measured. The first dose of heparin (loading dose) was administered immediately after the first transseptal puncture (directly in the sheath), with 100 IU/kg in the RIV group and 50 IU/kg in the WFR group (maximum dose of 10,000 IU); the reduced dose in the control group was based on prior group experience and literature data. 8-10 After that, the ACT was systematically measured every 30 minutes, aiming to maintaining it between 300 and 400 seconds. Additional doses of intravenous heparin were given whenever the ACT was below 300 seconds, calculated according to the formula created and tested by the group. 11 RIV group: → Hep Dose (IU) = Weight (Kg) x CI * 2 VRF group: → Hep Dose (IU) = Weight (Kg) x CI * 3 *CI= Correction Index ACT (sec) CI* 150 – 200 75 201 – 250 50 251 – 300 25 > 301 0 The removal of the sheaths was performed still in the operating room after protamine sulfate infusion (5,000 IU). Statistical analysis Data for all variables were evaluated for normality through Histogram and D`Agostino & Pearson’s Test. Continuous variables were described as mean and standard deviation and compared using unpaired Student's t test, except for the variable “baseline INR” (data evaluated as “non-normal”), which was compared using the Mann-Whitney test. Categorical variables were described as absolute numbers and percentages in relation to the sample and compared using Fischer's exact test. The level of statistical significance was set at 5%. GraphPad Prism 7.0e software was used for statistical analysis. Results The clinical characteristics of the groups were similar, including the CHA 2 DS 2 -VASC score, presence of structural heart disease and predominance of paroxysmal AF. At the end of the procedure, 100% isolation of the PVs in both groups was demonstrated. The percentage of patients who received linear ablation of the LA and the cavo-tricuspid isthmus was similar, but the ablation of fragmented CFAEs was more frequent in the WFR group, probably due to the progressive abandonment of this technique in recent years. There was no statistically significant difference regarding the total procedure time (Table 1). According to the described protocol, no patient had intracavitary thrombus at the TEE on the day before the procedure. It is noteworthy that no patient was excluded from this study due to LA thrombus. There were no deaths. Primary outcomes: One patient had procedure-related ischemic stroke in the RIV group, evolving with mild dysarthria in the immediate postoperative period, with spontaneous resolution within 48 hours without further sequelae (Figure 3). This patient had paroxysmal AF without structural heart disease or risk factors for TE events (CHA 2 DS 2 -VASC = 0). No thromboembolic events occurred in the WFR group. Major bleeding occurred in 2 patients in the WFR group: 1 hemopericardium with cardiac tamponade and 1 large hematoma at the femoral puncture site. The first case was controlled by pericardiocentesis, volume replacement and administration of protamine sulfate. The second case required blood transfusion and longer hospital stay. Both were discharged without further complications. Major bleeding – retroperitoneal hematoma – occurred in the RIV group and required surgical intervention (drainage) due to uncontrollable pain, and the patient was discharged without sequelae. Secondary outcomes: Only one puncture site hematoma, clinically not relevant, was observed in theWFR group; none was observed in the RIV group. As expected, the baseline INR was higher in theWFR group (2.5±0.03 vs . 1.2±0.02; p<0.0001), but there was no difference in baseline ACT between the WFR and RIV groups (123.7 ± 3 vs. 118 ± 4.2; p = 0.34). The mean ACT level during the procedure was adequate in both groups, within the recommended range and similar in 437