ABC | Volume 114, Nº3, March 2020

Original Article Silva et al. AF ablation with rivaroxaban Arq Bras Cardiol. 2020; 114(3):435-442 of apixaban or warfarin. The combined outcome of death, stroke or bleeding was similar (22/318 pts vs. 23/315 pts; p = 0.0002 for noninferiority). Brain magnetic resonance imaging, after the procedure, showed similar rates of “silent” cerebral ischemic lesions. 17 In the AEIOU trial, Reynolds MR et al. described similar bleeding rates and no stroke in 3 groups – uninterrupted edoxaban, interrupted (interruption of one dose) edoxaban and warfarin. 18 Ameta-analysis that included 7400 pts from15 observational and 1 randomized studies reported a trend toward lower rate of TE events in patients receiving rivaroxaban compared to warfarin (p = 0.052), with similar bleeding complications (1.15% vs. 1.66%; p = 0.23). 19 Sawhney V et al. compared DOAC (64% rivaroxaban) to warfarin, uninterrupted in 1884 AF ablation procedures, and found no difference between the groups in relation to the primary outcome consisting of death, TE or major bleedings (2.2% vs. 1.4% p = 0.2). 20 With these now more consistent results, catheter ablation of AF under uninterrupted use of warfarin, dabigatran or rivaroxaban is now class I recommendation in the latest expert consensus (HRS, EHRA, ECAS, APHRS, SOLAECE), published in 2017. 1 In our service, which has 14 years of experience in AF ablation, with current 50 to 100 procedures/year, after a long period using uninterrupted warfarin (therapeutic INR) for AF ablation, we chose rivaroxaban as an alternative based on the presented results, in a major adaptation to our routine, to the preoperative group protocol and drug pharmacokinetics. The dose taken on the previous night allowed the procedure to be performed on the following day, with the patient within the drug therapeutic window and, at the same time, outside its peak of action. Moreover, the next dose, to be taken on the day of the ablation, would be administered a few hours after the end of the procedure, an adequate period to observe complications. The low overall rates of adverse events reported in both groups was in agreement with the abovementioned literature results. The low rate of hemorrhagic events in the RIV group was noteworthy, even those related to venous access, performed by conventional puncture without the aid of ultrasound (US). This tool has been used to guide venipuncture in patients using anticoagulants. Data from a meta-analysis (4 observational studies) showed a 60% and 66% reduction in major and minor vascular complication rates, respectively, with the use of US. 21 However, randomized trials have not yet confirmed these data. Yamada et al. randomized 320 patients for punctures guided or not by US (Ultra-Fast Trial); they reported shorter time to puncture, less fluoroscopy use, fewer inadvertent arterial punctures and less local postoperative pain when using US, but without significant difference regarding major (vascular) complications. 22 In the present series, one should consider that the approach of our group regarding the accesses – only 3 femoral punctures, without jugular punctures or intracardiac echocardiography (larger sheaths) – may have contributed to low rates of vascular complications. Moreover, one cannot rule out that the US, if used to guide the punctures, would have prevented such complications. On the other hand, ischemic stroke occurred in one patient in this group, a fact that had not been observed with warfarin throughout the group's experience. We considered the event as occasional, as it statistically corresponds to the rates reported in the literature. The main fear of using rivaroxaban is the lack of a direct “antidote” in case of bleeding complications, especially cardiac tamponade, a potentially lethal event, if not treated quickly. This study did not allow us to assess this risk situation because no cardiac tamponade occurred. In studies with available DOAC so far, although some have reported greater drainage in cases of cardiac tamponade, there were no significant differences in the management of these complications or mortality compared to warfarin. In the J-CARAF (Japanese AF ablation registry), in contrast, there was a lower rate of pericardial effusions that required drainage with DOAC than with warfarin (p < 0.05). 23 In general, in situations of major bleedings with warfarin or DOAC, supportivemeasures (saline replacement and vasoactive drugs), reversal of heparin (protamine sulfate), eventual use of prothrombin complex or Factor VII, and immediate drainage by pericardiocentesis are recommended, and the service should be prepared for the immediate approach of such complications. Certainly, the availability of a direct reversing agent would bring a greater sense of safety to the procedure, but the potential risk of thromboembolic complications should be considered when reversing anticoagulation completely after extensive RF applications to the left atrial endocardium. In the RE-CIRCUIT trial, bleeding complications with dabigatran were treated without the use of the specific direct reversing agent, idarucizumab, despite its availability in the centers involved in the study. 16 It is generally agreed that regardless of the elected periprocedural anticoagulation strategy, intravenous heparin should be administered before or immediately after the first transseptal puncture at doses that maintain the ACT levels between 300 and 400 seconds. 1,12 Previous studies have shown that patients on continuous warfarin use reach the ACT target levels faster and with lower heparin doses compared to those who transitioned to unfractionated heparin for ablation. 8-10 In case of uninterrupted DOAC use for ablation, more recent studies report that higher doses of heparin are required. 9 Due to these data, we used a loading dose and additional doses (formula described above) of heparin in patients in the WFR group. Our findings showed that, as with enoxaparin, rivaroxaban patients received higher doses of heparin to achieve adequate levels of ACT compared to those using uninterrupted warfarin. Well-controlled heparin replacement in these patients, using the formula previously tested in the group, also prevented large extrapolations in ACT levels (over 400 seconds), which may have influenced the low incidence of hemorrhagic events. Study limitations Potential limitations include: (1) retrospective, nonrandomized study; (2) unlike the WFR group, the baseline INR in the RIV group was not necessarily collected on the day before the procedure, but randomly in the weeks or days preceding it; however, this consideration may have no impact due to the low influence of DOACs on INR; (3) the fact that there was no cardiac tamponade in the RIV group made it impossible for us to conclude on the severity of this hemorrhagic complication in this group of patients or to compare their approach to the control group; (4) regarding cerebral ischemic events, the study was limited to clinical data, and no routine imaging study was performed to investigate the so-called silent ischemic lesions, previously described 440