Comparison of effectiveness and safety between uninterrupted direct oral anticoagulants with and without switching to dabigatran in atrial fibrillation ablation

Abstract Introduction Recent studies have demonstrated the feasibility of uninterrupted direct oral anticoagulants (DOACs) with a temporary switch to dabigatran (“dabigatran bridge”) for atrial fibrillation (AF) ablation. We compared the effectiveness and safety between uninterrupted DOACs with and without the “dabigatran bridge” in patients taking factor Xa inhibitors. Methods AF patients on factor Xa inhibitors (rivaroxaban/apixaban/edoxaban) undergoing catheter ablation were eligible (n = 348). Brain MRI was performed within 2 days after the procedure to detect silent cerebral events (SCEs). Rivaroxaban/apixaban/edoxaban were uninterruptedly used in 153 patients (Group 1); these DOACs were switched to dabigatran on the day of AF ablation in 195 patients (Group 2). After propensity score matching, the unfractionated heparin (UFH) amount and the activated clotting time (ACT) kinetics during the procedure, the SCE incidence, and the follow‐up complications (30 days, thromboembolism and major/minor bleeding) in the two groups were compared. Results Group 2 had higher initial ACT value and shorter time to optimal ACT (>300 seconds) than Group 1 (184 ± 36 s vs 145 ± 22 s, and 34 ± 29 s vs 43 ± 34 s, P < .05, respectively). Group 2 tended to require less amount of UFH to achieve optimal ACT than Group 1, but the total amount of UFH for the procedure was comparable. Group 2 had lower SCE incidence than Group 1 (16.2% vs 26.4%, P < .05). The prevalence of follow‐up complications was unchanged between the two groups. Conclusions Switching to dabigatran on the day of AF ablation decreases preclinical thromboembolic events with similar bleeding risk to uninterrupted factor Xa inhibitors.


| INTRODUC TI ON
Catheter ablation is a common intervention for atrial fibrillation (AF) but serious complications especially clinically apparent cerebral embolism and cardiac tamponade, remain a concern.1,2 Therefore, periprocedural oral anticoagulation (OAC) should be optimized.
Direct oral anticoagulants (DOACs) became the practical standard for periprocedural OAC in AF ablation. In recent randomized trials, the uninterrupted use of DOACs demonstrated the incidence of major bleeding almost similar to, or even lower than, the uninterrupted use of vitamin-K antagonist (VKA).3-6 An expert consensus statement has recommended uninterrupted DOACs for periprocedural OAC in AF ablation.7 Idarucizumab, an antidote of direct thrombin inhibitor (dabigatran), is now available worldwide.8 However, andexanet alfa, an antidote of factor Xa inhibitors (rivaroxaban, apixaban, and edoxaban), is available in only limited countries.9 Serious bleeding may lead to worse outcomes in the presence of an irreversible anticoagulant. Dabigatran seems to be beneficial in countries without andexanet alfa, especially when used without interruption during the procedure. Recent studies have proposed the uninterrupted use of DOACs with a temporary switch to dabigatran ("dabigatran bridge") in patients undergoing AF ablation and taking factor Xa inhibitors. 10,11 This OAC strategy seems to compensate for the drawbacks of the uninterrupted use of factor Xa inhibitors in the absence of andexanet alfa. However, this has not been fully evaluated in clinical practice.
In this study, we compared the effectiveness and safety between uninterrupted DOACs with and without "dabigatran bridge" in patients taking factor Xa inhibitors prior to AF ablation.

| Study population
This is a nonrandomized single center prospective study of AF  Before May 2017, rivaroxaban, apixaban, or edoxaban was uninterruptedly used during the procedure (Group 1). Idarucizumab became available in May 2017, but andexanet alfa has not yet been approved here in Japan. For safety reasons, dabigatran was preferred for use at least during AF ablation in the case of adverse bleeding. Thus, since May 2017, rivaroxaban, apixaban, and edoxaban were continued until one day before AF ablation (the day of hospital admission) and then the anticoagulants were switched to dabigatran on the day of the procedure (Group 2). If a patient's CrCl was less than 50 mL/min, the low dose (dabigatran 110 mg b.i.d.) was selected. Dabigatran was maintained after the procedure. Figure 1A shows the schematic diagram of anticoagulation protocol for each group.

| Oral anticoagulation therapy
A proton pump inhibitor was prescribed in all patients and was continued at least for the 30-day follow-up period after the procedure to minimize the risk of esophageal injury.

| Ablation procedure
In paroxysmal AF patients, pulmonary vein isolation (PVI) was performed using either cryoballoon ablation (CBA, single short freezing for 180 seconds in each PV) or radiofrequency catheter ablation (RFCA) in the first session. RFCA was selected in patients with common PV or large PV ostium (>28 mm) based on the LA/PV anatomy evaluated by cardiac-computed tomography imaging before the procedure. In persistent AF patients, only RFCA was used for PVI in the first session. In all patients, only the PVI strategy was used for the first AF ablation. In the second session for both types of AF, the incomplete line of PVI was repaired with RFCA if necessary, and additional posterior wall isolation (linear ablation at the roof and bottom between left and right PV) and superior vena cava isolation were performed.
The CBA procedure was achieved using electroanatomical mapping (EnSite NavX, Abbott) and fluoroscopic guidance to position the cryoballoon catheter. In the RFCA procedure, PVI was achieved using a focal "point-by-point" catheter approach, delivering radiof-

| Anticoagulation during procedure
Baseline activated clotting time (ACT) was measured before a first injection of unfractionated heparin (UFH). Then, a bolus of UFH (80-120 IU/kg) was administered. ACT was measured every 10-20 minutes after the first UFH shot, and additional UFH boluses (20-60 IU/kg) were administered before transseptal puncture to reach optimal ACT (>300 seconds) for AF ablation. 15 The time from the baseline ACT measurement until reaching optimal ACT and the amount of UFH to achieve optimal ACT was calculated. During the procedure, ACT was measured every 20-30 minutes; additional UFH boluses were administered to maintain the ACT > 300 s. The total amount of UFH required for the procedure was also calculated.

| Brain MRI
Brain MRI was performed in all patients within 2 days after the procedure using a 1.5 Tesla (T) scanner (Achieva 1.5T Nova Dual; Philips Healthcare) with an 8-channel brain coil, or a 3T scanner   17 Clinically relevant bleeding events that did not fulfill ISTH criteria for major bleeding events were defined as minor bleeding events.

| Statistical methods
Continuous variables, represented as mean ± standard deviation, were compared using unpaired t tests. Categorical data, expressed as frequencies and percentages, were compared using chi-square tests.
Because of the heterogeneity of the groups' sample sizes and patient characteristics, propensity score matching was performed by the nearest neighbor technique to reduce the effect of potential confounding factors. Propensity scores were estimated from a logistic model and matched using a caliper width equal to 0.2 of the standard deviation of the logit of propensity score. Variables that potentially affect the incidence of SCE, thromboembolisms, and bleeding events were selected; predictors for propensity score matching were age, sex, type of AF, CHADS 2 score, left-atrial diameter, and type of procedure (RFCA or CBA).
All tests were two sided, and a P < .05 was considered statistically significant. Statistical analyses were performed using JMP11 (SAS Institute).  Table 1 summarizes patients' characteristics before and after propensity score matching.

| UFH amount and ACT kinetics
Initial ACT before first heparin injection significantly increased in Group 2 (184 ± 36 s vs 145 ± 22 s, Figure 2A), and the time to reach optimal ACT (>300 seconds) decreased in Group 2 (34 ± 29 s vs 43 ± 34 s, Figure 2B). The amount of UFH required to achieve the optimal ACT tended to decrease in Group 2 ( Figure 2C), but the total amount of UFH between the two groups remained unchanged ( Figure 2D). Group 2 achieved optimal ACT more quickly than Group 1; this finding is associated with the higher initial ACT value.

| SCE
The incidence of SCE significantly decreased in Group 2 (16.2%), compared to Group 1 (26.4%, Figure 3). Among Group 1, the incidence of SCE was 23% in rivaroxaban, 39% in apixaban, and 19% in edoxaban. Apixaban seems to have higher incidence of SCE than rivaroxaban and edoxaban, but the difference did not reach statistical significance (P = .09). We also examined the incidence of SCE in patients taking dabigatran for periprocedural period; dabigatran was uninterruptedly used during the procedure (uninterrupted dabigatran group). The incidence of SCE was compared between Group 2 and the uninterrupted dabigatran group. After propensity score matching, 61 patients in each group were selected for analysis. Patient characteristics after propensity score matching were shown in Table S1.
Uninterrupted dabigatran group had lower incidence of SCE than Group 2 (18% vs 23%) but no statistical difference was observed (P = .50); "dabigatran bridge" and uninterrupted dabigatran therefore have a similar impact on SCE.

| Follow-up complications
All patients were followed up at least for 30 days and none of them dropped out. Two major bleeding events were observed in Group 1 but none in Group 2. One major event in Group 1 was intracranial hemorrhage (epidural hematoma) in a patient with uninterrupted rivaroxaban (15 mg), causing death. This happened 1 week after discharge although brain MRI (within 2 days after the procedure)

| D ISCUSS I ON
This study compared uninterrupted DOACs with and without switching to dabigatran ("dabigatran bridge") in patients taking factor Xa inhibitors prior to AF ablation. The results demonstrated that uninterrupted DOACs with "dabigatran bridge" showed higher initial ACT value, shorter time to achieve ACT > 300 s, and lower incidence of MRI-detected SCE than uninterrupted DOACs without "dabigatran bridge" (uninterrupted factor Xa inhibitors). However, follow-up complications between the two groups were comparable.

| Comparison with previous studies
Recent studies have evaluated the effectiveness and safety of "dabigatran bridge" for AF ablation by comparing with minimally-interrupted DOACs strategy.10,11 On the other hand, DOACs were uninterruptedly used during the procedure in both groups of our study. Because a recent guideline/expert consensus statement recommends uninterrupted use of warfarin and DOACs for the periprocedural period,7 the comparison between different uninterrupted DOACs strategies might be an advantage of this study. After hospital discharge, dabigatran was switched back to the original DOAC prescribed before the procedure in previous studies. Contrarily, the switched dabigatran was continued at least for the follow-up period in this study. A previous study assessed adverse events for 8 TA B L E 1 Patient characteristics before/after propensity score matching Before propensity score matching After propensity score matching Antiplatelets, n (%) 13 (8) 14 (7) .649 12 (9) 9 (7) .495 AAD, n (%) 24 (16) 27 (14) .631 23 (17) 17 (13) .304 Low dose NOAC, n (%) 23 (15) 38 (19) .276 22 (16) 33 (24) .096 Ablation procedure impact on activated partial thromboplastin time (aPTT), mainly related to the activity of intrinsic pathway, than factor Xa inhibitors.18 ACT is also related to the intrinsic pathway activity and the use of dabigatran on the day of the procedure would increase initial ACT value, shortening the time to optimal ACT value, which may prevent microthrombus formation during the procedure. In this study, different intraprocedure ACT kinetics between the groups may, at least partly, contribute to the difference in the incidence of SCE, however the mechanisms why "dabigatran bridge" just on the day of the proce- Although no thromboembolic events were observed in either group, the lower incidence of SCE in Group 2 suggests the potential benefit to minimize procedure-related thromboembolic risk.
Recent studies demonstrated that uninterrupted dabigatran, but not factor Xa inhibitors, decreased the incidence of major bleeding during the periprocedural period of AF ablation when compared to uninterrupted VKA. Although there were more bleeding events in Group 1 than in Group 2, there was no significant difference in the prevalence of follow-up complications in this study; the clinical efficacy and safety were statistically comparable between the two groups. This is likely attributable to the small number of patients.
Nevertheless, the use of dabigatran on the day of AF ablation still seems to have an advantage in case of procedure-related adverse bleeding in the countries where idarucizumab, but not andexanet alfa, is clinically available.

| Limitation
This is a nonrandomized single-center study of a small number of patients. Treatment adherence of DOACs was not carefully evaluated, and poor adherence in some patients could have caused inadequate anticoagulation. The regular dose of rivaroxaban in Japan is different from that in Europe/North America. However, patient characteristics between the two groups in this study were well-matched by propensity score, and represent average age, BMI, and CHADS2 score typical of Japanese AF patients undergoing catheter ablation; the results may be more applicable to Japanese patients. Acute SCE lesions reportedly regress during the follow-up period, but we did not perform brain MRIs during the chronic phase. We performed diffusion-weighted MRI (DWI) imaging with ADC map but did not perform T2-weighted axial fluid-attenuated inversion recovery (FLAIR) imaging, which is a potential drawback of this study. A lack of MRI imaging prior to the procedure is also one of the limitations.

| CON CLUS IONS
Uninterrupted DOACs with "dabigatran bridge" minimizes the preclinical thromboembolic events with similar bleeding risk to uninterrupted factor Xa inhibitors. However, it still has an advantage in case of procedure-related adverse bleeding in the countries without Andexanet alfa.