How minimally interrupted direct oral anticoagulants affect intraprocedural anticoagulation during atrial fibrillation ablation? Insights from a Japanese single‐center retrospective study

Abstract Background Data are still lacking regarding the effects of minimally interrupted direct oral anticoagulants (MID) on the intensity of intraprocedural anticoagulation of atrial fibrillation (AF) ablation. Methods A total of consecutive 269 patients who undergone AF ablation were eligible for the study. All oral anticoagulants (OACs) were discontinued just one dose before the procedure except warfarin. We assessed the total required dose of UFH and time‐to‐target ACT > 300 seconds (TTA) for each of direct oral anticoagulant (DOAC) groups compared with the uninterrupted warfarin group. Results DOACs were used in 86% of the patients in the present study (dabigatran group (DG)‐17%, rivaroxaban group (RG)‐30%, apixaban group (AG)‐29%, and edoxaban group (EG)‐10%). DG and EG used comparable dose of total UFH to WG (WG vs DG; 206 ± 53 U/kg vs 231 ± 63 U/kg; P = .664, vs EG; 239 ± 67 U/kg; P = .335), while RG and AG required higher total UFH (WG vs RG; 206 ± 53 U/kg vs 270 ± 63 U/kg; P < .001, vs AG; 263 ± 62 U/kg; P < .001). TTA was significantly longer in RG (RG:73 ± 28 minutes vs WG:51 ± 25 minutes; P = .001), AG (AG:64 ± 26 minutes vs WG:51 ± 25 minutes; P = .02), and EG (EG:67 ± 34 minutes vs WG:51 ± 25 minutes; P = .02) than WG, whereas DG was comparable to WG (DG:51 ± 29 minutes vs WG:51 ± 25 minutes; P = NS). Especially, only RG demonstrated significantly slower increase in ACT than WG (P = .013). In the multivariate analysis, warfarin or dabigatran use, age > 75 years, and body weight < 60 kg are clinical predictors for achieving TTA within 60 minutes (TTA‐60). Conclusion MID‐dabigatran was comparable to uninterrupted warfarin, whereas MID‐factor Xa inhibitors were not. MID is a feasible protocol; however, we should be aware of its effect on the intraprocedural anticoagulation and differences among DOACs in the responsiveness to heparin.


| INTRODUC TI ON
Atrial fibrillation (AF) is a critical risk factor for cerebral thromboembolism and mortality. 1 While the Atrial Fibrillation Follow-Up Investigation of Rhythm Management (AFFIRM) study 2 failed to demonstrate the superiority of rhythm control over rate control in terms of mortality, several studies after the AFFIRM trial demonstrated that rhythm control with AF ablation led to improvements in quality of life, 3 exercise tolerance, 4 and mortality. 5 Specifically, catheter ablation for AF is more effective in maintaining sinus rhythm than antiarrhythmic drug therapy alone. Current consensus and guidelines 6,7 recommend the consideration of AF ablation as a firstline therapy in patients with symptomatic AF.
One of the important issues in AF ablation is periprocedural stroke. For example, in 2006, periprocedural stroke was noted in 1.1% of patients. 8 However, the risk of stroke reduced to 0.098%. 9 The improvement in thromboembolic complications has been owing to several factors, including continuation with vitamin K antagonists (VKAs) as a periprocedural anticoagulation protocol, 10  Calkins et al demonstrated the superiority of uninterrupted dabigatran to warfarin regarding major bleeding events in the RE-CIRCUIT trial. 13 Continuous apixaban and edoxaban therapies were also comparable to uninterrupted warfarin in the AXAFA-AFNET 5 and ELIMINATE-AF trials. 14,15 Although these studies proved the noninferiority or the superiority of uninterrupted direct OACs compared with uninterrupted VKA therapy in terms of feasibility and safety, concerns remain regarding serious bleeding complications because antidotes for factor Xa inhibitors are not still readily available in many countries including Japan.
Minimally interrupted direct OAC (MID) is a proposed alternative periprocedural protocol for AF ablation that is expected to reduce the risk of major bleeding complications. 16 The MID protocol is based on skipping one or two doses of OACs before ablation.
Although MID is assumed to be a safe oral anticoagulation protocol for reducing major bleeding complications and to be equivalent to uninterrupted VKA therapy for its capability of effective intraprocedural anticoagulation, real-world data on the feasibility of MID in daily clinical practice of AF ablation remain limited. Thus, the aim of the present study was to elucidate the effects of MID on intraprocedural anticoagulation during AF ablation and to demonstrate the feasibility and safety of MID in periprocedural OAC management for AF ablation. In the current study, serial changes in ACT and timeto target ACT (TTA) were used as surrogate markers to assess the suitability of intraprocedural anticoagulation using different direct OACs, including dabigatran, rivaroxaban, apixaban, and edoxaban, in comparison with uninterrupted VKA therapy.

| Study design and population
This was a retrospective, single-center, observational study. The study protocol adhered to the Declaration of Helsinki and was ap-

| Periprocedural oral anticoagulation management
All OACs, whether VKAs or direct OACs, were discontinued except for a single morning dose on the day before the ablation; therefore, there was a minimum of 24 hours between the last OAC dose and ablation, including the once daily direct OACs rivaroxaban and edoxaban. To ensure compliance with the protocol during the hospital stay for AF ablation, OACs were withheld or otherwise administered under the observation by nursing stuff. VKA therapy was continued only if prothrombin time (PT)-international normalized ratio (PT-INR) was below the therapeutic range, which was 2.0-3.0 for patients younger than 75 years of age and >1.6 for those over 75 years of age old according to the current guidelines of the Japanese Circulation Society. 17 Patients who skipped more than two doses of OACs before the procedure were excluded from the analysis.

| Anticoagulation during the procedure
Unfractionated heparin (UFH) was used for intraprocedural anticoagulation. Intravenous UFH was administered at 100 U/kg as an initial bolus immediately after the initial transseptal puncture. ACT was measured every 20 minutes after the initial UFH administration, and additional UFH was administered at 50-60 U/kg every 20 minutes to achieve and maintain an ACT of >300 seconds based on the measured ACT.

| AF ablation
Transesophageal echocardiography (TEE) was performed in all patients within one week before the ablation to exclude left atrial thrombus and valvular heart diseases. Transthoracic echocardiography (TTE) was required within 1 month before the ablation to evaluate cardiac function and the next day after the procedure to assess for cardiac complications due to the ablation. The following TTE parameters before ablation were used for baseline: left atrial diameter, left atrial volume index, and left ventricular ejection fraction. TTE the day after the ablation was allowed if no TTE was performed before the ablation. Plasma coagulation markers including PT and activated partial thromboplastin time were measured before each session.
Venous access was established through the right or left femoral vein and the right internal jugular vein. Cryoenergy was applied for 180-240 seconds to each PV. After successful PV isolation, isoproterenol (10-20 μg/mL) was administered.
At least 30 minutes after successful PV isolation, electrical activity in each PV was reconfirmed by pacing on the inside and the antrum of each PV to confirm entrance/exit block and by intravenous adenosine triphosphate to unveil dormant conductions. If AF remained after successful PV isolation, intracardiac cardioversion (10-30 J) was performed to restore sinus rhythm. In patients with nonparoxysmal AF, additional linear ablations were performed for the LA roof, bottom, and mitral isthmus. If macroreentrant atrial tachycardias (AT) were inducible with atrial burst pacing from the coronary sinus electrodes, additional RF ablations were performed to eliminate inducible AT. All ablation procedures were performed under conscious or deep sedation with an initial bolus of pentazocine and hydroxyzine pamoate and continuous dexmedetomidine administration. After completion of the ablation, protamine (30 mg) was administered to reverse the effect of heparin, and all sheaths were removed. All patients were followed at 1 and 6 months after the ablation.

| Statistical analysis
All continuous variables were expressed as means ± standard deviation or medians with interquartile ranges. All categorical variables were reported as number (percentage) of patients. Unpaired Student's t test and one-way analysis of variance (ANOVA) were used to compare continuous variables. Serial changes in ACT were evaluated by repeatedmeasures ANOVA. Categorical variables were compared with the chi-squared test or Fisher's exact test. A P-value of <.05 was considered statistically significant. Logistic regression analysis was performed for the evaluation of clinical predictors of TTA within 60 minutes (TTA-60).
All results were analyzed with SPSS base 11.0J for Windows.

| Patient population and OAC distribution
All patients were followed at 1 and 6 months. Of a total of 272 patients who underwent AF ablation during the study period, three patients were excluded because of the lack of preprocedural OACs due to hemodialysis (n = 2) and poor adherence (n = 1). Consequently, a total of 269 patients were included in the final analyses. The baseline characteristics of the study cohort are presented in Table 1. Warfarin was used in 14% of the patients, whereas direct OACs were used in 86% of the patients and included dabigatran (17%), rivaroxaban (30%), apixaban (29%), and edoxaban (10%). The mean age was significantly higher in the apixaban and edoxaban groups compared with the other groups.
There was a difference in sex distribution, especially in the dabigatran group. Inappropriate OAC doses were observed in the rivaroxaban (9.9%; overdose/underdose, 2/6), apixaban (10%; overdose/underdose, 2/6), and edoxaban (15%; overdose/underdose, 4/0) groups. The body weight was significantly lower in the apixaban group. No significant differences in CHADS 2 scores or cardiac function on TTE were observed among the groups. The treatment interruption was significantly longer for the OACs with once daily dosing (warfarin, rivaroxaban, and edoxaban) than those with twice daily dosing (dabigatran and apixaban) (mean treatment interruption, 25.3 ± 3.4 and 16.9 ± 2.2 hours for once daily and twice daily dosing, respectively; P < .0001). The details related to interruption durations are shown in Table 1.

| UFH and mean ACT
There were significant differences in mean initial and additional UFH doses among the OACs, although the initial and mean additional UFH doses were predetermined at 100 and 50-60 U/kg, respectively. Compared with the warfarin group (98 ± 12 U/kg), the rivaroxaban (103 ± 12 U/kg, P = .035) and apixaban (106 ± 13 U/ kg, P = .001) groups required higher doses for the initial UFH dose.

| Mean TTA and transition of the change in ACT
As shown in Table 2, the mean TTA was comparable between the dabigatran and warfarin groups. Factor Xa inhibitors (rivaroxaban, apixaban, and edoxaban) demonstrated significantly longer TTA than warfarin.
The time course of ACT values measured overtime for each direct OAC is shown in Figure 1. Compared with the warfarin group, there were no statistically significant differences in the dabigatran (P = .77; Figure 1A), apixaban (P = .21; Figure 1C), and edoxaban (P = .73; Figure 1D) groups. However, only the rivaroxaban group demonstrated a late catch-up type change in ACT and a significantly slower increase in ACT compared with the warfarin group (P = .013, Figure 1B).

| Clinical predictors for TTA-60
Female sex, body weight of ≤60 kg, age of ≥75 years, and warfarin and dabigatran use were more frequent in the TTA-60 group, whereas rivaroxaban use in the non-TTA-60 group was significantly more frequent than that in the TTA-60 group (Table 3). Based on these results, the clinical predictors for TTA-60 were evaluated by multivariate analysis. As shown in Table 4, body weight of <60 kg, age of >75 years, and warfarin and dabigatran use were clinical predictors for TTA-60.
Conversely, rivaroxaban use was a risk factor for non-TTA-60.

| Complications
Pericardial effusion (PE), including hemopericardium, occurred in five patients, with an incidence of 1.86%. Rivaroxaban and warfarin were used in two and three patients, respectively. The characteristics of these patients are presented in Table 5. The rate of warfarin use and body weight were significantly higher in the PE group than the non-PE group, whereas there were no significant differences in any of the other baseline characteristics between the two groups.
One patient with PE required surgical repair, whereas the remaining patients recovered by pericardiocentesis and anticoagulation reversal with protamine.
Thromboembolism within 30 days was recognized in one patient in the rivaroxaban group (0.37%). The female patient complained of mild dizziness 3 weeks after the procedure. Brain magnetic resonance imaging revealed a subacute stroke in the right parietal lobe that was not related to her symptoms ( Figure 2). The patient recovered shortly without any residual neurological deficits.

| D ISCUSS I ON
Accumulating evidence and advances in technologies and skills established AF ablation as a cornerstone therapeutic approach.
Currently, direct OACs play a central role in anticoagulation therapy for AF by replacing VKAs due to their noninferiority/superiority in the prevention of ischemic stroke and the lower risk of bleeding compared with VKAs. [18][19][20][21][22] However, evidence is limited regarding the outcomes of concomitantly used intraprocedural heparin and direct OACs during AF ablation. Of note, Asians are at a higher risk of intracranial hemorrhage than non-Asians during treatment with VKAs. 23 Furthermore, AF ablation is currently indicated even for elderly patients who are prone to bleeding. 24 In the era of AF ablation, the safety of periprocedural direct OAC dosing should be elucidated.
Several studies demonstrated the characteristics of MID for AF ablation by comparing with uninterrupted warfarin [25][26][27] and reported that uninterrupted warfarin consistently demonstrated shorter TTA values than MID. However, the study design in these trials was advantageous for uninterrupted warfarin. Konduru et al, 25 Bassiouny et al, 26  In the present study, rivaroxaban and apixaban required significantly higher doses of UFH compared with warfarin throughout the procedure, whereas dabigatran and edoxaban did not. Conversely, previous studies reported that dabigatran required higher heparin doses. 26,29 The disagreement between the current and the two previous studies  Warfarin-use 3 (60) 34 (13) .020 Values are mean ± SD or the number(%). Abbreviations are shown in  33 We believe that MID should be considered as an alternative periprocedural direct OAC dosing protocol, especially for patients with higher bleeding risk; however, differences in responsiveness to UFH should be taken into account for optimal intraprocedural anticoagulation.

| LI M ITATI O N S
We acknowledge several limitations in the current study. First, this was a retrospective observational study performed at a single-center in Japan. Oral anticoagulation was performed in accordance with the guidelines of the Japanese Circulation Society. Thus, the practices were different from those in European, American, and Asian countries. The patients were not blinded, randomized, or adjusted based on their background characteristics. Second, the patient enrollment was consecutive; however, three patients not on OACs were excluded. Third, the anticoagulant effects of the direct OACs could not be accurately estimated by ACT. We recognize that ACT is merely a surrogate marker for intraprocedural anticoagulation. Fourth, we could not determine that the MID protocol and TTA-60 guaranteed the prevention of future clinical thromboembolic or bleeding events.
Fifth, we recognize underdoses of factor Xa inhibitors may confound the results, though they were not frequent (7.5% in RG, and 7.6% in AG). Furthermore, all patients were Japanese, who were fundamentally at a lower risk for TE (mean CHADS 2 , 1.1 ± 1.0). Therefore, it is prudent to confirm these results in other populations. A randomized, prospective, multicenter study to validate the safety of the MID protocol for AF ablation is required. The patient complained of mild dizziness 3 weeks after the AF ablation. The brain MRI images reveal a round-shaped lesion (arrows) in the right parietal lobe suggesting subacute stroke. No abnormal findings are observed in cerebellum or brain stem

| CON CLUS IONS
MID is a feasible option for periprocedural anticoagulation strategy for AF ablation. However, differences in responsiveness to UFH among different direct OACs should be taken into account to achieve optimal intraprocedural anticoagulation. The present study elucidated a clear difference between dabigatran, a direct thrombin inhibitor, and other factor Xa inhibitors.

ACK N OWLED G EM ENT
We appreciate all staff members of the catheterization laboratory at the Cardiovascular Center at Kyoto Katsura Hospital for their dedication to caring.

CO N FLI C T O F I NTE R E S T S
All authors declare no conflict of interests for this article.