Safety and efficacy of uninterrupted treatment with edoxaban or warfarin during the peri‐procedural period of catheter ablation for atrial fibrillation

Abstract Background The real‐world safety and efficacy of uninterrupted anticoagulation treatment with edoxaban (EDX) or warfarin (WFR) during the peri‐procedural period of catheter ablation (CA) for atrial fibrillation (AF) are yet to be investigated. Methods We conducted a two‐center experience, observational study to retrospectively investigate consecutive patients who underwent CA for AF and received EDX or WFR. We examined the incidence of thromboembolic and bleeding complications during the peri‐procedural period. Results The EDX and WFR groups included 153 and 103 patients, respectively (total: 256 patients). Demise or thromboembolic events did not occur in either of the groups. The incidence of major bleeding in the EDX and WFR groups was 0.7% and 2.9%, respectively. The total incidence of major/minor bleeding in the EDX and WFR groups was 7.8% and 8.7%, respectively. Of note, the incidence of bleeding complications in the uninterrupted WFR strategy group was markedly high in patients with an estimated glomerular filtration rate (eGFR) <30 (75%) or a HAS‐BLED score ≥3 (60%). Patients with eGFR ≥30 and a HAS‐BLED score ≤2 had a lower incidence of bleeding (<10%), regardless of the administered anticoagulation drug (EDX or WFR). Conclusions This study confirmed the safety and efficacy of uninterrupted anticoagulation therapy using EDX or WFR in real‐world patients undergoing CA for AF. Patients with severely impaired renal function and/or a higher bleeding risk during uninterrupted therapy with WFR were at a prominent risk of bleeding. Therefore, particular attention should be paid in the treatment of these patients.


| INTRODUC TI ON
Atrial fibrillation (AF) is one of the clinically relevant arrhythmias commonly encountered in daily clinical practice. 1 Owing to an increased risk of thromboembolism, anticoagulation treatment is recommended for patients with AF in the European and Asia Pacific countries, including Japan. 2 Vitamin K antagonists, such as warfarin (WFR), have been used for a long time to prevent thromboembolic events in patients with AF. However, the benefits of this treatment are occasionally counterbalanced by the risk of bleeding. 3  Apart from pharmacological therapy, catheter ablation (CA) is another therapeutic option for AF. 5 However, the risk of fatal bleeding and thrombosis during the peri-procedural period of CA often results in controversy regarding the optimal usage of anticoagulation drugs. 6 Thus far, only a few studies have assessed the safety and efficacy of EDX during the peri-procedural period of CA worldwide. 7 Therefore, we sought to investigate the incidence of bleeding and thromboembolic complications of uninterrupted treatment with EDX during the peri-procedural period of CA, vs those of WFR in the real-world Japanese population.

| Study subjects
This study complied with the tenets of the Declaration of Helsinki, and a locally appointed ethics committee approved the research protocol. We conducted a two-center experience, observational study to retrospectively investigate consecutive patients who underwent CA for AF and received anticoagulation treatment with WFR or EDX at the New Tokyo Hospital and Ageo Central General Hospital between January 2015 and December 2018. The decision to proceed with CA was based on the consensus of arrhythmia specialists according to the guidelines of the Japanese Circulation Society. 8 Clinical information regarding the patient characteristics and conditions was acquired from the medical records. The CHADS 2 and HAS-BLED scores were calculated as previously described. 9,10 Peri-procedural complications, such as major/minor bleeding and thromboembolic episodes, were also evaluated from the medical records. Major bleeding was defined based on the criteria established by the International Society on Thrombosis and Haemostasis. 11 Minor bleeding was defined as nonmajor bleeding that did not necessitate a surgical procedure or blood transfusion, but necessitated intervention such as a manual compression, additional resting time, laboratory testing or image inspection. The observational period for peri-procedural complications was 1 month prior to and 2 months after the procedure. The selection of anticoagulation treatment was at the physicians' discretion. Patients who received treatment with EDX were classified in the EDX group. For those with a creatinine clearance level <50 mL/min, low body weight (<60 kg) or concomitant use of potent P-glycoprotein inhibitors, a low dose (30 mg) of EDX was administered (39.2%). The remaining patients (60.8%) received a high dose (60 mg) of EDX. On the day of the CA procedure, EDX was administered within 8 hours prior to or after the procedure, based on the operators' discretion. Patients who received treatment with WFR were classified in the WFR group. For patients aged >70 years in this group, WFR was adopted to maintain a prothrombin time-international normalized ratio (PT-INR) of 2.0-3.0 or 1.6-2.6. 8 The administration of WFR was continued on the procedural day, except when the PT-INR exceeded the target range. In both groups, anticoagulation therapy was administered for at least 4 weeks prior to and 2 months after the procedural day and was continued on the procedural day.

| CA procedure
Transesophageal echocardiography was performed in most patients (97%) on the day prior to the procedure. Dexmedetomidine hydrochloride, pentazocine, and fentanyl citrate were used as intravenous anesthetics. A bolus of unfractionated heparin (80-120 U/kg) was injected prior to performing the septal puncture. Subsequently, the administration of unfractionated heparin (500-3000 U) was continued to achieve an active clotting time (ACT) of 300-350 seconds every 15 minutes during the procedure. CARTO (Biosense Webster Inc) or ENSITE (Abbott) were used as the three-dimensional mapping system. All patients who underwent initial ablation, and those who underwent secondary ablation because of pulmonary vein reconduction, were subjected to pulmonary vein isolation. Superior vena cava isolation, cavo-tricuspid isthmus ablation, and left atrial posterior wall isolation were performed in patients in whom nonpulmonary vein foci were induced or additional ablation was deemed necessary by the operator. Following termination of the procedure, hemostatic compression was performed for at least 10 minutes until adequate hemostasis was achieved.
The strategy for radiofrequency CA at the New Tokyo Hospital and Ageo Central General Hospital was as follows: (a) a BeeAT The strategy for cryoballoon ablation (Arctic Front Advance™ 28 mm cryoballoon; Medtronic, Inc) was as follows: (a) a BeeAT cardioversion catheter (Japan Lifeline) was placed through a 7 Fr sheath into the coronary sinus using the jugular vein approach; (b) two 8.5 Fr sheaths or one 8.5 Fr sheath and one 7 Fr sheath were inserted using the femoral vein. One sheath was exchanged for a steerable 15 Fr sheath (FlexCath™ Advance; Medtronic, Inc) after performing the septal puncture; and (c) a 4-pole or 10-pole catheter was placed for ventricular pacing through the femoral vein.

| Statistical analysis
Continuous variables are expressed as the mean ± standard deviation or median (interquartile range). The baseline clinical data were compared between the groups using the Student's t test or Mann-Whitney U test. Categorical variables are presented as the frequency and percentage using Fisher's exact test. The odds ratio for the risk of complications was calculated via logistic regression analysis. A value of P < .05 denoted statistical significance. All statistical analyses were performed using the SPSS software (IBM SPSS Statistics for Windows, version 25.0; IBM Corporation).

| Patient characteristics
A total of 256 patients who underwent CA for AF between January 2015 and December 2018 were selected. Patients with mechanical heart valves (n = 2), ischemic heart disease receiving dual antiplatelet therapy (n = 1), and kidney failure undergoing hemodialysis (n = 1) were included in the present study. The EDX and WFR groups included 153 and 103 patients, respectively. Overall, the time in therapeutic range for WFR was 73.7% from 2 months before to 3 months after the procedure. In the EDX group, 47 patients (31%) received treatment within 8 hours prior to CA, while the remaining 106 (69%) were treated within 8 hours after CA. The baseline patient characteristics are shown in Table 1. There were no significant differences between the EDX and WFR groups in terms of age, sex, height, and body weight. The rate of a CHADS 2 score ≥2 was significantly lower in the EDX group than the WFR group (19% vs 31%, respectively, P = .026). The patients in the WFR group tended to have a history of congestive heart failure (11% vs 3%, respectively, P = .008) and higher levels of B-type

| Complications
The complications which occurred during the peri-procedural period are shown in Table 2. There was no occurrence of thromboembolic events in either of the groups.  Figure S1). Table 3 includes the results of a subanalysis conducted to assess whether the timing of EDX administration on the day of CA (pre-CA or post-CA) affected the incidence rate of complications.
However, it should be noted that the only one major bleeding episode recorded occurred in a patient "pretreated" with EDX.
Subsequently, we sought to investigate the risk factors associated with bleeding complications in the treatment groups (Table 4).
In the WFR group, patients with bleeding events had a lower estimated glomerular filtration rate (eGFR) (45 ± 28 vs 67 ± 17, respectively, P = .001) and a higher HAS-BLED score (P = .004) compared with those without complications. In contrast, in the EDX group, there were no statistical differences in any parameters. Figure 1 illustrates that the incidence of bleeding complications in the WFR group was markedly high in patients with an eGFR <30 (75%) or a HAS-BLED score >3 (60%), while those with an eGFR ≥30 and a HAS-BLED score ≤2 had a lower incidence of bleeding (<10%) regardless of the administered anticoagulation drug. The odds ratio for the risk of bleeding complications in the WFR group patients with an eGFR <30 and/or a HAS-BLED score ≥3 vs those without was 46.5 (confidence interval: 4.1-517.2, P = .002).

| D ISCUSS I ON
The main findings of the present, real-world, observational study were as follows. Firstly, uninterrupted anticoagulation treatment with EDX or WFR during the peri-procedural period of CA prevented the occurrence of thromboembolism in real-world patients receiving CA for AF. Secondly, uninterrupted anticoagulation therapy with EDX or WFR was overall safe based on the low incidence of serious bleeding complications noted in both groups. Thirdly, patients with renal failure (eGFR <30) and/or higher bleeding risk (HAS-BLED score ≥3) who received uninterrupted anticoagulation therapy with WFR were at a markedly high risk of bleeding complications. In the present observational study, we examined the real-world Japanese population, including patients with mechanical heart valves, ischemic heart disease receiving dual antiplatelet therapy, and kidney failure undergoing hemodialysis. These patients were excluded from the randomized clinical studies, such as the ELIMINATE AF trial. Our real-world observational study confirmed the efficacy of uninterrupted anticoagulation therapy with EDX or WFR during CA for AF in such high-risk patients (thromboembolism: 0%, in both groups). The incidence of major bleeding complications in the EDX and WFR groups was only 0.7% and 2.9%, respectively, warranting the safety of uninterrupted anticoagulation treatment in real-world practice. In our study, the total rate of major and minor bleeding complications in the EDX and WFR groups was 7.8% and 8.7%, respectively. The latter rate was relatively higher than that reported in the ELIMINATE AF trial, probably reflecting a higher bleeding risk in the WFR group in our study. The Fushimi registry, a Japanese allcomer registry consisting of 3,282 patients with AF, showed a similar trend to that observed in our study in terms of patient characteristics: higher comorbidity rates (diabetes mellitus, congestive heart failure, chronic kidney disease, and nonparoxysmal AF) and higher HAS-BLED and CHADS 2 scores in the WFR group compared with the DOAC group, and the absence of an oral anticoagulant group. 16 Thus, we believe that the results of our study can be applied to the real-world clinical practice.

TA B L E 2 Incidence of thromboembolic and bleeding complications
In our study, it was also intriguing that the total rate of major and The results showed a low composite incidence of F I G U R E 1 Incidence of major/minor bleeding complications stratified according to the eGFR and HAS-BLED score. A, The bar graphs represent the incidence of bleeding complications (left, edoxaban group; right, warfarin group) stratified according to the eGFR (≤29, 30-49, ≥50) in each group. Of note, patients with an eGFR <30 receiving uninterrupted therapy with warfarin were at a prominent risk of bleeding complications. B, The bar graphs represent the incidence of bleeding complications (left, edoxaban group; right, warfarin group) stratified according to the HAS-BLED score (0, 1-2, ≥3) in each group. Of note, patients with an HAS-BLED score ≥ 3 receiving uninterrupted therapy with warfarin were at a prominent risk of bleeding complications. EDX, edoxaban; eGFR, estimated glomerular filtration rate; WFR, warfarin thromboembolism and major bleeding (0.2%). 21 Hirao et al reported that administration of rivaroxaban within 8 hours prior to the CA procedure was associated with a higher incidence of major bleeding than administration 8-28 hours prior to the CA procedure (18% vs 0%, respectively, P = .026). 22 In the ELIMINATE AF trial, all patients with uninterrupted EDX therapy received the medicine in the evening (after the CA procedure), and reported a relatively low incidence of major bleeding (2.7%). 7 In the present study, EDX was administered within 8 hours prior to or after the procedure at the operators' discretion. As shown in Table 3, the timing of EDX administration (pre-CA or post-CA) did not significantly affect the incidence of bleeding complications. However, the only one event of major bleeding recorded in the EDX group occurred in a patient who received pre-CA treatment. In addition, more heparin was needed to maintain an ACT of 300-350 seconds in the EDX group than the WFR group.
A previous study revealed that more heparin was needed in patients with anticoagulation therapy using EDX than those using WFR. This study indicated that the difference in the half-lives of WFR and EDX affects the ACT. 23 Moreover, similar results have been reported for apixaban, rivaroxaban, and dabigatran. 24 Our study produced similar results and supported the findings of other studies.

| Limitations
There were several limitations in this study. Firstly, this was a retrospective, two-center experience, observational study involving a relatively small number of patients. However, our study could show clinically relevant results with statistical significance, as other studies enrolling a relatively small number of patients. 15,25 Secondly, there could be an operator bias. The operators' experience with CA ranged from at least 100 to >2000 procedures. The rates of procedural complications, such as cardiac tamponade and false femoral aneurysm, are associated with the skill of an operator. Thirdly, patients with insufficient adherence to treatment were excluded from the analysis, which may have induced a patient-selection bias.

| CON CLUS ION
This real-world observational study confirmed the safety and efficacy of uninterrupted anticoagulation treatment during the periprocedural period of CA for AF. Uninterrupted therapy with EDX or WFR resulted in lower rates of thromboembolic and bleeding complications. Patients with severely impaired renal function and/or a higher bleeding risk during uninterrupted therapy with WFR were at a prominent risk of bleeding complications. Therefore, utmost attention should be paid in the treatment of these patients to prevent bleeding complications.

ACK N OWLED G EM ENTS
NAI (a translation company, Kanagawa, Japan) supported the English revision of this manuscript.

CO N FLI C T O F I NTE R E S T
All authors have no conflict of interest relevant to the topic of this manuscript.