Author + information
- Received February 1, 2018
- Revision received August 8, 2018
- Accepted August 13, 2018
- Published online December 17, 2018.
- Takehiro Kimura, MDa,
- Shin Kashimura, MDa,
- Takahiko Nishiyama, MDa,
- Yoshinori Katsumata, MDa,
- Kohei Inagawa, MDb,
- Yukinori Ikegami, MDb,
- Nobuhiro Nishiyama, MDa,
- Kotaro Fukumoto, MDa,
- Yoko Tanimoto, MDb,
- Yoshiyasu Aizawa, MDa,
- Kojiro Tanimoto, MDb,
- Keiichi Fukuda, MDa and
- Seiji Takatsuki, MDa,∗ ()
- aDepartment of Cardiology, Keio University School of Medicine, Tokyo, Japan
- bDepartment of Cardiology, National Hospital Organization, Tokyo Medical Center, Tokyo, Japan
- ↵∗Address for correspondence:
Dr. Seiji Takatsuki, Department of Cardiology, Keio University School of Medicine, 35 Shinanomachi, Shinjuku-ku, Tokyo, 160-8582, Japan.
Objectives This randomized study compared uninterrupted rivaroxaban therapy with warfarin therapy as prophylaxis against catheter ablation (CA)-induced asymptomatic cerebral infarction (ACI) and identified the risk factors of rivaroxaban.
Background The reported incidence of ACI during CA for atrial fibrillation (AF) remains at 10% to 30%, and periprocedural oral anticoagulation could affect this incidence.
Methods Patients with nonvalvular AF undergoing radiofrequency CA were randomly assigned to receive either uninterrupted rivaroxaban or warfarin as periprocedural anticoagulation therapy. CA was performed after at least 1 month of adequate anticoagulation. Cerebral magnetic resonance imaging (MRI) was performed within 2 weeks before and 1 day after CA to detect ACI.
Results A total 132 patients were enrolled; 127 (median: 60.0 years of age; 83.5% males; 64.6% incidence of paroxysmal AF) complied with the study protocol and were analyzed; 64 patients received rivaroxaban, and 63 patients received warfarin. The rates of CA-induced ACI in the rivaroxaban group (15.6% [10 of 64 patients]) were similar to those in the warfarin group (15.9% [10 of 63 patients]; p = 1.000). No thromboembolic events developed; no differences in major or nonmajor bleeding rates were observed between the 2 drug groups (3.1% vs. 1.6%, respectively, or 18.8% vs. 19.0%, respectively). Multiple regression analysis indicated that the presence of deep and subcortical white matter hyperintensity (p = 0.002; odds ratio [OR]: 5.323) and the frequency of cardioversions (p = 0.016; OR: 1.250) were associated with the incidence of ACI.
Conclusions No notable differences were found between the incidence of CA-induced ACI in the rivaroxaban group and that in the warfarin group in this randomized study.
- asymptomatic cerebral infarction
- atrial fibrillation
- cardiac magnetic resonance imaging
Catheter ablation (CA) is an established therapy for atrial fibrillation (AF), particularly for symptomatic patients, before the AF becomes permanent (1). However, periprocedural complications remain an issue. The incidence of CA-associated complications, including stroke (0.23%) and transient ischemic attack (0.71%), are estimated to be 4.54% (2). Estimated rates of periprocedural asymptomatic cerebral infarction (ACI) vary, with reported rates of 6.3% (3), 12.2% (4), 14.3% (5), or 38.4% (6) depending on the CA procedure or condition. Therefore, ACI may be affected by procedural rather than patient-related factors (7). Furthermore, interrupting anticoagulation during AF ablation may be associated with higher rates of ACI (8,9). Considering that elderly patients with ACI are at greater risk for dementia and cognitive impairment than those without ACI (10), ACI should not be ignored.
Direct oral anticoagulant agents have been approved since 2008 and are widely used perioperatively during CA for AF because of their minimal interaction with other drugs or food, the nonrequirement for monitoring, and the rapid onset of anticoagulant effects. The incidence of perioperative complications requiring hospital admission during CA for AF was reportedly 8.9% and 14.0% with rivaroxaban and warfarin treatment, respectively (11). Rivaroxaban potentially shortens the duration of hospital admission. However, only limited data exist regarding ACI incidence with different anticoagulant agents during the perioperative period (12). In this randomized study, either uninterrupted rivaroxaban or warfarin was administered to patients with AF undergoing CA, and their efficacy and safety were compared, particularly in regard to the occurrence of ACI.
This randomized, open-label, assessor-blind, parallel-group study was conducted at 2 Japanese institutions, Keio University Hospital and Tokyo Medical Center, between July 2014 and January 2016. Rivaroxaban and warfarin were perioperatively administered to patients with nonvalvular AF undergoing CA and were compared according to their efficacy and safety. The study was conducted according to Declaration of Helsinki tenets and Ethical Guidelines for Medical and Health Research Involving Human Subjects (established by the Japanese Ministry of Health, Labor and Welfare). The study protocol was approved by the ethics committees of Keio University Hospital and Tokyo Medical Center (UMIN000013341). Written informed consent was obtained from all patients. Criteria included patients: 1) who had nonvalvular AF; 2) those undergoing CA for AF; 3) those with left atrial diameter of ≤55 mm; and 4) those between 20 and 80 years of age. Patients were excluded if they: 1) had experienced a stroke or transient ischemic attack at 6 months before enrollment; 2) had major bleeding at 12 months before enrollment; 3) had undergone coronary artery bypass grafting or other major surgery at 6 months before enrollment; 4) had experienced myocardial infarction at 2 months before the scheduled CA date; 5) had a creatinine clearance of <30 ml/min upon enrollment; 6) required treatment with ≥2 antiplatelet drugs; 7) showed contraindications for rivaroxaban or warfarin therapy; and 8) were deemed ineligible by the physician in charge.
The sample size was not based on statistical power calculation but on a 2-year recruitment period at the study sites for this exploratory study. Approximately 150 patients were randomized to 2 treatment arms at a 1:1 ratio, with approximately 10% discontinuation rate.
The primary endpoint was new rates of ACI, and the secondary endpoint was occurrence of major bleeding or thromboembolic events (myocardial infarction, ischemic stroke, systemic embolism, or vascular death) during the 30-day observation period after CA. Transient ischemic attack and cerebral infarction were considered thromboembolic events, whereas pericardial effusion and hematoma were considered hemorrhagic complications. Hematoma requiring blood transfusion or surgical intervention and pericardial bleeding necessitating drainage (cardiac tamponade) were classified as major bleeding events. Mild hematoma and pericardial hemorrhage not requiring surgical intervention were considered nonmajor bleeding events.
Cerebral cardiac magnetic resonance imaging
Cerebral magnetic resonance imaging (MRI) was performed twice during the periprocedural period, that is, within 2 weeks before (pre-MRI) and the day after (post-MRI) CA, to obtain ACI-associated cerebral pathology findings. The average interval between pre- and post-MRI was 3.7 ± 4.2 days. Cerebral MRI was performed using a 1.5-T scanner (Signa, GE Healthcare, Milwaukee, Wisconsin). The imaging protocol included T1-weighted imaging sequences (TR/TE, 520/8 ms; NEX, 1; FOV, 20 cm; matrix, 256 × 192; slice, 5/2.5 mm); T2-weighted imaging sequences (TR/TE, 4,000/103 ms; other parameters were the same as those for T1-weighted imaging); fluid-attenuated inversion recovery (FLAIR) sequences (TI/TR/TE, 8,802/101/2,200 ms; other parameters same as those for T1-weighted imaging); diffusion-weighted imaging (DWI) sequences (TR/TE, 5000/65 ms; 1 b = 1,000; FOV, 25 cm; matrix, 128 × 128; other parameters same as those for T1-weighted imaging); and time-of-flight sequences (TR/TE/FA, 25/3/20°; matrix, 256 × 160; slice, 1.2 mm). Lacunar infarction was identified as focal lesions with diameters of ≥2 mm with hyperintensity on a T2-weighted image accompanied by ill-defined irregular margins, low signals on T1-weighted images, and high signals on FLAIR images. A newly developed ACI was diagnosed when a lesion suspected of being an infarction from an imaging test was not observed before initiating treatment and had neither neurological signs nor symptoms, including bilateral differences in deep tendon reflexes and suspected cerebrovascular dementia, nor subjective symptoms, including transient ischemic attack (13). ACI diagnoses were confirmed by comparing pre- and post-CA DWI (Figure 1).
According to Japanese Brain Dock Guidelines of 2008 (14), deep and subcortical white matter hyperintensities (Figures 2A and 2B) were determined based on 5 grade categories: grade 0 = absence; grade 1 = punctate foci with <3-mm diameter or extended perivascular foci; grade 2 = punctate or discrete foci with ≥3-mm diameter on subcortical and deep white matter; grade 3 = confluent foci with an indistinct boundary on subcortical and deep white matter; and grade 4 = widely distributed confluence on most of the white matter. Grades 2, 3, and 4 deep and subcortical white matter hyperintensities were considered severe. Periventricular hyperintensities (Figures 2C and 2D) were also assessed by using 5 grade categories: grade 0 = absent or “periventricular rims” only; grade 1 = localized lesions such as “periventricular caps”; grade 2 = lesions extending along the whole periventricular area; grade 3 = irregular periventricular hyperintensity extending into the deep white matter; and grade 4 = periventricular hyperintensity extending throughout deep and subcortical white matter. Grades 2, 3, and 4 periventricular hyperintensities were considered severe. Images were assessed by 2 independent radiologists blinded to patient information, and both observers reached a consensus regarding each set of images.
Anticoagulation strategies and ablation procedures
Patients were randomized to receive either rivaroxaban or warfarin at a 1:1 ratio at ≥30 days before CA and underwent transesophageal echocardiography at 7 days before CA to detect left atrium (LA) or left atrial appendage (LAA) thrombus. Patients randomized to uninterrupted warfarin were maintained at a ≥2.0 international normalized ratio (INR) (in cases with patients ≥70 years of age, an INR ≥1.6), whereas those randomized to rivaroxaban should have had a documented INR of ≤2.0 (in cases with patients ≥70 years of age, an INR of ≥1.6) before the initiation of rivaroxaban therapy. A standard Japanese guideline-approved 15-mg rivaroxaban dose was administered once daily in the evening (15,16). Rivaroxaban was initiated for patients not treated with warfarin or those with a prothrombin time-INR (PT-INR) of <2.0 on the day of randomization. Warfarin was discontinued in patients with PT-INR of ≥2.0, whereas rivaroxaban was initiated in those with PT-INR of <2.0. Patients at a high risk for bleeding and with a creatinine clearance rate of 30 to 49 ml/min, ≥75 years of age, or weighing ≤50 kg received another Japanese guideline-approved 10-mg standard dose once daily for safer execution of the study (17). Warfarin was administered once daily in the evening, although the dose was controlled to attain a PT-INR of 2.0 to 3.0. Patients ≥70 years of age received a dose controlled to attain PT-INR of 1.6 to 2.6 according to the Japanese guidelines for the management of anticoagulant and antiplatelet therapy in cardiovascular disease (18). Rivaroxaban or warfarin treatment was not interrupted throughout the periprocedural period.
During CA, heparin was intravenously administered at a dose of 100 U/kg of body weight just before atrial septal puncture to reduce the risk of ACI (8). Activated clotting time was measured at 20-min intervals and maintained at 300 to 350 s throughout CA. During the procedure, circumferential pulmonary vein isolation was performed by using a 3-dimensional electroanatomical mapping system, and radiofrequency (RF) current deliveries were applied through an irrigated-tip catheter (Thermocool SF; Biosense Webster, Inc., Irvine, California), using RF energy (30 to 35 W, except for areas near the esophagus where 25 W was applied). When the temperature probe in the esophagus reached 38°C, the RF energy was turned off. At least 2 long sheaths, 1 for the ablation catheter and the other for the lasso catheter to confirm pulmonary vein isolation, were advanced into the LA. Box isolation, including all 4 pulmonary veins and the left atrial posterior wall or circumferential pulmonary vein isolation, was performed in patients with persistent AF. Cavotricuspid isthmus ablation and mitral isthmus ablation were performed during AF ablation as needed. After the procedure was completed, heparin was neutralized with protamine sulfate. The sheath was removed, and hemostasis was achieved. For observational purposes, patients rested in the supine position for 5 h. Anticoagulant therapy was resumed on the evening of that day.
A portable device was used to perform electrocardiography twice daily and whenever patients had symptoms throughout the observation period.
Patient characteristics were compared using Fisher exact test for nominal variables and Mann-Whitney U test for ordered and continuous variables. Fisher exact test was used to compare the rates of ACI in the rivaroxaban group with those in the warfarin group. To identify risk factors that were independently associated with periprocedural ACI, we conducted univariate and multivariate analyses using a logistic regression model. Multivariate analysis was conducted using stepwise backward elimination (maximum number of iterations: 20). Considering the sparseness of outcome events, variables with missing data, which would adversely affect statistical power, were excluded from the multivariate analysis. The significance levels for all tests were set at p values of <0.05. Statistical analyses were conducted using SPSS version 23.0 statistics (IBM Corp., Armonk, New York) and R version 3.0.2 (R Foundation for Statistical Computing, Vienna, Austria).
A total of 132 patients with nonvalvular AF were enrolled, 5 of whom dropped out; 127 patients complied with the study protocol and were analyzed. The median age of the study population was 60.0 years old (interquartile range 52.0 to 66.0 years of age), with 106 men (83.5%) and 82 women (64.6%) with paroxysmal AF. Concurrent illness at enrollment included congestive cardiac failure in 7 of 127 patients (5.5%), hypertension in 63 of 127 (49.6%), diabetes mellitus in 13 of 127 (10.2%), and dyslipidemia in 27 of 127 (21.3%).
Comparison between the rivaroxaban and warfarin groups
Sixty-four patients in the rivaroxaban group and 63 in the warfarin group underwent CA and head MRI before and after the procedure. All patients who underwent pre- and post-procedural MRI were subjected to intergroup comparisons of efficacy and safety, which we refer to as the full analysis set (Figure 3). Among patients in the rivaroxaban and warfarin groups, 84.4% and 84.1%, respectively, underwent AF ablations for the first time. Patient characteristics are shown in Table 1. No differences in age, sex, body mass index, AF type, blood pressure, left atrial appendage flow velocity, or degree of spontaneous echo contrast were observed between both treatment groups. PT-INR, activated partial thromboplastin time, and mean activated clotting time were significantly greater in the warfarin group than in the rivaroxaban group, whereas anti-Xa activity, factor X activity, and unfractionated heparin dose during CA were significantly greater in the rivaroxaban group than in the warfarin group. No significant differences regarding ACI incidence and the number of lesions were found between the rivaroxaban (15.6% [10 of 64 patients]) and warfarin (15.9% [10 of 63 patients]; p = 1.000) groups (Figures 4 and 5⇓⇓).
No symptomatic thromboembolic events occurred in either group. Cardiac tamponade, successfully treated with pericardial drainage, and clinically significant puncture site hematomas occurred in 1 patient each in the rivaroxaban group, whereas a puncture site hematoma developed in 1 patient in the warfarin group. Furthermore, 12 patients in each group exhibited nonmajor bleeding. Despite no significant intergroup difference, complication rates tended to be higher in the warfarin group (17.5%; 95% confidence interval [CI]: 9.1% to 29.1%) than in the rivaroxaban group (6.3%; 95% CI: 1.7% to 15.2%; p = 0.059). The incidence of puncture site hematomas were significantly higher in the warfarin (7 patients [12.1%]) than in the rivaroxaban (1.6%) group (Table 2). No cerebrovascular or cardiovascular event or other adverse event was considered serious. At 1 month after CA, AF symptoms disappeared in 49 rivaroxaban patients (76.6%) and 51 warfarin patients (81.0%), whereas AF was clinically alleviated in 12 rivaroxaban patients (18.8%) and 10 warfarin patients (15.9%), with no significant intergroup differences.
Degree of interobserver reproducibility in image interpretation
The degree of interobserver reproducibility (kappa score) while interpreting cerebral MRI was as follows: 92.9% (0.72) for ACI presence or absence, 92.9% (0.74) for the number of lesions (ACIs), 99.2% (0.89) for periventricular hyperintensities, 97.6% (0.92) for deep and subcortical white matter hyperintensities, and 93.7% (0.66) for lacunar infarctions, suggesting favorable interobserver reproducibility.
Factors associated with ACI
We evaluated factors associated with ACI in a dataset obtained from the 127 patients undergoing CA (Figure 3). Baseline patient characteristics of the 20 patients who developed ACI and the 107 patients who did not are shown in Table 3. Regarding patient-related factors, age at enrollment, blood urea nitrogen level, systolic blood pressure, diastolic blood pressure, and D-dimer levels obtained just before anticoagulant agent administration 1 day before ablation were significantly greater in patients who developed ACI than in those who did not. Rates of severe periventricular hyperintensities, severe deep and subcortical white matter hyperintensities, and lacunar infarctions were significantly greater in patients who developed ACI than in those who did not. Procedural factors showed that electrical cardioversion was performed significantly more often in patients who developed ACI than in those who did not.
Blood pressure data were not available for 4 patients with new onset ACI and were excluded from variables in the multivariate model to maintain statistical power. Univariate and multivariate stepwise logistic regression analyses revealed that severe deep and subcortical white matter hyperintensities and number of cardioversions were factors associated with ACI occurrence (Table 4).
Anticoagulation regimen and ACI
In the rivaroxaban group, ACI was observed in 7 of 57 patients (12.3%; 95% CI: 5.1% to 23.7%) receiving 15-mg doses and in 2 of 6 patients (33.3%; 95% CI: 4.3% to 77.7%) receiving 10-mg doses (p = 0.201). Patients treated with 10 mg of rivaroxaban (n = 6) were older than patients treated with 15 mg of rivaroxaban (n = 57) (median 67.5 years of age [IQR: 60.0 to 73.0] vs. 59.0 years of age [IQR: 51.0 to 64.0], respectively; p = 0.025). Height and body weight were lower in patients treated with 10 mg of rivaroxaban (height: 160.0 cm [IQR: 155.5 to 165.9 cm] vs. height: 170.1 cm [IQR: 164.6 to 175.8 cm]; p = 0.018; weight, 59.5 kg [IQR: 46.0 to 65.0 kg] vs. 70.0 kg [IQR: 63.0 to 78.0 kg]; p = 0.012), and creatinine clearance was also lower in this treatment group (46.5 ml/min [IQR: 41.0 to 49.0 ml/min] vs. 85.0 ml/min [IQR: 72.0 to 104.0 ml/min]; p < 0.001). In the warfarin group, ACI was observed in 7 of 55 patients (12.7%; 95% CI: 5.3% to 24.5%) and in 3 of 8 patients (37.5%; 95% CI: 8.5% to 75.5%) with target INRs of 2.0 to 3.0 and 1.6 to 2.6, respectively (p = 0.016). ACI incidence rate tended to be higher in low-dose patients in both groups, although it was not significant.
During CA for AF, perioperative uninterrupted rivaroxaban and warfarin treatments resulted in comparable rates of ACI. Severe deep and subcortical white matter hyperintensities detected by head MRI before CA and the number of cardioversions during CA for AF were clarified as risk factors of ACI.
Several findings have recently been published regarding differences in perioperative ACI incidences with anticoagulant drugs, including direct oral anticoagulant agents, in patients undergoing CA for nonvalvular AF. ACI rates with dabigatran were 26.7% and 10% with warfarin (p < 0.05) (11). The odds ratio (OR) of silent cerebral ischemic lesions for dabigatran versus warfarin was 2.287 (p = 0.042), with no significant differences for warfarin versus rivaroxaban, apixaban, and edoxaban (19). Discontinuing periprocedural oral anticoagulant agents was an independent predictor for ACI (OR: 2.58; p < 0.05) (9). Moreover, uninterrupted rivaroxaban, interrupted apixaban, and uninterrupted warfarin resulted in ACI rates of 16.4% (9 of 55 patients), 20.0% (10 of 50 patients), and 18.8% (13 of 69 patients), respectively (20). Kirchhof et al. (21) reported that apixaban therapy was comparable with vitamin K antagonists in patients undergoing atrial ﬁbrillation ablation at risk of stroke with respect to bleeding, stroke, acute brain MRI lesions, and cognitive function. The incidence of ACI was 27.2% (44 of 162 patients) for uninterrupted apixaban and 24.8% (40 of 161 patients) for uninterrupted vitamin K antagonists (p = 0.635) (21). Previous randomized studies reported post-ablation ACI incidence rates of 6.3% to 43.2% (Table 5). Studies involving uninterrupted anticoagulation showed similar ACI incidence rates of 15.7% to 20.0%.
Other safety profiles
The VENTURE AF study showed that rates of puncture site hematomas (nonmajor bleeding events) for rivaroxaban were 6.5% (8 of 123 patients) and 8.3% (10 of 121 patients) vitamin K antagonists (12). In another study, puncture site complication rates were reported to be 1.1% (2 of 188 patients), 1.1% (2 of 176 patients), and 3.1% (6 of 192 patients) for rivaroxaban, dabigatran, and vitamin K antagonists, respectively (22). In our study, thromboembolic and bleeding events were comparable between both groups. Additionally, the incidence of CA-associated puncture site hematomas was significantly lower in the rivaroxaban group than in the warfarin group, suggesting that, during AF ablation, perioperative anticoagulation management with uninterrupted rivaroxaban may be safer than that with uninterrupted warfarin.
Risk factors for ACI
Our results suggested that severe deep and subcortical white matter hyperintensities may be a patient-related factor for ACI and that the number of cardioversions may be a procedural factor for ACI. ACI may be a multifactorial CA complication. Furthermore, the following risk factors have been identified for ACI: age, spontaneous echo contrast, complex fractionated atrial electrograms ablation (4), low left ventricular ejection fraction, perioperative coronary artery imaging (23), RF ablation duration, preoperative MRI findings (numerous white matter lesions) (6), lower mean perioperative activated clotting time, perioperative electrical or pharmaceutical cardioversion (5), high incidence of arterial hypertension (24), aging (25), spontaneous echo contrast, and treatment duration before heparin injection (26).
Although some studies reported that cardioversion was significantly associated with ACI rates (5,11,26), others reported no such association (6,27,28). Here, both univariate and multivariate analyses showed that number of cardioversions was significantly associated with ACI rates. Procedure duration may (26,29) or may not (4,5,11,30) significantly affect ACI incidence. In this study, procedure duration tended to be longer in patients who developed ACI (mean: 233.5 min) than in those who did not (mean: 197.0 min), although the results were insignificant. Procedural difficulties related to patient conditions could affect cardioversion and procedure duration. However, we found no distinct association between cardioversion frequency and procedure duration (Pearson correlation coefficient: 0.349). Given that no heterogeneity in the physician’s performance of CA was observed between the 2 institutions, institutional factors for CA were not considered during statistical analysis.
Age may (4,20,25,31) or may not (3,5,8,19,26,29,30,32) serve as an ACI predictor. Whether hypertension may (11,20,23,24,33) or may not (3-5,8,11,19,25-32,34-36) predict ACI incidence remains unclear. However, systolic and diastolic hypertension increase the risk for symptomatic cerebral infarction (37). In this study, univariate analyses indicated that age, systolic blood pressure, and diastolic blood pressure significantly influenced ACI incidence. These factors may need further evaluation using larger datasets.
Cerebral white matter lesions are related to symptomatic stroke or symptomatic cerebral infarction recurrence (38,39). Kobayashi et al. (40) classified asymptomatic cerebral lesions showing high signals on MRI during subcortical silent brain infarctions (focal T2 hyperintensities of >3 mm with correlative T1 hypointensity), focal white matter T2 hyperintensities (similar to subcortical silent brain infarction but without correlative T2 hypointensity), and periventricular hyperintensities; they reported that subcortical silent brain infarction was the most significant risk factor for stroke. In this study, univariate analyses suggested that both periventricular hyperintensities and deep and subcortical white matter hyperintensities were related to CA-associated ACI, whereas multivariate analyses suggested that deep white matter lesions were more closely related to CA-associated ACI than periventricular hyperintensities.
Anti-Xa activities of rivaroxaban can be used as a coagulation index, although which may not be so clinically significant as PT-INR during warfarin use. With the anti-Xa activity level observed in this study, rivaroxaban could exert a rate of ACI occurrence similar to that of warfarin.
No signiﬁcant differences in left atrial appendage filling flow (LAAF) and left atrial appendage emptying flow (LAAE) was observed between patients who developed ACI and those who did not. However, LAAF and LAAE tended to be lower in patients those who developed ACI than in those who did not. Similar findings were reported by Ichiki et al. (23).
Data for the quality of warfarin therapy in relation to information on INR and time in the therapeutic range were unavailable in this study because of the short follow-up period.
Although the location of new lesions may have concerned tissue abnormalities underlying white matter hyperintensities, the association between the new ACI lesion location and hyperintense areas before CA could not be identified.
Multivariate analyses may have been underpowered because of sparse outcome events (n = 20); therefore, some significant factors could have been missed in the final model. A larger sample size would allow for the identification of significant predictive factors for ACI.
During CA for AF, no significant differences in ACI incidence were observed between the uninterrupted perioperative use of rivaroxaban and warfarin. No thromboembolic events developed, and no differences in major or nonmajor bleeding incidences were observed between both drugs.
COMPETENCY IN MEDICAL KNOWLEDGE: No significant differences between ACI rates with uninterrupted rivaroxaban and those with warfarin anticoagulation were observed during the perioperative period of catheter ablation for nonvalvular atrial fibrillation.
TRANSLATIONAL OUTLOOK: Further studies are needed to identify procedural factors, for example, catheter ablation or other ACI-associated conditions.
Supported by Bayer Yakuhin, Ltd., Japan. All authors have reported that they have no relationships with industry relevant to the contents of this paper to disclose.
All authors attest they are in compliance with human studies committees and animal welfare regulations of the authors’ institutions and Food and Drug Administration guidelines, including patient consent where appropriate. For more information, visit the JACC: Clinical Electrophysiology author instructions page.
- Abbreviations and Acronyms
- asymptomatic cerebral infarction
- atrial fibrillation
- catheter ablation
- cardiac magnetic resonance
- diffusion-weighted imaging
- fluid-attenuated inversion recovery
- prothrombin time-international normalized ratio
- Received February 1, 2018.
- Revision received August 8, 2018.
- Accepted August 13, 2018.
- 2018 American College of Cardiology Foundation
- Calkins H.,
- Hindricks G.,
- Cappato R.,
- et al.
- Cappato R.,
- Calkins H.,
- Chen S.A.,
- et al.
- Gaita F.,
- Caponi D.,
- Pianelli M.,
- et al.
- DeSimone C.V.,
- Madhavan M.,
- Ebrille E.,
- Rabinstein A.A.,
- Friedman P.A.,
- Asirvatham S.J.
- Di Biase L.,
- Gaita F.,
- Toso E.,
- et al.
- Muller P.,
- Halbfass P.,
- Szollosi A.,
- et al.
- Ichiki H.,
- Oketani N.,
- Ishida S.,
- et al.
- Sawada T.,
- Taneda J.,
- Okamoto K.,
- et al.
- ↵The Japan Brain Dock Society. Brain dock guideline 2008. Available at: http://jbds.jp/doc/guideline2008.pdf. Accessed December 6, 2017.
- Tanigawa T.,
- Kaneko M.,
- Hashizume K.,
- et al.
- Ogawa S.,
- Ikeda T.,
- Kitazono T.,
- et al.
- ↵Japanese Circulation Society. Guidelines for management of anticoagulant and antiplatelet therapy in cardiovascular disease. JCS. 2009. Available at: http://www.j-circ.or.jp/guideline/pdf/JCS2009_hori_d.pdf. Accessed December 6, 2017.
- Nakamura K.,
- Naito S.,
- Sasaki T.,
- et al.
- Yoshimura A.,
- Iriki Y.,
- Ichiki H.,
- et al.
- Kirchhof P.,
- Haeusler K.G.,
- Blank B.,
- et al.
- Rillig A.,
- Meyerfeldt U.,
- Tilz R.R.,
- et al.
- Herm J.,
- Fiebach J.B.,
- Koch L.,
- et al.
- Herrera Siklody C.,
- Deneke T.,
- Hocini M.,
- et al.
- Gaita F.,
- Leclercq J.F.,
- Schumacher B.,
- et al.
- Miyao S.,
- Takano A.,
- Teramoto J.,
- Takahashi A.
- van Swieten J.C.,
- Kappelle L.J.,
- Algra A.,
- van Latum J.C.,
- Koudstaal P.J.,
- van Gijn J.
- Kobayashi S.,
- Okada K.,
- Koide H.,
- Bokura H.,
- Yamaguchi S.