Comparison with Previous Studies
SCE is potentially associated with unfavorable neuropsychological outcomes. Previous studies demonstrated that the uninterrupted OAC minimized the risk of SCEs during AF ablation compared to interrupted OAC;4 expert consensus and guidelines first recommend the use of NOACs or VKA without interruption during the procedure.12 However, recent studies have reported that the incidence of MRI-detected SCEs still hovers at 2% to 27.2 % in patients on uninterrupted OAC for AF ablation.5,13-17 In the present study, the SCE incidence remains at 23.1% in total, and was almost similar to or even higher than the incidence in the previous studies.
Different protocols of intra-procedural UFH administration may affect the incidence of SCEs despite of the same ACT target (>300 sec) during the procedure. Kirchhof et al. have reported that the SCEs at postprocedural MRI were identified in 84 out of 323 patients on uninterrupted apixaban or VKA (26.0%); a bolus of UFH (100 IU/kg, based on body weight) was first administered prior to transseptal puncture, as in the present study.13 On the other hand, Di Biase et al. have demonstrated that the SCEs were detected in only 2 out of 86 patients on uninterrupted rivaroxaban (2.3%); the designated amount of UFH was first injected (10,000/8,000 IU in male/female, irrespective of body weight) before transseptal puncture.15 The amount of the first UFH shot seems to be higher in the previous study by De Biase et al. than in the present study (4000-6000IU, based on Japanese body weight), which likely shortens the time to reach the optimal ACT. UFH boluses were intermittently administered during the procedure in the previous and the present studies, however the transseptal sheaths were also continuously infused with UFH saline in the previous study by De Biase et al.,15 which may advantage keeping ACT levels constant and avoiding clot formation within the sheaths.
Different diagnostic criteria determined by brain MRI may also affect the SCE incidence. Guijian et al. reported that SCE incidence was 8% when using the definition of both abnormal diffusion-weighted image and fluid-attenuated inversion recovery (FLAIR) image, but when considering all abnormal diffusion-weighted images, the incidence rose to 27%.18 We diagnosed SCE using only diffusion-weighted images according to a neuroimaging expert’s recommendation, which may increase sensitivity.11
Previous studies have reported the risk factors for SCEs, including age, spontaneous echo contrast, complex fractionated atrial electrograms ablation, low left ventricular ejection fraction, perioperative coronary artery imaging, low mean perioperative ACT, electrical/pharmaceutical cardioversion, and arterial hypertension, etc.19-24However, these factors were identified in patients taking only VKA, but not NOACs. Kimura et al. demonstrated that the presence of deep and subcortical white matter hyper-intensity and the frequency of cardioversions were associated with SCEs in patients taking rivaroxaban or VKA without interruption during the procedure.17 In this study, LAD, baseline ACT before UFH injection, and time to reach the optimal ACT were significant risk factors for SCEs. To the best of our knowledge, we first demonstrated that intra-procedural ACT kinetics significantly affect the incidence of SCEs in patients on uninterrupted NOACs/VKA for AF ablation. Nevertheless, the information on this issue is still largely limited.
The ACT kinetics during the procedure are affected by the peri-procedural anticoagulation strategy.10Microthrombus can form immediately after a sheath/catheter is inserted into the body. Therefore, it is better for patients to achieve the optimally anticoagulated condition as quickly as possible. The present study demonstrated that different anticoagulants had different impacts on the intra-procedural ACT. Martin et al. demonstrated that at baseline, NOACs prolonged ACT differently; ACT was longer with dabigatran and shorter with apixaban,25 being consistent with our results. Dabigatran may have the advantage of decreasing the risk of SCE in terms of baseline ACT prolongation. Martin et al. also demonstrated that ACT prolongation in response to UFH was significantly smaller in rivaroxaban/apixaban than in dabigatran/VKA,25 which may underlie the longer time needed to reach the optimal ACT in rivaroxaban/apixaban in our study. Modifying the intraprocedural dosing of UFH may decrease the time to reach the optimal ACT and reduce the incidence of SCEs. However, the optimal ACT target (>300 sec) was validated in VKA-treated patients and may not be directly applied to those on NOACs; a specific regimen/monitoring of UFH in each NOAC may be required.
In the present study, apixaban and warfarin had the higher incidence of SCEs than the other anticoagulants although the difference did not reach statistical significance. Accumulating evidence has demonstrated a unique anticoagulation profile in each anticoagulant. In this study, the number of patients was relatively higher in dabigatran group than in the other groups likely because of the favorable outcome of dabigatran in RE-CIRCUIT study.26 Physicians may use and/or choose an anticoagulant in a patient with a specific clinical background based on the evidence; apixaban and warfarin tended to be prescribed in the more elderly patients who had higher thromboembolic risk in the era of uninterrupted NOACs/VKA.