Circulation Journal
Online ISSN : 1347-4820
Print ISSN : 1346-9843
ISSN-L : 1346-9843
Arrhythmia/Electrophysiology
Silent Cerebral Ischemic Lesions After Catheter Ablation of Atrial Fibrillation in Patients on 5 Types of Periprocedural Oral Anticoagulation – Predictors of Diffusion-Weighted Imaging-Positive Lesions and Follow-up Magnetic Resonance Imaging –
Kohki NakamuraShigeto NaitoTakehito SasakiKentaro MinamiYutaka TakeEri GotoSatoru ShimizuYoshiaki YamaguchiNaoko SuzukiToshiaki YanoMichiharu SengaKoji KumagaiKenichi KasenoNobusada FunabashiShigeru Oshima
Author information
JOURNAL FREE ACCESS FULL-TEXT HTML

2016 Volume 80 Issue 4 Pages 870-877

Details
Abstract

Background: The aim of this study was to identify the predictors of silent cerebral ischemic lesions (SCIL) after catheter ablation of atrial fibrillation (AF) and to determine whether SCIL develop into cerebral infarcts in patients with 5 types of oral anticoagulants (OAC).

Methods and Results: We retrospectively studied 286 consecutive patients (median, 67 years; 208 male; paroxysmal/persistent/long-standing persistent AF [LSP-AF], 147/90/49) who received periprocedural OAC and underwent MRI after the procedure. Warfarin (n=46) was continued, while dabigatran (n=47), rivaroxaban (n=89), apixaban (n=87), and edoxaban (n=17) were discontinued on the day of the procedure. I.v. heparin was infused to maintain an activated clotting time of 300–350 s during the procedure. Fifty-eight SCIL in 40 patients (14.0%) were identified on diffusion-weighted MRI. On multivariate logistic analysis, LSP-AF and dabigatran use were significant positive predictors of SCIL (OR, 2.912 and 2.287; P=0.006 and 0.042, respectively). Among 34 patients with 49 SCIL undergoing follow-up MRI, 45 (91.8%) of the lesions disappeared and 4 lesions developed into chronic cerebral infarcts. The SCIL with development into infarcts had a larger lesion diameter than those without (median, 6.55 mm vs. 4.2 mm; P=0.002).

Conclusions: LSP-AF and dabigatran use were independent risk factors for post-ablation SCIL in patients with uninterrupted warfarin and interrupted non-vitamin K antagonist OAC, but the majority of SCIL disappeared. (Circ J 2016; 80: 870–877)

Percutaneous catheter ablation has been widely used as a curative treatment option for drug-refractory atrial fibrillation (AF).1 It is, however, an invasive therapeutic procedure involving the cardiovascular system, and thus carries a risk for thromboembolic events.2 Adequate periprocedural anticoagulation with warfarin plays an important role in preventing thromboembolism in patients undergoing AF ablation.37 Non-vitamin K antagonist oral anticoagulants (NOAC) have been recently used as an alternative to warfarin for periprocedural anticoagulation management,8,9 and the safety and efficacy of both uninterrupted1014 and interrupted1517 NOAC treatment during AF ablation have been demonstrated. Nevertheless, silent cerebral emboli, which result in no clinically apparent neurologic deficits and are identified on brain magnetic resonance imaging (MRI), are relatively common, and were observed in 4.3–50% of patients with interrupted oral anticoagulants (OAC),1829 and in 0–18.2% of patients with uninterrupted OAC,11,30,31 who underwent either radiofrequency (RF) catheter or cryoballoon ablation of AF. Further, it has been reported that the majority of silent cerebral ischemic lesions (SCIL) identified on diffusion-weighted (DW) MRI (DWI) after AF ablation, that is, 72–100% of the lesions, were undetectable on follow-up MRI.22,23,26,27,32 Most of these previous studies with post-ablation MRI consisted of patients receiving the periprocedural OAC warfarin. The factors predicting SCIL and the development of the lesions with regard to the different types of OAC used during the periprocedural period of AF ablation have not yet been fully investigated. The purpose of this retrospective observational study was therefore to identify the predictors of SCIL detected on DWI after catheter ablation of AF among patients receiving the periprocedural OAC warfarin, dabigatran etexilate, rivaroxaban, apixaban, or edoxaban, and to determine whether post-ablation DWI-positive lesions develop into established cerebral infarcts on follow-up MRI.

Editorial p 814

Methods

Subjects

This study retrospectively enrolled 286 consecutive patients who received OAC prior to catheter ablation of AF and underwent MRI on the next day after the ablation procedure at Gunma Prefectural Cardiovascular Center from July 2014 to October 2015. OAC consisted of warfarin (group W; n=46), dabigatran etexilate (group D; n=47), rivaroxaban (group R; n=89), apixaban (group A; n=87), or edoxaban (group E; n=17). The exclusion criteria were age <20 years, any mechanical prosthetic heart valve, hemodialysis, pregnancy, allergy to heparin or history of heparin-induced thrombocytopenia, and absolute and relative contraindications to MRI including the presence of metallic implants and severe claustrophobia. All patients provided written informed consent for catheter ablation of AF and MRI.

All patients underwent transthoracic echocardiography to assess atrial and ventricular size, left ventricular function, and presence of structural heart disease using standard parasternal and apical views, and transesophageal echocardiography to rule out any intracardiac thrombi on the day of the procedure or the day before.

Periprocedural Anticoagulation

The warfarin dosage was adjusted to achieve a therapeutic international normalized ratio (INR), which was 1.6–2.6 in patients aged ≥70 and 2.0–3.0 in patients aged <70.33 Warfarin was given in the evening and was continued at the adjusted doses throughout the periprocedural period. Although 22 patients (47.8%) had subtherapeutic INR on the day before the procedure, those patients were included in the analysis. The NOAC used in this study, that is, dabigatran at a dose of 110 or 150 mg twice daily, rivaroxaban at 10 or 15 mg once daily in the morning, apixaban at 2.5 or 5 mg twice daily, and edoxaban at 30 or 60 mg once daily in the morning, were administered until the day before the procedure, discontinued on the day of the procedure, and restarted on the next morning after the procedure. In the NOAC groups, 113 patients (47.1%) received lower NOAC doses: 31, 36, 30, and 16 patients in groups D, R, A, and E, respectively. Neither unfractionated heparin nor low-molecular-weight heparin was used during the interruption of the NOAC before the procedure.

Immediately after the femoral venous puncture during the ablation procedure, an initial heparin bolus of 5,000 units in group W and 10,000 units in the NOAC groups was given i.v., followed by a continuous infusion (500–1,000 units/h) and additional bolus (1,000–5,000 units) of heparin to maintain the activated clotting time (ACT) between 300 and 350 s. The ACT was measured every 10 min until ACT reached 300 s and thereafter every 30 min. The i.v. heparin was stopped at the end of the procedure and protamine was given to partially reverse the anticoagulation effect of heparin at the operator’s discretion. After achieving hemostasis of the femoral venous puncture sites and verifying the absence of any pericardial effusions on echocardiography, all patients received a continuous infusion of heparin at a dose of 10,000 units per 24 h until the next morning after the procedure.

Catheter Ablation

All patients received deep conscious sedation throughout the procedure using propofol or dexmedetomidine, and pentazocine. Pulmonary vein (PV) isolation (PVI) was performed using either RF or cryothermal energy, and bidirectional conduction block between the left atrium (LA) and PV was verified in all the PV during sinus rhythm. Linear ablation along the cavotricuspid isthmus was performed using a 3.5-mm-tip externally irrigated ablation catheter or an 8-mm-tip non-irrigated ablation catheter in all patients, and a superior vena cava (SVC) isolation was additionally performed if any triggers of AF originating from the SVC were documented.

RF Ablation

A multi-electrode catheter was positioned in the coronary sinus (CS). After the transseptal puncture, a 3.5-mm-tip externally irrigated ablation catheter (ThermocooL® SmarttouchTM, Biosense Webster, Diamond Bar, CA, USA or FlexAbilityTM or CoolFlexTM, St. Jude Medical, St. Paul, MN, USA) and two 7-Fr decapolar circular mapping catheters were introduced into the LA through three 8-Fr sheaths (SL0, St. Jude Medical). The circumferential RF lesions extensively encircling the ipsilateral superior and inferior PV were created by a point-by-point ablation, guided by a double Lasso technique and 3-D computed tomography or MRI reconstruction integrated into an electroanatomic mapping system (CARTO® 3, Biosense Webster or EnSite NavXTM, St. Jude Medical). Each application of RF energy was delivered for 40–60 s with a maximum temperature of 42℃. The RF power output was limited to 35 W along the right PVI line and 40 W along the anterior PVI line of the left PV. Along the posterior line of the left PV, the RF power output and duration were limited to 30 W and 30 s, respectively. The esophageal temperature was monitored using an esophageal temperature probe (SensiTherm, St. Jude Medical or Esophastar, Japan Lifeline, Tokyo, Japan) and an esophageal temperature limit of 41℃ was defined to interrupt RF energy delivery. Thirty-five patients (12.2%) underwent additional substrate modification for the AF including continuous fractionated atrial electrogram ablation and/or linear ablation in the atria after PVI. Electrical cardioversion restored sinus rhythm if the AF did not terminate after these procedures.

Cryoablation

Multi-electrode catheters were positioned in the CS and right ventricle. Electrical cardioversion restored sinus rhythm if AF was present at the beginning of the ablation procedure. After the transseptal puncture, a 15-Fr steerable sheath (FlexCath Advance® Steerable Sheath, Medtronic, Minneapolis, MN, USA) was advanced into the LA. The steerable sheath was continuously infused with heparinized saline (4 units/ml) at a rate of 20 ml/h. A 23- or 28-mm second-generation cryoballoon catheter (Arctic Front Advance® Cardiac ablation catheter, Medtronic) was introduced through the sheath into the LA using a 15- or 20-mm diameter circular mapping catheter (Achieve® Mapping catheter, Medtronic) as a guide, which was inserted via the central lumen of the cryoballoon catheter. Optimal PV occlusion by wedging the cryoballoon was verified by injecting contrast medium via the lumen of the cryoballoon catheter. Cryoenergy was then delivered with a maximum duration of 180 ms for each application. If PVI could not be completed with multiple cryoenergy applications using the cryoballoon, touch-up ablation for the remaining gaps was performed using an 8-mm-tip cryoablation catheter (Freezor® Max Cardiac Cryoablation Catheter, Medtronic) or 3.5-mm-tip externally irrigated RF ablation catheter. To detect any right phrenic nerve injury during the cryoablation of the right-sided PV, the right phrenic nerve was constantly paced from the SVC and the fluoroscopic diaphragm movement during spontaneous breathing was monitored.

Brain MRI

All patients underwent brain MRI with a 1.5-T MR unit (Achieva Nova Dual, Philips Medical Systems, Einthoven, Netherlands) on the next day after the procedure. The imaging protocol included the following 4 pulse sequences, which were acquired in the axial plane through the anterior and posterior cerebral commissures: (1) DW sequence with a single-shot echo-planar imaging technique (repetition time/echo time [TR/TE], 3,000/78 ms; matrix, 128×128; field of view [FOV], 240×240 mm; b values, 0 and 1,000s/mm2; slice thickness/gap, 5.0/1.0 mm; no. signals averaged [NSA], 2); (2) fluid-attenuated inversion recovery (FLAIR) sequence (TR/TE, 8,000/100 ms; inversion time, 2,300 ms; matrix, 288×157; FOV, 240×190 mm; slice thickness/gap, 5.0/1.0 mm; NSA, 2); (3) T2-weighted spin echo sequence (TR/TE, 4,832/100 ms; matrix, 336×196; FOV, 240×190 mm; slice thickness/gap, 5.0/1.0 mm; NSA, 2); and (4) T1-weighted spin echo sequence (TR/TE, 450/12 ms; matrix, 288×159; FOV, 240×190 mm; slice thickness/gap, 5.0/1.0 mm; NSA, 2).

The DW sequence is the most sensitive tool for the early detection of cerebral ischemia and is capable of detecting subtle changes within minutes of an ischemic event.34,35 The apparent diffusion coefficient (ADC) map is calculated using the DW datasets with the 2 different b-values applying linear regression methods, and rules out a T2 shine-through artifact that can be falsely confounded with restricted diffusion. The ADC of a hyperintense lesion identified on DWI is reduced. Therefore, in the present study, a new cerebral ischemic lesion associated with the ablation was defined as a hyperintense lesion on DWI, corresponding to a reduced ADC map. The number of new ischemic lesions for each patient was counted, and the size and anatomic location of the lesions were analyzed.

The follow-up MRI protocol included the same DW, FLAIR, and T1- and T2-weighted spin-echo sequences as obtained at the initial MRI. Development of the DWI-positive lesions into established cerebral infarcts was determined based on the presence of a corresponding FLAIR hyperintensity, and, in contrast, a complete reversal of the lesions was determined based on the lack of a corresponding FLAIR hyperintensity. All MRI were reviewed by experienced radiologists blinded to the patient characteristics.

Periprocedural Examinations and Follow-up

Neurological assessment was performed on the day before and after the ablation procedure. The incidence of symptomatic thromboembolic events, including an ischemic stroke, transient ischemic attack, and systemic embolus, was evaluated. Major bleeding complications were defined as cardiac tamponade or pericardial effusion requiring drainage, intracranial and gastrointestinal hemorrhage, hemothorax, retroperitoneal bleeding, any bleeding requiring blood transfusion, and pseudoaneurysm, arteriovenous fistula, or hematoma requiring any intervention. Minor bleeding complications were defined as pericardial effusion, pseudoaneurysm, arteriovenous fistula, or hematoma not requiring drainage or intervention and rebleeding at the venous puncture sites.

All patients were followed up in the outpatient clinic every 1–2 months. OAC were continued for at least 3 months, and thereafter the decision to discontinue the OAC was left to the discretion of the physician in charge, considering the risk of thromboembolic events, that is, CHA2DS2-VASc score (congestive heart failure, hypertension, diabetes mellitus, vascular disease [myocardial infarction, aortic plaque, and peripheral vascular disease], age of 65–74 years, female sex [1 point for presence of each], age ≥75 years, and stroke/transient ischemic attack [2 points]).

Statistical Analysis

Normally distributed continuous variables are expressed as mean±SD and non-normally distributed continuous variables as median (IQR). Comparative analysis among the 5 OAC was done using 1-way analysis of variance with a Tukey’s post-hoc analysis or Kruskal-Wallis test with a post-hoc Mann-Whitney U-test corrected for multiple comparisons using the Bonferroni method. Categorical variables are expressed as the absolute value and percentage, and comparative analysis was done using chi-squared test. Student’s t-test or Mann-Whitney U-test were used to compare differences between patients with and without SCIL. Multivariate logistic regression modeling was carried out to identify significant predictors of SCIL using potential confounders with clinical significance or significant association on univariate analysis. P<0.05 was considered statistically significant. Statistical analysis was carried out using SPSS Statistics 22 (SPSS, Chicago, IL, USA).

Results

Subject demographics and baseline characteristics are listed in Table 1. The median age was 67 years and the proportion of men was 72.7%. Groups W, D, R, A, and E consisted of 46, 47, 89, 87, and 17 patients, respectively. Group D had a significantly lower D-dimer level (P<0.01), group R had a lower percentage of structural heart diseases and CHADS2 score (congestive heart failure, hypertension, age ≥75 years, diabetes mellitus [1 point for presence of each], and stroke/transient ischemic attack [2 points]) (P<0.05), and group A had a lower CHADS2 score (P<0.05) than group W. More patients in group E underwent cryoballoon ablation than those in the other groups (Table 2). Regarding intraprocedural anticoagulation, group W had a significantly smaller heparin dosage for the ablation and higher ACT than the others (P<0.05).

Table 1. Subject Baseline Characteristics vs. OAC Type
  All
(n=286)
Warfarin
(n=46)
Dabigatran
(n=47)
Rivaroxaban
(n=89)
Apixaban
(n=87)
Edoxaban
(n=17)
P-value
Age (years) 67 (60–72) 69 (63–75) 66 (57–70) 65 (60–72) 67 (61–71) 68 (60.5–75) 0.094
Male gender 208 (72.7) 32 (69.6) 41 (87.2) 65 (73.0) 57 (65.5) 13 (76.5) 0.106
Body weight (kg) 65.1±11.8 61.8±10.8 67.2±11.4 66.1±11.9 65.2±12.7 62.5±9.7 0.159
Type of AF             0.432
 Paroxysmal 147 (51.4) 17 (37.0) 27 (57.4) 51 (57.3) 44 (50.6) 8 (47.1)  
 Persistent 90 (31.5) 18 (39.1) 11 (23.4) 24 (27.0) 30 (34.5) 7 (41.2)  
 Long-standing persistent 49 (17.1) 11 (23.9) 9 (19.1) 14 (15.7) 13 (14.9) 2 (11.8)  
Structural heart disease 36 (12.6) 11 (23.9) 8 (17.0) 7 (7.9)* 10 (11.5) 0 (0) 0.032
Hypertension 172 (60.1) 35 (76.1) 26 (55.3) 53 (59.6) 49 (56.3) 9 (52.9) 0.182
Diabetes mellitus 46 (16.1) 12 (26.1) 6 (12.8) 14 (15.7) 13 (14.9) 1 (5.9) 0.268
Heart failure 28 (9.8) 7 (15.2) 5 (10.6) 4 (4.5) 11 (12.6) 1 (5.9) 0.198
Stroke/TIA 24 (8.4) 6 (13.0) 7 (14.9) 4 (4.5) 5 (5.7) 2 (11.8) 0.117
CHADS2 score 1 (0–2) 1.5 (1–2) 1 (0–2) 1 (0–2)* 1 (0–2)* 1 (0–2) 0.017
 0 86 (30.1) 6 (13.0) 16 (34.0) 27 (30.3) 31 (35.6) 6 (35.3)  
 1 107 (37.4) 17 (37.0) 16 (34.0) 37 (41.6) 32 (36.8) 5 (29.4)  
 2 63 (22.0) 14 (30.4) 7 (14.9) 23 (25.8) 15 (17.2) 4 (23.5)  
 >2 30 (10.5) 9 (19.6) 8 (17.0) 2 (2.2) 9 (10.3) 2 (11.8)  
CHA2DS2-VASc score 2 (1–3) 3 (1–4) 2 (1–3) 2 (1–3) 2 (1–3) 2 (1–3) 0.056
HAS-BLED score 1 (1–2) 2 (1–2) 1 (1–2) 1 (1–2) 1 (1–2) 1 (0.5–2.5) 0.194
D-dimer (μg/ml) 0.3 (0.2–0.4) 0.4 (0.3–0.5) 0.3 (0.2–0.3)** 0.3 (0.2–0.5) 0.3 (0.2–0.4) 0.4 (0.3–0.45) 0.008
BNP (pg/ml) 73.5
(34.3–146.3)
85.2
(46.0–147.8)
72.3
(34.4–146.0)
58.7
(31.4–117.5)
66.8
(30.4–210.0)
75.5
(52.8–171.0)
0.294
Use of antiplatelet drugs 31 (10.8) 7 (15.2) 4 (8.5) 8 (9.0) 10 (11.5) 2 (11.8) 0.799
LAD (mm) 42 (37–45) 43 (39–47) 42 (39–45) 40 (36–44.5) 42 (37–46) 41 (37.5–44) 0.245
LVEF (%) 60 (55–65) 60 (50–66) 60 (55–65) 65 (60–65) 60 (55–65) 60 (55–65) 0.709
LAA flow velocity (cm/s) 51 (28–78) 40.5 (23–62.5) 56 (26–81) 53.5 (34–80) 49 (28–85) 36 (23–81) 0.227
Spontaneous echo contrast 75 (26.2) 15 (32.6) 12 (25.5) 20 (22.5) 24 (27.6) 4 (23.5) 0.777

Data given as mean±SD, median (IQR) or n (%). **P<0.01, *P<0.05 vs. warfarin. AF, atrial fibrillation; BNP, B-type natriuretic peptide; CHA2DS2-VASc score, congestive heart failure, hypertension, diabetes mellitus, vascular disease (myocardial infarction, aortic plaque, and peripheral vascular disease), age of 65–74 years, female sex (1 point for presence of each), age ≥75 years, and stroke/transient ischemic attack (2 points); HAS-BLED score, hypertension (systolic blood pressure >160mmHg), abnormal renal and liver function, stroke, bleeding tendency/predisposition, labile international normalized ratios (if on warfarin), age >65 years, drugs (antiplatelets or nonsteroidal anti-inflammatory drugs) or excess alcohol (1 point for presence of each); LAA, left atrial appendage; LAD, left atrial diameter; LVEF, left ventricular ejection fraction; OAC, oral anticoagulant; TIA, transient ischemic attack.

Table 2. Procedure-Related Parameters vs. OAC Type
  All
(n=286)
Warfarin
(n=46)
Dabigatran
(n=47)
Rivaroxaban
(n=89)
Apixaban
(n=87)
Edoxaban
(n=17)
P-value
Total procedure time
(min)
141
(122–162)
132.5
(109–161)
142
(125–162)
146
(126.5–165.5)
134
(120–162)
146
(120–164)
0.611
Type of LA procedure             0.068
 PVI alone 251 (87.8) 35 (76.1) 42 (89.4) 80 (89.9) 77 (88.5) 17 (100)  
 PVI and substrate
modification
35 (12.2) 11 (23.9) 5 (10.6) 9 (10.1) 10 (11.5) 0 (0)  
LA ablation energy source             <0.001
 RF 253 (88.5) 41 (89.1) 44 (93.6) 80 (89.9) 79 (90.8) 9 (52.9)*,††,‡‡,§§  
 Cryothermal 33 (11.5) 5 (10.9) 3 (6.4) 9 (10.1) 8 (9.2) 8 (47.1)*,††,‡‡,§§  
Total heparin dosage
(units)
13,180
(11,367.5–
15,555)
8,815
(6,232.5–
10,592.5)
13,450
(11,900–
151,10)**
15,100
(12,535–
16,785)**
13,750
(11,800–
15,720)**
12,800
(12,090–
14,925)**
<0.001
Mean ACT during the
procedure (s)
315
(303–334)
336
(319–351)
314
(306–332)**
307
(297–322)**
314
(302.5–331.5)**
316
(306.5–334)
<0.001
Maximum ACT during
the procedure (s)
337.5
(323–361)
367.5
(345.5–406)
342
(319–360)**
329
(316–349.5)**
333
(323–354)**
341
(329–350)*
<0.001
Electrical cardioversion
during the procedure
151 (52.8) 24 (52.2) 25 (53.2) 50 (56.2) 41 (47.1) 11 (64.7) 0.643
Exchanging catheters
over the transseptal
sheaths
54 (18.9) 7 (15.2) 9 (19.1) 18 (20.2) 17 (19.5) 3 (17.6) 0.968

Data given as mean±SD, median (IQR) or n (%). *P<0.05, **P<0.01, vs. warfarin, ††vs. dabigatran, ‡‡vs. rivaroxaban, §§vs. apixaban. ACT, activated clotting time; LA, left atrium; OAC, oral anticoagulant; PVI, pulmonary vein isolation; RF, radiofrequency.

No symptomatic thromboembolic events or deaths occurred in this series. The incidence of major and minor bleeding complications did not significantly differ among the 5 groups (Table 3).

Table 3. Events Associated With AF Ablation vs. OAC Type
  All
(n=286)
Warfarin
(n=46)
Dabigatran
(n=47)
Rivaroxaban
(n=89)
Apixaban
(n=87)
Edoxaban
(n=17)
P-value
Symptomatic stroke/TIA/systemic
embolus
0 0 0 0 0 0  
SCIL 40 (14.0) 4 (8.7) 11 (23.4) 8 (9.0) 14 (16.1) 3 (17.6) 0.142
Major complications 2 (0.7) 0 0 1 (1.1) 0 1 (5.9) 0.169
 Cardiac tamponade 1 (0.3) 0 0 0 0 1 (5.9)  
 Retroperitoneal bleeding 1 (0.3) 0 0 1 (1.1) 0 0  
Minor complications without any
intervention
16 (5.6) 2 (4.3) 3 (6.4) 6 (6.7) 5 (5.7) 0 0.453
 Pericardial effusion not requiring
drainage
2 (0.7) 0 1 (2.1) 1 (1.1) 0 0  
 Groin hematoma 2 (0.7) 0 0 2 (2.2) 0 0  
 Rebleeding at the venous puncture
sites
12 (4.2) 2 (4.3) 2 (4.3) 3 (3.4) 5 (5.7) 0  
Total bleeding complications 18 (6.3) 2 (4.3) 3 (6.4) 7 (7.9) 5 (5.7) 1 (5.9) 0.451

Data given as n (%). SCIL, silent cerebral ischemic lesion. Other abbreviations as in Table 1.

Predictors of SCIL

Initial MRI was performed 22.5 h (IQR, 20–27 h) after the ablation procedure. Forty patients (14.0%) had a total of 58 SCIL detected on DWI: 23 patients with 1 SCIL, 16 with 2 SCIL, and 1 with 3 SCIL. Groups W, D, R, A, and E included 4 (8.7%), 11 (23.4%), 8 (9.0%), 14 (16.1%), and 3 patients (17.6%) with SCIL, respectively (Table 3).

The baseline characteristics and procedure-related parameters during the ablation were compared between the patients with and without SCIL (Table 4). The percentage of the AF types (paroxysmal, persistent, or long-standing persistent AF [LSP-AF]) significantly differed between the patients with and without SCIL (P=0.014). On multivariate logistic regression analysis using the types of AF, OAC, and procedure in the LA, and total procedure time, the LSP-AF and use of dabigatran were significant positive predictors of SCIL (OR, 2.912 and 2.287; 95% CI: 1.359–6.240 and 1.030–5.074; P=0.006 and 0.042, respectively; Table 5). In group W, all patients with SCIL had a therapeutic INR on the day before the procedure. In the NOAC groups, 20 (55.6%) of the 36 SCIL-positive patients received lower dose of NOAC: 6, 3, 8, and 3 patients in groups D, R, A, and E, respectively.

Table 4. Baseline Characteristics and Procedure Parameters vs. Presence of SCIL
  Patients with SCIL
(n=40)
Patients without SCIL
(n=246)
P-value
Age (years) 68 (61–72) 67 (60–72) 0.509
Male gender 28 (70.0) 180 (73.2) 0.676
Body weight (kg) 63.5±11.3 65.4±11.9 0.363
Type of AF     0.014
 Paroxysmal 19 (47.5) 128 (52.0)  
 Persistent 8 (20.0) 82 (33.3)  
 Long-standing persistent 13 (32.5) 36 (14.6)  
Structural heart disease 3 (7.5) 33 (13.4) 0.296
Hypertension 22 (55.0) 150 (61.0) 0.474
Diabetes mellitus 6 (15.0) 40 (16.3) 0.841
Heart failure 4 (10.0) 24 (9.8) 0.572
Stroke/TIA 4 (10.0) 20 (8.1) 0.440
CHADS2 score 1 (0–2) 1 (0–2) 0.914
CHA2DS2-VASc score 2 (1–3) 2 (1–3) 0.639
Type of OAC     0.142
 Warfarin 4 (10.0) 42 (17.1)  
 Dabigatran 11 (27.5) 36 (14.6)  
 Rivaroxaban 8 (20.0) 81 (32.9)  
 Apixaban 14 (35.0) 73 (29.7)  
 Edoxaban 3 (7.5) 14 (5.7)  
>3 weeks OAC before procedure 35 (87.5) 218 (88.6) 0.504
Antiplatelet drugs 4 (10.0) 27 (11.0) 0.557
D-dimer (μg/ml) 0.3 (0.2–0.5) 0.3 (0.2–0.4) 0.500
BNP (pg/ml) 101.0 (44.8–179.3) 69.8 (33.0–141.5) 0.145
LAD (mm) 43 (37.5–48) 41 (37–45) 0.268
LVEF (%) 65 (55–65) 60 (55–65) 0.954
LAA flow velocity (cm/s) 47.5 (27–77) 52 (28–78) 0.737
Spontaneous echo contrast 13 (32.5) 62 (25.2) 0.331
Total procedure time (min) 153 (123.5–168) 138 (121–159) 0.082
Type of LA procedure     0.033
 PVI alone 31 (77.5) 220 (89.4)  
 PVI and substrate modification 9 (22.5) 26 (10.6)  
Type of LA ablation energy source     0.460
 RF 34 (85.0) 219 (89.0)  
 Cryothermal 6 (15.0) 27 (11.0)  
Total heparin dosage (units) 13,995 (12,100–15,587.5) 12,905 (11,235–15,517.5) 0.177
Mean ACT during procedure (s) 317 (304–336) 315 (303–333) 0.604
Maximum ACT during procedure (s) 336 (322–371) 338 (323–361) 0.931
Electrical cardioversion during procedure 25 (62.5) 126 (51.2) 0.185
Exchanging catheters over the transseptal sheaths 10 (25.0) 44 (17.9) 0.286

Data given as median (IQR) or n (%). Abbreviations as in Tables 1–3.

Table 5. Multivariate Predictors of SCIL
  OR 95% CI P-value
Type of AF
 Paroxysmal     Reference
 Persistent     0.298
 Long-standing persistent 2.912 1.359–6.240 0.006
Type of OAC
 Warfarin     Reference
 Dabigatran 2.287 1.030–5.074 0.042
 Rivaroxaban     0.324
 Apixaban     0.090
 Edoxaban     0.365
Type of LA procedure
 PVI alone     Reference
 PVI and substrate modification     0.244
Total procedure time (min)     0.314

Abbreviations as in Tables 1–3.

Characteristics of SCIL

Groups W, D, R, A, and E had 7, 15, 15, 17, and 4 SCIL, respectively, and the maximum diameter and distribution of the lesions are given in Table 6. The maximum lesion diameter was similar among the 5 groups (P=0.586). Thirty-four (58.6%) of the lesions were located in the right hemisphere and 24 (42.4%) were in the left hemisphere, and the predominant location of the lesions was the frontal lobe, cerebellum, and parietal lobe.

Table 6. SCIL Maximum Diameter and Anatomic Location vs. OAC Type
  All
(n=58)
Warfarin
(n=7)
Dabigatran
(n=15)
Rivaroxaban
(n=15)
Apixaban
(n=17)
Edoxaban
(n=4)
P-value
Maximum diameter (mm) 4.3 (3.7–5.2) 4.3 (4.0–5.7) 4.0 (3.5–5.0) 4.5 (3.6–5.2) 4.3 (4.05–5.2) 5.45 (3.65–6.6) 0.586
Anatomic location
 Frontal lobe 21 (36.2) 3 (42.9) 6 (40.0) 7 (46.7) 3 (17.6) 2 (50.0)  
 Temporal lobe 1 (1.7) 0 0 0 0 1 (25.0)  
 Parietal lobe 12 (20.7) 1 (14.3) 1 (6.7) 3 (20.0) 7 (41.2) 0  
 Occipital lobe 6 (10.3) 1 (14.3) 3 (20.0) 1 (6.7) 1 (5.9) 0  
 Cerebellum 13 (22.4) 2 (28.6) 5 (33.3) 3 (20.0) 3 (17.6) 0  
 Radiate crown 3 (5.2) 0 0 1 (6.7) 1 (5.9) 1 (25.0)  
 Corpus callosum 1 (1.7) 0 0 0 1 (5.9) 0  
 Basal ganglia 1 (1.7) 0 0 0 1 (5.9) 0  

Data given as median (IQR) or n (%). Abbreviations as in Tables 2,3.

Follow-up MRI

Follow-up MRI was available in 34 (85.0%) out of the 40 patients with SCIL on initial MRI, and was performed 124 days (IQR, 47–191 days) after the ablation procedure. Among 49 SCIL (group W, n=5; group D, n=14; group R, n=10; group A, n=16; group E, n=4) in the 34 patients, 45 (91.8%) of the lesions disappeared on follow-up MRI. Groups R, A, and E included 1, 1, and 2 lesions with development into chronic cerebral infarcts, respectively, while no lesions in Groups W and D developed into chronic infarcts. SCIL that developed into chronic infarcts had a significantly larger lesion diameter on initial DWI than those without (6.55 mm; IQR, 5.4–8.6 mm vs. 4.2 mm; IQR, 3.6–4.7 mm, P=0.002).

Discussion

The present study has identified the predictors of DWI SCIL after catheter ablation of AF in patients receiving periprocedural OAC (uninterrupted warfarin and interrupted NOAC). LSP-AF and dabigatran use were independent risk factors for SCIL after AF ablation. Further, on follow-up MRI in the SCIL-positive patients 91.8% of the lesions disappeared, while 8.2% of the lesions developed into chronic cerebral infarcts under continuous OAC following the ablation.

Predictors of SCIL

The occurrence of thromboembolisms during the periprocedural period of AF ablation is associated with various factors: pre-existing microthrombi loosened by catheter manipulation, which were undetectable on pre-ablation transesophageal echocardiography; clot or char formation by RF-induced thermal denaturization of blood proteins; intraprocedural thrombus formation at the catheter, guidewire, and sheath; microbubble generation by introduction of catheters and contrast angiocardiography; post-procedural thrombus formation due to endothelial denudation, and atrial stunning after restoration of sinus rhythm.20,36 Previous studies with post-ablation MRI have reported some predictors of SCIL after AF ablation, such as older age,23,30,37 non-paroxysmal AF,29 higher CHA2DS2-VASc score,31 presence of spontaneous echo contrast,30 lower ACT,20,25,30 extensive LA ablation,30 cardioversion,20,25 and exchanging catheters over a transseptal sheath37 during the procedure. All these studies consisted of patients with uninterrupted or interrupted warfarin during the periprocedural period. In a prospective multicenter study describing the safety and feasibility of periprocedural uninterrupted apixaban,11 all DWI were negative for new silent cerebral ischemia in 29 patients undergoing RF ablation. The present study included a total of 240 patients (83.9%) receiving periprocedural OAC consisting of the 4 types of NOAC in addition to 46 patients with warfarin. Thereby, it was shown that OAC type was independently associated with risk of post-ablation SCIL.

One possible reason for dabigatran use being an independent predictor of SCIL is the bias in the sample selection process. The present study was not randomized, and differences in baseline characteristics were seen between the groups.

Another possible explanation is that warfarin was continued, while the 4 NOAC were all discontinued on the day of the procedure. This OAC protocol may have more greatly facilitated periprocedural microthrombus formation in the dabigatran group, compared with the warfarin group. In a study by Di Biase et al, uninterrupted therapeutic warfarin reduced the risk of SCIL, compared with interrupted warfarin bridged with low-molecular-weight heparin.29 Thus, periprocedural continuation of the NOAC may have further reduced the incidence of SCIL in the present NOAC groups.

Nevertheless, the reason why dabigatran was the only predictor of SCIL among the NOAC remains unclear. This may be due to the different mechanisms of the anticoagulant action: direct thrombin inhibition or factor Xa inhibition. Some preclinical studies showed that direct thrombin inhibitors enhanced the thrombin generation and hypercoagulability at low dose, whereas factor Xa inhibitors did not exhibit such a phenomenon.38,39 Whether these differences in the activation of coagulation according to the type of NOAC contribute to the clinically significant differences in the occurrence of SCIL requires further investigation in clinical studies.

In this series, the patients with LSP-AF had a significantly higher risk of SCIL, compared with those with paroxysmal AF. This is in line with the Di Biase et al study, in which only warfarin was used for periprocedural anticoagulation management.29

Follow-up MRI of SCIL

The majority of SCIL initially detected on DWI were undetectable on follow-up MRI, which was consistent with previous studies.22,23,26,27,32 Acute cerebral ischemic lesions can be occasionally reversible if the DWI hyperintensities reflect less severe hypoperfusion and cytotoxic edema.40,41 Deneke et al showed that post-ablation DWI-positive lesions with a maximum diameter ≤10 mm had a tendency to completely reverse, while lesions with a diameter >10 mm developed into permanent tissue injury.22 No SCIL had a maximum lesion diameter >10 mm in the present study, but the SCIL that developed into chronic infarcts had a larger lesion diameter on DWI than those without (median, 6.55 mm vs. 4.2 mm; P=0.002).

Study Limitations

The present study had some limitations. First, these data were derived from a single center. Second, this study was retrospectively designed, which introduced a sample selection bias for each OAC. Thus, it would be premature to conclude that dabigatran was an independent risk factor for SCIL. In addition, the reliability of the multivariate regression analysis was hampered by the sample size and number of patients with SCIL. The sample size was relatively small in each group, especially in the edoxaban group. Therefore, further understanding of the impact of OAC on the occurrence of post-ablation SCIL will require investigation in a prospective randomized study with a larger number of patients in each group. Finally, pre-ablation MRI was not performed to evaluate pre-existing cerebral ischemic lesions. Thus, the SCIL may partly represent ischemic events that occurred before the ablation. Most SCIL, however, had no obvious hyperintensity on the FLAIR and T2-weighted spin-echo sequences during initial MRI, suggesting that the occurrence of SCIL was likely to be associated with the ablation.

Conclusions

LSP-AF and dabigatran use may be independent risk factors for SCIL after catheter ablation of AF among patients receiving periprocedural OAC involving uninterrupted warfarin and interrupted dabigatran, rivaroxaban, apixaban, and edoxaban. On follow-up MRI, 91.8% of the lesions disappeared and few SCIL developed into chronic cerebral infarcts on continuous OAC following the ablation.

Disclosure

The authors declare no conflicts of interest.

References
 
© 2016 THE JAPANESE CIRCULATION SOCIETY
feedback
Top