Apixaban is increasingly used for stroke prevention in patients with atrial fibrillation. Data about the safety of left atrial radiofrequency ablation procedures under continuous apixaban therapy are lacking. We performed a matched-cohort study of patients undergoing left atrium ablation procedures for atrial fibrillation or left atrial flutter. For each patient on apixaban, 2 patients on phenprocoumon were matched by age, gender, and type of arrhythmia. The primary safety end point was a composite of bleeding, thromboembolic events, and death. We identified 105 consecutive patients (35 women; mean age 63 years) on apixaban and matched 210 phenprocoumon patients (70 women, mean age 64 years). The primary end point was met in 11 patients of the apixaban group and 26 patients of the phenprocoumon group (10.5% vs 12.3%, p = 0.71). Major bleeding complications occurred in 1 patient of the apixaban group and 1 patient of the phenprocoumon group (1% vs 0.5%, p >0.99). Minor bleeding complications were observed in 10 patients of the apixaban group and 25 patients of the phenprocoumon group (9.5% vs 11.9%, p = 0.61). No patient in either group experienced a thromboembolic event and no patient died. In patients on apixaban, no clinical variable was predictive for bleeding complications. Left atrial ablation procedures under continuous oral anticoagulation with apixaban are feasible and as safe as under continuous oral anticoagulation with phenprocoumon.
Apixaban, an oral factor X inhibitor, is the first novel oral anticoagulant (NOAC) that showed superiority over warfarin with respect to a combined bleeding and thromboembolic end point. Furthermore, apixaban levels are less dependent on renal function than those of dabigatran and rivaroxaban. For patients with symptomatic atrial fibrillation (AF), rhythm control (achieving and maintaining sinus rhythm) is the main therapeutic goal and radiofrequency catheter ablation has emerged as standard therapy. During ablation procedures, the periprocedural bleeding risk has to be weighed against the risk of thromboembolism and stroke, in particular given the high thrombogenicity of the ablation lesions. Previous studies have demonstrated that oral anticoagulation with vitamin K antagonists should be continued periprocedurally. There is only limited data for patients taking NOACs, and data regarding apixaban are lacking. Using a single-center clinical registry, we investigated the safety of continuous periprocedural apixaban during left atrial ablation procedures and compared the safety profiles of apixaban and phenprocoumon in a matched-cohort study.
Methods
Patients were derived from our clinical database for patients undergoing catheter ablation at the German Heart Center Munich. Patients were included in the apixaban group if they (1) underwent ablation in the left atrium and (2) were on apixaban therapy during the ablation procedure, with apixaban started at least 4 weeks before ablation. Patients were eligible until March 7, 2014. Patients of the phenprocoumon group were derived from the same database and were on continuous phenprocoumon therapy during ablation (and at least 4 weeks before ablation). Patients on phenprocoumon were matched by age, gender, and type of rhythm disorder in a 2:1 ratio with patients of the apixaban group.
All patients underwent dual-source computed tomography or transesophageal echocardiography to rule out intracardiac thrombus before ablation. All patients provided written informed consent to imaging and ablation procedures. Ablation procedures were performed in fasting state under analgo-sedation (fentanyl-sodium and disoprivan). Vascular access was obtained through a femoral vein. In addition, an arterial 4-F femoral line was placed at the discretion of the operator for monitoring purposes (mainly in patients with persistent AF requiring more extensive ablations). A steerable 6-F catheter (EP-XT, C.R. Bard, Inc., Lowell, Massachusetts) was placed in the coronary sinus. The left atrium was accessed by a single transseptal puncture (or through a patent foramen ovale), and both a steerable sheath (Agilis; St. Jude Medical, St. Paul, Minnesota) and the ablation catheter were introduced through the single puncture (or persisting foramen ovale) into the left atrium (LA). A circular mapping catheter was then placed in the LA through the steerable sheath, and a 3-dimensional reconstruction of the LA obtained using Carto 3 (Biosense Webster, Diamond Bar, California) or NavX Velocity (St. Jude Medical, St. Paul, Minnesota). Ablations were performed using a 3.5 mm irrigated-tip radiofrequency ablation catheter. In paroxysmal AF, circumferential en-bloc pulmonary vein isolation was performed. In persistent AF, a sequential ablation approach with pulmonary vein isolation and substrate modification (ablation of complex fractionated atrial electrograms) was used. Atrial tachycardias either as the presenting arrhythmia or arising during ablation were mapped by entrainment maneuvers and activation sequence, and lesions (focal or linear) were placed according to the identified arrhythmogenic mechanism. If no sinus rhythm was achieved during ablation, direct current cardioversion was performed. The same experienced operators performed ablation procedures in both groups.
On the day of the ablation, patients on apixaban received a morning dose of 2.5 mg apixaban. Patients on phenprocoumon received the usual maintenance dose the evening before ablation. Baseline activated clotting time (ACT) was measured after completing vascular access. After completing access to the left atrium, heparin bolus of 50 to 60 IU/kg was administered intravenously. ACT was then checked 15 minutes after bolus injection and every 30 minutes thereafter. Target ACT was 250 to 300 seconds in both groups.
Sheaths were routinely removed 4 hours after the end of the procedure. A compression bandage was applied for 2 to 4 hours. Patients on apixaban received their standard evening dose (after removal of sheaths). Phenprocoumon patients received their standard maintenance dose. All patients received a physical examination after removal of the compression bandage. Duplex ultrasound of the vascular access site was performed the following day. Echocardiography was performed at the end of the ablation procedure and on the following day. Patients were usually discharged 2 days after the procedure. Oral anticoagulation was continued for at least 3 months.
The primary end point was a composite of bleeding, thromboembolic events, and death during the hospital stay, as previously described. Bleedings were classified according to criteria proposed in the Randomized Evaluation in PCI Linking Angiomax to Reduced Clinical Events (REPLACE-2) study, including pericardial effusion. Major bleedings were defined as (1) any bleeding requiring blood transfusion, (2) retroperitoneal bleedings or hematomas requiring surgical intervention, (3) pericardial effusions requiring drainage or surgical intervention, or (4) intracranial hemorrhage. Minor bleedings were defined as (1) hematomas >5 cm not requiring transfusion or interventions or (2) pericardial effusions >5 mm not requiring an intervention. Thromboembolic events were defined as (1) cerebrovascular accidents or transient ischemic attack. In case of a clinical suspicion, a cranial computed tomography of the brain was performed and the diagnosis had to be confirmed by a neurologist. Furthermore, thromboembolic events included (2) deep vein thrombosis and pulmonary embolism and (3) peripheral arterial embolism.
Dichotomous traits were compared using chi-square tests and continuous traits were compared using t tests or nonparametric Wilcoxon tests, wherever appropriate. To account for the influence of procedure duration and patients’ body weight on required heparin dose, we constructed a linear regression model with heparin dose (natural log-transformed for normality) as a dependent variable and study group, body weight, and procedure duration as independent variables. To identify predictors of bleeding complications, we constructed logistic regression models with the occurrence of any bleeding complication as the dependent variable. A 2-tailed p value of <0.05 was regarded significant in all analyses. Statistical analyses were carried out using R version 3.0.2 ( http://www.r-project.org/ ) (with the library MatchIt ( http://gking.harvard.edu/matchit ) for matching) and SPSS version 21 (SPSS Inc., Chicago, Illinois).
Results
A total of 105 consecutive patients taking apixaban were matched with 210 patients on phenprocoumon. Clinical characteristics of the study sample are given in Table 1 . The study sample consisted of elderly patients (mean age 63 and 64 years in apixaban and phenprocoumon groups, respectively), and paroxysmal AF was the most common arrhythmia (54%). Cardiovascular risk profile was generally similar in both groups, although chronic kidney disease was somewhat more frequent in the apixaban group (10% vs 4%, p = 0.04).
| Variable | Apixaban group (n=105) | Phenprocoumon group (n=210) | p-value |
|---|---|---|---|
| Age (years) | 63 ± 12 | 64 ±10 | 0.59 (matched) |
| Women | 35 (33%) | 70 (33%) | matched |
| Type of arrhythmia | |||
| Paroxysmal AF | 57 (54%) | 114 (54%) | matched |
| Persistent AF | 28 (27%) | 56 (27%) | matched |
| Left atrial tachycardia | 20 (19%) | 40 (19%) | matched |
| Body mass index (kg/m²) | 27.5 ± 5.0 | 27.1 ± 4.3 | 0.58 |
| Hypertension | 63 (60%) | 120 (61%) | 0.83 |
| Diabetes mellitus | 10 (10%) | 23 (11%) | 0.70 |
| Stroke | 8 (8%) | 15 (7%) | 0.88 |
| Coronary artery disease | 15 (16%) | 22 (11%) | 0.22 |
| Chronic kidney disease | 10 (10%) | 7 (4%) | 0.04 |
| CHADS 2 score | 1.2 ± 1.1 | 1.0 ± 0.9 | n.a. |
| Median CHADS 2 score ∗ | 1 (0, 2) | 1 (0, 2) | 0.09 |
| CHA 2 DS 2 -VASc score | 2.1 ± 1.8 | 2.0 ± 1.5 | n.a. |
| Median CHA 2 DS 2 -VASc score ∗ | 2 (1, 3) | 2 (1, 3) | 0.78 |
| LV ejection fraction (%) | 58 ± 7 | 56 ± 6 | 0.06 |
| Left atrial diameter (mm ∗ ) | 44 (39, 49) | 44 (40, 49) | 0.59 |
| Medication | |||
| Aspirin | 23 (22%) | 31 (15%) | 0.10 |
| Clopidogrel | 6 (6%) | 3 (1%) | 0.03 |
| Aspirin+Clopidogrel | 4 (4%) | 2 (1%) | 0.80 |
| ACEI/ARB | 38 (36%) | 108 (51%) | 0.01 |
| Betablocker | 93 (89%) | 196 (93%) | 0.15 |
| Apixaban 10 mg/d | 100 (95%) | n.a. | n.a. |
| Apixaban 5 mg/d | 5 (5%) | n.a. | n.a. |
The procedural data are provided in Table 2 . Presenting rhythm, ablation sites, and procedure duration did not differ relevantly between groups. Baseline ACT was similar in both groups (139 seconds and 143 seconds in apixaban and phenprocoumon groups, respectively, p = 0.16). Mean ACT during procedure was lower in the apixaban group than in the phenprocoumon group (median 258 seconds vs 288 seconds, p <0.001), whereas at the same time heparin doses were higher in the apixaban than in the phenprocoumon group (median 9,450 units vs 8,000 units, p <0.001). In a regression model accounting for procedure duration and patients’ body weight, mean heparin doses were 12% higher in the apixaban group than in the phenprocoumon group (95% confidence interval 6% to 19%, p <0.001).
| Variable | Apixaban group (n=105) | Phenprocoumon group (n=210) | p-value |
|---|---|---|---|
| Presenting rhythm | |||
| Sinus rhythm | 60 (57%) | 106 (51%) | 0.30 |
| Atrial fibrillation | 30 (29%) | 60 (29%) | 0.96 |
| Atrial tachycardia | 15 (14%) | 42 (20%) | 0.20 |
| Ablation location | |||
| PVI | 98 (93%) | 187 (89%) | 0.22 |
| CFAE | 29 (28%) | 51 (25%) | 0.55 |
| Left atrial lines | 36 (34%) | 76 (38%) | 0.54 |
| Intraprocedural CV | 19 (18%) | 44 (23%) | 0.35 |
| Procedure duration, min ∗ | 145 (105, 208) | 139 (93, 183) | 0.11 |
| RF time (min ∗ ) | 40 (25, 55) | 43 (30, 68) | 0.19 |
| Anticoagulation | |||
| Baseline INR | 1.05 ± 0.1 | 2.1 ± 0.4 | <0.001 |
| Baseline ACT (sec) | 139 ± 27 | 143 ± 20 | 0.16 |
| Mean ACT (sec) | 258 ± 26 | 288 ± 34 | <0.001 |
| Maximum ACT (sec) | 283 ± 45 | 311 ± 37 | <0.001 |
| Heparin (units) ∗ | 9450 (7250, 11750) | 8000 (6250, 10100) | <0.001 |
The safety end points are listed in Table 3 . The primary composite end point of bleeding and thromboembolic events occurred in 12 patients of the apixaban group and 26 patients of the phenprocoumon group (11.4 vs 12.3%, p >0.99). The only major bleeding complication in the apixaban group occurred in a patient with a combination of pseudoaneurysm and arteriovenous fistula at the femoral access site, which required surgical repair after 2 unsuccessful attempts of manual compression (no blood transfusion). The patient was discharged in good clinical condition. Minor bleeding complications occurred in 10 patients (9.5%) in the apixaban group, that is minor pericardial effusions (5 to 8 mm) or groin hematomas (>5 cm in diameter) not requiring blood transfusion or intervention.
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
