Periprocedural management of anticoagulation for atrial fibrillation catheter ablation in direct oral anticoagulant–treated patients

Abstract

Background

Guidelines recommend performing atrial fibrillation (AF) catheter ablation without interruption of a direct oral anticoagulants (DOACs) and to administer unfractionated heparin (UFH) for an activated clotting time (ACT) ≥300 seconds, by analogy with vitamin K antagonist (VKA). Nevertheless, pharmacological differences between DOACs and VKA, especially regarding ACT sensitivity and UFH response, prevent extrapolation from VKA to DOACs.

Hypothesis

The level of anticoagulation at the time of the procedure in uninterrupted DOAC–treated patients is unpredictable and would complicate intraprocedural UFH administration and monitoring.

Methods

This prospective study included interrupted DOAC–treated patients requiring AF ablation. Preprocedural DOAC concentration ([DOAC]), intraprocedural UFH administration, and ACT values were recorded. A cohort of DOAC‐treated patients requiring flutter catheter ablation was considered to illustrate [DOAC] without DOAC interruption.

Results

Forty‐eight patients underwent AF and 14 patients underwent flutter ablation, respectively. In uninterrupted DOAC–treated patients, [DOAC] ranged from ≤30 to 466 ng/mL. When DOAC were interrupted, from 54 to 218 hours, [DOAC] were minimal (maximum: 36 ng/mL), preventing DOAC‐ACT interference. Anyway, ACT values were poorly correlated with UFH doses (R ² = 0.2256).

Conclusions

Our data showed that uninterrupted DOAC therapy resulted in an unpredictable and highly variable initial level of anticoagulation before catheter ablation. Moreover, even with DOAC interruption preventing interference between DOAC, UFH, and ACT, intraprocedural UFH monitoring was complex. Altogether, our exploratory results call into question the appropriateness of transposing UFH dose protocols, as well as the relevance of ACT monitoring in uninterrupted DOAC–treated patients.

1. INTRODUCTION

Direct oral anticoagulants (DOACs), which include dabigatran, rivaroxaban, apixaban, and edoxaban, are recommended for stroke prevention in patients with atrial fibrillation (AF) and have progressively replaced vitamin K antagonists (VKA) as the new standard of care.1

On top of long‐term anticoagulation, guidelines recommend AF catheter ablation as first‐line therapy in patients with symptomatic paroxysmal or persistent AF.1 This invasive procedure entails a risk of major bleeding, including hemopericardium, and a transient but high procedure‐related risk for thromboembolic events that varies from 1% to 5%.2, 3 Therefore, rigorous anticoagulation throughout the procedure is crucial.

To achieve this goal, latest guidelines recommend performing AF ablation without interruption of oral anticoagulation and to administer intraprocedural unfractionated heparin (UFH) in an adjusted dose to achieve and maintain a target activated clotting time (ACT) of ≥300 seconds.4 Indeed, in both interrupted and uninterrupted VKA–treated patients, an ACT <300 seconds is an independent predictor of major thromboembolic complications.5 In uninterrupted DOAC–treated patients, the same ACT ≥300 seconds usually proposed for intraprocedural UFH is based on very limited evidence, mostly extrapolated from VKA‐treated populations. Although a few recent studies report that uninterrupted DOAC therapy during AF ablation is feasible and safe, the critical point of intraprocedural anticoagulation management has not been addressed and protocols on UFH administration and ACT monitoring are lacking.6, 7, 8, 9 However, several arguments support that UFH management cannot be directly extrapolated from VKA‐treated patients: the initial level of anticoagulation varies according to DOAC concentration, DOAC interferes with the ACT, and the sensitivity of the ACT to DOAC as well as the response to UFH administration in DOAC‐treated patients strongly differ from that of VKA‐treated patients and also among DOACs.10, 11

This study aims to illustrate intraprocedural UFH management complexity in uninterrupted DOAC–treated patients. For that, we report data from two different populations of patients, to build a thought model: (1) DOAC‐treated patients undergoing AF catheter ablation after DOAC interruption, as it used to be recommended in previous guidelines; and (2) DOAC‐treated patients undergoing flutter catheter ablation, usually performed without DOAC interruption. We analyzed our practice to show that in uninterrupted DOAC–treated patients, the initial level of anticoagulation before catheter ablation is unpredictable and that, even in situations without interaction between DOAC and ACT, UFH administration remains challenging.

2. METHODS

This was a post hoc analysis of the Preprocedural Concentration of Direct Oral Anticoagulants (CORIDA) study (http://www.ClinicalTrials.gov NCT02643992), a prospective, multicenter cohort study including DOAC‐treated patients requiring any invasive procedure.12 The aim of the main study was to investigate the factors influencing preprocedural DOAC concentration. In the present study, we focused on patients who required catheter ablation, and we specifically analyzed periprocedural management of anticoagulation. The CORIDA study was approved by the research ethics boards (Comité de Protection des Personnes Ile de France 1, ref. 2013–13 272 and Comité interne CHU UCL Namur 09/2014, ref. B039201422406). All patients were informed before inclusion, and the physician in charge of the patient signed the nonopposition form.

2.1. Patients and catheter ablation procedures

Patients from the CORIDA study were included in this study if they were taking dabigatran, rivaroxaban, or apixaban and required catheter ablation procedures, namely pulmonary vein isolation and cavotricuspid isthmus ablation, the cornerstone of AF and flutter catheter ablation procedures, respectively. For patients undergoing AF ablation, pulmonary vein isolation could be associated with creation of a cavotricuspid isthmus line or other left atrium lines (linear lesions or posterior‐wall isolation). Radiofrequency energy was the dominant energy source used for both AF and flutter ablation.

2.2. Preprocedural coagulation testing

Patients had one blood sample taken in the electrophysiology laboratory just before the beginning of the procedure for DOAC concentration measurement, as previously described.11

2.3. Intraprocedural anticoagulation management during AF catheter ablation

Patients were managed according to local practice and current guidelines, and the study had no impact on procedures. Intravenous UFH was administered as a bolus prior to or immediately following trans‐septal puncture during AF catheter ablation procedures. The ACT was measured using the Hemochron® Signature Elite Whole Blood Microcoagulation System (Accriva Diagnostics, San Diego, CA) and ACT‐LR cartridges, every 15 to 20 minutes after the UFH bolus administration, and additional UFH was administered to maintain ACT ≥300 seconds. UFH bolus and additional doses were left at the physician's discretion. Intravenous protamine administration at the end of the procedure was left at the physician's discretion.

2.4. Data collection

For each patient, demographic, clinical, and biological data were collected: age, sex, weight, history of stroke, CHADS₂ and CHA₂DS₂‐VASc scores, and creatinine clearance calculated using the Cockcroft‐Gault formula. The following characteristics of periprocedural DOAC management were recorded: DOAC type, duration of DOAC interruption, use of preprocedural low‐molecular‐weight heparin (LMWH) bridging, and preprocedural DOAC concentration. The duration of DOAC interruption was recorded as the exact time from the last drug intake to the procedure. The following characteristics related to the catheter ablation procedure were recorded: type of arrhythmia, procedure length, UFH bolus, UFH requirement for the procedure, ACT after initial UFH bolus, and maximal ACT achieved during the procedure. Periprocedural complications were noted, including major and minor bleeding events according to the International Society on Thrombosis and Haemostasis (ISTH), thromboembolic events (stroke, transient ischemic attack, and symptomatic venous thromboembolism), major cardiovascular events; and death.13 They were assessed from DOAC interruption to the end of hospitalization.

2.5. Statistical analysis

Descriptive statistics are presented as median median [minimum‐maximum] for quantitative variables and number (percentage) for qualitative ones. Linear regression analysis was performed with the Pearson correlation. A P value of 0.05 was considered significant.

3. RESULTS

3.1. Characteristics of patients undergoing AF catheter ablation

A total of 48 DOAC‐treated patients from the CORIDA study underwent AF catheter ablation. Their characteristics are shown in the Table 1. The majority of patients were treated with rivaroxaban (62.5%) and apixaban (25%), followed by dabigatran (12.5%). More than a quarter had a CHA₂DS₂‐VASc score ≥ 3, and 3 patients had a history of stroke. Four (8%) ablations were aborted because of left atrium appendage thrombus visualized on transesophageal echocardiography performed at the time of the procedure.

Table 1.

Patient characteristics

Variable	AF Ablation, n = 48	Flutter Ablation, n = 14
Age, y	67 [40–81]	67 [57–83]
Male sex	31 (65)	11 (79)
Weight, kg	80 [43–185]	91 [64–120]
CrCl <50 mL/min	2 (4)	0 (0)
History of stroke	3 (6)	0 (0)
Dabigatran/rivaroxaban/apixaban	6 (12.5)/30 (62.5)/12 (25)	2 (14)/8 (57)/4 (29)
CHA2DS2‐VASc score	2 [0–4]	2 [0–3]
CHADS2 score	1 [0–3]	1 [0–2]
Antiarrhythmic drugs
Amiodarone	15 (31)	4 (29)
Verapamil	0 (0)	2 (14)
Diltiazem	0 (0)	0 (0)
β‐Blocker	30 (67)	6 (43)
Antiplatelet agents	6 (12.5)	2 (14)
Type of arrhythmia
Paroxysmal	21 (48)	5 (36)
Persistent	22 (50)	9 (64)
Long‐standing persistent	1 (2)a	0 (0)
Rhythm at the beginning of the procedure
Sinus rhythm	25 (57)	6 (43)
Arrhythmia	19 (43)a	8 (57)
AF catheter ablation length, min	141 [79–452]	NA
Duration of DOAC interruption, h	110 [44–218]	6.5 [2.5–28]
DOAC concentration, ng/mL	≤30 [≤30–36]	170 [≤30–466]
LMWH bridging	45 (94)	0 (0)

Abbreviations: AF, atrial fibrillation; CHADS₂, congestive HF, HTN, age ≥ 75 y, DM, prior stroke/TIA/TE; CHA₂DS₂‐VASc, congestive HF, HTN, age > 75 y, DM, stroke/TIA, vascular disease, age 65–74 y, sex category (female); CrCl, creatinine clearance; DM, diabetes mellitus; DOAC, direct oral anticoagulant; HF, heart failure; HTN, hypertension; IQR, interquartile range; LMWH, low‐molecular‐weight heparin; max, maximum; min, minimum; NA, not applicable; TE, thromboembolism; TIA, transient ischemic attack.

Data are presented as n (%) or median [min–max].

^aPercentage was based on the number of patients who effectively underwent AF catheter ablation (n = 44).

3.2. Preprocedural DOAC management for AF catheter ablation

The median duration of DOAC interruption before AF catheter ablation was 110 [54–218] hours; DOAC interruption was >72 hours for all but 3 patients. Preprocedural bridging with therapeutic LMWH was performed in all these patients, mainly with enoxaparin twice daily (58%), but also tinzaparin once daily (31%) and nadroparin once daily. The median duration between the last LMWH injection and AF catheter ablation was 24.5 [12.5–55] hours.

3.3. Preprocedural coagulation testing in patients undergoing AF catheter ablation

The median DOAC concentration measured just before the beginning of the AF catheter ablation was ≤30 [≤30–36] ng/mL, with only 6% of the measurements remaining >30 ng/mL.

3.4. Intraprocedural anticoagulation during AF catheter ablation

The majority of patients (57%) received an initial UFH bolus of 5000 IU, resulting in a median ACT of 237 [204–368] seconds; 14% of them received 4000 IU UFH, with a median ACT of 210 [196–230] seconds and others received 6000 to 10 000 IU UFH, with a median ACT of 295 [210–340] seconds. Adjusted to weight, the initial UFH bolus dose of 57 [38–110] IU/kg increased the ACT to 245 [196–368] seconds, and 18% of the patients reached the target ACT of ≥300 seconds. Correlation between body weight–adjusted UFH bolus and subsequent ACT value was poor (R ² = 0.2256, P = 0.0013; Figure 1). An additional UFH bolus of 2000 [1000–8000] IU, or 23 [8–87] IU/kg, was administered in 86% of the entire cohort. The median maximal ACT then achieved was 309 [206–>400] seconds. Thus, 58% of the patients reached ACT ≥300 seconds during the procedure. The median total UFH requirement was 90 [46–136] IU/kg. None of the patients received protamine at the end of the procedure.

Figure 1.

Correlation of ACT achieved after the initial UFH bolus with body weight–adjusted UFH bolus dose. Scatter plot of ACT (seconds) vs body weight–adjusted UFH bolus dose (IU/kg), with linear regression line (full line) and ACT ≥300 seconds (dotted line). Abbreviations: ACT, activated clotting time; UFH, unfractionated heparin

3.5. Periprocedural outcomes of patients undergoing AF catheter ablation

Two patients had minor bleeding (2 groin hematomas). In these patients, preprocedural DOAC concentrations were 24 and 32 ng/mL and maximal ACT values were 269 and 330 seconds, respectively. No thromboembolic event, major cardiovascular event, or death was recorded.

3.6. Flutter catheter ablation

Fourteen DOAC‐treated patients from the CORIDA study underwent flutter catheter ablation. Their characteristics are shown in Table 1. DOACs were not discontinued or were discontinued shortly before flutter catheter ablation, with a median duration of interruption of 6.5 [2.5–28] hours, resulting in a wide range of DOAC concentrations, from ≤30 to 466 ng/mL, at the time of procedure. Moreover, 43% of the DOAC concentrations remained >200 ng/mL, 64% remained >100 ng/mL, and only 7% were ≤30 ng/mL. Neither bleeding nor thromboembolic complications occurred.

4. DISCUSSION

There are two main findings from our study that focused on DOAC‐treated patients undergoing catheter ablation. First, uninterrupted DOAC resulted in a large and unpredictable range of DOAC concentrations ie, highly variable level of anticoagulation before intraprocedural UFH administration. Second, even after DOAC interruption, the effect of UFH bolus administration was not foreseeable, underscoring the well‐known need for an accurate monitoring of UFH to reach the targeted intraprocedural level of anticoagulation ie ACT ≥300 seconds. Therefore, because interferences between DOAC, UFH, and ACT are increasingly described, our results point out the limit of intraprocedural UFH monitoring in the presence of DOAC, whereas optimal anticoagulation is needed to prevent both thromboembolic and bleeding events.

The initial level of anticoagulation before AF catheter ablation is a crucial parameter to optimize intraprocedural UFH administration. Indeed, it has been shown that uninterrupted VKA–treated patients have a lower UFH requirement and are more likely to achieve an ACT ≥300 seconds with a single bolus, compared with interrupted VKA–treated patients.14 This strategy reduced the occurrence of periprocedural thromboembolic and bleeding complications.6 Whereas the initial level of anticoagulation is measurable in VKA‐treated patients, based on international normalized ratio values, it remains unknown in uninterrupted DOAC–treated patients, where DOAC concentration is not routinely measured. Moreover, it is highly variable, as we reported that a last DOAC intake within 24 hours resulted in a wide range of DOAC concentrations, from ≤30 to 466 ng/mL, at the time of the procedure. This large interindividual pharmacokinetic variability showed that the initial level of anticoagulation varies substantially; thus, the initial UFH dose should not be fixed, unless accepting an unpredictable intraprocedural anticoagulation.

We showed that a strategy based on DOAC interruption that ensures undetectable DOAC concentration at the time of AF catheter ablation still resulted in a highly interindividual variability in UFH response, with a poor correlation between UFH dose and ACT levels and fewer than two‐thirds of the patients achieving ACT ≥300 seconds during the procedure. Indeed, ACT is accurate for monitoring UFH effects, as a strong and reproducible correlation between ACT and UFH anti‐Xa activity was repeatedly reported.15, 16 However, several factors affect the anticoagulant response to fixed or weight‐adjusted UFH dose.17, 18 Bloemen et al. described a large interindividual disparity of UFH pharmacodynamic effect with coefficients of variation of 24% to 43%.19 Moreover, like us, Gabus et al. showed that more than half of the patients with interrupted VKA and LMWH bridging undergoing AF catheter ablation remained with ACT <300 seconds at 20 minutes after the initial UFH bolus.20 Mulry et al. showed that the linear relationship between UFH dose and ACT was limited, as the slope of this line varied from patient to patient.21 This unpredictable response to UFH administration underscores the need for a reliable monitoring to achieve and maintain an appropriate intraprocedural level of anticoagulation during AF catheter ablation.

Because ACT is not specific to UFH, multiple interferences between DOAC, UFH, and ACT question the reliability of the monitoring. First, as discussed above, the initial level of anticoagulation in uninterrupted DOAC–treated patients varies substantially. Second, the ACT has poor performance to reflect DOAC concentration (ie, initial level of anticoagulation before UFH administration). When blood samples were spiked with increasing concentrations of apixaban, rivaroxaban, and dabigatran ranging from 0 to 500 ng/mL, the sensitivity of the ACT strongly differed between DOACs. 10, 11 Van Ryn et al. showed that the dose‐dependent increase in the ACT was linear with dabigatran concentrations ≤250 ng/mL, but then flattened.22 With apixaban and rivaroxaban, the flat dose–response effect was observed at >100 ng/mL. Thus, high concentration of DOAC, especially of anti‐Xa, could be associated with a normal or poorly prolonged ACT.23 Moreover, DOAC impact varies according to the device used, Hemochron® ACT‐LR, using celite as activator, is the most sensitive to DOAC.10

Third, UFH effect on the ACT varies according to the uninterrupted oral anticoagulant. DOAC‐treated patients required more UFH and reached the ACT target ≥300 seconds later compared with VKA.24 Furthermore, UFH effect on the ACT also varies according to the DOAC agent. Nagao et al. showed that the total UFH requirement for the procedure was higher in apixaban‐treated patients than in patients receiving dabigatran or rivaroxaban, and then VKA.25 On the contrary, Yamaji et al. reported that higher UFH bolus were required for apixaban and rivaroxaban compared with dabigatran.26 Thus, these results confirm interference between DOAC, UFH, and ACT.

Last, in uninterrupted VKA patients, the ACT reflects the effects of both VKA and UFH, because the ACT strongly correlates with the international normalized ratio, and UFH and VKA have additive effects.27, 28, 29 In DOAC‐treated patients, the ACT may reflect the anticoagulant effect of neither DOAC nor UFH, potentially leading to a mismanagement of intraprocedural anticoagulation. In particular, we might administer too much UFH during AF catheter ablation to achieve a target ACT that has, in fact, no value in the presence of uninterrupted DOAC–treated patients. The consequences of this mismanagement are unclear. The recent studies assessing AF catheter ablation with uninterrupted DOAC reported a low incidence of bleeding and thromboembolic events, similar to uninterrupted VKA, but UFH management remained unknown.7, 8, 9 One could argue that the potential UFH overdose induced by DOAC‐ACT and UFH‐DOAC interference is not relevant, as major bleeding events occur infrequently, nevertheless, blinded intraprocedural anticoagulation may make it harder to manage complications.

Although our study is not designed to modify clinical practice, it aims to draw attention to an unknown and neglected aspect of AF catheter ablation. Cardiologists and electrophysiologists should become aware of the evidence gaps regarding intraprocedural anticoagulation and ACT monitoring in uninterrupted DOAC–treated patients. In the end, the ACT‐based goal‐directed strategy is, in fact, a blinded strategy for these patients. For now, vigilance regarding bleeding and thrombosis should be reinforced while waiting for more laboratory and clinical data regarding intraprocedural anticoagulation.

4.1. Study limitations

There are several limitations to our study. First, the sample size is limited. Second, ACT values were not measured in the flutter ablation group, as patients did not receive intraprocedural UFH. However, our aim was not to compare ACT values according to DOAC concentrations, but to show the high interindividual variability of DOAC concentration with uninterrupted DOAC. Moreover, in the AF ablation group, the ACT was not measured before UFH administration. Third, bridging was widely used in our cohort, as it was previously recommended.6 Nevertheless, our aim was not to assess the relevance of a strategy including DOAC interruption and bridging, but to use such a strategy as a model of situation where residual DOAC concentration cannot interfere with ACT monitoring. Finally, these results obtained with the Hemochron® device cannot be extrapolated to other points of care because of wide variability between instruments and reagents.

5. CONCLUSION

Our data showed that uninterrupted DOAC therapy resulted in an unpredictable and highly variable initial level of anticoagulation before catheter ablation. Moreover, they highlighted that even with DOAC interruption preventing interference between DOAC, UFH, and ACT, intraprocedural UFH monitoring was complex. Altogether, our exploratory results call into question the appropriateness of transposing UFH dose protocols, as well as the relevance of ACT monitoring in uninterrupted DOAC–treated patients. Complementary data are urgently required to optimize UFH administration and monitoring in uninterrupted DOAC–treated patients undergoing AF catheter ablation.

Martin A‐C, Lessire S, Leblanc I, et al. Periprocedural management of anticoagulation for atrial fibrillation catheter ablation in direct oral anticoagulant–treated patients. Clin Cardiol. 2018;41:646–651. 10.1002/clc.22944