When should anticoagulation be 

stopped after atrial fibrillation ablation?

Paulus Kirchhof
Director, Institute of Cardiovascular Sciences, University of Birmingham, UK
SWBH and UHB NHS Trusts, Birmingham, UK
AFNET, Münster, Germany
http://www.birmingham.ac.uk/staff/profiles/cem/CVRS/Kirchhof-Paulus.aspx
email: p.kirchhof@bham.ac.uk
 (When) should anticoagulation be 
stopped after atrial fibrillation ablation?

Paulus Kirchhof
Director, Institute of Cardiovascular Sciences, University of Birmingham, UK
SWBH and UHB NHS Trusts, Birmingham, UK
AFNET, Münster, Germany
http://www.birmingham.ac.uk/staff/profiles/cem/CVRS/Kirchhof-Paulus.aspx
email: p.kirchhof@bham.ac.uk
Financial disclosures (past and current)
Consulting Fees/ Ownership/Partnership/
Speaker’s Bureau Research Grants
Honoraria Employee

3M Medica None None British Heart Foundation (BHF)

AstraZeneca European Union
Bayer Fondation Leducq
Boehringer Ingelheim German Federal Ministry for Education and Research (BMBF)
Boston Scientific German Research Foundation (DFG)
Bristol Myer Squibb 3M Medica
Cardiome Cardiovascular Therapeutics
Daiichi-Sankyo Daiichi Sankyo
Johnson & Johnson Pfizer / BMS
MEDA Pharma MEDA Pharma
Medtronic Medtronic
Merck OMRON
Otsuka St Jude Medical
Pfizer
Sanofi Aventis Patents (filed by University of Birmingham)
Servier WO 2015140571
Siemens WO 2016012783
Takeda
Progress in the management of patients with AF

ESC AF guidelines, Eur Heart J 37:2893-962.(2016)

Currently available oral anticoagulants

Kirchhof P. Lancet 390:1873-87 (2017)

Anticoagulation with DOACs / NOACs is the standard of care
2% of all patients with AF receiving NOACs will suffer a major bleeding event per year of therapy

Ruff CT, et al. Lancet. 383:955-962 (2014)

Camm AJ et al. Eur Heart J; 37:1145-53.(2016)
Kirchhof P et al. J Am Coll Cardiol 72:141-53.(2018)
AF, anticoagulation, and dementia
longitudinal analysis of 444106 patients with AF in Sweden, 74 years, 56% male, analysed for dementia

propensity matched cohorts

factors associated with dementia
age (per decade) 2.19
female sex 1.04
prior stroke / ICH 1.34 / 1.2
VKA / NOAC use 0.62 / 0.48
AAD use 0.72

Friberg L, Rosenqvist M. Eur Heart J published on line November (2017)

B.D.

67 year old man with hypertension, diabetes, normal kidney function, and symptomatic
paroxysmal AF has a scheduled admission for an AF ablation. He is currently on a NOAC,
bisoprolol, valsartan, metformin, and atorvastatin, He asks whether the ablation can be
dangerous, and whether he can stop the blood-thinners at some point after the ablation.

What do you suggest to the patient?

1. Stop NOAC 48 hours prior to the ablation – TAMPONADE!
2. Change to phenprocoumon a week before admission – VKA can be antagonised.
3. Skip the morning dose of apixaban – otherwise the transseptal puncture is unsafe.
4. Continue all medication as planned without periprocedural or long-term changes
5. The NOAC can be stopped a year after AF ablation unless AF recurs.
COMPARE: Continuous vs bridged VKA in AF ablation

8 major bleeds (0.8%) 3 major bleeds (0.4%)

7 pericardial effusions (0.9%) 2 pericardial effusions (0.5%)
174 minor bleeds (22%) 22 minor bleeds (4%)
Di Biase L, et al. Circulation. 129:2638-2644 (2014)
AF ablation on continuous VKA or NOAC?
674 patients randomised,
mean age 64 (58, 70) years
209 (33%) female

Patients at risk of
stroke (one of age
> 75 years, heart
failure,
hypertension,
diabetes, prior
stroke) with an
indication for atrial
fibrillation ablation

49 sites in Europe and the USA

650 patients planned

www.axafa.af-net.eu, Di Biase L, et al. Europace 19: 132-138. (2017)

AF ablation on continuous VKA or NOAC?
10%
Death

Stroke or TIA

Major Bleed (TIMI)

Major Bleed (ISTH)

5%

0%
)
)
)

)
)

)
13
17
18

48
15

18

07

40
(1
(3
(3

(7
(3

(3

(1

(7
an
n
an

AC
A

A
ra
VK

VK

VK

VK
ab
ab

O
at

ox

N
ix

ig
Ap

ab

ar
iv
D

AXAFA-AFNET 5 RE-CIRCUIT VENTURE-AF Combined

Cappato R et al. Eur Heart J; 36:1805-11.(2015)

Calkins H et al. N Engl J Med; 376:1627-36.(2017)
Kirchhof P et al. Eur Heart J 39:2942-55.(2018)
Hohnloser SH et al. Eur Heart J (2019)
What about the morning dose? AEIOU
300 patients undergoing AF ablation randomised to uninterrupted versus minimally interrupted (holding 1 dose)
periprocedural apixaban. Endpoints included clinically significant bleeding, major bleeding, and nonhemorrhagic
stroke or systemic embolism (SE) from the time of ablation through 30 days.

Uninterrupted apixaban  Interrupted apixaban  Risk difference

(n=150) (n=150)
Stroke 0 0
Major bleeding 1.3% 2.1% -0.7%
p = NS
Clinically significant 11.3% 9.7% 1.7% [-5.5% to 8.8%]
bleeding p = NS

Reynolds MR et al. JACC Clin Electrophysiol 4:580-8.(2018)

Subtle cognitive dysfunction after AF ablation?
150 patients undergoing comprehensive cognitive testing 2 and 90
days after ablation of paroxysmal AF (PAF, 60), persistent AF (PeAF,
30), or SVT (30), also 30 AF patients not undergoing AF ablation (Ctrl),
no clinical stroke / TIA
37 patients undergoing AF ablation, high-resolution
diffusion weighted brain magnetic resonance imaging
(HR-DWI MRI) within 48 hours after ablation

Cognitive Dysfunction

56 lesions in 16/37 patients

28% of AF patients at day 2
15% of AF patients at day 90 Medi C et al. J Am Coll Cardiol; 62:531-9.(2013)
Herm J, et al. Circ Arrhythm Electrophysiol. 6:843-850 (2013)
AF ablation on continuous VKA or NOAC?
  All patients Apixaban  VKA P value
(n=323) (n=162)  (n=161)

No lesion 239 (74.0%) 118 (72.8%) 121 (75.2%) 0.635

Exactly one lesion 46 (14.2%) 27 (16.7%) 19 (11.8%) 0.211

Exactly two lesions 21 (6.5%) 7 (4.3%) 14 (8.7%) 0.111

More than two lesions 17 (5.3%) 10 (6.2%) 7 (4.3%) 0.463

  All patients Apixaban VKA

Cognitive function  28 (26, 29) 28 (26, 29) 28 (26, 29)
(MoCA)
Change in MoCA 1.0 (-1.0, 2.0), 0.0 (-1.0, 2.0), 1.0 (-1.0, 2.0),
p<0.001* n=301 n=296
At least mild cognitive  141/607 75/305 (24.6%) 66/302 (21.9%)
impairment (MoCA <26) (23.2%), -5.1% -9.2%
-7.2%, p=0.005*

Kirchhof P et al. Eur Heart J 39:2942-55.(2018)

Stopping OAC in patients after AF ablation:
Current Consensus
AF ablation conveys an intrinsic thromboembolic risk (similar to cardioversion)
Consensus suggests to continue oral anticoagulation for 4-8 weeks / 1-2 months after AF ablation in
patients without stroke risk factors.
Consensus suggests to continue anticoagulation indefinitely in patients with stroke risk factors.

EHRA consensus statement, Sticherling C et al. Europace 17:1197-214.(2015)

Hansen ML et al. Europace 17:18-23.(2015)
ESC AF guidelines, Eur Heart J 37:2893-962.(2016)
HRS/EHRA/Solaece AF ablation consensus statement, Calkins H et al. Heart Rhythm 15:e275-e444.(2017)
Meta Analysis: NOACs may be safer than warfarin in patients
undergoing cardioversion

Kotecha D et al. J Am Coll Cardiol 72:1984-6.(2018)

Stopping OAC after AF ablation? What we say and what we do.
Surveys of Canadian EP cardiologists and of European EP specialists in 2012 and 2013
Danish nationwide analysis of event rates in 4050 AF patients after a first AF ablation (mean age 59 years, IQR 53-65,
years 2001 – 2011 analysed), considering stroke and bleeding in anticoagulated and non-anticoagulated patients.

• 95% in Canada said that they will discontinue

OAC after AF ablation in patients without
stroke risk factors
• 16% in Europe said they would discontinue
OAC after AF ablation in patients with stroke
risk factors
• 50% of ‘general’ AF patients without stroke
risk factors are continued on OAC

Lip GY et al. Europace 14:741-4.(2012)

Mardigyan V et al. Can J Cardiol 29:219-23.(2013)
Kirchhof P et al. Europace 16:6-14.(2014)
Karasoy D et al. Eur Heart J 36:307-14a.(2015)
Verma A et al. Am Heart J 197:124-32.(2018)
The complex relation between AF and stroke risk
AF without associated disorders has minimal
risk of stroke in young patients – so the story is
not AF in and of itself.
Many comorbidities have a stroke risk
independent of AF – so strokes in patients with
AF plus these disorders are not necessarily
consequent to AF.
Co-morbidities can produce atrial
pathophysiological alterations that promote
thrombus formation in the LA/LAA (“atrial
cardiomyopathy”) which is additive to the
prothrombotic abnormalities consequent to AF.
Thus, in AF, the stroke risk should increase:
As the AF burden increases, and,
As the comorbidities increase. Jim Reiffel, personal communication (2018)
ESC AF guidelines, Eur Heart J 37:2893-962.(2016)
Goette A et al. Europace 18:1455-90.(2016)
Stopping OAC in patients after AF ablation:
Open questions
AF ablation conveys an intrinsic thromboembolic risk (like cardioversion)
Consensus suggests to continue oral anticoagulation for 4-8 weeks / 1-2 months after AF ablation in
patients without stroke risk factors.
Consensus suggests to continue anticoagulation indefinitely in patients with stroke risk factors.
Are AF ablation patients different from other patients? If so, is this due to the patient or the procedure?
The precise thromboembolic risk during the later phases of left atrial wound healing (e.g.3-8 weeks
after ablation) is not known.
Energy types (RF / Cryo, no signal for differences in peri-procedural events)
With or without TOE / intracardiac echo (no signal for peri-procedural events)
Paroxysmal or chronic forms of AF (potentially slightly higher stroke risk in those with chronic AF)
Intensity of AF ablation (PVI only or PVI with additional ablation targets and extended lesions)
Bleeding risk on anticoagulation is relatively constant over time (outside of the periprocedural period)
Stopping OAC after AF ablation in patients without risk factors?

We may need a trial of stopping oral anticoagulation in patients without

stroke risk factors comparing stopping OAC / DOAC
2 weeks vs. 2 months after AF ablation.
Stopping OAC after AF ablation in patients with risk factors?
Danish nationwide analysis of event rates in 4050 AF patients after a first AF ablation (mean age 59 years, IQR 53-65,
years 2001 – 2011 analysed), considering stroke and bleeding in anticoagulated and non-anticoagulated patients.
Thromboembolic risk appeared lower than in a matched cohort of non-ablated patients (HR 0.53).

Karasoy D et al. Eur Heart J 36:307-14a.(2015)

Stopping OAC after AF ablation in patients with risk factors: OCEAN
Randomisation of 1572 patients with at least one of the CHADSVa risk factors one year after successful AF ablation to
either NOAC therapy (rivaroxaban 15 mg OD) or antiplatelet therapy (low-dose aspirin OD)

Primary outcome: Composite of stroke, systemic

embolism, and incident ‘silent stroke lesions’ on MRI
imaging
Key secondary outcomes: Major bleeding, TIA, MoCA
and MMSE (assessing cognitive function)

At least 3x24 hour

Holter without AF
Verma A et al. Am Heart J 197:124-32.(2018)
B.D.

67 year old man with hypertension, diabetes, normal kidney function, and symptomatic
paroxysmal AF has a scheduled admission for an AF ablation. He is currently on a NOAC,
bisoprolol, valsartan, metformin, and atorvastatin, He asks whether the ablation can be
dangerous, and whether he can stop the blood-thinners at some point after the ablation.

What do you suggest to the patient?

1. Stop NOAC 48 hours prior to the ablation – TAMPONADE!
2. Change to phenprocoumon a week before admission – VKA can be antagonised.
3. Skip the morning dose of apixaban – otherwise the transseptal puncture is unsafe.
4. Continue all medication as planned without periprocedural or long-term changes
5. You can participate in the OCEAN trial a year after successful AF ablation
 Thank you

Download Professor Paulus Kirchhof

- My AF app (for AF patients) and
- AF Manager app (for physicians / HCP) from Apple or Android p.kirchhof@bham.ac.uk
Store to get IT support tools underpinning integrated AF www.birmingham.ac.uk/paulus-kirchhof
managment based on the 2016 ESC AF guidelines

ESC AF guidelines, Kirchhof P et al. Eur Heart J 37:2893-962.(2016)

HRS/EHRA/Solaece AF ablation consensus statement, Calkins
supported H et al. Heart
by European Union,Rhythm 15:e275-e444.(2017)
grant agreement No 633196 [CATCH ME]