Europace (2020) 22, 1337–1344 CLINICAL RESEARCH doi:10.1093/europace/euaa132 Ablation for atrial fibrillation Predictors of recurrence of atrial fibrillation within the first 3 months after ablation Matthias Daniel Zink 1,2, Winnie Chua3, Stef Zeemering2, Luigi di Biase4, Bayes de Luna Antoni5, Callans David6, Gerhard Hindricks7, Karl Georg Haeusler8, Hussein R. Al-Khalidi9, Jonathan P. Piccini9, Lluı́s Mont10, Jens Cosedis Nielsen11, Luis Alberto Escobar5, Joseph de Bono12, Isabelle C. Van Gelder13, Tom de Potter14, Daniel Scherr15,2, Sakis Themistoclakis16, Derick Todd17, Paulus Kirchhof3, and Ulrich Schotten2* 1 Department of Internal Medicine I, University Hospital RWTH Aachen, Aachen, Germany; 2Department of Physiology, Cardiovascular Research Institute Maastricht (CARIM), Universiteitsingel 50, 6229 ER Maastricht, Netherlands; 3Institute of Cardiovascular Sciences, University of Birmingham, Birmingham, UK; 4Department of Medicine (Cardiology), Albert Einstein College of Medicine at Montefiore Hospital, Montefiore-Einstein Center for Heart & Vascular Care New York, NY, USA; 5Autonomous University of Barcelona and Institut Català Ciències Cardiovasculars (ICCC)-St. Pau Hospital, Barcelona, Spain; 6Cardiology Division, Hospital of the University of Pennsylvania, Philadelphia, PA, USA; 7 University of Leipzig, Heart Center Leipzig, Leipzig, Germany; 8Department of Neurology, University Hospital Würzburg, Würzburg, Germany; 9Department of Cardiac Electrophysiology, Duke University Medical Center, Duke Clinical Research Institute, Durham, NC, USA; 10Arrhythmia Section, Universitat de Barcelona, Hospital Clinic, Barcelona, Catalonia, Spain; 11Department of Cardiology, Aarhus University Hospital, Aarhus, Denmark; 12Department of Cardiology, University Hospitals Birmingham NHS Foundation Trust, Birmingham, UK; 13Department of Cardiology and Thorax Surgery, UMCG Thorax Center, University of Groningen, Groningen, The Netherlands; 14 Department of Cardiology, Electrophysiology section, Cardiovascular Center, OLV Hospital, Aalst, Belgium; 15Department of Cardiology, Medical University of Graz, Graz, Austria; 16Unit of Electrophysiology and Cardiac Pacing, Dell’Angelo Hospital, Mestre-Venice, Italy; and 17Department of EP, Devices and ICC, Liverpool Heart and Chest Hospital, Liverpool, UK Received 29 September 2019; revised 19 January 2020; editorial decision 14 April 2020; online publish-ahead-of-print 28 July 2020 Aims Freedom from atrial fibrillation (AF) at 1 year can be achieved in 50–70% of patients undergoing catheter ablation. Recurrent AF early after ablation most commonly terminates spontaneously without further interventional treat- ment but is associated with later recurrent AF. The aim of this investigation is to identify clinical and procedural factors associated with recurrence of AF early after ablation. ................................................................................................................................................................................................... Methods and We retrospectively analysed data for recurrence of AF within the first 3 months after catheter ablation from the results randomized controlled AXAFA–AFNET 5 trial, which demonstrated that continuous anticoagulation with apixaban is as safe and as effective compared to vitamin K antagonists in 678 patients undergoing first AF ablation. The pri- mary outcome of first recurrent AF within 90 days was observed in 163 (28%) patients, in which 78 (48%) patients experienced an event within the first 14 days post-ablation. After multivariable adjustment, a history of stroke/tran- sient ischaemic attack [hazard ratio (HR) 1.54, 95% confidence interval (CI) 0.93–2.6; P = 0.11], coronary artery dis- ease (HR 1.85, 95% CI 1.20–2.86; P = 0.005), cardioversion during ablation (HR 1.78, 95% CI 1.26–2.49; P = 0.001), and an age:sex interaction for older women (HR 1.01, 95% CI 1.00–1.01; P = 0.04) were associated with recurrent AF. The P-wave duration at follow-up was significantly longer for patients with AF recurrence (129 ± 31 ms vs. 122 ± 22 ms in patients without AF, P = 0.03). ................................................................................................................................................................................................... Conclusion Half of all early AF recurrences within the first 3 months post-ablation occurred within the first 14 days post- ablation. Vascular disease and cardioversion during the procedure are strong predictors of recurrent AF. P-wave duration at follow-up was longer in patients with recurrent AF. ................................................................................................................................................................................................... Trial Clinicaltrials.gov identifier NCT02227550 registration 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 䊏 Keywords Atrial fibrillation • Ablation • AXAFA • Early recurrence • Blanking period • Apixaban ................................................................................................................................................................................................... *Corresponding author. Tel: þ31 (0)43 3881077. E-mail address: schotten@maastrichtuniversity.nl C The Author(s) 2020. Published by Oxford University Press on behalf of the European Society of Cardiology. V This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/4.0/), which permits non-commercial re-use, distribution, and reproduction in any medium, provided the original work is properly cited. For commercial re-use, please contact journals.permissions@oup.com 1338 M.D. Zink et al. Study population and inclusion/exclusion What’s new? • In this retrospective analysis of the randomized control criteria In this study, we considered 678 patients in the AXAFA trial and excluded AXAFA-AFNET 5 trial, 28% of patients had recurrent atrial fi- patients who were still in AF after ablation (N = 30) and patients who had brillation (AF) within the first 3 months post-ablation. Half of no complete 3 months of follow-up data including rhythm status or date all recurrences occurred within the first 14 days. of first recurrence of AF after ablation (N = 59, Supplementary material • Vascular disease and the need for cardioversion during ablation online, Figure S1). Excluded patients showed no difference in epidemio- were strong predictors of recurrent AF. logic data, symptoms, or items of CHA2DS2-VASc score as compared to • P-wave duration of patients with recurrent AF was longer at patients who were not excluded (Supplementary material online, Table the 3 months of follow-up. S1). Participants with previous ablation or surgical therapy for AF were excluded by AXAFA inclusion and exclusion criteria. All patients received pulmonary vein isolation. In a subset of patients, one or more additional techniques were performed: linear lesion Introduction (N = 11), mitral isthmus line (N = 3), left atrial roof lesion (N = 9), coro- nary sinus ablation (N = 1), ablation of complex fragmented atrial electro- Catheter ablation has become an established rhythm control therapy grams (N = 20), and ablation left atrial appendage (N = 1). This data were in patients with symptomatic atrial fibrillation (AF).1,2 Freedom from not considered for further statistical analysis due to low numbers. AF and maintenance of sinus rhythm at 1-year post-ablation can be Treatment decisions, as energy type or ablation technique, was at discre- tion of the treating physician. In 19 patients, a cavotricuspid ablation was achieved in 50–70% of patients undergoing catheter ablation. Thus, a performed in addition to pulmonary vein isolation. substantial number of patients would require repeat ablation due to In a subset of patients with available electrocardiogram (ECG) symptomatic recurrent AF.2 Identification of patient groups at higher recordings in sinus rhythm, we analysed the association of P-wave risk for developing recurrent AF can help in developing preventive duration, inter-atrial block, and P-wave morphology at baseline strategies, and tailoring of rhythm control therapy post-ablation. (n = 383), pre-ablation (n = 355), and at 3 months of follow-up Many AF recurrences occur early after AF ablation and terminate (n = 496) with recurrent AF. spontaneously without further interventional treatment.3 While many patients with early recurrences of AF do not develop recurrent AF beyond the so-called ‘blanking period’, early recurrences have Electrocardiogram analysis Digital 12-lead ECGs recorded at baseline, at the index ablation, and been shown to be associated with late and symptomatic recurrences R follow-up after 3 months were analysed with GeoGebra 5.0V software. of AF.3–9 Digital calipers were used to measure P-wave duration, determined from Therefore, the aim of this study was to identify the clinical the earliest to the last P-wave signal in any frontal lead. All measurements and procedural-related factors associated with the first recur- were performed by a blinded assessor and corroborated by a second ex- rences of AF within the first 3 months following ablation in the pert observer when there was ambiguity. multicenter AXAFA–AFNET 5 trial.3,10 Because signs of advanced P-wave morphology was assessed by observation of the inferior leads atrial remodelling such as dilated atria or prolonged P-wave du- (II, III, and aVF).13,14 Interatrial block (IAB) was defined as bimodal (posi- ration have been associated with new-onset AF, we analysed the tive/negative) P-wave morphology with an initial upward deviation from association of P-wave characteristics in a subset of patients for the baseline and a final downward deviation towards the baseline recurrence of AF.11 (Supplementary material online, Figure S2). Normal atrial conduction was considered as a P-wave duration of 120 ms regardless of morphology; partial IAB was defined as a P-wave >_120 ms without bimodal morphol- ogy; and advanced IAB was defined as a P-wave >_120 ms with bimodal Methods morphology.15 Study population AXAFA was a prospective multicentre randomized parallel-group design, Statistical methods comparing apixaban to vitamin K antagonist (VKA) in a cohort of AF The primary outcome, recurrence of AF within the first 3 months after patients undergoing catheter ablation.12 Participants were randomized ablation, was defined as any episode of AF (paroxysmal, persistent, long- 1:1 to VKA or apixaban therapy at baseline, stratified by investigation site lasting, or permanent AF) or atrial flutter, which lasted for at least 30 s, and type of AF (paroxysmal vs. persistent/long-standing persistent). occurring within the 3 months blanking period according to the 2017 Additional drug treatment or procedural decisions (e.g. type of ablation HRS/EHRA/ECAS/APHRS/SOLAECE expert consensus statement.1 The technique) were at the discretion of the treating physician. All patients date of first occurrence of primary outcome was determined by ECG or underwent 24-h Holter monitoring 3 months after ablation. long-term ECG (N = 83), AF symptoms (N = 74), or by other ways The trial was conducted by the steering committee in cooperation (N = 6: implanted devices, self-measurement, unspecified). All patients with Atrial Fibrillation NETwork (AFNET e.V.), Münster, Germany with clinical recurrences of AF were included. Baseline characteristics of (www.kompetenznetz-vorhofflimmern.de) in accordance with the decla- patients with and without recurrent AF within the first 3 months after ab- ration of Helsinki and the International Conference on Harmonization lation were assessed. Good Clinical Practice Guidelines (ICH-GCP), conducted in Europe and For statistical analysis, SPSS (SPSS Statistics version 25, IBM North America and approved by local ethic review boards Corporation, USA) was used. Descriptive statistics for continuous varia- (NCT02227550). The protocol of this analysis was approved by the bles were summarized as means (standard deviations) or medians (25th, steering committee and supported by AFNET. CONSORT statement 75th percentiles), and for categorical variables as counts (percentages), and diagram are described in the original manuscript.10 respectively. Continuous variables were compared using Student’s t-test, Predictors of recurrent AF within 3 months after ablation 1339 or Mann–Whitney U test, after checking for normality using the Kolmogorov–Smirnov test. Categorical variables were compared using Pearson’s v2 test. Clinical, procedural, and therapy-related factors which may influence recurrence of AF were analysed using a univariate Cox model as well as multivariable Cox regression model analysis. Candidate variables were selected using forward selection (entry criterion P = 0.05, removal P = 0.1). Variables considered were randomized treatment (VKA, apixaban), AF type (paroxysmal, persistent), sex (male, female), age (years), interaction of sex and age, height (cm), hypertension, diabetes mellitus, symptomatic heart failure, history of stroke/transient ischaemic attack (TIA), confirmed coronary artery disease, time since AF diagnosis (months), type of ablation technique (cryoballoon, radiofrequency, other), need for cardioversion during procedure, duration of procedure after transseptal puncture (minutes), antiarrhythmic drug usage of beta- blocker, Class IC, or Class III. Kaplan–Meier cumulative risk plots and the Figure 1 Cumulative frequency of first AF recurrence after log-rank tests were used to estimate the proportion of patients free from catheter ablation (n = 163). Vertical black line indicates Day 14. AF after the initial procedure. A two-sided P-value of < 0.05 was consid- Follow-up and first AF recurrence are displayed as median and q1: ered statistically significant. 25th percentile; q3: 75th percentile. AF, atrial fibrillation. Results disease (HR 1.85, 95% CI 1.20–2.86; P = 0.005), the need for car- Study population dioversion during the procedure (HR 1.77, 95% CI 1.26–2.49; We analysed data from 589 patients with a median follow-up period P = 0.001), and an age:sex interaction (HR 1.01, 95% CI 1.00–1.01; of 92 [25th percentile (q1) 86, 75th percentile (q3) 99] days after ab- P = 0.04) indicating higher risk of recurrent AF for young men and lation. Recurrent AF was observed in 163 (28%) patients with a me- older women. dian time to recurrence of 14 (q1 6, q3 43) days. Among the 163 Kaplan–Meier cumulative risk curves indicated that the type of ab- patients with recurrent atrial arrhythmia, 91% (n = 148) had recur- lation technique did not influence AF recurrences (log-rank P = 0.16, rent AF only, 5% (n = 8) had recurrent atrial tachycardia or atrial flut- Figure 2A). Coronary artery disease (log-rank P = 0.02, Figure 2B), his- ter only, and 4% (n = 7) had both. Overall, 48% of recurrences tory of stroke/TIA (log-rank P = 0.01, Figure 2C), and the need for car- occurred within the first 14 days post-ablation (n = 78; Figure 1). dioversion during the ablation procedure (log-rank P < 0.001, Compared to patients without AF recurrence, patients with recur- Figure 2D) were associated with recurrent AF. rent AF were older, more likely to be female, had a history of stroke/ TIA, coronary artery disease, had higher CHA2DS2-VASc scores, re- Interatrial block and P-wave duration quired electrical cardioversion during ablation more often and were In patients who were in sinus rhythm at the time of ablation, the pres- less likely to have used sodium channel blocking antiarrhythmic drugs ence of advanced IAB or P-wave duration shorter of 90 ms in any before ablation (Table 1). After ablation, patients with a recurrence of available ECG (baseline, ablation, and follow-up) was not associated AF reported higher (n = 16), lower (n = 97), or unchanged (n = 42) with recurrent AF (Supplementary material online, Table S2). There EHRA score as compared to symptoms before ablation (missing, was no difference for IAB (partial or advanced) for recurrent AF at n = 8). In total, 56 patients with recurrence of AF had no AF-related baseline, ablation and follow-up. When the ECG at the end of follow- symptoms after ablation (EHRA score = 1). up was analysed (i.e. including patients, who underwent AF ablation in AF), two parameters were associated with recurrent AF. Patients Factors associated with recurrent atrial with recurrent AF showed a higher proportion of P-wave prolonga- fibrillation tion at follow-up compared to the day of ablation [no recurrence: In univariate analyses (adjusted by site, Table 2), age per year [hazard n = 108 (42%) vs. AF recurrence: n = 38 (57%), P = 0.03, ratio (HR) 1.02, 95% confidence interval (CI) 1.00–1.04; P ¼ 0.02], Supplementary material online, Table S2], and a higher absolute P- history of stroke/TIA (HR 1.68, 95% CI 1.02–2.77; P ¼ 0.04), coro- wave duration compared to patients without recurrent AF (No re- nary artery disease (HR 1.53, 95% CI 1.01–2.33; P < 0.001), the currence: 122 ± 22 ms vs. AF recurrence: 129 ± 31 ms, P = 0.03, need for electrical cardioversion during ablation (HR 1.76, 95% CI Supplementary material online, Table S2). 1.26–2.46; P < 0.001), longer duration of ablation after transseptal puncture (steps per minute, HR 1.01, 95% CI 1.00–1.01; P = 0.005) were associated with recurrent AF. Type of ablation technique (with Discussion cryoballoon as reference) did not show a significant association be- tween use of radiofrequency ablation and recurrent AF (HR 2.85, Main findings 95% CI 0.35–23.48; P ¼ 0.33). This analysis of factors associated with recurrent AF in 589 The forward selection process in multivariate analyses identified AXAFA-AFNET 5 patients undergoing a first ablation of AF identi- three predictors associated with increased risk of recurrent AF fied present vascular disease and the need for electrical cardiover- within the first 3 months after ablation (Table 3): coronary artery sion during the ablation procedure as strong predictors for 1340 M.D. Zink et al. Table 1 Characteristics of baseline, AF, ablation, AF recurrence, and follow-up data. No recurrence, AF recurrence, P-value N (%)/mean 6 SD N (%)/mean 6 SD ...................................................................................................................................................................................................................... Patients 426 (72%) 163 (28%) Randomization apixaban group 212 (50%) 87 (53%) 0.43b Sex, male 294 (69%) 97 (60%) 0.03a,b Age (years) 63 ± 9 65 ± 9 0.04a,c Height (cm) 175 ± 10 175 ± 10 0.96c Weight (kg) 88 ± 17 89 ± 18 0.41c BMI (kg/m2) 28.72 ± 4.78 29.02 ± 5 0.47c RR systolic (mmHg) 135 ± 19 137 ± 19 0.37c RR diastolic (mmHg) 80 ± 12 81 ± 12 0.21c Atrial fibrillation, paroxysmal 267 (63%) 91 (56%) 0.13b AF diagnosis (months) 56 ± 67 58 ± 63 0.17c EHRA (baseline) IIa 119 (30%) 35 (23%) 0.14b IIb 141 (35%) 49 (32%) III 129 (32%) 64 (42%) IV 9 (2%) 5 (3%) NYHA No 295 (69%) 102 (63%) 0.12b I 44 (10%) 22 (14%) II 79 (19%) 31 (19%) III 8 (2%) 8 (5%) Prior cardioversion (pharmacological or electrical) 279 (66%) 114 (70%) 0.31b Age 65–74 years 150 (35%) 74 (45%) 0.02a,b Older than 74 years 36 (9%) 16 (10%) 0.6b History of stroke /TIA 26 (6%) 20 (12%) 0.01a,b Hypertension 385 (90%) 144 (88%) 0.47b Diabetes mellitus 43 (10%) 24 (15%) 0.11b Symptomatic heart failure 95 (22%) 38 (23%) 0.79b Coronary artery disease 44 (10%) 29 (18%) 0.01a,b Cardiomyopathy 51 (12%) 22(13%) 0.62b CHA2DS2-VASc score 1.4 ± 0.7 1.6 ± 0.8 0.03a,c Valvular disease 52 (12%) 13 (8%) 0.14b COPD 26 (6%) 10 (6%) 0.99b Heart rate (index) (/min) 68 ± 19 73 ± 23 0.1c QRS (ms) 97 ± 16 98 ± 19 0.88c PQ (ms) 177 ± 33 181 ± 36 0.3c QTc (ms) 435 ± 38 436 ± 47 0.93c Hospitalization (days) 2.3 ± 1.6 2.6 ± 2.9 0.31c Hospitalization after ablation (days) 1.6 ± 1.1 1.8 ± 2.1 0.43c Duration of ablation (min) 146 ± 57 147 ± 55 0.73c Duration after trans septal puncture (min) 120 ± 49 121 ± 51 0.96c Type of ablation technique Cryoballoon 126 (30%) 49 (30%) 0.11b Radiofrequency 263 (62%) 108 (66%) Other 37 (9%) 6 (4%) Fluoroscopy time (min) 14.9 ± 11.8 15.1 ± 11.6 0.69c Heparin total dose (IU) 12849 ± 5792 12004 ± 5588 0.1c Need for cardioversion during procedure 111 (26%) 68 (42%) <0.001a,b Antihypertensive medication 250 (59%) 104 (64%) 0.26b Sodium channel blocker 105 (25%) 26 (16%) 0.02a,b Antiarrhythmic class III 98 (23%) 28 (17%) 0.12b Continued Predictors of recurrent AF within 3 months after ablation 1341 Table 1 Continued No recurrence, AF recurrence, P-value N (%)/mean 6 SD N (%)/mean 6 SD ...................................................................................................................................................................................................................... Beta blockers 305 (72%) 113 (69%) 0.59b Diuretics 137 (32%) 61 (37%) 0.23b Statin 159 (37%) 55 (34%) 0.42b Total amount of drugs 4 ± 2.2 3.9 ± 2.1 0.45c CAD, history of coronary arteria disease; CABG, coronary arterial bypass graft; FU, measurement at 3-month follow-up period; SD, standard deviation; TIA, transient ischaemic attack. a P-value <0.05. b Pearson’s v2 test. c Mann–Whitney U test. Table 2 Univariate analysis for recurrence of AF within the first 3 months after ablation Reference category (unit) Hazard ratio 95% CI P-value .................................................................................................................................................................................................................... Randomization group, VKA 0.91 0.67–1.25 0.57 AF type, persistent 1.26 0.89–1.77 0.19 Age (years) 1.02 1.00–1.04 0.02a Sex, female 1.33 0.95–1.86 0.10 Age (years)  sex (female) 1.01 1.00–1.01 0.06 Height (cm) 0.98 0.97–1.00 0.08 Hypertension 1.13 0.67–1.90 0.65 Diabetes mellitus 1.35 0.86–2.11 0.19 Symptomatic heart failure 0.84 0.55–1.27 0.40 History of stroke/TIA 1.68 1.02–2.77 0.04a Coronary artery disease 1.53 1.01–2.33 <0.001a AF diagnosis (months) 1.00 0.99–1.00 0.65 Type of ablation technique Cryoballoon – 0.009a Radiofrequency 2.85 0.35–23.48 0.33 Other 0.46 0.28–0.78 0.004a Need for eCV during procedure 1.762 1.26–2.46 <0.001a Duration of ablation after trans septal puncture (min) 1.01 1.00–1.01 0.005a Betablocker 0.97 0.68–1.37 0.85 Sodium channel blocker 0.77 0.48–1.23 0.27 Antiarrhythmic drug Class III 0.69 0.45–1.07 0.09 Hazard ratios were quantified for potential clinical and procedural factors associated with recurrence of AF. Data were adjusted for ‘site’ of ablation. AF, atrial fibrillation; CI, confidence interval; eCV, electrical cardioversion; TIA, transient ischaemic attack; VKA, vitamin K antagonist. a Marks significant factors. recurrence of AF after AF ablation. Notably, half of all observed AF TIA, a higher CHA2DS2-VASc score and more often underwent car- recurrences occurred within the first 14 days after ablation. dioversion during catheter ablation (Table 1). Although limited to analysis of P-wave duration at the end of Our finding that 28% of all patients had recurrent AF is in the range follow-up, a longer P-wave duration appeared to be associated of reported recurrences in the first 3 months after ablation (19– with recurrent AF after ablation. 56%).8,9,16 The wide range can be explained due to varying AF definitions, observing intervals and efforts for AF diagnosis, and may include a temporal effect related to improved ablation techniques Baseline findings and timely course of like pressure sensitive and irrigated ablation catheters or second- atrial fibrillation generation cryoballoon in our cohort. The improvement in AF- In our cohort of AF patients after first ablation with at least one risk related symptoms in our cohort for patients with recurrence of AF is factor for stroke, patients with recurrent AF were older, more often in line with previous reported improvement of symptoms17 and qual- female, had more often coronary artery disease, a history of stroke/ ity of life.18 1342 M.D. Zink et al. Table 3 Multivariate analysis (Cox regression, adjusted for ‘site’ of ablation) for recurrence of AF within the first 3 months after ablation Hazard ratio 95% CI P-value .................................................................................................................................................................................................................... History of stroke/TIA 1.54 0.93–2.56 0.1 Coronary artery disease 1.85 1.20–2.86 0.005a Need for cardioversion during procedure 1.77 1.26–2.49 0.001a Age (years)  sex (female) 1.01 1.00–1.01 0.04a Variables included in the forward selection are randomization (VKA, apixaban), AF type (paroxysmal, persistent), sex (female), age (steps per year), interaction of sex (female) and age (steps per year), height (steps in cm), hypertension, diabetes mellitus, symptomatic heart failure, history of stroke/TIA, confirmed coronary artery disease, AF knowl- edge (steps per months), type of ablation energy used (cryoballoon, radiofrequency, other), need for cardioversion during procedure, duration of procedure after transseptal puncture (steps in min), antiarrhythmic drug usage of betablocker, Class IC, or Class III. AF, atrial fibrillation; TIA, transient ischaemic attack; VKA, vitamin K antagonist. a Marks significant factors. A B C D Figure 2 Kaplan–Meier cumulative risk plots of (A) type of ablation energy used, (B) coronary artery disease, (C) history of stroke/TIA, and (D) the need for cardioversion during procedure. AF, atrial fibrillation; TIA, transient ischaemic attack. Predictors of recurrent AF within 3 months after ablation 1343 Predictors of recurrence of atrial fibrillation measured by imaging techniques or 12-lead ECG. We analysed re- The interaction of higher age and female sex, history of stroke/TIA, peatedly recorded P-wave parameters and ECG sign for IAB as sur- coronary artery disease, and the need for cardioversion during pro- rogate marker for atrial remodelling and its association to recurrence cedure were predictors for recurrence of AF. Comparable risk of AF. We found longer P-waves in patients with recurrent AF in our factors were reported before for structural heart disease,7 and car- cohort. Due to the high number of patients undergoing the proce- dioversion during procedure.6 According to previous studies, several dure in AF, this difference was only found when analyzing ECGs at factors failed in prediction for recurrent AF in our cohort like hyper- the end of the 3 months of follow-up (Supplementary material online, tension,6 AF type,7 age, or the duration of ablation.16 In addition to Table S2). There is evidence for a change of P-wave indices due to vascular disease as a marker for an atrial cardiomyopathy, the need surgical ablation procedure as effect of reverse electrical atrial for cardioversion during procedure may be considered as indication remodelling.21 However, prolonged P-wave duration and increased for an advanced substrate of maintaining AF. dispersion have been shown association with recurrent AF after AF Procedural effects like type of ablation technique used and dura- ablation.22,23 Furthermore, prolonged P-wave duration is known to tion of ablation showed no differences in comparison of the groups be associated with enlarged atria,15 atrial conduction disturbances and but an effect on rate of AF recurrence in univariate analysis while in associated with AF development.11 Based on our data, the presence multivariate analysis failed to do so. Similar findings for duration of ab- of advanced IAB as a predictor for new-onset atrial fibrillation,11 did lation were reported before.3,19 The fact that type of AF, type of abla- not add to the prediction of recurrent AF within the first 3 months af- tion and procedural time showed no association to primary ter ablation but should be seen as hypothesis generating. Upcoming outcome, can be explained, as the treatment was at discretion of investigation with primary outcome for AF recurrence in context of treating physicians. Thus, ablation was performed as needed for indi- advanced IAB and extended follow-up period are warranted. vidual success. Therefore, it is not surprising that a well-chosen abla- tion strategy also leads to comparable outcome for recurrence of AF Limitations in advanced types of AF as with early stages of AF. Independently, the The data obtained from this multicentre study conducted in need for cardioversion shows a significant association to recurrence European and North America was collected in a structured and pre- of AF. This could be a sign of more advanced atrial remodelling and defined way. Unfortunately, the data lack information on biomarkers probably more ablation is needed, because an AF trigger or a not ab- and imaging data which might add more information for atrial struc- lated substrate still maintains the ongoing arrhythmia. The results for tural remodelling. Continuous rhythm monitoring in this cohort was ablation technique used, is in line with the randomized trials compar- not provided after ablation, therefore reported rate of recurrence is ing the two main energy sources (cryoballoon ablation and radiofre- related to clinical AF, because asymptomatic episodes could be quency ablation)20 and showed in the investigated cohort no missed by chance. Repeated ECG recording is available before and af- significant difference. Notably, because used therapy was at the dis- ter the procedure, and in the follow-up period. Therefore, detailed cretion of the treating physician this cohort consists of rarely used 12-lead ECG analysis was performed as a surrogate marker for atrial types of ablation techniques (PVAC N = 3, Laser ablation N = 42). In electrical remodelling. There is little evidence for association of pre- Kaplan–Meier cumulative risk plot a lower dynamic of AF recur- dictors of recurrent AF within the first 3 months with later recur- rences compared to cryoballoon and radiofrequency ablation was rence. Future studies on predictors of early AF recurrences are revealed for laser ablation technique. therefore certainly warranted. As because many patients presented in AF, the P-wave analysis and Potential clinical implications advanced IAB was not available in all patients. Although, P-wave analysis points towards an association of recurrent AF with advanced Given the available data for recurrence of AF in comparable investiga- IAB the analysis is potentially biased by analyzing the healthier tions with smaller cohorts, this investigation allows higher confidence patients in our cohort and underestimates the association compared in identified predictors for recurrent AF within the first 3 months af- to existing evidence,11,13,14,24–26 because in particular higher degree ter catheter ablation. While vascular disease was previously de- of atrial remodelling is probably associated with higher chance of pre- scribed as a risk factor for recurrent AF, our data identified as novel senting in AF at any time point. As a matter of fact, advanced IAB was aspect the need for cardioversion during procedure as a potential a marker of AF much later than 3 months in previous investigations.11 and easily to assessable predictor for recurrent AF after a first AF ab- Data of this study were obtained for a comparison of Apixaban vs. lation. Prolonged P-wave duration at the end of the blanking period VKA, thus the primary objective and hypothesis of this retrospective in patients with recurrent AF might as well be considered as a marker analysis was not defined a priori and no placebo control was included; of increased risk of recurrent AF. Recurrence of AF within the blank- therefore, our results should be considered as descriptive and not as ing period is a strong and independent predictor of long-term abla- a confirmatory investigation. Because no 1-year follow-up data is tion success. Therefore, our finding that half of patients with early available, a comparison of long-term ablation success, linked to recur- recurrent AF have their first recurrence within Day 1 to Day 14 after rence within the blanking period is not possible. ablation, followed a robust recurrent AF frequency until 3 months after AF ablation, the blanking period of 3 months might be reconsidered. Conclusion Electrocardiogram parameter Atrial remodelling is independently associated with new-onset AF Half of all early AF recurrences occurred within the first 14 days after and higher number of AF recurrences after ablation and can be first ablation. Ablation technique source, duration of ablation, 1344 M.D. Zink et al. hypertension, type of AF, diabetes mellitus, or antiarrhythmic drug 5. Andrade JG, Khairy P, Macle L, Packer DL, Lehmann JW, Holcomb RG et al. Incidence and significance of early recurrences of atrial fibrillation after cryobal- usage showed no effect on recurrence of AF within the blanking pe- loon ablation: insights from the multicenter Sustained Treatment of Paroxysmal riod. Present vascular disease (coronary artery disease or stroke/ Atrial Fibrillation (STOP AF) trial. Circ Arrhythm Electrophysiol 2014;7:69–75. TIA) and the need for cardioversion during the procedure are strong 6. Themistoclakis S, Schweikert RA, Saliba WI, Bonso A, Rossillo A, Bader G et al. Clinical predictors and relationship between early and late atrial predictors for recurrent AF within the blanking period. P-wave dura- tachyarrhythmias after pulmonary vein antrum isolation. Heart Rhythm 2008; tion at follow-up was longer in patients with recurrent AF in the first 5:679–85. 3 months after ablation. 7. Bertaglia E, Stabile G, Senatore G, Zoppo F, Turco P, Amellone C et al. Predictive value of early atrial tachyarrhythmias recurrence after circumferential anatomical pulmonary vein ablation. Pacing Clin Electrophysiol 2005;28:366–71. 8. Takayama K, Hachiya H, Iesaka Y, Hirao K, Isobe M. Early recurrence after long- Supplementary material standing persistent atrial fibrillation ablation. Int Heart J 2018;59:321–7. 9. Oral H, Knight BP, Ozaydin M, Tada H, Chugh A, Hassan S et al. Clinical signifi- Supplementary material is available at Europace online. cance of early recurrences of atrial fibrillation after pulmonary vein isolation. J Am Coll Cardiol 2002;40:100–4. 10. Kirchhof P, Haeusler KG, Blank B, De Bono J, Callans D, Elvan A et al. Apixaban Funding in patients at risk of stroke undergoing atrial fibrillation ablation. Eur Heart J AXAFA–AFNET 5 is an investigator-initiated trial. Sponsor of the trial is 2018; doi:10.1093/eurheartj/ehy176. AFNET. AXAFA–AFNET 5 was partially funded by BMS/Pfizer, the 11. Enriquez A, Sarrias A, Villuendas R, Ali FS, Conde D, Hopman WM et al. New-onset atrial fibrillation after cavotricuspid isthmus ablation: identification of DZHK (German Centre for Cardiovascular Research), and by the BMBF advanced interatrial block is key. Europace 2015;17:1289–93. (German Ministry of Education and Research) to AFNET. Furthermore, 12. Di Biase L, Callans D, Hæusler KG, Hindricks G, Al-Khalidi H, Mont L et al. this work was supported by the European Union (CATCH ME: Rationale and design of AXAFA-AFNET 5: an investigator-initiated, randomized, Characterizing Atrial fibrillation by Translating its Causes into Health open, blinded outcome assessment, multi-centre trial to comparing continuous Modifiers in the Elderly, No. 633196, the ITN Network AFibTrainNet, apixaban to vitamin K antagonists in patients undergoing atrial fibrillation cathe- ter ablation. Europace 2017;19:132–8. No. 675351, and the ERACoSysMED H2020 ERA-NET Cofund project 13. Wu JT, Zhao DQ, Li FF, Wu R, Fan XW, Hu GL et al. Advanced interatrial block Systems medicine for diagnosis and stratification of atrial fibrillation), predicts recurrence of atrial fibrillation after accessory pathway ablation in British Heart Foundation (FS/13/43/30324 to P.K.), Leducq Foundation to patients with Wolff-Parkinson-White syndrome. Clin Cardiol 2019;42:806–11. P.K., and the Netherlands Heart Foundation (RACE V) to U.S. and I.v.G. 14. Bayes de Luna A, Escobar-Robledo LA, Aristizabal D, Weir Restrepo D, Mendieta G, Masso van Roessel A et al. Atypical advanced interatrial blocks: defi- AFNET, DZHK, BMS/Pfizer; AXAFA–AFNET 5; NCT02227550. nition and electrocardiographic recognition. J Electrocardiol 2018;51:1091–3. 15. Bayes de Luna A, Platonov P, Cosio FG, Cygankiewicz I, Pastore C, Baranowski Conflict of interest: U.S. received honoraria from Johnson & Johnson R et al. Interatrial blocks. A separate entity from left atrial enlargement: a consen- and is consultant of EP solution. He is scientific director and shareholder of sus report. J Electrocardiol 2012;45:445–51. YourRhythmics BV. He received a research grant from Medtronic. M.D.Z. 16. Chang SL, Tsao HM, Lin YJ, Lo LW, Hu YF, Tuan TC et al. Characteristics and received speaker/fees from BMS Pfizer, Educational Grant DGK (German significance of very early recurrence of atrial fibrillation after catheter ablation. J Society of Cardiology sponsored by St. Jude Medical), educational support Cardiovasc Electrophysiol 2011;22:1193–8. 17. Bjorkenheim A, Brandes A, Magnuson A, Chemnitz A, Svedberg L, Edvardsson N by Biotronik. K.G.H. reports speaker fees/consultant honoraria from Bayer. et al. Assessment of atrial fibrillation-specific symptoms before and 2 years after Boehringer Ingleheim, BMS/Pfizer, Daiichi Sankyo, Medtronik, Biotronik and atrial fibrillation ablation: do patients and physicians differ in their perception of Edwards Lifesciences, and study grants from Bayer and Sanofi. P.K. received symptom relief? JACC Clin Electrophysiol 2017;3:1168–76. research support for basic, translational, and clinical research projects from 18. Kuck KH, Furnkranz A, Chun KR, Metzner A, Ouyang F, Schluter M et al. Cryoballoon or radiofrequency ablation for symptomatic paroxysmal atrial fibril- the European Union, British Heart Foundation, Medical Research Council lation: reintervention, rehospitalization, and quality-of-life outcomes in the FIRE (UK), and German Centre for Cardiovascular Research, from several drug AND ICE trial. Eur Heart J 2016;37:2858–65. and device companies active in atrial fibrillation and has received honoraria 19. Budzianowski J, Hiczkiewicz J, Burchardt P, Pieszko K, Rzezniczak J, Budzianowski from several such companies in the past. P.K. is listed as inventor on two P et al. Predictors of atrial fibrillation early recurrence following cryoballoon ab- lation of pulmonary veins using statistical assessment and machine learning algo- patents held by University of Birmingham (Atrial Fibrillation Therapy WO rithms. Heart Vessels 2019;34:352–9. 2015140571, Markers for Atrial Fibrillation WO 2016012783). All other 20. Kuck KH, Brugada J, Furnkranz A, Metzner A, Ouyang F, Chun KR et al. authors report no conflict of interest. Cryoballoon or radiofrequency ablation for paroxysmal atrial fibrillation. N Engl J Med 2016;374:2235–45. 21. Nassif M, Krul SP, Driessen AH, Deneke T, Wilde AA, de Bakker JM et al. Data availability Electrocardiographic P wave changes after thoracoscopic pulmonary vein isola- Access to data (raw and derived) is regulated by the AXAFA steering tion for atrial fibrillation. J Interv Card Electrophysiol 2013;37:275–82. committee. For a reasonable request please contact the corresponding 22. Jiang H, Lu Z, Lei H, Zhao D, Yang B, Huang C. Predictors of early recurrence author. and delayed cure after segmental pulmonary vein isolation for paroxysmal atrial fibrillation without structural heart disease. J Interv Card Electrophysiol 2006;15: 157–63. References 23. O’Donnell D, Furniss SS, Dunuwille A, Bourke JP. Delayed cure despite early re- 1. Calkins H, Hindricks G, Cappato R, Kim YH, Saad EB, Aguinaga L et al.; currence after pulmonary vein isolation for atrial fibrillation. Am J Cardiol 2003; Document Reviewers. 2017 HRS/EHRA/ECAS/APHRS/SOLAECE expert con- 91:83–5. sensus statement on catheter and surgical ablation of atrial fibrillation. Europace 24. Wu JT, Long DY, Dong JZ, Wang SL, Fan XW, Yang HT et al. Advanced intera- 2018;20:e1–160. trial block predicts clinical recurrence of atrial fibrillation after catheter ablation. 2. Kirchhof P, Benussi S, Kotecha D, Ahlsson A, Atar D, Casadei B et al. 2016 ESC J Cardiol 2016;68:352–6. Guidelines for the management of atrial fibrillation developed in collaboration 25. Escobar-Robledo LA, Bayes-de-Luna A, Lupon J, Baranchuk A, Moliner P, with EACTS. Europace 2016;18:1609–78. Martinez-Selles M et al. Advanced interatrial block predicts new-onset atrial fi- 3. Andrade JG, Khairy P, Verma A, Guerra PG, Dubuc M, Rivard L et al. Early recur- brillation and ischemic stroke in patients with heart failure: the “Bayes’ rence of atrial tachyarrhythmias following radiofrequency catheter ablation of Syndrome-HF” study. Int J Cardiol 2018;271:174–80. atrial fibrillation. Pacing Clin Electrophysiol 2012;35:106–16. 26. Baranchuk A, Yeung C. Advanced interatrial block predicts atrial fibrillation re- 4. Nalliah CJ, Lim TW, Kizana E, Qian P, Kovoor P, Thiagalingam A et al. Clinical sig- currence across different populations: learning Bayes syndrome. Int J Cardiol nificance of early atrial arrhythmia type and timing after single ring isolation of 2018;272:221–2. the pulmonary veins. Europace 2015;17:1038–44.