Outcomes of surgical ablation for atrial fibrillation in on- versus off-pump coronary artery bypass grafting

Abstract

OBJECTIVES

A considerable number of patients undergoing coronary artery bypass grafting surgery suffer from atrial fibrillation and should be treated concomitantly. This manuscript evaluates the impact of on-pump versus off-pump bypass grafting on the applied lesion set and rhythm outcome.

METHODS

Between January 2017 and April 2020, patients who underwent combined bypass grafting and surgical ablation for atrial fibrillation were consecutively enrolled in the German CArdioSurgEry Atrial Fibrillation registry (CASE-AF, 17 centres). Data were prospectively collected. Follow-up was planned after one year.

RESULTS

A total of 224 patients were enrolled. No differences in baseline characteristics were seen between on- and off-pump bypass grafting, especially not in type of atrial fibrillation and left atrial size. In the on-pump group (n = 171, 76%), pulmonary vein isolation and an extended left atrial lesion set were performed more often compared to off-pump bypass grafting (58% vs 26%, 33 vs 9%, respectively, P < 0.001). In off-pump bypass grafting a box isolating the atrial posterior wall was the dominant lesion (72% off-pump vs 42% on-pump, P < 0.001). Left atrial appendage management was comparable in on-pump versus off-pump bypass grafting (94% vs 91%, P = 0.37). Sinus rhythm at follow-up was confirmed in 61% in the on-pump group and in 65% in the off-pump group (P = 0.66). No differences were seen in in-hospital or follow-up complication-rates between the two groups.

CONCLUSIONS

In coronary artery bypass grafting patients undergoing concomitant atrial fibrillation ablation, our data suggests that the technique applied for myocardial revascularization (off-pump vs on-pump) leads to differences in the ablation lesion set, but not in safety and effectiveness.

Recent Medicare data revealed a remarkable prevalence of atrial fibrillation (AF) of 28.4% in patients undergoing cardiac surgery.

Graphical Abstract

INTRODUCTION

Recent Medicare data revealed a remarkable prevalence of atrial fibrillation (AF) of 28.4% in patients undergoing cardiac surgery. While the prevalence is higher in mitral valvular disease as compared to coronary artery disease (CAD), it still affects a considerable number of patients undergoing coronary artery bypass grafting (CABG) surgery with up to 20% [1]. These patients are exposed to a substantial risk of stroke, heart failure development and mortality due to their arrhythmia. Concomitant ablation in CABG surgery is therefore associated with over 25% improved survival [2]. This has been recognized by current guidelines for the treatment of AF, which recommend concomitant ablation in this patient population [3, 4].

While off-pump CABG surgery techniques (i.e. without the use of extracorporeal circulation) showed beneficial outcomes for certain patient subgroups with CAD [5], there should be consensus that the chosen technique or access should not limit the application or comprehensiveness of additional therapies which are also indicated. While extracorporeal circulation and cardioplegic arrest allows for endocardial ablation, the application of cryo- and/or radiofrequency energy, and an extensive lesion set up to a biatrial Maze IV, off-pump CABG surgery procedures are limited to epicardial ablation techniques including pulmonary vein isolation (PVI) or a box lesion isolating the PVs and the posterior left atrial (LA) wall. Further, the epicardial approach on the beating heart might reduce the ability to achieve transmural lesions due to the heat sink effect and potentially limits the efficacy of certain energy sources [6].

We expect that comprehensiveness of lesions differ depending on the technique used for myocardial revascularization in patients undergoing concomitant AF ablation, which could potentially impact rhythm outcome in terms of sinus rhythm (SR) restoration. We therefore performed a subgroup analysis of data provided by the nationwide, prospective, observational, multicentre German CArdioSurgEry Atrial Fibrillation Registry (CASE-AF). The aim was to understand the impact of the technique used for myocardial revascularization on the applied lesion set and rhythm outcome in patients undergoing concomitant AF ablation.

MATERIALS AND METHODS

Between January 2017 and April 2020, patients who underwent surgical ablation for AF were consecutively enrolled in the CASE-AF register [7–9]. This multicentre, observational, cohort study aims to prospectively collect data on patients undergoing surgical ablation as a stand-alone or concomitant procedure in 17 German cardiac surgery departments. The study was registered into the ClinicalTrials.gov database [ClinicalTrials.gov Identifier: NCT03091452 (https://clinicaltrials.gov/study/NCT03091452)] and was approved by the local ethics committees (Ethik-Kommission der Landesärztekammer Rheinland-Pfalz in Mainz, reference number: 837.536.15 (10304), date of approval: 06.05.2016). The Institute for Heart Attack Research [Institute für Herzinfarktforschung (IHF), Ludwigshafen, Germany] supervised the register. The study was performed in accordance with the Declaration of Helsinki. Data management, monitoring and statistical analyses were performed by the IHF. The classification of AF was adapted from the guidelines of the European Society of Cardiology (ESC) as paroxysmal, persistent, or long-standing persistent [10]. Recurrence was defined as any documented occurrence of AF > 30 s after the blanking period (i.e. the first 3 months after the ablation during which a recurrence of AF is not considered a failure of the procedure).

An electronic case report form on a web-based platform was used for data collection. Data were collected pre-, intra- and post-operatively in the hospital and from a telephone interview after one year. Informed consent was obtained from each patient before the procedure.

The CASE-AF registry is purely descriptive and not confirmative. Categorical variables were presented as absolute numbers and percentages and were compared by the Pearson chi-squared or Fisher’s exact test. The normality of distribution of continuous variables was tested using the Shapiro–Wilk test and Quantil-Quantil-Plot. Continuous variables were presented as median and interquartile range (IQR) or mean and standard deviation and were compared by Mann–Whitney–Wilcoxon test. All tests were 2-tailed, and P-values <0.05 were considered statistically significant. P-values presented in tables should be interpreted in a descriptive sense as measure of association between parameters. Analysis were performed using SAS statistical software package, version 9.4 (Cary, NC, USA).

RESULTS

A total of 1000 patients were enrolled in the register, of which 224 patients underwent combined CABG surgery and surgical ablation. Extracorporeal circulation was used in 171 (76%) patients and 53 (24%) patients were operated off-pump. Baseline characteristics are summarized in Table 1, showing no significant differences between the two groups, especially not in type of AF, LA size and pre-operative treatment.

Table 1:

Baseline characteristics.

	On-pump (n = 171) percentage (number), Mean ± SD		Off-pump (n = 53) percentage (number), Mean ± SD		P-value
Patient characteristics
Male	84	(143/171)	87	(46/53)	0.58
Age (years)	71.1 ± 7.4	n = 170	69.2 ± 7.9	n = 53	0.15
BMI (kg/m2)	28.6 ± 4.6	n = 171	28.7 ± 5.6	n = 53	0.79
COPD	12	(21/170)	11	(6/53)	0.84
Stroke	9	(16/170)	15	(8/53)	0.24
TIA	3	(5/170)	8	(4/53)	0.22
Diabetes mellitus	36	(62/170)	30	(16/53)	0.40
Hypertension	88	(149/170)	87	(46/53)	0.87
OSAS	3	(5/170)	6	(3/53)	0.40
Kidney disease	22	(38/170)	19	(10/53)	0.59
Thyroid disorder	8	(14/170)	13	(7/53)	0.29
Prior MI	19	(32/170)	32	(17/53)	0.042
CHA2DS2-Vasc Score	3.5 ± 1.5	n = 151	3.7 ± 1.6	n = 49	0.27
HAS-BLED Score	2.6 ± 1.0	n = 169	2.5 ± 1.0	n = 53	0.94
LAD (mm)	46 ± 9	n = 136	43 ± 7	n = 28	0.17
LVEF (%)	51 ± 11	n = 171	50 ± 12	n = 51	0.49
LAA-Thrombus	10	(11/113)	4	(2/47)	0.35
AF characteristics
Paroxysmal AF	68	(116/171)	70	(37/53)	0.79
Persistent AF	18	(31/171)	19	(10/53)	0.90
Long-standing persistent AF	14	(24/171)	11	(6/53)	0.61
EHRA-Score I-IIa	65	(72/110)	58	(19/33)	0.41
EHRA-Score IIb-IV	35	(38/110)	42	(14/33)	0.41
Number of AAD used	1 ± 1	n = 171	1 ± 1	n = 53	0.12
Refractory to amiodarone	6	(5/85)	9	(3/32)	0.68
Electrical cardioversions	0 ± 1	n = 171	0 ± 1	n = 53	0.38
Prior ablation attempt	4	(6/162)	6	(3/53)	0.69
Oral anticoagulation	82	(140/171)	85	(45/53)	0.61

AAD: antiarrhythmic drugs; AF: atrial fibrillation; BMI: body mass index; COPD: chronic obstructive pulmonary disease; LAA: left atrial appendage; LAD: left atrial diameter; LVEF: left ventricular ejection fraction; MI: myocardial infarction; OSAS: obstructive sleep apnoea syndrome; TIA: transient ischaemic attack. Bold values denote statistical significance at the p < 0.05 level.

Intraoperative variables are shown in Table 2. Due to the different approaches, the two groups were substantially different for almost all aspects of the ablation techniques. In the off-pump group, all ablations were conducted epicardially, whereas, in the on-pump group, epi- and endocardial ablation was performed (P < 0.001). All off-pump ablations used exclusively radiofrequency energy, while in the on-pump group radiofrequency and cryoablation, or a combination of both were used (P = 0.001). The concept of ablation lines showed significant differences as well: in the on-pump group, PVI and additional LA linear lines, implying an extensive left atrial lesion set, were performed more often than in off-pump CABG surgery (58% vs 26% and 33% vs 9%, respectively, P < 0.001). In contrary, a box lesion isolating the PVs and the LA posterior wall was the dominant lesion set in off-pump CABG surgery (72% vs 42% P < 0.001). However, the application of right atrial lines (8% on-pump vs 4% off-pump, P = 0.40) and exclusion of the left atrial appendage (LAA, 94% on-pump vs 91% off-pump, P = 0.37) were comparable in both groups.

Table 2:

Periprocedural data.

	On-pump (n = 171) percentage (number)		Off-pump (n = 53) percentage (number)		P-value
Re-do procedure	1	(2/171)	0	(0/53)	1.00
Energy emitted
Endocardial	24	(41/171)	0	(0/53)	< 0.001
Epicardial	81	(139/171)	100	(53/53)	< 0.001
Energy source
Radiofrequency	83	(141/170)	100	(52/52)	0.001
Cryoablation	19	(32/170)	0	(0/52)	< 0.001
Ablation lines
Pulmonary vein isolation	58	(88/153)	26	(12/47)	< 0.001
Box isolation	42	(64/153)	72	(34/47)	< 0.001
Left atrial lines	33	(51/153)	9	(4/47)	< 0.001
Right atrial lines	8	(12/153)	4	(2/47)	0.40
LAA management	94	(160/170)	91	(48/53)	0.37

AF: atrial fibrillation; LAA: left atrial appendage. Bold values denote statistical significance at the p < 0.05 level.

As shown in Table 3, the overall rate of in-hospital major cardiac and cerebrovascular events (MACCE) tended to be numerically higher in the on-pump group without significant difference (4% vs 2%, P = 1.00). The same accounts for the implantation rate of pacemaker and ICD (3% vs 0%, P = 1.00, 1% vs 0%, P = 1.00). Postoperative usage of antiarrhythmic drugs (AAD) Class III was significantly higher in the on-pump group (46% vs 21%, P = 0.001), not leading to differences in postoperative cardioversion (13% vs 15%, P = 0.71) or in SR at discharge (69% vs 69%, P = 0.94).

Table 3:

In-hospital outcomes.

	On-pump (n = 171) percentage (number)		Off-pump (n = 53) percentage (number)		P-value
MACCE	4	(6/168)	2	(1/53)	1.00
Mortality	2	(4/168)	2	(1/53)	1.00
Stroke	1	(2/168)	0	(0/53)	1.00
TIA	0	(0/168)	0	(0/53)	–
AV-block 3rd degree	2	(3/168)	0	(0/53)	1.00
Sinus arrest	2	(3/168)	0	(0/53)	1.00
Cardioversion	13	(22/168)	15	(8/53)	0.71
Pacemaker	3	(5/168)	0	(0/53)	1.00
ICD	1	(1/168)	0	(0/53)	1.00
Medication at discharge
Oral anticoagulation	84	(139/165)	74	(39/53)	0.08
AAD class I/III	46	(76/165)	21	(11/53)	0.001
Amiodarone	97	(71/73)	100	(11/11)	1.00
Betablocker	79	(130/165)	79	(42/53)	0.94
Digitalis	1	(1/165)	6	(3/53)	0.045
Rhythm at discharge
Sinus rhythm	69	(112/163)	69	(36/53)	0.94
Atrial fibrillation	22	(36/163)	29	(15/52)	0.32
Other	9	(15/163)	2	(1/52)	0.08

AAD: antiarrhythmic drugs; AV: atrioventricular; ICD: intracardiac defibrillator; MACCE: major adverse cardiac and cerebrovascular events; TIA: transient ischaemic attack. Bold values denote statistical significance at the p < 0.05 level.

Follow-up was completed in 99% [on-pump, median follow-up 410 days (381, 543)] and 94% [off-pump, median follow-up 516 days (446, 618)] of patients (Table 4). SR at follow-up was documented in 61% in the on-pump group, and in 65% in the off-pump group [P = 0.66, adjusted OR 0.83 (0.41–1.67)]. Fifty-four percent of the on-pump patients vs 60% of the off-pump patients were in SR off AAD Class I or III [P = 0.41, adjusted OR 0.74 (0.37—1.46)] (Fig. 1). SR was restored in 76% in paroxysmal AF in the off-pump cohort, and in 65% in the on-pump cohort (P = 0.24). In non-paroxysmal AF patients, off-pump CABG surgery revealed 40% SR versus 53% in the on-pump group (P = 0.38) (Fig. 2). However, these differences did not reach significance level. Readmissions due to AF (18%, P = 0.11) and re-ablations in the later time course (3%, P = 0.59) occurred only in the on-pump group.

Table 4:

Follow-up.

	On-pump (n = 171) percentage (number), Median (IQR)		Off-pump (n = 53) percentage (number), Median (IQR)		P-value
Follow-up completed	99	(170/171)	94	(50/53)	0.042
Time to follow-up (days)	410	(381, 543)	516	(446, 618)	<0.001
Mortality	8	(13/170)	6	(3/50)	1.00
Method of follow-up
ECG	79	(119/151)	91	(41/45)	0.06
Holter	28	(42/151)	31	(14/45)	0.67
7-Day Holter	0	(0/42)	0	(0/14)	–
Device	9	(14/151)	0	(0/45)	0.043
Rhythm
Freedom from AF	61	(99/162)	65	(31/48)	0.66
Freedom from AF off AAD	54	(87/162)	60	(29/48)	0.41
Symptomatic recurrence	14	(19/140)	22	(10/45)	0.16
EHRA IIb-IV	8	(10/119)	2	(1/41)	0.29
Cardioversion	3	(4/147)	7	(3/46)	0.36
Repeated ablation	3	(5/150)	0	(0/45)	0.59
Complications
Stroke	2	(3/153)	2	(1/47)	1.00
TIA	1	(1/151)	0	(0/47)	1.00
Pacemaker	6	(10/156)	2	(1/47)	0.68
ICD	3	(5/156)	2	(1/47)	1.00
Rehospitalization	42	(65/154)	32	(15/47)	0.21
Cardiovascular	25	(16/65)	7	(1/15)	0.28
AF	18	(12/65)	0	(0/15)	0.11
Not cardiovascular	43	(28/65)	40	(6/15)	0.83
Unknown	14	(9/65)	53	(8/15)	0.002
Therapy
Oral anticoagulation	71	(112/157)	72	(34/47)	0.89
AAD Class I/III	12	(19/157)	6	(3/47)	0.34
Amiodarone	92	(12/13)	100	(3/3)	1.00
Betablocker	83	(131/157)	91	(43/47)	0.17
Digitalis	3	(4/157)	2	(1/47)	1.00

AAD: antiarrhythmic drugs; AF: atrial fibrillation; TIA: transient ischaemic attack. Bold values denote statistical significance at the p < 0.05 level.

Figure 1:

Freedom from atrial fibrillation at follow-up. AAD: antiarrhythmic drugs.

Figure 2:

Freedom from atrial fibrillation per type. AF: atrial fibrillation.

DISCUSSION

In this manuscript, we evaluate whether off-pump performed CABG surgery affects the lesion set applied in concomitant AF ablation and if this potential difference would reflect on an inferior rhythm outcome when compared to on-pump CABG surgery procedures. We found no significant differences in SR at follow-up, despite substantial differences in the applied lesion sets. To the best of our knowledge, this is the first study comparing outcomes in patients undergoing concomitant myocardial revascularization and AF ablation, with and without the use of cardiopulmonary bypass.

The importance of concomitant AF ablation is proven and has been recommended by current guidelines [4]. In line with the goal to become less invasive, off-pump strategies for surgical myocardial revascularization have evolved during the last decade to avoid potential consequences of cardioplegic arrest with cross-clamping the aorta and/or extracorporeal circulation. While data on outcome and complications comparing off-pump versus on-pump CABG surgery is still not entirely conclusive [5], this registry showed that the chosen strategy might have substantial impact on the lesion set and its comprehensiveness used in concomitant AF ablation.

The Maze IV lesion sets reflects the gold standard for surgical ablation by isolating the triggers responsible for AF induction, as well as by interrupting potential macro re-entries which sustain AF, and is therefore most effective in all types of AF [11]. However, this lesion set can be hardly reproduced in an off-pump procedure. This reflects also the results of this study, revealing a more extensive lesion being applied mainly in on-pump procedures. Considering this limitation in lesion set, we would have expected inferior rhythm outcome in off-pump procedures. However, the vast majority of patients in both groups presented with paroxysmal AF, in whom the isolation of focal triggers in the vicinity of the PVs is the cornerstone for ceasing the arrhythmia [12, 13]. As such, the ablation strategy of most centres would be PVI, which can be equally effectively applied in on-pump and off-pump procedures. Moreover, the off-pump group of this registry received mainly a box lesion due to the used device-design, which additionally isolates the LA posterior wall and should therefore eliminate even more potential triggers compared to PVI alone. Very few patients received a trigonal or right atrial intercaval line. This beneficial effect might have contributed to the superior overall results in the off-pump group by providing a more effective lesion set for patients with paroxysmal AF, who reflected the majority of the cohort. However, it was less effective in persistent/long-standing persisting AF compared to the more extensive lesion set applied in the on-pump procedures. This underscores that isolation of the PVs might be a solid lesion set for effective rhythm control in patients undergoing CABG surgery procedures. However, additional lines, such as the left atrial isthmus lesion, will contribute to success in patients with more advanced AF, and a deviation from the off-pump ablation strategy should be considered to be able to apply a more comprehensive lesion set.

There might be additional factors contributing to the overall success rate of off-pump ablation. While data is not entirely conclusive, several meta-analyses of studies reporting on the incidence of postoperative AF for off-pump versus on-pump CABG surgery, confirmed a lower occurrence rate when extracorporeal circulation and cardiac arrest are avoided [14, 15]. Reduction of myocardial oedema, inflammatory processes and a more stable electrolyte balance might equalize the limitation in applying a more extensive lesion set in off-pump CABG surgery in the early postoperative period. This might allow for sufficient electrical and anatomical remodelling, which plays a role in mid-term rhythm outcome as well. Also notable, while in on-pump procedures frequently a decision was made between PVI and a more extensive left atrial lesion set including a box lesion, in off-pump surgery a box lesion was the dominant lesion set applied. This might be inferior to a left atrial Maze pattern on the one hand, but superior to PVI on the other.

Further, a fundamental advantage of ablation on a beating heart is the potential of testing for bidirectional block as electrophysiological end-point and lesion verification. While ablation during cardioplegic arrest reduces the heat sink effect caused by intracavity blood flow and thus should increase device efficacy in creating transmural lesion, end-point measurements for the entire lesion set might not be possible and therefore existing gaps can hardly be detected.

Besides differences in lesion sets depending on off-pump or on-pump ablation, the LAA was managed in the same high percentage of patients in both groups. LAA closure in CABG surgery is associated with reduced mortality and thromboembolic events [16]. The LAA is not only the main source of thrombus formation causing ischaemic stroke, but is also a known trigger source for AF [17–19] Moreover, strokes associated with AF are more severe and devastating compared to other etiologies [20]. Oral anticoagulation therapy is effective in preventing thrombus formation, but its use can be contraindicated and might be associated with a risk of bleeding. Therefore, efficacious LAA exclusion will not only significantly reduce the postoperative stroke rate in AF patients, but should also reflect on the rhythm outcome. With the availability of epicardial clip devices, a beating heart procedure does not limit the surgeon in managing the LAA, which is confirmed by the data of this registry.

There are several limitations when evaluating registry data. As such it is impossible to define which exact lesion might have contributed to success or failure. Since the ablation was performed according to each centres and surgeon’s clinical practice, there was no standardized approach for each group or for different types of AF. Further, only the minority of patients received Holter evaluation, and no information on other supraventricular arrhythmias than AF was available, which both is recommended by the current guidelines. Follow-up by telephone contact and rhythm monitoring with mainly only ECG, does not provide accurate coverage, which might result in a distorted picture of the actual recurrence rate. However, since the goal of this study was rather to compare two different treatment approaches than revealing specific success rates that can be achieved, those errors affected both arms the same way.

A time-to-event analyses would have been an appropriate way to analyse the two groups, especially if the follow-up time is different. Due to a substantial lack on information of the exact timing to first recurrence, a survival analysis with interval-censored survival times using all information about the timepoint of the first recurrence was used instead.

While registry data usually is less standardized, it nonetheless reflects the real-world practice in a more realistic way by the lack of specific inclusion-/exclusion criteria and of defined treatment characteristics.

Finally, a certain number lost to follow-up occurs commonly in registry data which increases the likelihood of the study being underpowered. The purely descriptive and non-confirmative character of a registry leaves the risk of type II error for non-significant differences. It needs to be emphasized that this data should be interpreted in a descriptive sense.

CONCLUSION

In patients undergoing concomitant AF ablation, our data suggests that the technique applied for myocardial revascularization (off-pump vs on-pump) leads to differences in the applied lesion set, but not in safety and in effectiveness. This suggests that the majority of patients with CAD and AF can be treated equally successful within the opportunities and limitations of each operative technique. Both procedures were effective and safe. According to this registry, the limited lesion set in off-pump procedures did not lead to differences in outcome. A box lesion, which is as feasible in off-pump as in on-pump procedures, is more effective than PVI alone and reveals acceptable rhythm outcome in this patient cohort.

Glossary

ABBREVIATIONS

AF

Atrial fibrillation

CABG

Coronary artery bypass grafting

CAD

Coronary artery disease

CASE-AF

CArdioSurgEry Atrial Fibrillation Registry

ESC

European Society of Cardiology

IHF

Institute für Herzinfarktforschung

LA

Left atrium

LAA

Left atrial appendage

PVI

Pulmonary vein isolation

SR

Sinus rhythm

FUNDING

This work was supported by a grant from AtriCure Europe BV, De I 260, 1101 EE, Amsterdam Z. O.

Conflict of interest: none declared.

DATA AVAILABILITY

Since the registry is on-going, partial data may be provided to qualified researchers through a direct request, including proposed research question(s), to oezdemir@stiftung-ihf.de. Patient level data will not be made available.

Author contributions

Mindy Vroomen: Resources; Writing—original draft; Writing—review & editing. Ulrich Franke: Resources; Writing—review & editing. Jochen Senges: Conceptualization; Funding acquisition; Methodology; Project administration; Resources; Writing—review & editing. Ivar Friedrich: Resources; Writing—review & editing. Theodor Fischlein: Resources; Writing—review & editing. Thorsten Lewalter: Conceptualization; Methodology; Writing—review & editing. Taoufik Ouarrak: Data curation; Formal analysis; Methodology; Software; Validation; Writing—review & editing. Bernd Niemann: Resources; Writing—review & editing. Andreas Liebold: Resources; Writing—review & editing. Thorsten Hanke: Conceptualization; Funding acquisition; Methodology; Project administration; Supervision; Writing—review & editing. Nicolas Doll: Conceptualization; Funding acquisition; Methodology; Project administration; Supervision; Writing—review & editing. Marc Albert: Resources; Supervision; Writing—review & editing.

Reviewer information

Interactive CardioVascular and Thoracic Surgery thanks Peter Alston, Kelly Christy, Daniel A. Grandmougin and the other anonymous reviewers for their contribution to the peer review process of this article.