Ablation of paroxysmal atrial fibrillation during coronary artery bypass grafting: 12 months’ follow-up through implantable loop recorder

Abstract

Objectives: The study aimed to identify responders to atrial fibrillation (AF) ablation, through continuous subcutaneous monitoring in patients with paroxysmal atrial fibrillation (PAF), who underwent epicardial pulmonary vein isolation (PVI) concomitantly with coronary artery bypass grafting (CABG). Methods: Seventy-two patients aged 61.6 ± 4.7 years with PAF underwent epicardial PVI with bipolar radiofrequency during CABG. Conduction block was confirmed by pacing. At the end of the procedure, the implantable loop recorder (ILR) for continuous monitoring was implanted in all patients. Follow-up data were collected through the ILR telemetry. Patients with an AF burden (AF%) ≪ 0.5% were considered AF free (responders). Patients with AF% > 0.5% were classified as non-responders. The AF episodes stored by the implanted device were visually inspected by the investigators to confirm the arrhythmia. The data were collected each month during 1-year follow-up. Results: No procedure-related complications occurred either for ablation or for the monitoring device. At the first post-ablation follow-up (1 month) during the blanking period, 37 patients (51%) were AF free, that is, with AF% ≪ 0.5%. At the end of the blanking period (3rd follow-up), 44 (61%) patients were AF free. At 12 months’ follow-up, 52 (72%) patients were AF free. Among 20 (28%) patients with AF recurrence, six (30%) patients were completely asymptomatic. There were no ischaemic strokes during the 1-year follow-up. Conclusion: Concomitant AF ablation during CABG is effective in the treatment of AF, as assessed through 1 year of continuous monitoring. Use of subcutaneous monitors is safe and accurate for AF detection, clinically relevant in identifying responders and non-responders and managing the medical therapies accordingly.

Ischaemic heart disease, Atrial-fibrillation ablation, Continuous subcutaneous monitoring, Coronary artery bypass grafting

1 Introduction

Many studies demonstrated that preoperative atrial fibrillation (AF) in patients undergoing heart surgery is associated with major adverse cardiac events and can reduce survival [1–3]. Nowadays, surgical ablation of AF is the standard recommended concomitant procedure, which leads to improving quality of life, reducing the risk of stroke and heart failure and improving survival [4].

Nevertheless, it is well known that patients with symptomatic AF may also have asymptomatic episodes [5], which can increase after ablation [6,7], and there is a poor correlation between symptoms and the underlying heart rhythm [8,9].

Therefore, it is quite risky to rely on patient-reported symptoms, as it can lead to misclassification of patients with/without real recurrences of the arrhythmia. Misclassifying patients with/without AF recurrences, in its turn, may negatively influence clinical decisions regarding antiarrhythmic and anti-thrombotic treatments.

Intermittent Holter electrocardiograph (ECG) or event recorders are the only tools currently used in post-ablation monitoring, combined with patient-reported symptoms. However, intermittent methods also have relevant drawbacks in identifying correctly responders and non-responders to therapy, due to their low sensitivity in detecting paroxysmal AF (PAF) episodes [10–12].

New technologies for continuous AF monitoring with minimally invasive procedures become now available in clinical practice [13]. The data of the Reveal XT Performance Trial (XPECT) study demonstrated that subcutaneous continuous monitoring is highly sensitive in detecting AF episodes and highly accurate in measuring the duration of AF in terms of daily burden [14]. Our group has already developed experience in the management of patients submitted to percutaneous ablation and subcutaneous continuous monitoring [15].

Thus, the aim of this study was to identify responders to AF ablation through continuous subcutaneous monitoring in patients with PAF, who underwent epicardial PVI concomitantly with coronary artery bypass grafting (CABG).

2 Materials and methods

2.1 Patient populations

Between February 2008 and March 2009, concomitant AF ablation was performed in 72 consecutive patients with ischaemic heart disease. Indications for concomitant procedure were symptomatic, paroxysmal drug-refractory AF (at least two antiarrhythmic drugs had been attempted) with documented monthly sustained episodes and indication for CABG, according to the American College of Cardiology/American Heart Association (ACC/AHA) guidelines for CABG surgery. All patients with a record of heart surgery and those requiring concomitant valve surgery were excluded from this study.

PAF was defined as the occurrence, in the previous 6 months, of one or more episodes lasting ≪ 7 days, all episodes terminating spontaneously [1].

The baseline assessment included clinical evaluation, standard laboratory tests, 12-lead ECG, transthoracic, trans-oesophageal echocardiography and coronary angiography.

This study was approved by the local Ethic Committee and conducted in compliance with the protocol and in accordance with standard operating procedures of the study. All patients signed the Informed Consent Form for participation in the study.

2.2 Operation technique

The ablation procedure was performed using an irrigated bipolar radiofrequency (RF) ablation system (Cardioblate BP, Model 60821; Medtronic, Minneapolis, MN, USA). Pulmonary-vein isolation was performed in all patients. Briefly, after onset of cardiopulmonary bypass, when the heart was still beating, the dissection of pericardial reflections and eventually the interatrial groove was performed and the bipolar device was clamped around the atrial cuff containing the inflow of the right pulmonary veins (PVs). RF energy was delivered to create the encircling ablation. On the left side, after the heart was lifted, the encircling ablation was created in a similar manner. The ligament of Marshall was surgically dissected. Care was taken to apply the ablation on the atrial myocardium rather than on the PV itself. All patients received at least two encircling lesions on the left and on the right side. In cases of complex anatomy, additional overlapping linear lesions were performed. No further connecting lines to the mitral annulus or the left-atrial appendage were made. The left-atrial appendage was untouched. After cardioplegic arrest, a standard CABG procedure was performed.

2.3 Conduction-block validation

The assessment of intra-operative conduction block was performed in all patients. Two monopolar electrodes (Medtronic Inc., Minneapolis, MN, USA) were attached to the distal portion of each PV vein pair and paced prior to and after ablation. Conduction block was validated when it was not possible to pace the heart anymore through 20-mA pulses.

2.4 The implanted device

The implantable loop recorder (ILR) (Reveal XT; Medtronic Inc., Minneapolis, MN, USA) for continuous monitoring was implanted in the parasternal area of the chest at the end of the surgical operation. The requirement for defining the exact final position was an R-wave amplitude ≥0.4 mV assessed through the Vector Check.

The implanted device continuously classifies the heart rhythm of the patient [16]. This classification is made through the analysis of the beat-to-beat variability of cardiac cycles on a 2-min ECG strip. The device stores the amount of AF per day (Daily AF Burden, h in AF in 1 day) and the AF Burden of the overall follow-up period, defined as the percentage of time in AF (AF%). In addition, the ECG is stored for visual confirmation of AF episodes. By accumulating data from multiple follow-up sessions, it is possible to discern the trend in the Daily AF Burden in 1 year.

Patients were provided with the Patient Assistant, a tool that allows each patient to store the ECG through the implanted device during symptoms: data were collected to analyse heart rhythm during symptomatic events. Fig. 1 shows the ablation technique and ILR implantation.

Fig. 1

Operation technique and ILR implantation. (A) Epicardial right PVI with irrigated bipolar radiofrequency; (B) epicardial left PVI with irrigated bipolar radiofrequency; (C) ILR implantation.

2.5 Definition of responders

Patients with an AF% ≪ 0.5% were considered AF free (responders). The cut-off of 0.5% corresponds to a maximum cumulative time in AF of 3.6 h in 1 month and to more than 99.5% of the time spent in sinus rhythm during the overall follow-up period (1 month). Patients with AF% > 0.5% were classified as non-responders: AF was visually verified by investigators through the analysis of the stored ECGs. Fig. 2 shows an example of the trend in the daily AF burden in a responder and a non-responder.

Fig. 2

Trend in the Daily AF Burden (hours per day) in a responder (panel A) and a non-responder (panel B). The responder only had some sporadic and short-lasting (≪10 min each) episodes during follow-up.

2.6 Postoperative management and follow-up

All patients were kept on antiarrhythmic drug (AAD) therapy before surgery. After the surgical operation, patients discontinued AAD therapy. All patients received oral anticoagulation therapy for a minimum of 6 months. At the 6-month follow-up, all patients with CHADS₂ score = 0 and all responders with a CHADS₂ score = 1 discontinued oral anticoagulation.

The data stored by the ILR were collected every month during the 12-month follow-up. In patients with recurrences, the telemetric data and the stored ECGs were used to tailor the antiarrhythmic therapy and/or to guide a percutaneous ablation procedure.

2.7 Statistical analysis

Results are expressed as mean values ± SD or as numbers and percentages, as appropriate. Student’s t-test was used for comparison of continuous data and chi-square test was used for comparison of categorical data. The AF events were then categorised by month intervals commencing from the date of procedure to the completion of the 12-month study period. Point prevalence of AF for the entire cohort was calculated for each month interval for both asymptomatic and symptomatic AF events. Analysis of variance (ANOVA) was used to compare the difference in prevalence from the beginning of the study to the end.

A p value ≤ 0.05 was regarded as statistically significant. All analyses were performed by means of the Statistical Package for Social Sciences (SPSS) (SPSS Inc., Chicago, IL, USA) software package.

3 Results

3.1 Patient populations, surgical and early postoperative period data

This study enrolled 72 patients (mean age 62 ± 5 years, 58 males), followed up for 12 months after CABG with concomitant AF ablation. All the patients had ischaemic heart disease and a history of PAF with mean duration of 5.4 ± 4.2 years. Table 1 shows the preoperative patient characteristics.

Table 1

Baseline characteristics of the patient population.

CABG with concomitant PVI using irrigated bipolar RF ablation was performed in all patients. The ablation procedure was done during on-pump, while the heart was beating. The mean time of ablation was 2.9 ± 0.7 min. At least two applications of the clamp were performed per lesion. The mean number of applications was 2.8 ± 0.5 and 3.2 ± 0.9 for left and right PVI, respectively.

After a double encircling ablation, absence of atrial capture (conduction block) during pacing from PV was reached in 128 (89%) out of 144 PV pairs. In the remaining 16 (11%) PV pairs without conduction block, additional lesions were performed until the block was finally achieved. Seven patients required multiple RF application due to the complex anatomy of the right- (five patients) and left- (two patients) sided PVs Finally, complete disconnection of the PVs from the LA was successfully achieved in all patients.

The average conduits number during CABG was 2.9 ± 1.2. The mean cardiopulmonary bypass duration was 102 ± 5.5 min, with a mean time of aorta cross-clamping of 69.5 ± 4.2 min. At the end of the surgical procedure, the ILR was implanted in the parasternal area, which required 10.2 ± 5.4 min on average.

In the early postoperative period, re-operation due to bleeding was required in two (3%) patients. No ablation- or monitoring-device-implantation-related complications occurred.

During hospital stay, there were 12 (17%) patients who required either pharmacological (seven patients) or electrical (five patients) cardioversion due to prolonged paroxysms of AF. The mean intensive care unit stay was 2.7 ± 1.4 days and the mean hospital stay was 8.2 ± 3.4 days. All patients were discharged in sinus rhythm.

3.2 Follow-up data

No patients were excluded during monitoring-device (ILR) implantation and follow-up due to sensing issues and all data were analysed by the investigators. Each patient had 1, 3, 6 and 12 months’ follow-up ILR data collection.

Oral anticoagulation was discontinued 6 months after ablation in all responders with a CHADS₂ score = 0 (37, 51%) and CHADS₂ score = 1 (seven, 10%). All patients with a CHADS₂ score ≥ 2 remained on oral anticoagulation independently of AF recurrences. There were no ischaemic strokes during the 1-year follow-up.

3.3 Responders at blanking period, 6 months and 1-year follow-up

At the first post-ablation follow-up examination (1 month) during the blanking period, 37 patients (51%) were AF free, that is, AF% ≪ 0.5%. At the end of the blanking period (third follow-up), 44 patients (61%) were AF free.

Of the 72 patients, 49 (68%) were AF free 6 months after the operation. At the 12-month postoperation follow-up, 52 out of 72 (72%) were AF free. None of these patients was treated with antiarrhythmic medication. Fig. 3 shows the percentage of responders, month by month, from the day of the operation up to the 12-month follow-up.

Fig. 3

The percentage of responders, month by month, from the day of the surgical procedure to the 12-month follow-up.

3.4 Non-responders

At the 12-month follow-up after surgery, five (7%) patients out of 72 had atrial tachycardias recorded by the device; all were on AAD therapy. Patients with AF% > 0.5% and without any arrhythmia other than AF accounted for 17 (24%). Of these, 14 (83%) were on AAD. The overall number of patients with atrial arrhythmia recurrences was 20 (28%): six patients (30%) were completely asymptomatic.

Twelve (17%) patients underwent catheter ablation of AF with the Carto XP (Biosense Webster) system. In four (6%) patients, the second ablation was performed to eliminate atypical atrial flutter. One patient had typical atrial flutter, which required cavo-tricuspid isthmus ablation. A third procedure was performed only in one patient due to atypical atrial flutter. In all the 12 patients who underwent at least two procedures, recovery of conduction was documented in at least one PV.

After the additional procedures, only 10 (14%) patients still had AF recurrences and two (3%) patients had atrial tachycardia episodes, with AF% > 0.5%.

3.5 Trend of AF burden

The AF% decreased significantly (p ≪ 0.0001) during the follow-up period, even in non-responders. Fig. 4 shows the trend of the monthly AF%, considering the overall population of responders and non-responders. These trends show a plateau from 6 months onwards after the last procedure. A similar trend in AF% was observed in the non-responders’ group, but with higher values. The trend is a plateau for these patients from the fifth month on.

Fig. 4

The trend of the monthly AF%, considering the overall population of responders and non-responders. These trends show a plateau from 6 months onwards after the last procedure.

3.6 Symptomatic episodes

Patients were provided with the patient assistant to store the ECG during palpitations only. Patients recorded 992 episodes during the study period: 685 (69%) were not confirmed as AF episodes on the analysis of the corresponding ECG. AF was confirmed only in 31% of symptomatic episodes. Fig. 5 shows heart rhythm corresponding to the symptomatic episodes, as classified by the investigators.

Fig. 5

The heart rhythm corresponding to the symptomatic episodes, as classified by the investigators.

4 Discussion

This study is the first one assessing the success rate of epicardial AF ablation during concomitant CABG through continuous monitoring in patients with PAF and ischaemic heart disease. The success rate (72%) of epicardial PVI during CABG at 1 year of follow-up was lower compared with previously published data [17,18]. Such a difference can be explained by the more accurate and reliable diagnostics of continuous monitoring, including asymptomatic events.

The prevalence of AF in patients undergoing CABG without valve surgery varies from 1.7% to 7.2% [19,20]. There are not enough data concerning concomitant AF ablation during CABG without valve surgery, especially in patients with PAF. Some studies showed that preoperative PAF in patients requiring coronary revascularisation reduced long-term survival, that is, it is necessary to perform AF ablation even if AF has paroxysmal form [2,3]. Our study included 72 ‘non-valves’ patients who underwent concomitant AF surgery with PVI isolation only, and we achieved a good success rate.

This study is also the first one showing the monthly evolution of AF% after the surgical operation, including the blanking period. This post-ablation period is definitely a critical period, during which patients can have AF even if they will become responders. looking at the trend, the peak rate of responders is seen nearly 4–5 months after ablation; then, after 6 months, the rate stabilises. It is therefore essential to wait 6 months before classifying a patient as a responder and making relevant decisions on anti-thrombotic therapy.

To our knowledge, the majority of studies regarding concomitant AF ablation during surgery estimated the efficacy of ablation through 24- or 48-h Holter monitoring [17,20,21]. Although such methods are recommended to be used in current consensus for AF management, there may be an increase of asymptomatic events after AF ablation [6,22], which could potentially increase the risk of stroke due to undetected, silent AF. Moreover, correction of the ischaemia also minimises the patients’ complaints and can mask AF events. That is why it is very important to use continuous monitoring, especially regarding discontinuation of anticoagulation therapy. In our study, we stopped oral anticoagulation therapy only in responders with CHADS₂ score ≤1 after 6 months’ follow-up. We did not have any stroke during follow-up.

We proved that symptoms are unreliable markers to judge the success of the procedure in terms of AF recurrences. Due to the intermittent nature of the arrhythmia and the inconsistency of symptoms, the only reliable and accurate method of correctly classifying each patient as a responder or non-responder is through continuous monitoring. Thus, continuous monitoring can be referred to as the most authoritative method for making medical decisions in accordance with objective data.

We defined responders as patients with an AF% ≪ 0.5%. This definition is reasonable and is in line with some previous findings in the field of AF and implantable devices. Many previous studies used a cut-off of 5 min in the Daily AF Burden (total time in AF during 1 day) to evaluate the presence or absence of clinically relevant AF [10,11,23,24]. This method allows the detection of any AF episode longer than 5 min and prevents the storage of artefacts, as artefacts mainly influence short episodes [13]. The use of this approach was a compromise between the risk of losing short runs of AF and filtering out short-lasting artefacts. A similar method was also used by Botto et al. [11], where they discriminated between long-lasting episodes (>24 h) and any AF episode, described as the episodes providing a burden > 5 min in 1 day.

The device we used in this investigation analyses ECG strips of 2 min each and calculates the irregularity of the cardiac cycle in this time interval; consequently, the duration of each AF episode is a multiple of 2 min. A daily AF burden of 5 min corresponds to an AF% = 0.347%, and an AF% = 0.5% corresponds to 7.2 min in 1 day: these two values are comparable. Recently, the TRENDS study [25] showed that a daily AF burden longer than 5.5 h in 1 day can significantly increase the risk of stroke. This value would correspond to an AF% = 0.764% during a 1-month follow-up period. Thus, we come to the conclusion that an AF% = 0.5% is comparable to these thresholds; being almost in the middle of the range, this value may be a good compromise in establishing which patients are responders to therapy.

Botto et al. [11] demonstrated that the length of each episode is a relevant clinical parameter and should also be combined with the standard risk factors for stroke (CHADS₂ score) to define the risk of ischaemic stroke in patients with PAF. The aim of our study was not to correlate the amount of AF with the risk of stroke, but only to assess the efficacy of AF ablation. That is why we analysed the data in terms of monthly AF burden and not in terms of AF duration or maximum daily AF burden.

The investigators could analyse and inspect all the data collected, thanks to the strict requirements at implantation: R-wave ≥ 0.4 mV during mapping. The data collected through the implanted device were also used to decide about antiarrhythmic therapy after the blanking period in patients with recurrences. However, the study was not designed or powered to make any conclusions regarding AAD therapy.

4.1 Study limitations

This study was observational, non-randomised. Although there were no stroke events during the study period, this study was not powered to investigate this topic, and we had only a 1-year follow-up. We stopped oral anticoagulation therapy only in patients with CHADS₂ score = 0 (51%) and in AF-free patients with CHADS₂ score = 1 (10%). According to Botto findings [11], patients with CHADS₂ score = 1 and no AF anymore have a very low risk of stroke.

This is a single-centre research and we cannot extend our results to general clinical practice. However, thanks to that, we ensured homogeneous management of each procedure. On the other hand, our results cannot be compared with those of multicentre trials.

AF% is calculated as the mean value of Daily AF Burden over the follow-up period; hence, patients with different AF patterns (frequent short-lasting episodes or sporadic long-lasting episodes) might show a similar value. However, as our data were collected on a monthly basis, the threshold of 0.5% is the guarantee that every responder had ≪3.6 h of AF in 1 month.

5 Conclusions

Concomitant AF ablation during CABG is effective in treating AF, as assessed through 1 year of continuous monitoring. The use of subcutaneous monitors is safe and accurate for AF detection, clinically relevant in identifying responders, non-responders and managing medical therapies.