Efficacy, safety, and long-term survival of concomitant valve replacement and bipolar radiofrequency ablation in patients aged 70 years and older: A comparative study with propensity score matching from a single-centre


 Background Concomitant bipolar radiofrequency ablation and valve replacement in the elderly remains controversial. In the current study, we aimed to compare the outcomes of concomitant valve replacement and bipolar radiofrequency ablation with valve replacement alone in elderly patients with atrial fibrillation (AF). Methods This was a retrospective study of patients aged ≥ 70 years who underwent valve replacement with or without bipolar radiofrequency ablation in a single-centre between January 2006 and March 2015. The early postoperative results and long-term clinical outcomes were compared after propensity score matching. Results A total of 34 pairs of patients (73.94 ± 2.64 years old; 34 in the AF with ablation group and 34 in the AF without ablation group) were enrolled in the propensity score matching analysis. There were no significant differences between the two matched groups in terms of surgical mortality (5.88% vs. 2.94%, P = 0.555) and major postoperative morbidity. Kaplan–Meier analysis revealed a significantly better overall survival in the AF with ablation group compared to the AF without ablation group (P = 0.009). Cumulative incidence curves showed a lower incidence of cardiovascular death in the AF with ablation group (P = 0.025, Gray’s test). Patients in the AF with ablation group had a reduced incidence of stroke compared to patients in the AF with ablation group (P = 0.009, Gray’s test). The freedom from AF after 5 years was 58.0% in the AF with ablation group and 3.0% in the AF without ablation group. Conclusions The addition of bipolar radiofrequency ablation is a safe and feasible procedure, even in patients aged ≥ 70 years, with a better long-term survival and a reduced incidence of stroke compared to valve replacement alone. These findings suggest that bipolar radiofrequency ablation should always be considered as a concomitant procedure for elderly patients with AF who require cardiac surgery. However, a large-scale, prospective, multi-centre, randomized study should be performed in the future to fully validate our findings.

Patients with advanced age may have an increased frequency of adverse side effects to drugs used for treating AF [8]. In addition, age is deemed to be an independent predictor of stroke in patients with AF [9].
Age of≥70 years is frequently accompanied by multiple co-morbidities, and is a signi cant risk factor associated with worse postoperative results [10] [11] [12] [13]. Thus, whether concomitant BRFA should be performed in elderly patients is controversial. Several previous studies have indicated similar success rates and similar low complication rates in the elderly compared to younger patients undergoing concomitant Cox-maze surgery [14] [15]. In an analysis from the Polish National Registry of Cardiac Surgery Procedures [14], neither clinical bene t nor harm was demonstrated in a subgroup of patients > 70 years with concomitant surgical ablation during the 12-year study period. Furthermore, Funatsu et al. [16] set inclusion criteria for adding surgical ablation, including an age of < 70 years in their clinical practice. However, patients with severe conditions may prefer to choose a simpli ed procedure without surgical ablation. Furthermore, no previous clinical studies have speci cally compared the clinical results of concomitant surgical ablation for AF vs. valve replacement (VR) alone in the elderly, and few studies have assessed the effect of surgical ablation on long-term outcomes [17]. Therefore, in the current study, we aimed to compare the clinical results between concomitant BRFA and no ablation in patients with VR aged ≥ 70 years after adjusting for severity by propensity score matching (PSM).

Study Populations
Between January 2006 and March 2015, a total of 362 consecutive patients (aged ≥ 70 years) with valve disease and AF underwent cardiac surgery at our institution. Among them, 48 patients underwent VR with concomitant Cox-maze procedure (42 underwent BRFA procedure and 6 underwent Cox-maze III procedure). The inclusion criteria were as follows: 1) Patients who underwent VR and/or BRFA for AF; and 2) patients with biological valves. The exclusion criteria were as follows: 1) Patients with mechanical valves; 2) patients who had received repeated cardiac surgeries; 3) patients who had received other forms of surgery for AF; and 4) patients who had not completed a follow-up of at least 5 years, unless death occurred. Finally, 276 patients were included in the nal analyses. Preoperative demographic and clinical data, perioperative outcomes, and follow-up results were collected for all patients and retrospectively analysed. Of the 276 patients, 42 patients underwent VR with concomitant BRFA, and the remaining 234 patients underwent VR alone. All patients had persistent or long-lasting persistent AF, the diagnosis of which was con rmed by 24-hour Holter monitoring before surgery. PSM was performed to achieve balance in covariates between groups. The process of patient selection is illustrated in Figure 1.
The study was approved by the Ethics Committee of Union Hospital, Fujian Medical University, and the study adhered to the Declaration of Helsinki. The requirement for written informed consent from the patients or their guardians was waived due to the retrospective nature of the study.

Surgical Procedures
Decisions regarding the indication for AF ablation were made according to the preference of the heart surgeon at our institution (according to the patients' EuroSCORE and other characteristics). In general, patients with a EuroSCORE ≥ 10, severe valvular heart disease requiring emergency surgery, LA diameter > 55 mm, or long-lasting (> 15 years) persistent AF were not considered suitable candidates for concomitant BRFA.
All surgical procedures were performed through a median sternotomy under general anaesthesia with hypothermic cardiopulmonary bypass (CPB). We used the Medtronic Cardioblate G2 Surgical Ablation System (Medtronic Inc., Minneapolis, MN) with a bipolar radiofrequency clamp for surgical ablation in accordance with the method introduced by Sie et al. (Figure 2). [18] In cases where there was a thrombus in the left atrial, atrial thrombectomy was performed rst after aortic cross-clamping. If there was no thrombus, right atrial ablation was performed with the heart beating, and left atrial ablation was performed during cardioplegic arrest. The right pulmonary was isolated, and the circumferential line around the right pulmonary veins was ablated. Two ablation lines were made from the right atrial appendage to the right atrial free wall and the tricuspid annulus (the junction between the anterior and septal lea ets of the tricuspid valve). The incision on the free wall of the right atrium was extended from the atrioventricular groove to the interatrial groove. Two ablation lines were created from the right atriotomy to the inferior vena cava and the superior vena cava. Another ablation line was created around the ori ce of the coronary sinus. After aortic cross-clamping, the left pulmonary vein was isolated, and the ligament of Marshall was interrupted. The left atrium was entered via the incision parallel to the interatrial groove, and the circumferential line around the left pulmonary veins was ablated. Other ablation lines were created between the left atrial appendage and the left superior pulmonary vein, between the bilateral superior and inferior pulmonary veins, between the right inferior pulmonary vein and the posterior mitral annulus. The left auricle was nally ligated. Each radiofrequency ablation was performed at least three times. The valve prosthesis was then implanted. The temporary epicardial pacing wire was placed routinely after cardiac surgery.

Postsurgical Treatment
The postoperative care was similar for both groups. Unless contraindicated, amiodarone infusion was given 5 mg/kg in the rst hour and 0.6 mg/kg/h over the following 72 hours, and 200 mg orally twice a day for the next 3 weeks. For patients undergoing BRFA for AF, amiodarone infusion was given 200 mg orally daily until the end of the rst 6 months to reduce the risk of early atrial arrhythmia recurrence. Sotalol or metoprolol was given as appropriate to patients with a history of side effects of amiodarone, or other clinical contraindications, in order to control the ventricular rate. Patients were advised to continue use of warfarin for 3 months after surgery, unless there was a speci c contraindication. The duration of prothrombin was regularly monitored; the target international normalized ratio was 1.5-2.0 for aortic valve replacement and 2.0-2.5 for mitral or tricuspid valve replacement. In cases where the patient underwent valve replacement alone, and if no attempts were made to restore and maintain sinus rhythm by means of ablation, β-blocker drugs were administered to control the ventricular rate, and warfarin was used for anticoagulation.

Outcome Measures
The primary outcomes included all-cause death, cardiac death, sinus rhythm restoration, and stroke events. The secondary outcomes of interest were surgical mortality and major postoperative morbidity.
Surgical mortality was de ned as death during the same hospitalization, and cardiac death was de ned as death due to malignant arrhythmia, stroke, congestive heart failure, or sudden unexplained death. Neurological complications are de ned here as embolic or hemorrhagic strokes. Stroke was de ned as an ischemic event caused by cerebral embolism. After 6 months of surgical ablation, any episode of AF, atrial utter, or atrial tachycardia lasting more than 30 seconds recorded by electrocardiogram or 24-hour Holter monitoring, was considered as recurrence of AF. Long-term survival was measured as time (months) to death from the date of surgery.

Follow-up
Follow-up visits at discharge were made 3, 6, and 12 months postoperatively and annually thereafter. The patients' health history, a physical examination, a 12-lead electrocardiogram, and an echocardiography were obtained at each visit. Patients with palpitations or documented atrial arrhythmias by 12-lead electrocardiography underwent 24-hour Holter monitoring for further con rmation. The period between the day after surgery and death, loss to follow-up, or the prede ned date of March 2020 was de ned as the follow-up duration.

Statistical Analysis
Continuous variables are expressed as mean ± standard deviations and categorical variables are presented as count, unless otherwise noted. Differences were compared using the Student's t-test or Mann-Whitney U test for continuous variables, and the chi-square test for categorical variables. One-toone pair matching was performed using the caliper-matching method with the variables shown in Table 1, and a 0.02 propensity score tolerance was imposed on the maximum propensity score distance. The overall survival and AF-free survival were analysed using Kaplan-Meier analysis and log-rank methodology. Cox regression was used to test for differences in survival between groups, and hazard ratios (HR) and 95% con dence intervals (CI) were computed. The incidence of postoperative stroke in the matched population was estimated using the cumulative incidence function with death as a competing risk. The hazard of cardiac death was also analysed using the competing risk method, with non-cardiac death as the competing event. The differences between cumulative incidence curves in the competing risk analysis were compared using Gray's tests. Fine and Gray competing risk regression was used to analyse the determinants of postoperative stroke and cardiac death. The sub distribution hazard ratio (SHR) and 95% CI were computed. A two-sided P-value < 0.05 was considered statistically signi cant. All statistical analyses were performed using SPSS 26.0 and R 3.6.2.

Baseline Characteristics
Patient characteristics and comparison of the data before and after matching are shown in Table 1. At our centre, approximately 13.3% of patients (aged ≥ 70 years) with valve disease with AF underwent concomitant Cox-maze surgery. PSM resulted in two groups of 68 patients (34 pairs), with similar baseline characteristics and a similar risk pro le. The mean age of THE 34 pairs of patients was 73.94 ± 2.64 years. Before PSM, the duration of preoperative AF was signi cantly longer among patients with VR alone than among patients who underwent VR/BRFA. Moreover, age, smoking history, liver dysfunction, left ventricular diastolic diameter, and liver dysfunction were signi cantly different between the two groups. days, P = 0.001) compared to the AF without ablation group. There was no signi cant difference in hospital mortality between the two propensity-matched groups (5.88% vs. 2.94%, P = 0.555). Two patients died in the with ablation group (ventricular tachycardia in 1 case, low cardiac output syndrome and multiple organ failure in 1 case), and 1 patient died of pneumonia in the AF without ablation group. Other adverse events, such as pulmonary complications, neurological complications, renal failure requiring dialysis, low cardiac output, reoperation for bleeding, and permanent pacemaker (PPM) implantation were similar between the two matched groups. A PPM was implanted to treat third-degree atrioventricular block in three (7.14%) patients in the AF with ablation group within 6 months of surgery compared to one (2.38%) patient in the AF without ablation group (P = 0.303). The remaining eight patients who underwent concomitant BRFA and were not successfully matched experienced no severe perioperative complications or death, except for one patient who had a PPM implanted due to thirddegree atrioventricular block. The incidence of PPM insertion in the AF with ablation group (n = 42, before PSM) was 9.52%.

Late Outcomes in the Matched Population
Follow-up was complete after a median of 85.00 months (interquartile range [IQR], 58-98 months).

Long-term Survival and Cardiovascular Death
Kaplan-Meier analysis revealed signi cantly better overall survival (P = 0.009; Figure 3) in the AF with ablation group during clinical follow-up (Cox regression: HR, 0.479; 95% CI, 0.235-0.974; P = 0.042). The 5-and 10-year survival rate was 82% and 58% for the AF ablation group compared to 68% and 15% for the AF without ablation group. Cardiovascular death occurred in 6 (17.6%) patients in the AF with ablation group, and 17 (50.0%) patients in the AF without ablation group. Cumulative incidence curves also showed a lower incidence of cardiovascular death in the AF with ablation group (P = 0.025, Gray's test; Figure 4). The unadjusted and adjusted SHRs evaluating the impact of additional BRFA procedures on the risk of cardiovascular death were 0.337 (95% CI, 0.135-0.843; P = 0.020) and 0.264 (95% CI, 0.093-0.753; P = 0.013), respectively.

Stroke Events and Sinus Rhythm Restoration
Patients in the AF with ablation group had a reduced incidence of stroke events compared to patients in the AF without ablation group (P = 0.009, Gray's test; Figure 5) during clinical follow-up. Stroke events occurred in 5 (14.7%) patients in the AF with ablation group compared to 18 (52.9%) patients in the AF without ablation group. The unadjusted and adjusted SHRs evaluating the impact of additional BRFA procedures on the risk of stroke events were 0.343 (95% CI, 0.137-0.859; P = 0.022) and 0.219 (95% CI, 0.076-0.631; P = 0.005), respectively. The AF with ablation group had better AF-free survival than the AF without ablation group (P 0.0001; Figure 6). Freedom from AF after 5 years was 58.0% in the AF with ablation group compared to 3.0% in the AF without ablation group, and some patients developed recurrent AF over time.

Discussion
Valve disease with AF in the elderly will become more common as the elderly population progressively increases. However, surgical treatment of AF at the time of cardiac surgery is not yet widely practiced in elderly patients. Many surgeons believe that the surgical risks are increased, owing to the additional aortic cross-clamp and CPB time required for the addition of the Cox-maze procedure.
As a simpler surgical technique that only requires a few minutes of operative time, BRFA has been gradually used to create lines of conduction block in patients with AF who undergo other open-heart surgeries. Our study aimed to assess the e cacy, safety, and long-term survival of BRFA compared to VR without AF ablation in elderly patients (aged ≥ 70 years) with valve disease and AF.
The results of our matched control study indicate that satisfactory performance can be achieved by BRFA concomitant with VR, even in elderly patients, who are considered by many surgeons to be at high risk of ablation failure.
An important nding of our study is that the addition of BRFA procedures did not signi cantly increase the risk of common postoperative complications and postoperative mortality. One possible explanation is that better early postoperative hemodynamic stability associated with the treatment of AF. This result demonstrates that the addition of the BRFA procedure has the potential to eliminate the risks of increased CPB time. Our ndings are in line with those previous studies. A study reported by Ad N and colleagues analysed the safety and e cacy of concomitant surgical AF ablation according to patients' age, and showed that the performance of the Cox-maze III/IV procedure for patients aged ≥ 75 years did not increase the operative risk. [19] Another analysis of the Cox-maze procedure by Kuh JH and colleagues indicated that patients aged ≥ 70 years obtained satisfactory results without major morbidity or mortality, although this study lacked a control group for comparison. [20] In our study, the differences in clinical outcomes were illustrated by comparison of matched elderly patients who underwent valve replacement alone versus those who underwent an ablation concomitant valve replacement procedure. Furthermore, we observed some early clinical results that have not been reported in previous studies. We found that BRFA procedures for AF were correlated with an increased risk of prolonged mechanical ventilation after surgery. The length of hospital stay and ICU stay were also signi cantly longer. These ndings may be related to impaired lung function due to the increased aortic cross-clamp and CPB time required for the addition of the BRFA procedure, which also prolongs the length of hospital stay and ICU stay. Another possible reason might be that patients in the AF with ablation group require a longer hospital stay to monitor the heart rhythm and heart rate.
The main indications for PPM implantation were a third-degree atrioventricular block, complications of which were due to ablation lines, which may damage the atrioventricular node [21]. Thus, biatrial ablation led to a signi cant PPM implantation rate. In our study, BRFA did not signi cantly increase the rate of PPI after surgery, which may be due to the fact that our surgeon was particularly careful to avoid damage of the atrioventricular node in the ablation area. The requirement for PPM has been reported to range from 4.5% to 13.3% in several studies [22][23] [24]. In our cohort, 9.52% of patients received a PPM in the 6 months after surgery, which matched with the previous studies mentioned above. This suggests that relatively older age may not be a factor of the increased PPM requirements.
Another notable nding is that elderly patients had better long-term outcomes of RPFA concomitant with VR compared to VR without AF ablation. We show that the proportion of patients who achieved sinus rhythm was signi cantly higher in the AF with ablation group compared to the AF without ablation group, both at discharge and during follow-up. Moreover, the cumulative freedom from recurrence of AF was 65.2% at 5 years, which was comparable to that shown in previous studies conducted in experienced centres. Suwalski P et al. reported that the rate of sinus rhythm maintenance in 27 patients > 70 years old with a concurrent Cox-maze procedure was 89% at a mean follow-up of 51 months, [15] whereas our analysis resulted in a rate of 65.2%; this difference was likely driven by individual differences in patients and heterogeneity among ablation procedures. Furthermore, similar to our analysis, the investigators observed low in-hospital mortality. The study by Lee R. demonstrated that 78% of the patients (age 65.9 ± 12.2 years) who underwent surgical ablation were free of AF after 47 months of follow-up, and also proved that restoration of sinus rhythm improved survival. [25] In terms of long-term survival associated with concomitant ablation, the investigators reported a HR of 0.390, whereas the current study resulted in a very similar HR of 0.479. We also showed that additional BRFA procedures are associated with a low probability of stroke events and a signi cant decrease in cardiovascular death events and long-term overall mortality. We believe that restoration of sinus rhythm and excision of the left atrial appendage help to prevent stroke, which may lead to a remarkable reduction in long-term mortality. It has also been con rmed that chronic AF is associated with reduced survival after valve replacement [26]. The long-term success in our study argues strongly for a more widespread adoption of this currently underutilized treatment option for the elderly.
Previous clinical results of the concomitant BRFA with VR in the elderly were mostly compared to those in the younger group. Besides, previous studies have generally focused on patients with mitral valve surgery, and the outcomes of surgical ablation in patients with other types of valve surgery have not been well researched. What are the effects of different valve replacement types on clinical outcomes of bipolar radiofrequency ablation for atrial brillation? This is a question that deserves further exploration. The truth is that at present, this is beyond the scope of our study due to the limited number of elderly patients undergoing valvular and atrial brillation surgery in our center. Although several previous studies have described the e cacy and safety of concomitant surgical ablation with VR in elderly patients aged ≥ 70 years, to the best of our knowledge, our study is the rst to compare matched elderly patients undergoing valve replacement alone versus an ablation concomitant valve replacement procedure. Our ndings are robust because the duration of follow-up was adequate. Furthermore, we performed a matching-adjusted indirect comparison to remove observed differences in the patients' baseline characteristics between the two groups. Indeed, PSM is being increasingly used to reduce possible confusion in observational studies [27], which can provide clinicians with relatively reliable comparative evidence to assist with decision-making. With no prospective randomized studies in the literature, this study might provide some reference value for clinical practice. We believe that the survival bene ts of surgical radiofrequency ablation far outweigh the harmful effects of prolonged myocardial ischemic time and increased surgical complexity. Furthermore, the favourable in-hospital and long-term outcomes observed in our study add to the evidence that age alone should not be considered as a limiting factor for the concomitant BRFA procedure. However, further research is needed to con rm and validate the evidence.
There are several limitations to the current study. Firstly, this is a retrospective clinical analysis involving a small sample size at a single-centre. Due to these factors such as individual patient characteristics or surgeon's preference/expertise, the majority of elderly patients do not accept ablation surgeries, and prefer conservative treatments for AF; for this reason, it was di cult to enrol a su cient number of patients in this research. Moreover, we had to set strict inclusion criteria in order ensure the rigor of the research, which also reduced the number of suitable participants. Secondly, although we used PSM, confounds and selection biases between the two groups cannot be eliminated. Therefore, large-scale prospective randomized studies are warranted to evaluate BRFA in elderly patients. Further research is also needed to assess the quality of life of concomitant VR and BRFA in patients aged ≥ 70 years. Lastly, no comparison data for a younger population is presented in our study.

Conclusion
In this study, we compared the outcomes between concomitant BRFA and no ablation in patients with valve replacement aged ≥ 70 years. Our results indicate that the addition of BRFA is both safe and feasible, even in elderly patients, with a better long-term survival and a reduced incidence of stroke events compared to valve replacement alone. These ndings suggest that BRFA should always be considered as a concomitant procedure for AF in elderly patients who require to undergo cardiac surgery. However, a large-scale, prospective, multi-centre, randomized study should be performed in the future to fully validate our ndings. Abbreviations AF: atrial brillation; BRFA: bipolar radiofrequency ablation; HR: hazard ratios; SHR: subdistribution hazard ratio; CI: con dence intervals; VR: valve replacement; CPB: cardiopulmonary bypass; ICU: intensive care unit; PPM: permanent pacemaker.

Declarations Acknowledgments
We highly acknowledge the contribution by the participating doctors: Xiao-fu Dai, Qi-min Wang, Zhongyao Huang, Han-fan Qiu, Xue-Shan Huang. Also, we are also deeply indebted to to Sheng Chen and his colleagues. Availability of data and materials All data generated or analyzed during this study are included in this published article Ethics approval and consent to participate The present study was approved by the ethics committee of Union Hospital, Fujian Medical University and adhered to the tenets of the Declaration of Helsinki.

Consent for publication
Due to the retrospective nature of the present study, written informed consent from the patients or their guardians was waived.