Biatrial ablation vs. left atrial concomitant surgical ablation for treatment of atrial fibrillation: a meta-analysis

Transcription

1 Europace (015) 17, doi: /europace/euu0 CLINICAL RESEARCH Ablation for atrial fibrillation Biatrial ablation vs. left atrial concomitant surgical ablation for treatment of atrial fibrillation: a meta-analysis Kevin Phan 1,, Ashleigh Xie 1, Yi-Chin Tsai 3, Narendra Kumar 4, Mark La Meir 4,5, and Tristan D. Yan 1,6 * 1 The Collaborative Research (CORE) Group, Macquarie University Hospital, Macquarie University, Technology Place, Sydney, Australia; Sydney Medical School, The University of Sydney, Sydney, Australia; 3 The Prince Charles Hospital, Chermside, Australia; 4 Department of Cardiothoracic Surgery and Cardiology, Academic Hospital Maastricht and Cardiovascular Research Institute Maastricht, Maastricht, The Netherlands; 5 University Hospital Brussels, Brussels, Belgium; and 6 Department of Cardiothoracic Surgery, Royal Prince Alfred Hospital, University of Sydney, Sydney, Australia Received 30 April 014; accepted after revision 18 July 014; online publish-ahead-of-print 1 October 014 Aims Surgical ablation performed concomitantly with cardiac surgery has emerged as an effective curative strategy for atrial fibrillation (AF). Left atrial (LA) lesion sets for ablation have been suggested to reduce procedural times and post-surgical bradycardia compared with biatrial (BA) lesions. Given the inconclusive literature regarding BA vs. LA ablation, the present meta-analysis sought to assess the current evidence.... Methods Electronic searches were performed using six databases from their inception to December 013, identifying all relevant and results randomized trials and observational studies comparing BA vs. LA surgical ablation AF patients undertaking cardiac surgery. In 10 included studies, 5 patient results were available for analysis to compare BA (n ¼ 888) vs. LA (n ¼ 1337) ablation. Sinus rhythm prevalence was higher in the BA cohort compared with the LA cohort at 6-month and 1-month follow-up, but similar beyond 1 year. Permanent pacemaker implantations were higher in the BA cohort, but 30-day and late mortality, neurological events, and reoperation for bleeding were similar between BA and LA groups.... Conclusions Biatrial and LA ablations produced comparable 30-day and late mortality but LA was associated with significantly reduced permanent pacemaker implantation rates. Biatrial ablation appeared to be more efficacious than LA ablation in achieving SR at 1 year, but this difference was not maintained beyond1year. Trends appear to be driven by the preferential selection of long-standing and persistent AF patients for the BA approach. Future randomized studies of adequate follow-up are required to validate risks and benefits of BA vs. LA surgical ablation Keywords Atrial fibrillation Surgical ablation Biatrial Left atrial ablation Meta-analysis Introduction Atrial fibrillation (AF) is the most common sustained arrhythmia, associated with increased risk of mortality and stroke. With an estimated annual cost of $6.65 billion in the United States, and an estimated prevalence ranging from.7 to 6.1 million in the United States, 1 treatment and management of AF represents a significant healthcare and economic burden. Surgical ablation has emerged as a more effective curative strategy compared with anti-arrhythmic therapy, with the aim of interrupting the multiple, disorganized re-entrant circuits that underlie AF pathology. The traditional Cox-Maze III procedure represents the gold-standard treatment for AF and achieves complete isolation with biatrial (BA) lesions, min of cardiopulmonary bypass (CPB) and median sternotomy. The Cox-Maze procedure is usually performed concomitantly with other cardiac procedures such as mitral valve repair (MVR) or coronary artery bypass grafting (CABG). Recent enhancements in the understanding of AF triggers, such as pulmonary veins and left atrium, have led to a surge in alternative ablation modalities (Cox-Maze IV) * Corresponding author. Tel: +61 () ; fax: +61 () address: tristanyan@annalscts.com Published on behalf of the European Society of Cardiology. All rights reserved. & The Author 014. For permissions please journals.permissions@oup.com.

2 Meta-analysis of BA vs. LA surgical ablation 39 What s new? Recent studies have supported the safety and superior efficacy of surgical ablation compared with catheter-based techniques, but there are limited data ascertaining the effect of different energy sources and lesion sets. To our knowledge, this is the first meta-analysis of comparative studies focusing on the clinical and mortality outcomes of biatrial (BA) vs. left atrial (LA) surgical ablation for atrial fibrillation (AF). The primary findings from this analysis include that BA ablation was more efficacious than LA ablations in achieving sinus rhythm up to 1 year follow-up, but outcomes were comparable for longer-term follow-up. Left atrial ablation was associated with significantly lower permanent pacemaker implantations, with comparable 30-day and late mortality rates to BA ablation. The findings of our analysis suggest that BA ablation should be used judiciously in high-risk AF patients; however, this remains to be confirmed in future prospective randomized studies which are adequately powered. using different energy sources such as radiofrequency, cryo-energy, and microwaves. In order to reduce procedural time, postoperative bradyarrhythmias and promote less extensive lesions, newer techniques have adopted ablation of the left atrium only. 3 5 However, there are concerns that left atrial (LA) lesion sets are less efficacious compared with BA approaches, particularly in patients with right-sided AF triggers. 6,7 Given the inconclusive literature regarding the efficacy and clinical outcomes of BA vs. LA lesion sets, the present meta-analysis sought to assess the current evidence base. Methods Literature search strategy Electronic searches were performed using Ovid Medline, PubMed, Cochrane Central Register of Controlled Trials (CCTR), Cochrane Database of Systematic Reviews (CDSR), ACP Journal Club, and Database of Abstracts of Review of Effectiveness (DARE) from their date of inception to December 013. To achieve the maximum sensitivity of the search strategy, we combined variants of the terms: atrial fibrillation AND ablation AND biatrial OR left atrial as either key words or MeSH terms. The reference lists of all retrieved articles were reviewed for further identification of potentially relevant studies, assessed using the inclusion and exclusion criteria. Selection criteria Eligible comparative studies for the present systematic review and meta-analysis included those in which patient cohorts compared BA vs. LA surgical ablation outcomes for AF treatment. Studies with fewer than 0 patients in each treatment arm were excluded. When institutions published duplicate studies with accumulating numbers of patients or increased lengths of follow-up, only the most complete reports were included for quantitative assessment at each time interval. Reference lists were also hand-searched for further relevant studies. All publications were limited to those involving human subjects and in the English language. Abstracts, case reports, conference presentations, editorials, reviews, and expert opinions were excluded. Data extraction and critical appraisal All datawere extracted from article texts, tables, and figures. Three investigators (K.P., Y.T., A.X.) independently reviewed and assessed the quality of each retrieved article. The risk of bias was assessed using the Downs and Black checklist 8 for randomized and observational studies. Discrepancies between the reviewers were resolved by discussion and consensus. The final results were reviewed by the senior investigators (M.L., T.D.Y.). Statistical analysis Clinical outcomes were assessed using a standard meta-analysis technique, with the odds ratio (OR) used as a summary statistic. Both fixed- and random-effects models were tested and used to calculate the pooled OR or weighted mean differences (WMDs) for the surgical literature. Since similar results were obtained, only results of the random-effect model are presented. x tests were used to study heterogeneity between trials. I statistic was used to estimate the percentage of total variation across studies, owing to heterogeneity rather than chance, with values.50% considered as substantial heterogeneity. If there was substantial heterogeneity, the possible clinical and methodological reasons for this were explored qualitatively. Publication bias of the major outcomes of this meta-analysis was detected by Egger s regression test. All P values were two-sided. All statistical analysis was conducted with Review Manager Version 5..1 (Cochrane Collaboration, Software Update). Results Literature search A total of 445 references were identified through six electronic database searches. Manual search of reference lists yielded four new studies. After exclusion of duplicate or irrelevant references, 38 potential relevant articles were retrieved. After manual search of reference lists, and applying the inclusion and exclusion criteria, 10 studies were selected for analysis (Supplementary material online, Figure S1). The study characteristics are summarized in Supplementary material online, Figure S1. In these 10 studies, 5 patient results were available for analysis to compare BA (n ¼ 888) vs. LA (n ¼ 1337) lesion sets for surgical ablation for AF treatment. Study characteristics are summarized in Table 1. Quality assessment In the 10 studies included in this meta-analysis, there were two prospective randomized trials, 14,15 one propensity score analysed study, 11 two prospective observational studies, 4,10 four retrospective observational studies, 3,5,9,13 and one study which was originally designed as a retrospective but halfway though transitioned into a prospective observational study. 1 The risk of bias in each study according to reporting, external validity, selection bias, and power are summarized using the Downs and Black Quality Assessment Checklist in Supplementary material online, Table S1. Three studies used cryoablation energy, 3,10,11 four studies used radiofrequency energy, 4,13 15 and three studies used a combination of different energy sources including radiofrequency, cryoablation

3 40 Table 1 Study characteristics First Year Country Study Type of n n Mean follow-up Type of Lesion set Cardiac surgery AAD therapy protocol author period study (BA) (LA) (months) ablation USA 007 R, OS R RF, CY, MW PVI+ + LAA MVR NR Pecha Germany R, OS days RF, CY PVI + LAA CABG, AVR, MVR, NR TVR 10 Scandinavia P, OS R CY PVI + LAA MVR, ASD, CABG, AVR 11 South Korea 006 PSA CY PVI MVR, AVR, TVR, CABG, ASD, VSD USA R/P a RF, CY PVI CABG, MVR, AVR, TVR Amiodarone used selectively in early post op for atrial arrhythmias 100 mg amiodarone for postop AF/atrial arrhythmias, tapered in 1 month Patients discharged on AADs and withdrawn if in SR at 3 month follow-up Deneke 13 Germany NR R, OS RF PVI + LAA MVR, AVR, CABG NR 14 China P, RCT RF PVI + LAA + RAA MVR, AVR, CABG mg amiodarone for 6-months postoperation. If contraindicated, 80 mg sotolol was used instead 15 India P, RCT RF PVI + LAA MVR, TVR, DVR, OMV 4 00 Turkey 001 P, OS RF PVI + LAA + RAA MVR, TVR, CABG, AVR Takami Japan NR R, OS (BA), (LA) CY PVI + LAA + RAA MVR, CABG, AVR, TVR If patient in AF 7 days postoperation, then amiodarone was given for months 00 mg amiodarone given to patients for 3 months. Patients still in AF afterwards underwent cardioversion a Originally retrospectively, but transitioned into prospective study halfway through; R, retrospective; OS, observational study; PSA, propensity-score analysis; BA, biatrial; LA, left-atrial; R, range; AAD, antiarrhythmic drug; RF, radiofrequency; CY, cryoablation; MW, microwave; NR, not reported; PVI, pulmonary vein isolation; CS, cut-and-sew; LAA, left atrial appendage occlusion; RAA, right atrial appendage occlusion; MVR, mitral valve replacement, CABG, coronary artery bypass grafting; AVR, aortic valve replacement; TVR, tricuspid valve repair/replacement; ASD, atrial septal defect; ventral septal defect; OMV, open mitral valvotomy; DVR, double valve replacement. NR K. Phan et al.

4 Meta-analysis of BA vs. LA surgical ablation 41 Table Description of lesion success and rhythm success from included studies First Left-sided ablation lesion description Right-sided ablation lesion description Definition(s) for success author... PVI only or PVI + lesion to mitral valve annulus (+LAAO). Lesions to tricuspid valve annulus + R atrial free wall + intercaval line. Pecha PVI + box lesion + LAA + isthmus isolation. Intercaval lesion + isolation of cavotricuspid isthmus + right atrial appendage and terminal crest isolation Only done in patients with persistent/ long-standing AF. Deneke Epicardial approach Ipsilateral pair-wise PVI + connecting lesions between R/L PVs + line to mitral annulus on posterior LA. Endocardial approach Single box PVI + line from PV to LAA + additional line from PV to MV annulus posteriorly + epicardial coronary sinus ablation done on the opposite side of MV annular lesion. LAmazeprocedure: boxlesionoffourpvs + MVannulus lesion connected to PVI (isthmus lesion). This approachwasusedforrecent-onsetorparoxysmalaf. PV ostial circumferential ablation + interconnecting ablation lines between PVs on same side + line across LA posterior wall, to excised LAA + LA isthmus line connecting left inferior PV to mitral annulus LAA amputation + circumferential lesion at base + Isolation of right PVs + Left PVs encircled, with connecting line between both sets via roof + line left PVs to the posterior mitral annulus performed. PVI using encircling ablation around all 4 PVs + cryoablation applied to posterior MV annulus + LAA opened and left PVs isolated from LA + LAA base encircling ablation LAA amputated or sutured + circumferential lesion at base + line from LAA to left superior PV created + isolation of right PVs completed by circular ablation + Connecting line made between the two sets of PVs near to LA roof + line from left PVs to the posterior mitral annulus performed Takami LAA excision + Cryoablation delivered between upper and lower left PVs from base of excised LAA + line made from left lower atrial incision edge to posterior mitral valvular annulus and interatrial septum. BA lesion setdone accordingto the full Cox-MazeIII lesion set using surgical endocardial cryoablations in both atria. All patients had LAA ligation. Isolation of cavotricuspid isthmus + added vertical line toward superior vena cava + Isthmus isolation lines made from the lowest part of the right atriotomy + posterior line toward posterior tricuspid annulus BA ablation chosen for patients with long-standing, symptomatic AF or young patients, or right sided cardiac operations. LA lesions + right-sided lesions: lines from the inferior vena cava to tricuspid annulus BA ablation included LA lesion + endocardial right-sided ablation: intercaval line + right atrial isthmus line + septal line. Superior vena cava to inferior vena cava line + lateral wall to anterior-medial tricuspid annulus lesion + medial free-wall to anterior-medial tricuspid annulus + cavotricuspid isthmus ablation with excision of right atrial appendage. RA free wall is ablated in a line from superior to inferior vena cava + another line to posterior tricuspid annulus + medial wall of RA to anterior tricuspid annulus Right atrial appendage excised + line from superior and inferior caval cannulation sites endocardially + lesion performed from right atrial appendage to anterior tricuspid leaflet + posterior incision at the atrio-ventricular groove to posterior tricuspid annulus + line performed on the right side of interatrial septum up to coronary sinus + line from inferior vena cava and up to posterior tricuspid annulus. Right atrial appendage excision + cryoablation delivered towards right atrial free wall + tricuspid ring at base of posterior leaflet Freedom from AF or atrial flutter at 3, 6, 9, 1 months. Freedom from AADs not incorporated. NR Normal SR or pacing rhythm, and freedom from AF and Class III anti-arrhythmic drugs at 1-month follow-up Freedom from AF/atrial flutter/ atrial tachyarrhythmia off anti-arrhythmic medications at 3, 6, 1, 4-month follow-up Freedom of AF/atrial flutter/atrial tachycardia off AADs at 1 months Freedom from AF/typical right atrial flutter/atypical atrial macro-re-entry at 1 months Freedom from AF/atrial flutter on follow-up Success defined as freedom from AF/ atrial flutter/junctional rhythms/ ventricular fibrillation at 3 6 months (composite) and midterm Freedom from AF/ atrial arrhythmias/ atrial flutter at 3, 6, 1, 4 months Sinus rhythm success as freedom from AF/junctional rhythm and pacemakers at.6 month follow-up BA, biatrial; LA, left atrial; PVI, pulmonary vein isolation; LAA, left atrial appendage; LAAO, left atrial appendage occlusion; R, right; L, left; AF, atrial fibrillation; AAD, anti-arrhythmic drugs; NR, not reported; SR, sinus rhythm; MV, mitral valve. and microwave. 5,9,1 Concomitant CABG surgery was reported in seven studies, 3,4,10 14 while concomitant valvular was performed in all included studies. 3 5,9 15 Descriptions of the BA and LA lesion sets employed in each study is summarized in Table. Five studies had mean follow-up periods beyond 1 year. 3,11,13 15 Eight studies reported SR outcomes at 1 months or longer. 4,9 15 Freedom from AF at 1 months was reported in six studies. 4,9 1,15 The indications for anti-arrhythmic drugs (AADs) varied, and was clearly described in six studies (Table 1). Two studies 10,11 defined ablation success as sinus rhythm, freedom from AF, and freedom from AAD at 1 months, which is consistent with the outcomes recommended by the Heart Rhythm Society (HRS), European Rhythm Association, and

5 4 K. Phan et al. Table 3 Baseline characteristics of included studies Baseline characteristics n (%), BA n (%), LA OR/WMD (95% CI) Heterogeneity P for overall effect BA vs. LA... P value (95% CI) I... Age (years + SD) (.81,0.9) Male 39 (44.9) 447 (5.5) 0.81 (0.65,1.00) LVEF (%) 58 (100) 549 (100) 1.17 (.7,0.06) LAD (mm) 665 (100) 746 (100) 0.49 (0.36,1.34) Diabetes 5 (1.) 78 (14.3) 1.03 (0.69,1.54) Heart failure 89 (39.4) 185 (40.) 0.96 (0.6,1.5) Neurological events 8 (6.1) 67 (8.3) 0.75 (0.47,1.0) Hypertension 108 (9.0) 4 (54.5) 0.53 (0.38,0.74) Paroxysmal AF 56 (11.5) 155 (31.4) 0.34 (0.3,0.50) , Permanent/persistent AF 504 (88.6) 418 (69.0) 3.05 (.13,4.35) 0.8 0, Operative parameters n, BA n, LA WMD (95% CI) Heterogeneity P for overall effect... P value (95% CI) I... Cross-clamp (min) (0.49,1.74) CPB (min) (7.6,4.3) BA, biatrial; LA, left atrial; LVEF, left ventricular ejection fraction; LAD, left atrial diameter; AF, atrial fibrillation; CPB, cardiopulmonary bypass; WMD, weighted mean difference; OR, odds ratio; SD, standard deviation. European Cardiac Arrhythmia Society Consensus. 16 Definitions of success for each included study are summarized in Table. Baseline patient and operational characteristics The baseline patient and operational characteristics are summarized in Table 3. The average age ranged between 37 and 68 years and 36 and 69 years for BA and LA groups, with no significant difference (P ¼ 0.3). The proportion of males was similar between the BA and LA cohorts (44.9 vs. 5.5%; P ¼ 0.05). There were also no differences between BA and LA groups in terms of LA diameter (LAD) (P ¼ 0.6), diabetes (P ¼ 0.88), prior heart failure (P ¼ 0.85), and neurological events (P ¼ 0.3). The LVEF was significantly higher in the LA group although by a small margin of 1.17% (P ¼ 0.04). The proportion of patients with hypertension was also higher in the LA group (54.5 vs. 9%; P ¼ 0.000). Compared with the LA group, patients undergoing BA ablation were more likely to have permanent/persistent AF (88.6 vs. 69%;, ) and less likely to have paroxysmal AF (11.5 vs. 31.4%; P, ). There were no significant differences in cross-clamp duration between BA and LA ablation cohorts (P ¼ 0.07). However, CPB duration was longer for the BA compared with LA ablation by approximately 16 minutes (P ¼ 0.000). Assessment of efficacy Prevalence of sinus rhythm The numberof patients insr at discharge was comparablebetween BA and LA groups (70.0 vs. 74.7%; OR 0.8; 95% CI ; P ¼ 0.1; I ¼ 0%). However, SR prevalence was higher in the BA group compared with the LA group in four studies 3,9,11,15 reporting at 6-month (74 vs. 64%; OR 1.43, 95% CI ; P ¼ 0.03; I ¼ 0%) and four studies 9 1 reporting at 1-month follow-up (77 vs. 70%; OR 1.79, 95% CI ; P ¼ 0.005; I ¼ 37%). For patients with follow-up beyond 1 year, SR prevalence in BA and LA groups was similar (59 vs. 64%; OR 1.03; 95% CI ; P ¼ 0.87; I ¼ 6%). The weighted average mean follow-up for studies reporting SR beyond 1 year was 34.3 months. These results are summarized in Figure 1. Freedom from atrial fibrillation For surgical ablative treatment, freedom from AF at 1 months is a recommended outcome by the HRS, European Heart Rhythm Association (EHRA), and European Cardiac Arrhythmia Society (ECAS) consensus. In the present meta-analysis, freedom from AF was significantly higher in the BA compared with the LA ablation group (77.8 vs. 71.1%; OR 1.6; 95% CI ; P ¼ 0.0; I ¼ 37%). Assessment of safety Mortality Mortality outcomes at 30 days were reported in 8 out of 1 included studies. The 30-day mortality rates for BA ablation were not significantly different to LA ablation (4.1 vs..5%; OR 0.85; 95% CI ; P ¼ 0.14; I ¼ 4%). Late mortalities were reported in five included studies, and also showed similar risks between BA and LA ablation cohorts (5.4 vs. 5.3%; OR 1.1; 95% CI ; P ¼ 0.87; I ¼ 66%). Significant heterogeneity was observed in the late mortality comparison. The results for 30-day and late mortality are summarized in Figure. Permanent pacemaker implantations Permanent pacemaker implantations were significantly higher in the BA vs. LA ablation cohorts at latest follow-up (7.0 vs. 5.4%; OR 1.75; 95% CI ; P ¼ 0.008; I ¼ 0%). No significant heterogeneity was detected in this comparison. The results are summarized in Figure 3A.

6 Meta-analysis of BA vs. LA surgical ablation 43 BA Events Total LA Events Study or Subgroup Total Discharge SR Takami Heterogeneity: t = 0.00; c = 1.0, df = 4 (P = 0.88); I =0% Test for overall effect: Z = 1.8 (P =0.0) Weight % % % % % M-H, Random, 95% CI 7.5% 0.81 [0.5,.61] 0.81 [0.30,.14] 1. [0.51,.96] 0.81 [0.48, 1.38] 0.68 [0.38, 1.1] 0.81 [0.59, 1.1] Year month SR Takami % 1.5 [0.9, 5.37] % 1.11 [0.46,.68] % 1.43 [0.80,.53] % 1.80 [1.03, 3.13] % 1.49 [1.05,.1] Heterogeneity: t = 0.00; c = 0.96, df = 3 (P = 0.81); I =0% Test for overall effect: Z =.3 (P =0.03) 1 year SR % 0.96 [0.47, 1.94] % 15.5% 8.7% 1.54 [0.91,.6] 1.76 [0.70, 4.45].60 [1.43, 4.71] % 1.64 [1.09,.47] Heterogeneity: t = 0.06; c = 4.59, df = 3 (P = 0.0); I =35% Test for overall effect: Z =.36 (P =0.0) >1 year SR % 1.3 [0.50, 3.0] 008 Deneke % 31.% 33.1% 0.7 [0.34, 1.51] 0.79 [0.45, 1.38] 1.54 [0.90,.63] % 1.03 [0.70, 1.51] Heterogeneity: t = 0.04; c = 4.06, df = 3 (P = 0.6); I =6% Test for overall effect: Z = 0.16 (P =0.87) M-H, Random, 95% CI Favours LA Favours BA Figure 1 Forest plot of sinus rhythm prevalence at discharge, 6-month, 1 year and.1 year follow-up, showing summary of ORs with 95% confidence intervals for included studies. Neurological events The prevalence of neurological events, including strokes, transient ischaemic attacks, and thromboembolisms, was found to be comparable between BA and LA groups (1.8 vs..0%; OR 1.7; 95% CI ; P ¼ 0.63; I ¼ 19%). The results are summarized in Figure 3B. Reoperation for bleeding The incidence of reoperations for bleeding was not significantly different between LA and BA groups (4.0 vs. 3.0%; OR 1.30; 95% CI ; P ¼ 0.63; I ¼ 3%), with no significant heterogeneity detected. The results are summarized in Figure 3C. Publication bias Egger s test detected no publication bias for the major outcomes of this meta-analysis: sinus rhythm prevalence beyond 1 year (P ¼ 0.78; t ¼ 0.3), late mortality (P ¼ 0.44; t ¼ 0.88), permanent pacemaker implantations (P ¼ 0.59; t ¼ 0.58), neurological events (P ¼ 0.60; t ¼ 0.58), and reoperations for bleeding (P ¼ 0.9; t ¼ 0.09).

7 44 K. Phan et al. Discussion BA Events Total LA Events Study or Subgroup Total 30-day mortality Heterogeneity: t = 0.04; c = 6.4, df = 6 (P = 0.40); I =4% Test for overall effect: Z = 1.49 (P = 0.14) Weight Over the past two decades, ablative strategies have revolutionized the way AF patients are treated. Surgical ablation was pioneered by Cox in 199 using a cut and sew approach with median sternotomy to disrupt chaotic re-entry circuits. This later paved the way for its progression into a myriad of ablation techniques, from alternative energy sources such as radiofrequency and cryoenergy to innovative minimally invasive video-assisted thoracoscopic 17 0 and endocardial catheter approaches. 1 Recent meta-analyses have supported the safety and superior efficacy of surgical ablation,3 compared with catheter-based techniques. 4 However, the relative efficacy of BA vs. LA lesions sets remains a decision-making challenge for modern clinicians and surgeons. The current study is the first meta-analysis of comparative studies assessing the benefits and risks of BA vs. LA surgical ablation for AF. The association of SR prevalence and freedom from AF with improved quality of life and survival outcomes in AF patients has been well documented. 5,6 Long-term restoration of SR is particularly necessary in AF patients where rate-control and pharmacological strategies have not provided adequate relief. In the present meta-analysis, SR prevalence was superior for BA at 6 (75 vs. 65%) and 1 months (77 vs. 70%). Freedom from AF was also significantly higher in the BA ablation cohort at 1 months (78 vs. 71%). However beyond 1 year, SR outcomes were comparable between BA and LA cohorts, consistent with previous studies such as by et al. 9 The M-H, Random, 95% CI 8.0%.43 [0.1, 7.71] Year % 5.6 [0.4, ] 18.1% 3.60 [0.73, 17.61] 15.6% 0.96 [0.17, 5.35] %.66 [0.58, 1.17] % 0.09 [0.00, 1.70] 14 8.% 1.36 [0.39, 4.76] % 1.70 [0.85, 3.39] Late mortality %.43 [0.1, 7.71] %.6% 1.5% 8.7%.11 [0.36, 1.4] 4.69 [1.00,.09] 0.17 [0.01, 3.1] 0.34 [0.14, 0.81] % 1.1 [0.30, 4.1] 6 1 Heterogeneity: t = 1.40; c = 11.91, df = 4 (P = 0.0); I = 66% Test for overall effect: Z = 0.16 (P = 0.87) M-H, Random, 95% CI Favours BA Favours LA Figure Forest plot of early 30-day and late mortality, showing summary ORs with 95% confidence intervals for included studies. drop in freedom from AF in the BA group is likely because higher-risk patients, including long-standing/persistent AF, were preferentially chosen for the more extensive BA lesion set, and thus may translate into poorer rhythm and clinical outcomes at longer term follow-up. This would also justify differing results between the present study and a previous meta-analysis, which suggested 10% higher success rate in the BA group for up to 3 years postoperatively. While the late occurrence of the drop in freedom of AF in the present meta-analysis was surprising, this may be evidence supporting the inherent efficacy of the BA ablation approach, which was mitigated beyond 1 months by poorer rhythm outcomes of long-standing or persistent AF patients. Given the comparable outcomes beyond 1 year and significantly shorter CPB, the current evidence suggests that additional right-sided lesions do not confer long-term benefits with regards to efficacy. However, BA ablation may still represent an appropriate strategy for high-risk AF patients where immediate rhythm control is required. Atrial fibrillation has consistently been associated with elevated perioperative mortality and stroke risk in patients undergoing cardiac procedures. 6 As such surgical ablation for AF treatment is predominantly performed with concomitant cardiac surgery, with comparable short and long-term mortality risks demonstrated in recent meta-analyses. 7 However, these studies were limited to assessing the mortality outcomes of ablation with various energy sources and lesion sets, thus making it difficult to ascertain the relative mortality risks of BA and LA approaches. In the present study, 30-day

8 Meta-analysis of BA vs. LA surgical ablation 45 Study or Subgroup BA Events Total LA Events Total Weight M-H, Random, 95% CI Year (A) Permanent Pacemakers Takami % 1.36 [0.11, 16.05] % 6.18 [0.9, ] Not estimable % 1.87 [0.79, 4.4] % 1.39 [0.40, 4.85] % 4.67 [0.6, 85.35] Pecha % % Heterogeneity: t = 0.00; c =.95, df = 6 (P = 0.8); I =0% Test for overall effect: Z =.65 (P = 0.008) (B) Neurological Events % 9.1% 9.% 13.7% Heterogeneity: t = 0.3; c = 4.91, df = 4 (P = 0.30); I =19% Test for overall effect: Z = 0.48 (P = ) and late mortality outcomes were not significantly different between BA and LA cohorts, ranging from 3 5%. At 30-day follow-up there was a trend of higher mortality in the BA group by 64% compared with the LA ablation group, however this did not reach significance. This is likely related to the fact that the majority of studies recruited higher risk, long-standing, persistent AF patients for the BA approach (89 vs. 69%), while recent-onset and paroxysmal AF patients were selected for the LA approach. While late mortality outcomes were heterogeneous (I ¼ 66%), we attribute this partially to the varying follow-up periods of the included studies, ranging from 1 to 4 years. Additionally, other complications such as neurological events and reoperation for bleeding were not significantly different between BA and LA cohorts (Figure 3). Therefore, the additional 0.% 1.03 [0.41,.60].17 [1.11, 4.4] 1.75 [1.16,.66] Not estimable [0.4, ].64 [0.58, 1.04] [0.00, 1.76] 0.80 [0.07, 9.1] % 1.18 [0.35, 4.03] % 1.7 [0.48, 3.33] (C) Reoperation for Bleeding %.43 [0.1, 7.71] % 0.49 [0.04, 5.60] % 10.8% 11.0% 1.7% 0.79 [0.1, 3.00] 0.1 [0.01,.17] [0.64, ] 3.45 [0.60, 19.69] % 1.30 [0.45, 3.75] 3 17 Heterogeneity: t = 0.55; c = 4.91, df = 5 (P = 0.19); I =3% Test for overall effect: Z = 0.49 (P = 0.63) M-H, Random, 95% CI Favours BA Favours LA Figure 3 Forest plot postoperative complications (A) permanent pacemakers, (B) neurological events, (C) reoperation for bleeding, showing summary ORs with 95% confidence intervals for included studies. right-sided lesion to LA ablation can be performed safely with no increases in the risk of mortality. Previous reports have raised concerns surrounding the elevated postoperative bradyarrhythmias ensued by BA lesion sets, requiring permanent pacemaker implantations. In a recent univariate and multivariate logistic regression analysis of 594 patients by Pecha et al., 5 the BA lesion set was found to be the only predictor of permanent postoperative pacemaker implantations within 30-day follow-up. However, the evidence base is inconsistent, with Gillinov et al. 7,8 reporting similar implantation rates of 5 1% for mixed LA and BA Cox-Maze IV procedures to full Cox-Maze III procedures. In the present meta-analysis, permanent pacemaker implantations were significantly higher in the BA group compared with the LA group,

9 46 K. Phan et al. thus justifying prior reported concerns of postoperative iatrogenicallyinduced bradycardia and other arrhythmias. 9 The postoperative bradycardia is not likely to be due to injury to the sinus-atrial (SA) or atrioventricular nodes, since individual sinus beats of normal SR arise from the region of the anatomic SA node as well as specific sites away from the anatomic SA node. 30 Instead, the higher incidence of pacemaker implantations in the BA group can be explained by this cohorthaving a more severe formof AF, thus resulting inincreased likelihood of sinus node dysfunction. However, due to the heterogeneity of techniques used in the included studies, such as PVI, radiofrequency, and cryoablation, the potential effect of different energy sources on pacemaker implantation rates must not be excluded. Overall, our findings suggest that the BA ablation approach is associated with increased risk of pacemaker implantations and should be used judiciously in high-risk AF patients. Strengths, limitations, and future directions This present study assessed the risks and benefits of studies, which directly compared BA and LA ablation cohorts, rather than the pooling of data from a mixture of comparative and non-comparative studies. This methodology will minimize potential bias based on different degrees of follow-up, sample size, clinical endpoint, and AF definitions between BA and LA ablation groups. However, a number of limitations should be acknowledged and the results of the present study should be interpreted with caution. Firstly, this meta-analysis includes prospective randomized trials as well as non-randomized comparative studies. Accordingly, observational studies with intrinsic sources of patient selection, surgical and publication bias were included in the present meta-analysis. From Table, a variety of different lesion sets for both LA and BA ablation across the included studies can be observed. The definition of sinus rhythm success and freedom from AF varied between the studies. Several studies 3,4,9,13 15 did not clearly state exclusion of patients on anti-arrhythmic medications from the success group. As such, these definitions of success is inconsistent with the HRS consensus statements. The lack of control for the use of antiarrhythmic drugs on follow-up is a source of heterogeneity, making it difficult to definitively assess the benefits and risks of either approach, thus undermining the validity of the current evidence base. The diversity of surgical ablation energy sources and the types of concomitant cardiac surgical procedures may also account for the significant heterogeneity detected in cross-clamp and CPB durations. It is also acknowledged that the unstandardized reporting of longterm follow-up for the included studies also reduces the robustness of the long-term outcomes (.1 year) of this meta-analysis. Furthermore, not all outcomes were reported according to HRS/EHRA/ ECAS consensus guidelines, further contributing to the difficulty in comparing unstandardized clinical endpoints. Large prospective registry studies and randomized trials are required to adequately address the issue of which lesion set is best for surgical ablation for AF. One of the ongoing projects of the Cardiothoracic Surgical Trials Network is a comparative effectiveness randomized trial, 31 with a nested randomized comparison between PVI and BA maze lesion sets for persistent and long-standing AF patients undergoing mitral valve surgery. A design for a randomized study of surgical ablation with subgroup comparison between BA, LA and PVI lesion sets as well as energy source has also been proposed, 3 with an estimated sample size of 51 patients in each group and 80% statistical power. The outcomes of such studies should provide greater insight into the role of the lesion set in influencing the longer-term efficacy of surgical ablation for AF. Conclusions Biatrial and LA ablations produced comparable 30-day and late mortality but LA was associated with significantly reduced permanent pacemaker implantation rates. Biatrial ablation appeared to be more efficacious than LA ablation in achieving SR at 1 year; however, such differences were not maintained beyond 1 year follow-up. Trends appear to be driven by the preferential selection of long-standing and persistent AF patients for the BA approach. Further validation is required by future randomized studies with larger patient sample sizes and longer follow-up duration. Supplementary material Supplementary material is available at Europace online. Conflict of interest: none declared. References 1. Go AS, Mozaffarian D, Roger VL, Benjamin EJ, Berry JD, Blaha MJ et al. Heart disease and stroke statistics 014 update: a report from the American Heart Association. Circulation Cox JL, Schuessler RB, D Agostino HJ Jr., Stone CM, Chang BC, Cain ME et al. The surgical treatment of atrial fibrillation. III. Development of a definitive surgical procedure. J Thorac Cardiovasc Surg 1991;101: Takami Y, Yasuura K, Takagi Y, Ohara Y, Watanabe T, Usui A et al. Partial maze procedure is effective treatment for chronic atrial fibrillation associated with valve disease. J Card Surg 1999;14: M, Akpinar B, Sanisoglu I, Sagbas E, Bayindir O. Intraoperative saline-irrigated radiofrequency modified Maze procedurefor atrial fibrillation. Ann Thorac Surg 00; 74:S Pecha S, Schafer T, Yildirim Y, Ahmadzade T, Willems S, Reichenspurner H et al. Predictors for permanent pacemaker implantation after concomitant surgical ablation for atrial fibrillation. J Thorac Cardiovasc Surg 014;147: Barnett SD, Ad N. Surgical ablation as treatment for the elimination of atrial fibrillation: a meta-analysis. J Thorac Cardiovasc Surg 006;131: Gillinov AM. Choice of surgical lesion set: answers from the data. Ann Thorac Surg 007;84: Downs SH, Black N. The feasibility of creating a checklist for the assessment of the methodological quality both of randomised and non-randomised studies of health care interventions. J Epidemiol Community Health 1998;5: LK, CedolaSR, CoganJ, JiangJ, YangJ, TakayamaHet al. Rightatriallesions donot improve the efficacy of a complete left atrial lesion set in the surgical treatment of atrial fibrillation, but they do increase procedural morbidity. J Thorac Cardiovasc Surg 013;145:356 61; discussion A, Peterffy M, Kallner G. Learningwhatworks insurgical cryoablationof atrial fibrillation: results of different application techniques and benefits of prospective follow-up. Interact Cardiovasc Thorac Surg ;13: JB, Bang JH, Jung SH, Choo SJ, Chung CH, Lee JW. Left atrial ablation versus biatrial ablation in the surgical treatment of atrial fibrillation. Ann Thorac Surg ;9: ; discussion PM, Kruse J, Shalli S, Ilkhanoff L, Goldberger JJ, Kadish AH et al. Where does atrial fibrillation surgery fail? Implications for increasing effectiveness of ablation. J Thorac Cardiovasc Surg 010;139: Deneke T, Khargi K, Voss D, Lemke B, Lawo T, Laczkovics A et al. Long-term sinus rhythm stability after intraoperative ablation of permanent atrial fibrillation. Pacing Clin Electrophysiol ;3: J, Meng X, Li H, Cui Y, Han J, Xu C. Prospective randomized comparison of left atrial and biatrial radiofrequency ablation in the treatment of atrial fibrillation. Eur J Cardiothorac Surg ;35:116.

10 Meta-analysis of BA vs. LA surgical ablation V, Kumar S, Javali S, Rajesh TR, Pai V, Khandekar J et al. Efficacy of three different ablative procedures to treat atrial fibrillation in patients with valvular heart disease: a randomised trial. Heart Lung Circ 008;17: Calkins H, Kuck KH, Cappato R, Brugada J, Camm AJ, Chen SA et al. 01 HRS/ EHRA/ECAS Expert Consensus Statement on Catheter and Surgical Ablation of Atrial Fibrillation: recommendations for patient selection, procedural techniques, patient management and follow-up, definitions, endpoints, and research trial design. Europace 01;14: Wudel JH, Chaudhuri P, Hiller JJ. Video-assisted epicardial ablation and left atrial appendage exclusion for atrial fibrillation: extended follow-up. Ann Thorac Surg 008; 85: Pruitt JC, Lazzara RR, Ebra G. Minimally invasive surgical ablation of atrial fibrillation: the thoracoscopic box lesion approach. J Interv Card Electrophysiol 007;0: La Meir M, Gelsomino S, Lucà F, Pison L, Colella A, Lorusso R et al. Minimal invasive surgery for atrial fibrillation: an updated review. Europace 013;15: Fragakis N, Pantos I, Younis J, Hadjipavlou M, Katritsis DG. Surgical ablation for atrial fibrillation. Europace 01;14: Mahapatra S, LaPar DJ, Kamath S, Payne J, Bilchick KC, Mangrum JM et al. Initial experience of sequential surgical epicardial-catheter endocardial ablation for persistent and long-standing persistent atrial fibrillation with long-term follow-up. Ann Thorac Surg ;91: Phan K, Xie A, Tian DH, Shaikhrezai K, Yan TD. Systematic review and meta-analysis of surgical ablation for atrial fibrillation during mitral valve surgery. Ann Cardiothorac Surg 014;3: Phan K, Xie A, La Meir M, Black D, Yan TD. Surgical ablation for treatment of atrial fibrillation in cardiac surgery: a cumulative meta-analysis of randomised controlled trials. Heart 014;100: Kearney K, Stephenson R, Phan K, Chan WY, Huang MY, Yan TD. A systematic review of surgical ablation versus catheter ablation for atrial fibrillation. Ann Cardiothorac Surg 014;3: Corley SD, Epstein AE, DiMarco JP, Domanski MJ, Geller N, Greene HL et al. Relationships between sinus rhythm, treatment, and survival in the Atrial Fibrillation Follow-Up Investigation of Rhythm Management (AFFIRM) Study. Circulation 004; 109: Wolf PA, Mitchell JB, Baker CS, Kannel WB, D Agostino RB. Impact of atrial fibrillation on mortality, stroke, and medical costs. Arch Intern Med 1998;158: Cheng DC, Ad N, Martin J, Berglin EE, Chang BC, Doukas G et al. Surgical ablation for atrial fibrillation in cardiac surgery: a meta-analysis and systematic review. Innovations (Phila) 010;5: Gillinov AM, Bhavani S, Blackstone EH, Rajeswaran J, Svensson LG, Navia JL et al. Surgery for permanent atrial fibrillation: impact of patient factors and lesion set. Ann Thorac Surg 006;8:50 13; discussion Robertson JO, Cuculich PS, Saint LL, Schuessler RB, Moon MR, Lawton J et al. Predictors and risk of pacemaker implantation after the Cox-maze IV procedure. Ann Thorac Surg 013;95:015 0; disussion Worku B, PakSW, Cheema F, Russo M, Housman B, VanPattenDet al. Incidence and predictors of pacemaker placement after surgical ablation for atrial fibrillation. Ann Thorac Surg ;9: Gillinov AM, Argenziano M, Blackstone EH, Iribarne A, DeRose JJ Jr., Ailawadi G et al. Designing comparative effectiveness trials of surgical ablation for atrial fibrillation: experience of the Cardiothoracic Surgical Trials Network. J Thorac Cardiovasc Surg ;14:57 64 e.