Article Text
PDF
Abstract
Introduction Concomitant surgical ablation is a treatment modality for patients with atrial fibrillation (AF) undergoing cardiac surgery, however, its efficacy and clinical outcomes are not well established. The present study is the first cumulative meta-analysis of randomised controlled trials (RCT) on clinical outcomes of surgical ablation versus no ablative treatment in all patients with cardiac surgery.
Methods Electronic searches were performed using six databases from their inception to October 2013, identifying all relevant RCTs comparing surgical ablation versus no ablation in patients with AF undertaking cardiac surgery. Data were extracted and analysed according to predefined clinical endpoints.
Results Sixteen relevant RCTs were identified for the present study. Higher prevalence of sinus rhythm in the surgical ablation group was evident at all ≥12 month follow-up (OR, 6.72; 95% CI 4.88 to 9.25; p<0.00001). There were no significant differences between surgical ablation versus no ablation in terms of mortality (OR, 1.05; 95% CI 0.66 to 1.68; p=0.83), pacemaker implantations (OR, 0.88; 95% CI 0.51 to 1.51; p=0.64), and neurological events (OR, 0.86; 95% CI 0.37 to 2.04; p=0.74). Cumulative meta-analysis demonstrated that these trends have remained consistent over the years, with recent studies narrowing the CIs of the summary estimates.
Conclusions The evaluation of the current randomised trials demonstrates that concomitant surgical ablation and cardiac surgery is safe and effective at restoring sinus rhythm.
Statistics from Altmetric.com
Introduction
Atrial fibrillation (AF) is the most common cardiac arrhythmia, characterised by chaotic electrical activity and the lack of coordinated contractions in the atria. In addition to our understanding of the increasing mortality rates over the past two decades, AF has a projected prevalence of more than 12 million by 2050 in the USA alone.1––3 The first-line use of antiarrhythmic drugs to control AF has been shrouded amidst controversy, due to reports of limited efficacy, poor patient compliance and contraindications.4 ,5 As such, AF treatment has primarily been focused on electrically isolating the underlying pathology.
The traditional Cox–Maze procedure was pioneered by Dr James L Cox and represents the gold standard treatment for AF with the highest reported success rates.6 This procedure consists of a series of complex biatrial incisions via median sternotomy and cardiopulmonary bypass, and aims to interrupt the multiple re-entrant circuits and fibrillatory conduction of AF. Despite its effectiveness, the Cox–Maze has not been universally adopted due to the technical complexity involved, increased operation durations and high risk of bleeding.7––9 The use of alternative lesion sets and energy sources, such as radiofrequency and cryothermal energy, have simplified the maze procedure by replacing the complex surgical incisions with lines of transmural necrosis.
Randomised studies and systematic reviews to date have not provided conclusive evidence regarding the role of surgical ablation in a diverse cardiac surgical population.10––13 In order to explore the indications, risks, benefits and effects of technique and lesion sets, the present cumulative meta-analysis aims to summarise the available randomised evidence regarding the clinical outcomes of surgical ablation in AF patients undergoing cardiac surgery.
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 October 2013. To achieve the maximum sensitivity of the search strategy, we combined the terms: ‘atrial fibrillation’ AND ‘ablation OR pulmonary vein OR maze OR radiofrequency’ AND ‘randomised controlled trial’ as either key words or medical subject headings (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. Expert academic cardiothoracic surgeons (MLM, TDY) were consulted as to whether they knew of any unpublished data.
Selection criteria
Eligible randomised controlled trials (RCT) for the present systematic review and meta-analysis included those in which patient cohorts underwent any cardiac surgery concomitantly with surgical ablative treatment of atrial fibrillation. Studies that did not include sinus rhythm or AF-free survival as endpoints 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. All publications were limited to those involving human subjects. For articles not in English, further clarification was sought from the English-translated abstract and original authors to see if inclusion and exclusion criteria were met. Abstracts, case reports, conference presentations, editorials, reviews and expert opinions were excluded.
Data extraction and critical appraisal
All data were extracted from article texts, tables and figures, and assessed qualitatively using tools recommended by the Cochrane Collaboration for risk of bias, by two reviewers. Agreement was measured using simple agreement and kappa statistics. Discrepancies between the two reviewers were resolved by discussion and consensus. The final results were reviewed by senior investigators.
Statistical analysis
Clinical outcomes were assessed using standard and cumulative meta-analysis techniques, with the OR as a summary statistic of the raw data extracted from each included study. For standard and cumulative meta-analysis, both fixed-effect and random-effect models were tested. When both models provided comparative outcomes, then random-effects model statistics were reported. χ2 Tests were used to study heterogeneity between trials. I2 statistic was used to estimate the percentage of total variation across studies, owing to heterogeneity rather than chance, with values greater than 50% considered as substantial heterogeneity.14 If there was substantial heterogeneity, the possible clinical and methodological reasons for this were explored qualitatively.
Evidence of publication bias was sought using methods of Egger et al15 and Begg and Mazumdar16 Contour-enhanced funnel plot was performed to aid in interpretation of the funnel plot. Possible asymmetry was investigated using trim-and-fill analysis.17
All p values were twosided. All statistical analyses were conducted with Review Manager V.5.2.1 (Cochrane Collaboration, Software Update, Oxford, UK) and Comprehensive Meta-Analysis 2.2 (Biostat, Englewood, New Jersey, USA).
Results
Literature search
A total of 1126 references were identified through six electronic database searches. Manual search of reference lists yielded three new studies. After exclusion of duplicate or irrelevant references, 871 potentially relevant articles were retrieved. After detailed evaluation of these articles, 35 studies remained for assessment. After applying the selection criteria, 16 RCTs were selected for analysis (see online supplementary figure 1). The agreement κ statistic for study inclusion was good (κ=0.88; SE, 0.064). In these 16 studies, 1082 patients underwent procedures that involved cardiac surgery with surgical ablation (SA) (cut and sew (CS)+SA group; n=607) or without surgical ablation (CS group; n=475) (table 1).
Summary of RCTs comparing CS+SA versus CS-only surgical treatment in patients with AF
Baseline and operation characteristics
Males accounted for 24–68% of patients undergoing CS+SA and 20–82% undergoing CS alone (weighted mean, WM, 52% vs 55%; p=0.22). The average age ranged between 19years and 74 years, and 37years and 75 years old (WM, 62.6 vs 63.6; p=0.97) for CS+SA and CS groups, respectively. There was also no significant difference between CS+SA and CS for LVEF and left atrial diameter (LAD). Baseline characteristics for hypertension, diabetes and prior strokes were not significantly different between treated and control arms (see online supplementary table 1).
Cardiopulmonary bypass (CBP) time was significantly longer when CS surgery was performed concomitantly with surgical ablation. With the exception of three studies which did not report CBP, all other studies reported CBP times which demonstrated longer average CBP for the CS+SA group (see online supplementary table 2). Similarly, cross-clamp time was significantly longer for the CS+SA group compared to CS.
Quality assessment
All the included studies were RCTs.18–30 ,33 ,44 Seven studies had >50 patients,19 ,22–24 ,27 ,28 ,44 while the remaining studies had less than 50 patients.18 20 21 25 26 29 30 33 34 Six studies used radiofrequency ablation,18 ,19 ,22 ,27 ,29 ,30 four studies used Cox–Maze CS,20 ,21 ,24 ,25 three studies used cryoablation,23 ,28 ,44 five studies reported patients undergoing pulmonary vein isolation,20 ,21 ,24 ,28 ,44 and two studies used microwave ablation.26 ,34 Results for coronary artery bypass grafting (CABG) were reported in eight studies,22 ,23 ,25 ,26 ,29 ,33 ,34 ,44 while valvular surgery was studied in all RCTs except Pokushalov et al.33
The following outcomes are consistent with those recommended by the Heart Rhythm Society, European Rhythm Association and European Cardiac Arrhythmia Society consensus.31 ,32 Four studies reported follow-up of greater than 1 year,24 ,25 ,28 ,33 while the remaining studies had follow-up of 12 months or less.18–23 ,26 ,27 ,29 ,30 ,33 ,34 ,44 Sinus rhythm was the primary endpoint in 13 studies,18 ,21–30 ,34 ,44 while AF-free survival was the primary endpoint in three studies.25 ,26 ,33 Antiarrhythmic drugs (AAD) were indicated for patients with restored sinus rhythm (SR) from cardioversion in four studies,19 ,24 ,25 ,26 for postoperative AF or flutter in four studies,20 ,21 ,23 ,28 given to all patients prophylactically, and weaned off in four studies,19 ,24 ,25 ,26 and not used at all in two studies (table 2).33 ,34
Freedom from atrial fibrillation and antiarrhythmic drugs at 1 year for cardiac surgery with or without concomitant surgical ablation
Using tools recommended by the Cochrane Collaboration for the risk of bias (see online supplementary figure 2), a high risk of ‘other bias’ was concluded for Pokushalov et al,33 for using a more liberal cut-off value for AF burden compared to guidelines. High risk of ‘other bias’ was also concluded for Knaut et al,34 who confirmed lesion transmurality visually and not based on electrophysiological mapping. Based on the risk of bias assessment, study-removal sensitivity analysis was performed. However, this did not lead to any significant changes in the reported outcomes.
Assessment of safety
Mortality
The risk of 30-day all-cause mortality was not significantly different between CS+SA and CS groups at 30 days (5.3% vs 3.8%; OR, 1.23; 95% CI 0.65 to 2.39; p=0.53; I2=0%; table 3). Furthermore, all-cause mortality was also not significantly different (8.7% vs 7.6%; OR, 1.05; 95% CI 0.66 to 1.68; p=0.83; I2=0%; table 3). No significant heterogeneity was observed in these two comparisons. Cumulative meta-analysis of 30-day mortality and all-cause mortality outcomes showed that the OR and point values have stabilised, with decreasing CIs (figure 1) time.
Summary of clinical outcomes using standard meta-analysis techniques
Cumulative forest plot of the OR of 30-day mortality (A) and all-cause mortality (B) in patients with atrial fibrillation with surgical ablation (CS+SA) or without ablation (CS). The cumulative estimate of the OR and 95% CI of each trial corresponds to the middle of the squares and the horizontal line. On each line, the cumulative number of events as a fraction of the cumulative total number randomised is shown for both treatment and control groups. CS, cut and sew; SA, surgical ablation.
Neurological events
Neurological events were reported in 10 out of 16 included RCTs, with comparable results between CS+SA and CS groups (4.9% vs 5.8%; OR, 0.86; 95% CI 0.37 to 2.04; p=0.74; I2=29%; table 3). Even up to recent studies, the cumulative evidence demonstrates no difference in stroke between CS+SA and CS groups (figure 2).
Cumulative forest plot of the OR of incidence of neurological events (A), pericardial tamponades (B), and pacemaker implantations (C) in patients with atrial fibrillation with surgical ablation (CS+SA) or without ablation (CS). The cumulative estimate of the OR and 95% CI of each trial corresponds to the middle of the squares and the horizontal line. On each line, the cumulative number of events as a fraction of the cumulative total number randomised is shown for both treatment and control groups. CS, cut and sew; SA, surgical ablation.
Pacemaker implants
Thirteen out of 16 studies reported outcomes for pacemaker implantation. Overall, there was no difference in pacemaker implantations whether surgical ablation was performed or not (5.8% vs 8.3%; OR, 0.88; 95% CI 0.51 to 1.51; p=0.64; I2=0%; table 3).
Other morbidities
The frequency of perioperative tamponade, or pericardial effusion (figure 2), was significantly less for CS+SA compared with CS (2.1% vs 9.0%; OR, 0.25; 95% CI 0.08 to 0.82; p=0.02; I2=0%; table 3). While reoperative bleeding was previously suggested to be a risk factor for surgical ablation, this study showed no such difference between the treated and control groups (7.6% vs 5.6%; OR, 1.28; 95% CI 0.64 to 2.55; p=0.48; I2=0%). Additionally, AAD at follow-up was comparable between both groups (31.3% vs 44.2%; OR, 0.51; 95% CI 0.23 to 1.12; p=0.09; I2=76%).
Assessment of efficacy
The number of patients in SR at discharge was significantly higher in the CS+SA group compared to CS group (62.7% vs 26.6%; OR, 7.64; 95% CI 4.04 to 14.45; p<0.00001; I2=67%). The CS+SA group also had a significantly higher proportion of patients in SR compared to CS only at 3-month (62.5% vs 29.8%; OR, 4.79; 95% CI 2.79 to 8.23; p<0.00001; I2=56), 6-month (62.6% vs 27.4%; OR, 5.44; 95% CI 3.44 to 8.61; p<0.00001; I2=37%), and ≥12-month (66.7% vs 26.1%; OR, 6.72; 95% CI 4.88 to 9.25; p<0.00001; I2=0%) follow-up periods (table 3). Subgroup analysis of the different surgical ablation techniques demonstrated no significant difference affecting SR outcomes (see online supplementary figure 4). Cumulative meta-analysis demonstrated that earlier studies had higher heterogeneity and favoured surgical ablation over cardiac surgery only with respect to SR prevalence (figure 3).
Cumulative forest plot of the OR of sinus rhythm (SR) in patients with atrial fibrillation with surgical ablation (CS+SA) or without ablation (CS) at discharge (A), 6-month (B) and ≥12-month follow-up. The cumulative estimate of the OR and 95% CI of each trial corresponds to the middle of the squares and the horizontal line. On each line, the cumulative number of events as a fraction of the cumulative total number randomised is shown for both treatment and control groups. *p<0.0001. CS, cut and sew; SA, surgical ablation.
Publication bias
Inspection of the funnel plot (see online supplementary figure 3) did not show significant asymmetry for all-cause mortality. Trim-and-fill analysis suggested that four studies were missing, however, this did not significantly affect the overall effect size with Begg's test score of p=0.3674 and an Egger's test score of p=0.0654 (see online supplementary figure 3A). Sensitivity analyses for all-cause mortality showed similar results for all subgroups but when poor or small trials were excluded. These results suggest that publication bias was not a significant influencing factor.
Funnel plot for 1-year sinus rhythm outcomes showed significant asymmetry, confirmed by Trim-and –fill analysis showing seven missing studies (see online supplementary figure 3B). This significantly affected effect size, with Begg's test score of p=0.03434 and Egger's test score of 0.00272. These results suggest that publication bias was a possible influencing factor. However, subsequent sensitivity analyses for 1-year sinus rhythm showed similar results for all subgroups but when poor or small trials were excluded.
Discussion
Research efforts into the underlying pathophysiology and therapies for AF have empowered clinicians with the knowledge and tools to tackle this significant healthcare burden. For patients with AF and concurrent indications for CS, concomitant SA emerged as an effective treatment option. The efficacy and clinical outcomes of concomitant SA and CS are not well established. Thus, the present study aims to report the clinical outcomes of CS+SA versus CS alone via a cumulative meta-analysis of RCTs.
The restoration of SR in AF patients represents a crucial therapeutic strategy, given that SR is a profound, independent predictor of quality of life and survival.35––38 The present meta-analysis demonstrated a higher SR prevalence in the CS+SA group at discharge, 3-month, 6-month, and beyond 12-month follow-up compared to the CS group, consistent with previous studies.10 ,28 Furthermore, the present cumulative meta-analyses demonstrates that both magnitude and precision of long-term >1 year SR point values have remained essentially unchanged since the RCTs published in 2004. There is consistent evidence that concomitant surgical ablation is superior to cardiac surgery only in terms of long-term SR restoration in AF patients.
In the present review, acceptable 30-day mortality rates (range: 0–15%) and all-cause mortality rates (range: 0–17%) were reported in the included RCTs. No significant difference was found between CS+SA and CS groups in terms of 30-day and all-cause mortality. Furthermore, cumulative meta-analysis showed that these trends have stabilised for the last decade, suggesting maturation of trials used to evaluate surgical ablation over time. This suggests that further small randomised trials will not substantially add to the evidence base, but rather emphasises the need for pragmatic, well-designed randomised studies with adequate power and sample sizes.
Previous meta-analyses by Khargi et al12 and Cheng et al10 have argued that permanent pacemaker implantations remained high following concomitant surgical AF ablation, suggesting that postoperative iatrogenically induced bradycardia and arrhythmias still warrant concern.10 ,12 However, there was significant heterogeneity in their analysis due to the inclusion of non-RCTs, and the need for permanent pacemakers have been reported anywhere from 6% to 56% in the literature, depending on lesion patterns and energy sources used.39––43 This was not reflected in the present meta-analysis, with no difference in permanent pacemaker implantations in the CS+SA group compared to the CS-only group.
AF is a major risk factor for strokes and thromboembolism, and thus elimination of AF via surgical ablation may have a protective effect against these risks. From 10 RCTs and 735 patients, no significant difference was found in terms of neurological events between the CS+SA and CS alone groups. However, the effect of anticoagulation therapy on stroke outcomes was difficult to ascertain, which varied among the included RCTs. This may account for underestimated stroke incidence, and therefore the meta-analysis results should be viewed with caution. Overall, the current study demonstrates that superior SR can be achieved with surgical ablation without additional risk of neurological events.
There are several limitations to this cumulative meta-analysis. Subgroup analysis based on the type of cardiac surgery, energy source or lesion set of surgical ablation was not feasible due to the low number of RCTs in each subgroup, the lack of comparative data for different lesion sets, as well as type of AF pathology. These outcomes will be crucial for resolving the current debate in the literature, propelled by conflicting results on the efficacy of surgical ablation in patients with paroxysmal and persistent AF.44 ,45 Furthermore, AF monitoring was mainly based on ECG and 24 h Holter methods at 3-month or 6-month intervals, which may not have detected paroxysmal or asymptomatic recurrent episodes. Small sample size and inadequate statistical power of the included studies are also significant limitations. Potential reporting and publication bias and the varying skill-sets and experiences of surgeons may further exaggerate the overall effect size of the endpoints studied in this meta-analysis. Additionally, the narrow range of clinical outcomes were reported by the RCTs, with only six out of the 16 included studies reporting outcomes at follow-up >1 year.
To overcome these key constraints and inadequate power of available RCTs in the evidence base, we propose a potential design for a large pragmatic randomised study. The study endpoints will be established in accordance with the recommendations of consensus statements from the American Heart Association, American College of Cardiology, and European Heart Rhythm Association.1 ,32 The primary endpoint hypothesis is that surgical ablation for atrial fibrillation would result in greater restoration of sinus rhythm at 1-year follow-up compared to cardiac surgery without ablation. Other endpoints studied should include mortality rates, neurological events, anticoagulation, antiarrhythmic therapies, and the need for implantable cardioverter defibrillator and pacemaker devices. These outcomes should be assessed at 1-year and 5-year follow-ups.
According to previous publications and the current meta-analysis, we assumed that the sinus rhythm restoration rate would be 70% in the ablation group and 30% in the non-ablation cardiac surgery group. The sample size required to detect this difference with 95% CI and 80% statistical power is 23 patients in each group, with effect size of 0.75. However, to facilitate subgroup analysis with regards to energy source (CS vs radiofrequency vs cryoablation vs microwave) and lesion sets (biatrial vs left atrial vs pulmonary vein isolation), the level of significance α should be adjusted to α* the overall type I error rate, resulting in α* value of 0.42%. Furthermore, the smallest primary endpoint proportion within the subgroup categories was 57.5% rather than 70%.28 To detect this, smaller effect size of 0.4 with the adjusted overall error rate requires a sample of 218 in each group. The expected attrition rates due to drop-out and failure to obtain endpoint information at 1-year and 5-year follow-up should also be compensated for, increasing the required sample size for each group to 251. As such, this proposed trial design will be adequately powered to detect any efficacy differences between various energy sources and lesion sets used in concomitant surgical ablation and cardiac surgery.
In summary, we conclude that surgical ablation is a viable treatment for AF during concomitant cardiac surgery without increased mortality or morbidity risks. Short-term, mid-term and long-term prevalence of SR are significantly improved in patients who undergo surgical ablation. However, limited sample sizes and inadequate power of available randomised evidence means that there is still uncertainty surrounding choice of technique and lesion sets, which should be addressed in large, multi-institutional and adequately powered randomised studies.
Key messages
-
What is already known on this subject?
-
Surgical ablation is often performed during cardiac surgery to treat atrial fibrillation. However, there is limited evidence regarding the efficacy and morbidities associated with surgical ablation in concomitant cardiac surgery.
-
What this study adds?
-
This is the first cumulative meta-analysis of randomised evidence on clinical outcomes of surgical ablation versus no ablative treatment in all cardiac surgical patients. This study demonstrates that concomitant surgical ablation and cardiac surgery is safe and effective at restoring sinus rhythm. Cumulative meta-analysis suggests that further small randomised studies are unlikely to substantially add to the current evidence base.
-
How might this impact on clinical practice?
-
Concomitant surgical ablation and cardiac surgery should be considered as a safe and efficacious treatment method for atrial fibrillation. Unanswered questions surrounding the choice of energy source and lesion sets should be addressed in future, pragmatic randomised trials.
References
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Files in this Data Supplement:
- Data supplement 1 - Online supplement
Footnotes
-
Contributors KP, TDY conceptualised the idea. KP, AX carried out the database search, data extraction and data analysis. KP, AX, ML and TDY contributed to the reporting. ML, DB and TDY provided expert opinions and supervision. All authors have read and approved the final version of the manuscript.
-
Funding The authors received no specific funding of this project.
-
Competing interests None.
-
Provenance and peer review Not commissioned; externally peer reviewed.
-
Data sharing statement All data analysed within this study are available from the corresponding papers. All data used is visible within this paper and the online supplementary figures.
Linked Articles
- Heartbeat