Minimally invasive thoracoscopic surgical ablation and hybrid ablation in AF patients: results from a single-arm systematic review and meta-analysis Authors: Fouad Hanna, Ali Dway, Hamza A Abdul-Hafez, Yousef M Husseiny, Ahmed Youssef Hassan, Mazen Negmeldin A Yassin, Basel Saber Hussein Ayoub, Hadeel Jameel Ayesh, Omneya Kandil, Abdelrahman Elgendy, Nada G Hamam, Abdelrahman El-Helbawy Published: 2026 ABSTRACT AbstractBackgroundAtrial fibrillation (AF) is the most frequent cardiac arrhythmia in adults, with presentations from asymptomatic to severe cardiovascular morbidities. Cardiac ablation treats patients resistant to antiarrhythmic drugs (AAD) using approaches such as bilateral surgical ablation, which permits targeting of superior vena cava lesions but entails longer procedures and higher complication risk, and hybrid ablation that combines epicardial and endocardial techniques.MethodWe performed a meta-analysis per PRISMA guidelines, including observational and interventional single-arm studies of patients undergoing bilateral surgical or hybrid ablation in single- and double-setting subgroups. Studies were selected by predefined criteria, and data were extracted independently by 12 investigators. Searches covered multiple databases up to March 29, 2025.ResultThirty-two studies with 1,871 patients were evaluated. Overall complication rates were 0.11 (95% CI: 0.08–0.14) for surgical ablation and 0.13 (95% CI: 0.095–0.17) for hybrid ablation, with no significant subgroup difference (p = 0.093). Arrhythmia-free rates were 0.81 (95% CI: 0.69–0.89) for surgical and 0.76 (95% CI: 0.70–0.85) for hybrid ablation (p = 0.305). Mean operative times were 160.9 min (95% CI: 135 − 186.9) for surgical and 192.65 min (95% CI: 109.3–275.9) for hybrid ablation (p = 0.069). Bleeding rates were 0.03 (95% CI: 0.02–0.04) and 0.049 (95% CI: 0.03–0.07), respectively (p = 0.06). Mortality, stroke, conversion to sternotomy, and hospital length of stay showed no significant differences.ConclusionSurgical ablation yielded a slightly higher arrhythmia-free rate and fewer total complications, while hybrid ablation had shorter operative time; subgroup (single vs. double setting) differences were not significant.Supplementary InformationThe online version contains supplementary material available at 10.1186/s13019-026-04135-4.Keywords: Atrial fibrillation, Hybrid ablation, Bilateral thoracoscopic ablation Introduction The most common cardiac arrhythmia in adults is atrial fibrillation (AF), which is estimated to have a prevalence of 60 million cases all over the world [1]. The clinical manifestations of AF may be asymptomatic or present with chest pain, palpitations, and dyspnea. Moreover, patients have the risk of developing significant cardiovascular morbidities, including cardiogenic shock, ischemic stroke, and sudden cardiac death [2]. These negative outcomes place a substantial burden on global healthcare systems, and therefore, it is warranted to comprehensively evaluate the management of symptomatic AF [3, 4]. According to ESC/ACC/AHA guidelines, AF management includes rhythm control (via cardioversion or ablation) plus rate control, anticoagulation, and risk factor modification. Antiarrhythmic drugs (AADs) remain a cornerstone in rhythm management but are limited by poor long-term efficacy, frequent adverse effects, and life-threatening toxicities [3, 5, 6]. The Cox-Maze IV procedure remains the surgical treatment of choice for paroxysmal or persistent AF, particularly in patients undergoing concomitant cardiac surgery [5, 7]. However, the need for cardiopulmonary bypass and a high complication rate hinder the approach for many patients [8]. Therefore, minimally invasive ablation approaches were needed to avoid the risks associated with traditional surgical ablation. Catheter ablation has emerged as a superior alternative to medical and surgical therapy for paroxysmal or persistent AF patients, particularly those who failed medical treatment and have contraindications to such therapies [5]. However, the long-term maintenance of sinus rhythm for non-paroxysmal (persistent and long-standing persistent) AF is less than optimal, and the need for repeat ablation procedures is common [9–12]. Considering the limitations and challenges, hybrid ablation has been developed with the aim of combining the advantages of thoracoscopic surgical ablation (direct visualization of anatomical structures and possibility to isolate epicardial lesions) and endocardial catheter ablation (possibility to isolate endocardial lesions). Several studies have reported favorable outcomes of hybrid ablation for persistent AF, establishing it as a valuable treatment approach for non-paroxysmal AF [13, 14]. However, the ideal access and timing of epicardial and endocardial ablation remain debatable. Hybrid ablation can be performed by a unilateral or bilateral approach, with significant variations between the two accesses. Bilateral thoracoscopic approach allows targeting of superior vena cava lesions, although it is associated with a longer procedural duration and an increased risk of complications [15]. On the other hand, a unilateral approach minimizes surgical trauma and is associated with a shorter procedural duration [15]. Comparative studies of both thoracoscopic accesses are lacking, with the bulk of findings derived from case series or retrospective cohort studies with a small sample size. Additionally, the timing of epicardial and endocardial ablation ranges from same-day to staged procedures. It has been suggested that delaying endocardial ablation after epicardial ablation may allow for resolution of tissue inflammation, which impedes conduction across linear epicardial ablation and thus increases the risk of atrial arrhythmia recurrence [16, 17]. Similarly, few studies have compared timing strategies for the thoracoscopic approach, leaving a knowledge gap that remains to be filled. In the paroxysmal and persistent AF population, the optimal approach remains unclear. Data derived mainly from retrospective cohort studies and case series are conflicting. Therefore, we performed this systematic review and meta-analysis to pool efficacy and safety data from the available literature, with a primary focus on comparing the outcomes of bilateral thoracoscopic surgical ablation versus hybrid ablation techniques. Methods Guidelines and ethics This systematic review and meta-analysis were carried out according to the International Prospective Register of Systematic Reviews (PROSPERO ID: RD420251107935) [18]. The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement and the Cochrane Handbook of Systematic Reviews and Meta-analyses of Interventions (version 5.1.0) were strictly followed during the development of this study [19]. Literature review Information Sources and Search Strategy We comprehensively searched PubMed’, Web of Science’, ‘Scopus’, and Cochrane Central Register of Controlled Trials (from inception to August, 2025). The keywords used for the participants were [‘‘Atrial Fibrillation’’ or ‘‘AF’’]. This was combined with the keywords for the procedures, which included [‘‘Minimally Invasive Cardiac Ablation,’’ ‘‘Thoracoscopic Ablation,’’ ‘‘Video-Assisted Thoracoscopic Surgery,’’ ‘‘VATS,’’ ‘‘Mini-Maze Procedure,’’ ‘‘Cox-Maze Procedure,’’ ‘‘Surgical Ablation,’’ ‘‘Hybrid AF Therapy,’’ or ‘‘Convergent Ablation’’]. Selection process and data management All electronic databases’ records retrieved through the literature search were imported into EndNote (Clarivate Analytics, PA) to remove duplicates. Next, all obtained citations were exported to Rayyan software (Qatar Computing Research Institute, 2016, Doha, Qatar) for initial title and abstract screening [20] and subsequent full-text screening of eligible titles and abstracts. Two authors performed the screening process, and the results were thoroughly revised. Disagreements between reviewers were resolved through discussion with a senior author. Data were extracted to standardized Google Sheets [Alphabet Inc. (Google), USA]. Subsequently, a senior author reviewed the sheets to identify any inconsistencies in the collected data. The extracted data comprised the following items: [1] characteristics of the included studies [2], characteristics of the included studies’ population, and [3] outcome measures, like sinus rhythm conversion, arrhythmia-free survival, AF recurrence, and complications (i.e., stroke, bleeding, phrenic nerve injury, mortality). Eligibility criteria A research question was formulated using the Population, Intervention, Comparison, and Outcome (PICO) [21] method to determine our study group. This was followed by a set of inclusion and exclusion criteria, with no limitations according to age, gender, race, country, or date of publication. Every two authors screened the studies independently, and conflicts were resolved by discussion. Studies were included in our systematic review and meta-analysis if they satisfied the following criteria: (1) Population: adult patients (> 18 years of age) suffering from paroxysmal, persistent, or long-standing Atrial Fibrillation indicated for surgical correction. (2) Intervention: Studies that included bilateral radiofrequency Thoracotomy Surgical Ablation or Single and multiple setting Hybrid Ablation. (3) Outcomes: Studies that reported at least one of the following outcomes: sinus rhythm conversion, arrhythmia-free survival, AF recurrence, and complications (i.e., stroke, bleeding, phrenic nerve injury, mortality). (4) Study design: RCTs, non-RCTs, prospective or retrospective observational studies published in English. We excluded book chapters, editorials, letters, conference abstracts, case reports, and reviews. Outcome definition The primary outcome was procedural success, which was defined as patient freedom from arrhythmia. Freedom from arrhythmia was assessed from the end of the procedure (for the surgical ablation group) or after the final stage of therapy (for the hybrid ablation group). It was characterized by the absence of any symptomatic or asymptomatic episodes of atrial fibrillation during follow-up monitoring, which could also allow for the discontinuation of antiarrhythmic medications [22]. Additional primary outcomes related to resource utilization were the duration of the operative procedure and the length of hospital stay. The secondary outcomes were related to postoperative adverse events, which were as follows: Bleeding incidence, stroke, and conversion to sternotomy. Risk of bias assessment The risk of bias (ROB) of all included studies was precisely assessed through 2 blinded reviewers using the NIH Quality Assessment Tool, which is explicitly designed for single-arm studies. This tool is composed of nine main domains: Clarity of stating the Objective, Eligibility Criteria of each study, Representative of the Study Population, Consistency of the Intervention, Validity of the Outcomes, Blinding of Outcome Assessors, follow-up Rates and Attrition, Statistical Analysis, and Handling of Missing Data. Authors’ judgements for each domain were grouped as ‘Low risk,’ Some concerns,’ or ‘High risk’ of bias. A third expert author discussed and resolved any differences in data extraction and bias assessment. Strategy for data synthesis The Comprehensive Meta-Analysis (CMA) software (Version 3, Biostat, Inc.) was used to perform the statistical analysis. The data synthesis should have focused on the synthesis of pooled summary estimates of each outcome separately in both the surgical ablation and hybrid ablation cohorts due to the nature of the included studies, which were all single-arm (they did not have a direct comparator group). To compute the rate of events (arrhythmia-free survival, complication rates), the event rate was determined as the percentage of patients who experienced the event. Random-effects model was used as the pooled analysis to take into consideration the heterogeneity among the studies, where it is assumed that the true effect sizes differ across the populations that are included. A logit transformation was used to calculate the pooled estimates and 95% confidence intervals. This transformation ensures that the variance is stabilized and that the pooled proportions remain within the 0 to 1 range. In case of continuous outcomes (e.g., operative time, length of stay in a hospital), the pooled mean estimate was calculated using the mean value presented in each of the studies. As with the dichotomous results, a random-effects model was used, and the findings are presented as the pooled mean with a 95% CI. Heterogeneity was evaluated and taken as statistically significant when the p-value was below 0.10. The Q (chi-square test) of Cochran and the degree of freedom (DF) were also determined to test the heterogeneity. The hybrid ablation group was pre-specified to undergo subgroup analysis to determine the impact of the timing of the procedures (single-setting and double-setting) on the results. Tests of statistical significance (p-value < 0.05) were made between subgroups. It was assumed that the study-to-study variance ( tau squared) was equal across all subgroups - this was calculated within subgroups and then pooled across subgroups. When one fails to find a statistically significant difference between subgroups, this should not be construed as an indication that the effect is identical in subgroups because statistical power to find subgroup differences is usually low. To investigate potential sources of between-study heterogeneity, univariable random-effects meta-regression analyses were performed using the method-of-moments (MM) estimator with Z-distribution inference, analyses were limited to outcome–intervention combinations with at least 10 contributing studies (k ≥ 10), consistent with recommendations for adequate moderator estimation. Continuous moderators were examined: mean patient age and mean left ventricular ejection fraction (LVEF). Both were grand-mean centred prior to analysis to reduce collinearity with the intercept and to yield clinically interpretable regression coefficients (β). Moderator significance was assessed using Cochran’s Qmodel test, with statistical significance set at p < 0.05. Residual between-study heterogeneity was quantified by τ² and I² after moderator adjustment. The proportion of between-study variance explained by each model was estimated using the R² analogue. Given the exploratory nature and the modest number of studies per model, findings are considered hypothesis-generating. The focus was made on the size of the point estimate and the accuracy of the 95% CI instead of the use of p-values alone. All the tests were two-sided, and a p-value of less than 0.05 was regarded as significant, except the test of heterogeneity (p < 0.10). Results Selection and characteristics (search results) The literature search identified a total of 4428 studies: Cochrane Central Register of Controlled Trials (n = 23), Scopus (n = 1192), Web of Science (n = 1050), and PubMed (n = 2163). After removing duplicates (n = 1377), 3051 studies were obtained for title-and-abstract screening. According to our inclusion/exclusion criteria, we excluded 2427 papers, and 624 studies were obtained for full-text screening. The full-text screening was done, and only 32 papers were eligible for inclusion from the 4 databases. Finally, 32 studies with a total of 1,871 patients were included in this meta-analysis after full-text review [23–43]. (Fig. 1) A PRISMA flow chart of our search through different databases Studies characteristics and baseline demographics The 32 studies focus on single-arm studies for patients undergoing bilateral surgical or hybrid ablation in single and double settings. (Table 1). All baseline demographics of 1,871 patients are described in Table 2. (Tables 2 and 3) Table 1Study characteristicsStudy authorPublication yearCountryStudy designType of interventionAim of the studyInclusion CriteriaGulkarov 2019 [1]2019USARetrospective cohortHybrid ablationTo show that the hybrid procedure is a relatively safe and effective option for patients with PAFPatients were symptomatic and had either persistent or long‑standing persistent AF that had failed at least one antiarrhythmic drug therapyHarlaar 2021 [2]2021NetherlandsRetrospective cohortBilateral thoracoscopic ablationTo assess the long-term outcomes of thoracoscopic pulmonary vein and box ablation in LSPAFPatients with symptomatic LSPAF who underwent thoracoscopic ablation at the Leiden University Medical CenterLa Meir 2013 [3]2013NetherlandsRetrospective cohortBilateral thoracoscopic ablationTo demonstrate the beneficial effect of minimally invasive RF surgical ablation on left atrial remodeling using 2D speckle-tracking echocardiographyParoxysmal AF treated successfully by RF bipolar surgical ablation for lone AF with a history (> 6 months) of paroxysmal recurrent AF with symptoms, in stable sinus rhythm in the 48 h preceding ablationMa 2021 [4]2021ChinaRetrospective cohortHybrid ablationTo investigate the immediate electrophysiological characteristics following modified ablation (3 circular and 3 linear lesions) via a unilateral approachPatients with LsPAF who underwent the single-stage hybrid procedure (unilateral thoracoscopic procedure + percutaneous EP examination and catheter ablation)Maclean 2020 [5]2020UKRetrospective cohortHybrid ablationTo provide a real-world insight into the risks and benefits of the convergent procedure in patients with longstanding persistent AFPatients with longstanding persistent atrial fibrillationGuo 2019 [6]2019ChinaRetrospective cohortBilateral thoracoscopic ablationTo explore an effective, reproducible, and less invasive surgical approach for lone AFLone AF patients with symptoms who had failed AAD therapy; unresponsive to cardioversion or catheter ablation; history of stroke/cerebral thrombosis; LVEF > 30%La Meir 2011 [7]2011NetherlandsRetrospective cohortBilateral thoracoscopic ablationTo present the results of minimally invasive surgical treatment for lone AF (LAF)Patients with LAF who underwent video-assisted bilateral RF ablation for LAFPojar 2014 [8]2014Czech RepublicRetrospective cohortBilateral thoracoscopic ablationTo report experience with minimally invasive thoracoscopic irrigated RF box-lesion ablation of AFPatients requiring non-pharmacologic treatment of AF; preference for minimally invasive surgery; earlier failure of catheter PVI; failed to maintain sinus rhythm on ≥ 1 AADTan 2021 [9]2021ChinaProspective cohortHybrid ablationTo evaluate the relationship between intraprocedural arrhythmia termination and long-term outcomes of hybrid ablation in long-standing persistent AFSymptomatic PsAF or LS-PsAF; age < 75; ≥1 failed catheter ablation if AF duration < 1 year; LA > 50 mm if AF 1–3 year; any duration > 3 year eligibleAl-Jaziri 2019 [10]2019NetherlandsProspective cohortHybrid ablationTo study efficacy and safety of single-stage hybrid AF ablation in persistent AF or paroxysmal AF with failed endocardial ablationSymptomatic persistent or LSPAF; or paroxysmal AF with ≥ 2 failed catheter ablations. Excluded: prior cardiac surgery, BMI ≥ 35Belluschi 2021 [11]2021ItalyProspective cohortBilateral thoracoscopic ablationTo assess the long-term durability of surgical pulmonary vein isolation (PVI) beyond 7 yearsHighly symptomatic paroxysmal stand-alone AF, refractory to ≥ 1 class I/III AAD, no concomitant structural disease. Excluded: chest wall deformities, previous thoracotomies, severe lung disease, obesity (BMI ≥ 35)Bisleri 2013 [12]2013ItalyProspective cohortHybrid ablationTo assess electrophysiological modifications related to the hybrid approach and their correlation with mid-term resultsPatients with long-standing persistent AFBudera June 2017 [13]2017Czech RepublicProspective cohortHybrid ablationTo summarize complications of all thoracoscopic ablation procedures and compare with similar recent papersNo other severe structural heart disease; pericardial adhesions and secondary AF causes were contraindicationsBudera September 2017 [14]2017Czech RepublicCohort studyHybrid ablationTo assess safety and mid-term efficacy of two-staged hybrid ablation using the Versapolar RF surgical catheterSymptomatic non-paroxysmal AF without significant structural, valvular, and coronary heart diseaseBulava 2017 [15]2017Czech RepublicProspective cohortBilateral thoracoscopic ablationTo analyze efficacy of epicardially created lesions using bipolar RF energy and describe typical reconduction sitesSymptomatic drug-refractory persistent or long-standing persistent AFDe Maat 2015 [16]2015NetherlandsRetrospective cohortBilateral thoracoscopic ablationTo report the long-term efficacy and safety of bilateral surgical PVI in patients with paroxysmal lone AFSymptomatic paroxysmal AF without concomitant structural disease, refractory to class I/III AAD or failed transcatheter PVIEdgerton July 2009 [17]2009USARetrospective cohortBilateral thoracoscopic ablationTo examine results of minimally invasive surgery incorporating RF ablation lines (Dallas lesion set) in persistent and long-standing persistent AFPatients with persistent AF or long-standing persistent AFEdgerton December 2009 [18]2009USAProspective cohortBilateral thoracoscopic ablationTo evaluate outcomes of a new minimally invasive technique with bilateral PVI, mapping, ablation of ganglionic plexi, and LAA excisionPatients with paroxysmal AF, persistent AF, or long-standing AF (grouped per HRS/EHRA/ECAS Expert Consensus Statement)Gehi 2013 [19]2013USACohort studyHybrid ablationTo evaluate 1-year outcomes of a hybrid epicardial-endocardial technique for AF ablationAF with disabling symptoms, failed class I/III AAD, and one of: failed endocardial ablation; persistent AF with large LA or structural disease; or long-standing persistent AFGeuzebroeka 2015 [20]2015NetherlandsRetrospective cohortBilateral thoracoscopic ablationTo describe results and learning curve of a totally thoracoscopic left atrial Maze (TT-Maze) performed by one surgeonPatients who underwent TT-Maze; excluded those with prior heart or lung surgeryChoi 2023 [21]2023KoreaRetrospective cohortBilateral thoracoscopic ablationTo determine the minimum number of ablation lines for complete PV isolation during thoracoscopic ablationPersistent or long-standing persistent AF, and medically refractory paroxysmal AFDe Maat 2013 [22]2013Netherlands & ItalyRetrospective cohortBilateral thoracoscopic ablationTo report midterm efficacy and safety of surgical minimally invasive PVI (SMI PVI) for symptomatic drug-refractory AFSymptomatic drug-refractory paroxysmal or persistent AF, or after failed transcatheter PVIMagni 2021 [23]2021NetherlandsProspective cohortHybrid ablationTo evaluate efficacy and safety of first-line single-stage hybrid ablation for (long-standing) persistent AFPatients undergoing single-stage hybrid ablation as first-line treatment for PsAF or long-standing PsAF (January 2015–December 2018)Mannakkara 2022 [24]2022UKRetrospective cohortHybrid ablationTo evaluate outcomes following convergent ablation at a single UK tertiary centreTwo-stage convergent ablation for AF (2012–2019) at St. Thomas’ Hospital, London; had trialled or were unsuitable for class I/III AADMie 2014 [25]2014ChinaProspective cohortBilateral thoracoscopic ablationTo evaluate a novel thoracoscopic ablation procedure for lone AFSymptomatic AF aged 30–81, drug-refractory, failed catheter ablation, history of stroke/embolism, or LVEF ≥ 30%Saini 2017 [26]2017USAProspective cohortBilateral thoracoscopic ablationTo provide data on long-term outcomes of minimally invasive surgical ablation of AFMedically refractory symptomatic AF who had failed ≥ 1 AAD or catheter ablation, or were not candidates for catheter ablationSebaga 2025 [27]2025FranceRetrospective cohortHybrid ablationTo describe outcomes of patients with HFrEF and long-standing PsAF who underwent the convergent procedurePatients with HFrEF referred for long-standing PsAF convergent ablation at two hospital centresWojtaszczyk 2017 [28]2017PolandProspective cohortHybrid ablationTo evaluate safety and effectiveness of hybrid ablation in long-term observationDrug-refractory symptomatic PsAF or long-standing PsAFWudel 2008 [29]2008USARetrospective cohortBilateral thoracoscopic ablationTo evaluate long-term efficacy of video-assisted thoracoscopic PVI and left atrial appendage exclusionSymptomatic intermittent AF who underwent video-assisted thoracoscopic PVI and LAA exclusionYan 2022 [30]2022ChinaRetrospective cohortBilateral thoracoscopic ablationTo determine safety and efficacy of a modified off-pump biatrial mini-maze procedure for long-standing persistent AFPatients with long-standing PsAF who underwent modified mini-maze using bipolar RF ablationSalzburg 2019 [31]2019SwitzerlandProspective cohortBilateral thoracoscopic ablationTo report experience with the AF HeartTeam approach for surgical stand-alone AF ablationFailed catheter ablation or not suitable for such; contraindication to anticoagulationCompier 2016 [32]2016NetherlandsProspective cohortBilateral thoracoscopic ablationTo evaluate feasibility of bipolar thoracoscopic epicardial LA box isolation with bidirectional conduction block in longstanding persistent AFConsecutive patients with symptomatic drug-refractory longstanding persistent AF Study characteristics Table 2Demographic patient characteristicsAuthorTotal patientsAge, mean (± SD)Male, n (%)DM, n (%)HTN, n (%)Coronary Heart Disease, n (%)Heart Failure, n (%)Gulkarov 2019 [1]3162.67 (10.88)21 (67.7)11 (35.5)31 (100)7 (22.6)22 (71)Harlaar 2021 [2]7758.9 (7.7)60 (77.9)2 (3)39 (51)7 (9)N/ALa Meir 2013 [3]3364.6 (6.9)28 (84.8)N/A18 (54.5)N/AN/AMa 2021 [4]4061.4 (9.4)30 (75)N/AN/AN/A31 (77.5)Maclean 2020 [5]4368.7 (8.5)32 (74.4)6 (14.0)N/A5 (11.6)N/AGuo 2019 [6]5455.5 (9.0)39 (72.2)6 (11.1)26 (48.1)3 (5.6)N/ALa Meir 2011 [7]2867.1 (9.1)24 (85.7)0 (0)14 (50.0)N/AN/APojar 2014 [8]4157.6 (8.0)31 (75.6)N/AN/AN/AN/ATan 2021 [9]5059.9 (8.2)38 (76)N/AN/A18 (36)N/AAl-Jaziri 2019 [10]5057 (9)38 (76)4 (8)23 (46)2 (4)2 (4)Belluschi 2021 [11]5055 (11.2)42 (84)N/AN/AN/A1 (2)Bisleri 2013 [12]4562.3 (9.8)33 (73.3)14 (31.1)34 (75.6)N/AN/ABudera June 2017 [13]11261.4 (8.8)74 (66)21 (19)69 (62)8 (7)22 (20)Budera September 2017 [14]4255 (10.0)42 (76)N/A13 (33)N/AN/ABulava 2017 [15]3058 (9.0)30 (86.66)N/AN/AN/AN/ADe Maat 2015 [16]4255 (10)32 (76)N/A14 (33)N/AN/AEdgerton July 2009 [17]11459.5 (10.6)79 (69.3)19 (19.2)62 (61.6)26 (25.3)N/AEdgerton December 2009 [18]3058 (9.0)26 (86.7)N/AN/AN/AN/AGehi 2013 [19]10162.9 (9.6)79 (78.2)19 (19.2)62 (61.6)26 (25.3)29 (28.7)Geuzebroeka 2015 [20]8259.9 (8.6)58 (71)N/AN/A0 (0)N/AChoi 2023 [21]2058.7 (10.17)19 (95)4 (20)8 (40)N/A6 (30)De Maat 2013 [22]8654 (11.0)67 (78)N/A39 (35)0 (0)0 (0)Magni 2021 [23]4957 (8.5)44 (89.5)6 (12.2)21 (42.9)5 (10.2)4 (8.2)Mannakkara 2022 [24]6761.7 (11.3)59 (88.1)12 (17.9)26 (38.8)N/A7 (11.9)Mie 2014 [25]5559.2 (15.4)39 (70.9)N/AN/AN/AN/ASaini 2017 [26]10962.7 (9.3)60 (55)13 (11.9)65 (59.6)109 (100)N/ASebaga 2025 [27]4359 (21.47)37 (86)4 (9)25 (58)N/A6 (14)Wojtaszczyk 2017 [28]9054.8 (9.8)70 (78)12 (13.3)60 (66.8)N/AN/AWudel 2008 [29]2262.7 (8.9)15 (68.2)N/A19 (86.4)1 (4.5)1 (4.5)Yan 2022 [30]10262.1 (8.1)74 (72.5)16 (15.7)56 (54.9)N/AN/ASalzburg 2019 [31]4265 (7)14 (33)13 (31)32 (76)12 (29)6 (19)Compier 2016 [32]2558 (7)19 (76)1 (4)7 (28)3 (12)N/AAbbreviations:DM = Diabetes mellitus; HTN = Hypertension; CAD = Coronary artery disease; HF = Heart failure; SD = Standard deviation; N/A = Not available Demographic patient characteristics Abbreviations:DM = Diabetes mellitus; HTN = Hypertension; CAD = Coronary artery disease; HF = Heart failure; SD = Standard deviation; N/A = Not available Table 3Cardiac-related patient characteristicsAuthorTotal patientsPersistent AF, n (%)Long-standing persistent AF, n (%)Paroxysmal AF, n (%)LA size (mm), mean (SD)LV Ejection fraction, mean (SD)Duration of AF (years), mean (SD)Re-ablated Patients, n (%)Gulkarov 2019 [1]3116 (51.6)15 (48.3)0 (0)43 (4.6)48.6 (15.5)2.2 (2.5)N/AHarlaar 2021 [2]770 (0)77 (100)N/A46 (13)N/A4 (3.3)12 (16)La Meir 2013 [3]330 (0)0 (0)33 (100)N/A60.1 (7.1)6.3 (2.1)N/AMa 2021 [4]400 (0)40 (100)0 (0)46 (4.6)60 (7.1)4.3 (2.6)12 (33)Maclean 2020 [5]4343 (100)N/A0 (0)47.09 (5.9)51.8 (10.4)3 (2.2)N/AGuo 2019 [6]5412 (22.2)40 (74.1)2 (3.7)42.6 (5)63.4 (4.8)4.1 (4.7)7 (13)La Meir 2011 [7]285 (17.8)9 (32.2)14 (50)42.2 (5.2)60.1 (7.1)12.1 (3.7)13 (46.4)Pojar 2014 [8]419 (22.0)8 (19.5)24 (58.5)45.3 (5.9)N/A5.1 (4.15)N/ATan 2021 [9]50N/A44 (88.0)6 (12)46.1 (7)62 (9.1)8.7 (5.9)15 (30)Al-Jaziri 2019 [10]5034 (68)11 (22)5 (10)40 (11)55 (6)5.2 (4.8)25 (50)Belluschi 2021 [11]50N/AN/A50 (100)41 (6.2)60 (4.6)N/A28 (56)Bisleri 2013 [12]45N/A45 (100)N/A51.3 (9.7)56.3 (7.0)6.9 (5.7)N/ABudera June 2017 [13]11237 (33)59 (53)16 (14)45.2 (5.5)55 (10.7)3.1 (3.8)13 (12)Budera September 2017 [14]420 (0)0 (0)42 (100)42 (5)58 (6)6.3 (12.6)9 (21)Bulava 2017 [15]3010 (33.33)20 (66.7)N/A52N/AN/AN/ADe Maat 2015 [16]42N/AN/A42 (100)42 (5)58 (6)~ 2 year (median only) †9 (21)Edgerton July 2009 [17]11432 (28.1)22 (19.3)60 (52.6)47.2 (8.3)54.2 (9.5)5.9 (5.5)24 (21.1)Edgerton December 2009 [18]3010 (33.3)20 (66.7)0 (0)52 (N/A)N/AN/AN/AGehi 2013 [19]10147 (46.5)37 (36.6)17 (16.8)51 (10)50.2 (9.8)5.9 (5.5)36 (35.6)Geuzebroeka 2015 [20]82N/AN/A41 (50)45 (6)56 (9)5.3 (IQR) †31 (38)Choi 2023 [21]20N/A19 (95)1 (5)46.47 (7.95)56.6 (8.3)7.3 (5.8)3 (15)De Maat 2013 [22]8612 (14)N/A74 (86)42 (6)56 (6)N/A †15 (17)Magni 2021 [23]4932 (65.3)17 (34.7)0 (0)N/A52.4 (7.5)3.3 (3)N/AMannakkara 2022 [24]6767 (100)0 (0)0 (0)N/A49.1 (9.4)3.2 (2.7)26 (38.8)Mie 2014 [25]5514 (25.45)3 (5.45)38 (69.09)N/AN/A7.6 (7.2)N/ASaini 2017 [26]10949 (44.95)0 (0)60 (55.04)N/A53.8 (8.9)3.6 (4)24 (22)Sebaga 2025 [27]430 (0)43 (100)0 (0)N/A35.3 (9.2)2 (4)N/AWojtaszczyk 2017 [28]9039 (43)51 (57)0 (0)45 (6)48.6 (9.7)4.5 (3.7)18 (20)Wudel 2008 [29]220 (0)0 (0)22 (100)34.7 (11.5)54.8 (3.6)3.7 (3.4)7 (31.8)Yan 2022 [30]1020 (0)102 (100)0 (0)45.7 (5.4)63.1 (6.3)5.7 (5.0)N/ASalzburg 2019 [31]42N/AN/A15 (35.7)48 (10)60 (40–77)N/AN/ACompier 2016 [32]250 (0)25 (100)0 (0)45 (5)N/A1.8 (0.9)3 (12)Abbreviations: PsAF = Persistent atrial fibrillation; LSPAF = Long-standing persistent AF; PAF = Paroxysmal AF; LA = Left atrium; LVEF = Left ventricular ejection fraction; AF dur = AF duration; N/A = Not available; IQR = Interquartile range Cardiac-related patient characteristics Abbreviations: PsAF = Persistent atrial fibrillation; LSPAF = Long-standing persistent AF; PAF = Paroxysmal AF; LA = Left atrium; LVEF = Left ventricular ejection fraction; AF dur = AF duration; N/A = Not available; IQR = Interquartile range Risk of bias assessment Risk of bias was assessed using the NIH Quality Assessment Tool. The methodological quality assessment of the included studies indicated a generally low risk of bias, with approximately 78% of the cohort achieving an overall “low risk” rating, while only 10% and 12% were categorized as high or unclear risk, respectively. High rigor was consistently observed across several domains, particularly regarding the clarity of research objectives (D1), representative study populations (D3), and the standardized delivery of interventions (D6). However, systemic high risk was identified in the blinding of participants and outcome assessors (D8 and D12), reflecting a universal limitation likely inherent to the nature of the interventions studied. Furthermore, nearly 50% of the studies exhibited high or unclear risk regarding sample size adequacy (D5), suggesting potential underpowering in a portion of the literature. Despite these limitations and the identification of three specific studies (Edgerton 2 2009, Maclean 2020, and Wojtaszczyk 2017) with an overall high risk of bias, the high degree of compliance in most methodological domains suggests that the synthesized evidence provides a reliable basis for the review’s conclusions. (Fig. 2) (Supplementary Fig. 1). Risk of Bias assessment plot Operative time A meta-analysis of single-arm studies evaluated the overall mean operative time in minutes for patients undergoing bilateral surgical or hybrid ablation for treating paroxysmal and persistent AF. (Fig. 3) Operative time Forest plot Surgical ablation analysis includes fourteen studies with 821 patients. Based on a random effects model, the overall mean of operative time among the surgical ablation group was 160.9 min (95% CI: 135–186.9). Significant heterogeneity was present (P < 0.00001, I2 = 99%). In the subgroup analysis of Hybrid ablation by the type of setting (double setting vs. single setting), seven studies with 389 patients were analysed. The overall mean operative time in the hybrid ablation group was 192.65 min (95% CI: 109.3–275.9) using a random effects model. Also, significant heterogeneity was present (p < 0.00001, I2 = 99%). Subgroup mean operative time varied; four studies (n = 199) contributed data to the double setting subgroup, while three studies (190 participants) contributed to the single setting subgroup. 125.5 min for the double setting (95% CI: 15.3–235.8), 281.9 min for the single setting (95% CI:154.7–409.2). Each subgroup displayed significant heterogeneity (I2, 99.4 to 99.9%), with no significant difference in operative times across different settings (Q = 3.31, d.f. = 1, p = 0.06). The overall mean operative time was higher in the hybrid ablation group compared to the surgical ablation group (mean = 192 min VS 160 min, respectively). Arrhythmia free A meta-analysis of single-arm studies evaluated the overall event rate of arrhythmia-free patients undergoing bilateral surgical or hybrid ablation for treating paroxysmal and persistent AF. (Fig. 4) Arrhythmia Free Forest plot Surgical ablation analysis includes sixteen studies with 863 patients. The overall rate of arrhythmia-free patients who went for surgical ablation was 83% (Event rate: 0.83, 95% CI: 0.73–0.90%), using a random effects model. Significant heterogeneity was present (P < 0.00001, I2 = 99%). In the subgroup analysis of Hybrid ablation by the type of setting, thirteen studies with 700 patients were analysed. The overall arrhythmia-free rate in the hybrid ablation group was 75% (Event rate: 0.75, 95% CI: 0.69–0.81%), using a random effects model. Also, significant heterogeneity was present (p < 0.00001, I2 = 99%). Subgroup arrhythmia-free event rate varied, six studies (309 participants) in the double setting subgroup, and seven studies (391 participants) in the single setting subgroup. The event rate % ranged from 75% (95% CI: 0.69–0.81%) for the double setting to 78% (95% CI: 70–85%) for the single setting. Each subgroup still displayed heterogeneity (p = 0.01, I2, 64 to 66%), with no significant difference in arrhythmia-free rate across different settings (Q = 1.05, d.f. = 1, p = 0.305). The overall event rate of arrhythmia-free was higher in the surgical ablation group compared to the hybrid ablation group (Event rate: 83% vs. 75%, respectively). Arrhythmia-free rates were assessed over a weighted mean follow-up of 30.0 months for surgical ablation (19 studies, n = 1,161) and 20.4 months for hybrid ablation (13 studies, n = 703), with monitoring performed predominantly via serial ECG and 24-hour Holter recording at structured intervals across all included studies (Supplementary Table 1). Overall complications A meta-analysis of single-arm studies evaluated the overall event rate of complications for patients undergoing bilateral surgical or hybrid ablation for treating paroxysmal and persistent AF. (Fig. 5) Overall complications Forest plot Surgical ablation analysis includes nineteen studies with 1154 patients. The overall rate of complications among patients who underwent surgical ablation was 11% (Event rate: 0.115, 95% CI: 0.08–0.14%), using a random effects model. Regarding the variation of effect size across populations, there was heterogeneity observed (P = 0.003, I2 = 53.5%). In the subgroup analysis of Hybrid ablation by the type of setting, twelve studies with 689 patients were analysed. A total of 85 patients had complications, and the overall complication rate was 13% (Event rate: 0.13, 95% CI: 0.09–0.17%), using a random effects model. Also, significant heterogeneity was present (p = 0.001, I2 = 63.5%). Subgroup arrhythmia-free event rate varied. Six studies (309 participants) contributed data to the double setting subgroup, and six studies (380 participants) contributed data to the single setting subgroup. The event rate % ranged from 17% (95% CI: 10–25%) for the double setting to 9% (95% CI: 6–15%) for the single setting. Each subgroup still displayed heterogeneity (p = 0.03, I2 = 57.7% to p = 0.09, I2 = 47.3%), with no significant difference in overall complications rate across different settings (Q = 2.83, d.f. = 1, p = 0.093). The overall event rate of complications was higher in the hybrid ablation group compared to the surgical ablation group (Event rate: 13% vs. 11%, respectively). The 11% overall complication rate for surgical ablation comprised phrenic nerve injury (1.7%), pleural or pericardial effusion (0.9%), bleeding requiring re-exploration (0.9%), pneumothorax (0.5%), stroke or TIA (0.5%), conversion to sternotomy (0.6%), and mortality (0.3%) (Supplementary Table 2). Mortality A meta-analysis of single-arm studies evaluated the overall event rate of mortality for patients undergoing bilateral surgical or hybrid ablation for treating paroxysmal and persistent AF. (Fig. 6) Mortality Forest plot Surgical ablation analysis includes eighteen studies with 1134 patients. The overall event rate of mortality for patients who went for surgical ablation was 1% (Event rate: 0.013, 95% CI: 0.008–0.023%), using a random effects model. There was heterogeneity observed (P = 0.003, I2 = 53.5%). In the subgroup analysis of Hybrid ablation by the type of setting, eleven studies with 583 patients were analyzed. A total of 3 patients died, and the overall mortality rate was 1% (Event rate: 0.013, 95% CI: 0.006–0.027), using a random effects model. Also, there was no heterogeneity observed (p = 0.99, I2 = 0.00%). Subgroup arrhythmia-free event rate varied, four studies (172 participants) in the double setting subgroup, and seven studies (411 participants) in the single setting subgroup. However, there was no significant difference between the pooled overall mortality rate among double and single setting subgroups (Q = 0.05, d.f. = 1, p = 0.813). Each subgroup remained with no heterogeneity observed (p = 0.99, I2 = 0.00% to p = 0.99, I2 = 0.00%). There was no difference in mortality rate between the surgical ablation and hybrid ablation group (Event rate: 1% vs. 1%, respectively). Bleeding A meta-analysis of single-arm studies evaluated the overall event rate of bleeding for patients undergoing bilateral surgical or hybrid ablation for treating paroxysmal and persistent AF. (Supplementary Fig. 2) Surgical ablation analysis includes seventeen studies with 1109 patients. The overall event rate of bleeding among patients who went for surgical ablation was 2% (Event rate: 0.02, 95% CI: 0.01–0.03), using a random effects model. No heterogeneity was observed (P = 0.804, I2 = 0.00%). In the subgroup analysis of Hybrid ablation by the type of setting, eight studies with 484 patients were analyzed. A total of 19 patients had bleeding, and the overall complication rate was 4% (Event rate: 0.04, 95% CI: 0.03–0.07), using a random effects model. Also, no heterogeneity was present (p = 0.343, I2 = 11.2%). Subgroup bleeding event rate varied, three studies (154 participants) in the double setting subgroup, and five studies (330 participants) in the single setting subgroup. Also, Subgroup analysis indicated different event rate % ranged from 7% (95% CI: 4–13%) for the double setting to 3% (95% CI of 1–6%) for the single setting. Each subgroup remained with no heterogeneity (p = 0.29, I2 = 18.9% to p = 0.72, I2 = 0.00%), with no significant difference in overall bleeding rate across different settings (Q = 3.37, d.f. = 1, p = 0.06). There was no difference in the bleeding rate between the surgical ablation and hybrid ablation group (Event rate: 2% vs. 4%, respectively). Length of hospital stay A meta-analysis of single-arm studies evaluated the overall mean length of hospital stay for patients undergoing bilateral surgical or hybrid ablation for treating paroxysmal and persistent AF. (Supplementary Fig. 3) Surgical ablation analysis includes fifteen studies with 907 patients. The overall mean length of hospital stay among the surgical ablation group was 6.02 days (95% CI: 5.2–6.8), based on a random effects model. Significant heterogeneity was present (P < 0.00001, I2 = 98.2%). Hybrid ablation Subgroup analysis was performed according to the type of setting; six studies with 286 patients were analysed. The overall mean length of hospital stay in the hybrid ablation group was 5.6 days (95% CI: 4.6–6.6), using a random effects model. Also, significant heterogeneity was present (p < 0.00001, I2 = 95.14%). Subgroup mean length of hospital stay varied, three studies (156 participants) in the double setting subgroup, and three studies (130 participants) in the single setting subgroup. 5.3 days for the double setting (95% CI: 3.9–6.8), 5.9 days for the single setting (95% CI: 4.4–7.3). Heterogeneity in the single setting decreased to 31% with a p value of 0.23, suggesting a more consistent effect within this subgroup, while the double setting subgroup remained significantly heterogeneous (p < 0.00001, I2 = 97.6). Variations across different settings were not significant in length of hospital stay (Q = 0.274, d.f. = 1, p = 0.6). There was no difference in the mean length of hospital stay (days) between the surgical ablation and the hybrid ablation group (Mean: 6 days vs. 5.6 days, respectively). Stroke A meta-analysis of single-arm studies evaluated the overall mean of stroke for patients undergoing bilateral surgical or hybrid ablation for treating paroxysmal and persistent AF. (Supplementary Fig. 4) Surgical ablation analysis includes sixteen studies with 995 patients. The overall event rate of stroke among patients who went for surgical ablation was 1% (Event rate: 0.018, 95% CI: 0.011–0.030%), using a random effects model. No heterogeneity was observed (P = 0.920, I2 = 0.00%). In the subgroup analysis of Hybrid ablation by the type of setting, eleven studies with 404 patients were analyzed. A total of 6 patients had strokes, and the overall stroke rate was 2% (Event rate: 0.02, 95% CI: 0.01–0.04), using a random effects model. Also, no heterogeneity was present (p = 0.79, I2 = 0.00%). Subgroup analysis of stroke event rate was consistent, with four studies (194 participants) in the double setting subgroup and four studies (210 participants) in the single setting subgroup. However, there was no significant difference between the pooled overall stroke rate among double setting and single setting subgroups (Q = 0.059, d.f. = 1, p = 0.808) (ER: 2%, 95% CI: 0.008–0.05 vs. ER: 2%, 95% CI: 0.010–0.06; respectively). Each subgroup remained with no heterogeneity observed (p = 0.88, I2 = 0.00% to p = 0.36, I2 = 5.84%). There was no difference in stroke rate between the surgical ablation and the hybrid ablation group (Event rate: 1% vs. 2%, respectively). Conversion to sternotomy A meta-analysis of single-arm studies evaluated the overall event rate of conversion to sternotomy for patients undergoing bilateral surgical or hybrid ablation for treating paroxysmal and persistent AF. (Supplementary Fig. 5) Surgical ablation analysis includes fifteen studies with 974 patients. The overall event rate of conversion to sternotomy among patients who underwent surgical ablation was 1% (Event rate: 0.017, 95% CI: 0.010–0.030), using a random effects model. No heterogeneity was observed (P = 0.920, I2 = 0.00%). In the subgroup analysis of Hybrid ablation by the type of setting, eight studies with 501 patients were analyzed. The overall event rate of conversion to sternotomy in the hybrid ablation group was 1% (Event rate: 0.01, 95% CI: 0.008–0.03), using a random effects model. Also, no heterogeneity was present (p = 0.97, I2 = 0.00%). Subgroup analysis included four studies (221 participants) in the double setting subgroup and four studies (280 participants) in the single setting subgroup. The event rate was 2% (95% CI: 0.009–0.05) for the double setting and 1% (95% CI: 0.004–0.03) for the single setting. Each subgroup showed no heterogeneity (p = 0.92, I2 = 0.00% to p = 0.87, I2 = 0.00%), with no significant difference in conversion to sternotomy rate across different settings (Q = 0.606, df. = 1, p = 0.436). There was no difference in the conversion to sternotomy rate between the surgical ablation and hybrid ablation groups (Event rate: 1% vs. 1%, respectively). Meta-regression analysis Univariable random-effects meta-regression was performed across 12 models to assess whether mean patient age or mean LVEF explained between-study heterogeneity. No moderator reached statistical significance (all p > 0.05). Full model parameters are reported in Table (5). Table 5Meta-regression results: influence of age and LVEF on outcomes of surgical and hybrid cardiac ablation — random-effects modelsOutcomeInterventionModeratorkβ CoefficientSE95% CI Lower95% CI UpperZ-valuep-valueTau²I² (%)Q (Model)R² AnalogArrhythmia-Free survivalHybridAge130.04220.055-0.06550.150.770.4420.234166.850.590.00LVEF120.00380.028-0.0510.05870.140.8910.272470.230.020.00SurgicalAge16-0.01090.083-0.17350.1517-0.130.8951.120388.480.020.14LVEF120.17040.1187-0.06220.4031.440.1511.315489.752.060.00ComplicationsHybridAge12-0.04980.056-0.15960.06-0.890.3740.313764.280.790.00LVEF110.01060.0273-0.04290.0640.390.6990.176746.540.150.00SurgicalAge19-0.07030.0463-0.1610.0204-1.520.1290.186349.022.310.00LVEF14-0.03720.0604-0.15560.0812-0.620.5380.230456.430.380.00MortalitySurgicalAge18-0.01220.0914-0.19130.1669-0.130.894000.020.00LVEF13-0.08570.1048-0.29110.1198-0.820.414000.670.00Operative time (min)SurgicalAge14-1.9133.8067-9.37415.548-0.50.6152717.6999.120.250.00LVAD†10-9.76466.5692-22.63993.1107-1.490.1372124.5698.962.210.22k = number of studies; SE = standard error; CI = confidence interval; Tau² = between-study variance; I² = proportion of total variance due to heterogeneity; Q = Cochran’s Q-statistic for model significance; R² Analog = proportion of between-study variance explained by the model. All models used random-effects method-of-moments (MM) estimation with Z-distribution. p-values < 0.05 are bolded in red* Operative Time / LVEF model covariate is labelled ‘Centred LVAD’ in source data. † For the LVAD moderator in Operative Time, β represents minutes per unit increase in LVAD score Meta-regression results: influence of age and LVEF on outcomes of surgical and hybrid cardiac ablation — random-effects models k = number of studies; SE = standard error; CI = confidence interval; Tau² = between-study variance; I² = proportion of total variance due to heterogeneity; Q = Cochran’s Q-statistic for model significance; R² Analog = proportion of between-study variance explained by the model. All models used random-effects method-of-moments (MM) estimation with Z-distribution. p-values < 0.05 are bolded in red * Operative Time / LVEF model covariate is labelled ‘Centred LVAD’ in source data. † For the LVAD moderator in Operative Time, β represents minutes per unit increase in LVAD score For arrhythmia-free survival, neither age nor LVEF significantly predicted outcomes in the hybrid (age: β = 0.042, p = 0.442; LVEF: β = 0.004, p = 0.891) or surgical groups (age: β = −0.011, p = 0.895; LVEF: β = 0.170, p = 0.151). Residual heterogeneity remained high across all four models (I² = 66.9–89.8%), and the R² analogue was negligible, except for the surgical LVEF model (R² = 0.14). For overall complications in the surgical group, age showed a non-significant negative trend (β = −0.070, p = 0.129) and LVEF was non-significant (β = −0.037, p = 0.538). Neither moderator was significant in the hybrid group (age: p = 0.374; LVEF: p = 0.699). R² analogues were zero across all complication models. Surgical mortality showed no association with age (β = −0.012, p = 0.894) or LVEF (β = −0.086, p = 0.414). For operative time, neither age (β = −1.91 min/year, p = 0.615) nor LVEF (β = −9.76 min/unit, p = 0.137) was significant, although the LVEF model yielded the highest R² across all analyses (R² = 0.22). Overall, the meta-regression analyses did not identify age or LVEF as significant moderators of any outcome, and substantial residual heterogeneity persisted in all models, suggesting that unmeasured study-level factors — such as lesion set, or follow-up duration — are the principal drivers of between-study variability. Discussion In this systematic review and meta-analysis, we evaluated single-arm bilateral surgical ablation and hybrid ablation for paroxysmal and persistent atrial fibrillation (AF), reporting pooled procedural rates and outcomes, rather than offering head-to-head comparisons. Our study’s primary findings show that bilateral surgical ablation is associated with higher rates of freedom from arrhythmia (83% vs. 75%), and an improved safety profile with a lower complication rate overall (10% vs. 13%) and bleeding rates (2% vs. 4%). In addition, hybrid ablation had a shorter mean operative time (151 min vs. 192 min, respectively), particularly for the single-stage procedure. Of note, the critical outcomes, such as mortality, stroke, conversion to sternotomy, and lengths of hospital stay, were low in both groups, indicating that both were relatively safe in these aspects. In terms of effectiveness, the surgery arm had a better arrhythmia-free survival, highlighting the effectiveness of a pure surgical strategy. This finding presents a clear picture, with the majority of recent literature comparing hybrid ablation to catheter ablation, hybrid techniques showing increased long-term success, especially in long-standing persistent or persistent AF [15, 44]. (Table 4) Table 4Comparison of procedural outcomesOutcomeSurgical ablationHybrid ablation (Single setting)Hybrid ablation (Double setting)Arrhythmia-free Rate83% (95% CI: 73–90%)78% (95% CI: 70–85%)72% (95% CI: 61–80%)Overall Complication Rate11% (95% CI: 8–14%)9% (95% CI: 6–15%)17% (95% CI: 10–25%)Mortality Rate1% (95% CI: 0.8–2.3%)1% (95% CI: 0.6–3%)1% (95% CI: 0.3–4%)Bleeding Rate2% (95% CI: 1–3%)3% (95% CI: 1–6%)7% (95% CI: 4–13%)Stroke Rate1.8% (95% CI: 1.1–3.0%)2% (95% CI: 1.0–6%)2% (95% CI: 0.8–5%)Conversion to Sternotomy1.7% (95% CI: 1.0–2.9%)1% (95% CI: 0.4–3%)2% (95% CI: 0.9–5%)Operative Time (minutes, mean)160.9 (95% CI: 135–186.9)281.9 (95% CI: 154.7–409.2)125.5 (95% CI: 15.3–235.8)Hospital Stay (days, mean)6.02 (95% CI: 5.2–6.8)5.9 (95% CI: 4.4–7.3)5.3 (95% CI: 3.9–6.8)Summary of pooled results for arrhythmia-free survival, complications, and other key metrics across the three treatment strategies. Data are presented as % (95% CI) for proportions and mean (95% CI) for continuous variables Comparison of procedural outcomes Summary of pooled results for arrhythmia-free survival, complications, and other key metrics across the three treatment strategies. Data are presented as % (95% CI) for proportions and mean (95% CI) for continuous variables Our study demonstrates that hybrid ablation is an improvement upon endocardial catheter ablation but has yet to reach the level of effectiveness of a comprehensive surgical procedure. These findings align with a meta-analysis by Varzaly et al., which reported a rate of 79.4% sinus rhythm with hybrid ablation [14]. Our subgroup analysis surprisingly favored single-setting hybrid procedures, which demonstrated a higher arrhythmia-free rate (78%) compared to double-setting procedures (72%). This suggests that a comprehensive one-stage procedure may provide a more robust lesion set, despite requiring a longer operative time. Although hybrid ablation aims to balance invasiveness and procedure duration [14], this approach does not consistently result in shorter operative time, with single-setting procedures averaging a substantial 281.9 min and double-setting procedures a quite moderate 125.5 min. Despite this, the mean length of hospital stay was not substantially different (6.0 days surgical vs. 5.6 days hybrid). Of safety concern, higher rates of total complications and bleeding in the hybrid group are also reported in other research work [45, 46]. Far more critically, the double-setting hybrid subgroup developed greater rates of complications (17% vs. 9%) and bleeding (7% vs. 3%) than the single-setting subgroup, suggesting that the cumulative risk of two separate procedures may surpass that of a single, longer operation. Reassuringly, mortality was similarly low in both groups at 1%, as corroborated by other evidence of no difference in short-term mortality [44, 47]. These findings have direct clinical applicability, implicating a trade-off between efficacy and procedural ease. In those where maximal efficacy is the overriding consideration, surgical ablation alone may be the best strategy, consistent with data for excellent efficacy of the Cox-maze procedure’ [47]. Hybrid ablation is nonetheless a valuable option, especially for those with long-standing persistent AF or failed catheter ablations, as stated in the 2024 ESC Guidelines [48, 49]. Our data on single vs double-setting procedures add another dimension to this decision, with the suggestion that a single-setting approach, although demanding, may offer enhanced efficacy with fewer morbidity. Such a decision should be undertaken by a multidisciplinary heart team after considering patient-specific factors and institutional experience as noted in the 2023 ACC/AHA/ACCP/HRS guidelines [10, 50]. However, these findings should be considered in the context of the study limitations. First, the review lacks formal statistical reporting of heterogeneity, which may influence the validity of conclusions. Besides, clinical heterogeneity because of combined patient populations and approaches, data pooling, and focus on short and medium-term outcomes limits the scope and relevance of the outcomes in the long term. Transcending these limitations guides some key directions of future research. There is an imperative need for large, multicenter randomized controlled trials (RCTs) between modern surgery and hybrid strategies with standardized end-point definitions and long-term follow-up. Next-generation research should also adopt AF burden as a first-line endpoint for a more complete assessment of efficacy, include patient-reported outcomes, compare these approaches to more recent technologies like pulsed field ablation (PFA) [51], and conduct robust cost-effectiveness analysis. Overall, this meta-analysis suggests that bilateral surgical ablation has more freedom from arrhythmia and a better safety profile than hybrid ablation in paroxysmal and persistent AF patients. Hybrid ablation is, however, also a practical treatment option, with great variability in outcome based on the technique of procedure (single vs. double-setting). These results underscore the current efficacy, safety, and complexity of procedures, balance, and underscore the need to make the decisions of the procedure patient-specific and the necessity to conduct additional high-quality studies to optimize the technical methods of the procedures, to select patients optimally, and to find the long-term role of each of the mentioned techniques in managing atrial fibrillation. Impact on clinical practice The results demonstrate a distinct trade-off of procedural efficacy and complexity, which directly affects the clinical decision-making of patients with paroxysmal and persistent atrial fibrillation. In patients for whom the highest effect is of prime importance, a bilateral surgical ablation is potentially a more effective approach, with a higher rate of arrhythmia freedom (83% vs. 75%), and a better safety profile with a lower overall complications rate (10% vs. 13%) and bleeding (2 vs. 4%). On the other hand, hybrid ablation is also an effective therapeutic alternative, especially in complicated cases, such as in patients with persistent AF over a long duration or after the catheter ablations have failed, as suggested by current guidelines [48, 49]. The decision of the hybrid approach is also subtle; a single-setting procedure might be more effective and less prone to morbidity despite an extended length of operation than the double-setting approach. Finally, these findings support the ACC/AHA/ACCP/HRS guideline recommendation [51] of 2023 that the choice of treatment is to be individualized by a multidisciplinary heart team with careful consideration of patient and institution-specific factors versus institutional knowledge and the unique risk-benefit profile of each method. Conclusion This meta-analysis indicates a higher probability of maintenance of sinus rhythm with a superior profile of safety by bilateral surgical ablation compared to hybrid ablation. This conclusion, however, must be balanced against the inherent limitations of the study, such as the absence of formal statistical reporting on heterogeneity and the clinical variability due to pooling of different patients and short to medium term follow-up. The limitations underscore a need to conduct large, multicenter randomized controlled trials with standardized endpoints and long-term follow-up to conclusively compare modern surgical and hybrid methods in the future. Future studies should also consider AF burden as the key endpoint to a more specific efficacy evaluation, integrate patient-reported outcomes, make comparisons with new technologies such as pulsed field ablation, and conduct sound cost-effectiveness analyses. This type of evidence of high quality is essential to maximize patient selection, improve the techniques of the procedures, and to establish the long-term role of each approach in the treatment of atrial fibrillation. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1 Supplementary Material 2 Supplementary Material 3 Supplementary Material 4 Supplementary Material 5 Supplementary Material 6 Abbreviations Antiarrhythmic drugs European Society of Cardiology American College of Cardiology American Heart Association Atrial Fibrillation Body mass index Confidence interval Degree of freedom Diabetes mellitus Heart failure with reduced ejection fraction Hypertension Inconsistency statistics Lone atrial fibrillation Long-standing persistent atrial fibrillation Left ventricular ejection fraction National Institutes of Health Paroxysmal atrial fibrillation Population, Intervention, Comparison, and Outcome Preferred Reporting Items for Systematic Reviews and Meta-Analyses International Prospective Register of Systematic Reviews Persistent atrial fibrillation Pulmonary vein isolation Cochran’s Q statistic Randomized controlled trial Radiofrequency Risk of bias Standard deviation Venous thromboembolism Author contributions All authors contributed to the study. Conception and design were carried out by Fouad Hanna. Material preparation and data collection were conducted by Fouad Hanna, Ali Dawy, Hamza A. Abdul-Hafez, Yousef M. Husseiny, Ahmed Youssef, Mazen Negmeldin, Basel Saber, Hadeel Jameel Ayesh, Omneya Kandil, Abdelrahman Elgendy, Nada Gamal, and Abdelrahman El-Helbawy. Data analysis was performed by Abdelrahman Elgendy. The first draft of the manuscript was written by Abdelrahman El-Helbawy and Fouad Hanna. The introduction was authored by Mazen Negmeldin, the methods by Nada Gamal, the results by Abdelrahman El-Helbawy, and the discussion by Ali Dawy and Omneya Kandil. All authors provided comments on previous versions of the manuscript. Revisions were carried out by Omneya Kandil, with additional corrections and revisions by Abdelrahman El-Helbawy and Fouad Hanna. Figures and tables were prepared by Hamza A. Abdul-Hafez, Abdelrahman El-Helbawy, and Fouad Hanna. All authors reviewed and approved the final manuscript. Funding Not available. Data availability All the data are available when the editor requests. Declarations Ethics approval and consent to participate This study was conducted following the ethical standards of the institutional research committee and registered with PROSPERO under registration number (CRD420251107935). Competing interests Fouad Hanna, Ali Dawy, Hamza A. Abdul-Hafez, Yousef M. Husseiny, Ahmed Youssef, Mazen Negmeldin, Basel Saber, Hadeel Jameel Ayesh, Omneya Kandil, Abdelrahman Elgendy, Nada Gamal, and Abdelrahman El-Helbawy have no conflicts of interest or financial ties to disclose. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Fouad Hanna and Ali Dway are Joint-First Authors. Fouad Hanna and Ali Dway are equal contributors to this work and are designated as co-first authors. References Associated Data Supplementary Materials Supplementary Material 1 Supplementary Material 2 Supplementary Material 3 Supplementary Material 4 Supplementary Material 5 Supplementary Material 6 Data Availability Statement All the data are available when the editor requests.