Hybrid convergent ablation versus endocardial catheter ablation for atrial fibrillation: a systematic review and meta-analysis of randomised control trials and propensity matched studies

Atrial fibrillation (AF) is the most common cardiac arrythmia, affecting approximately 33 million people worldwide [1]. Only 3% of all AF is associated with concomitant cardiac disease [2]. Endocardial catheter ablation (ECA) remains the mainstay of intervention in AF, however, its success may be limited by its inability to create transmural endocardial ablation lines [3‐5]. Surgical ablation was initially developed with cut and sew lesions, first described by Cox et al. in 1991 [6]. With the advent of thoracoscopic radiofrequency devices, surgical treatment of AF has shifted from open heart to minimally invasive procedures to isolate the pulmonary veins, occlude the left atrial appendage and create epicardial ablation lines [5].

The success rate of a surgical approach is superior to that of ECA in persistent AF [7]. The advantage of the surgical approach is the ability to visualize and target the left atrial structures, especially the posterior wall [8]. There are limitations of endocardial approaches. Anatomically, the posterior wall is difficult to isolate from the endocardial approach [8]. Furthermore, extensive endocardial ablation risks thermogenic damage to surrounding structures such as the phrenic nerve or esophagus [5]. Strengths of endocardial ablation is its ability to map AF substrates and develop electrophysiological (EP) endpoints. Hybrid convergent ablation (HCA) garners the strengths of both approaches by combining a subxiphoid (surgical) approach to target the posterior left atrial wall, followed by completion endocardial ablation [8]. The ability to ablate the posterior wall, validate these lesions, identify further lesions and ablate further arrhythmogenic substrates makes this approach effective for rhythm maintanence [5]. Unfortunately, much of the convergent experience comes from single-centre retrospective studies which are prone to bias [5, 9].

The primary aim of this study is to assess the freedom from AF in a HCA cohort compared to an ECA cohort alone. The secondary aim of this study is to assess the incidence of significant complications such as tamponade, sternotomy, esophageal injury, atrio-esophageal fistulae post procedurally and post procedural mortality. In order to account for bias associated with retrospective studies, we included randomized control trials (RCT) and propensity score-matched (PSM) analyses only.

This systematic review and meta-analysis of RCT’s and PSM studies aims to compare the efficacy of HCA to ECA. The inclusion of unmatched retrospective analyses into meta-analysis adds a significant degree of heterogeneity. Firstly, these studies have inherent patient related bias. Their retrospective design means inability to control for baseline variables that impact procedural efficacy, principally, number of previous ablations, LA dimensions, BMI, duration of AF and use of AAD [19‐22]. Secondly, patients who undergo hybrid procedures are more likely to have complex arrhythmogenic substrates, have failed previous ablations, and are therefore more resistant to further ablations [19, 21, 22]. As a result, the effectiveness of HCA may be underestimated in these studies. Lastly, previous meta-analyses included studies with a variety of surgical techniques (subxiphoid/ unilateral or bilateral thoracoscopic/mini-thoracotomy/sternotomy), which vary in terms of their efficacy and risk profile [9, 18, 20, 22]. The resultant heterogeneity in these meta-analyses was significant [8, 22]. Mhanna et al. reported a heterogeneity of 77% in their primary outcome (freedom from AF at last follow-up), and Zhang et al. reported a heterogeneity of 86% in the same outcome [9, 18]. A strength of the present study was the inclusion of RCT’s and PSM studies only. All patients underwent the same procedure and had similar pre-operative characteristics. The resultant heterogeneity was significantly smaller than previously published meta-analyses (I² = 0) for the primary outcome of interest.

The FFAF in the HCA cohort was significantly higher than the ECA cohort and is higher than the FFAF in previously published meta-analyses. Mhanna et al. reported a FFAF of 70% with an OR of 1.48 (95%CI 1.13–1.94, P < 0.01) favoring HCA [8]. Zhang et al. reported a FFAF in their HCA cohort of 57%, with an OR of 2.10 (95% CI 0.45–9.88), however, this failed to reach significance [18]. The main cohort of patients undergoing HCA in this study as well as previous meta-analysis was persistent AF, and the higher freedom from AF highlights the efficacy of HCA in this cohort.

Three studies reported FFAF off AAD. Subgroup analysis suggests that FFAF was 50% in the HCA cohort compared to 26% in the ECA cohort. Accordingly, reintervention rates were also low with HCA cohorts less likely to require subsequent DCCV. Some studies also suggest that HCA cohorts are less likely to require further ECA in the future, however this data was not reported by all included studies [17]. Our study also demonstrated a significantly lower attrition in FFAF in the HCA cohort versus ECA cohort, with an aggregate FFAF of 85% versus 61% at 12 months. The use of AAD can be associated with long-term side effects and cessation of use is beneficial. The continued decline of FFAF in ECA cohorts off AAD can be attributed to incomplete pulmonary vein isolation (PVI) or failure to address lesions that lie outside the PVI, thereby relying on AAD to maintain sinus rhythm [17]. The transmural lesions created by a hybrid convergent approach account for the higher FFAF off antiarrhythmic therapy, reduced need for further cardioversions and lower attrition rate over time observed in this study.

Three studies performed the HCA in the same sitting, and one study performed a staged HCA. Varzaly et al. performed a meta-analysis of rhythm maintenance following hybrid ablation, and found no significant differences between a staged or sequential approach [23]. Advantages of a sequential approach are immediate identification of lesion gaps that can be corrected by catheter ablation and shorter procedural times [23]. Advantages of a staged approach is that it allows time for lesions to mature and edema to regress, identifying definite lesions of further endocardial ablation 23. There is a paucity of literature exploring the utility of a staged procedure compared to a concomitant procedure, and future RCT’s are being conducted in the area to address this question.

Indication for AF in three studies was persistent AF, and one study performed convergent procedures on patients with paroxysmal AF. The recurrence rate for AF is usually dependent on the duration of AF, as the pathological mechanisms for persistent AF and long-standing persistent AF are more complex than paroxysmal AF [24]. The approach of PVI may not be effective enough for longstanding AF and further ablations are necessary. Patients may also need to undergo repeated procedures exposing them to the deleterious effects of radiation exposure, complications and cost [24]. The major strategic advantage of HCA over catheter ablation alone is the ability to attain a broad area of ablation across the entire posterior left atrial wall [25, 26]. PVI alone may not be sufficient in persistent AF as substrates are more likely to be located in non-pulmonary vein regions [25, 26]. Furthermore, the left atrial appendage, a further substrate for AF, can be closed with an additional thorascopic port [5].

Complications were more common in HCA cohort, with 9.4% of patients the HCA cohort and 1.6% in the ECA cohort reporting a post-procedural adverse event. Overall, mortality was rare, reported in only one patient (0.005%) in the HCA cohort secondary to a gastrointestinal bleed. Varzaly et al. reported a similar overall complication rate across 22 studies of 6.5%, with a low mortality rate of 0.2% reflecting the overall safety of the procedure [23]. Overall mortality is low in literature, with only one study documenting a high mortality rate [19]. The HCA group in that study experienced three deaths (in 24 patients) and these sudden deaths were attributed to the type of technology used (unipolar ablation) and the approach (pericardioscopic) [19]. The unipolar device has since been redesigned with an electrocardiogram sensing tip and an irrigation tip to reduce the rate of complications [4]. As HCA is an evolving field, future peri- and post-operative protocols may result in a lower complication rate. Delurgio et al. commented that all four postoperative effusions were potentially avoidable with postoperative non-steroidal therapy [15].

Cessation of oral anticoagulation therapy (OAC) post procedurally is an important outcome as the bleeding risk on OAC is not negligible. Themisocclatis et al. observed a 2% risk of major haemorrhage in patients on warfarin following catheter ablation [27]. The cessation of OAC depends on the maintenance of SR post procedurally and the CHADS-VASC2 score. The cessation of OAC post catheter ablation has been well studied. A meta-analysis by Liu et al. demonstrated a similar cumulative thromboembolic rate on and off OAC post ablation of 1.1% and 1.4% respectively [28]. Additionally, the rate of haemorrhagic complications in the group off OAC was significantly lower [28]. Convergent ablation hafs a higher FFAF than catheter ablation alone, expectedly the cessation of OAC would be higher in these cohorts. There is a paucity of evidence exploring this. Studies assessing post procedural success of minimally invasive surgical ablation demonstrate that the prevalence of OAC use was higher in the surgical cohorts, however this was confounded by patient bias whereby surgical cohorts have larger LAA dimensions and CHADS-VASC2 scores [29]. Lauritzen et al. demonstrated that OAC’s can be safety ceased 12 months post-surgical ablation, when patients have SR maintenance and a CHADS-VASC2 score less than 2 p30]. Future RCT’s exploring OAC cessation after convergent ablation would be beneficial.

There is also a paucity of evidence exploring the efficacy of convergent ablation in high-risk patients such as those with a high BMI and a previous history of cardiac surgery. Obese patients face greater risks of complications from ablation procedures due to their comorbidities p31]. During hybrid ablation, haemodynamic intolerance, stroke risk and ventilation may pose an issue [31]. Additionally, procedure times are usually longer and radiation exposure greater in obese patients [31]. Furthermore, the presence of epicardial fat may attenuate the energy delivered to the left atrial wall, diminishing the effectiveness of the procedure. These difficulties have not translated to a lower FFAF following thorascopic ablation [31, 32]. Patients with a previous history of cardiac surgery may have adhesions limiting exposure of the left atrium, and in which case the procedure may not be feasible. As the convergent procedure becomes more widespread and these cases are encountered, future cohort studies and RCT’s will investigate the issue further.

There were limitations to the present meta-analysis. A small number of studies, with small patient numbers, were included in the meta-analysis as a result of the inclusion of RCT’s and PSM studies only. Secondly, rhythm monitoring post-procedurally varied, with one study utilising an internal loop recorder and three utilising Holter monitoring and ECG’s. The definition of FFAF also varied between the studies, with three studies defining recurrence as 30 s of AF outside the blanking period and one study defining recurrence as an AF burden > 1% of the time. Again, this impacts the FFAF and adds systematic bias to the review. Finally, there was a paucity if individual patient data (IPD) to aggregate the KM curves past the one-year mark. Both Jan et al. and Kress et al. reported small patient numbers past the one year so deriving conclusions on AF free survival past the one-year mark is tenuous.

Future RCT’s, with larger patient cohorts and including LAA exclusion will further consolidate this data. As HCA is an emerging technology, trials with subgroup analysis assessing the advantage LAA exclusion or the vein of Marshall ablation are warranted. Also, pulsed field ablation as a part of convergent ablation is a novel technology with high efficacy, and future trials assessing this in larger patient populations will be beneficial [33].

HCA is associated with a significantly higher FFAF than ECA alone and a higher likelihood of rhythm maintenance off AAD. HCA, however, is also associated with more post-procedural complications than ECA alone. The overall risk of mortality remains low. Future RCT’s with larger patient cohorts comparing the HCA and ECA may further validate these results.