Globally, atrial fibrillation (AF) is the most commonly observed sustained cardiac arrhythmia, particularly in patients with structural heart diseases [1]. According to the European Society of Cardiology guidelines for the management of AF [2], catheter ablation (CA) is a class I treatment recommended for AF. However, because of the high rate of AF recurrence, most patients who receive CA should undergo warfarin or non-vitamin K antagonist oral anticoagulant (NOAC) therapy to reduce the risk of ischemic stroke or systemic embolism.

Although warfarin and NOACs reduce the risk of ischemic stroke and systemic embolism, they are associated with an increased risk of bleeding events [3]. Device-based left atrial appendage closure (LAAC) is a therapeutic modality for stroke prevention in patients with nonvalvular AF, particularly in those with a higher risk of bleeding determined using HAS-BLED scores. Recently, Wintgens et al. [4] reported that CA and LAAC could be combined into a one-stop therapy that can provide a straightforward strategy in maintaining rhythm and preventing stroke. Similar results have been reported by other studies with a small sample size and clinical studies performed at a single center [5, 6]. In 2015, the expert consensus statement of the European Heart Rhythm Association/European Association of Percutaneous Cardiovascular Intervention [7] suggested that combining the two left atrial interventions is a valuable and practical approach because of the common aspects of trans-septal puncture, anesthesia, and anticoagulation therapy.

However, patients are still recommended to undergo warfarin or NOACs after LAAC for at least 45 days because of the endothelialization of the left atrial appendage. During this time, it is unknown whether patients who are suitable for LAAC and have high HAS-BLED scores are associated with a high risk of bleeding or all-cause mortality. Moreover, postoperative complications, such as residual flow, are relatively common in patients who undergo LAAC, and postoperative complications increase the risk of bleeding and other adverse events. To the best of our knowledge, no meta-analysis has reported the postoperative complications in patients with a high risk of bleeding using one-stop therapy. Therefore, to evaluate the efficacy and safety of this new strategy, we performed a comprehensive meta-analysis of the contemporary literature.

The meta-analysis was performed according to the preferred reporting items for systematic reviews and meta-analysis statement and registered in the PROSPERO database, an International prospective register of systematic reviews. (CRD42018106746, https://​www.​crd.​york.​ac.​uk/​PROSPERO/​display_​record.​php?​RecordID=​106746).

In patients with nonvalvular AF, the risk of stroke increases 5-fold compared with that in patients with sinus rhythm [19]. CA is an efficacious therapy for normalizing rhythm in patients with symptomatic AF. However, because of the high rate of AF recurrence, it is recommended that such patients additionally receive oral anticoagulant therapy. LAAC has emerged as an alternative to long-term anticoagulation therapy with warfarin or NOACs. LAAC has an efficacy similar to warfarin for all-cause stroke prevention [20]. Phillips et al. [17] pooled the data of patients with a high risk of stroke from real-world Watchman LAAC registries and demonstrated the feasibility and safety of combined CA and LAAC therapy. However, the long-term outcomes of this new therapeutic strategy have not been well established. Accordingly, single-center clinical studies [6] have been performed, but the sample size used in these studies is relatively small. Postoperative complications, such as residual flow, are relatively common and influence reportedly prognosis in patients who undergo LAAC. Thus, we performed this meta-analysis to study the postoperative complications and outcomes of the combined CA and LAAC therapy.

The combined therapy has been receiving increasing attention as a potentially important treatment strategy; therefore, we concentrated on the primary adverse events, such as pericardial effusion and bleeding events. Although there are no comparative study data on one-stop therapy unlike CA or LAAC alone, these results were interestingly similar to those recently reported [21] on the safety and efficacy of LAAC alone; however, the rate of bleeding events was lower than that reported in our meta-analysis. The reason for this discrepancy might be the routine use of anticoagulants during the CA periprocedural period. Minor bleeding events accounted for 84% of the total events in our pooled data, and no patient died of pericardial effusion during follow-up. These results support the conclusion that the combined CA and LAAC therapy is safe during the periprocedural period.

Thrombus is located at the LAA in approximately 90% patients with nonvalvular AF [22]. Moreover, LAAC [23] and surgical LAA excision [24] are effective ways to prevent stroke in patients with AF. However, Noelck et al. [25] reported that LAA devices are associated with high rates of procedure-related harm and thus require further evaluation. In our analysis, the most common adverse event during the periprocedural period was residual flow (rate, 9.11%). However, it still remains controversial whether residual leakage can cause adverse cardiac events. Although the findings of the PROTECT AF trial and other studies did not establish a relationship between the minimal residual flow and adverse events, including thromboembolism [26, 27], and are not associated with other adverse cardiac events [28], these results are in contrast with those of the surgical literature, in which residual flows are associated with thrombus formation and other adverse clinical events [29, 30]. In addition, the results of recent studies suggest that residual flow or leak may result in a predisposition to device-related thrombus [31, 32]. In the present analysis of the periprocedural subgroup, we found a potential relationship between a high incidence of bleeding events and residual flow (P = 0.03). Meanwhile, patients in the subgroup with a higher periprocedural residual flow rate may have a higher possibility of all-cause mortality than those in the subgroup with a lower periprocedural residual flow rate (3.39% vs. 1.42%, P = 0.12); however, no patient directly died of residual flow. Minimal residual flow (< 5 mm) does not require special treatment or clinical intervention [20, 33]; however, the proportion of patients with high residual flow (> 5 mm) was very low, and the data in these studies were based on a limited number of events and follow-up period. In the present study, we found a higher incidence of bleeding events in patients with high residual flow. One potential treatment option for this includes the continuation of a course of warfarin or NOACs therapy for at least 45 days; furthermore, patients may continue the anticoagulation therapy if major residual flow (> 5 mm) is detected. Additionally, minor or major residual flow in the perioperative period was not related to their number during the average TEE follow-up (5.18 months) and average maximum-event reported time (2.26 months). Interestingly, the endothelialization of LAA may take 45 days and reduce the reflux of the LAAC device [34]; however, the actual time may be longer than we previously estimated. Interestingly, a recent study enrolled patients who received one-stop therapy, and the results demonstrate that the combination strategy is independently associated with the new peri-device leak [18]. Further research is needed to evaluate the possibility that CA leads to the lengthening of the endothelialization process. In summary, residual flow in the perioperative period does not lead to a poor prognosis, but the potential perioperative bleeding risk during continuous anticoagulant use should be of concern. In addition, transesophageal echocardiogram assessment is recommended at 6-month intervals if any peri-device leak is documented [18]. Furthermore, the relationship between residual leak and thrombus or other cardiac adverse events should be studied for advancement in clinical research, especially for patients with major residual flow (> 5 mm) and those who have received one-stop therapy.

In the present study, the mean follow-up was approximately 2 years, and the rate of all-cause mortality was 2.15% during this time. The incidence of bleeding and embolism events was 6.95 and 5.24%, respectively (follow-up period of approximately 2 years). In our pooled data of 50 bleeding events, only seven patients experienced significant bleeding events, such as bronchiectasis, carcinoma, Rendu–Osler–Weber disease, gastrointestinal bleeding, knee haematoma, hematuria, and groin haematoma, and no patients experienced or died from haemorrhagic stroke. The reason for this might be the marked decrease in the use of anticoagulant drugs after hybrid therapy (from 93.59 to 11.61%). A total of 31 embolism events occurred during the follow-up period; only six patients directly received device embolization. Although it was unclear if there was a relationship between device-related thrombus events and device embolization based on our pooled data, a previous meta-analysis [35] demonstrated that the rates of all-cause stroke/systemic embolism and ischemic stroke/systemic embolism increased in patients who developed device-related thrombus events, and the rates of haemorrhagic stroke also increased in these patients. Thus, further research is required to determine the association between implant device and embolism occurrence.

One issue is the tendency of higher peri-procedural residual flow or embolism that could be related to a concomitant inflammatory process promoted by ablating the left pulmonary veins. However, the risk of these major adverse events was seemingly lower than the overall risk of stroke by approximately 5% per year in patients with nonvalvular AF. Our data with regard to adverse events were similar to those reported by the early stage of PROTECT AF study [20], which in turn were similar to those of the previous PREVAIL and PROTECT AF trials and of some recently published meta-analyses [36, 37]. After the perioperative period, we further analyzed the relationship between the presence of residual flow during TEE follow-up and other adverse events. The subgroup analysis suggested that increased residual flow rate leads to a higher rate of bleeding events (P = 0.08); thus, patients with residual flow during TEE follow-up may have higher risk of bleeding events. Importantly, the presence of residual flow on TEE follow-up was not statistically associated with other adverse events during follow-up, including all-cause mortality, bleeding or embolism events, and AF recurrence. Thus, the safety and efficacy of combined CA and LAAC treatment were acceptable in our study.

Recently, the Watchman and Amplatzer devices have demonstrated a high rate of device-related thrombosis as reported by Fauchier et al. [38], and this drew the attention of researchers. They found that the incidence of device-related thrombosis in patients with LAA imaging was 7.2% per year and that of ischemic stroke was 4.0% per year. These results were significantly higher than those reported in previous studies [39] as well as in the present study. These discrepancies could be caused by a number of factors. Oral anticoagulation at discharge was considered to be a protective factor, but the proportion of patients who received oral anticoagulation was much lower than that in our study (33.47% vs. 93.59%). Moreover, the CHA₂DS₂-VASc score in our study was 2.2–4.4, much lower than the mean value in the study by Fauchier et al. (4.5 ± 1.5) [38].

AF recurrence after CA is a significant problem because the percentage of patients who are free from AF recurrence is not satisfactory [40]. The most common method of ablation is pulmonary vein isolation; however, the rate of AF recurrence was as high as 32.89% during the follow-up in our meta-analysis. In the subgroup analysis, patients with a higher AF recurrence rate also had a higher risk of embolism (P = 0.03) and residual flow (P = 0.04). These results indicate the potential effect of long-term residual leak on hemodynamics that may influence the maintenance of sinus rhythm. Conversely, complete occlusion seems to reduce the incidence of AF recurrence [41]. Electrical isolation of the LAA may reduce AF recurrence [42]. It was believed that the effects of complete occlusion on electrical remodeling of the LAA would provide a new method for the comprehensive treatment of AF in future.

Finally, compared with NOACs, LAAC delivers the expected results, making it a better cost-effective treatment strategy for the secondary prevention of stroke in patients with AF since the last 10 years [43]. Meanwhile, serious bleeding risk associated with oral anticoagulant continuation after ablation success seems to outweigh the benefits of thromboembolic risk reduction [44]. Based on the high bleeding events, it seems that one-stop therapy has better cost-effectiveness ratio in patients with a higher rate of bleeding risk. However, more randomized studies on the cost-effectiveness of this one-stop therapy are warranted to test our results.

Our results demonstrate that the use of CA and LAAC in a single procedure is safe and efficacious. However, there is some risk of bleeding in patients with LAA residual flow when warfarin or NOACs are continuously used. Additionally, our results suggest that patients with AF recurrence suffer from embolism events or LAA residual flow during follow-up. Finally, the average reported time of maximum residual flow events suggests that the process of LAA endothelialization takes longer than that was previously thought, and more studies should be performed to evaluate whether CA can lead to a prolongation of device endothelialization.