Introduction

Atrial fibrillation (AF) is the most common arrhythmia affecting more than 33 million people worldwide in 2010 [1]. AF is associated with an increase in stroke risk and mortality and its prevalence is rising [2]. While vitamin K antagonists such as Warfarin and non-vitamin K oral anticoagulants (NOAC) reduce stroke and mortality in AF, they substantially increase the risk for major bleedings [3]. In AF patients with a high risk for stroke and high risk for bleeding and/or contraindication to OAC, left atrial appendage occlusion (LAAO) may address both issues [4]. The five-year outcome data of the randomized controlled trials PROTECT-AF (WATCHMAN Left Atrial Appendage System for Embolic Protection in Patients With Atrial Fibrillation) and PREVAIL-AF (Evaluation of the WATCHMAN LAA Closure Device in Patients With Atrial Fibrillation Versus Long Term Warfarin Therapy) support a non-inferiority of LAAO to OAC with a significant decrease in hemorrhagic stroke [5]. However, comorbidities such as heart failure and chronic kidney disease impair clinical outcomes in patients with AF [6]. In fact, in patients undergoing transcatheter edge-to-edge repair (TEER), AF is associated with more bleeding, HF hospitalization, and mortality [7]. The mitral valve is of particular interest in LAAO because of the anatomical proximity and the effects on left atrial anatomy. Left atrial size is a predictor of cardiac death in heart failure before [8].

To our knowledge, this remains a field with no to little evidence, even though in a single-center study of 122 TEER patients, around 50% would also qualify for LAAO [9]. Currently, the first studies are evaluating a combined approach of both procedures even when no data on outcomes on LAAO in patients with mitral disease is available [10].

The German LAArge registry is an independently financed prospective, non-randomized registry of patients undergoing LAAO.

In this study, we aim to investigate the influence atrial myopathy due to mitral valve disease may have on in-hospital and long-term safety outcomes in patients after LAAO.

Methods

Data collection and LAArge registry

The non-profit organization “Institut für Herzinfarktforschung” (IHF, Ludwigshafen, Germany) manage and oversee the multicenter German left atrial appendage occlusion registry (LAArge). Thirty-eight centers participated in this prospective, non-randomized study. No funding from industry sources was used for the project. From all enrolled patients, written informed consent was obtained. The privacy measures and data collection have been described previously [11]. Using a web-based electronic case report form, baseline characteristics, procedural data as well as in-hospital data were collected and checked for plausibility. All transmitted data were encrypted and stored on servers maintained by IHF. Echocardiographic follow-up was documented at the standard follow-up at each participating site after LAAO, usually 3–6 months. The IHF conducted the 1-year follow-up by reports from the implanting center and via a standardized phone interview. The study was carried out according to the declaration of Helsinki and approved by the ethics committee of the State Chamber of Medicine in Rhineland-Palatinate, Germany.

Mitral valve disease definition and procedural methods

Mitral valve disease (MVD) was defined as pre-existing mitral regurgitation and/or TEER or surgical MV repair (SMVR) combined with a left atrial (LA) diameter of > 45 mm in women and > 48 mm in men. The reasoning was to include any irregular mitral valve anatomy, due to pathological MR or surgically or percutaneously treated MV. In addition, grading of mitral regurgitation after TEER becomes increasingly difficult. Patients were either allocated to the MVD group or the no-MVD group if the criteria were met. The detailed procedural methods have been described previously [11]. Patients were screened and enrolled following current guidelines and best medical practice [12]. Generally, AF patients with significant stroke risk and contraindication to anticoagulation were included. Device selection and pre-procedural imaging were left to the operator’s preference. The antithrombotic regime and procedural protocol were conducted according to the implanting center. Transesophageal echocardiography (TEE) was carried out to exclude patients with intracardiac thrombus and define anatomy for technical feasibility. Procedures were carried out in light sedation using propofol or general anesthesia.

Outcomes

In-hospital data included serious adverse events such as device embolization, peri-device leak, bleeding, stroke, and groin complications among others. Predefined safety outcomes are given in supplemental Table 1. Implantation success and adverse device events were defined as given in the Munich consensus document [13]. The primary outcome was a combined outcome of death, stroke, and systemic embolism after a year. Adverse events were collected for in-hospital events and follow-up events. Routine TEE follow-up was carried out but not mandatory to report as many implanting centers may have referred the patients to local centers for TEE follow-up.

Statistical analysis

Normally distributed continuous data are given as means ± of the standard deviation (SD), otherwise shown as medians with interquartile ranges (25th and 75th percentiles). Categorical data are presented in relative percentage and absolute values. Fisher’s exact test was used for rates of in-hospital complications. Statistical differences between both groups were compared using either a Chi-square test or the Mann–Whitney–Wilcoxon test. The 12-month event rates of death, the composite outcome of death and stroke, and a composite of death, stroke, and systemic embolism were calculated by the Kaplan–Meier method. The outcomes were compared between age groups using the log-rank test. Hazard ratios (HR) with 95% confidence intervals (CI) were estimated using Cox regression without adjustment. All statistical comparisons were two-sided, and p values < 0.05 were considered statistically significant. Analyses were performed using the Statistical Analysis System (SAS, Version 9.4, SAS Institute Inc., Cary, NC, USA).

Results

Baseline characteristics

A total of 555 patients were included in this study. The MVD cohort included 52 Patients and the no-MVD cohort 476 patients (Table 1). The MVD group was significantly more likely to be older (78.2 years ± 7.0 vs 75.9 years ± 8.0, p = 0.036) and female (59.6% vs 37.8%, p = 0.002). In the MVD cohort, 5.8% had a history of TEER and 11.5% had a history of SMVR. Cardiomyopathy (21.2% vs. 5.9%, p < 0.001) and chronic kidney disease (59.6% vs 36.8%, p = 0.001) were significantly higher in the MVD group. CHA2DS2-VASc (5.1 ± 1.5 vs. 4.5 ± 1.5, p = 0.009) and HAS-BLED-Score (4.4 ± 1.1 vs. 3.8 ± 1.1, p = 0.003) were significantly higher in MVD patients compared with the no-MVD group.

Table 1 Baseline characteristics
Full size table

Left atrial appendage anatomy

Data for the left atrial appendage anatomy are presented in Table 2. The MVD cohort presented with significantly larger LA diameter (53 mm vs. 48 mm, p < 0.001) as well as larger LAA ostia (at 135° 23.0 mm (20.5, 26.0) vs 20.0 mm (18.0, 23.0), p = 0.002). There were no significant differences in LA sludge or LAA thrombus formation. The chicken wing morphology occurred the most in both groups (34.6% vs 46.6%, p = 0.10), followed by the windsock (13.5% vs 16.2%, p = 0.61) and cauliflower (17.3% vs 15.5%, p = 0.74) morphology and less often by the cactus (5.8% vs 9.9%, p = 0.33) morphology. Patients with MVD were more likely to have an atypical (classification not applicable) LAA morphology (28.8% vs 11.7%, p < 0.001).

Table 2 Left atrial appendage anatomy
Full size table

Procedural data

Implant success was high in both cohorts with 96.2% and 97.9%, respectively. There was a significant difference in subpar device position in the MVD group [3.8% (2/52) vs 0.2% (1/476), p < 0.001] compared with the no-MVD cohort. No difference in device selection was reported. The Watchman device was the most common device (50% and 42.0%), followed by the Amplatzer Cardiac Plug (19.2% and 29.8%) and the Amplatzer Amulet (25.0% and 25.6%). Device embolization occurred in 1/52 procedures in the MVD cohort and 6/476 cases in the no-MVD cohort (p = 0.69). Peri-device leak and left–right shunt occurred in both groups without significant differences. No peri-device leaks > 5 mm were observed in either cohort (Table 3).

Table 3 Procedural data
Full size table

In-hospital safety data

The data for in-hospital safety are presented in Table 4. The MVD cohort had a longer hospital stay after the procedure compared with the no-MVD cohort [3 days (2, 6) vs 2 days (2, 3), p = 0.001]. There were no MACCE (death, stroke, or myocardial infarction) in the MVD cohort and 3 events in the no-MVD cohort. Other severe complications occurred in the MVD group with 5.8% and with 2.9% in the no-MVD cohort (p = 0.23). These included pericardial effusion with need for interventional drainage (3.58% vs 2.10%, p = 0.33), AV-Fistula (1.9% vs. 0.84%, p = 0.11), and severe bleedings (3.8% vs. 0.84%, p = 0.11) without significant differences between both groups. Moreover, moderate complications (13.5% vs 9.9%, p = 0.47) and overall complications (19.2% vs 12.8%, p = 0.20) were comparable between both cohorts.

Table 4 In-hospital safety data
Full size table

Follow-up safety data

The Follow-up data are shown in Table 5. The follow-up rate was high in both groups with 92.3% in the MVD cohort and 98.7% in the no-MVD cohort. Device embolization was comparable in both groups (1.9% vs. 2.3%, p = 0.86). There were no differences in major complications including stroke, myocardial infarction, and moderate or severe bleeding. Follow-up echocardiography data were obtained in 42.3% and 31.9% of the patient at a mean follow-up duration of 163 and 97 days. No differences were observed in LA thrombi (4.5% vs 6.0%, p = 1.00) or overall peri-device leak (27.3% vs 16.6%, p = 0.24). Leaks over 5 mm were not observed in either cohort. One-year mortality was numerically higher in the MVD cohort (17.8% vs 11.5%, p = 0.19) as well as a combined outcome of death, stroke, and systemic embolism (20.3% vs 12.4%, p = 0.13) without reaching statistical significance (Fig. 1). Significant predictors of the combined outcome after adjusting were peripheral artery disease (HR 2.41, 95% CI 1.46–3.98, p < 0.001) and chronic kidney disease (HR 3.46, 95% CI 2.02–5.93, p < 0.001) but not the presence or history of MVD (Table 6). Further data on safety outcomes (Table S1) and antithrombotic therapy (Table S2) are provided in the supplement.

Table 5 Follow-up safety data
Full size table
Fig. 1
figure 1

Kaplan–Meyer estimates of event-free survival (death, stroke, systemic embolism) in patients with the presence of mitral valve disease (MVD) or not (no MVD). p values < 0.05 considered significant

Full size image
Table 6 Adjusted cox regression for predictors of death, stroke, systemic embolism (67/528) in patients undergoing left atrial appendage occlusion
Full size table

Discussion

Main findings

Patients with atrial myopathy and mitral valve disease referred for LAAO have more comorbidities than patients without a history of mitral valve disease. Even though larger LA and LAA orifices are observed, procedure success is high and comparable to patients without MVD. Cardiovascular events after one year are more common in patients with mitral valve disease but are driven by comorbidities.

Clinical characteristics and observed events

In patients with mitral valve disease, the occurrence of new-onset atrial fibrillation is high, reaching over 48% at a 10-year follow-up, and is associated with increased cardiac mobility [14]. In a more recent cohort of 2425 patients with degenerative mitral regurgitation, AF was associated with increased overall mortality [15]. In our study, MVD in a cohort of patients with atrial fibrillation is associated with numerically higher mortality. Both conditions are expected to rise with an aging population and increasing comorbidity.

We did not find that the presence of MVD and atrial myopathy is a predictor of cardiovascular events after LAAO. In line with previous data, comorbidities such as chronic kidney disease are stronger factors for mortality in AF [16]. These results are an important safety signal for patients with MVD as LAAO can be conducted in these patients without additional risk to the basic risk from comorbidities.

Another shared trait of both conditions is atrial myopathy with an increased size of both LA and LAA. In fact, in a prospective imaging study using both computed tomography and echocardiography, the most significant determinants of LA enlargement were AF and mitral regurgitation [17]. In our study, atypical morphology of the LAA was observed at a higher rate in those with MVD and atrial myopathy. LAA remodeling has been described in patients with persistent AF [18]. While this has not been reported in MVD, a higher volume load may lead to further remodeling.

However, it must be noted that atrial myopathy is a complex trait that is reflected in electrocardiograms, magnetic resonance imaging, strain imaging, and invasive electrophysiologic studies [19]. In the absence of this data, we chose the combination of LA enlargement and mitral valve disease in atrial fibrillation based on a study that reported that MR in HFpEF reflected LA myopathy even in the absence of AF [20]. Further studies have shown that the combination of mitral valve disease, atrial fibrillation, and left atrial enlargement is highly associated with LA myopathy [21].

Yet, implant success was high in both groups without a significant difference (96.2% vs. 97.9%). The complex anatomy and larger LAA observed with MVD and atrial myopathy are not associated with longer procedures or enhanced support with general anesthesia. While we observed a higher incidence of subpar device position with MVD, event rates are small and should be carefully considered. Device embolization and peri-device leak > 5 mm were rare in both groups and comparable, suggesting that operators adjusted for the different anatomy in patients with mitral valve disease. Postoperative hospitalization was longer in patients with MVD undergoing LAAO which reflects the higher burden of comorbidities such as chronic kidney disease (59.6% vs 36.8%), cardiomyopathy (21.2% vs 5.9%), and anemia (34.6% vs 21.6%).

Due to the proximity of the mitral valve to the orifice of the LAA, there is a concern that higher blood flow or regurgitant jet may impair endothelialization of the LAAO device once implanted [22]. Theoretically, a higher incidence of device-related thrombi (DRT) would be the result.

To our knowledge, there is no data on outcomes of patients with mitral valve disease after LAAO. Recently, a single-center case series of 122 AF patients undergoing TEER reported that around 50% would qualify for LAAO [9]. However, the study considered patients with a CHA2DS2-VASc score ≥ 3 and no contraindication to OAC as potential candidates. The recent European Society of Cardiology (ESC) guidelines on AF recommend LAAO only in patients with contraindication to OAC [4]. In our cohort, no differences in LA thrombi were observed; however, this may be due to low numbers of patients with MVD and even lower with complete echocardiographic follow-up data.

In addition, the impact of MVD on DRT is difficult to assess since larger LAA size itself is a strong predictor. For instance, in an analysis of 1739 patients from PREVAIL-AF, PROTECT-AF, and the continued registries, a 6% higher risk of DRT per mm LAA diameter increase (OR 1.06 per mm increase; 95% CI 1.01–1.12; p = 0.019) was observed [23]. A larger LAA diameter was also a predictor of DRT in the European EWOLUTION registry [24].

Strengths and limitations

There are some caveats to be considered. No standardized process for patient selection, implanting procedure, or postprocedural management. Due to the observational study design, confounding factors cannot be excluded. Additionally, we were limited by the available data in the registry and had to rely on diameter for LA and LAA sizing without more sensitive volumetric values such as left atrial volume index. While our definition of patients with AF, MVD, and enlarged LA sizes probably reflects atrial myopathy, we cannot report sophisticated data such as magnetic resonance imaging or data from electroanatomic mapping.

Echocardiographic follow-up data are insufficient for conclusive incidences of DRT and peri-device leak. Adverse events were reported by the implanting center and maybe therefore unreliable. Also, this was a purely interventional study without any data on rhythm or rate control. Therefore, no association regarding AF treatment can be described in our registry. We also have limited data on the temporal relation of events, for instance, whether pericardial effusion was a result of transseptal puncture or device release.

Our study has some unique strengths as this is an industry-independent study under real-life conditions. This is also the very first work to systematically report the outcomes of patients with atrial myopathy due to mitral valve disease in LAAO. We can also report the 1-year mortality of this cohort. Additionally, this study uses different devices and employs a standardized 1-year follow-up independent of the implanting centers.

Conclusion

Patients with MVD undergoing LAAO show atrial myopathy with larger LA diameter and LAA orifices. However, procedural success is not impaired and safety outcomes are comparable to patients without MVD. No differences in DRT or peri-device leak were observed and 1-year mortality is largely driven by comorbidities that must be addressed in management. Left atrial enlargement and MVD were not independent predictors of cardiovascular events.