Changes in left atrial appendage orifice following percutaneous left atrial appendage closure using three-dimensional echocardiography

Percutaneous left atrial appendage (LAA) occlusion is increasingly performed in patients with atrial fibrillation and long-term contraindications for anticoagulation. Our aim was to evaluate the effects of LAA occlusion with the Watchman device on the geometry of the LAA orifice and assess its impact on the adjacent left upper pulmonary vein (LUPV) hemodynamics. We included 50 patients who underwent percutaneous LAA occlusion with the Watchman device and had acceptable three-dimensional transesophageal echocardiography images of LAA pre- and post-device placement. We measured offline the LAA orifice diameters in the long axis, and the minimum and maximum diameters, circumference, and area in the short axis view. Eccentricity index was calculated as maximum/minimum diameter ratio. The LUPV peak S and D velocities pre- and post-procedure were also measured. Patients were elderly (mean age 76 ± 8 years), 30 (60%) were men. There was a significant increase of all LAA orifice dimensions following LAA occlusion: diameter 1 (pre-device 18.1 ± 3.2 vs. post-device 21.5 ± 3.4 mm, p < 0.001), diameter 2 (20.6 ± 3.9 vs. 22.1 ± 3.6 mm, p < 0.001), minimum diameter (17.6 ± 3.1 vs. 21.3 ± 3.4 mm, p < 0.001), maximum diameter (21.5 ± 3.9 vs. 22.4 ± 3.6 mm, p = 0.022), circumference (63.6 ± 10.7 vs. 69.6 ± 10.5 mm, p < 0.001), and area (3.1 ± 1.1 vs. 3.9 ± 1.2 cm2, p < 0.001). Eccentricity index decreased after procedure (1.23 ± 0.16 vs. 1.06 ± 0.06, p < 0.001). LUPV peak S and D velocities did not show a significant difference (0.29 ± 0.15 vs. 0.30 ± 0.14 cm/s, p = 0.637; and 0.47 ± 0.19 vs. 0.48 ± 0.20 cm/s, p = 0.549; respectively). LAA orifice stretches significantly and it becomes more circular following LAA occlusion without causing a significant impact on the LUPV hemodynamics.


Introduction
The left atrial appendage (LAA) is the most common source of thrombi in patients with non-valvular AF, accounting for approximately 90% of cases [1]. Based on current guidelines, percutaneous LAA occlusion can be considered in selected patients with non-valvular AF who are at increased risk of stroke and have contraindications to long-term anticoagulation [2]. Based on recent randomized clinical trials that have demonstrated procedural safety and efficacy for ischemic stroke prevention with percutaneous LAA occlusion [3,4], the US Food and Drug Administration has approved the Watchman device (Boston Scientific, Maple Grove, Minnesota) for use in percutaneous LAA occlusion. The Watchman is the only device specifically approved for LAA occlusion in the United States and has been increasingly utilized in real-world clinical practice [5].
Although complications are rare, the LAA closure procedure is not entirely without risk. Potential damage to left pulmonary veins and the circumflex coronary artery is possible [6], these complications have been described mainly due to the Amplatzer cardiac plug device [7,8]. There are also cases reports of erosion and perforation of the pulmonary 1 3 artery with the Watchman struts [9,10], as well as pulmonary artery-LAA fistula formation after the Watchman implantation [11]. However, the impact of the Watchman device on the LAA orifice and adjacent left upper pulmonary vein (LUPV) is not well studied except in an animal model [12].
Accurate assessment of anatomic LAA characteristics is crucial for correct sizing and safe placement of LAA closure devices [13]. Transesophageal echocardiography (TEE) is recommended before and during Watchman implantation to measure the LAA size [14]. The depth and orifice diameters of the LAA are usually measured using two-dimensional (2D) TEE, however it has been shown that three-dimensional (3D) TEE is more accurate than 2D TEE and provides measurements that correlate better with those obtained by computed tomography [15][16][17][18]. The aim of this study was to evaluate the effects of LAA occlusion with the Watchman device on the geometry of LAA orifice using 3D TEE and assess its impact on the adjacent LUPV hemodynamics.

Study population
This was a retrospective observational study. A total of 64 patients with non-valvular AF, high risk of stroke (CHA 2 DS 2 -VASc score ≥ 2) and contraindications to longterm oral anti-coagulation underwent percutaneous LAA occlusion with the Watchman device between March 2019 and December 2019 at Yale-New Haven Hospital. Of these, twelve patients without 3D acquisitions of the LAA and two patients whose quality of 3D images was suboptimal for analysis were excluded. Our final study cohort included 50 patients. Patients' clinical characteristics were obtained from electronic medical records. The study was approved by the Yale University Institutional Review Board.

LAA occlusion procedure
All implantation procedures were performed under TEE guidance. The size of the Watchman device implanted was determined by the maximum LAA ostium and depth measurements assessed by 2D TEE according to the recommendations by the manufacturer. TEE was used to assess the correct device position, ensuring a compression of 8-20% of its original size before release.

Transesophageal echocardiography
Periprocedural 3D TEE was performed immediately before and after Watchman placement using the Acuson SC2000 machine equipped with the Z6MS transducer (Siemens Medical Solutions USA, Inc, Mountain View, CA) (Fig. 1). TEE was performed according to a standard protocol that includes the 2D assessment of the LAA in the mid-esophageal views at 0°, 45°, 90°, and 135°. LUPV flow was assessed using color and pulse-wave Doppler. Two-dimensional and color Doppler echocardiography was used to assess the presence of peri-device leaks, defined as a residual flow of any size detected with TEE at the end of the procedure. Three-dimensional TEE images of the LAA were obtained using a single beat acquisition with a volume size large enough to include the entire LAA ( Fig. 1A and B). The 3D data sets were digitally stored for offline analysis. The LUPV peak systolic (S) and diastolic (D) velocities were measured offline in the echolab workstation.

Three-dimensional echocardiography analysis
3D images of LAA pre-and post-device placement were analyzed using the multiplanar reconstruction mode of TomTec Image-Arena, 4D Cardio-View Software (TomTec Imaging Systems GmbH, Unterschleissheim, Germany). As shown in Fig. 2, the left circumflex coronary artery (Cx) and the Coumadin ridge were used as landmarks to ensure that the LAA ostium was measured at the same level in the pre-and post-implant images. The LAA long axes were positioned to allow visualization of the LAA orifice in the short axis. We measured 4 LAA orifice diameters: diameter 1 in the long axis, diameter 2 in its orthogonal view, and the minimum and maximum diameters in the short axis. Circumference and orifice area were also measured in the short axis. Eccentricity index was calculated as maximum/ minimum diameter ratio to assess the LAA orifice geometry. An eccentricity index = 1 represents a perfect circle [16]. Measurements were taken at end-systole because LAA orifice dimensions are greatest in this cardiac cycle phase [19].
We investigated the reproducibility of 3D analysis of LAA orifice in 10 randomly selected images (5 images pre-and 5 post-device placement). Measurements were done by a first observer blinded to initial measurements 1 month apart and by a second independent observer blinded to measurements obtained by the first observer. Both observers are level III echocardiographers with training in 3D echocardiography.

Statistical analysis
Categorical variables were expressed as numbers and percentages, and continuous variables were expressed as means ± standard deviations or median and interquartile range (IQR) according to their distribution. Comparisons of continuous variables pre-and post-Watchman placement and between 3D TEE versus 2D TEE measurements were done using the paired Student's t-test. Comparisons between groups were done using Student's t-test or the post-Watchman placement to ensure that the LAA ostium was measured at the same level. Diameter 1 (D1) is the orifice diameter measured in the long axis and diameter 2 (D2) is the orifice diameter in its orthogonal view. Minimum and maximum diameters, circumference, and area were measured in the short axis. LAA left atrial appendage Mann-Whitney U test. The intraclass correlation coefficient (ICC) and their 95% confident intervals (CI) were estimated to assess intra-and inter-observer variability of 3D measurements of LAA orifice. Statistical significance was determined by a p-value < 0.05. Statistical analysis was performed using IBM SPSS Statistics software, version 23.
The percent increase was 18.8% for diameter 1, 7.3% for diameter 2, 21.0% for minimum diameter, 4.2% for maximum diameter, 9.4% for circumference and 25.8% for area. We compared the percent increase in area-the LAA orifice dimension that showed the greatest percentage increase after the Watchman implant-according to the AF type and the implanted device size. The percent increase in area was not statistically different in patients with paroxysmal AF than in those with other types of AF ( Eccentricity index decreased and showed a value closer to l after device placement (pre-device 1.23 ± 0.16 vs. postdevice 1.06 ± 0.06, p < 0.001), which means that the LAA orifice changed from an ellipsoid geometry to a more circular shape.
Peri-device leak was noted in 11 (22%) patients after device implantation, all size ≤ 2 mm. There were no significant differences in any of the LAA orifice dimensions (both pre-and post-Watchman implantation), in the clinical variables, or in the device size when comparing the patients with peri-device leak with those without residual flow ( Table 2). The post-device eccentricity index tended to be higher, meaning a less circular LAA orifice, in patients with peri-device leak compared with those without leak although the difference was not statistically significant (1.07 ± 0.10  Table 3 shows the results of the reliability analysis of 3D measurements of LAA orifice. Intra-observer reliability was good to excellent for diameter 1, and excellent for diameter 2, minimum diameter, maximum diameter, circumference, and area. Inter-observer reliability was good to excellent for all parameters. Table S1 in the Supplementary Material shows the reliability analysis separating the pre-and postdevice measurements. Intra-observer agreement was better in the post-device measurements, mainly for diameter 1 and minimum diameter, while the inter-observer agreement was better in the pre-device measurements.

Discussion
In this study, we assessed the dimensions and geometry of the LAA orifice pre-and immediately post-implantation of the Watchman device using 3D echocardiography. Furthermore, we compared the physiology of the LUPV by measuring its velocities. Our study found: (1) significant stretching of the LAA orifice following implantation of the device, (2) the LAA orifice geometry changed from an oval to a more circular shape, and (3) the LAA occlusion did not have a significant impact on the LUPV hemodynamics.
As has been demonstrated in previous pathological studies [20,21] and through the 3DE evaluation of LAA [17], the shape of the LAA orifice is consistently elliptical rather than circular. Current percutaneous closure devices, however, have a circular shape to fill or cover the ostium. When implanting a circular device over an oval orifice, the device must be oversized to achieve complete occlusion and avoid device dislodgement. This oversizing can have an impact on the dimensions and geometry of the LAA, as we observed in this study. Our results showed that there was a significant increase in the dimensions of the LAA orifice with an average percentage increase of 25.8% of its area independent of the type of AF and the size of the implanted device. Because the LAA is a very flexible structure and its ostium must seat a circular device, the LAA orifice can be expected to change from an oval to a more circular shape, as our results demonstrated (Fig. 1D). It has been suggested that peridevice residual flow after LAA closure might be attributed  to incomplete sealing of the LAA orifice caused by the mismatch between the round shape of closure devices such as the Watchman device and the usually oval shape of the orifice [22]. However, it seems to us that the change in LAA orifice geometry towards a more circular shape with the implanted device could prevent the presence of peri-device leaks and likely explains the low prevalence of peri-device leaks we found (22%) which was lower than previously reported [13,23]. The finding that there were no differences in the LAA orifice dimensions between patients with and without residual flow and that patients without peri-device leak tended to have a lower post-device eccentricity index, that is, a more circular LAA orifice, supports this theory.
Our results contrast with previous findings that patients with peri-device leaks have larger dimensions of the LAA orifice [24] which could be due to the fact that our cohort mainly included patients with paroxysmal AF with smaller LAA compared to those of the patients included in the previous study and that we only evaluated the immediate post-procedure TEE images. All leaks were not significant (< 5 mm) according to the previously accepted definition [25]. For successful percutaneous LAA occlusion and to avoid any complications, it is necessary to take into account not only the anatomy of the LAA but its spatial relationship to neighboring structures [13]. Oversizing the device runs the risk of overlapping or impinging upon neighboring structures, including the left pulmonary veins, the coronary arteries, and the mitral valve.
The LAA orifice is in close proximity to the LUPV, the mean distance between the rim of the ostium and the LUPV was 11.1 ± 4.1 mm in an autopsy study [21], therefore, there is a potential risk of LUPV compression when the LAA is occluded with any percutaneous device.
In a study that assessed the effects of the PLAATO (percutaneous LAA transcatheter occlusion) device, no evidence of disruption to pulmonary vein inflow or mitral valve function was found [26]. On the other hand, there is a report of a patient who developed compression around 50 percent of the left inferior pulmonary vein following LAA occlusion with an Amplatzer cardiac plug device [7]. In our cohort, the Watchman device had no impact on the functional integrity of the LUPV even though the LAA orifice showed significant expansion, which supports the previous findings reported in a study carried out in a canine model [12]. In this animal study, it was observed that the Watchman device did not obstruct or affect the LAA adjacent structures, resulting in a favorable surface recovery. On the other hand, the oversized Amplatzer Cardiac Plug device could potentially jeopardize the LAA neighboring structures and lead to delayed healing.
A new-generation Watchman FLX device has recently been introduced. Compared to the previous generation of Watchman that was used in this study, the Watchman FLX device can treat both smaller and larger LAA ostia (ranging from 14 to 31.5 mm) and it is 10 to 20% shorter in length enabling implantation even in shallow LAAs [27]. Studies that include the use of this new device are needed to understand how its design impact the dimensions and shape of the LAA orifice and its neighboring structures.

Limitations
The limitations of this study include that we only evaluated the impact of LAA occlusion with a single percutaneous device, since it is the only one that is implanted in our center, therefore the results cannot be extrapolated to the other percutaneous devices available worldwide. Second, we measured the changes in LAA only immediately after device implantation, however, due to the healing and remodeling process these changes may be different in the longterm follow-up. Third, our study population was relatively small and we only included those patients with 3D images available, leading to selection bias. Finally, the effect of the changes of the LAA orifice geometry on left atrial function remains unknown.

Conclusion
Our study shows that LAA occlusion with the Watchman device modifies the geometry and stretches the LAA orifice significantly becoming more circular following but does not result in any detrimental hemodynamic changes of the adjacent LUPV.

Supplementary Information
The online version contains supplementary material available at https:// doi. org/ 10. 1007/ s10554-022-02525-y. Data availability The data underlying this article will be shared on reasonable request to the corresponding author.

Author contributions
Code availability Not applicable.

Declarations
Conflict of interest Dr. Sugeng has received a research grant from and is a speakers' bureau member for Canon Medical; has received a research grant from and is a speakers' bureau and advisory board member for Siemens Healthineers; has received a research grant from and is an advisory board member for Philips Healthcare; and has received