Evaluation of endothelialization of left atrial appendage occluder by three-dimensional transesophageal echocardiography and its inuence factors

Background This study evaluated surface endothelialization of Watchman occluder (Boston Scientic) through comparing relevant indicators before and left atrial appendage closure (LAAC) in 68 non-valvular atrial brillation (AF) patients. Methods Patients were monitored preoperatively by transthoracic echocardiography (TTE), transesophageal two-dimensional and three-dimensional echocardiography (2D-TEE, 3D-TEE), intraoperatively by 3D-TEE and angiocardiography, and followed at 3 and 6 months after operation by TEE. Result It showed that the inner diameters of LAA ostium measured by TEE at 0°, 90° and 135° before operation were correlated with corresponding maximum diameters measured by angiocardiography. The depth of LAA ostium measured at 90° was correlated with the maximum depth measured by angiocardiography. The left atrial diameter, left ventricular diameter and left ventricular free wall E/E' decreased postoperatively. Compared T-test showed that BNP level and persistent AF might be risk factors for spontaneous echo contrast (SEC) or thrombosis of LAA. Multiple linear regression identied persistent AF, preoperative LAA spontaneous echo contrast or recent thrombotic event, and history of embolism in other parts were the inuence factors of occluder endothelialization. Conclusion: LAAC (±s) and was compared by Mann-Whitney U test. CHA 2 DS 2 -VASc scores were described as median, upper and lower quartiles and were compared by Kruskal-Wallis One-way ANOVA. Correlation between data measured by TEE and by angiocardiography were analyzed by Pearson correlation analysis. Changes of cardiac function before, 3 months after and 6 months after surgery were compared by One-way ANOVA with Tukey’s multiple comparison test. Categorical data was compared by Fisher’s exact test. Multiple linear regression was used to analyze the inuence factors of postoperative surface endothelialization of occluder. P < 0.05 indicated statistical signicance.


Background
Atrial brillation (AF) is the most common type of arrhythmia in clinical practice. In Chinese population the prevalence of AF is 0.42% in people aged between 35-59 years old and 1.83% in people over 60 years old [1]. AF leads to changes in the hemodynamics of the heart cavity and increases the possibility of atrial thrombosis. A study showed that more than 90% of atrial thrombi caused by non-valvular atrial brillation (NVAF) originate from the left auricle [2]. NVAF is an independent risk factor for ischemic stroke. The risk of thrombotic events is 5% per year. The risk of stroke in AF patients is 4-5 times higher than that in non-AF patients [3], and the 5-year stroke rate of AF patients is 20%-25% [4]. At present, clinical prevention and treatment of NVAF patients with ischemic stroke still mainly rely on long term use of anticoagulants. However, the clinical application and e cacy of anticoagulants are limited by many factors such as dosage, patient compliance and adverse reactions. According to the literature, among the NVAF patients with CHA 2 DS 2 -VASc score ≥ 2 in Chinese population, 36.5% of them receive anticoagulants [5], and 44.4% of the anticoagulant users withdraw their medication within one year [6].
Left atrial appendage closure (LAAC) is a surgical procedure that blocks the ostium of left atrial appendage (LAA) to prevent thrombosis and thrombus detachment, so as to reduce the risk of stroke. LAAC has been recommended by relevant international guidelines as a new technology to prevent ischemic stroke in NVAF patients [7,8], and its long-term prognosis is relatively reliable [9][10][11]. Since the Watchman LAA occluder (Boston Scienti c, USA) appeared on the market in China in 2014, LAAC has developed rapidly in China, and the application of real-time transesophageal three-dimensional echocardiography (RT-3D-TEE) in interventional cardiosurgery has gradually increased. It does not obstructs surgical eld of vision and plays important roles in intraoperative monitoring and postoperative follow-up. The purpose of this study is to evaluate the degree of occluder endothelialization after LAAC by 3D-TEE and discuss its in uence factors based on the follow-up information of LAAC patients.

Patients
This study enrolled 68 patients with the diagnosis of AF and were scheduled for LAAC using Watchman (Boston Scienti c, USA) occluder in our hospital from January, 2017 to September, 2019, including 48 males and 20 females, aged 45-86 years old (63.68 ± 9.43 years). Among them, 54 patients (79.41%) had persistent AF, 14 (20.59%) had paroxysmal AF. Thirty-one (45.59%) patients had spontaneous echo contrast (SEC) of LAA blood ow or previous thrombosis history, and were treated with anticoagulant until meeting the surgical criteria. The Inclusion criteria of the study were: (1) age ≥ 18 years; (2) with NVAF, including persistent and paroxysmal types; (3) CHA 2 DS 2 -VASc score ≥ 1; (4) unable to take anticoagulants orally for a long time or had high risk of bleeding with anticoagulants. The exclusion criteria were: (1) cardiac insu ciency (EF < 30%); (2) preoperative TEE showed thrombosis or suspicious thrombosis in LAA (except for thrombus disappeared after treatment); (3) atrial septal disease (defect, aneurysm, space-occupying lesion, etc.); (4) complicated with severe hepato-renal insu ciency, coagulation dysfunction, arterial aneurysm, cardiac valvulopathy, pulmonary insu ciency, or severe lesions of other organs. The age, gender, BNP, CHA 2 DS 2 -VASc score and history of the patients were collected. This study was approved by the Ethics Committee of Yantai Yuhuangding Hospital, and informed consent of all patients were obtained.

Preoperative ultrasonography
All patients were examined for LAA thrombus by transthoracic echocardiography (TTE) and TEE before operation using Philips EPIQ 7c Color Doppler Ultrasound Diagnostic System, with S5-1 transthoracic probe (frequency 1-5 MHz) and X7-2t transesophageal 3D probe (frequency 2-7 MHz). The inner diameter and depth of the ostium of LAA were measured at 0°, 45°, 90° and 135°. The inner diameter was de ned as the distance between the ostium of the circum ex branch of the left coronary artery in LAA wall and about 2 cm below the extension ridge of the left superior pulmonary vein (LSPV). The depth was de ned as the distance between the middle point of the radial line of the LAA ostium and the tip of LAA ( Figure   1). The shape and lobulation of LAA were displayed in multi plane by 3D imaging technology.

Angiocardiography
After achieved anesthetization of patient, atrial septal puncture was performed via the femoral vein, and the catheter was sent for LAA angiocardiography to measure the maximum diameter and maximum depth of LAA ostium using the Philips Allura Xper ED20 digital at-panel X-ray system.

Intraoperative ultrasonography
Appropriate Watchman occluder (Boston Scienti c, USA) was selected according to the maximum inner diameter of the LAA ostium measured by RT-3D-TEE. The 3D con guration, size, and lobulation of LAA and its spatial position relationship with LSPV and mitral valve were observed. The atrial septum was punctured under the guidance of RT-3D-TEE. The puncture site was selected according to the position of LAA, which is usually posterior inferior to the oval fossa. The occluder was released at the right angle and position guided by RT-3D-TEE and angiocardiography. RT-3D-TEE was used to observe the real-time location, con guration, and shoulder protrusion of the occluder, its spatial relationship with the cardiac valve and the LSPV ostium, as well as any residual ow and pericardial effusion ( Figure 2). Successful occlusion should meet the PASS criteria [12]: (1) Position: device was distal to or at ostium of LAA; (2) Anchor: xation anchor engaged or device was stable as determined by a 'tug test'-retracting the deployment knob and letting go, to assess return to original position; (3) Size: device compressed 8%-30% of original size; (4) Seal: device covered well, and residual ow was < 3 mm as seen on TEE.

Follow-ups
Patients were followed at 3 months and 6 months after operation for ECG, TTE and TEE to observe symptom improvement, location of occluder, residual ow of atrial septum and any thrombosis on the surface of occluder. Thickness of the endothelium on the occluder surface was also measured ( Figure 3).
Statistical analysis SPSS 20.0 software was used for statistical analysis. Measurement data was described as mean ± standard deviation (±s) and was compared by Mann-Whitney U test. CHA 2 DS 2 -VASc scores were described as median, upper and lower quartiles and were compared by Kruskal-Wallis One-way ANOVA. Correlation between data measured by TEE and by angiocardiography were analyzed by Pearson correlation analysis. Changes of cardiac function before, 3 months after and 6 months after surgery were compared by One-way ANOVA with Tukey's multiple comparison test. Categorical data was compared by Fisher's exact test. Multiple linear regression was used to analyze the in uence factors of postoperative surface endothelialization of occluder. P < 0.05 indicated statistical signi cance.

Results
Operation outcomes LAAC for all 68 AF patients were successful, in which 53 cases (77.9%) were completely occluded. Sixtyfour patients (94.12%) underwent AF radiofrequency ablation at the same time. Fifteen cases (22.1%) had small amount of residual ow during the operation, with ow stream < 3 mm. By 3 months after operation, 4 out of the 15 cases still had residual ow, and another 17 cases developed residual ow as well. Among the 68 patients, shoulder protrusion occurred in 26 (38.2%) cases during operation and in 18 (26.5%) cases by 3 months post operation (including 12 standing cases and 6 newly developed cases). Thrombus on the surface of occluder was observed in 4 (5.8%) patients at 3 months follow-up, which still persisted in 2 patients by 6 months post operation. Pericardial effusion was found in 1 case (1.5%) during the operation, which disappeared after 3 months. There were no serious complications such as device detachment during or after the operation.

Changes of cardiac function after LAAC
Summarized in Table 1 was the comparison of cardiac function indicators before, 3 months, and 6 months after the operation. The left atrial diameter (LAD), left ventricular diameter (LVD) and E/E' (peak transmitral velocity-to-distance of left ventricular free wall mitral annulus movement ratio) signi cantly decreased with time (P < 0.05), and the differences were signi cant between any two groups (P < 0.05).

Risk factors of LAA SEC or thrombosis
Patients were divided into SEC/thrombosis group and non-SEC group. Patient's age, gender, smoking history, hypertension, stroke, type of AF, CHA 2 DS 2 -VASc score, and BNP level were compared between the two groups. We found signi cant difference in BNP level and persistent AF, indicating that high BNP level and persistent AF might be risk factors for SEC or thrombosis of LAA (Table 2).

In uence factors of endothelialization of occluder surface after LAAC
We compared the thickness of endothelialization of occluder surface 3 months and 6 months after LAAC. Result showed increased thickness of endothelialization at 6 months (1.33±0.23 mm) than 3 months  Table 3, we found that persistent AF, preoperative SEC of LAA blood ow or recent thrombotic events, and history of embolism in other parts were the independent in uence factors of endothelialization of occluder surface (P < 0.05).

Discussion
In AF patients, the left atrium loses normal systolic function and reshape of LAA occurred, which slows down blood ow, increases risks of thrombosis and stroke. According to the AHA/ACC/HRS guideline for the management of patients with AF, anticoagulants (class IA) are recommended for patients with CHA 2 DS 2 -VASc score ≥ 2 to prevent thrombus [13], while the 2016 ESC guideline recommends use of anticoagulants (class IA) for male with CHA 2 DS 2 -VASc score ≥ 2 and for women with CHA 2 DS 2 -VASc score ≥ 3 [7]. However, the clinical application and e cacy of anticoagulants are limited by many factors such as dosage, patient compliance, and adverse reactions. ESC guideline recommends LAAC (class IIB) for patients with contraindications to long-term anticoagulation treatment [7]. The 5-year PREVAL [9] and PROTECT AF study [10], which enrolled 1114 patients including 732 implanted with LAA occluder, reported that Watchman occluder provided similar protective e cacy against ischemic stroke with that of warfarin in AF patients (HR 1.71, P = 0.080). In addition, the occluder group had signi cantly reduced risks of hemorrhagic stroke (HR 0.20, P = 0.002), heart related mortality (HR 0.59, P = 0.027) and unexplained mortality (HR 0.73, P = 0.035) than warfarin group [11]. The long-term e cacy of other brands of occluder still required further study. In recent years, LAAC has been recommended as a new technology for treating thromboembolic events in AF patients by international guidelines [7,8].
In this study, 15 out of 68 AF patients treated with Watchman occluder had intraoperative residual ow of less than 3 mm and another 17 cases developed residual ow 3 months after the operation. All the 17 cases had received radiofrequency ablation before the implantation of occluder. Therefore, the development of residual ow was likely due the slight displacement of occluder caused by restored active systolic function of LAA after surgery, or due to subsiding of edema in the extension ridge of LSPV. Since the residual ows were all less than 3 mm, no clinical interventions were needed and patients were further followed.
RT-3D-TEE was used for preoperative evaluation and intraoperative monitoring in this study. Several studies have found that the accuracy of 3D-TEE was higher than that of 2D-TEE in pre-LAAC evaluation [14,15], and that 3D-TEE was highly correlated with measurements by intraoperative angiocardiography, therefore can be routinely used for intraoperative monitoring of LAAC [16]. In this study, we found correlation between the inner diameters of LAA ostium measured in 0°, 90° and 135° section by preoperative TEE and the maximum diameter measured by angiocardiography (P < 0.05), and the TEE measurement in 135° section was closest to the angiocardiography measurement. The depth measured by TEE in 90° section was correlated with the maximum depth measured by angiocardiography (P < 0.05).
In AF patients, the contractility of left atrium decreases. As it lasts, the electrical remodeling, tissue remodeling and ion channel remodeling of the left atrium become more serious, which leads to a series of electrophysiological dysfunction and attenuates the connection between atrial myocytes and the signal transduction between atrial cells [17,18]. It has been reported that the peak strain of each left atrium wall in AF patients was signi cantly lower than that in non-AF patients, and that the pressure and load of the left atrium did not improve signi cantly shortly after LAAC [19]. The reporter believed that although LAAC could be used as a preventive measure to reduce the risk of thromboembolism, it did not fundamentally change the heart rate of AF patients, therefore had no signi cant impact on the mechanical function of left atrium. In our study, 64 patients (94.12%) underwent AF radiofrequency ablation at the same time. We found gradually reduced LAD, LVD and E/E' at 3 months and 6 months post operation, suggesting improved function of left atrium and left ventricle in AF patients after improving the AF rhythm. Six cases developed shoulder protrusion 3 months after operation, which was likely attributed to the improvement of systolic function of left atrium and LAA after restoring the sinus rhythm. As the shoulder protrusion did not exceed 1/3 of the inner diameter of the occluder, no clinical intervention was applied.
Evaluating the process of endothelialization on the occluder surface after LAAC is important for the postoperative treatment of AF patients. Delayed endothelialization might be an in uence factor of thrombosis after device implantation [20]. Therefore, anti-thrombotic therapy must be carried out before the complete endothelialization of the occluder surface to prevent device-related thrombus event. Previous studies on the endothelialization of LAA occluder were limited to animal experiments, while only a few human cases were analyzed. Massarenti et al. [21] reported that a patient with hereditary hemorrhagic telangiectasia (HHT) suffered stroke 10 months after the implantation of Watchman occluder, which showed no endothelialization of the occluder surface after removing the device. HHT is an autosomal dominant hereditary disease with an incidence rate of about 1/5000. In HHT patients, selective expression of the genes encoding TGF-β pathway proteins leads to vascular endothelialization dysfunction, disorders of vascular smooth muscle development, and endothelial remodeling, which caused incomplete endothelialization of the occluder surface. Animal studies have found that complete endothelialization of LAA occluder occurred about 3 months after operation [22], and complete endothelialization of both sides of the occlusion device for atrial septal defect occurred 3 months after operation as well [23]. Therefore, in this study, we selected the time point of 3 months after LAAC for measuring the thickness of the surface endothelialization of the occluder. We found no statistically signi cant difference between the thickness of surface endothelialization 3 months and 6 months after LAAC. It is considered that the short-term effect on the area of occluder surface endothelialization after LAAC is greater than on its thickness. Previous research has reported incomplete surface endothelialization of the Amplatzer Cardiac Plug (St. Jude Medical) in 61% of LAAC patients 10 ± 6 months after the operation [24]. Multiple linear regression analysis found that persistent AF, SEC or recent thrombotic event, and thrombosis in other parts were independent factors in uencing the degree of occluder surface endothelialization. Granier et al. [24] de ned incomplete endothelialization as residual permeability on cardiac computed tomography without peridevice leak on TEE at follow-up after implantation of Watchman device (Boston Scienti c) or Amplatzer Cardiac Plug (St. Jude Medical), and found incomplete endothelialization might be related to diabetes, persistent AF and implantation of oversized occluder. This was consistent with our nding that persistent AF was an in uencing factor of occluder surface endothelialization. AF increases thrombogenic tendency by ful lling the Virchow's triad [25]: alterations in normal blood ow, endothelium dysfunction, and abnormal coagulationanticoagulation-brinolytic system. In addition, in patients with persistent AF, the level of endothelial factor such as vWF (produced by endothelial cells and induced secretion by in ammatory response, oxidative stress, activation of renin-angiotension system) was signi cantly higher than that in patients with paroxysmal AF [26]. Electrical cardioversion and radiofrequency ablation could restore sinus rhythm and at the same time change endothelial function. Prosperi-Porta et al. [27] reported a case of secondary multiple organ thrombosis after LAAC. Autopsy showed that the thrombi formed earlier than LAAC. This was consistent with our nding that abnormal coagulation-brinolysis function in AF patients, combined with the difference in the e cacy of oral anticoagulants, increased the risk of thrombosis, which might be a potential predictor for delayed endothelialization after LAAC. Besides, alteration of BNP level could affect SEC and thrombosis of LAA. Therefore, routine monitoring of BNP level before surgery and during the follow-ups could evaluate the left atrial blood ow status, and predict the long-term endothelialization status of the occluder surface after LAAC.
Currently, there is no gold standard for the evaluation of occluder endothelialization after LAAC. In this study, only 3D-TEE was used for evaluation of endothelialization, and the evaluation time was set according to previous animal studies. However, there is no relevant literature discussing the difference in endothelialization time of occluder between human and animal, which still requires further study.

Conclusion
In this study, we found that 3D-TEE played an important role in preoperative evaluation and postoperative monitoring of transcatheter LAAC. The measurements by 3D-TEE were well correlated with those by intraoperative angiocardiograph. After the operation, the cardiac function of AF patients gradually improved. Persistent AF, SEC of LAA blood ow or recent thrombotic event, and the history of embolism in other parts were in uence factors of the endothelialization of occluder.   Tables   Table 1. Comparison of parameters measured before operation, 3 months after and 6 months after left atrial appendage closure.
Before operation 3 months post operation 6 months post operation F P