Between May 2010 and January 2017, 82 patients with non-valvular AF (64.6% permanent, 4.9% persistent, 30.5% paroxysmal) with high embolic risk (CHA2DS2VASc score ≥2) and ineligible for OAC due to contraindication or failure of the therapy to prevent thromboembolic events, were selected for percutaneous LAA closure. Written informed consent was obtained from all patients.

Stroke was defined as a focal neurological impairment of sudden onset, and lasting more than 24 hours (or leading to death) and of presumed vascular origin (World Health Organization, 2006).

Bleeding complications were defined and classified according to the Thrombolysis In Myocardial Infarction system.26 Major bleeding was defined as intracranial hemorrhage or blood loss associated with ≥5 g/dl decrease in hemoglobin concentration, minor bleeding as observed blood loss associated with decrease in hemoglobin concentration of ≥3 and <5 g/dl or ≥4 g/dl decrease in hemoglobin concentration with no observed blood loss, and minimal bleeding as any clinically overt sign of hemorrhage associated with a <3 g/dl decrease in hemoglobin concentration.

Chronic kidney disease was defined according to the five stages of the Kidney Disease: Improving Global Outcomes classification.27

Patients were considered to be ineligible for OAC if (1) warfarin or other OAC was contraindicated due to the patient's comorbidities; (2) their international normalized ratio (INR) was labile or impossible to monitor; (3) blood dyscrasia secondary to hematological disease was present; (4) the patient had a history of intracranial bleeding or severe non-intracranial bleeding (not controlled by OAC or antiplatelet therapy); or (5) OAC was shown to be ineffective due to identification of thrombus in the LAA or the occurrence of thromboembolic events despite appropriate OAC.

Before the procedure, all patients underwent TEE after vascular filling with 500 ml of saline solution to characterize the morphology and to measure the depth and diameters of the LAA and to exclude the presence of thrombus. Implantation was guided by three-dimensional TEE under deep sedation or by ICE under local anesthesia, with the probe located in the left atrium.

In cases guided by TEE, standard views (0°, 45°, 90° and 135°) were used to guide the procedure and to confirm the dimensions of the LAA.

In cases guided by ICE, two femoral accesses were used: left to introduce the ICE catheter, and right to perform LAA closure. The ICE catheter was initially placed in the right atrium and rotated clockwise and tilted back to visualize the fossa ovalis and to guide transseptal puncture inferior and posterior in the fossa. The interatrial septum puncture was then dilated three times with the device delivery sheath to facilitate the passage of the ICE catheter to the left atrium (LA) through a single transseptal puncture. The ICE probe was then advanced with a slight anterior tilt to the LA, over the stiff guidewire in the pulmonary veins. In the LA, the ICE probe was positioned parallel to the LAA, rotated clockwise and tilted posteriorly, and three views were acquired, similar to those used for TEE: mitral view, without angulation of the ICE probe (corresponding to the 80-120° TEE view); aortic view, tilting the probe back (corresponding to the 0-50° TEE view); and posterior view, with the probe tilted sharply and parallel to the left superior pulmonary vein (corresponding to the 110-135° TEE view). ICE was used to guide transseptal puncture, cannulation of the delivery sheath in the LAA, controlled release of the device in the LAA, and assessment of the position of the radiopaque marker bands before final release of the occlusion device.

The two imaging techniques were compared in terms of device implantation success rates, two-dimensional and Doppler visualization, procedure and fluoroscopy time, complications, and length of hospital stay. Two occlusion devices were used: the AMPLATZER Cardiac Plug (ACP) or Amulet (St. Jude Medical, Plymouth, MN, USA); and the WATCHMAN (Boston Scientific, Plymouth, MN, USA). Device size was selected in accordance with TEE and angiographic measurements of the LAA in two views (left 20°/cranial 20°, and left 20°/caudal 20°). In cases guided by ICE, device size was selected on the basis of TEE measurements (on the previous day), angiographic measurements, and those acquired by ICE.

The therapeutic protocol following LAA closure consisted of dual antiplatelet therapy (aspirin 100 mg daily and clopidogrel 75 mg daily) for one month and single antiplatelet therapy for six months. Antithrombotic therapy could thereafter be discontinued at the discretion of the attending physician. All patients underwent transthoracic echocardiography (TTE) on the day after the procedure to exclude pericardial effusion, device dislodgement or migration, and peri-device flow. TEE was repeated one month following the procedure to screen for signs of incomplete endothelialization (peri-device flow ≥3 mm), formation of thrombus on the device, device dislodgement or migration, or signs of compression of the left superior pulmonary vein or the circumflex coronary artery, as indicated by changes in wall motion in the irrigated territory.

Clinical follow-up was maintained for 23±1 months, recording the occurrence of adverse events including death and stroke or transient ischemic attack (TIA), and with repeat TTE at three, six, nine and 12 months post-procedure. In the event of complications detected on the echocardiographic assessment, further control TEE exams could be scheduled at the discretion of the echocardiographer.

Complete clinical and echocardiographic follow-up was achieved and no patients were lost to follow-up, the cohort at the end of follow-up consisting of 82 patients.

A descriptive analysis was performed of continuous variables, which were calculated as mean and standard deviation, and of categorical variables, which were described as absolute and relative frequencies. Relative frequencies were expressed as percentages rounded to one decimal place. IBM SPSS version 20.0 was used for the statistical analysis.

The study population consisted of 82 patients with a history of non-valvular AF, of whom 64.6% were male and mean age was 74±8.0 years. The mean CHA2DS2VASc stroke risk score in this population was 4.7±1.4 and the mean HAS-BLED bleeding score was 3.3±1.0 (≥3 in 67.1% of patients). A history of ischemic stroke and severe bleeding was observed in 41.5% and 59.7% of patients, respectively.

The main characteristics of the study population and contraindications for OAC are shown in Table 1 and Figure 1. The most frequent indication for LAA closure was major bleeding under OAC (63.0% under warfarin and the remainder under NOACs). Another important indication was labile INR, which included patients with multiple bleeding episodes without clinical relevance (ecchymosis or hematoma on the limbs) associated with difficulty in monitoring INR or in maintaining it within the therapeutic range (2.0-3.5) (>50% of measurements in consecutive assessments) and those who were unable to have their INR monitored at least once per month, together with the physician's decision not to continue to prescribe warfarin. It should be noted that most cases of labile INR occurred in the early part of the study (2010-2013), when there was less use of NOACs.

Figure 1.

Contraindications for oral anticoagulation. Figures for labile INR precede the introduction of the novel oral anticoagulants.

(0.07MB).

Of the 82 patients selected, percutaneous LAA closure was not possible in three (3.6%), two due to the small size of the LAA and in the third due to the presence of severe venous disease; the procedural success rate was thus 96.3%. The mean device size implanted was 23.0±3.0 mm.

Table 2 presents the main technical characteristics of the LAA closure procedure.

With regard to the imaging modality used during the procedure, 56 implantations (68.2%) were guided by TEE under deep sedation and 26 (31.7%) by ICE under local anesthesia. In recent years (2016-2017) ICE was used more frequently, in 80% of procedures. The imaging modality was not chosen on the basis of the anatomical complexity of the interatrial septum or of the LAA. In our sample, none of the procedures guided by ICE needed to be converted to TEE during the procedure.

The LAA and the implantation procedure were adequately visualized by both methods (Table 3). Procedure and fluoroscopy time were shorter and there were fewer venous access complications in procedures guided by ICE (p<0.001). There were no differences between the methods in closure rates, occurrence of leaks or existence of residual interatrial communication at six months. When the last 10 TEE-guided procedures were compared to the last 10 guided by ICE, statistically non-significant trends were seen in closure success rates (90% vs. 100%, respectively, p=0.07), occurrence of leaks (30% vs. 10%, p=0.09), existence of residual interatrial communication at six months (70% vs. 50%, p=0.06), and complications, both periprocedural and during clinical follow-up (30% vs. 20%, p=0.06).

In the first 24 hours after the procedure there were two major bleeding events (2.4%) with cardiac tamponade requiring pericardiocentesis, and two cases of upper airway bleeding (2.4%) related to tracheal intubation, resolved by conservative measures. Eight small inguinal hematomas (9.8%) at the venous puncture site and one femoral artery pseudoaneurysm (1.2%) were also recorded, all resolved by minimally invasive measures.

No other major periprocedural complications were recorded, including device embolization, stroke/TIA or death.

Table 4 presents the major complications that occurred during clinical follow-up.

It should be noted that the patient with thrombus on the ACP device (with no associated shunt) had dual antiplatelet therapy replaced by OAC and that the thrombus regressed completely after six months, with no associated complications, after which the patient returned to single antiplatelet therapy.

A peri-device leak with color jet <3 mm was detected in seven cases (8.5%) but was not confirmed by TEE at three months. There were thus no persistent leaks (>1 month).

During clinical follow-up (23±1 months, minimum one month and maximum 79 months), there was one ischemic stroke and two major gastrointestinal bleeding events requiring transfusion. Four (4.9%) non-cardiovascular deaths were recorded, three from colon cancer and one from septic shock unrelated to endocarditis 10 months after LAA closure. Clinical follow-up was ≥1 year in 55% of patients. Figure 2 shows all-cause death in the population throughout the study period.

Figure 2.

Survival curve for all-cause mortality during clinical follow-up.

(0.06MB).

Table 5 presents a comparison between antithrombotic therapy at baseline and one month after LAA closure.

In this population, the mean CHA2DS2VASc stroke risk score was 4.7±1.4, while the mean HAS-BLED bleeding risk score was 3.3±1.0. Embolic and bleeding events were less frequent than expected from the observed CHA2DS2VASc (0.6% vs. 6.3%; p<0.001) and HAS-BLED (1.2% vs. 4.1%. p<0.001) scores (Figures 3 and 4).

Figure 3.

Comparison of actual and expected cardioembolic events in the study population.

(0.14MB).

Figure 4.

Comparison of actual and expected bleeding events in the study population.

(0.1MB).