Very delayed hemorrhagic pericardial effusion following percutaneous left atrial appendage occlusion

Abstract

Pericardial effusion is a serious complication following percutaneous left atrial appendage occlusion and is associated with adverse outcomes. While acute effusions occur from procedure associated macro-perforations, delayed effusions are less well understood. Reported here is a case of a hemorrhagic pericardial effusion presenting almost 6 months following left atrial appendage occlusion and successfully treated with corticosteroids.

Learning objective

Delayed pericardial effusions following left atrial appendage occlusion occur due to microperforations. These pericardial effusions can cause hemopericardium and pericardial inflammation. Anti-inflammatory therapy may promote resolution and prevent adverse sequalae.

Introduction

Pericardial effusion (PE) is a rare but serious complication following left atrial appendage occlusion (LAAO) [1,2]. Contemporary studies demonstrate PE incidence of 0.51–2 % following LAAO [[1], [2], [3], [4], [5]], but predominantly feature acute or sub-acute PEs detected within days following LAAO. PEs in the acute phase are hemorrhagic from procedure-associated macroperforation. Sub-acute PEs have been hypothesized to occur from device-associated microperforations leading to hemopericardium and pericardial inflammation [6,7]. Few reports have described delayed PEs presenting months following LAAO [3,[5], [6], [7]].

While many PEs result in tamponade requiring emergent percutaneous or surgical evacuation, medical strategies to prevent recurrence or curb the development of effusive constrictive pericarditis are less well established. Applying the principles of pericarditis treatment, successful management has been achieved with the use of anti-inflammatory medications [6,7]. Herein we describe a patient who presented approximately 6 months following LAAO from the Amulet device with a hemorrhagic PE without hemodynamic tamponade and who was managed successfully with corticosteroid therapy.

Case report

This case involved a 75-year-old male with persistent, symptomatic atrial fibrillation (A.FIB) that was recurrent despite multiple direct current synchronized cardioversions (DCCV) and radiofrequency ablation. With a CHADS2-VASC score of 5 and HAS-BLED score of 4, he had been placed on rivaroxaban for cardioembolic stroke prevention but had undergone two hospitalizations for gastrointestinal (GI) bleeding requiring transfusion. Due to concern for multiple sources of GI bleeding and recurrent bleeding events, the patient was referred for LAAO.

Cardiac computed tomography angiography (CTA) was used for pre-procedural planning (Fig. 1). The maximum landing zone width of the left atrial appendage (LAA) measured 21 mm, the minimum diameter measured 20.5 mm, and the LAA depth measured 19 mm. A trivial PE was present on pre-procedural transthoracic echocardiogram (TTE). The procedure was performed under conscious sedation using midazolam and fentanyl. Access was obtained in the right femoral vein. Intravenous heparin was given to achieve an activated clotting time of >250 s. Intracardiac echocardiography was used for implant imaging (Fig. 1). Transseptal puncture was performed using a Versacross system (Baylis Medical, Ontario, Canada) and an 8F Acuson Acunav probe (Siemens Healthineers, Cary, North Carolina) was placed in the LAA across a single transseptal tract. Contrast injection using a 6F pigtail catheter further defined LAA anatomy (Fig. 1). A 25 mm Amulet device (Abbott, Lake County, Illinois) was then implanted and deployed after confirmation of appropriate release criteria (Fig. 1). Post procedure TTE performed 5 h after the procedure revealed trivial PE unchanged from the previous examination and the patient was discharged on dual antiplatelet therapy (DAPT) with aspirin and clopidogrel.

Fig. 1.

Cardiac computed tomography angiography before (A,C) and after (B,D) LAAO. Intra-cardiac echocardiography before (E) and after (F) LAAO. Contrast angiography before (G) and after (H) LAAO.

LAA, left atrial appendage; LAAO, left atrial appendage occlusion.

Due to recurrence of symptomatic A.FIB necessitating rhythm control, the patient was placed on apixaban for 3 weeks, clopidogrel was discontinued, and successful DCCV was performed 6 weeks post-LAAO. Aspirin was continued with an additional 4-week course of apixaban after cardioversion. Then the patient was switched back to DAPT. At 5 months post-LAAO, the patient experienced symptomatic A.FIB once again and underwent another successful DCCV and was continued on DAPT.

A surveillance cardiac CT scan was planned at 4 months, in line with the European Heart Rhythm Association/European Association of Percutaneous Coronary Interventions expert statement recommending imaging within 6–24 weeks post-LAAO with either CT or transesophageal echocardiogram (TEE) and in accordance with the post-procedural imaging protocol of the CHAMPION-AF trial but not performed until 6 months after implant and demonstrated a moderate PE (Fig. 2). The attenuation coefficient of this fluid (57 ± 14.0 Hounsfield units) was consistent with an exudative hemorrhagic effusion [8]. CT imaging with multiplanar reconstruction demonstrated complete apposition of the Amulet device with the LAA orifice along with absence of peri-device leak or device-related thrombus (Fig. 1). TEE was not expected to provide additional relevant information and was thus not performed. TTE the following day confirmed moderate PE without tamponade (Fig. 3A, B, C). Since the patient had completed a 6-month course of clopidogrel, he was continued on aspirin alone. The patient remained asymptomatic and inflammatory biomarker testing was not performed at the time. Nonetheless, the patient, who weighs 75 kg, was empirically placed on a 5-week prednisone taper (80 mg/day in week 1, 60 mg/day in week 2, 40 mg/day in week 3, 20 mg/day in week 4, and 10 mg/day in week 5) to treat potential pericardial inflammation from hemopericardium. Non-steroidal anti-inflammatory drugs were avoided due to the patient's history of GI bleeding. Repeat TTE at the completion of the steroid course showed near resolution of the effusion (Fig. 3D, E, F).

Fig. 2.

Cardiac computed tomography angiography showing moderate pleural effusion (arrows).

Fig. 3.

TTE one day after computed tomography angiography confirmed moderate pericardial effusion without tamponade (asterisk in A, B, C). Repeat TTE after completion of the steroid course showed near resolution of the effusion (D, E, F).

TTE, transthoracic echocardiogram.

Discussion

PE is an important complication following LAAO. The PROTECT-AF trial demonstrated an incidence of PE requiring intervention at 4.8 % [1]. However, with newer generation devices and increased operator experience, a significantly decreased incidence of 1.9 % was reported in the subsequent PREVAIL trial and European/US registries have demonstrated PE rates of 1.24–1.50 % [[1], [2], [3], [4], [5]]. In fact, PE incidence in our LAAO experience was 1.35 % [3]. Periprocedural PEs have been associated with prolonged length of stay, lower likelihood of discharge on antiplatelet therapy predisposing to device thrombus and stroke, tamponade requiring percutaneous/surgical intervention, and rarely death [1,2]. As such, clinical vigilance and careful surveillance is required.

The majority of periprocedural PEs following LAAO are hemorrhagic, manifest within hours to days following LAAO, and are attributed to macroperforation either from transseptal puncture, manipulation of the delivery system, or device instrumentation. Delayed PEs are far less common and only four case reports describe a PE identified at 42–180 days following LAAO [6,7,9,10]. Such delayed PEs are thought most commonly to occur from device anchor microperforation. Such microperforations may lead to hemopericardium, pericardial inflammation, and exudative effusions [6,7].

Clinical factors pre-disposing to post-LAAO PEs include female sex, obesity, older age, decreased albumin, elevated CHADS2-VASC score, prior bleeding events, and pre-procedural antiplatelet therapy [1,2]. Our elderly patient with high thromboembolic risk, multiple episodes of GI bleeding, and concomitant antiplatelet/anticoagulant therapy was at high risk for developing PE. Sinus rhythm and paroxysmal A.FIB have also been independently associated with PE [1]. Mechanistically, LAA contraction during sinus rhythm may exert greater force on device anchors leading to microperforation. While our patient had a high A.FIB burden, he underwent two DCCV attempts following LAAO and the transient return to sinus rhythm may have increased risk for PE development. At this time however, no association has been identified between DCCV and risk of pericardial effusion post LAAO.

This case featured an Amulet device for LAAO, which has been associated with a higher rate of PE compared to the WATCHMAN device (2.4 % vs 1.2 %) [4]. Device design can provide insight into the mechanism of delayed PEs. The Amulet device employs a dual sealing mechanism of a plug and a disc with the former inserted in the LAA and the latter covering the LAA orifice. Oversizing with a 25 mm Amulet device was performed to allow for appropriate device compression. The circumferential fixation anchors of the Amulet are designed to prevent device embolization but are suspected of predisposing to microperforations in this case. Of note, the anchors on the Amulet device are about 2.2 times longer than those on the WATCHMAN and are hypothesized to be the cause of the higher incidence of PEs with the former [3]. PE from the Amulet disc perforating the LA wall or left circumflex artery have been reported and these mechanisms would likely result in a more acute PE, requiring invasive management [10]. However, in this case CT imaging did not demonstrate erosion of the disc in a nearby cardiac structure, which combined with the indolent presentation leads us to believe that the PE occurred from fixation anchor-induced microperforation over the course of months.

Given the absence of cardiac tamponade, there was no indication for pericardiocentesis and thereby PE fluid studies were not performed. Our suspicion for a hemorrhagic PE was based on the attenuation of pericardial fluid on CTA [8]. Unfortunately, no guidelines exist for the management of PE following LAAO. Despite the absence of pericarditis symptoms and hemodynamic stability, we were concerned that the introduction of blood within the pericardial space would lead to inflammation that may result in PE expansion and/or formation of effusive constrictive pericarditis. In fact, Khan et al. and Nandekeolyar et al. both report delayed PEs following LAAO leading to effusive constrictive pericarditis treated successfully with anti-inflammatory regimens of colchicine and steroids without requiring pericardial stripping; however, the latter case required steroids after colchicine alone proved insufficient [7,9]. To ward off aforementioned consequences early, we proceeded with an aggressive anti-inflammatory steroid taper and the patient had resolution of PE with no signs or symptoms of constriction at steroid completion.

Conclusion

Delayed PE following LAAO may occur due to microperforations leading to hemopericardium and pericardial inflammation. Anti-inflammatory therapy may promote resolution and prevent adverse sequalae.

CRediT authorship contribution statement

NA and RA contributed equally to data curation; writing – original draft preparation; writing – review & editing. MF, NT and LD contributed to manuscript review & editing. SF contributed to conceptualization; manuscript review & editing: project supervision.

Consent statement

Written informed consent was obtained from the patient for publication of this case report, including accompanying images.

Sources of funding

No sources of funding to report.

Declaration of competing interest

Dr. Filby is a consultant for Boston Scientific. The remaining authors have no personal or institutional interest with regards to the authorship and/or publication of this manuscript.