Fig. 18.1
Delayed presentation with pericardial effusion in a 45 year-old female, 1 month following combined atrial fibrillation ablation procedure and left atrial appendage closure using ACP device. (a) Subxiphoid transthoracic echocardiographic window showing circumferential effusion (arrow) that is amenable to a subxiphoid approach to pericardiocentesis. (b) Flouroscopic image showing typical appearance of 0.035″ wire in pericardial space. ACP device indicated by white arrow
Fig. 18.2
Acute pericardial effusion in 74 year-old female with clinical tamponade within 12 h of LAA closure with ACP device. Patients body habitus and location of the effusion necessitated a parasternal approach to pericardiocentesis. Tip of micropuncture needle indicated by black arrow. ACP device indicated by black arrow
Fig. 18.3
Portex 11 cm length 18G tuohy spinal needle (b) and trochar (a). Micropuncture kit with 7 cm 21G Needle (c), 0.018 in. stiff wire (d), and stiffened dilator (e)
Frank cardiac perforation during LAA occlusion procedures is a rare event, with individual cases reported in most large series [5, 6], suggesting a frequency of <0.5 % (Fig. 18.4). Perforation will typically lead to an immediate pericardial effusion and clinical tamponade. Blood loss can be large and life-threatening. Following pericardiocentesis , it is recommended that the blood drawn from the pericardial space be immediately administered into the central venous access to try to minimize blood loss and subsequent need for transfusions. The standard approach for dealing with a cardiac perforation is surgical repair, particularly if the patient is a reasonable operative candidate. Based on a single case experience of the authors (IC and KW) and anecdotal reports, we would recommend that reversal of anticoagulation and administration of clotting factors should not be performed until the operative team is in place and emergency sternotomy is possible. The latter agents may result in formation of large amounts of clot in the pericardial which can occlude the pericardial drain but without sealing of the perforation site. This can lead to tamponade that cannot be rescued with manipulation of the pericardial drain.
Fig. 18.4
LAA perforation during LAA occlusion procedure to seal chicken wing shaped appendage . (a) LAA angiography shown at baseline. (b) Photograph taken at the time of emergency sternotomy to deal with perforation of the LAA (indicated by white arrow) caused by a wire perforation
Rare instances of percutaneous management of a LAA perforation have been reported using an Amplatzer PFO or ASD closure devices [11]. In such cases, the perforation was likely caused by the delivery sheath, and the perforation was recognized with the sheath in the pericardial space, allowing delivery of the device through the sheath with deployment of the traditional left atrial disk in the pericardial space and the right atrial disk in the LAA. It would be prudent to recommend that such attempts should only be made by experienced operators and with the availability of the cardiothoracic surgical team in case these maneuvers are not successful.
Device Embolization
Device embolization is one of the most feared complications of LAA closure procedures. The device embolizes initially to the left atrium, and will typically make its way across the mitral valve into the left ventricle. For small and medium-sized devices in the absence of significant aortic valve disease, the device may subsequently embolize to the thoracic or abdominal aorta. If lodged in the left ventricle, the device may become entrapped in the left ventricular outflow tract (LVOT) or the mitral valve apparatus.
Although the frequency of device embolization is low, it has remained a stubbornly consistent event in most clinical trials and registries of LAA occlusive devices. A meta-analysis by Bajaj et al. reported an overall embolization rate of 3.9 % [12], but more contemporary trials and registries with second generation devices report a rate between 0 and 2 % [4–7]. Clinical presentation with device embolization may either occur acutely in close temporal relationship to the index procedure or be delayed several months. Variation in the recognition of this complication beyond the time of the index procedure is likely a function of the ultimate site to which the device embolizes. Embolization to the thoracic or abdominal aorta may be clinically silent, whereas embolization to the left ventricle with the subsequent risk of LVOT obstruction or disturbance of the mitral valve apparatus is more likely to present acutely with symptoms.
LAA device embolization is a highly morbid complication (Table 18.1) [1, 13–15]. Percutaneous retrieval of the device is technically very challenging due to the presence of barbs on the devices that are used to secure the device in the LAA. These can cause the device to become embedded in adjacent tissue in the heart or aorta. In addition, they complicate percutaneous retrieval in that the device has to be optimally snared to collapse the barbs before removal into a sheath, reducing the risk of injury to adjacent structures and the vascular access site. Collapsing the barbs into the retrieval sheath appears to be more difficult with the WATCHMAN device compared to the ACP device. As a result, surgical removal has been required for many of the cases reported in the literature, particularly when the site of embolization is the left ventricle and a WATCHMAN device has been used.
Table 18.1
Sample of studies reporting device embolization following LAA closure
Study | Device | Timing | Embolization | Outcome |
|---|---|---|---|---|
Aminian et al. (Aminian, 2014 #2) | ACP | 1 month | LVOT/mitral apparatus | Death |
Case report | ||||
Reddy et al. (Reddy, 2011 #10) | WATCHMAN | Procedure | LVOT | Perc removala |
PROTECT–AF Trial | 45-day follow-up | Thoracic aorta | Perc removal | |
CAP registry | 45-day follow-up | Abdominal aorta | Surgical removal | |
Gupta P et al. (Gupta, 2013 #1) | ACP | 1 day | LVOT | Surgical retrieval |
Case report | ||||
Pisani P et al. (Pisani, 2014 #3) | ACP | 6 months | LVOT/mitral apparatus | Surgical retrieval |
Case report |
Despite the challenges of percutaneous retrieval, a number of successful retrievals have been performed, including two cases by one of the authors (JS). The ACP device appears easier to retrieve percutaneously compared to the WATCHMAN device, likely related to the fact that the device can be fully collapsed and completely withdrawn into a large arterial sheath, regardless of which endscrew of the device is grasped. We have been successful in retrieving embolized ACP devices from the left atrium and the left ventricle, both of which were observed to have embolized on transthoracic echocardiogram the day after procedure. In the case where an embolized 24 mm ACP device was lodged in the left atrium (Fig. 18.5), a large 16F Checkflo 80 cm sheath was advanced into the left atrium across the transseptal puncture from the procedure the day before. A 5F EBU guide was telescoped through the sheath and to help direct a 25 mm gooseneck snare onto the embolized device. The snare was cinched on the distal endscrew of the device, which was completely pulled into the 16F sheath before removing the sheath from the left atrium and body.
Fig. 18.5
Percutaneous retrieval of embolized 24 mm ACP device lodged in the left atrium. (a) Transthoracic echocardiogram images showing the embolized device (arrows). (b) Flouroscopic image showing a 25 mm gooseneck snare, which had been delivered through a 5F EBU guide within a 16F 80 cm Checkflo sheath, grabbing the distal endscrew of the ACP device. (c) The device was pulled in against the 16F sheath in a coaxial manner. (d) The device was then pulled in firmly, with the lobe partially retrieved. (e) The lobe was entirely retrieved. (f) The disk was entirely retrieved
In the case where an embolized 28 mm ACP device was lodged in the left ventricular outflow tract (Fig. 18.6), we advanced a 14F 80 cm sheath into the ascending aorta from common femoral arterial access. A 5F JR4 guide was delivered into the left ventricle over a wire, and a 10 mm gooseneck snare was placed through the JR4 guide to direct the snare to the distal endscrew. Once captured, the device was pulled across the aortic valve in a vertical position (to avoid the hooks from disrupting the valve leaflets and adjacent tissue, as the hooks point from the lobe towards the disk). A transvenous pacer for temporary pacing of the patient was required for this case as there was transient heart block during the short time that the device was pulled across the aortic valve. The device was partially recaptured into the 14F sheath with the snare, but subsequently came off the snare. A bioptome was then used to grab the distal endscrew within the arterial sheath, and completely withdraw the device into the sheath, before removing the arterial sheath completely. In both cases, the patients tolerated the procedures well without any complications.
