Despite 25-plus years of surgical attempts to ligate or occlude and remove the left atrial appendage (LAA) at the time of open-chest cardiac procedures in patients with atrial fibrillation (AF), the development of percutaneous techniques to provide occlusion or ligation of the LAA is a very recent phenomenon. Long-term observational data suggest reduced embolic stroke rate in patients who have had effective closure of the LAA, and more recent evidence points to an increased stroke risk if closure is not 100% complete. With this in mind, and supported by AF stroke reduction data provided by clinical trials with the WATCHMAN LAA closure device (Boston Scientific), there was interest in my practice to seek a safe and effective alternative for long-term anticoagulation (OAC), specifically in patients who had previously undergone catheter ablation for AF.
In August 2013, we began performing suture ligation with the subxiphoid LARIAT Suture Delivery Device (SentreHEART, Inc.) in patients who were high risk or absolutely contraindicated to OAC, despite knowing this was off-label use of the device for ligation of the LAA. Our initial results were quite supportive of the technology. Three years and 50 LARIAT cases later, we are now enrolling patients in the multicenter aMAZE trial, which is evaluating LAA ligation as a tool to reduce recurrence rates after catheter ablation for persistent atrial fibrillation.
Upon FDA approval of the WATCHMAN device in May 2015, our EP program quickly aligned partnership with cardiac anesthesiology, TEE echocardiography, neurology, and interventional cardiology to build support for a hospital-wide LAA closure program. After Vanderbilt Heart’s agreement to start the program, our first group of patients targeted for therapy were those I had followed in clinic for many years post ablation, who were struggling to tolerate long-term anticoagulation, or were deemed high risk for adverse bleeding. Following this, we began to obtain referrals for patients with stroke and bleeding, end-stage renal disease patients with adverse bleeding, patients with high fall risk unable to tolerate OAC, and even those unable to adhere to long-term systemic OAC. Presenting at neurology grand rounds, the Vanderbilt Neurology Stroke Conference, a renal transplant conference, and the warfarin clinic has helped garner more widespread support for this therapy. We have now performed over 90 implants, and I have worked as a WATCHMAN proctor starting up new sites throughout the Southeast and Central United States.
An example case of WATCHMAN device implantation at the Vanderbilt Heart and Vascular Institute involves a 64-year-old African-American male who presented with a history of coronary artery disease, a previous stent in the left circumflex artery, hypertension with left ventricular hypertrophy, diastolic heart failure, end-stage renal disease on hemodialysis followed by renal transplantation, and a history of persistent AF. He had an ablation in 2013, followed by repeat ablation for atypical flutter in 2015. His CHA2DS2-VASc score was 4, HAS-BLED score was 3, and he had recurrent GI hemorrhage on rivaroxaban. Although he previously tolerated warfarin, he was referred for WATCHMAN device implantation after recurrent melena. His pre-op TEE showed an ostial dimension of 2.1 x 2.2 cm and a LAA depth maximum of 4.0 cm. Angiograms from his implantation are shown in Figures 1-4. Typically, the LAA ostium is more eccentric or elliptical, and when viewed from 3D TEE, I describe that it appears similar to the entry into a snail’s shell. Initial deployment of a 30-mm device had a slight shoulder and atypical orientation of the feet of the device. We recaptured and redeployed the device with perfect seal. Aggressive tug testing to ensure stability was performed with good fixation. He has since come off anticoagulation after 45 days, and is doing well in follow-up.
Building a successful LAA closure program certainly requires some forethought. All subjects have had a CHA2DS2-VASc score of 3 or higher, and generally some adverse bleeding event or factor increasing their HAS-BLED score to 3 or higher. We have obtained a signed letter from all referring noninterventional cardiologists or other specialty physicians, and precertified patients through insurance before performing the WATCHMAN implant, with a recent single exception of an inpatient with cirrhosis and history of intracranial bleeding in persistent flutter who would not be listed for liver transplantation until her AF and stroke risk were managed. We have also enrolled in the NCDR-LAAO Registry, which has affected how our TEE reports are structured, based on data elements required for the registry. We have generally performed all cases on therapeutic warfarin, with an INR goal of 2-2.5 on the day of the procedure, followed by the obligatory 45 days on aspirin 81 mg plus warfarin. Intraoperative TEE guidance by cardiac anesthesiology has provided great support for safe and effective delivery of the therapy; all subjects have obtained the 45-day follow-up TEE. We have seen no thrombus on the device with this approach, and have transitioned our patients to aspirin and clopidogrel for 6 months post implant. On rare occasions, we have had to deviate from this due to clinically adverse bleeding; however, no patient has had an adverse embolic event after a successful WATCHMAN implant. To date, I have not used NOACs as part of the short-term anticoagulation bridge.
The implantation technique also certainly relies on safe transseptal puncture skills. In my experience proctoring, this has been the #1 issue that new operators have struggled with at times. Having multiple secondary approaches to transseptal access is critical when presented with lipomatous hypertrophy, previous ASD or PFO closure, thick septum from multiple prior transseptal procedures, or difficult orientation of the fossa. This may include use of a spiral wire, Inoue wire (Toray), or Radiofrequency NRG® Transseptal Needle (Baylis Medical) in the case, tailoring the transseptal needle bend and sheath selection to obtain proper inferior and posterior fossa orientation for transseptal. In the majority of cases, I use an SL-0 sheath (St. Jude Medical) with a standard curve NRG needle, but I contour the needle to obtain a 90° bend. If transseptal access is not in the ideal location, I will re-stick; the new operator should have a very low threshold to do this, as forcing an implant from an anterior or superior stick is challenging and potentially dangerous. Post procedure, we transition to a 14 French short sheath, or pull and hold with a figure of 8 stitch in the EP lab and reverse heparin. All patients get a post-op PA/lateral chest x-ray prior to discharge on the morning after implant to check the device position. Interestingly, it can be quite challenging to find, even for a radiologist at times. When there is a question, we perform a quick fluoroscopy in obese patients prior to discharge.
Tracking the patients in the Vanderbilt Left Atrial Appendage Closure Registry (VaLAAR) has been helpful to allow surveillance of follow-up assessments and ensure all subjects get a one-year follow-up TEE. Patient satisfaction at 45 days, when systemic anticoagulation can be discontinued, has been very high. My personal satisfaction with the procedure has also been extremely high. For straightforward anatomy windsock LAA with good depth, I generally anticipate a 20-minute implant time or less, starting greater than 90% of cases with a double curve WATCHMAN Access Sheath (WAS). Vascular access can always be a surprising challenge; insertion of a 16 Fr valved sheath can be helpful to stabilize the torqueability of the WATCHMAN device in patients who are obese or have tortuous venous anatomy. Recapturing and redeploying the device is a critical skill that becomes straightforward after approximately 10-20 implants. Closure of short broccoli, chicken wing LAA, or LAA with less than 17 mm ostial diameter is certainly possible, after gaining an appreciation for how the device can be contoured into the LAA during deployment. Generally, the larger the device, the more compliant and forgiving it is. We have generally favored oversizing the device when adequate depth is available, and recommend a 6 Fr angled pigtail catheter for implantation with hand injection LAA angiogram before choosing a device size. Complete lobes of the LAA are frequently missed on TEE and can provide an additional pathway to begin device deployment, allowing use of a larger device than considered possible by TEE screening.
The main challenge associated thus far with our WATCHMAN LAA closure program has been the insurance company denials stating the device is “experimental.” Pre-certification is denied because it is lumped into a code including PFO closure for AF (which is always denied), or because when screening a patient for the device, we find an ostial dimension of only <14 mm or >31 mm. We have recently begun implanting the Amulet device (St. Jude Medical, now Abbott Medical), which provides options for these anatomic restrictions to WATCHMAN. Lastly, LAA thrombus on the day of the implant can be a challenge, and often requires higher intensity OAC, additional antiplatelet therapy, or transition to dabigatran or another NOAC, with rescheduling after excluding an active LAA clot.
Disclosure: Dr. Ellis has no conflicts of interest to report regarding the content herein. Outside the submitted work, he reports being a consultant for Boston Scientific, AtriCure, SentreHEART, and Medtronic, and research grants with Boston Scientific, AtriCure, and Medtronic.
Videos also available at: http://www.eplabdigest.com/video