Author + information
- Received November 27, 2007
- Revision received February 26, 2008
- Accepted March 11, 2008
- Published online September 9, 2008.
- Anne S. Kanderian, MD⁎,
- A. Marc Gillinov, MD†,
- Gosta B. Pettersson, MD, PhD†,
- Eugene Blackstone, MD† and
- Allan L. Klein, MD, FACC⁎,⁎ ()
- ↵⁎Reprint requests and correspondence:
Dr. Allan L. Klein, Department of Cardiovascular Medicine, Cleveland Clinic, 9500 Euclid Avenue, Desk F15, Cleveland, Ohio 44195
Objectives We sought to determine which surgical technique of left atrial appendage (LAA) closure is most successful by assessing them with transesophageal echocardiography (TEE).
Background Atrial fibrillation is a risk factor for stroke, with 90% of clots occurring in the LAA. Several surgical techniques of LAA closure are used to theoretically reduce the stroke risk, with varying success rates.
Methods A total of 137 of 2,546 patients who underwent surgical LAA closure from 1993 to 2004 had a TEE after surgery. Techniques consisted of either excision or exclusion by sutures or stapling. The TEE measurements included color Doppler flow in the LAA and interrogation for thrombus. Patent LAA, remnant LAA (residual stump >1 cm), or excluded LAA with persistent flow into the LAA were identified as unsuccessful closure.
Results Of the 137 patients, 52 (38%) underwent excision and 85 (62%) underwent exclusion (73 suture and 12 stapler). Only 55 of 137 (40%) of closures were successful. Successful LAA closure occurred more often with excision (73%) than suture exclusion (23%) and stapler exclusion (0%) (p < 0.001). We found LAA thrombus to be present in 28 of 68 patients (41%) with unsuccessful LAA exclusion versus none with excision. At time of TEE, 6 patients with successful LAA closure (11%) and 12 with unsuccessful closure (15%) had evidence of stroke/transient ischemic attack (p = 0.61).
Conclusions There is a high occurrence of unsuccessful surgical LAA closure. Of the various techniques, excision appears to be the most successful.
- atrial fibrillation
- left atrial appendage
- surgical left atrial appendage closure
- left atrial appendage thrombus
Atrial fibrillation (AF) is an arrhythmia of epidemic proportion and a potent source of cardioembolic events. The annual risk of stroke in patients with AF is 5% and increases with concomitant risk factors such as hypertension, left ventricular dysfunction, age, and valvular disease (1,2). Several studies have confirmed that the source of intracardiac thromboembolism in patients with AF is the left atrium and, more specifically, the left atrial appendage (LAA). In fact, in nonrheumatic AF, up to 90% of clots in the left atrium originate in the LAA (3,4). Randomized trials have established that warfarin is effective in reducing the stroke rate in patients with AF (1). However, the use of anticoagulation medications is not without limitations, adverse events, and contraindications (5,6).
It has been proposed by some investigators that closure of the LAA will decrease the stroke risk in patients with AF (3,7). Surgical closure of the LAA has been practiced since the 1930s, primarily in patients with mitral valve disease (3). In fact, current guidelines suggest obliteration of the LAA during mitral valve surgery (8). The surgical Maze procedure for AF originally advocated by Cox also incorporates excision of the LAA (9). Recently, the LAAOS (Left Atrial Appendage Occlusion) study demonstrated that LAA occlusion by suture or stapling at the time of coronary artery bypass grafting is safe and can be performed without lengthening the time of surgery or increasing the rate of post-operative bleeding (10). Currently, there are few centers and surgeons that routinely close the LAA during cardiac surgery. This reluctance on the part of surgeons may relate to lack of standardized surgical techniques and limited data concerning the effectiveness of the variety of techniques currently used.
There are several surgical techniques used to close the LAA, and they consist of either excising or excluding the appendage. Excision is performed by removal of the LAA, either by scissors or an amputating stapling device. Exclusion of the LAA is performed by closing the orifice into the LAA cavity with the appendage remaining attached. This technique is performed by various methods of suturing (running suture, pursestring or external ligation) or by stapling.
Although these surgical techniques are simple to apply, there is uncertainty regarding their reproducibility and effectiveness (11–15). Katz et al. (11) reported that surgical LAA ligation frequently is incomplete. Of 50 patients who underwent mitral valve surgery and LAA ligation by running suture technique, 18 (36%) had incomplete ligation detected by transesophageal echocardiography (TEE). The incomplete ligation was characterized as the presence of persistent color flow Doppler between the LAA and the left atrium. Another study by Garcia-Fernandez et al. (7) showed that 10.3% of patients who underwent LAA ligation by double suture technique during mitral valve replacement had incomplete ligation.
In recent years, TEE has become the standard tool for assessment of the LAA. Because there are several techniques used to surgically excise or exclude the LAA, our objective was to use TEE as a means to determine which current surgical technique is most successful at closing the LAA. Our hypothesis was that surgical LAA excision is superior to exclusion.
At the Cleveland Clinic, 2,546 patients underwent surgical LAA closure from 1993 to 2004. Many of those patients had no continued follow-up. As a result, from our TEE database, we identified 137 post-operative patients from the cohort who had a complete TEE with color Doppler interrogation of the LAA. The mean time to TEE was 8.1 ± 12 months. Nine experienced cardiovascular surgeons performed the surgeries. Techniques used to close the LAA were: 1) excision (by scissors or an amputating stapling device); and 2) exclusion (by suture or stapler). Indications for post-operative TEE consisted of pre-cardioversion for AF (n = 63), endocarditis (n = 31), mitral valve assessment (n = 17), AF ablation (n = 5), stroke or transient ischemic attack (TIA) (n = 4), tricuspid valve assessment (n = 4), aortic valve assessment (n = 2), left ventricular thrombus (n = 2), pericardial disease (n = 2), embolic events (n = 2), atrial septal defect (n = 1), aortic fistula (n = 1), aortic dissection (n = 1), heart failure (n = 1), and right atrial mass (n = 1).
A standard TEE was performed in all patients, and the LAA was evaluated in multiple views (16). Color Doppler was applied across the LAA to assess the presence of flow between the left atrium and the closed LAA. The presence of thrombus in the left atrium or the LAA was also documented. The LAA was classified as: 1) successful closure; 2) patent LAA; 3) excluded LAA with persistent flow into the appendage; or 4) remnant LAA. Patent LAA was defined as a persistent communication of the LAA with the left atrium due to dehiscence of suture or staple (Fig. 1). Excluded LAA with persistent flow into the appendage was defined as the presence of a color flow jet between the left atrium and LAA despite the 2-dimensional appearance of an obliterated LAA (Fig. 2). Remnant LAA was defined as a residual stump or pouch remaining in the LAA >1 cm in maximum length after closure (Fig. 3). Unsuccessful LAA closure was characterized as the presence of a patent LAA, excluded LAA with persistent flow into the appendage, or remnant LAA. Successful closure was defined as the absence of all the aforementioned findings. All the TEEs were reanalyzed by the investigators, with special emphasis placed on evaluating the LAA. The intrareader and interreader variability of classifying successful LAA closure was 98% and 97%, respectively.
Continuous data are expressed as mean ± standard deviation and were compared with the use of the 2-tailed Student t test. Categorical variables were compared with the Fisher exact test. A p value <0.05 was considered statistically significant. The SPSS 9 statistical software package (SPSS Inc., Chicago, Illinois) was used.
A total of 137 patients were included in the study. The mean age was 65 ± 12 years. Fifty-two patients (38%) underwent excision (41 by scissors and 11 by an amputating stapling device), and 85 (62%) underwent exclusion, of which 73 of these (86%) were by suture exclusion and 12 (14%) by stapler exclusion with the LAA remaining attached. Table 1 depicts baseline characteristic of patients undergoing surgical LAA elimination during cardiac surgery.
Closure success of LAA
The key finding of the study was that only 55 of 137 patients (40%) had successful LAA closure. Successful LAA closure occurred more often with excision of the LAA (73%) versus suture exclusion (23%) and stapler exclusion (0%), p < 0.001 (Table 2).
Among the patients with LAA excision (n = 52), a remnant LAA (residual stump >1 cm) was present in 14 (27%). Of patients who had suture exclusion (n = 73), 6 (8%) had a patent LAA, 6 (8%) had a remnant LAA, and 44 (61%) had an excluded LAA with persistent flow into the appendage. Of patients who had stapler exclusion (n = 12) of the LAA, 2 (17%) had a patent LAA, 7 (58%) had a remnant LAA, and 3 (25%) had an excluded LAA with persistent flow into the appendage. Of note, none of the attempts to perform stapler exclusion of the LAA were successful.
Clinical and echocardiographic variables were analyzed to assess whether they were predictive of successful surgical LAA closure (Table 3). As expected, LAA excision was predictive of successful procedural outcome (p < 0.001). Excluding the LAA by either suture or stapler techniques was more likely to predict unsuccessful LAA closure (p < 0.001 and p = 0.002, respectively). We found that the Maze procedure was also associated with successful LAA closure, likely due to the majority of patients having concomitant LAA excision. Previous investigators have hypothesized that increased left atrial size and/or area were likely to predict unsuccessful LAA closure; however, these variables were not predictive of unsuccessful closure in our study, as previously demonstrated by Garcia-Fernandez et al. (7).
No patients who had LAA excision and a residual stump had evidence of LAA thrombus within the remnant LAA. However, 28 patients who had unsuccessful LAA exclusion had LAA thrombus detected by TEE (Fig. 4). Of patients who had suture exclusion, LAA thrombus was present in 2 of 6 patients with patent LAA, 1 of 2 patients with remnant LAA and persistent flow into the appendage, and 20 of 44 patients with excluded LAA and persistent flow into the appendage (46%). Of patients who had stapler exclusion, LAA thrombus was present in 1 of 2 patients with patent LAA, 2 of 2 patients with remnant LAA and persistent flow into the appendage, and 2 of 3 patients with excluded LAA and persistent flow into the appendage.
At the time of TEE, patients were assessed for history of stroke or TIA after their original surgery. There were a total of 18 patients (13%) who experienced stroke/TIA (6 with LAA excision, 11 with suture exclusion, and 1 with stapler exclusion, p = NS). Of the 55 patients with successful LAA closure, 6 (11%) had stroke/TIA versus 12 of the 82 patients (15%) with unsuccessful LAA closure, p = 0.61. There were 3 additional patients who had evidence of peripheral embolic events. Of patients who had unsuccessful LAA closure, 4 (30%) had evidence of LAA thrombus.
Currently, there is tremendous interest in closure of the LAA by the use of surgical or percutaneous techniques. The results of our study show that, with current surgical techniques, LAA management is unsuccessful in nearly 60% of patients. Of the various techniques, excision of the LAA is most effective (success rate of 73%); however, there is a likelihood of leaving a residual stump. Although we found no thrombus present in residual stumps, this has been classified in the literature as unsuccessful closure and, theoretically, residual LAA tissue could still pose a risk for harboring thrombus. A high percentage of patients with suture exclusion of the LAA had persistent flow into the appendage, as documented by color Doppler from the LA and the LAA (60%), and a high percentage of those with stapler exclusion had a persistent LAA stump >1 cm (58%). We report a greater rate of unsuccessful LAA closure than what has been previously reported in the literature. Our study is different in that a variety of different surgical techniques were used, including various suture techniques, to close off the LAA. Additionally, our population studied was more than double the size than what was reported in previous studies.
Persistent flow into the LAA after exclusion is indicative of persistent communication and, theoretically, thrombi can traverse this communication and embolize. This development is quite concerning, particularly because the LAA is more apt to thrombose when it is partially closed, as blood is more stagnant. The prevalence of LAA thrombus in appendages with persistent flow was high (46% in suture exclusion and 67% in stapler exclusion). However, it remains to be determined whether appendages with residual flow or a residual stump are associated with increased risk of emboli. Nevertheless, it would seem unwise to discontinue anticoagulation in a patient with LAA thrombus and a persistent communication with the left atrium as demonstrated by persistent flow into the appendage.
Currently, there are devices designed to percutaneously occlude the LAA: the PLAATO (Percutaneous Left Atrial Appendage Transcatheter Occlusion) device (Appriva Medical Inc., Sunnyvale, California) (17), the WATCHMAN left atrial appendage system (Atritech, Inc., Plymouth, Minnesota) (18), and the Amplatzer septal occluder device (AGA Medical Corp., Plymouth, Minnesota) (19). Preliminary studies have shown that deploying these devices is feasible, and the short-term follow-up of the PLAATO device seems to be promising, with a high successful LAA occlusion rate up to 6 months. However, further long-term studies are necessary to determine continued efficacy and safety.
This study did demonstrate a trend towards decreased incidence of stroke/TIA in patients with successful LAA closure; however, this was not statistically significant, probably because the sample size was small. Although in theory, closing the LAA may translate into a decreased stroke rate in patient with AF, there are still concerns and controversies. For instance, concern exists for increased post-operative bleeding when excision of the LAA is performed. There has also been some apprehension regarding deterioration of hemodynamics with LAA elimination (20–22). Studies have shown that the LAA plays a role in regulating volume status by several physiologic functions, such as mediating thirst, modulating the relationship between pressure and volume, improving left atrial compliance, improving cardiac output, and releasing atrial natriuretic peptide (22). Few studies conducted in patients who underwent the Maze procedure along with bilateral appendage removal demonstrated attenuated secretion of atrial natriuretic peptide and water retention (23,24). The risks and benefits of LAA closure in different populations of patients have yet to be determined.
This study has certain limitations. Because it was a retrospective study on patients who had a TEE after surgical LAA closure, there may be a selection bias, and patients having a TEE may not be representative of the entire population. We included all potential definitions of unsuccessful closure including remnant LAA, which has not always been used in other studies. Additionally, surgeons use different techniques in closing the LAA, and this nonrandomized study does not account for inherent bias. Stroke and TIA outcomes obtained in this study were only assessed until the time of TEE.
From this study, we conclude that when surgical LAA closure is performed, excision of the appendage is the most reliable method. Our study raises the concern of discontinuing anticoagulation in patients with AF who have had surgical LAA closure due to the high rate of unsuccessful closure. If anticoagulation medication is to be discontinued, consideration should be given to performing a TEE to ensure successful LAA closure. Further studies are indicated to determine whether patients who undergo LAA closure demonstrate a reduction in thromboembolic events.
Dr. Gillinov is a consultant for AtriCure, Medtronic, and Edwards and is a speaker for Boston Scientific and St. Jude.
This work was funded by the State of Ohio Third Frontier Project.
- Abbreviations and Acronyms
- atrial fibrillation
- atrial natriuretic peptide
- left atrial appendage
- transesophageal echocardiography
- transient ischemic attack
- Received November 27, 2007.
- Revision received February 26, 2008.
- Accepted March 11, 2008.
- American College of Cardiology Foundation
- Manning W.J.,
- Silverman D.I.,
- Katz S.E.,
- et al.
- Brass L.M.,
- Krumholz H.M.,
- Scinto J.M.,
- Radford M.
- Wehinger C.,
- Stollberger C.,
- Langer T.,
- Schneider B.,
- Finsterer J.
- Garcia-Fernandez M.A.,
- Perez-David E.,
- Quiles J.,
- et al.
- Bonow R.O.,
- Carabello B.A.,
- Kanu C.,
- et al.
- Katz E.S.,
- Tsiamtsiouris T.,
- Applebaum R.M.,
- Schwartzbard A.,
- Tunick P.A.,
- Kronzon I.
- Ostermayer S.H.,
- Reisman M.,
- Kramer P.H.,
- et al.
- Al-Saady N.M.,
- Obel O.A.,
- Camm A.J.
- Benjamin B.A.,
- Metzler C.H.,
- Peterson T.V.