There remains no doubt that atrial fibrillation (AF) is associated with an increased stroke risk (1). It is also clear that the vast majority of thrombi in patients with non-valvar AF are located in the left atrial appendage (LAA) (2). Multiple randomized trials have demonstrated a mortality and stroke rate reduction with anticoagulation using a vitamin K antagonist (3). More recently, the superiority of LAA closure using a nitinol cage (Watchman device) over anticoagulation with a vitamin K antagonist has been shown regarding all-cause mortality (driven by a lower rate of intracranial hemorrhage), disabling strokes and long-term bleeding (disregarding the up-front rate of pericardial bleeding) (4). Therefore, in most countries, LAA closure has established itself as the treatment of choice in patients with a high or prohibitive bleeding risk, despite the fact that its utility has been shown in a different patient population, those with a low bleeding risk who are able to tolerate anticoagulation. Optimally, a LAA closure device would seal the LAA completely leaving no potentially thrombogenic LAA tissue behind and not cause thrombus formation both of which may diminish and, perhaps, off-set the (stroke prevention) benefits of the procedure. The reality is that neither applies to the current technology. In fact, regardless of the device or trial, device-associated thrombus formation (DTF) has been reported. It has been less clear if and to what degree the discovery of this finding increases the stroke risk. Previously, based on the observed device-associated thrombus rate of 20/478 patients (4.2%) in PROTECT-AF of whom 3 had a stroke prior to detection, the device-thrombus associated stroke risk has been estimated to be 0.3% per 100 patient years. In other words, it has been assumed that the risk of stroke caused by DTF per year is only 0.3% (5). From a different perspective, however, one could assume that the risk of a stroke in the presence of DTF is at least 3/20 (=15%) not taking into account that some of the other reported strokes at follow-up may have been related to unknown DTF (or thrombi no longer seen on the device as they have embolized). Hence, the DTF rate has not received as much attention as it, perhaps, should. Therefore, the findings of the recently published manuscript by Fauchier et al. in the Journal of the American College of Cardiology are very important (6). Data from 469 consecutive patients who underwent LAA closure, in the overwhelming majority with the Watchman device, Amplatzer Cardiac Plug or Amulet, were retrospectively analyzed with focus on the incidence and consequence of DTF.
First, data confirm previous findings of DTF with an incidence for the Watchman device (5.5%) similar to that previously reported in an analysis of several publications (2–6%) (7) and for the Amplatzer Cardiac Plug/Amulet device (11%) also similar to that reported in prior analyses (2–18%) (7). Of note, this numerical difference between the devices did not reach statistical significance. It should be mentioned that the accuracy of imaging for thrombus detection has not been validated; distinction of thrombus versus pronounced endothelialization or tissue proliferation is not always clear (8) and the true incidence of DTF may, hence, be over- or underestimated.
Second, and more importantly, the authors also confirm our suspicion that DTF is not a finding to be complacent about because the stroke/systemic embolism rate is higher than if no thrombus is present. Of the 26 patients who were found to have DTF, 4 (15%) had a stroke compared to 10 of 313 (3%) who did not have DTF (adjusted hazard ratio: 4.4, P=0.04, on multivariable analysis DTF was an independent risk factor for stroke apart from age and prior history of stroke). Incidentally, the stroke rate (15%) in those patients with DTF was no different from the aforementioned 15% observed in patients with DTF in the PROTECT-AF trial. Though the statistical power to prove that DTF may cause cerebral or systemic embolic events and that the risk of these events is higher in patients who are found to have DTF was limited in this study, it supports our concept of pathophysiology, that intravascular or intra-cardiac thrombi can cause embolization and are, therefore, potentially dangerous. Nevertheless, the findings should not be misused to prematurely discredit an effective procedure but should be viewed in the following context: the alternative to LAA occlusion, anticoagulation, also does not eliminate the risk of LAA thrombus. In patients on therapeutic anticoagulation for AF while undergoing transesophageal echocardiography for planned ablation, the reported rate of LAA thrombi remains 0.3–3.6% (9-16). This risk will not diminish over time during anticoagulation, whereas the risk of DTF is likely to decrease after device endothelialization. Moreover, the prevalence of LAA thrombus is much higher in patients with AF who are not or inadequately anticoagulated. For example, the prevalence of LAA thrombus was 12% in 600 patients with AF of <48 hours duration whose anticoagulation was subtherapeutic and 15% in a study combining transesophageal echocardiography and autopsy data (17). Most importantly, despite the potential for DTF, the event rate after LAA closure has been very low in both randomized trials and a number of large registries, thus supporting its efficacy (4,18-20).
How can we avoid DTF? To answer this question, a better understanding of what causes or promotes it is needed. In the meantime, intuitively, one would suspect the usual suspects, a thrombogenic endovascular surface, stasis and blood thrombogenicity. A thrombogenic endovascular surface is likely to remain at least until, and maybe beyond, complete endothelialization. How long does complete endothelialization take? In an aninmal (dog) model, this takes approximately one month (21). However, the speed of endothelialization in humans is not known and is difficult to study on a larger scale. In one study of four humans who underwent an autopsy, on average, 417 days after the procedure, the device appeared to have been covered with endocardium (21). Whether consistent endothelialization occurs within the first few months after implantation in humans is unknown. In this context, assessing endothelialization based on permeability of the device after implantation by CT scanning in patients who did not have a peri-device leak (assessed by transesophageal echocardiography), incomplete endothelialization was reported in 61% at a mean follow-up of 10 months (22) thereby suggesting that it may take longer than previously thought. Stasis may promote thrombus formation in AF. The observed higher stroke rates in patients with AF and mitral stenosis (23) and lower than expected rates in those with AF and mitral regurgitation (24) support this notion. Spontaneous echo contrast (SEC), an indicator of stasis, is an independent predictor for strokes in AF (25). Similarly, a reduction in LAA peak flow velocities, a surrogate for stasis, is an independent risk factor for stroke in patients with AF (26,27). Though the relationship between stasis and thrombus formation has only been shown in patients with AF who have not undergone LAA closure, it is likely that this would contribute to DTF. Thrombogenicity, i.e., the propensity of a patient’s blood to clot, is difficult to gauge. Parameters that may be associated with thrombus formation are frequently elevated in patients with AF. Fibrinogen (28), prothrombin fragments 1 and 2 (29), D-dimers (28), thrombi-antithrombin complexes (29), platelet microparticles (30), beta-thromboglobulin (28) and von Willebrand factor (31) are examples. In other words, AF may be associated with a “prothrombotic” state. Moreover, some parameters [e.g., D-dimer level, von Willebrand factor (vWF)] may predict the presence of LAA thrombus or clinical events regardless of treatment with anticoagulation (32-35). Furthermore, gene polymorphisms of the coagulation system [e.g., fibrinogen (36,37), factor XIII (38)] as well as of platelet function [e.g., integrin alpha 2 (39)], by virtue of increased prothrombotic agents (38), may cause a higher risk of thrombus formation (37) or clinical events (39) in patients with AF. Once again, whether this also applies to DTF is not clear. Some inflammatory mediators [e.g., C-reactive protein (CRP), soluble intercellular adhesion molecule-1 (sICAM-1), fibrinogen, interleukin 6 (IL-6), tumor necrosis factor (TNF)-alpha, CD-40 ligand] have been found to be elevated in patients with AF (40-44) and may promote LAA (and, perhaps, device-associated) thrombus formation. These aforementioned risk factors for LAA thrombus may, in the future, predict the propensity of a patient to develop DTF. However, even if it was possible to predict the likelihood of DTF (for example, by characterizing patients regarding stasis and thrombogenicity), what would be the therapeutic consequences? If patients classified at high risk of DTF should be treated differently, how should they be treated? One might imagine that anticoagulation with vitamin K antagonists or direct anticoagulants or heparin (e.g., enoxaparin) would be more likely to prevent thrombus formation than antiplatelet therapy. However, though not compared head-to-head in a randomized clinical trial, there are mixed results from available data with some studies suggesting no difference in the incidence of DTF regardless of treatment (double antiplatelet therapy, vitamin K antagonists or direct oral anticoagulants) (45,46). Not to mention that early treatment with anticoagulants may be associated with a higher risk of pericardial bleeding/effusion. Therefore, even if correct identification of patients at risk for DTF were possible, we currently do not know what the preventive/therapeutic consequence should be, other than that we do know that single antiplatelet therapy or no antiplatelet/anticoagulant therapy after device implantation, from the perspective of DTF, is not as safe as dual antiplatelet therapy or anticoagulation (the use of dual antiplatelet therapy or anticoagulation after the procedure had a protective effect regarding DTF in the study published by Fauchier et al.). Though it appears that DTF may not be benign, it remains to be determined if thrombus location, size and morphology matters. For example, it is conceivable that a small layered thrombus at the device LAA transition has a different embolic risk than a large mobile thrombus.
Third, the clinical event rate (death: 6.9%, ischemic stroke: 4.0%, major hemorrhage: 3.8%) at just over 1 year (mean follow-up 13 months) was higher than reported in the pivotal randomized trial [e.g., death and ischemic stroke per 100 patient-years was 3.0% and 2.2%, respectively, in the device group in PROTECT-AF (47)]. The reason for this is not clear. However, it is likely, at least in part, related to the patients’ baseline characteristics. For example, in PROTECT-AF, the mean age in the intervention group was 72 years, the incidence of prior transient ischemic attacks or strokes 18% and none of the patients were considered to have a contraindication to anticoagulation. This compares with a mean age of 75 years and incidence of prior ischemic strokes of 41% and a proportion of patients considered to have a contraindication to anticoagulation of 73% in Fauchier et al.’s study. Under real world circumstances, given the overall poorer health, a higher rate of adverse events is to be expected. This is also supported by data from other larger registries. For example, in EWOLUTION, at one-year follow-up in the just over 1,000 patients who underwent LAA closure included in this registry, the mortality was 9.8% (18). Hence, in real life, the expected bench mark regarding outcomes after LAA closure should be adjusted according to patients’ baseline health.
In conclusion, the authors are to be commended for systematically examining the prevalence and relevance of DTF. The data was acquired without support from industry. It suggests that this phenomenon remains common and should not be taken lightly. It would seem that surveillance imaging (e.g., echo or CT) would be prudent in the period when no or incomplete endothelialization is anticipated. However, when, how often and how long this surveillance is necessary is unclear. It further advocates the temporary use of double antiplatelet therapy or anticoagulation in those patients who are likely to tolerate it. Finally, it demonstrates that in a real-world population, a high adverse event rate including death is to be expected due to the patients’ baseline age and co-morbidities.
Where to go from here? To further understand the pathophysiology of DTF and its implications as well as preventive strategies we should focus on further study of echocardiographic, clinical and biochemical risk factors in a prospective fashion, examine which thrombi are associated with the highest risk and determine in a prospective, optimally randomized, trial what the best preventive (and therapeutic) strategies may be. Meanwhile, device manufacturers should continue to explore designs that reduce thrombogenicity.
Conflicts of Interest: Horst Sievert has either ownership interests, stock options, received travel reimbursement, study honoraria or consultant fees for 4tech Cardio, Abbott, Ablative Solutions, Ancora Heart, Bavaria Medizin Technologie GmbH, Bioventrix, Boston Scientific, Carag, Cardiac Dimensions, Celonova, Cibiem, CGuard, Comed B.V., Contego, CVRx, Edwards, Endologix, Hemoteq, InspireMD, Lifetech, Maquet Getinge Group, Medtronic, Mitralign, Nuomao Medtech, Occlutech, pfm Medical, Recor, Renal Guard, Rox Medical, Terumo, Vascular Dynamics, Vivasure Medical, Venus, Veryan. The other author has no conflicts of interest to declare.
- Wolf PA, Abbott RD, Kannel WB. Atrial fibrillation as an independent risk factor for stroke: the Framingham Study. Stroke 1991;22:983-8. [Crossref] [PubMed]
- Blackshear JL, Odell JA. Appendage obliteration to reduce stroke in cardiac surgical patients with atrial fibrillation. Ann Thorac Surg 1996;61:755-9. [Crossref] [PubMed]
- Hart RG, Benavente O, McBride R, et al. Antithrombotic therapy to prevent stroke in patients with atrial fibrillation: a meta-analysis. Ann Intern Med 1999;131:492-501. [Crossref] [PubMed]
- Reddy VY, Doshi SK, Kar S, et al. 5-Year Outcomes After Left Atrial Appendage Closure: From the PREVAIL and PROTECT AF Trials. J Am Coll Cardiol 2017;70:2964-75. [Crossref] [PubMed]
- Reddy VY, Holmes D, Doshi SK, et al. Safety of percutaneous left atrial appendage closure: results from the Watchman Left Atrial Appendage System for Embolic Protection in Patients with AF (PROTECT AF) clinical trial and the Continued Access Registry. Circulation 2011;123:417-24. [Crossref] [PubMed]
- Fauchier L, Cinaud A, Brigadeau F, et al. Reply: Device-Related Thrombus After Percutaneous Left Atrial Appendage Occlusion in Patients With Atrial Fibrillation. J Am Coll Cardiol 2018;72:474-5. [Crossref] [PubMed]
- Lempereur M, Aminian A, Freixa X, et al. Device-associated thrombus formation after left atrial appendage occlusion: A systematic review of events reported with the Watchman, the Amplatzer Cardiac Plug and the Amulet. Catheter Cardiovasc Interv 2017;90:E111-21. [Crossref] [PubMed]
- Salaun E, Deharo JC, Habib G, et al. Extensive Endothelialization or Thrombus Related to New-Generation Left Atrial Appendage Occluders. JACC Clin Electrophysiol 2017;3:787-8. [Crossref] [PubMed]
- Khan MN, Usmani A, Noor S, et al. Low incidence of left atrial or left atrial appendage thrombus in patients with paroxysmal atrial fibrillation and normal EF who present for pulmonary vein antrum isolation procedure. J Cardiovasc Electrophysiol 2008;19:356-8. [Crossref] [PubMed]
- Puwanant S, Varr BC, Shrestha K, et al. Role of the CHADS2 score in the evaluation of thromboembolic risk in patients with atrial fibrillation undergoing transesophageal echocardiography before pulmonary vein isolation. J Am Coll Cardiol 2009;54:2032-9. [Crossref] [PubMed]
- Scherr D, Dalal D, Chilukuri K, et al. Incidence and predictors of left atrial thrombus prior to catheter ablation of atrial fibrillation. J Cardiovasc Electrophysiol 2009;20:379-84. [Crossref] [PubMed]
- McCready JW, Nunn L, Lambiase PD, et al. Incidence of left atrial thrombus prior to atrial fibrillation ablation: is pre-procedural transoesophageal echocardiography mandatory? Europace 2010;12:927-32. [Crossref] [PubMed]
- Wallace TW, Atwater BD, Daubert JP, et al. Prevalence and clinical characteristics associated with left atrial appendage thrombus in fully anticoagulated patients undergoing catheter-directed atrial fibrillation ablation. J Cardiovasc Electrophysiol 2010;21:849-52. [PubMed]
- Frenkel D, D’Amato SA, Al-Kazaz M, et al. Prevalence of Left Atrial Thrombus Detection by Transesophageal Echocardiography: A Comparison of Continuous Non-Vitamin K Antagonist Oral Anticoagulant Versus Warfarin Therapy in Patients Undergoing Catheter Ablation for Atrial Fibrillation. JACC Clin Electrophysiol 2016;2:295-303. [Crossref] [PubMed]
- Balouch M, Gucuk Ipek E, Chrispin J, et al. Trends in Transesophageal Echocardiography Use, Findings, and Clinical Outcomes in the Era of Minimally Interrupted Anticoagulation for Atrial Fibrillation Ablation. JACC Clin Electrophysiol 2017;3:329-36. [Crossref] [PubMed]
- Alqarawi W, Birnie DH, Spence S, et al. Prevalence of left atrial appendage thrombus detected by transoesophageal echocardiography before catheter ablation of atrial fibrillation in patients anticoagulated with non-vitamin K antagonist oral anticoagulants. Europace 2019;21:48-53. [PubMed]
- Malik R, Alyeshmerni DM, Wang Z, et al. Prevalence and predictors of left atrial thrombus in patients with atrial fibrillation: is transesophageal echocardiography necessary before cardioversion? Cardiovasc Revasc Med 2015;16:12-4. [Crossref] [PubMed]
- Bergmann MW, Ince H, Kische S, et al. Real-world safety and efficacy of WATCHMAN LAA closure at one year in patients on dual antiplatelet therapy: results of the DAPT subgroup from the EWOLUTION all-comers study. EuroIntervention 2018;13:2003-11. [Crossref] [PubMed]
- Reddy VY, Gibson DN, Kar S, et al. Post-Approval U.S. Experience With Left Atrial Appendage Closure for Stroke Prevention in Atrial Fibrillation. J Am Coll Cardiol 2017;69:253-61. [Crossref] [PubMed]
- Tzikas A, Shakir S, Gafoor S, et al. Left atrial appendage occlusion for stroke prevention in atrial fibrillation: multicentre experience with the AMPLATZER Cardiac Plug. EuroIntervention 2016;11:1170-9. [Crossref] [PubMed]
- Schwartz RS, Holmes DR, Van Tassel RA, et al. Left atrial appendage obliteration: mechanisms of healing and intracardiac integration. JACC Cardiovasc Interv 2010;3:870-7. [Crossref] [PubMed]
- Granier M, Laugaudin G, Massin F, et al. Occurrence of Incomplete Endothelialization Causing Residual Permeability After Left Atrial Appendage Closure. J Invasive Cardiol 2018;30:245-50. [PubMed]
- Salem DN, Stein PD, Al-Ahmad A, et al. Antithrombotic therapy in valvular heart disease--native and prosthetic: the Seventh ACCP Conference on Antithrombotic and Thrombolytic Therapy. Chest 2004;126:457S-482S. [Crossref] [PubMed]
- Nakagami H, Yamamoto K, Ikeda U, et al. Mitral regurgitation reduces the risk of stroke in patients with nonrheumatic atrial fibrillation. Am Heart J 1998;136:528-32. [Crossref] [PubMed]
- Transesophageal echocardiographic correlates of thromboembolism in high-risk patients with nonvalvular atrial fibrillation. The Stroke Prevention in Atrial Fibrillation Investigators Committee on Echocardiography. Ann Intern Med 1998;128:639-47. [Crossref] [PubMed]
- Takada T, Yasaka M, Nagatsuka K, et al. Blood flow in the left atrial appendage and embolic stroke in nonvalvular atrial fibrillation. Eur Neurol 2001;46:148-52. [Crossref] [PubMed]
- Santiago D, Warshofsky M, Li Mandri G, et al. Left atrial appendage function and thrombus formation in atrial fibrillation-flutter: a transesophageal echocardiographic study. J Am Coll Cardiol 1994;24:159-64. [Crossref] [PubMed]
- Mondillo S, Sabatini L, Agricola E, et al. Correlation between left atrial size, prothrombotic state and markers of endothelial dysfunction in patients with lone chronic nonrheumatic atrial fibrillation. Int J Cardiol 2000;75:227-32. [Crossref] [PubMed]
- Turgut N, Akdemir O, Turgut B, et al. Hypercoagulopathy in stroke patients with nonvalvular atrial fibrillation: hematologic and cardiologic investigations. Clin Appl Thromb Hemost 2006;12:15-20. [Crossref] [PubMed]
- Choudhury A, Chung I, Blann AD, et al. Elevated platelet microparticle levels in nonvalvular atrial fibrillation: relationship to p-selectin and antithrombotic therapy. Chest 2007;131:809-15. [Crossref] [PubMed]
- Conway DS, Heeringa J, Van Der Kuip DA, et al. Atrial fibrillation and the prothrombotic state in the elderly: the Rotterdam Study. Stroke 2003;34:413-7. [Crossref] [PubMed]
- Nozawa T, Inoue H, Hirai T, et al. D-dimer level influences thromboembolic events in patients with atrial fibrillation. Int J Cardiol 2006;109:59-65. [Crossref] [PubMed]
- Vene N, Mavri A, Kosmelj K, et al. High D-dimer levels predict cardiovascular events in patients with chronic atrial fibrillation during oral anticoagulant therapy. Thromb Haemost 2003;90:1163-72. [Crossref] [PubMed]
- Habara S, Dote K, Kato M, et al. Prediction of left atrial appendage thrombi in non-valvular atrial fibrillation. Eur Heart J 2007;28:2217-22. [Crossref] [PubMed]
- Ancedy Y, Berthelot E, Lang S, et al. Is von Willebrand factor associated with stroke and death at mid-term in patients with non-valvular atrial fibrillation? Arch Cardiovasc Dis 2018;111:357-69. [Crossref] [PubMed]
- Carter AM, Catto AJ, Grant PJ. Association of the alpha-fibrinogen Thr312Ala polymorphism with poststroke mortality in subjects with atrial fibrillation. Circulation 1999;99:2423-6. [Crossref] [PubMed]
- Bozdemir V, Kirimli O, Akdeniz B, et al. The association of beta-fibrinogen 455 G/A gene polymorphism with left atrial thrombus and severe spontaneous echo contrast in atrial fibrillation. Anadolu Kardiyol Derg 2010;10:209-15. [Crossref] [PubMed]
- Marín F, Corral J, Roldan V, et al. Factor XIII Val34Leu polymorphism modulates the prothrombotic and inflammatory state associated with atrial fibrillation. J Mol Cell Cardiol 2004;37:699-704. [Crossref] [PubMed]
- Roldán V, Marin F, Gonzalez-Conejero R, et al. Factor VII -323 decanucleotide D/I polymorphism in atrial fibrillation: implications for the prothrombotic state and stroke risk. Ann Med 2008;40:553-9. [Crossref] [PubMed]
- Marcus GM, Smith LM, Glidden DV, et al. Markers of inflammation before and after curative ablation of atrial flutter. Heart Rhythm 2008;5:215-21. [Crossref] [PubMed]
- Ederhy S, Di Angelantonio E, Dufaitre G, et al. C-reactive protein and transesophageal echocardiographic markers of thromboembolism in patients with atrial fibrillation. Int J Cardiol 2012;159:40-6. [Crossref] [PubMed]
- Qu YC, Du YM, Wu SL, et al. Activated nuclear factor-kappaB and increased tumor necrosis factor-alpha in atrial tissue of atrial fibrillation. Scand Cardiovasc J 2009;43:292-7. [Crossref] [PubMed]
- Pinto A, Tuttolomondo A, Casuccio A, et al. Immuno-inflammatory predictors of stroke at follow-up in patients with chronic non-valvular atrial fibrillation (NVAF). Clin Sci (Lond) 2009;116:781-9. [Crossref] [PubMed]
- Pamukcu B, Lip GY, Snezhitskiy V, et al. The CD40-CD40L system in cardiovascular disease. Ann Med 2011;43:331-40. [Crossref] [PubMed]
- Reddy VY, Mobius-Winkler S, Miller MA, et al. Left atrial appendage closure with the Watchman device in patients with a contraindication for oral anticoagulation: the ASAP study (ASA Plavix Feasibility Study With Watchman Left Atrial Appendage Closure Technology). J Am Coll Cardiol 2013;61:2551-6. [Crossref] [PubMed]
- Bergmann MW, Betts TR, Sievert H, et al. Safety and efficacy of early anticoagulation drug regimens after WATCHMAN left atrial appendage closure: three-month data from the EWOLUTION prospective, multicentre, monitored international WATCHMAN LAA closure registry. EuroIntervention 2017;13:877-84. [Crossref] [PubMed]
- Holmes DR, Reddy VY, Turi ZG, et al. Percutaneous closure of the left atrial appendage versus warfarin therapy for prevention of stroke in patients with atrial fibrillation: a randomised non-inferiority trial. Lancet 2009;374:534-42. [Crossref] [PubMed]