Bringing doctors together

or

A free, private network for verified physicians

Leading voices

Dr. Dorairaj Prabhakaran

General Cardiology

View

Dr. Pugazhendhi Vijayaraman

EP · Geisinger

View

Dr. Robert Schaller

EP · HUP

View

Dr. Trey Woods

General Cardiology · Emory Healthcare

View

Dr. Brahmajee Nallamothu

Interventional Cardiology · UMHS

View

Dr. Felix Yang

EP · NYU Langone

View

Dr. Jennifer Rymer

Interventional & Peripheral Cards · Duke University

View

Dr. Ami Bhatt

Adult Congenital Heart Disease · MGH

View

Dr. Gregory Marcus

EP · UCSF

View

Dr. Connor O'Brien

Cardiac Critical Care · UCSF Medical Center

View
  • Discover your community

  • Stay current

  • A profile that represents you

  • Find your people

About RoonPrivacyTerms

    Overview

    Primary Results of the SALAMANDER Registry: A Multicenter Observational Cohort Study.

    May 14, 2026DOI: 10.1161/jaha.125.046625Journal ArticleAHA

    Written by: Wojciech WańhaŁukasz KuźmaMariusz KowalewskiSylwia IwańczykJoanna GrygierJan JeskeMarco FrazzettoRadosław LitwinowiczMarian BuryszEwa Gaszewska-ŻurekClaudio SanfilippoTomasz BochenekMichał LelekKrystian WitaZenon HuczekAgnieszka Kapłon-CieślickaTomasz MazurekPiotr ScisłoJerzy SachaKrzysztof StruniawskiBrunon TomasiewiczFacehauwe MezqueWiktor KuliczkowskiKrzysztof BartuśSławomir DobrzyckiMichał ŚwięczkowskiAnna KuraszPaweł KraliszDawid MiśkowiecJarosław D KasprzakMarta ChameraMaksymilian MielczarekPiotr DrewlaRafał GałąskaFabrizio UgoMarco FranzinoBernardo CorteseFilippo L GurgoglioneLuigi VignaliGiorgio BenattiFederico BarocelliPiotr WacińskiPiotr NiezgodaAdam SukiennikJacek KubicaMario IannaconeKlaudia KowalskaŁukasz LewickiCarmelo GrassoDavide CapodannoFabrizio D'AscenzoOvidio de FilippoFrancesco BrunoRadosław GocołZbigniew KalarusPiotr SuwalskiMarek GrygierGrzegorz SmolkaGregory Y H LipWojciech Wojakowski

    Find these authors on Roon

    BC
    Dr. Bernardo Cortese
    Interventional & Complex Coronary Cards · CWRU

    Methods

    The data that support the findings of this study are available from the corresponding author upon reasonable request.The SALAMANDER registry is a real‐life, multicenter, observational study conducted across 16 high‐volume cardiac centers in Europe, with the primary objective of evaluating the safety and efficacy of LAAO using a variety of currently available technologies.14 Overall, 7 distinct devices have been assessed within the registry framework, representing the technological diversity of current clinical practice. These include (1) AtriClip (AtriCure, Mason, OH), a fully thoracoscopic, epicardial clip system applied externally to occlude the LAA via minimally invasive surgical access; (2) Watchman (Boston Scientific), a percutaneous, endocardial occluder delivered transseptally to seal the LAA from within the atrium; (3) Watchman FLX (Boston Scientific), the next‐generation version of the Watchman device, with enhanced conformability and anchoring features; (4) Amulet (Abbott, Plymouth, MN), a dual‐seal endocardial device comprising a lobe and a disc, suitable for complex LAA morphologies; (5) LARIAT (SentreHeart, Redwood City, CA), a hybrid endo‐epicardial system involving transseptal endocardial guidance and epicardial suture‐based ligation; (6) Lambre (Lifetech Scientific, Shenzhen, China), a self‐expanding, endocardial occluder with a distinctive anchoring mechanism and adaptability to varying appendage anatomies; and (7) Omega (Shape Memory Medical Inc., Santa Clara, CA), an endocardial device using shape memory.The patients' data were collected using an electronic case report form developed by the Polish Cardiac Society and were protected in accordance with the applicable Polish regulations, the General Data Protection Regulation, and institutional standard operating procedures (Data S1). The study was conducted in line with the principles of the Declaration of Helsinki and was registered at www.ClinicalTrials.gov (NCT05144958). The study adhered to Strengthening the Reporting of Observational Studies in Epidemiology guidelines for observational studies (Table S1). Ethical approval was obtained from the local Bioethics Committee of the Upper‐Silesian Medical Centre of the Silesian Medical University in Katowice, Poland (PCN/CBN/0052/52/139/I/22). Due to the noninterventional, observational nature of this prospective registry, the requirement for informed consent was waived. To minimize potential bias, all participating centers followed standardized data collection procedures with predefined event definitions and protocol‐based measures to ensure cross‐center consistency in reporting. In addition, regular data quality checks, source data verification, and central data validation were performed to ensure the completeness, accuracy, and uniform ascertainment of outcomes across sites.
    Patients in the study have been continuously recruited at each participating center to ensure unbiased inclusion of the eligible population. The enrollment period for this analysis spanned from October 29, 2010, to December 19, 2024. Inclusion criteria for the SALAMANDER registry comprise the following: (1) age ≥18 years; (2) diagnosis of paroxysmal, persistent, or permanent nonvalvular AF; (3) increased thromboembolic risk, defined as a CHA2DS2‐VASc score ≥2 in men or ≥3 in women; (4) elevated bleeding risk, particularly in patients with a HAS‐BLED score ≥3; (5) history of significant bleeding events; (6) documented bleeding complications while on OAC, including VKAs; (7) presence of comorbid conditions predisposing to bleeding, such as coagulopathies, thrombocytopenia, or end‐stage renal disease requiring dialysis; and (8) failure of oral anticoagulation due to stroke recurrence or associated cerebrovascular abnormalities such as cerebral aneurysm. Participants in this registry met all the inclusion criteria and did not meet any exclusion criteria. Study sites are listed in Table S2. The flowchart of the study design is presented in Figure 1.
    IQR indicates interquartile range.
    Clinical and procedural data are collected in accordance with the study protocol and predefined criteria for success and complications, thereby enabling a robust assessment of the safety and efficacy of LAA closure in real‐world clinical practice. The primary outcomes analysis focused exclusively on clinical events occurring during the index hospitalization (from the LAAO procedure until hospital discharge). Imaging outcomes were additionally assessed at a 1‐time control visit within 3 months after the procedure. This study uses structured, prespecified end points based on the definitions outlined in the Munich Consensus.15 As the SALAMANDER registry is ongoing, postdischarge clinical outcomes are not analyzed in this article, which focuses on baseline characteristics, procedural data, and in‐hospital outcomes. Longitudinal follow‐up outcomes (24 months) will be reported separately once the full data set becomes available to avoid bias related to unequal follow‐up duration and potential temporal changes in device selection and procedural strategies.
    The primary end point was technical success, defined as complete LAA exclusion without device‐related complications and with no residual peridevice leak >5 mm. Procedural success was defined as achievement of technical success without major periprocedural complications, including vascular injury, thromboembolic events (eg, stroke or air embolism), periprocedural death, or major bleeding. Major bleeding was defined as type 3 or higher according to the Bleeding Academic Research Consortium criteria. Ischemic stroke was defined as an acute focal neurological deficit presumed to be due to ischemia, persisting ≥24 hours, or <24 hours with imaging confirmation of infarction. Transient ischemic attack was defined as a rapid‐onset focal neurologic deficit due to cerebral ischemia, resolving within 24 hours. Systemic embolism was defined as an occlusive thrombotic event in a systemic artery with clinical manifestations. Effective device occlusion is defined as a residual jet ≤5 mm, evaluated intraoperatively or during control visit (transoesophageal echocardiography or computed tomography), depending on the specific device used. Given the heterogeneous nature of available LAAO technologies, including 7 distinct occluder systems—Watchman, Watchman FLX, Amulet, LARIAT, AtriClip, Lambre, and Omega—the study adopted uniform, device‐agnostic definitions of success. Despite differences in design, deployment strategy (endocardial, epicardial, or hybrid), and imaging control protocols, the endpoint definitions were standardized across all devices to enable valid interdevice comparisons. Procedural metrics, including implantation depth, compression ratio, residual leak, need for partial or full recapture, and imaging features (eg, spontaneous echo contrast, coverage of trabeculations), were systematically collected. Device‐specific parameters were reported using a harmonized data structure, and all echocardiographic data were evaluated by a central core laboratory to ensure consistency across the diverse technologies.
    Data were summarized using mean±SD and median (interquartile range) for continuous data, and count (percentage) for categorical data. Continuous variables were tested for distribution by the Kolmogorov–Smirnov test. For comparisons between 2 groups, the unpaired Student's t test or Mann–Whitney U test was applied, depending on data distribution. For comparisons across >2 groups, ANOVA or the Kruskal–Wallis test was used as appropriate. Categorical variables were compared using the χ2 test or Fisher's exact test. To identify independent predictors of procedural success and clinical outcomes, multivariable logistic regression analysis was performed. Variables with a P value <0.05 in univariable analysis were included in the multivariable model. Results are expressed as odds ratios with 95% CIs. Figures were created in BioRender (https://biorender.com/). Given the minimal extent of missingness and the fact that these variables were not used for building adjusted multivariable models at this stage of the analysis, no imputation procedures were performed.

    Results

    The study enrolled 1660 patients (median age, 76 [interquartile range, 70–81] years; 38% women). A total of 98.7% of patients had AF, and the remaining 1.3% were diagnosed with atrial flutter. The median LVEF was 55%. The most common comorbidities were hypertension (90%), chronic kidney disease (49.9%), and coronary artery disease (37.7%). A prior history of ischemic stroke was reported in 26.3% of patients and systemic embolic events in 12.1%. The median CHA2DS2‐VASc was 4 (interquartile range, 3–5). The median HAS‐BLED score was 4 (interquartile range, 3–4). Detailed characteristics of enrolled participants are presented in Table 1.
    Study Group Characteristics
    AF indicates atrial fibrillation; DOAC, direct oral anticoagulant; IQR, interquartile range; OAC, oral anticoagulation; and VKA, vitamin K antagonist.
    Patient characteristics according to the implanted device are presented in Table S3. The most common types of LAA shapes were chicken wing (63.5%), windsock (13.5%), and cauliflower (12.5%). Additional data on the characteristics of LAA are presented in Table S4.
    The most common indication for LAAO was significant bleeding, in 83.7% of enrolled patients. Among these bleeding events, 68.8% occurred during treatment with DOACs and 24.9% during therapy with VKAs. The most frequent bleeding sites were the lower (27.9%) and upper (21.7%) gastrointestinal tracts, and 17.4% had a history of hemorrhagic stroke. The second‐ and third‐most‐common indications were other reasons (8.4%) and failure of OAC (5.6%), respectively. A detailed description of the indications for LAAO is presented in Table 1.
    The most frequently used devices were Watchman FLX (39%), Amplatzer Amulet (30.1%), and Watchman (17.5%). General anesthesia was administered to 55.9% of enrolled participants, of which 32.6% received intubated anesthesia. The remaining procedural characteristics are presented in Table 1.
    Before the procedure, 91.7% of patients received OAC in the form of DOACs (48.9%), VKAs (17.1%), DOACs with single antiplatelet therapy (SAPT; 12%), triple antithrombotic therapy (5.7%) or low‐molecular‐weight heparin (4.1%).
    The most common antithrombotic regimens after the procedure were DAPT (50%), SAPT (27.3%), and DOACs (9.8%). The percentage distribution of various antithrombotic regimens is presented in Figure 2.
    DAPT indicates dual antiplatelet therapy; DOAC, direct oral anticoagulant; LMWH, low‐molecular‐weight heparin; SAPT, single antiplatelet therapy; TAT, triple antithrombotic therapy; and VKA, vitamin K antagonist.
    The most common antithrombotic regimen before the procedure, DOACs (48.9%), was most often switched to DAPT (45.4% of all DOAC‐treated patients) or SAPT (24.4% of all DOAC‐treated patients) postprocedurally, while in 18.8% of cases it was continued. Among patients receiving VKAs (17.1% of all patients before the procedure), 58.8% were switched to DAPT, 22.2% to SAPT, and 12% continued their previous regimen. The majority of patients receiving low‐molecular‐weight heparin before the procedure (4.1% of all patients) were switched to DAPT after the LAAO (70.6% of all low‐molecular‐weight heparin–treated participants). Among patients receiving triple antithrombotic therapy before the procedure, 65.3% were treated with DAPT and 31.6% with SAPT following LAAO.
    An illustrative distribution of changes in antithrombotic therapy is presented in Figure 3.
    DAPT indicates dual antiplatelet therapy; DOAC, direct oral anticoagulant; LMWH, low‐molecular‐weight heparin; SAPT, single antiplatelet therapy; TAT, triple antithrombotic therapy; and VKA, vitamin K antagonist. *AtriClip does not require antithrombotic therapy.
    The technical success rate was 95.5%, with the most common causes of failure being residual leak (2.2%) and tamponade (1.4%). The procedural success rate was 90.5%. The most frequent factors of procedural failure were vascular complications (3.7%), periprocedural death (1.1%), and major bleeding (0.7%). Among the 19 periprocedural deaths, the leading causes were cardiac tamponade (5 cases [26.3%]) and major bleeding (4 cases [21.1%]). Urgent cardiac surgery was required in 9 patients, including the need for device removal in 5 cases (55.6%). A detailed description of the primary outcomes is presented in Table 2. No statistically significant differences were observed among the occluders in technical and procedural success rates, as shown in Table 2. Detailed results of the univariate analysis are available in Tables S5 and S6.
    Primary Outcomes Overall and on the Basis of the Type of Device Used
    No statistically significant differences were observed in the frequency of primary outcomes after dividing the analyzed period into 2 time intervals. Year‐by‐year analyses demonstrated stable technical and procedural success rates, whereas the incidence of major bleeding and periprocedural death varied significantly over time (P<0.001 and P=0.003, respectively). Notably, these adverse events showed lower rates in more recent years compared with earlier peaks, albeit with marked interannual variability. Temporal trends in the frequency of primary outcomes and in the use of individual devices are presented in Tables S7 through S9.

    Discussion

    This study represents one of the largest European cohorts evaluating the outcomes of stand‐alone LAAO for thromboembolism prevention in patients with nonvalvular AF. Our findings demonstrate that the stand‐alone LAAO procedure is associated with a high technical (95.5%) and procedural (90.5%) success rate, regardless of the device type, reflecting the advances in device technology and operator expertise. Second, we observed a reduction in the use of antiplatelet and anticoagulant therapy in patients undergoing LAAO, highlighting the procedure's potential to minimize long‐term pharmacological burden and associated bleeding risks. These results reinforce the growing evidence supporting LAAO as a viable and effective alternative to long‐term OAC, especially in patients with contraindications or intolerance to anticoagulant therapy. Two general problems are observed in many patients with AF receiving OAC. The first is an inadequate level of anticoagulation due to poor compliance, leading to increased stroke risk, especially for DOAC. The second is an induction of bleeding, which leads to discontinuing the therapy. One real‐world study revealed that ≈30% of patients with nonvalvular AF discontinue OAC in the first 2 years after therapy admission.16 The most frequently documented reasons for discontinuing anticoagulation were bleeding (13%) and infrequent AF paroxysms (14%).17 In the United States, analysis of MarketScan data (12 129 patients) revealed that up to 47% of patients with nonvalvular AF discontinued anticoagulation therapy in an average of 120 days, most commonly due to bleeding.16 Hence, a reliable management option is needed for these patients. Previous real‐world studies, although based on smaller patient cohorts and using a narrower spectrum of devices, demonstrate technical and procedural success rates comparable with the outcomes achieved in our study. One Belgian study involving 457 patients undergoing transcatheter LAAO with the use of either the Watchman or Amplatzer cardiac plug/Amulet device reported a technical success rate of 97.1% and a procedural success rate of 95.9%.18 The EWOLUTION (Registry on WATCHMAN Outcomes in Real‐Life Utilization) registry, which included 1025 patients undergoing transcatheter LAAO with the Watchman device, demonstrated technical and procedural success rates of 98.5% and 97.3%, respectively.19
    In the ITALIAN‐FLX registry, which included 772 patients, there was a technical success rate of 100% and a procedural success rate of 97.3%.20 Large randomized clinical trials have shown technical success rates that, while comparable, remain lower: 88% in PROTECT‐AF and 90% in PRAGUE‐17.12, 21 The high rates of technical and procedural success reported in both real‐world studies, such as ours, and in randomized clinical trials underscore the safety and feasibility of transcatheter LAAO. In addition, a number of studies showed good performance of external epicardial occlusion devices both in animal studies and clinically implanted concomitantly to open heart surgery and thoracoscopic ablation, revealing no safety issues and nearly 100% of LAA closure with no pouch and stable long‐term results.22, 23, 24, 25, 26, 27, 28 Different epicardial (AtriClip) and endocardial devices (Watchman, Amulet), are in various stages of development or clinical use with good results in experienced centers. Some doubts about LAAO effectiveness have been formulated in terms of transcatheter techniques.29, 30, 31 There is a debate whether it is related to acute complications, need for bridging anticoagulation, mid‐ and long‐term device leakage, or thromboembolism originating from other parts of the left atrium.32, 33 Epicardial clips are able to avoid those weaknesses, while limitations associated with the thoracoscopic intervention, such as the need for short induction and intubation, severely depressed lung function, and previous surgery with left pleural cavity opening, are self‐evident. The optimal antithrombotic regimen following LAAO remains debatable. Despite growing experience with the procedure and, therefore, increasing implantation success rates, there is still no universally accepted postimplantation protocol. Various strategies have been proposed, ranging from DAPT to short‐term OAC, depending on patient bleeding risk profiles. The current evidence on the optimal antithrombotic strategy following LAAO remains conflicting, especially since studies include patients with widely varying bleeding risks and use inconsistent regimens. There is no general consensus on the most effective/safe approach, and clinical practice can vary significantly between centers.34 While some studies support the use of low‐dose DOACs for reducing thromboembolic and bleeding events, others report no significant advantage over simplified regimens such as SAPT or even no therapy.35 A network meta‐analysis of 41 studies36 found that DOAC monotherapy was associated with the lowest rates of thromboembolic events and the highest risk of major bleeding, while DOAC treatment was superior to VKA in reducing all‐cause death by 61%, and DAPT was associated with 50% lower risk of thromboembolic events compared with SAPT. Clinical trial evidence comparing antithrombotic regimes is limited. The ADALA (Low‐Dose Direct Oral Anticoagulation Versus Dual Antiplatelet Therapy After Left Atrial Appendage Occlusion) trial identified that low‐dose DOAC, in the form of apixaban 2.5 mg twice daily, was a superior therapeutic option, reducing the risk of the primary end point (major bleeding, thromboembolic events, and device‐related thrombosis) by 81% at 3‐month follow‐up.37 Conversely, there are some observational data. For example, findings from the LOGIC (Left Atrial Appendage Occlusion in Patients With Gastrointestinal or Intracranial Bleeding) registry demonstrated no significant differences in overall death, cardiovascular death, thromboembolic events, or bleeding risk between patients receiving SAPT or no antithrombotic therapy and those treated with standard antithrombotic regimens.38 In the EWOLUTION study, the use of DAPT was comparable (60.8% versus 50% in our cohort), while SAPT was less common (6.9% versus 27.3% in our study).19 In contrast, John et al reported DAPT use in 72% of their patients, with limited use of DOACs (3.7% versus 9.8% in our cohort).18 In the SALAMANDER registry, discharge antithrombotic therapy varied not only between centers but also at the operator level (as the choice of postprocedural regimen was left to the operator's discretion). This likely reflects real‐world decision making that is strongly shaped by the individual patient's clinical profile, procedure‐related anatomic considerations, and implant quality (optimal anatomic sealing with stable device positioning). In many cases, the predominant driver is a bleeding‐prone phenotype (frequent prior major bleeding, particularly gastrointestinal or intracranial), in whom even short‐term OAC may be considered unacceptable, favoring DAPT or SAPT. Conversely, a distinct subgroup undergoes LAAO due to perceived OAC failure, that is, thromboembolic events (including stroke) occurring despite anticoagulation, which may prompt some operators to prefer a period of OAC or DOAC‐based therapy after the procedure, even for long‐term use. These patient‐level contrasts, combined with center/operator‐specific decisions, plausibly explain the observed heterogeneity in post‐LAAO regimens. This heterogeneity also mirrors differences between landmark trials. In PROTECT AF and PREVAIL, the postimplant protocol was anchored in short‐term warfarin plus aspirin followed by a transition to antiplatelet therapy, reflecting a strategy that assumes temporary anticoagulation is feasible in many patients. In contrast, PRAGUE‐17 enrolled a high‐risk population and used a more antiplatelet‐forward post‐LAAO approach, which is closer to the DAPT‐dominant practice seen in contemporary bleeding‐driven cohorts such as SALAMANDER.10, 11, 12
    Device selection (endocardial versus epicardial) for LAAO is usually determined by LAA anatomy, bleeding risk profile, feasibility of antithrombotic therapy, and institutional expertise.8 Endocardial systems require transseptal access and implantation within the LAA lumen, necessitating at least a temporary course of anticoagulant or antiplatelet therapy. In contrast, epicardial techniques achieve LAA exclusion without leaving permanent material within the LAA. This fundamental distinction often allows simplification or avoidance of long‐term anticoagulation.39 Moreover, epicardial systems require pericardial access or surgical exposure, which introduces procedure‐specific risks such as pericardial bleeding, effusion, tamponade, and injury to adjacent structures.8 Endocardial devices, by contrast, are implanted via a fully percutaneous transseptal approach and are therefore associated with a lower immediate procedural risk profile. Their periprocedural complications are more commonly related to transseptal puncture, device embolization, or access‐site bleeding.40 In clinical practice, these differences directly influence patient selection and the individualization of periprocedural and postprocedural management strategies.

    Limitations

    The SALAMANDER registry is an observational, real‐world multicenter cohort study. The AF population has been recruited across 2 countries (Poland and Italy), but as with any observational cohort study, there are limitations, including common limitations to observational cohort studies, inadequacy of data recording/capture, and selection bias. Despite selecting high‐volume reference centers to ensure uniform procedural quality, certain data elements were not collected systematically across sites. In this context, a notable example is the lack of detailed information on imaging modalities used during percutaneous LAAO procedures, which reflects the inherent challenges of comprehensive data acquisition in large observational registries. The large overall study cohort is a strength of this study; however, the marked imbalance in sample sizes between device groups limits the ability to reliably compare outcomes, particularly rare events such as tamponade, and therefore warrants cautious interpretation. Finally, as the present cohort is derived exclusively from European populations, the generalizability of these findings to other regions may be limited, given the known ethnic differences in AF‐related complication (ie, stroke and bleeding),41, 42 LAA morphology,43 and the so‐called “East Asian” paradox in relation to antithrombotic therapy effectiveness and safety44; therefore, similarly large prospective cohorts from different populations globally are warranted.

    Conclusions

    In one of the largest European cohorts, we show that the stand‐alone LAAO procedure is associated with a high technical and procedural success rate, with potentially comparable outcomes across device types. The reduction in the use of antiplatelet and anticoagulant therapy in patients undergoing LAAO reduces long‐term OAC use and associated bleeding risks.

    Sources of Funding

    The SALAMANDER study was initiated and funded by the Scientific Platform of the Polish Cardiac Society. The funding body had no role in study design, data collection, data analysis, data interpretation, or article writing.

    Disclosures

    Dr Lewicki has consulted and lectured for Abbott and Boston Scientific outside the present work. Dr Grasso has consulted, lectured, and served as an advisory board member for Abbott, Boston Scientific, and Edwards outside the present work. Dr Grygier has consulted, lectured, and served as an advisory board member for Boston Scientific outside the present work. Dr Smolka has consulted and lectured for Abbott and Boston Scientific outside the present work. Dr Kapłon‐Cieślicka has consulted, served as an advisory board member, received honoraria for lectures, and received travel grants from Adamed, AstraZeneca, Bayer, Bausch Health, Boehringer Ingelheim, KRKA, Pfizer, Polpharma, and Servier outside the present work. Dr Kralisz consulted and lectured for Boston Scientific outside the present work. Dr Waciński has consulted, lectured, and served as an advisory board member for Philips outside the present work. The remaining authors have no disclosures to report.

    Supporting information

    Data S1
    Tables S1–S9
    BCDr. Bernardo Cortese

    Dr. Bernardo Cortese

    Cardiology · Interventional & Complex Coronary
    Case Western Reserve University
    bernardo