Atrial fibrillation (AF) is the most common cardiac arrhythmia with a prevalence of 1-2%. Without antithrombotic treatment, the annual risk of a cardioembolic event is 5-6%. The source of a cardioembolic event is a thrombus, which usually forms in the left atrial appendage (LAA). The prevention of cardioembolic events involves treatment with anticoagulant drugs, which were limited to, until recently, vitamin K antagonists (e.g. warfarin). Anticoagulant treatment with warfarin can lead to adverse bleeding events, some of which can be life threatening. Recently, two new options for thrombus prevention have been developed. The first is the novel anticoagulants (NOAC), which were associated with slightly better safety profiles due to a lower frequency of intracranial bleeding in large randomized trials . The second option involves interventional occlusion of the LAA. The aim of this project is to compare the LAA occlusion intervention to NOAC pharmacological treatment in a randomized multicenter study of AF patients at high risk of a cardioembolic event.

Interventional left atrial appendage closure vs. novel anticoagulation agents in high-risk patients with atrial fibrillation (PRAGUE-17 study)

INTRODUCTION Atrial fibrillation (AF) is the most common sustained cardiac arrhythmia with a prevalence of 1-2% in the general population; additionally, the incidence increases with age. (1) Among the elderly 75 years and older, its incidence is 12%. (2) AF is associated with a five-fold risk of a cardioembolic event; approximately 20% of all strokes are due to AF. (3) The annual risk of stroke in AF patients is 5 -6 % in AF patients without antithrombotic treatment. Moreover, a stroke due to AF is typically more severe, i.e. associated with worse neurological deficits. (4) Not only does the incidence of AF increase with age, the risk of strokes in AF patients also increases with age with age: in patients between 50-59 years, the risk of stroke is about 1.5%, while in patients between 80 - 90, the risk increases to about 23%. (5) The source of an embolic stroke is a thrombus, which forms in the left atrium (LA) due to insufficient mechanical contraction of LA. Most often thrombi arise in the left atrial appendage (LAA); according to data from autopsies and findings from transesophageal echocardiography (TEE) in AF patients, approximately 90% of all left atrial thrombi arise in the LAA. (6) In addition to insufficient mechanical contraction of LA, the pro-coagulant status of patients also plays an important role in thrombus development. A patient's risk profile is characterized according to the CHA2DS2-VASc score. Each letter presents one recognized risk factor (C - congestive heart failure or left ventricular dysfunction, H - hypertension, A - age > 75 years (2 points), D - diabetes, S - history of stroke (2 points), V - present vascular disease (atherosclerosis), A - age > 65 years, Sc- female sex. In the CHA2DS2-VASc scoring system, each risk factor is evaluated as either 1 or 2 points. The sum of all points presents the annual risk of a cardioembolic event, in percent. The prevention of strokes or other systemic cardioembolic events (thrombus can embolize to peripheral artery of upper or lower extremity, artery supplying gastrointestinal tract, renal artery etc.) involve treatment using anticoagulant drugs, which were, until recently, vitamin K antagonists only (e.g. warfarin). These drugs block the effect of vitamin K in the liver, which in turn blocks the production of coagulant factors II, VII, IX, and X. As a result, the coagulation cascade is slowed. The efficacy of warfarin is estimated using International Normalized Ratio (INR); the effective INR to prevent thrombus formation in AF patients is between 2.0 and 3.0. The amount of vitamin K in food varies, which can lead to variations in INR despite a stable warfarin dose. This represent one of the disadvantages of warfarin treatment. In large randomized trials, in which all patients are very carefully followed, only 65% of INR values were successfully maintain between 2.0 and 3.0. (7) In real clinical praxis, it was shown that only 50% of all INR values are within the desired range. (8) This can easily lead to both overtreatment and undertreatment of patients. Vitamin K antagonism is thought to be effective when 70% of INR values are maintained in the therapeutic range. (9) Another disadvantage of warfarin treatment involves dietary restrictions.

Recently, several new drugs that can block coagulation have been discovered, and are appropriately named novel anticoagulants (NOAC). These new drugs include direct thrombin (factor IIa) inhibitors, such as dabigatran, and inhibitors of factor Xa, such as rivaroxaban and apixaban. In contrast to warfarin, they block a single step in the coagulation cascade. The dosing of these drugs does not have to be adjusted based on laboratory tests, like warfarin. Stable concentrations of these drugs are achieved by once or twice daily dosing, according to their half-lives. Dosing only needs to be adjusted when there is significant comorbidity (i.e. renal or liver function) and no dietary restrictions are needed. In large prospective multicenter studies, NOAC were non-inferior to warfarin, i.e. prevention of systemic cardioembolic events was similar between patients treated by warfarin and NOAC. (10) (11) (12) (13) Additionally, NOAC were associated with slightly better safety profiles relative to the lower frequency of intracranial bleeding. (10) (11) (12) (13)

The prevention of thrombus formation in the fibrillating atrium by anticoagulation has been shown to be effective; however, adverse bleeding events can occur with both warfarin (even in the therapeutic range) and NOAC. The frequency of any (major or minor) bleeding during NOAC treatment is approximately 15-18% per year; severe bleeding can lead to hospitalization or even death.

The risk of bleeding can be estimated using the HAS-BLED scoring system. Each letter presents a risk factor for bleeding (H - hypertension, A - abnormal liver or renal function, S - history of stroke, B - history of bleeding, L - labile INR, E - elderly patients, and D - drugs, alcohol consumption). Patients with HAS-BLED scores > 3 are considered to be at high risk of bleeding.

The limitation of anticoagulative drugs was one of reasons behind the development of nonpharmacological treatments for AF. Because about 90% of left atrial thrombi occur in the LAA; (6) the occlusion of the LAA is the main non-pharmacological option for prevention of cardioembolic event from the LA. LAA occlusion is performed using a percutaneous approach; the catheter is inserted through the femoral vein via a transseptal puncture into the LA. The LAA is visualized and closed using a special device. Currently, two different devices are present on the market (Watchman, Boston Scientific and Amplatzer Cardiac Plug, St. Jude Medical). The largest trial published regarding LAA occlusion was the PROTECT trial. (14) In this prospective randomized multicenter trial, occlusion of the LAA using a Watchman device was compared with standard warfarin treatment. The LAA occlusion was non-inferior to warfarin; the incidence of cardioembolic events and death was similar in both groups (3% of patients per year in the LAA occlusion group vs. 4.9% of patients per year in the warfarin group). The major disadvantage of LAA occlusion was related to peri-procedural and device-related complications, which were present in 7.4% of patients per year. Fifty-five percent of the reported adverse event occurred on the day of the procedure and the most often was significant pericardial effusion, which was present in 4.8% patients. In summary, LAA occlusion can be associated with significant risk of peri-procedural complications; additionally, while it is reduced, the risk of thrombus formation in the LA is not eliminated. In other study assessing LAA occlusion, periprocedural and device-related complications occurred in 6 - 8%. (15) Despite the risk, LAA occlusion presents a very promising alternative to anticoagulant treatment for patients with a very high risk of bleeding, or in those contraindicated for warfarin or NOAC. For patients with the highest CHA2DS2-VASc and HAS-BLED scores, no studies have been published to date, and no studies are in progress according to the randomized trial study registry. The inclusion criterion in the PROTECT study was a CHADS2 score > 1, with approximately 1/3 of patients having a CHADS2 of 1 only. (14) The anticoagulation treatment is well established in the majority of patients with AF in standard clinical practice; however, despite the anticoagulation treatment, systemic cardioembolic event can occur and bleeding risks can be very serious. In the largest study with rivaroxaban, the ROCKET study, in which the drug was compared to warfarin, the annual incidence of stroke was in 1.7 - 2.2% of patients and the annual incidence of any bleeding (both major and minor) was 15% (11) Similarly, in the ARISTOTLE study, in which another NOAC apixaban was compared to warfarin, the annual incidence of cardioembolic event was 1.27 - 1.6%, and the annual risk of major or clinically significant bleeding was 4 - 6%.(10) The best treatment for those AF patients at highest risk, i.e.

patients with high CHA2DS2-VASc scores and a very high bleeding risk, is still not known. Moreover, the number of such patients has been increasing due to the increasing age of the general population.

Therefore, treatment of these patients presents a serious problem for clinical practice and a search for the optimal solution is warranted. The safety profile for NOAC seems to be slightly better than warfarin; nonetheless, the bleeding risk with NOAC is not negligible, which is evident by comparing the CHA2DS2-VASc and HAS-BLED scoring systems: some factors increase the risk for both thrombotic and bleeding events.

The Aim of our project is to compare LAA occlusion with NOAC in a randomized multicenter study of AF patients at the greatest risk. Such randomized study has never been done.

PATIENTS AND METHODS Patients with AF with an indication for anticoagulation will be eligible for the study. Inclusion criteria will be the presence of AF (paroxysmal, persistent or permanent), signed informed consent and one of the following criteria

1. history of significant bleeding (i.e. bleeding which required intervention or hospitalization), even in the absence of anticoagulation treatment at the time of the bleeding event, or
2. a cardioembolic event, which occurred on anticoagulation, or
3. a high risk profile of the patient, defined as a CHA2DS2-VASc score ≥ 3 and a HAS-BLED score ≥ 2

Exclusion criteria will be as follows:

1. thrombus in the LA or LAA;
2. mechanical valve prosthesis;
3. mitral stenosis;
4. previous LAA ligation during cardiac surgery;
5. life expectancy less than 2 years;
6. comorbidities other than AF, which present an indication for anticoagulation;
7. patent foramen ovale with atrial septal aneurysm
8. mobile plaque in the aorta;
9. symptomatic atherosclerosis of the carotid artery;
10. pericardial effusion greater than 10 mm;
11. clinically significant bleeding within the 30 days prior to the scheduled procedure;
12. stroke or other cardioembolic event within the 30 days prior to the scheduled procedure;
13. acute coronary syndrome within the 90 days prior to the scheduled procedure,
14. gravidity,
15. significant valvular disease,
16. creatinine clearance less than 30 ml/min

A routine TEE will not be done on all patients before randomization. If the patient is randomized to the LAA occlusion arm, a TEE examination will be done to exclude the presence of a thrombus in the LA or LAA before the invasive procedure. The presence of a thrombus in the LAA or LA is an additional exclusion criterion (patients with a thrombus in the LAA will be removed from analyses).

Next, a cardiac CT will be done to assess the morphology of the LAA, and a blood sample will be drawn for assessment of natriuretic peptides and cytokines.

Patients will be randomized to the interventional LAA occlusion (group A) or NOAC treatment (group B) in a 1/1 ratio, the randomization will be done using web-based randomization software. The software is designed to respect the CHA2DS2VASc characteristics of patients, i.e., the goal is to have comparable CHA2DS2VASc in both arms at the end of the study. The randomization process will be done outside all participating centers and therefore it will be masked from study subjects and site personnel.

For those randomized to group A (LAAO group), an interventional LAA occlusion will be performed. The type of LAA occlusion system will be left on the discretion of the implant center; however, only Amulet device (St. Jude Medical) and the Watchman device (Boston Scientific) are available for use in this study. The goal will be to achieve approximately half of patients implanted with the Watchman system (Boston Scientific) and the other half implanted with the Amulet system (St. Jude Medical). The protocol and LAA imaging for the procedure will use the current praxis of the surgeon: the procedure can be done under TEE, ICE, or angiographic visualization and control, depending on the experience and preference of the particular center and surgeon.

After the procedure, all patients will be treated with a combination of aspirin 100 mg (once daily) and clopidogrel 75 mg (once daily) for three months. After three months, a control TEE will be done. If the TEE finding is satisfactory (i.e., the absence of a thrombus on the device, and the absence of peri-device leakage or a leakage line of less than 5 mm, and a satisfactory position of the device as assessed by the surgeon), the clopidogrel will be discontinued, however, the aspirin (100 mg, once daily) will be continued. Additional transesophageal echocardiographies are not planed unless indicated by the clinical status of the patient.

In some patients, the duration of dual antiplatelet treatment could be individualized taking into consideration the particular risks of that patients. In patients with very high bleeding risk (e.g. patients with a history of repeated bleedings), the duration of dual antiplatelet treatment could be shortened to 6 weeks. After that, if a control TEE yields satisfactory results, clopidogrel could be discontinued and patients could be left on aspirin monotherapy only. Six-weeks of dual antiplatelet treatment has been recently shown to be sufficient after LAA occlusion. (16) On the other hand, in patients with a significant thrombotic risk (e.g. with a history of repeated cardioembolic events, but without a history of overt bleeding), the duration of dual antiplatelet treatment could be prolonged to 6 months. The decision (i.e., the assessment of the individual risk profile of an implanted patient) and the patient-specific approach will be left to the discretion of the treating physician.

Experience with the LAA occlusion procedure In studies published regarding LAA occlusion, the experience of the interventional cardiologist (learning curve) and staff was shown to be very important; i.e. the number of complications during the first implantations were significantly higher compared to those that followed. (14, 17) To prevent interventional lack of experience from affecting outcomes, the other two teams will start performing the procedures before the initiation of the study, with the goal of doing at least 5- 10 LAA occlusions before the start of the study. This first 5-10 LAA occlusion will be done not according to the inclusion criteria of our study, but the patients will be enrolled according to the current standard criteria and guidelines for LAA occlusion. Additionally, the first 5-10 procedures in the study will be done with the assistance of an interventional cardiologist experienced with LAA occlusion.

For those randomized to group B (NOAC group), patients will be treated using a registered NOAC (rivaroxaban, apixaban or dabigatran, but preferentionally apixaban) at the dose recommended by the manufacturer (in a case of apixaban 5 mg or 2.5 mg twice daily) and adjusted relative to comorbidities (body weight, age, and renal function).

Outpatient follow-up will be done at 6 weeks, 3, 6, 9, and 12 months after randomization and then every 6 months. At the 6-month follow-up, a blood sample will be drawn for cytokines measurement.

Endpoints: The primary endpoint of the study will be the occurrence of any of the following events within 24 months following randomization:

(1) stroke or TIA (any type), (2) systemic cardioembolic event, (3) clinically significant bleeding 4) cardiovascular death (5) a significant peri-procedural or device-related complication.

A stroke is defined as the sudden onset of a focal neurologic deficit, from a nontraumatic cause, in a location consistent with the territory of a major cerebral artery and categorized as either ischemic, hemorrhagic, or unspecified. An imaging examination (CR, MRI) will be recommended. A TIA is defined according to current standard criteria, i.e., as a stroke with relief of symptoms within 24 hours. Systemic cardioembolic events will be defined as an acute vascular occlusion of an extremity or organ, documented with imaging (CT angiography, percutaneous interventional angiography), surgery, or autopsy.

Clinically significant bleeding will be a composite of major and non-major clinically significant bleeding. Clinically significant major bleeding will be defined according to the criteria of the International Society on Thrombosis and Hemostasis (ISTH), as clinically overt bleeding accompanied by one or more of the following: a decrease in hemoglobin level ≥ 20 g/l over a 24-hour period, transfusion of 2 or more units of packed red cells, bleeding at a critical site (intracranial, intraspinal, intraocular, pericardial, intramuscular with compartment syndrome, or retroperitoneal), or fatal bleeding. (18) This definition is the same as that used recently in all trials comparing NOAC with warfarin. Clinically significant non-major bleeding will be defined similarly as it was in the ROCKET trial, i.e., any bleeding requiring hospitalization or invasive procedure, which doesn´t meet the ISTH major criteria.

Peri-procedural and device-related complications will be evaluated as significant if there is (1) a pericardial effusion requiring drainage, surgery, or pericardiocentesis, (2) a cardioembolic event during the procedure, (3) significant peri-procedural bleeding (i.e., major vascular bleeding requiring surgical revision, or blood concentrate transfusion), (4) device embolization, or (5) a thrombus on the device with a consequent cardioembolic event.

The secondary endpoints will be the incidence of each component of the primary end-point; quality of life (assessed by patients using Health-related quality of life questionnaire, i.e. the questionnaire which was used in the PROTECT study)(19), minor bleeding, a comparison of economic costs, an analysis of changes in cytokine concentrations (natriuretic peptides and other inflammatory and pro-thrombotic cytokines) before the procedure and at 6 months, and the relationship of LAA morphology to success of the LAA occlusion and the risk of systemic cardioembolic events.

Minor bleeding will be defined as any bleeding reported by the patient, which did not meet the criteria for major or clinically significant non-major bleeding. Economical comparisons will involve comparing all costs for treatment in both groups (i.e. drugs, device, and all hospitalizations) for the whole study period. Because the LAA is an important source of natriuretic peptides, and in some patients, the production of these hormones can be eliminated after LAA occlusion, natriuretic peptides and other inflammatory and thrombotic cytokines will be measured before, and 6 months after randomization and the concentrations will be compared.

Statistics For any randomized trial, the power and sample size calculation are critical determinants of the overall success of the trial and the interpretation of findings. The sample size calculation was based on the following assumptions: the power of the test = 0.8, statistical significance border = 0.05. Based on information from large previous trials with NOAC, the investigators assume that 13% of patients in the NOAC group and 10% of patients in the interventional group will achieve the primary endpoint per year. (15) With regard to the border for non-inferiority, the investigators expect 18% of patients to achieve the primary endpoint per year. This means, that for the non-inferiority hypothesis, 198 patients will have to be enrolled in each group, i.e. a total of 396 study participants. Data analysis will be performed based on intention-to-treat principles. For data description, standard descriptive statistical methods will be used: absolute and relative frequencies for categorical data and the median with 5-95% percentiles for continuous data. For categorical variables, statistical analysis will be done using the χ2 or the Fisher exact test; for continuous variables, the Student t-test or Mann-Whitney U test will be used. Kaplan-Meier curves will be calculated for visualizing the occurrence of end-points during follow-up. The influence of patient characteristics on the occurrence of end-points will be calculated using logistic regression and the Cox proportional risk model, when appropriate. Statistical analyses will be done using SPSS 22 software (IBM Corporation, 2013).

DISCUSSION: The incidence of AF increases with the increasing age of the population. In a cohort of 4000 people in the Framingham study, who were followed for more than 30 years, the lifetime risk of the occurrence of AF in a 40 year old man (without previous AF) was 26% and it was 23% for women. (20). The CHA2DS2-VASc score also rises with increasing age, which means the risk of cardioembolic (as well as bleeding) events also increase with age; therefore, the number of patients with a need for, but concurrently with a risk of anticoagulant treatment, increases. The safety profile of NOAC has been shown in large randomized studies to be slightly better compared to warfarin, mainly due to the slightly lower incidence of intracranial bleeding. LAA occlusion, which has been shown to be non-inferior to warfarin in a large study (14), can be a possible alternative in patients at high risk for bleeding. However, LAA occlusion has been associated with a significant number of complications, and therefore, the risk of interventional LAA occlusion is "acceptable" mainly in patients at highest risk for bleeding and stroke.Until now, published and ongoing studies have not focused on this particular group of high-risk patients; for example, an inclusion criterion for the PROTECT trial was only a CHADS2 >1 and 34% of all PROTECT patients had a CHADS2 score of 1.

(14) No studies comparing treatment with NOAC vs. LAA occlusion have been published. Because treatment for these particular patients is very complicated, any progress relative to this very complex set of patients is warranted and much needed.

Inclusion Criteria:

1. history of significant bleeding (i.e. bleeding which required intervention or hospitalization), even in the absence of anticoagulation treatment at the time of the bleeding event, or
2. a cardioembolic event, which occurred on anticoagulation, or
3. a high risk profile of the patient, defined as a CHA2DS2-VASc score ≥ 3 and a HAS-BLED score ≥ 2

Exclusion Criteria:

1. thrombus in the LA or LAA;
2. mechanical valve prosthesis;
3. mitral stenosis;
4. previous LAA ligation during cardiac surgery;
5. life expectancy less than 2 years;
6. comorbidities other than AF, which present an indication for anticoagulation;
7. patent foramen ovale with atrial septal aneurysm
8. mobile plaque in the aorta;
9. symptomatic atherosclerosis of the carotid artery;
10. pericardial effusion greater than 10 mm;
11. clinically significant bleeding within the 30 days prior to the scheduled procedure;
12. stroke or other cardioembolic event within the 30 days prior to the scheduled procedure;
13. acute coronary syndrome within the 90 days prior to the scheduled procedure,
14. gravidity,
15. significant valvular disease,
16. creatinine clearance less than 30 ml/min