International longitudinal registry of patients with atrial fibrillation and treated with rivaroxaban: RIVaroxaban Evaluation in Real life setting (RIVER)

Background Real-world data on non-vitamin K oral anticoagulants (NOACs) are essential in determining whether evidence from randomised controlled clinical trials translate into meaningful clinical benefits for patients in everyday practice. RIVER (RIVaroxaban Evaluation in Real life setting) is an ongoing international, prospective registry of patients with newly diagnosed non-valvular atrial fibrillation (NVAF) and at least one investigator-determined risk factor for stroke who received rivaroxaban as an initial treatment for the prevention of thromboembolic stroke. The aim of this paper is to describe the design of the RIVER registry and baseline characteristics of patients with newly diagnosed NVAF who received rivaroxaban as an initial treatment. Methods and results Between January 2014 and June 2017, RIVER investigators recruited 5072 patients at 309 centres in 17 countries. The aim was to enroll consecutive patients at sites where rivaroxaban was already routinely prescribed for stroke prevention. Each patient is being followed up prospectively for a minimum of 2-years. The registry will capture data on the rate and nature of all thromboembolic events (stroke / systemic embolism), bleeding complications, all-cause mortality and other major cardiovascular events as they occur. Data quality is assured through a combination of remote electronic monitoring and onsite monitoring (including source data verification in 10% of cases). Patients were mostly enrolled by cardiologists (n = 3776, 74.6%), by internal medicine specialists 14.2% (n = 718) and by primary care/general practice physicians 8.2% (n = 417). The mean (SD) age of the population was 69.5 (11.0) years, 44.3% were women. Mean (SD) CHADS2 score was 1.9 (1.2) and CHA2DS2-VASc scores was 3.2 (1.6). Almost all patients (98.5%) were prescribed with once daily dose of rivaroxaban, most commonly 20 mg (76.5%) and 15 mg (20.0%) as their initial treatment; 17.9% of patients received concomitant antiplatelet therapy. Most patients enrolled in RIVER met the recommended threshold for AC therapy (86.6% for 2012 ESC Guidelines, and 79.8% of patients according to 2016 ESC Guidelines). Conclusions The RIVER prospective registry will expand our knowledge of how rivaroxaban is prescribed in everyday practice and whether evidence from clinical trials can be translated to the broader cross-section of patients in the real world. Trial registration Unique identifier: NCT02444221. Registerd 14 May 2015; Retrospectively Registered. Electronic supplementary material The online version of this article (10.1186/s12959-019-0195-7) contains supplementary material, which is available to authorized users.


Introduction
Atrial fibrillation (AF) is the most common sustained arrhythmia reported in adult patients [1] and is associated with an at least a five-fold increase in the risk of stroke [2]. Vitamin K antagonists (VKAs) formed the cornerstone of thromboembolic prophylaxis in patients with AF for many years. The introduction of non-vitamin K anticoagulants (NOACs), dabigatran, rivaroxaban, apixaban and edoxaban, provided physicians with agents with comparative efficacy and reduced potential for bleeding compared with vitamin K antagonists, while removing the need for dose titration, periodic laboratory testing and dietary restrictions that are necessary with VKAs [3][4][5]. Evidence from phase III studies showed that rivaroxaban, one of several oral direct Factor Xa inhibitors, is noninferior to warfarin for the reduction of stroke or systemic embolism in patients with AF [6] and significantly reduces rates of intracranial and fatal haemorrhages, but not rates of bleeding overall. Based on these results, rivaroxaban is now licensed in more than 130 countries worldwide for stroke prevention in patients with AF.
Real-world studies (such as Global Anticoagulant Registry in the FIELD-AF [GARFIELD-AF] [7], ORBIT-AF I & II) [8] and GLORIA-AF [9,10] have demonstrated that NOAC use is increasing while post-marketing surveillance (such as Xar-elto® for Prevention of Stroke in Patients with Atrial Fibrillation [XANTUS]) [11] and national studies (such as the EXPAND) [12,13] have provided data on the safety and efficacy of rivaroxaban. However, most of these registries are disease-specific and evaluate treatment patterns and outcomes across different treatments. Therefore, additional drug-specific observational studies are needed to assess the clinical and health economic risk-benefits of specific NOACs and in particular, the factors that could potentially influence outcomes (e.g. compliance with prescribing guidelines, persistence and comorbidities) for each treatment alternative separately. This paper describes the design of the RIVaroxaban Evaluation in Real life setting (RIVER) registry and baseline characteristics of patients who received rivaroxaban as part of routine care. This paper also includes an initial analysis of the doses of rivaroxaban used in everyday practice as well as a description of the use of rivaroxaban in combination with antiplatelets in different clinical scenarios, such as patients with renal impairment.

Registry design
RIVER (NCT02444221) is an ongoing international, multicentre, prospective registry of patients with newly diagnosed non-valvular AF and at least one investigatordetermined risk factor for stroke. All patients in this study are initiated on an anticoagulant treatment with rivaroxaban at the start of the study, as part of routine care, for the prevention of thromboembolic stroke.
The goal of the RIVER registry is to provide insights into the clinical management and related outcomes of rivaroxaban-treated AF patients across a number of regions and across the spectrum of healthcare systems, in routine clinical management. Follow-up will be for a minimum of 2 years (and up to 3.5 years) after enrolment (Fig. 1). By capturing data from unselected patients treated in everyday practice, the registry has the potential to identify best practices as well as deficiencies in the use of rivaroxaban as an initial treatment for the prevention of thromboembolic events in patients with newly diagnosed AF. The decision to initiate treatment with rivaroxaban was based on clinicians' judgement. The study was designed such that the collection of data would not interfere with the clinical management of patients or with the prescribing behaviours of attending physicians. No specific treatments, tests, or procedures were mandated or withheld from patients, and patients were free to withdraw from the registry at any time. The registry will not only assess the rates of clinical outcomes such as stroke and systemic embolization (SE), all-cause mortality, other major cardiovascular events (Table 1) and bleeding complications, but will also evaluate therapy persistence (including discontinuation, interruption and changes to the treatment regimen). Patients' experiences using antithrombotic treatment will also be evaluated by the Anti-Clot Treatment Scale (ACTS) questionnaire.
Independent ethics committee and hospital-based institutional review board approvals were obtained, as necessary, for the registry protocol. The registry is being conducted in accordance with the Declaration of Helsinki, local regulatory requirements, and the International Conference on Harmonisation-Good Pharmacoepidemiological and Clinical Practice guidelines.
RIVER is an independent academic research initiative, sponsored by the Thrombosis Research Institute, London, UK and funded by an unrestricted research grant from Bayer AG, Berlin, Germany. The authors are responsible for the design and conduct of this study, data analyses and the final content of this paper.

Registry population and site selection
The study population consists of consecutive patients enrolled prospectively at sites where rivaroxaban was already routinely prescribed for stroke prevention before the start of the study. Patients, receiving rivaroxaban as part of usual care, were eligible for enrolment in RIVER if they were 18 years of age or older, had a new diagnosis of non-valvular AF within the past 6 weeks and at least one additional risk factor for stroke as identified by the investigator. These risk factors were identified by clinicians and were not restricted to those in stroke prevention guidelines. Patients with a transient reversible cause of AF, those for whom follow-up is not foreseen or possible and all patients participating in interventional studies were excluded. All patients provided written informed consent to participate. Sites were chosen to be representative of AF care settings and included patients from multiple settings: office-based specialists, hospital departments (neurology, cardiology, geriatrics, internal medicine, and emergency room), anticoagulant clinics and general or family practice settings.

Data capture
Patient visits are not mandated, but data collection using the electronic case report form (eCRF) occurs at 4-monthly intervals and captures all relevant data from the patients' medical notes (Fig. 1).
A summary of the evaluations performed at baseline and at follow-up visits are outlined in Additional file 1: Table S1. At entry into the study, data are collected on the care setting, patients' demographics, vital signs, medical history, the nature of AF (paroxysmal versus persistent versus permanent), symptoms of AF, relevant medications, including antiplatelet and other concomitant therapy and investigator-identified risk factors for stroke. The results from investigations at diagnosis (including ECG morphology, left ventricular ejection fraction [LVEF] measurement, full blood count, haemoglobin, platelet count and creatinine) are recorded as well as the treatment strategy for AF (rate or rhythm control) and the date and outcome following cardioversion. Information is collected on initial rivaroxaban dosing regimen, start and stop dates, changes in therapy, compliance concerns and the reason for suspending or terminating therapy sooner than intended (such as bleeding, patient decision, and/or physician decision). Patients' experiences using antithrombotic treatment are recorded on the Anti-Clot Treatment Scale (ACTS) questionnaire at 4, 12 and 24 months. Based on the data collected from the eCRF, healthcare resource consumption will be captured so that the economic burden of AF can be computed both overall and per patient per year from the perspective of the payer, e.g. national health service, public/private/statutory insurance etc.

Data management
Data are captured using an electronic case report form (eCRF) designed by eClinicalHealth Services, Stirling, UK and submitted electronically via a secure website to the registry-coordinating centre at the Thrombosis Research Institute who were responsible for checking the completeness and accuracy of data collected from medical records. All patients are assigned a unique identifier and personal identifiable data will be removed at the hospital source, ensuring anonymity and protecting confidentiality.
Data quality is assured through a combination of remote electronic monitoring and more conventional onsite monitoring (including source data verification in 10% of all cases), based on the validated quality assurance programmes developed in GARFIELD-AF [14]. The milestones for study eCRFs will be examined by the registry-coordinating centre at TRI, to ascertain completeness, accuracy and data queries sent to participating sites.

Statistical analysis
Continuous variables are expressed as mean ± standard deviation (SD). Categorical variables are expressed as frequencies and percentages. Statistical analysis was performed using SAS software version 9.4 (SAS Institute Inc., Cary, NC, USA).

Concomitant antiplatelet therapy
Just under one in five (17.9%, n = 904) patients who were initiated on rivaroxaban at the time of diagnosis of AF also received concomitant antiplatelet therapy. Use of antiplatelets was higher in patients with a CHA 2 DS 2 -VASc score of 2-9 (19.4%) than in those with a score of 0 or 1 (9.2%) ( Table 3); 229 (25.3%) had a coronary artery disease (CAD), peripheral artery disease (PAD) or carotid disease. The most frequently used antiplatelet was aspirin in 55.4%

Discussion
Observational or non-interventional studies play an important role in understanding and confirming the clinical effectiveness of interventions across the spectrum patients treated in routine practice.
RIVER is an independent treatment registry of patients from multiple care settings and patients were recruited from 17 countries. The RIVER registry provides an opportunity to examine the impact of different recruitment methodologies on the characteristics and treatment between the registries. RIVER will document the regional heterogeneity in the clinical presentation and assess the rate of stroke and systemic embolism and other outcomes with specific reference to the incidences of bleeding and therapy persistence. In addition to reporting on pre-specified clinical and economic outcomes, analyses from RIVER will be hypothesis generating, allowing the exploration of some aspects of the real world situation of AF treated with rivaroxaban and outcomes.
The degree of the patient involvement in the generation of data is expected to provide unique insights into the patients-reported outcomes. Patients' experiences using antithrombotic treatment are recorded on the Anti-Clot Treatment Scale (ACTS) questionnaire. It is important to recognise that registries differ in their design, recruitment strategies, care settings, geographical spread and duration of follow-up. Compared to other ongoing prospective registries, the RIVER registry has the potential to capture the medicosocial burden due to AF in large-scale populations by employing broad inclusion and exclusion criteria in widely representative populations of patients with AF who are treated with rivaroxaban, across a range of clinical settings and to capture long-term follow-up data. The value of RIVER registry is enhanced by a quality assurance programme and the supervision of an independent Audit Committee, which oversees site-dependent verification, remote site monitoring and electronic database monitoring to ensure data quality.
In the presence of all these observational studies, it is important to highlight that there are a number of important differences in the design (Additional file 1: Table  S2) and recruitment of patients (Table 4) in RIVER compared with these registries. RIVER is a drug-specific registry which allows studying of drug utilisation patterns including off-label use whereas other registries like GARFIELD-AF, ORBIT-AF I and II, GLORIA-AF and Dresden NOAC are disease-specific registries which allows studying of burden, management and adverse event profile for different treatments related to a specific disease. Patients in XANTUS study were enrolled from Europe, Israel and Canada whereas EXPAND and XAPASS are Japanese registries and ORBIT-AF registries are restricted to USA [19]. By contrast, RIVER has recruited patients more broadly from 17 countries (outside Asia) including Europe, the Middle East, Latin America, Canada and Australia. RIVER provides long-term follow-up data of 2 years where as XANTUS had 1-year follow-up data. RIVER is a more contemporary population in which physicians' now have experience using the NOACs. It is a known phenomenon that, after approval of new treatments, physicians start to apply these first in well-defined, carefully selected populations and expand it to other populations only after a period of time, which is reflecting learning curves based on increased comfort levels, communication of real world evidence, newly reported and previously unknown side effects or changes in guideline recommendations [28][29][30] . As a consequence, treatment patterns and outcomes may shift over time, which is an additional justification for more  , suggesting differences in the guideline recommendations, which are based on patient characteristics and ethnicities. In Japan, the regular dosage of rivaroxaban is 15 mg od, which is lower than the global recommended dosage of 20 mg od [33].

Limitations
Centres for recruitment of patients in RIVER are limited to the care settings where rivaroxaban is the standard of care and includes sites beyond those included in the clinical trials. Comparison of different treatments are not possible, as there is no comparator drug, while larger prospective disease registries such as GARFIELD-AF and drug registries such as GLORIA-AF and ORBIT-II have the potential to conduct comparative effectiveness studies between different treatment approaches in newly diagnosed patients with AF. Due to the observational design of the study, laboratory and other investigations could not mandated.

Strengths
RIVER is an independent academic research initiative. This prospective registry was designed with a novel approach to outcomes research, including the following features: sites representation of national AF care settings, randomised site selection, unselected eligible patients enrolled consecutively with a follow-up period of minimum 2 years and extensive monitoring and audit including source data verification of 10% of all eCRFs. Like other prospective registries, such as GARFIELD-AF [34], RIVER aims to overcome the inherent bias of retrospective cohort studies from national registries and claims databases [35][36][37][38][39].

Conclusion
The RIVER prospective registry will expand our knowledge of how rivaroxaban is prescribed in everyday practice and whether evidence from clinical trials can be translated to the broader cross-section of patients treated in the real world.

Additional file
Additional file 1: Table S1. Evaluations performed at Baseline and at Follow-Up Visits. Table S2. Study design of ongoing real-world studies of rivaroxaban.