Routine clinical practice in the periprocedural management of edoxaban therapy is associated with low risk of bleeding and thromboembolic complications: The prospective, observational, and multinational EMIT‐AF/VTE study

Abstract Background Guidance for periprocedural anticoagulant management is mainly based on limited data from Phase III or observational studies and expert opinion. Hypothesis EMIT‐AF/VTE was designed to document the risks of bleeding and thromboembolic events in more than 1000 patients on edoxaban undergoing diagnostic and therapeutic procedures in clinical practice. Methods Routine care in a multinational multicenter, prospective observational study. Participants were adult patients with atrial fibrillation and/or venous thromboembolism treated with edoxaban for stroke prevention or for secondary prevention in venous thromboembolic disease, undergoing a wide range of diagnostic and therapeutic procedures. Edoxaban therapy was interrupted periprocedurally at the treating physician's discretion. Patients were evaluated from 5 days pre‐ until 30 days postprocedure. Primary outcome was the incidence of International Society on Thrombosis and Haemostasis defined major bleeding; secondary outcomes included incidence of clinically relevant non‐major bleeding, acute coronary syndrome, and acute thromboembolic events. Results Outcomes and management are reported for the first procedures in 1155 unselected patients. Five cases of major bleeding (0.4%) and eight of clinically relevant non‐major bleeding (0.7%) were documented, five (38%) of which occurred outside the period of likely edoxaban effect (last edoxaban dose ≥3 days prior to bleeding). Five (0.4%) deaths from any cause, seven acute thromboembolic events (0.6%) including two cardiac deaths (0.2%) in six patients, and one acute coronary event (0.1%) occurred. Conclusions The periprocedural bleeding and acute thromboembolic event risks for patients treated with edoxaban were low. This can help inform both clinical routine and guidelines for the periprocedural management of edoxaban.


| INTRODUCTION
Non-vitamin K dependent oral anticoagulants (NOAC) have become the preferred and recommended long-term therapy 1,2 for stroke prevention in patients with atrial fibrillation (AF) and for the treatment and secondary prevention of venous thromboembolism (VTE). Compared to vitamin K antagonists (VKA), NOACs exhibit a rapid onset of action and a shorter effect duration, so a shorter interruption, if any, prior to invasive procedures might be expected. At least 10% of these patients require diagnostic and therapeutic procedures each year. 3 Guidance for clinicians is mainly based on pharmacokinetic data, Phase III studies involving cardiovascular procedures such as PCI and AF ablation, 4-6 expert opinion or guidelines (eg, from the European Heart Rhythm Association [EHRA] 7 or other specialist societies 8,9 ) and practical questions remain.
The periprocedural management of NOAC treatment focuses on reducing the risk of bleeding, without complicating the procedure with an increased risk of thromboembolic events. Current evidence suggests that if the bleeding risk is low, interruption of anticoagulation therapy may not be necessary. 10 However, this has mainly been studied in patients treated with VKA.
A recently published 11 prospective study using a predefined periprocedural management protocol for apixaban, dabigatran and rivaroxaban, found event rates for major bleeding (MB) at 30 days lower than the prespecified 2% for all three NOACs, whereas a different strategy resulted in low major bleeding rates in patients treated with dabigatran.
Previous studies with edoxaban have demonstrated noninferiority to warfarin for stroke prevention in AF and for VTE treatment/secondary prevention of thromboembolic events and shown a reduced MB or clinically relevant non-major bleeding (CRNMB) incidence, and in particular a lower incidence of potentially life-threatening intracranial bleeding. 12,13 Currently, however, periprocedural management data for edoxaban are only available from post hoc analyses of recent studies 14,15 Important questions remain on time and modalities of NOAC interruption, as there is minimal information to decide if current recommendations provide the optimal balance of bleeding and thrombo-embolic risk. 16 This first prospective study was designed to collect information about the periprocedural management of patients receiving edoxaban, and to document bleeding and thromboembolic events in a large unselected sample of more than 1000 patients on long-term edoxaban therapy.

| METHODS
A description of the objectives and design of the study has been previously published. 17 The study was a multicenter, prospective, observational, study conducted in Europe (NCT02950168) in accordance with the Declaration of Helsinki and with local IRB approvals. Written informed consent was obtained from participants prior to enrolment.
The periprocedural management of anticoagulant therapy was at the discretion of the investigator, including any decision whether to interrupt edoxaban therapy and the timing/duration of any interruption of edoxaban treatment. The site investigator determined the edoxaban patient management approach based upon individual patient findings, in the context of published guidelines and professional experience: no attempt was made to influence patient management by the study authors, study team or the sponsor.

| Patient recruitment
Enrolment commenced in December 2016 and was complete in July 2018. Patients were recruited from 326 centers from Belgium, Germany, Italy, Netherlands, Portugal, Spain, and the United Kingdom. Eligible patients were ≥18 years of age, had AF or VTE, were treated with edoxaban according to the local labels, were not enrolled in any other study concurrently, and underwent any type of diagnostic or therapeutic procedure. Data from the first procedure in each patient is reported.

| Observations
The observation period of the study started 5 days before the procedure and ended 30 days afterwards. To enhance data capture, patients received a memory aid booklet at enrolment into the study, which was reviewed at the end of the study. Date of visit, details of edoxaban treatment, and clinical outcomes were documented at 30 days after each procedure. No interruption of edoxaban therapy was defined if edoxaban was administered on each day of the observation period including at any time on the day of the procedure. Any interruption of edoxaban treatment was recorded as the number of days without administration of edoxaban, preprocedural and/or postprocedural. Any dose skipped before or on the day of the procedure was defined as preprocedural. Day one of follow-up was defined as the day of procedure.

| Statistical analysis
The study planned to document about 1415 procedures. This estimate was based on an assumed incidence of MB at 30 days of 1.5%, based on published data from the Dresden NOAC registry 20 which reported a 1.2% MB incidence during a 30 ± 5 day postprocedural follow up period, and another cohort study which identified a 1.8% MB incidence during a 30-day follow-up. 21 Assuming a 5% dropout rate, 1415 procedures were expected to allow for a reasonable estimation of the incidence of MB, with a precision of the 95% confidence interval of the event rate of 0.65%.
All clinical data was documented using the Medidata Rave Electronic Data Capture system (Medidata Solutions Inc., New York, New York). Binary, categorical, and ordinal parameters were summarized by means of absolute and percentage numbers. Numerical data were described by standard statistics. The statistical analyses were performed using SAS version 9.4 (SAS Institute, Cary, North Carolina).
All analyses were descriptive and exploratory.

| RESULTS
Informed consent for the EMIT-AF/VTE study was obtained from 1284 subjects, from whom data on 1441 procedures was collected. In 109 patients early enrolment followed by delay or cancellation of the index procedure led to missing inclusion criteria or incomplete documentation. Therefore, there were 120 procedures less than the planned 1415; this reduction in numbers of procedures changed the precision of the 95% confidence interval of the event rate from 0.65% to 0.70%. The evaluable data set contains complete subject baseline and complete periprocedural data from the first procedure for 1155 patients. Of 1155 first procedures, 1100 (95.2%) were prescheduled, 50 (4.3%) unplanned (emergency/urgent), and 5 (0.4%) unknown. One hundred and twenty-eight second or subsequent procedures were performed in these 1155 patients; data from these procedures will be reported separately. The derivation of this data set is summarized in Clinical characteristics and patient data: all subjects and grouped by EHRA procedural bleeding risk  Figure S1 details all procedural groups, including those procedural groups that were less than 10% of the total number of procedures, and the number of complications in each procedural group.

| Patient and procedure characteristics
In line with studies on other NOACs, 10 (Table S1).

| Outcomes
We report on the risks of bleeding and thromboembolic events from the first prospective observational multicenter multinational study on periprocedural management and outcomes of 1155 unselected procedures in patients managed with edoxaban therapy. A wide range of medical specialties was procedurally represented, and 23 clinically relevant complications were reported. Baseline characteristics were similar to those from other non-observational studies. 10,11,20 This investigation complements the PAUSE study, 11

| Comparison with published data
The first data on the periprocedural outcomes of NOAC were derived as sub-analyses from the large Phase III studies of these compounds. In a sub-study of the ARISTOTLE trial, 10  A sub-analysis of the RE-LY study 22 found an incidence of major peri- for example, included a higher percentage of high bleeding-risk procedures (33.5%) than EMIT (24.2%), but only EMIT included emergency procedures.
As EMIT is a study of unselected patients taking edoxaban, VTE was the indication for edoxaban administration in 8.6% of the patients. However, the low incidence of events did not allow us to distinguish between AF and VTE patients. PAUSE 11 was a pure AF patient trial that excluded patients with VTE as the indication for oral anticoagulation.
Although there are design differences between EMIT, DRESDEN, 20 and PAUSE, 11 all three studies suggest to physicians that for the majority of patients short-term interruption is the preferred approach in periprocedural anticoagulation management in order to minimize hemorrhagic or thromboembolic complications.

| Limitations of study
One limitation of this study is lack of a comparator arm and its restriction to one NOAC, since differences exist between the various NOACs. Further, no attempt was made to evaluate the impact of geographical location or ethnicity.
However, given the plethora of diagnostic and interventional procedures, the conduct of a global study/registry that would address the above limitations is logistically and economically not feasible. The EMIT data set, being unique in capturing observational rather than prescriptive management of patients on edoxaban in real life setting, complements the data from DRESDEN 20 and PAUSE. 11 The three studies together indicate that oral anticoagulation, when managed carefully, is associated with a low incidence of major hemorrhagic and significant ischemic events.
Observational studies often do not provide the same level of data quality as controlled studies. However, patients were provided with memory aids to support data collection (which were reviewed for completeness at the end of the procedural follow-up), critical data elements were reviewed at the patient level, and all critical events were centrally adjudicated. Offsetting any loss in quality because of missing data or lack of mechanisms to support compliance is the ability of an observational study to accurately reflect current clinical practice without external influence. Rather than adhering to a study protocol defined treatment regimen, physicians were free to follow their understanding of best practices for a given patient.

| CONCLUSIONS
In patients taking edoxaban for an approved indication, the periprocedural occurrence of MB and CRNMB as safety measures, and ATE and ACS as markers of efficacy, were low in the overall study and all patient subgroups. EMIT AF/VTE documents periprocedural edoxaban management, with short interruption times as a safe and effective approach that prevents adverse outcomes.
Further study is required to better understand the risk factors in those few patients who develop bleeding or ischemic complications.