Abstract
To estimate major clinical event rates for patients with atrial fibrillation (AF) and atherosclerotic disease treated with edoxaban in routine practice, and to evaluate how well such patients were represented in ENGAGE AF-TIMI 48, the seminal randomized trial comparing edoxaban against warfarin for AF. ETNA-AF-Europe is a prospective cohort of AF patients receiving edoxaban in routine care. We compared patients with coronary or peripheral artery disease (CAD/PAD) to: (1) those without CAD/PAD in ETNA-AF-Europe, and (2) CAD/PAD patients in ENGAGE AF-TIMI 48. Of 13,164 patients in ETNA-AF-Europe, 23.3% had CAD/PAD. Compared with those without, patients with CAD/PAD had higher rates of stroke/systemic embolism (0.87%/year vs. 0.59%/year; HR 1.5, 95%-CI 1.14–1.88), acute coronary syndrome (1.24%/year vs. 0.37%/year; HR 3.3, 95%-CI 2.60–4.27), major bleeding (1.06%/year vs. 0.81%/year; HR 1.3, 95%-CI 1.04–1.63), cardiovascular death (1.59%/year vs. 0.85%/year; HR 1.9, 95%-CI 1.54–2.26), and all-cause death (6.02%/year vs. 3.53%/year; HR 1.7, 95%-CI 1.55–1.89). Compared with CAD/PAD patients in ENGAGE-AF TIMI-48, those in ETNA-AF-Europe had fewer cardiovascular comorbidities, less prevalent aspirin use (20.2% vs. 50.3%), and lower rates of stroke/systemic embolism (0.87%/year vs. 1.5%/year), major bleeding (1.04%/year vs. 3.0%/year), and cardiovascular death (1.59%/year vs. 3.7%), but higher non-cardiovascular mortality (4.43%/year vs. 1.6%/year). In routine practice, deaths and bleeding were the most common events in edoxaban-treated patients with AF. This pattern was consistent between those with and without atherosclerosis. ENGAGE-AF TIMI-48 participants with CAD/PAD had substantially higher cardiovascular but lower non-cardiovascular risks than those treated in daily practice.
Trial registration number: NCT02944019 (ClinicalTrials.gov Identifier).
Graphical Abstract
Supplementary Information
The online version contains supplementary material available at 10.1007/s11239-026-03269-5.
Keywords: Clinical Trial, Phase IV; Coronary Artery Disease; Edoxaban; Hemorrhage; Peripheral Arterial Disease; Thromboembolism
Key points
What are the key findings?
Patients with atrial fibrillation and concomitant atherosclerotic disease treated in European daily practice have substantially lower cardiovascular yet higher non-cardiovascular risks than similar patients treated in the seminal randomized trial.
In routine practice, deaths—particularly those of non-cardiovascular causes—and bleeding were the most common event in edoxaban-treated patients with atrial fibrillation. This pattern was consistent between those with and without atherosclerosis.
What are the implications for future directions?
Our findings support the ongoing shift toward:
Better identification of patients who will develop thrombosis or bleeding
Reducing major and recurrent bleeding
Even though ENGAGE AF-TIMI 48’s population is under-representative of some patients treated in daily practice, we believe the implications are limited—since both the rates of thrombosis and bleeding associated with edoxaban were low. We believe two additional factors may have contributed to the low rates in routine practice:
Improved management of cardiovascular risk factors
Tailoring antithrombotic strategies to higher-risk individuals
Introduction
Coronary artery disease (CAD) and peripheral artery disease (PAD) are common in patients with atrial fibrillation (AF) [1, 2]. Together or in isolation, these conditions raise thrombotic risk, though the magnitude varies with clinical context [2–10]. Many of these patients also have conditions that increase bleeding risk [2, 7, 8]. Finding the optimum with antithrombotic treatment is therefore inherently complex.
Edoxaban is an antithrombotic that directly inhibits activated factor X. It has been evaluated in patients with AF across different stages of atherosclerotic disease in four randomized trials [3–5, 10]. Combined these trials showed that in patients with AF: (1) edoxaban is at least equally efficacious as warfarin at preventing thrombotic events but has the added benefit of causing ≥ 20% fewer major bleeding events [10]; (2) combining edoxaban with clopidogrel results in similar bleeding risks as a triple therapy regimen consisting of warfarin, clopidogrel, and aspirin in those who recently underwent a successful percutaneous coronary intervention [4]; (3) edoxaban monotherapy leads to 60–70% fewer bleeding events compared with dual pathway inhibition consisting of edoxaban and clopidogrel in those with stable CAD [3, 5].
The key limitation of the three trials in patients with established atherosclerosis (ENTRUST-AF PCI, EPIC-CAD, and PRAEDO AF) is imprecision: they lacked sufficient power to exclude clinically meaningful differences in thrombotic events between regimens [3–5]. Similarly, in ENGAGE AF–TIMI 48, only 20% of participants had CAD or PAD, leaving the subgroup too small to assess treatment-effect heterogeneity reliably [7, 8, 10]. In addition, all trials applied strict eligibility criteria and intensive follow-up, which limits generalizability to higher-risk patients encountered in routine practice [3–5, 7, 8, 10].
Based on these considerations, we conducted this post-hoc sub-study of a single-arm, prospective cohort study to address three key questions: First, what are the annual rates of major cardiovascular and death events in patients with AF and atherosclerotic disease treated with edoxaban in European routine practice? Second, how do these rates compare to those in AF patients without atherosclerotic disease? Finally, how well do the trial participants with atherosclerosis reflect those seen in everyday European practice?
Methods
Study design
In this study, we used Edoxaban Treatment in RoutiNe Clinical PrActice for Patients with Non-Valvular AF in Europe (ETNA-AF-Europe), a single-arm, non-interventional, multinational, prospective cohort study on more than 13,000 patients treated at least once with edoxaban. The details on the rationale and design of ETNA-AF-Europe (NCT02944019) have been described previously [11]. We selected ENGAGE AF–TIMI 48 as the comparator because, among trials of patients with AF and atherosclerotic disease, its study population most closely resembled that of ETNA-AF-Europe [11]. Our STrengthening the Reporting of OBservational studies in Epidemiology (STROBE) checklist is provided in Table S1 [12].
Study participants, period, setting, and sample size
In ETNA-AF-Europe, consecutive patients at family medicine offices or hospital departments (i.e., cardiology, neurology, neurosurgery, geriatric, and internal medicine departments and outpatient clinics) at 850 sites in ten European countries were invited to participate. In- or outpatients with an ongoing prescription of edoxaban were eligible for inclusion if AF was (re-)confirmed by electrocardiographical tracing (i.e. electrocardiogram, Holter monitoring, pacemaker, or a different implantable device) within the 12 months prior to enrolment. Non-adults (i.e., < 18 years of age) and those participating in an interventional study were excluded. The period of enrolment was from November 2016 to February 2018. We did not perform a power analysis for this sub-study.
Ethics approval and consent
ETNA-AF-Europe is performed in accordance with the Declaration of Helsinki. Its study protocol approvals were provided by competent authorities as well as by central and local ethics committees for each participating site, in all countries involved. All participants provided written informed consent.
Data collection and definitions of variables
Patients were followed up for 4 years; or for at least 2 years after discontinuing edoxaban, until withdrawal of consent, death, transfer to another institution, or loss to follow-up. Data were collected at baseline and yearly thereafter (± 2 months).
ETNA-AF-Europe was designed to mirror routine clinical practice. For most conditions, no rigid definition was provided. Instead, the treating clinicians judged whether clinical conditions like CAD and PAD were present. Errors in data collection were corrected by applying triangulation. Other baseline characteristics included a modified CHA2DS2-VASc (not including complex vascular plaque and a modified HAS-BLED score (alcohol use was defined as ≥ 1 unit per day, and determination of renal or hepatic disease was left to the discretion of the treating clinician). We defined off-label dosing of edoxaban in accordance with the Summary of Product Characteristics, that is, 30 mg once daily in patients without a dose reduction criterion (i.e., off-label reduced dose) and 60 mg once daily in those with one or more criteria (i.e., off-label standard dose) [13].
Thrombotic outcomes of interest were the composite of all stroke and systemic embolism, ischemic stroke, and acute coronary syndromes, which includes ST-elevation myocardial infarction (STEMI), non-ST-elevation myocardial infarction (NSTEMI), and unstable angina. Bleeding outcomes were the composite of International Society on Thrombosis and Haemostasis (ISTH) major bleeding and ISTH clinically relevant nonmajor (CRNM) bleeding [14], both event types in isolation, intracranial hemorrhage, hemorrhagic stroke, and major gastro-intestinal bleeding. Deaths were classified as all-cause, cardiovascular, or non-cardiovascular. We defined cardiovascular death as any death where: (a) the cause was adjudicated as cardiovascular; or (b) its cause was reported as unknown or was missing, but was preceded by a stroke, transient ischemic attack, systematic embolism, myocardial infarction, venous thromboembolism, or major bleeding within 30 days. All major bleeding, clinically relevant non-major bleeding, strokes, systemic embolisms, and deaths were adjudicated by an independent committee.
Statistical analysis
To provide estimates of the annual rates of cardiovascular events in patients with and without atherosclerotic disease in routine clinical practice, we first divided the ETNA-AF-Europe cohort into two subgroups by history of CAD/PAD. For both subgroups, we summarized the clinical characteristics and determined the events per 100 patient-years (hereafter referred to as %/year) of major cardiovascular events and deaths, including their 95% confidence intervals (CIs). We directly compared the two subgroups for each clinical outcome using univariable Cox proportional hazard models. We also summarized the baseline clinical characteristics and estimated annual event rates for those with and without CAD/PAD by edoxaban dose received. To assess potential sampling bias, we compared the clinical characteristics of patients included in our final dataset with those of excluded patients.
To assess how well patients with atherosclerotic disease in ENGAGE AF-TIMI 48 reflect those in ETNA-AF-Europe, we conducted two separate analyses using previously published data from ENGAGE AF-TIMI 48 as the comparator [7, 8]. For patients with CAD, we put the baseline characteristics and annual event rates of the two studies into perspective [8]. For patients with PAD, however, we limited our analysis to baseline characteristics only, as the low number of events in ENGAGE AF-TIMI 48 for the approved edoxaban dosing regimens (60 mg once daily or 30 mg once daily for eligible patients) precluded meaningful comparisons of event rates [7].
In all analyses, we used case deletion and pairwise deletion for mandatory (e.g. dose of edoxaban) and non-mandatory (e.g. serum creatinine) missing data, respectively. Baseline characteristics were summarized using summary statistics appropriate for the expected distribution of the data. We did not adjust for multiple comparisons; therefore, our results should be interpreted as exploratory. All statistical analyses were performed using the Statistical Analysis System (SAS) version 9.3 or higher (SAS Institute, Cary, North Carolina, USA).
Results
Patient selection
Between November 2016 and February 2018, 13,632 patients provided informed consent, 468 (3.4%) of whom were excluded from this sub-study. Of the 13,164 included patients, 1,701 (12.9%) prematurely discontinued participation for other reasons than death (12.9% of those without vs. 13.0% with CAD/PAD). The details on the flow of participants are provided in Fig. 1.
Fig. 1.
Flow of ETNA-AF-Europe participants. The patients in the green box form our sample population used in the analyses on patients with vs. without CAD/PAD in ETNA-AF-Europe, and those in the yellow boxes those used in our analyses on patients with either CAD or PAD in ETNA-AF-Europe vs. in ENGAGE AF-TIMI 48. CAD, coronary artery disease; PAD, peripheral artery disease. a Patients could have multiple reasons for their exclusion from the final dataset
With vs. without atherosclerotic disease in daily clinical practice
Baseline characteristics
Of 13,164 patients in ETNA-AF-Europe, 23.3% had CAD/PAD. Patients with CAD/PAD were older (mean age: 75.2 vs. 73.2 years), more often male (71.0% vs. 52.3%) and considered frail (13.3% vs. 9.9%) than those without. Cardiovascular comorbidities were more prevalent in the patients with CAD/PAD, resulting in a higher mean modified CHA2DS2-VASc score (3.8 vs. 3.0). They were also more likely to receive 30 mg of edoxaban (30.1% vs. 21.0%) and antiplatelet therapy on top of edoxaban (19.9% vs. 2.8%). More details are provided in Table 1.
Table 1.
Baseline clinical characteristics of patients treated in daily practice by atherosclerotic diseases
| With CAD/PAD (n = 3,069, 23.3%) |
Without CAD/PAD (n = 10,095, 76.7%) |
|
|---|---|---|
| Edoxaban 30 mg, n (%) | 925 (30.1%) | 2,117 (21.0%) |
| Off-label dosing of edoxaban, n (%) | 597 (19.5%) | 1,424 (14.1%) |
| - Of the reduced dose, n (%) | 283 (9.2%) | 536 (5.3%) |
| - Of the standard dose, n (%) | 314 (10.2%) | 888 (8.8%) |
| Ongoing use of antiplatelet therapy, n (%) | 611 (19.9%) | 284 (2.8%) |
| - Aspirin | 445 (14.5%) | 229 (2.3%) |
| - Clopidogrel | 210 (6.8%) | 50 (0.5%) |
| - Ticagrelor | 10 (0.3%) | 2 (< 0.1%) |
| Male, n (%) | 2,180 (71.0%) | 5,281 (52.3%) |
| Age [years], mean ± SD | 75.2 ± 8.5 | 73.2 ± 9.7 |
| - < 65, n (%) | 329 (10.7%) | 1,671 (16.6%) |
| - ≥ 65, < 75, n (%) | 968 (31.5%) | 3,490 (34.6%) |
| - ≥ 75, < 85, n (%) | 1,393 (45.4%) | 5,324 (52.7%) |
| - ≥ 85, n (%) | 379 (12.3%) | 1,003 (9.9%) |
| Current smoker, n (%) | 230 (7.5%) | 593 (5.9%) |
| BMI [kg/m2], mean ± SD | 28.0 ± 4.7 | 28.1 ± 5.2 |
| Creatinine clearance [mL/min],a mean ± SD | 68.2 ± 25.7 | 73.0 ± 25.9 |
| - < 30 mL/min, n (%) | 99 (3.2%) | 199 (2.0%) |
| - ≥ 30, < 50 mL/min, n (%) | 606 (19.7%) | 1,550 (15.4%) |
| - ≥ 50, < 80 mL/min, n (%) | 1,235 (40.2%) | 3,855 (38.2%) |
| - ≥ 80 mL/min, n (%) | 834 (27.2%) | 3,366 (33.3%) |
| - Missing/unknown, n (%) | 295 (9.6%) | 1,125 (11.1%) |
| Modified CHA2DS2-VASc, mean ± SD | 3.8 ± 1.5 | 3.0 ± 1.4 |
| - Low risk (0 for men, 1 for women), n (%) | 24 (0.8%) | 343 (3.4%) |
| - Intermediate risk (1 for men, 2 for women), n (%) | 169 (5.5%) | 1,630 (16.1%) |
| - High risk (≥ 2 for men, > 2 for women), n (%) | 2,779 (90.6%) | 7,897 (78.2%) |
| - Missing/unknown, n (%) | 97 (3.2%) | 225 (2.2%) |
| Modified HAS-BLED, mean ± SD | 2.8 ± 0.9 | 2.4 ± 1.0 |
| - Low risk (< 2), n (%) | 218 (7.1%) | 1,458 (14.4%) |
| - Intermediate risk (2 or 3), n (%) | 1,839 (59.9%) | 6,187 (61.3%) |
| - High risk (> 3), n (%) | 567 (18.5%) | 822 (8.1%) |
| - Missing/unknown, n (%) | 445 (14.5%) | 1,628 (16.1%) |
| Frail, n (%) | 407 (13.3%) | 1,003 (9.9%) |
| - Missing/unknown, n (%) | 196 (6.4%) | 684 (6.8%) |
| Coronary artery disease, n (%) | 2,860 (93.2%) | 0 |
| Unstable angina pectoris, n (%) | 199 (6.5%) | 0 |
| Myocardial infarction, n (%) | 567 (18.5%) | 0 |
| - With intervention, n (%) | 437 (14.2%) | 0 |
| Peripheral artery disease, n (%) | 432 (14.1%) | 0 |
| - Claudication at ≤ 200 m, n (%) | 301 (9.8%) | 0 |
| - Rest pain, n (%) | 69 (2.2%) | 0 |
| - Necrosis or gangrene of the limb, n (%) | 30 (1.0%) | 0 |
| Hypertension, n (%) | 2,564 (83.5%) | 7,591 (75.2%) |
| Congestive heart failure, n (%) | 994 (32.4%) | 1,048 (10.4%) |
| Valvular disease,b n (%) | 550 (17.9%) | 1,320 (13.1%) |
| Diabetes mellitus, n (%) | 963 (31.4%) | 1,918 (19.0%) |
| History of any stroke or transient ischemic attack, n (%) | 312 (10.2%) | 867 (8.6%) |
| History of major bleeding, n (%) | 42 (1.4%) | 76 (0.8%) |
| Intracranial, n (%) | 18 (0.6%) | 44 (0.4%) |
| Gastrointestinal, n (%) | 14 (0.5%) | 21 (0.2%) |
CAD, coronary artery disease; BMI, body mass index; PAD, peripheral artery disease; SD, standard deviation. a Values outside of the 5 − 150 mL/min range were considered missing; b Valvular heart disease was defined as at least moderate aortic or mitral regurgitation, aortic stenosis, or prior valve surgery (i.e., bioprosthesis replacement, valve repair, or valvuloplasty)
The differences between patients with and without CAD/PAD were directionally consistent between the two edoxaban doses. Of the four subgroups, those with CAD/PAD on 30 mg had the worst risk profiles. Table S2 provides full details.
Off-label dosing status was unknown for 1,910 patients (14.5%), largely because creatinine clearance was missing (10.8%). The prevalence of CAD and PAD was similar for excluded and included patients (20.3% vs. 21.7% and 3.4% vs. 3.3%, respectively), but excluded patients less often received the 30 mg dose (18.4% vs. 23.1%) or were dosed off-label (9.0% vs. 15.4%). Table S3 lists missing variables and Table S4 provides the full characteristics of excluded patients.
Clinical outcomes
Median follow-up was 1,452 days [interquartile interval 1,171–1,478 days] and was similar for patients with and without CAD/PAD.
Thrombotic event rates were low in both subgroups but were consistently higher in those with atherosclerotic disease. The annual rate of the composite outcome of stroke and systemic embolism was 50% higher in patients with CAD/PAD (0.87%/year vs. 0.59%/year; HR 1.5, 95%-CI 1.14–1.88). The difference was most pronounced for acute coronary events, which occurred at a threefold higher rate in those with concomitant disease (0.75%/year vs. 0.24%/year; HR 3.3, 95%-CI 2.60–4.27). Bleeding events were also uncommon in both subgroups but occurred more frequently in patients with CAD/PAD. The annual rate of major bleeding was 1.50%/year for patients with CAD/PAD compared with 1.06%/year for those without (HR 1.3, 95%-CI 1.04–1.63), with gastrointestinal bleeding accounting for about half of these events (0.52%/year vs. 0.35%/year; HR 1.50, 95%-CI 1.07–2.05). Intracranial bleeding was rare for both groups (0.22%/year vs. 0.17%/year; HR 1.3, 95%-CI 0.79–2.10). The most common event was all-cause death in either subgroup (6.02%/year vs. 3.53%/year; HR 1.7, 95%-CI 1.55–1.89), with most due to non-cardiovascular causes (4.43%/year vs. 2.67%/year; HR 1.7, 95%-CI 1.49–1.86). More details are provided in Fig. 2.
Fig. 2.
Major cardiovascular events in patients treated in daily practice by atherosclerotic disease status. Rates are in events per 100 patient-years (95% confidence intervals). CAD, coronary artery disease; CRNM, clinically relevant non major; NSTEMI, non-ST-elevation myocardial infarction; PAD, peripheral artery disease; STEMI, ST-elevation myocardial infarction
Bleeding rates were similar for patients with and without CAD/PAD on 30 mg edoxaban. For the other outcomes, the subgroup differences were directionally consistent between the two edoxaban doses (Fig. S1).
Daily practice vs. randomized trial
Coronary artery disease
The prevalence of CAD was near-identical between ETNA-AF-Europe and ENGAGE AF-TIMI 48 (21.7% vs. 21.0%). Patients treated in routine clinical practice were older than those in the randomized trial (mean age: 75.2 vs. 72.1 years) and they more often received 30 mg edoxaban than trial participants (30.4% vs. 24.5%), with a third of cases in daily practice receiving this dose off-label. Aspirin use was lower in routine practice (20.2% vs. 50.3%), as were rates of hypertension, heart failure, diabetes mellitus, and prior stroke or transient ischemic attack, yielding a lower mean CHA₂DS₂-VASc score (3.7 vs. 5.1). The average HAS-BLED score was identical in both settings (2.8). Table 2 provides the full details.
Table 2.
Baseline characteristics of patients with coronary or peripheral artery disease in daily clinical practice vs. in the randomised trial
| Coronary artery disease | Peripheral artery disease | |||
|---|---|---|---|---|
| ETNA-AF-Europe (n = 2,860, 21.7%) |
ENGAGE AF-TIMI 48 (n = 1,478, 21.0%)a |
ETNA-AF-Europe (n = 432, 3.3%) |
ENGAGE AF-TIMI 48 (n = 841, 4.0%)b |
|
| Edoxaban 30 mg, n (%) | 870 (30.4%) | 362 (24.5%) | 153 (35.4%) | 258 (30.7%) |
| Off-label dosing, n (%) | 555 (19.4%) | NRc | 104 (24.1%) | NRc |
| - Of the reduced dose | 267 (9.3%) | NRc | 51 (11.8%) | NRc |
| - Of the standard dose | 288 (10.1%) | NRc | 53 (12.3%) | NRc |
| Ongoing use of antiplatelets, n (%) | 579 (20.2%) | 744 (50.3%) | 84 (19.4%) | NR |
| - Aspirin | 424 (14.8%) | 689 (45.9%) | 62 (14.4%) | 272 (32.3%) |
| Male, n (%) | 2,035 (71.2%) | 1,137 (76.9%) | 314 (72.7%) | 634 (75.4%) |
| Age [years], mean ± SD, or median [Q1-Q3] | 75.2 ± 8.5 | 72.1 ± 9.0 | 77 [71–82] | 74 [68–79] |
| Current smoker, n (%) | 201 (7.0%) | 124 (8.3%) | 61 (14.1%) | 92 (11.0) |
| Weight [kg], mean ± SD, or median [Q1-Q3] | 81.9 ± 16.0 | 87.0 ± 19.0 | 80 [70–92] | 83 [73–95] |
| eGFR [mL/min/1.73m2], mean ± SD | 68.8 ± 19.7d, e | 72.9 ± 29.7 | 66.9 ± 20.3d, e | NR |
| Creatinine clearance [mL/min], mean ± SD | 68.1 ± 25.7e | NR | 65.9 ± 25.9e | NR |
| - ≤ 50 mL/min, n (%) | 661 (23.1%)e | NR | 121 (28.0%)e | 210 (25.0%) |
| CHA2DS2-VASc, mean ± SD | 3.7 ± 1.5f | 5.1 ± 1.3 | 4.6 ± 1.4f | 5.2 ± 1.3 |
| HAS-BLED, mean ± SD | 2.8 ± 0.9f | 2.8 ± 1.0 | 3.0 ± 1.0f | 2.8 ± 1.0 |
| Peripheral artery disease, n (%) | 223 (7.8%) | 118 (8.0%) | 222 (51.4%) | 502 (59.7%) |
| Hypertension, n (%) | 2,394 (83.7%) | 1,405 (95.1%) | 365 (84.5%) | 793 (94.3%) |
| Congestive heart failure, n (%) | 69 (33.9%) | 961 (65.0%) | 120 (27.8%) | 513 (61.0%) |
| Diabetes mellitus, n (%) | 885 (30.9%) | 669 (45.3%) | 178 (41.2%) | 377 (44.8%) |
| Prior stroke or transient ischemic attack | 283 (9.9%) | 382 (25.8%) | 55 (12.7%) | 267 (31.7%) |
eGFR, estimated glomerular filtration rate; NR, not reported; Q1-Q3, first to third quartile; SD, standard deviation. a Data from the patients prescribed the high dose of edoxaban (i.e. 60 mg once daily with a dose reduction to 30 mg once daily in eligible patients). Results were extracted from Table S1 of the original report [8]; b Data from all patients with peripheral artery disease in ENGAGE AF-TIMI 48, which includes patients on high-dose edoxaban, low-dose edoxaban, and warfarin. Results were extracted from Table 1 of the original report [7]; c Dose adjustments were performed at randomisation and during follow-up according to the approved dosing criteria. Off-label use is therefore expected to be rare; d eGFR was estimated using the CKD-EPI formula; e Values outside of the 5 − 150 range were considered missing; f In ETNA-AF-Europe, we used a modified version of the scores
CIs for event rates were unavailable for participants of ENGAGE AF-TIMI 48. Thrombotic events in daily practice were roughly half of those in the randomized trial: stroke and systemic embolisms occurred at 0.87%/year vs. 1.5%/year; ischemic stroke at 0.65%/year vs. 1.3%/year; and myocardial infarctions at 0.71%/year vs. 1.2%/year. Bleeding rates were also much lower in daily practice—half to a third of those seen in the randomized trial. Major bleeding was 1.04%/year for ETNA-AF-Europe patients compared with 3.0%/year for those in ENGAGE AF-TIMI 48. Intracranial hemorrhage rates were similar between the settings at 0.22%/year and 0.3%/year. The most striking difference was in mortality. Though all-cause death rates were similar between the settings (5.95%/year vs. 5.3%/year), their causes differed. Non-cardiovascular causes accounted for most deaths in daily practice (4.37%/year vs. 1.6%/year) while cardiovascular deaths predominated in the randomized trial (1.58%/year vs. 3.7%/year). Figure 3 provides the complete overview.
Fig. 3.
Major cardiovascular events in patients with coronary artery disease in daily practice vs. in the randomized trial. Rates are in events per 100 patient-years (95% confidence intervals). a Data from patients with coronary artery disease prescribed the high dose of edoxaban (i.e. 60 mg once daily with a dose reduction to 30 mg once daily in eligible patients). Results were extracted from Figure S4 of the original report (patients with established coronary artery stenosis or angina pectoris). Confidence intervals of the event rates were not available [8]
Peripheral artery disease
PAD was uncommon in both ETNA-AF-Europe and in ENGAGE AF-TIMI 48 (3.3% vs. 4.0%). Patients in daily clinical practice were less often treated with aspirin in addition to edoxaban (14.4% vs. 32.3%) than those in the randomized trial, but more frequently received an off-label edoxaban dose, with 51 of the 153 patients (33.3%) on this dose receiving it off-label. Notably, congestive heart failure (27.8% vs. 61.0%) and a prior stroke or transient ischemic attack (12.7% vs. 31.7%) were substantially less prevalent in daily practice than in the randomized trial. Further details are provided in Table 2.
As noted in our methods section, the number of events among patients with PAD in ENGAGE AF-TIMI 48 treated with the approved edoxaban dosing regimen was too small for meaningful comparisons.
Discussion
This sub-study of ETNA-AF-Europe, a non-interventional prospective cohort study on more than 13,000 patients prescribed edoxaban, has two main findings. First, patients with AF and concomitant atherosclerotic disease treated with edoxaban in European countries have substantially lower cardiovascular risks than participants of the seminal randomized trial comparing edoxaban to warfarin (i.e., ENGAGE AF-TIMI 48), but higher non-cardiovascular risks. Second, in routine care, the most common event types were bleeding and non-cardiovascular death, both for those with and without atherosclerotic disease. We examine these results below.
Different risk profiles between the two settings
The finding of different risk profiles between the two settings suggests ENGAGE AF-TIMI 48’s trial population is under-representative of some patients treated in daily practice. In this case, however, we believe the trial’s results are generalizable to most of these cases. First, because cardiovascular risks are lower in daily practice than in the trial—not higher. Second, because the annual rates of both thrombotic and bleeding events in daily practice are low. Given our design we cannot differentiate between low background risks and edoxaban’s treatment effect. However, these findings add to the available trial evidence [3–5], and further ease concerns about possible suboptimal treatment with edoxaban in this subpopulation.
Why patients in routine practice show markedly lower cardiovascular risk yet higher non-cardiovascular risk remains uncertain. The most straightforward explanation is selection. ENGAGE AF–TIMI 48 enrolled patients with a CHADS₂ score of at least 2 from multiple continents and excluded those at high bleeding risk, with liver disease, active or recent cancer, limited life expectancy, poor adherence, or substance dependence [10, 11]. Beyond these formal criteria, frail patients or those with competing risks are often excluded from trial participation [15]. By contrast, ETNA-AF Europe enrolled patients only at European sites and was explicitly designed to include unselected patients, reflecting its pragmatic design [11]. Selection alone, however, may not fully account for the observed differences. We believe two additional factors also contributed.
Improved management of cardiovascular risk factors
The first factor is better management of cardiovascular risk factors. Since ENGAGE-AF TIMI 48’s completion, better (largely pharmacological) interventions have become available to manage heart failure, diabetes mellitus, obesity, kidney disease, and obstructive sleep apnea [16]. Moreover, recent European Society of Cardiology guidelines—especially since 2016, when ETNA-AF-Europe began enrollment—increasingly emphasize comprehensive risk factor management [16–18].
Combined, these advances should reduce the background risks of cardiovascular outcomes for patients receiving such treatments, as supported by findings from contemporary reports [19–22]. Since we did not collect detailed information on risk factor management, our data only provide indirect support for this contention.
Tailoring antithrombotic strategies to higher-risk individuals
The second factor is clinicians tailoring antithrombotic strategies to the higher-risk individual. Our data show clinicians are avoiding combination therapy with an antiplatelet drug and more often dose edoxaban off-label. By doing so, they may be better at balancing the risks than if they were to adhere to the recommendations of guidance documents in all cases—a notion backed by multiple observations.
First, patients undergoing a percutaneous coronary intervention with placement of a newer-iteration stent benefit from a shorter period of dual antiplatelet therapy or dual pathway inhibition [3, 23, 24]. Second, though edoxaban’s dose reduction criteria are clinically useful and have been studied carefully, they oversimplify the underlying pharmacobiology and have only been validated for the more typical patients enrolled in ENGAGE AF-TIMI [10, 25–27, 48]. Third, edoxaban exerts its effect in a dose-response manner, with higher exposure leading to more bleeding events and lower exposure to more thrombosis once levels fall below a certain threshold [10, 28–31]. Fourth, using edoxaban in currently approved doses is likely to cause too high bleeding risks in some higher-risk populations, for instance, Japanese octogenarians deemed unsuitable for conventional doses [32] Last, indirect evidence from typical patients suggests selectively using off-label 30 mg instead of on-label 60 mg has major potential to improve outcomes [28, 29].
These arguments are not specific to edoxaban but apply to all currently approved DOACs [26, 31–44]. Although current dosing criteria and guideline recommendations are probably appropriate for most patients, we suggest clinicians should not be discouraged from tailoring the antithrombotic therapy to those at very high bleeding risk who were under-represented in randomized trials.
Shift to better identification of high-risk individuals and reducing bleeding risk
Even though patients with atherosclerotic disease in our study faced a threefold higher risk of acute coronary syndromes than those without, the absolute risks were limited. Rather, for both groups, the most common event types were deaths—particularly those of non-cardiovascular causes—and major or CRNM bleeding.
Although the literature is mixed, most studies support our findings across populations at different absolute risks of thrombosis and bleeding [22, 33, 45–51]. Considered together, a consistent pattern emerges: serious bleeding is more frequent than thrombotic events. We believe this supports the going shift toward: (1) better identification of patients who will develop thrombosis or bleeding, and (2) reducing major and recurrent bleeding—so those at very high bleeding risk are not denied anticoagulation, and interruption or discontinuation can be avoided. Detailed reviews discussing these priorities can be read elsewhere.
Limitations
Our study has important limitations. First, we indirectly compared results from ETNA-AF-Europe and ENGAGE AF-TIMI 48, which differed in eligibility criteria, time periods, regions, variable definitions, as well as timing, frequency and method of data collection. For example, ETNA-AF-Europe lacked standardized medical definitions for CAD or PAD. These differences may have affected our findings. Second, we only had access to aggregated data from ENGAGE AF-TIMI 48—without patient-level details or 95% confidence intervals for those with either CAD or PAD. Third, our sample may be biased since participants of ETNA-AF-Europe excluded from our analyses differed from those who were included. Additionally, many patients discontinued the study early. If these dropouts were at higher risk, we may have underestimated event rates in this population. Fourth, we lacked information on the timing of prior events (e.g., acute coronary syndromes, revascularizations) and could not reliably analyze outcomes for patients on antiplatelet therapy or off-label edoxaban doses due to limited statistical power. Last, as a post-hoc sub-study with no adjustment for multiple testing, our results carry a higher risk of false positives and should not be overinterpreted.
Conclusions
In routine European clinical practice, deaths—particularly from non-cardiovascular causes—and bleeding were the most common events in AF patients treated at least once with edoxaban. This pattern was consistent for those with and without atherosclerotic disease. Those with atherosclerosis treated in European daily practice have substantially lower cardiovascular yet higher non-cardiovascular risks than similar patients treated in ENGAGE AF-TIMI 48, the seminal trial comparing edoxaban with warfarin. Even though the trial population is under-representative of some patients treated in daily practice, we believe the implications are limited—since both the rates of thrombosis and bleeding associated with edoxaban were low.
Supplementary Information
Below is the link to the electronic supplementary material.
Acknowledgements
We thank the participants and all principal investigators for their contributions to ETNA-AF-Europe. E-M.F. had full access to all the data and takes responsibility for its integrity and the accuracy of the data analysis. The graphical abstract was designed using resources from Flaticon.com. These resources were made by either Freepik or hqrloveq.
Author contributions
P.K., R.D.C., and J.R.d.G. contributed to the design of ETNA-AF-Europe. T.A.C.d.V., J.J.K., J.R.d.G., and E-M.F. designed the present sub-study. E-M.F. performed the statistical analyses, and T.A.C.d.V. prepared the figures. T.A.C.d.V. drafted the manuscript; J.R.d.G., J.J.K., R.P., and M.E.W.H. revised subsequent versions. T.A.C.d.V. drafted the peer-review response. All other authors critically reviewed the manuscript and the peer-review response. All authors read and approved the final version.
Funding
This study was funded by Daiichi Sankyo. Daiichi Sankyo has reviewed the manuscript before submission.
Data availability
The data underlying this article were provided by Daiichi Sankyo by permission. Data sharing criteria and the procedure for requesting access can be accessed here: https://vivli.org/ourmember/daiichi-sankyo/.
Declarations
Competing interests
Financial: T.A.C.d.V. reports research support from Daiichi Sankyo; personal fees from Bristol Myers Squibb; and financial support for printing his doctoral thesis from Amsterdam University Medical Centers, CardioResearch Rijnstate, MEDCON International/CVKG.nl, and Daiichi Sankyo Nederland BV. J.R.d.G. is supported by an ERA4Health PERSUADE consortium grant, by the Information Technology for European Advancement (ITEA4) grant, number 21026, as well as the CVON/Dutch Heart Foundation grant 01-002-2022-0118 for the EmbRACE (Electrophysiological and Molecular Biomarkers in Atrial Fibrillation) project. He received research grant funding through his institution from Atricure Inc., Bayer, Boston Scientific, Daiichi Sankyo, Johnson & Johnson, Medtronic, and Philips. He has also received honoraria and consultancy fees from Atricure Inc., Bayer, Berlin-Chemie, Daiichi Sankyo, Johnson & Johnson, Menarini, Medtronic, Novartis, and Servier. M.E.W.H. received speakers fees from Roche Diagnostics, Bayer, BMS/Pfizer, Boehringer Ingelheim, Daiichi Sankyo and a research grant of the Dutch Thrombosis Federation. J.J.K., E.M.F., and R.S. are employees of Daiichi Sankyo. LP reports having received fees and honoraria from Daiichi Sankyo, SOTIO and Beckman Coulter. ID reports speaker fees from Daiichi Sankyo and Pfizer. JS has received consultant and/or speaker fees from Abbott, Bayer, Berlin-Chemie, Biosense Webster, Biotronik, Boehringer-Ingelheim, Boston Scientific, BMS, Daiichi Sankyo, Medscape, Medtronic, Menarini, Pfizer, Saja, Servier, and WebMD. He reports ownership of Swiss EP and CorXL. P.K. receives research support for basic, translational, and clinical research projects from European Union (BigData@Heart, grant agreement EU IMI 116074; AFFECT-AF, grant agreement 84770; MAESTRIA, grant agreement 965286), British Heart Foundation (PG/17/30/32961; PG/20/22/35093; AA/18/2/34218), Leducq Foundation, Medical Research Council (United Kingdom), and German Centre for Cardiovascular Research, from several drug and device companies active in AF, and has received honoraria from several such companies in the past, but not in the last three years. He is an inventor on two patents held by University of Birmingham (Atrial Fibrillation Therapy WO 2015140571, Markers for Atrial Fibrillation WO 2016012783). R.D.C. has received fees, honoraria and/or research funding from Sanofi-Aventis, Boehringer Ingelheim, Bayer, Bristol Myers Squibb/Pfizer, Daiichi Sankyo, Novartis, Portola, Roche, Menarini, and Guidotti. Other authors declare no financial interests.Non-financial: T.A.C.d.V is a member of the adjudication committee for the Left Atrial Appendage Exclusion for Prophylactic Stroke Reduction Trial (LeAAPS; NCT05478304), sponsored by AtriCure; the Low International Normalized Ratio to Minimize Bleeding with Mechanical Valves trial (LIMIT; NCT03636295) and the Direct Oral Anticoagulation versus Warfarin after Cardiac Surgery trial (DANCE; NCT04284839), both sponsored by the Population Health Research Institute. He also serves on the Data Safety Committee of the DUTCH Weight Control in Atrial Fibrillation Study (DUTCH-WAIST; NCT06184633), sponsored by Rijnstate Hospital. All other authors report no non-financial interests.
Non-Financial:T.A.C.d.V is a member of the adjudication committee for the Left Atrial Appendage Exclusion for Prophylactic Stroke Reduction Trial (LeAAPS; NCT05478304), sponsored by AtriCure; the Low International Normalized Ratio to Minimize Bleeding with Mechanical Valves trial (LIMIT; NCT03636295) and the Direct Oral Anticoagulation versus Warfarin after Cardiac Surgery trial (DANCE; NCT04284839), both sponsored by the Population Health Research Institute. He also serves on the Data Safety Committee of the DUTCH Weight Control in Atrial Fibrillation Study (DUTCH-WAIST; NCT06184633), sponsored by Rijnstate Hospital. All other authors report no non-financial interests.
Consent
All participants provided written informed consent.
Ethics approval
ETNA-AF-Europe is performed in accordance with the Declaration of Helsinki, and its study protocol approvals were provided by competent authorities, and central and local ethics committees for each participating site, in all countries involved.
Footnotes
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Michniewicz E, Mlodawska E, Lopatowska P, Tomaszuk-Kazberuk A, Malyszko J (2018) Patients with atrial fibrillation and coronary artery disease - double trouble. Adv Med Sci 63(1):30–35. 10.1016/j.advms.2017.06.005 [DOI] [PubMed] [Google Scholar]
- 2.Goto S, Bhatt DL, Röther J et al (2008) Prevalence, clinical profile, and cardiovascular outcomes of atrial fibrillation patients with atherothrombosis. Am Heart J 156(5):855–863, 863.e852. 10.1016/j.ahj.2008.06.029 [DOI] [PubMed] [Google Scholar]
- 3.Cho Min S, Kang D-Y, Ahn J-M et al (2024) Edoxaban Antithrombotic Therapy for Atrial Fibrillation and Stable Coronary Artery Disease. N Engl J Med 0(0). 10.1056/NEJMoa2407362 [DOI] [PubMed]
- 4.Vranckx P, Valgimigli M, Eckardt L et al (2019) Edoxaban-based versus vitamin K antagonist-based antithrombotic regimen after successful coronary stenting in patients with atrial fibrillation (ENTRUST-AF PCI): a randomised, open-label, phase 3b trial. Lancet 394(10206):1335–1343. 10.1016/s0140-6736(19)31872-0 [DOI] [PubMed] [Google Scholar]
- 5.Fukamachi D, Okumura Y, Matsumoto N et al (2022) Edoxaban monotherapy in nonvalvular atrial fibrillation patients with coronary artery disease. J Interv Cardiol 2022:5905022. 10.1155/2022/5905022 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Zelniker TA, Ruff CT, Antman EM, Giugliano RP (2019) The efficacy and safety of non-vitamin K antagonist oral anticoagulants in patients with atrial fibrillation and coronary artery disease: a meta-analysis of randomized trials. Eur Heart J Acute Cardiovasc Care 8(6):554–561. 10.1177/2048872618796990 [DOI] [PubMed] [Google Scholar]
- 7.Bonaca MP, Antman EM, Cunningham JW et al (2022) Ischaemic and bleeding risk in atrial fibrillation with and without peripheral artery disease and efficacy and safety of full- and half-dose edoxaban vs. warfarin: insights from ENGAGE AF-TIMI 48. Eur Heart J Cardiovasc Pharmacother 8(7):695–706. 10.1093/ehjcvp/pvab089 [DOI] [PubMed] [Google Scholar]
- 8.Zelniker TA, Ruff CT, Wiviott SD et al (2019) Edoxaban in atrial fibrillation patients with established coronary artery disease: insights from ENGAGE AF-TIMI 48. Eur Heart J Acute Cardiovasc Care 8(2):176–185. 10.1177/2048872618790561 [DOI] [PubMed] [Google Scholar]
- 9.Lip GY, Nieuwlaat R, Pisters R, Lane DA, Crijns HJ (2010) Refining clinical risk stratification for predicting stroke and thromboembolism in atrial fibrillation using a novel risk factor-based approach: the euro heart survey on atrial fibrillation. Chest 137(2):263–272. 10.1378/chest.09-1584 [DOI] [PubMed] [Google Scholar]
- 10.Giugliano RP, Ruff CT, Braunwald E et al (2013) Edoxaban versus warfarin in patients with atrial fibrillation. N Engl J Med 369(22):2093–2104. 10.1056/NEJMoa1310907 [DOI] [PubMed] [Google Scholar]
- 11.De Caterina R, Kelly P, Monteiro P et al (2019) Design and rationale of the Edoxaban Treatment in routiNe clinical prActice for patients with Atrial Fibrillation in Europe (ETNA-AF-Europe) study. J Cardiovasc Med (Hagerstown) 20(2):97–104. 10.2459/jcm.0000000000000737 [DOI] [PubMed] [Google Scholar]
- 12.von Elm E, Altman DG, Egger M et al (2008) The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. J Clin Epidemiol 61(4):344–349. 10.1016/j.jclinepi.2007.11.008 [DOI] [PubMed] [Google Scholar]
- 13.European Medicines Agency Lixiana: EPAR-Product Information. Committee for Medicinal Products for Human Use (CHMP); Jul 3, 2015. URL: https://www.ema.europa.eu/en/documents/product-information/lixiana-epar-product-information_en.pdf
- 14.Schulman S, Kearon C (2005) Definition of major bleeding in clinical investigations of antihemostatic medicinal products in non-surgical patients. J Thromb Haemost 3(4):692–694. 10.1111/j.1538-7836.2005.01204.x [DOI] [PubMed] [Google Scholar]
- 15.Soler-Espejo E, Rivera-Caravaca JM, Bru-Cánovas JD et al (2024) Patients with atrial fibrillation and common exclusion criteria from clinical trials are at high risk of clinical events: the Murcia AF project II (MAFP-II) cohort study. Intern Emerg Med 19(7):1941–1948. 10.1007/s11739-024-03701-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Van Gelder IC, Rienstra M, Bunting KV et al (2024) ESC Guidelines for the management of atrial fibrillation developed in collaboration with the European Association for Cardio-Thoracic Surgery (EACTS): Developed by the task force for the management of atrial fibrillation of the European Society of Cardiology (ESC), with the special contribution of the European Heart Rhythm Association (EHRA) of the ESC. Endorsed by the European Stroke Organisation (ESO). Eur Heart J 2024:ehae176. 10.1093/eurheartj/ehae176 [Google Scholar]
- 17.Hindricks G, Potpara T, Dagres N et al (2020) ,. ESC Guidelines for the diagnosis and management of atrial fibrillation developed in collaboration with the European Association for Cardio-Thoracic Surgery (EACTS): The Task Force for the diagnosis and management of atrial fibrillation of the European Society of Cardiology (ESC) Developed with the special contribution of the European Heart Rhythm Association (EHRA) of the ESC. Eur Heart J 2021;42(5):373–498.10.1093/eurheartj/ehaa612. [DOI] [PubMed]
- 18.Kirchhof P, Benussi S, Kotecha D et al (2016) 2016 ESC guidelines for the management of atrial fibrillation developed in collaboration with EACTS. Eur Heart J 37(38):2893–2962. 10.1093/eurheartj/ehw210 [DOI] [PubMed] [Google Scholar]
- 19.Komen JJ, Hunt NB, Pottegård A et al (2023) Heterogeneity after harmonisation: a retrospective cohort study of bleeding and stroke risk after the introduction of direct oral anticoagulants in four Western European countries. Pharmacoepidemiol Drug Saf 32(11):1223–1232. 10.1002/pds.5650 [DOI] [PubMed] [Google Scholar]
- 20.Connolly SJ, Eikelboom J, Joyner C et al (2011) Apixaban in patients with atrial fibrillation. N Engl J Med 364(9):806–817. 10.1056/NEJMoa1007432 [DOI] [PubMed] [Google Scholar]
- 21.Granger CB, Alexander JH, McMurray JJ et al (2011) Apixaban versus warfarin in patients with atrial fibrillation. N Engl J Med 365(11):981–992. 10.1056/NEJMoa1107039 [DOI] [PubMed] [Google Scholar]
- 22.Piccini JP, Patel MR, Steffel J et al (2025) Asundexian versus apixaban in patients with atrial fibrillation. N Engl J Med 392(1):23–32. 10.1056/NEJMoa2407105 [DOI] [PubMed] [Google Scholar]
- 23.Cassese S, Byrne RA, Tada T, King LA, Kastrati A (2012) Clinical impact of extended dual antiplatelet therapy after percutaneous coronary interventions in the drug-eluting stent era: a meta-analysis of randomized trials. Eur Heart J 33(24):3078–3087. 10.1093/eurheartj/ehs318 [DOI] [PubMed] [Google Scholar]
- 24.Lemesle G, Didier R, Steg PG et al (2025) Aspirin in patients with chronic coronary syndrome receiving oral anticoagulation. N Engl J Med. 10.1056/NEJMoa2507532 [DOI] [PubMed] [Google Scholar]
- 25.Steffel J, Collins R, Antz M et al (2021) 2021 European Heart Rhythm Association practical guide on the use of non-vitamin K antagonist oral anticoagulants in patients with atrial fibrillation. Europace 23(10):1612–1676. 10.1093/europace/euab065 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Douxfils J, Adcock DM, Bates SM et al (2021) 2021 update of the International Council for Standardization in Haematology recommendations for laboratory measurement of direct oral anticoagulants. Thromb Haemost 121(08):1008–1020. 10.1055/a-1450-8178 [DOI] [PubMed] [Google Scholar]
- 27.Moner-Banet T, Alberio L, Bart PA (2020) Does one dose really fit all? On the monitoring of direct oral anticoagulants: a review of the literature. Hamostaseologie 40(2):184–200. 10.1055/a-1113-0655 [DOI] [PubMed] [Google Scholar]
- 28.Steffel J, Ruff CT, Yin O et al (2021) Randomized, double-blind comparison of half-dose versus full-dose edoxaban in 14,014 patients with atrial fibrillation. J Am Coll Cardiol 77(9):1197–1207. 10.1016/j.jacc.2020.12.053 [DOI] [PubMed] [Google Scholar]
- 29.Zimerman A, Braunwald E, Steffel J et al (2024) Dose reduction of edoxaban in patients 80 years and older with atrial fibrillation: post hoc analysis of the ENGAGE AF-TIMI 48 randomized clinical trial. JAMA Cardiol 9(9):817–825. 10.1001/jamacardio.2024.1793 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Yin OQP, Antman EM, Braunwald E et al (2018) Linking endogenous factor Xa activity, a biologically relevant pharmacodynamic marker, to edoxaban plasma concentrations and clinical outcomes in the ENGAGE AF-TIMI 48 trial. Circulation 138(18):1963–1973. 10.1161/circulationaha.118.033933 [DOI] [PubMed] [Google Scholar]
- 31.Godino C, Mazza R, Gaspardone C et al (2025) Plasma levels measurement of the 4 direct oral anticoagulants in patients with atrial fibrillation at the time of acute thromboembolic and bleeding events. J Thromb Haemost. 10.1016/j.jtha.2025.06.031 [DOI] [PubMed] [Google Scholar]
- 32.Okumura K, Akao M, Yoshida T et al (2020) Low-dose edoxaban in very elderly patients with atrial fibrillation. N Engl J Med 383(18):1735–1745. 10.1056/NEJMoa2012883 [DOI] [PubMed] [Google Scholar]
- 33.Joosten LPT, van Doorn S, van de Ven PM et al (2023) Safety of Switching from a Vitamin K Antagonist to a Non-Vitamin K Antagonist Oral Anticoagulant in Frail Older Patients with Atrial Fibrillation: Results of the FRAIL-AF Randomized Controlled Trial. Circulation. 10.1161/circulationaha.123.066485 [DOI] [PubMed] [Google Scholar]
- 34.De Vriese AS, Caluwé R, Van Der Meersch H, De Boeck K, De Bacquer D (2021) Safety and efficacy of vitamin K antagonists versus rivaroxaban in hemodialysis patients with atrial fibrillation: a multicenter randomized controlled trial. J Am Soc Nephrol 32(6):1474–1483. 10.1681/asn.2020111566 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Pokorney SD, Chertow GM, Al-Khalidi HR et al (2022) Apixaban for patients with atrial fibrillation on hemodialysis: a multicenter randomized controlled trial. Circulation 146(23):1735–1745. 10.1161/circulationaha.121.054990 [DOI] [PubMed] [Google Scholar]
- 36.Freixa X, Cruz-González I, Cepas-Guillén P et al (2024) Low-dose direct oral anticoagulation vs dual antiplatelet therapy after left atrial appendage occlusion: the ADALA randomized clinical trial. JAMA Cardiol 9(10):922–926. 10.1001/jamacardio.2024.2335 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Connolly SJ, Ezekowitz MD, Yusuf S et al (2009) Dabigatran versus warfarin in patients with atrial fibrillation. N Engl J Med 361(12):1139–1151. 10.1056/NEJMoa0905561 [DOI] [PubMed] [Google Scholar]
- 38.Reilly PA, Lehr T, Haertter S et al (2014) The effect of dabigatran plasma concentrations and patient characteristics on the frequency of ischemic stroke and major bleeding in atrial fibrillation patients: the RE-LY Trial (Randomized Evaluation of Long-Term Anticoagulation Therapy). J Am Coll Cardiol 63(4):321–328. 10.1016/j.jacc.2013.07.104 [DOI] [PubMed] [Google Scholar]
- 39.Yoon CH, Park YK, Kim SJ et al (2014) Eligibility and preference of new oral anticoagulants in patients with atrial fibrillation: comparison between patients with versus without stroke. Stroke 45(10):2983–2988. 10.1161/strokeaha.114.005599 [DOI] [PubMed] [Google Scholar]
- 40.Bhagirath VC, Eikelboom JW, Hirsh J et al (2017) Apixaban-calibrated anti-FXa activity in relation to outcome events and clinical characteristics in patients with atrial fibrillation: results from the AVERROES trial. TH Open 1(2):e139–e145. 10.1055/s-0037-1613679 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Cannon CP, Bhatt DL, Oldgren J et al (2017) Dual antithrombotic therapy with dabigatran after PCI in atrial fibrillation. N Engl J Med 377(16):1513–1524. 10.1056/NEJMoa1708454 [DOI] [PubMed] [Google Scholar]
- 42.Rose M, Beasley BN, Center For Drug Evaluation And Research Application Number : 202155Orig1s000 Medical Review(s). US Food and Drug Administration. URL: https://www.accessdata.fda.gov/drugsatfda_docs/nda/2012/202155Orig1s000MedR.pdf [Accessed November 21, 2021]
- 43.Ruff CT, Giugliano RP, Braunwald E et al (2015) Association between edoxaban dose, concentration, anti-Factor Xa activity, and outcomes: an analysis of data from the randomised, double-blind ENGAGE AF-TIMI 48 trial. Lancet 385(9984):2288–2295. 10.1016/S0140-6736(14)61943-7 [DOI] [PubMed] [Google Scholar]
- 44.Zhang L, Yan X, Fox KAA et al (2020) Associations between model-predicted rivaroxaban exposure and patient characteristics and efficacy and safety outcomes in patients with non-valvular atrial fibrillation. J Thromb Thrombolysis 50(1):20–29. 10.1007/s11239-020-02077-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45.Lip GYH, Keshishian A, Li X et al (2018) Effectiveness and safety of oral anticoagulants among nonvalvular atrial fibrillation patients. Stroke 49(12):2933–2944. 10.1161/STROKEAHA.118.020232 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 46.Graham DJ, Baro E, Zhang R et al (2019) Comparative stroke, bleeding, and mortality risks in older Medicare patients treated with oral anticoagulants for nonvalvular atrial fibrillation. Am J Med 132(5):596-604.e511. 10.1016/j.amjmed.2018.12.023 [DOI] [PubMed] [Google Scholar]
- 47.Higa M, Morimoto T, Ikeda M, Ueda S (2025) Real-World Outcomes of Patients With Non-Valvular Atrial Fibrillation on Anticoagulants Ineligible for Phase III Trials of Direct Oral Anticoagulants ― A Retrospective Cohort Study ―. Circulation Reports ;advpub 10.1253/circrep.CR-25-0079 [DOI] [PMC free article] [PubMed]
- 48.Okumura K, Akao M, Yoshida T (2020) Low-dose edoxaban in very elderly patients with atrial fibrillation. N Engl J Med. 10.1056/NEJMoa2012883 [DOI] [PubMed] [Google Scholar]
- 49.Steinberg BA, Shrader P, Pieper K et al (2018) Frequency and outcomes of reduced dose non-vitamin K antagonist anticoagulants: results from ORBIT-AF II (the outcomes registry for better informed treatment of atrial fibrillation II). J Am Heart Assoc 7(4):e007633. 10.1161/jaha.117.007633 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 50.Larsen TB, Skjoth F, Nielsen PB, Kjaeldgaard JN, Lip GY (2016) Comparative effectiveness and safety of non-vitamin K antagonist oral anticoagulants and warfarin in patients with atrial fibrillation: propensity weighted nationwide cohort study. BMJ 353:i3189. 10.1136/bmj.i3189 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 51.Nielsen PB, Skjoth F, Sogaard M et al (2017) Effectiveness and safety of reduced dose non-vitamin K antagonist oral anticoagulants and warfarin in patients with atrial fibrillation: propensity weighted nationwide cohort study. BMJ 356:j510. 10.1136/bmj.j510 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The data underlying this article were provided by Daiichi Sankyo by permission. Data sharing criteria and the procedure for requesting access can be accessed here: https://vivli.org/ourmember/daiichi-sankyo/.