Atrial fibrillation (AF) more than doubles the five-year risk of stroke in middle-aged men and women.1 Prior cerebrovascular disease markedly amplifies the risk of recurrent stroke in patients with or without AF.1,2 Figure 1 shows the influence of AF and prior cerebrovascular disease (stroke or transient ischaemic attack [TIA]) on the estimated five-year risk of a composite of stroke, systemic thromboembolism, or TIA (most events were ischaemic strokes) for a 60-year-old individual, from a large cohort study conducted in the UK.1 These observations demonstrate the need for long-term treatment to reduce the risk of stroke in these individuals, particularly where AF and prior cerebrovascular disease co-exist.
Long-term oral anticoagulation is the mainstay of stroke prevention in at-risk subjects.3 Vitamin K antagonists (VKA), principally warfarin, have been the most commonly used method of oral anticoagulation, but the more recent availability of non-VKA oral anticoagulants (NOACs) has broadened treatment options for these patients.4 Edoxaban is a Factor Xa antagonist with a therapeutic indication for the prevention of stroke in adults with non-valvular AF and at least one other risk factor.5 This review summarises the clinical evidence-base for edoxaban from the randomised, controlled Effective Anticoagulation with Factor Xa Next Generation in Atrial Fibrillation–Thrombolysis in Myocardial Infarction 48 trial (ENGAGE AF-TIMI-48; NCT00781391).6
Design of the ENGAGE-AF TIMI-48 trial
The design of this trial7 is summarised by Khan and Thomas in the first article in this supplement. Briefly, a population of 21,105 patients with non-valvular AF were randomised in a double-blind, double-dummy manner to once-daily low-dose edoxaban (30 mg reduced to 15 mg in appropriate patients), once-daily high-dose edoxaban (60 mg reduced to 30 mg in appropriate patients), or to warfarin (target international normalised ratio [INR] 2.0–3.0) for a median of 2.8 years. The dose reductions were applied for patients with renal dysfunction (creatinine clearance 30–50 ml/min), weight ≤60 kg, or prescription of verapamil, dronedarone or quinidine (these drugs are strong P-glycoprotein inhibitors that might increase systemic exposure to edoxaban by inhibiting its metabolism8). Patients were also required to have a CHADS2 score of at least two, signifying elevated risk of stroke at baseline.9 The low-dose edoxaban regimen was not subsequently submitted for regulatory approval, and further discussion in this article will be about the approved high-dose regimen.
The primary efficacy outcome was time to stroke or systemic embolic event (SSE). A modified intention-to-treat (mITT) analysis was used first to evaluate non-inferiority of edoxaban versus warfarin; if this was significant, the ITT population was used to test for superiority (this included outcomes during a 30-day period for transition to open-label anticoagulant therapy after the end of randomised treatment). A composite of stroke, SSE or cardiovascular (CV) death was a prespecified secondary efficacy end point, and major bleeding was a prespecified primary safety end point.
Clinical outcomes related to stroke from the main trial analysis
A summary of main outcomes from ENGAGE-AF TIMI-48, with emphasis on outcomes related to stroke, is shown in Figure 2.6 Edoxaban was non-inferior to warfarin in terms of the primary end point, although a trend towards superiority did not reach significance (p=0.08). The risk of stroke/SSE/CV death, stroke/SSE/all-cause death, and death/intracranial haemorrhage were reduced significantly with edoxaban compared with warfarin.
Edoxaban was associated with reduced risk of haemorrhagic stroke versus warfarin (p<0.001). Hazard ratios (HRs) for ischaemic stroke, disabling/fatal stroke and fatal stroke did not differ significantly between warfarin and edoxaban. The reduced risk of haemorrhagic stroke, and the composite end point including intracranial haemorrhage described above, were consistent with significant reductions in major bleeding events for edoxaban versus warfarin, including any major bleed, fatal bleeds, any intracranial bleed and fatal intracranial bleeds.6
Cardiovascular death with edoxaban was reduced versus warfarin (HR 0.86, 95% confidence interval [CI] 0.77 to 0.97, p=0.013). There was no significant effect on SSE (not shown in figure 2).
Influence of prior cerebrovascular disease on stroke outcomes in ENGAGE-AF TIMI-48
Twenty-eight per cent of the ENGAGE-AF population had a prior history of stroke or TIA at baseline.10 In terms of risk factors for stroke, patients with prior stroke/TIA were on average younger (70.4 vs. 70.7 years), less likely to be aged ≥75 years (35.5% vs. 42.2%), with more renal impairment (creatinine clearance ≤50 ml/min; 20.3% vs. 18.2%) and weight ≤60 kg (13.8% vs. 8.4%), but were less likely to have heart failure (42.4% vs. 63.4%), diabetes (42.4% vs. 63.4%) or hypertension (26.6% vs. 39.9%) at baseline (all p<0.001). More patients with prior stroke/TIA had CHADS2 score 4–6 (67.0% vs. 5.1%, p<0.001), consistent with this risk factor contributing two points to the score, while other individual risk factors contribute only one point each. More patients with, versus without, prior stroke (29.0% vs. 23.9%, p<0.001) required a dose reduction for renal impairment, low body weight or receipt of a strong P-glycoprotein inhibitor (verapamil, dronedarone or quinidine).
The effects of edoxaban versus warfarin on the primary end point was similar in patients with prior stroke/TIA (HR 0.86, 95% CI 0.67 to 1.09) and without (HR 0.88, 95% CI 0.72 to 1.08), with a p for interaction (pint) of 0.86. A nominally significant reduction in CV death, stroke or SSE in favour of edoxaban was observed in patients with prior stroke/TIA (HR 0.82, 95% CI 0.69 to 0.97, p=0.019), although the HR for patients without prior stroke/TIA barely failed to achieve significance (HR 0.89, 95% CI 0.79 to 1.01, p=0.08), and pint was non-significant (0.38). Findings were similar for all-cause death, stroke or SSE, CV death, and major adverse CV events, with a marginally significant reduction in risk on the prior stroke/TIA group only, and no significant treatment interaction.
There was no significant heterogeneity in outcomes for stroke, or subtypes of stroke, in patients with or without prior stroke/TIA (figure 3). The risk of haemorrhagic stroke was reduced irrespective of stroke/TIA history, as in the main analysis, with no influence of prior stroke/TIA on the HRs for edoxaban versus warfarin. However, the absolute benefit of edoxoban was larger in the prior stroke/TIA group, consistent with the higher baseline risk of adverse stroke outcomes in this group. The reduction in haemorrhagic stroke with edoxaban versus warfarin was consistent with the maintained advantage for edoxaban in terms of reduced risk of bleeding events (figure 4).
The risk of net clinical benefit outcomes (death/stroke/systemic embolic event/major bleed, death or disabling stroke, death or disabling stroke or life-threatening bleed, and death or stroke or intracranial haemorrhage) were also reduced significantly in favour of edoxaban in the prior stroke/TIA group, but not in the lower-risk group without prior cerebrovascular diseases. Again, there were no significant interactions between stroke/TIA history and outcomes.
Other key subgroup analyses from ENGAGE-AF TIMI-48
Causes of death
Another subgroup analysis focused on mortality outcomes across treatment groups.11 There was a trend towards fewer deaths on edoxaban versus warfarin (773 vs. 839 deaths, p=0.08). Fewer fatal bleeds on edoxaban versus warfarin contributed importantly to the lower number of deaths with edoxaban (35 vs. 56, p=0.013). There were no significant differences between arms in the numbers of deaths for CV deaths unrelated to bleeding.
Elderly and frail patients
The ENGAGE-AF TIMI-48 population was stratified into subgroups by age for this subgroup analysis: 26% were aged <65 years, 34% were aged between 65 and 74 years, and 40% were aged 75 years or over.12 There were some important differences at baseline between these groups, particularly a lower proportion of some significant concomitant diseases with increasing age, e.g. diabetes (41%, 43% and 28%, respectively) and heart failure (70%, 63% and 45%, respectively). It is possible that the elderly group represented a healthy, motivated and compliant population: time in therapeutic range for warfarin was modestly higher in the elderly group (69.6%) relative to the younger groups (67.8% and 68.4%). The risk of stroke increased with age: the proportions with CHADS2 score ≥4 were 15%, 17%, and 32% and with CHA2DS2-VASc score ≥4 were 33%, 79%, and 88%, for the increasing age groups, respectively. HAS-BLED score increased more steeply with age (16%, 58%, and 56% had score ≥3), signifying an especially marked effect of age on bleeding risk.
There were no significant interactions between age and treatment outcomes. Hazard ratios for the primary end point (stroke or SSE) were between 0.83 and 0.94 across the three age categories, all 95% CI crossed unity, and pint was 0.84. Findings were similar for ischaemic stroke (HRs 0.90–1.16, pint=0.48), and fatal stroke (HRs 0.86–1.02, pint=0.89). Nominally significant reductions in favour of edoxaban (point estimates for HR and their 95% CI were all <1) occurred for the 65–74 year and ≥75 year groups for haemorrhagic stroke, intracranial haemorrhage, and major bleeding; however, 95% CI for the <65 year age group for these outcomes crossed unity, and pint did not approach significance for any outcome. These data were consistent with the higher risk of bleeding in older patients, and suggest that the benefit in the main analysis may have been driven, importantly, by treatment effects in older patients.
Nine hundred (4.3%) of the ENGAGE-AF TIMI-48 population were considered to be at risk of falls, based on prior history of falls, lower extremity weakness, poor balance, cognitive impairment, orthostatic hypotension, receipt of psychotropic drugs, severe arthritis, or dizziness.13 Unsurprisingly, more of this frail population were elderly (66% were aged ≥75 years vs. 39% for younger patients); they were more likely to suffer bone fractures, major bleeding events or death, but were not at increased risk of ischaemic events. There was no significant interaction between frailty status and treatment effects.
Anticoagulant treatment regimen and treatment history
Prior treatment with a VKA in ENGAGE-AF TIMI-48 was defined as more than 60 days of continuous receipt of one of these drugs.14 About two-fifths of the trial population (41%) were prior VKA users according to this definition, although 37% of the ‘VKA naïve’ group were receiving a VKA at randomisation. The VKA naïve group were less likely to be receiving aspirin (37% vs. 95% for the VKA experienced group) or thienopyridine (3.3% vs. 1.6%), and were more likely to be receiving amiodarone (15% vs. 10%) at baseline. Randomisation to edoxaban versus warfarin influenced the primary end point more strongly in VKA naïve (HR 0.71, 95% CI 0.56 to 0.90) compared with VKA experienced (HR 1.01, 95% CI 0.82 to 1.24) patients; pint was 0.028. Edoxaban was also more effective than warfarin in VKA naïve patients for stroke: VKA naïve – HR 0.69, 95% CI 0.54 to 0.88, VKA experienced – HR 1.06, 95% CI 1.85 to 1.31, pint=0.011; and ischaemic stroke: VKA naïve – HR 0.74, 95% CI 0.56 to 0.97, VKA experienced – HR 1.29, 95% CI 0.99 to 1.62, pint=0.004.
Haemorrhagic stroke was reduced similarly in both VKA history groups (HRs 0.56 and 0.53, respectively, pint=0.87). A significantly larger treatment effect occurred in VKA naïve patients for the primary outcome, stroke and ischaemic stroke, but no effect of VKA history on the reduction in haemorrhagic stroke. There was no influence of treatment history on MI, CV death and all-cause death for edoxaban versus warfarin. Edoxaban reduced major bleeding outcomes, e.g. HRs (95% CI) for major bleeding versus warfarin (VKA naïve/VKA experienced) were 0.80 (0.65 to 0.97)/0.81 (0.69 to 0.95), pint=0.90. Similar results were seen for other bleeding outcomes.
One-quarter of patients (25%) received a single antiplatelet inhibitor at baseline.15 Receipt versus non-receipt of these drugs did not influence the primary outcome significantly (HR 1.12, 95% CI 0.95 to 1.32), but increased the risk of major bleeding (HR 1.46, 95% CI 1.27 to 1.67). Receipt of a single antiplatelet inhibitor did not alter the relative efficacy of edoxaban and warfarin.
Anticoagulant therapy may be suspended for surgical interventions, and 43% of the ENGAGE-AF TIMI-48 population suspended study treatment for more than three days without open-label anticoagulant cover.16 Patients who interrupted anticoagulant were more likely to report subsequent stroke/SSE (15.42 vs. 0.26/100 patient-years) and major adverse CV events (60.82 vs. 0.36/100 patient-years, p<0 .001 for both). Another analysis evaluated the effects of stopping or continuing anticoagulant therapy on outcomes occurring between seven days before and 30 days after the surgery.17 Rates of SSE, major bleeds, or major bleeds plus clinically relevant non-major bleeding did not differ between treatment groups for patients who had discontinued or continued their anticoagulant regimen.
An analysis in 1,943 patients from four East Asian countries (Japan, China, Taiwan, and South Korea) demonstrated significant benefits for edoxaban versus warfarin for the primary end point (HR 0.53, 95% CI 0.31 to 0.90, p=0.02).18 Randomisation to edoxaban was associated with less major bleeding versus warfarin, as in the main analysis (higher dose: HR 0.61, 95% CI 0.41 to 0.89, p=0.011).
Trial participants from Latin America were at similar risk of stroke compared with patients elsewhere, but were at increased risk of mortality and intracranial haemorrhage.19 Treatment effects were broadly similar between patients from Latin America and other regions, except for a greater effect of edoxaban on haemorrhagic stroke in Latin America (HRs 0.16 vs. 0.64, pint=0.037).
Implications for care
Anticoagulant therapy with a VKA represented the standard of care for anticoagulant therapy for people with AF for many years, but management guidelines in this area now recommend use of a NOAC (including edoxaban) or VKA for patients with AF and risk factor(s) for stroke.20,21 The analyses from ENGAGE-AF TIMI-48 described above confirm the utility of edoxaban for stroke prevention, with equivalent efficacy with warfarin for preventing CV and embolic events, with a reduced risk of bleeding, including into the brain. The significantly reduced risk of major bleeding on edoxaban versus warfarin contributed to numerically reduced mortality on this treatment.
Importantly, the therapeutic profile of edoxaban is maintained in patients with a history of cerebrovascular disease and in elderly patients (including frail individuals). Advanced age and prior stroke/TIA are themselves potent risk factors for stroke, as reflected by their higher weighting compared with other risk factors in the CHA2DS2-VASc risk engine for stroke prediction.9
Edoxaban was as effective as warfarin for the prevention of stroke and systemic embolism, in patients with non-valvular AF, with or without prior cerebrovascular disease. In addition, edoxaban significantly reduced the risk of bleeding versus warfarin.
- The randomised ENGAGE-AF TIMI-48 trial compared the effects of edoxaban and warfarin on clinical outcomes in people with atrial fibrillation (AF) at elevated risk of stroke
- Edoxaban was as effective as warfarin for the prevention of stroke or systemic embolic events in people with AF, with a significantly reduced risk of major bleeds
- This benefit extended to people with prior stroke or transient ischaemic attack, who are at markedly elevated risk of stroke
Conflicts of interest
PG has received conference fees and travel grants, speaker fees and been Principal Investigator for NIHR commercial trials for companies involved with NOACs: Bayer, Boehringer Ingelheim, Daiichi Sankyo, and BMS/Pfizer.
A medical writer (Dr Mike Gwilt, GT Communications) provided editorial assistance, funded by the British Journal of Cardiology.
Lead Consultant Stroke Physician
Southend University Hospital NHS Trust, Prittlewell Chase, Westcliff-on-Sea, Essex, SS0 0RY.
Articles in this supplement
1. Insights in patients with atrial fibrillation and co-existing cardiovascular disease
3. Insights in the treatment and prophylaxis of venous thromboembolism
4. Insights from primary care: opportunities and challenges
1. Christiansen CB, Gerds TA, Olesen JB et al. Atrial fibrillation and risk of stroke: a nationwide cohort study. Europace 2016;18:1689–97. https://doi.org/10.1093/europace/euv401
2. Mohan KM, Wolfe CD, Rudd AG, Heuschmann PU, Kolominsky-Rabas PL, Grieve AP. Risk and cumulative risk of stroke recurrence: a systematic review and meta-analysis. Stroke 2011;42:1489–94. https://doi.org/10.1161/STROKEAHA.110.602615
3. Kirchhof P, Benussi S, Kotecha D et al. 2016 ESC guidelines for the management of atrial fibrillation developed in collaboration with EACTS. Eur Heart J 2016;37:2893–962. https://doi.org/10.1093/eurheartj/ehw210
4. Plitt A, Bansilal S. The nonvitamin k antagonist oral anticoagulants and atrial fibrillation: challenges and considerations. J Atr Fibrillation 2017;9:1547. https://doi.org/10.4022/jafib.1547
5. Daiichi Sankyo UK Limited. Lixiana 60 mg Film-Coated Tablets. Summary of Product Characteristics. Available at: https://www.medicines.org.uk/emc/product/6905/smpc [accessed February 2019].
6. Giugliano RP, Ruff CT, Braunwald E et al. Edoxaban versus warfarin in patients with atrial fibrillation. N Engl J Med 2013;369:2093–104. https://doi.org/10.1056/NEJMoa1310907
7. Ruff CT, Giugliano RP, Antman EM et al. Evaluation of the novel factor Xa inhibitor edoxaban compared with warfarin in patients with atrial fibrillation: design and rationale for the Effective aNticoaGulation with factor xA next GEneration in Atrial Fibrillation-Thrombolysis In Myocardial Infarction study 48 (ENGAGE AF-TIMI 48). Am Heart J 2010;160:635–41. https://doi.org/10.1016/j.ahj.2010.06.042
8. Mendell J, Zahir H, Matsushima N et al. Drug-drug interaction studies of cardiovascular drugs involving P-glycoprotein, an efflux transporter, on the pharmacokinetics of edoxaban, an oral factor Xa inhibitor. Am J Cardiovasc Drugs 2013;13:331–42. https://doi.org/10.1007/s40256-013-0029-0
9. Lip GY, Halperin JL. Improving stroke risk stratification in atrial fibrillation. Am J Med 2010;123:484–8. https://doi.org/10.1016/j.amjmed.2009.12.013
10. Rost NS, Giugliano RP, Ruff CT et al. Outcomes with edoxaban versus warfarin in patients with previous cerebrovascular events: findings from ENGAGE AF-TIMI 48 (Effective Anticoagulation With Factor Xa Next Generation in Atrial Fibrillation-Thrombolysis in Myocardial Infarction 48). Stroke 2016;47:2075–82. https://doi.org/10.1161/STROKEAHA.116.013540
11. Giugliano RP, Ruff CT, Wiviott SD et al. Mortality in patients with atrial fibrillation randomized to edoxaban or warfarin: insights from the ENGAGE AF-TIMI 48 trial. Am J Med 2016;129:850–857.e2. https://doi.org/10.1016/j.amjmed.2016.02.028
12. Kato ET, Giugliano RP, Ruff CT et al. Efficacy and safety of edoxaban in elderly patients with atrial fibrillation in the ENGAGE AF-TIMI 48 trial. JAMA 2016;5:e003432. https://doi.org/10.1161/JAHA.116.003432
13. Steffel J, Giugliano RP, Braunwald E et al. Edoxaban versus warfarin in atrial fibrillation patients at risk of falling: ENGAGE AF-TIMI 48 analysis. J Am Coll Cardiol 2016;68:1169–78. https://doi.org/10.1016/j.jacc.2016.06.034
14. O’Donoghue ML, Ruff CT, Giugliano RP et al. Edoxaban vs. warfarin in vitamin K antagonist experienced and naive patients with atrial fibrillation. Eur Heart J 2015;36:1470–7. https://doi.org/10.1093/eurheartj/ehv014
15. Xu H, Ruff CT, Giugliano RP et al. Concomitant use of single antiplatelet therapy with edoxaban or warfarin in patients with atrial fibrillation: analysis from the ENGAGE AF-TIMI48 trial. J Am Heart Assoc 2016;5:e002587. https://doi.org/10.1161/JAHA.115.002587
16. Cavallari I, Ruff CT, Nordio F et al. Clinical events after interruption of anticoagulation in patients with atrial fibrillation: an analysis from the ENGAGE AF-TIMI 48 trial. Int J Cardiol 2018;257:102–07. https://doi.org/10.1016/j.ijcard.2018.01.065
17. Douketis JD, Murphy SA, Antman EM et al. Peri-operative adverse outcomes in patients with atrial fibrillation taking warfarin or edoxaban: analysis of the ENGAGE AF-TIMI 48 trial. Thromb Haemost 2018;118:1001–08. https://doi.org/10.1055/s-0038-1645856
18. Yamashita T, Koretsune Y, Yang Y et al. Edoxaban vs. warfarin in East Asian patients with atrial fibrillation – an ENGAGE AF-TIMI 48 subanalysis. Circ J 2016;80:860–9. https://doi.org/10.1253/circj.CJ-15-1082
19. Corbalán R, Nicolau JC, López-Sendon J et al. Edoxaban versus warfarin in Latin American patients with atrial fibrillation: the ENGAGE AF-TIMI 48 trial. J Am Coll Cardiol 2018;72:1466–75. https://doi.org/10.1016/j.jacc.2018.07.037
20. National Institute for Health and Care Excellence. Atrial fibrillation: management. CG180. London: NICE, 2014. Available from: https://www.nice.org.uk/guidance/cg180
21. National Institute for Health and Care Excellence. Edoxaban for preventing stroke and systemic embolism in people with non-valvular atrial fibrillation. TA355. London: NICE, 2015 (reviewed 2018). Available from: https://www.nice.org.uk/guidance/ta355