Efficacy and safety of edoxaban in comparison with dabigatran, rivaroxaban and apixaban for stroke prevention in atrial fibrillation

Transcription

1 New Technologies, Diagnostic Tools and Drugs 981 Efficacy and safety of edoxaban in comparison with dabigatran, rivaroxaban and apixaban for stroke prevention in atrial fibrillation An indirect comparison analysis Flemming Skjøth 1,2 ; Torben Bjerregaard Larsen 1,2 ; Lars Hvilsted Rasmussen 1,2 ; Gregory Y. H. Lip 1,3 1 Thrombosis Research Unit, Aalborg University, Aalborg, Denmark; 2 Department of Cardiology, Aalborg AF Study Group, Cardiovascular Research Centre, Aalborg University Hospital, Aalborg, Denmark; 3 University of Birmingham Centre for Cardiovascular Sciences, City Hospital, Birmingham, UK Summary Large Phase 3 clinical trials for stroke prevention in atrial fibrillation (AF) have compared non-vitamin K antagonist oral anticoagulants (NOACs) against warfarin, with the edoxaban trial only recently reported. In the absence of head to head trials directly comparing these NOACs against each other, we compared the efficacy and safety of edoxaban to other agents by an indirect comparison analysis. We performed an indirect comparison analysis of edoxaban (2 dose strategies) against apixaban (1 dose), dabigatran etexilate (2 doses) and rivaroxaban (1 dose), for their relative efficacy and safety against each other. For high-dose edoxaban vs apixaban, there were no significant differences in efficacy, mortality, myocardial infarction and major bleeding. Apixaban was associated with less major or clinically relevant non-major bleeding (hazard ratio [] 0.79; 95% confidence interval [CI] ) and gastrointestinal bleeding ( 0.72; 95% CI ). For dabigatran 110 mg twice daily, there were no significant differences in the main efficacy or safety. Dabigatran 150 mg bid was associated with lower stroke/systemic embolism (SE) ( 0.75; ), stroke ( 0.73; ) and haemorrhagic stroke ( 0.48; ). There were no significant differences between high-dose edoxaban vs rivaroxaban for efficacy or mortality, but rivaroxaban had more major and/or clinically relevant non-major bleeding. When compared to low-dose edoxaban, apixaban was associated with lower stroke/se ( 0.70; ), stroke ( 0.70; ) and ischaemic stroke ( 0.65; ), but more major bleeding ( 1.47; ). For dabigatran 110 mg bid, there were no significant differences in the efficacy, but dabigatran 110 mg bid had higher major (and gastrointestinal) bleeding. Dabigatran 150 mg bid and rivaroxaban were associated with lower stroke/se and ischaemic stroke, but higher bleeding rates. In the present analysis, we have provided for the first time, comparisons of efficacy and safety of edoxaban against other NOACs. Notwithstanding the significant limitations of an indirect comparison analysis, some differential effects are evident with the NOACs for stroke prevention, allowing us to allow the prescriber a choice to be able to fit the drug to the patient clinical profile (and vice versa). Keywords Atrial fibrillation, indirect comparisons, stroke prevention, dabigatran, apixaban, rivaroxaban, edoxaban Correspondence to: Flemming Skjøth, Phd, MSc Forskningens Hus Aalborg University Hospital Søndre Skovvej Aalborg, Denmark Tel.: , Fax: fls@rn.dk Received: February 8, 2014 Accepted after minor revision: February 18, 2014 Prepublished online: February 28, 2014 doi: /th Thromb Haemost 2014; 111: Note: The review process for this paper was fully handled by Christian Weber, Editor in Chief. Introduction The availability of non-vitamin K antagonist oral anticoagulants (NOACs; previously referred to as novel or new oral anticoagulants) have changed the landscape for stroke prevention in atrial fibrillation (AF). Following the first three Phase 3 randomised clinical trials (RE-LY [1], ROCKET-AF [2], ARISTOTLE [3]) comparing the NOACs (dabigatran etexilate, rivaroxaban, and apixaban, respectively) against warfarin for stroke prevention in atrial fibrillation (AF) (1), one more randomised trial (ENGAGE-AF) has been published (4). The NOACs have shown similar or greater efficacy for stroke prevention in their respective clinical trials, as well as similar or less major bleeding compared to warfarin (5). Importantly, all NOACs result in a lower risk of haemorrhagic stroke or intracranial haemorrhage compared to warfarin. Drug characteristics of the various NOACs are summarised in Suppl. Table 1 (available online at www. thrombosis-online.com). A question remains which of the NOACs is better, in relation to efficacy and safety, and the only robust way to assess this issue would be a head-to-head clinical trial. In the absence of such head-to-head trials, one way of doing such a comparison is by Schattauer 2014 Thrombosis and Haemostasis 111.5/2014

2 982 Skjøth et al. Stroke prevention in atrial fibrillation Table 1: Baseline characteristics of the Phase 3 clinical trials comparing NOACs. Study characteristics Study design Number of patients Follow up period, months Randomised groups Age, years # Female, % CHADS 2, (mean) CHADS 2 3 6, % Paroxysmal AF, % Prior stroke, TIA or SE, % Heart failure, % Prior myocardial Infarction, % Diabetes, % Hypertension, % Medication Aspirin, % Vitamin K antagonist, % TTR, % (obtained in warfarin arm) Endpoint crude event rates (%/yr)(obtained in warfarin arm) Stroke or SE Death from any cause Myocardial infarction ISTH major bleeding RE-LY (Dabigatran) Randomised open label 18, Dose adjusted warfarin vs blinded doses of dabigatran (150 mg bid, 110 mg bid) 71.5 ± * ROCKET-AF (Rivaroxaban) Multi-centre randomised double- blind, double-dummy 14, Dose adjusted warfarin vs rivaroxaban 20 mg OD 73 [65 78] * 1.1* 3.4* ARISTOTLE (Apixaban) Randomised control doubleblind parallel arm * Based on safety population during treatment. AF, atrial fibrillation; bid, twice daily; OD, once daily; SE, systemic embolism; TIA, transient ischaemic attack. # RE-LY: mean±sd, ROCKET-AF, ARISTOTLE, ENGAGE-AF: median [inter-quartile range]., Not available. 18, Dose adjusted warfarin vs apixaban 5 mg bid 70 [63 76] * ENGAGE-AF (Edoxaban) Three-group, randomised, double-blind, double-dummy 21, Dose adjusted warfarin vs high-dose or low-dose edoxaban strategy 72 [64 78] * an indirect comparison analysis (6, 7), assuming sufficient compatibility and exchangeability of the data between the compared trials. The basic study characteristics of the four randomised trials (1-4) are summarised in Table 1. The ROCKET-AF study enrolled a population with more comorbidity than the other studies as evident by a higher mean CHADS 2 score. The ENGAGE-AF and ROCKET-AF trials had more patients with heart failure and with hypertension, compared to RE-LY and ARISTOTLE. Overall, the comorbidity in the ENGAGE-AF population was intermediate between RE-LY/ARISTOTLE and ROCKET-AF. The performance of warfarin in the respective studies reflected by time in therapeutic range (TTR) ranged from 55% in ROCKET-AF to 68% in ENGAGE-AF. Several indirect comparison analyses have been published based on the overall clinical trial results (8-11). In our indirect comparison analysis (8), we found a significantly lower risk of stroke and systemic embolism (by 26%) for dabigatran (150 mg twice daily [bid]) compared with rivaroxaban, as well as haemorrhagic stroke and non-disabling stroke. There were no significant differences for apixaban vs dabigatran (both doses) or rivaroxaban; or rivaroxaban vs dabigatran 110 mg bid in preventing stroke and systemic embolism. For ischaemic stroke, there were no significant differences between the novel agents. Major bleeding was significantly lower with apixaban compared with dabigatran 150 mg bid (by 26%) and rivaroxaban (by 34%), but not significantly different from dabigatran 110 mg bid. When compared with rivaroxaban, dabigatran 110 mg bid was associated with less major bleeding (by 23%) and intracranial bleeding (by 54%). These results were also confirmed in other subsequent indirect comparison analyses. Thrombosis and Haemostasis 111.5/2014 Schattauer 2014

3 Skjøth et al. Stroke prevention in atrial fibrillation 983 In ENGAGE-AF (4) both low dose (30 mg, reduced to 15 mg depending on patient characteristics) and high dose (60 mg, reduced to 30 mg depending on patient characteristics) strategy arms of edoxaban were found non-inferior to warfarin for the primary efficacy endpoint (stroke/systemic embolism [SE]). In the intention-to-treat analysis, there was a trend favouring high-dose edoxaban vs warfarin (hazard ratio [], 0.87; 97.5% confidence interval [CI] ; p = 0.08) and an unfavourable trend with low-dose edoxaban versus warfarin (, 1.13; 97.5% CI ; p = 0.10) for the primary endpoint. Major bleeding was significantly less in both arms of edoxaban [high-dose, 0.80; ; p<0.001; low-dose, 0.47; ; p<0.001). Both doses were associated with significantly lower cardiovascular mortality. Following the recent publication of the ENGAGE-AF trial, the aim of the present study is to perform an indirect comparison analysis of edoxaban against apixaban, dabigatran (2 doses) and rivaroxaban, for their relative efficacy and safety against each other. The present analyses with edoxaban would complement prior indirect comparison analyses that have been only confined to comparisons of apixaban, dabigatran (2 doses) and rivaroxaban (8-11). Methods Efficacy and safety from the ENGAGE-AF (4), RE-LY (1), ROCKET-AF (2) and ARISTOTLE (3) clinical trials were reviewed for comparability and consistency of definition (see Suppl. Table 2, available online at Endpoints included in this analysis were reported in ENGAGE-AF and at least two other studies. Non-disabling stroke and death from vascular causes was not reported in ARISTOTLE. Systemic embolism and major or clinically relevant non-major bleeding were not reported by RE-LY. Other location bleeding was not reported in ROCKET-AF. Gastrointestinal (GI) bleeding was only reported by number of events by ROCKET-AF, the risk ratio was therefore used in the analysis, as an approximation of the. Study results were reported on the intention-to-treat population for all included from RE-LY and the primary efficacy endpoint stroke and SE in ROCKET-AF, and for all efficacy in ARIS- TOTLE and ENGAGE-AF. Secondary efficacy from ROCKET-AF and safety from ROCKET-AF, ARISTOTLE, and EN- GAGE-AF were reported on the safety on-treatment population. Reported direct comparisons against warfarin from each study are presented in Figure 1, representing the information upon which the indirect comparisons are based. We used the so-called Bucher method (12) for indirect comparisons using warfarin as a common comparator, which is a statistical method for estimating hazard rate ratio and corresponding uncertainty. The method is recommended as a preferred method for indirect comparison, superior to informal methods, as comparison of confidence intervals (12). Details on the method are described in our previous paper (8) for an analogous indirect com- Figure 1: Forest plots for direct comparisons between NOAC s (apixaban, dabigatran (110 mg BID and 150 mg BID), rivaroxaban, highdose edoxaban (60 mg, adjusted to 30 mg), and low-dose edoxaban (30 mg, adjusted to 15 mg) vs warfarin. parison of the RE-LY, ROCKET-AF and ARISTOTLE clinical trials. The estimated indirect relies on a similarity assumption (13) that the study s would likely have been obtained if the study populations of the two clinical trials were interchanged. A natural requirement of fulfilling the similarity assumption is that the clinical studies are absolutely comparable in terms of population characteristics and conduct of experiment (as highlighted earlier, there are some differences between the trials per se, as shown in Table 1). Primarily, the study population of the ROCKET-AF study generally had more comorbidity, and with the ENGAGE-AF study intermediate. The ENGAGE-AF achieved a higher level of treatment efficiency in terms of TTR in the warfarin arm. The crude event rates for stroke or systemic embolism, all-cause death, myocardial infarction (MI), and major bleeding in the warfarin arms of the studies (ROCKET-AF only major bleeding) were fairly comparable. The stroke event rates from ROCKET-AF were slightly higher, likely due to the more comor- Schattauer 2014 Thrombosis and Haemostasis 111.5/2014

4 984 Skjøth et al. Stroke prevention in atrial fibrillation bid population, whereas the rates of all-cause death and MI were lower, likely due to the evaluation on the safety population. The main focus in this analysis was the indirect comparisons of edoxaban (both doses) vs dabigatran (both doses), rivaroxaban and apixaban. Stata 11.2 (StataCorp LP), R v ( ject.org) and Microsoft Excel 2003 for Windows was used for the statistical analyses and graphical presentation. Ethical approval was not required for this analysis. Results Figure 2: Forest plots for indirect comparisons between high-dose edoxaban (60 mg, adjusted to 30 mg) and apixaban, dabigatran (110 mg BID and 150 mg BID) and rivaroxaban. Indirect comparisons for efficacy and safety with high-dose edoxaban (60 mg adjusted to 30 mg) are summarised in Figure 2 (exact numbers in Table 2 A). When compared to high-dose edoxaban, apixaban was not significantly different in efficacy, mortality, MI and major bleeding. Apixaban was associated with significantly less major or clinically relevant non-major bleeding ( 0.79; ) and GI bleeding ( 0.72; ). For dabigatran 110 mg bid, there were no significant differences in the main efficacy or safety, although other location bleeding was higher with dabigatran 110 mg bid. Dabigatran 150 mg bid was associated with lower stroke/ SE ( 0.75; ), stroke ( 0.73; ) and haemorrhagic stroke ( 0.48; ) but more other location bleeding. There was no significant difference with edoxaban high dose compared to rivaroxaban for efficacy or mortality, but rivaroxaban was associated with more major bleeding ( 1.30; ) and major or clinically relevant non-major bleeding ( 1.20; ). Indirect comparisons for efficacy and safety with low-dose edoxaban (30 mg adjusted to 15 mg) are summarised in Figure 3 (exact numbers in Table 2 B). When compared to low-dose edoxaban, apixaban was associated with significantly fewer strokes/ SE ( 0.70; ), stroke ( 0.70; ) and ischaemic stroke ( 0.65; ), but more major bleeding ( 1.47; ), as well as more bleeding in other location. For dabigatran 110 mg bid, there were no significant differences in the efficacy. Regarding safety, dabigatran 110 mg bid was associated with more major bleeding ( 1.70; ), more GI bleeding ( 1.64; ), and more bleeding in other locations ( 2.35; ). Dabigatran 150 mg bid was associated with lower stroke/se ( 0.58; ), stroke ( 0.57; ) and ischaemic stroke ( 0.54; ) and non-disabling stroke, with higher risk of bleeding : major bleeding ( 1.98; ), GI bleeding ( 2.24; ), and bleeding in other locations ( 2.68; ). Rivaroxaban was associated with less stroke/se ( 0.78; ), stroke ( 0.75; ), ischaemic stroke ( 0.67; ) and SE ( 0.19; ). Rivaroxaban was also associated with less MI ( 0.68; ). Rivaroxaban was associated with more bleeding : major bleeding ( 2.21; Thrombosis and Haemostasis 111.5/2014 Schattauer 2014

5 Skjøth et al. Stroke prevention in atrial fibrillation 985 Table 2A: Indirect estimates of () by the Bucher method: Edoxaban high-dose strategy (60 mg, adjusted to 30 mg). Apixaban <-> Edoxaban 60 Dabigatran 110 <-> Edoxaban 60 Dabigatran 150 <-> Edoxaban 60 Rivaroxaban <-> Edoxaban 60 Efficacy Stroke or Systemic embolism Stroke ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) Ischaemic or uncertain type of stroke 0.92 ( ) 1.11 ( ) 0.76 ( ) 0.94 ( ) Haemorrhagic stroke 0.94 ( ) 0.57 ( ) 0.48 ( ) 1.09 ( ) Systemic embolism 1.34 ( ) 0.35 ( ) Non-disabling stroke 1.08 ( ) 0.78 ( ) 1.29 ( ) Mortality Death from any cause Death from vascular causes 0.97 ( ) ( ) ( ) ( ) ( ) ( ) ( ) Other Myocardial infarction 0.94 ( ) 1.37 ( ) 1.35 ( ) 0.86 ( ) Bleeding ISTH major bleeding Major or clinically relevant non-major bleeding ( ) ( ) 1.00 ( ) 1.16 ( ) ( ) ( ) Intracranial bleeding 0.89 ( ) 0.66 ( ) 0.85 ( ) 1.43 ( ) Gastrointestinal bleeding 0.72 ( ) 0.89 ( ) 1.22 ( ) 1.18 ( ) Other location bleeding 1.27 ( ) 1.52 ( ) 1.73 ( ), Not available. Table 2B: Indirect estimates of () by the Bucher method: Edoxaban low-dose strategy (30 mg, adjusted to 15 mg). Apixaban <-> Edoxaban 30 Dabigatran 110 <-> Edoxaban 30 Dabigatran 150 <-> Edoxaban 30 Rivaroxaban <-> Edoxaban 30 Efficacy Stroke or systemic embolism Stroke ( ) ( ) ( ) ( ) ( ) ( ) ( ) ( ) Ischaemic or uncertain type of stroke 0.65 ( ) 0.79 ( ) 0.54 ( ) 0.67 ( ) Haemorrhagic stroke 1.55 ( ) 0.94 ( ) 0.79 ( ) 1.79 ( ) Systemic embolism 0.70 ( ) 0.19 ( ) Non-disabling stroke 0.77 ( ) 0.55 ( ) 0.92 ( ) Mortality Death from any cause Death from vascular causes 1.02 ( ) ( ) ( ) ( ) ( ) ( ) ( ) Other Myocardial infarction 0.74 ( ) 1.08 ( ) 1.07 ( ) 0.68 ( ) Bleeding ISTH major bleeding Major or clinically relevant non-major bleeding ( ) ( ) 1.70 ( ) 1.98 ( ) ( ) ( ) Intracranial bleeding 1.40 ( ) 1.03 ( ) 1.33 ( ) 2.23 ( ) Gastrointestinal bleeding 1.33 ( ) 1.64 ( ) 2.24 ( ) 2.18 ( ) Other location bleeding 1.98 ( ) 2.35 ( ) 2.68 ( ), Not available. Schattauer 2014 Thrombosis and Haemostasis 111.5/2014

6 986 Skjøth et al. Stroke prevention in atrial fibrillation ), major or clinically relevant non-major bleeding ( 1.66; ), intracranial bleeding ( 2.23; ), and GI bleeding ( 2.18; ). Discussion Figure 3: Forest plots for indirect comparisons between low-dose edoxaban (30 mg adjusted to 15 mg) and apixaban, dabigatran (110 mg BID and 150 mg BID) and rivaroxaban. In the present analysis, we have provided comparisons of efficacy and safety of edoxaban against other NOACs. Notwithstanding the significant limitations of an indirect comparison analysis, some differential effects are evident with the NOACs for stroke prevention in AF. We show that the high-dose edoxaban regime was broadly comparable in efficacy to apixaban, but apixaban was associated with lower major or clinically relevant non-major bleeding or GI bleeding. High-dose edoxaban was broadly comparable in efficacy and safety to dabigatran 110 mg bid, but had lower efficacy compared to dabigatran 150 mg bid. There were no differences in efficacy between high dose edoxaban and rivaroxaban, but the latter was associated with more bleeding. Low dose edoxaban was less efficacious compared to apixaban, dabigatran 150 mg bid and rivaroxaban, but had less major bleedings and was generally more safe than all the four alternatives. The ENGAGE-AF trial (4) was the largest of the Phase 3 clinical trials of NOACs compared against warfarin for stroke prevention in AF. In contrast to prior trials, ENGAGE-AF had an adaptive dose design with the use of a 50% lower dose in the treatment arm in the presence 1 of the following: creatinine clearance (CrCl) ml/minute, body weight 60 kg, or concomitant verapamil or quinidine [strong P-gp inhibitors]). For the efficacy, high dose edoxaban was broadly comparable to apixaban, dabigatran 110 mg bid and rivaroxaban, but inferior to dabigatran 150 mg bid. For safety, high-dose edoxaban had similar major bleeding compared to dabigatran (both doses), less major bleeding compared to rivaroxaban, but more major bleeding compared to apixaban. Nonetheless, the high-dose edoxaban regime was associated with more GI bleeding compared to warfarin, as was seen with dabigatran 150 mg bid and rivaroxaban. Only apixaban was not associated with an increase in GI bleeding. In contrast, the low-dose edoxaban had poorer efficacy compared to apixaban, dabigatran 150 mg bid and rivaroxaban. Efficacy was broadly similar to dabigatran 110 mg bid. Safety was generally in favour of edoxaban. In ENGAGE-AF (4), the low-dose edoxaban strategy was associated with a numerical but non-significant increase in MI, with no significance seen for high dose edoxaban. In the HOKUSAI trial of venous thromboembolism treatment, a higher rate of MI was evident with edoxaban-treated patients compared to warfarin (14). A similar numerical increase in MI was seen with dabigatran (both doses) compared to warfarin in RE-LY (15), but this may be related to quality of anticoagulation control, with higher myocardial infarction rates where TTR was <65% compared to those >65% (16). However, the small numerical difference needs to be put in context of the large reduction in haemorrhagic stroke, mortality etc. seen with edoxaban in EN- GAGE-AF. Published post-marketing data in real world populations do not suggest any excess in MI compared to warfarin (17). Thrombosis and Haemostasis 111.5/2014 Schattauer 2014

7 Skjøth et al. Stroke prevention in atrial fibrillation 987 It is more likely that warfarin is a good therapy for preventing MI, rather than the higher MI rates resulting from a bad effect of NOACs per se (18). A recent meta-analysis of the NOAC trials shows a higher risk of MI with low-dose NOAC regimes (dabigatran 110 mg bid, edoxaban low-dose strategy) compared to warfarin (19). In the ENGAGE-AF trial (4), whilst both high- and low-dose edoxaban were non-inferior to warfarin for all stroke and systemic embolism (primary endpoint), this was driven a marked reduction in haemorrhagic stroke. Indeed, ischaemic stroke was similar to warfarin with high dose edoxaban, and increased (by 40%) with low-dose edoxaban compared to warfarin. Interestingly, all-cause and cardiovascular mortality was lower with both dose regimes of edoxaban. In the historical trials of warfarin vs placebo/control, the use of warfarin resulted in a significant reduction in all stroke (by 64%) and all-cause mortality (by 26%) (20). Thus, despite the higher ischaemic stroke rate with low-dose edoxaban compared to warfarin, the marked reduction in haemorrhagic stroke resulted in non-inferiority of low-dose edoxaban to warfarin for the primary endpoint of all stroke and SE. The safety of edoxaban was evident, with lower major bleeding and intracranial bleeding compared to warfarin with both edoxaban doses. This analysis extends previous studies (8-11) of indirect comparisons of NOACs. However, high quality anticoagulation with well controlled adjusted dose warfarin is associated with less stroke and serious bleeding (21-24), and this indirect comparisons cannot adjust for this, nor address this aspect. Of note patients in the highest quartile of average centre-based (not individual) TTR may have less benefit on stroke and systemic embolism with a novel drug compared to warfarin (25). Limitations All indirect comparison analyses have inherent limitations and are no substitutes for head-to-head randomised trials. Until such trials are done, indirect comparisons are one statistical method used to compare different drugs, and have been used to assist with regulatory submissions and/or cost effectiveness comparisons (26). Patient selection was different in the four trials, and quality of anticoagulation control (as reflected by average TTR) in the comparator (warfarin) arm was different in the four trials, particularly for ROCKET-AF (which had average TTR 55%) whilst the other three trials had better international normalized ratio control (average TTRs >60%). Also, RE-LY was a three-arm open clinical trial with powered direct comparisons between warfarin, dabigatran 150 mg bid and dabigatran 110 mg bid; in contrast, the other trials (ROCKET-AF, ARISTOTLE, ENGAGE-AF) were double-blind trials and depending on patient characteristics (e.g. moderate renal impairment), dose-adjustments were made to give lower doses to particular patient subgroups (5, 27). Thus, it is not entirely clear whether the indirect comparisons can be considered straightforward, and indeed, comparisons of this kind can only be considered to be hypothesis generating and the basis for a head-to-head trial. Notwithstanding the significant limitations of an indirect comparison analysis, some differential effects are evident with the NOACs for stroke prevention, allowing a prescriber a choice to be able to fit the drug to the patient clinical profile (and vice versa). In this respect, indirect comparisons allows an informed choice (for example) whether the prescriber is aiming for higher efficacy in reducing ischaemic stroke (e.g. dabigatran 150 mg bid) or greater safety with lowest major bleeding profile, especially in the elderly or those at high bleeding risk (e.g. dabigatran 110 mg bid, apixaban or edoxaban). Conclusion We have provided for the first time, a comparison of the efficacy and safety of edoxaban against other NOACs. In this analysis, high dose edoxaban was comparable in efficacy to apixaban, but apixaban had less bleeding. High dose edoxaban was broadly comparable in efficacy and safety to dabigatran 110 mg bid, but had lower efficacy compared to dabigatran 150 mg bid. There were no differences in efficacy between high-dose edoxaban compared to rivaroxaban. Low-dose edoxaban was less efficacious compared to the other drugs, but had less major bleeding. These results are hypothesis generating and should be confirmed in a head-to-head randomised trial. What is known about this topic? Since the initial Phase 3 randomised clinical trials comparing the non-vitamin K antagonist oral anticoagulants (NOACs) against warfarin for stroke prevention in atrial fibrillation (AF), one further Phase 3 clinical (ENGAGE-AF) has been published. In the absence of head to head trials directly comparing the NOACs against each other, an accepted method of comparing the efficacy and safety of edoxaban to other agents (dabigatran, rivaroxaban, apixaban) is by performing an indirect comparison analysis. What does this paper add? In the present analysis, we have provided for the first time, comparisons of efficacy and safety of edoxaban against other NOACs. For high-dose edoxaban vs apixaban, there were no significant differences in efficacy, mortality, myocardial infarction and major bleeding. For dabigatran 110 mg bid, there were no significant differences in the main efficacy or safety. Dabigatran 150 mg bid was associated with lower stroke/systemic embolism (SE), stroke and haemorrhagic stroke. There were no significant differences between high-dose edoxaban vs rivaroxaban for efficacy or mortality, but rivaroxaban had more bleeding. When compared to low-dose edoxaban, apixaban was associated with lower stroke/se and ischaemic stroke, but more major bleeding. For dabigatran 110 mg bid, there were no significant differences in the efficacy, but had more major (and gastrointestinal) bleeding. Dabigatran 150 mg bid and rivaroxaban were associated with lower stroke/se and ischaemic stroke, but higher bleeding rates. Schattauer 2014 Thrombosis and Haemostasis 111.5/2014

8 988 Skjøth et al. Stroke prevention in atrial fibrillation Conflicts of interest Prof Lip has served as a consultant for Bayer, Astellas, Merck, AstraZeneca, Sanofi, BMS/Pfizer, Daiichi-Sankyo, Medtronic, Biotronik and Boehringer Ingelheim, and has been on the speaker bureaus for Bayer, BMS/Pfizer, Boehringer Ingelheim, Daiichi- Sankyo, Medtronic and Sanofi. TBL and L have served as a speaker for BMS/Pfizer, Boehringer Ingelheim. FS none declared. References 1. Connolly SJ, Ezekowitz MD, Yusuf S, et al. Dabigatran versus warfarin in patients with atrial fibrillation. N Engl J Med 2009; 361: Patel MR, Mahaffey KW, Garg J, et al. Rivaroxaban versus warfarin in nonvalvular atrial fibrillation. N Engl J Med 2011; 365: Granger CB, Alexander JH, McMurray JJ, et al. Apixaban versus warfarin in patients with atrial fibrillation. N Engl J Med 2011; 365: Giugliano RP, Ruff CT, Braunwald E, et al. Edoxaban versus warfarin in patients with atrial fibrillation. N Engl J Med 2013; 369: Potpara TS, Lip GY. Novel oral anticoagulants in non-valvular atrial fibrillation. Best Pract Res Clin Haematol 2013; 26: Skjøth F, Larsen TB, Rasmussen LH. Indirect comparison studies are they useful? Insights from the novel oral anticoagulants for stroke prevention in atrial fibrillation. Thromb Haemost 2012; 108: Harenberg J, Marx S, Wehling M. Head-to-head or indirect comparisons of the novel oral anticoagulants in atrial fibrillation: What s next? Thromb Haemost 2012; 108: Lip GY, Larsen TB, Skjøth F, et al. Indirect comparisons of new oral anticoagulant drugs for efficacy and safety when used for stroke prevention in atrial fibrillation. J Am Coll Cardiol 2012; 60: Kansal AR, Sharma M, Bradley-Kennedy C, et al. Dabigatran versus rivaroxaban for the prevention of stroke and systemic embolism in atrial fibrillation in canada. Comparative efficacy and cost-effectiveness. Thromb Haemost 2012; 108: Rasmussen LH, Larsen TB, Graungaard T, et al. Primary and secondary prevention with new oral anticoagulant drugs for stroke prevention in atrial fibrillation: Indirect comparison analysis. Br Med J 2012; 345: e Mantha S, Ansell J. An indirect comparison of dabigatran, rivaroxaban and apixaban for atrial fibrillation. Thromb Haemost 2012; 108: Bucher HC, Guyatt GH, Griffith LE, et al. The results of direct and indirect treatment comparisons in meta-analysis of randomized controlled trials. J Clin Epidemiol 1997; 50: Song F, Loke YK, Walsh T, et al. Methodological problems in the use of indirect comparisons for evaluating healthcare interventions: Survey of published systematic reviews. Br Med J 2009; 338: b Buller, Decousus H, Grosso MA, et al. Edoxaban versus warfarin for the treatment of symptomatic venous thromboembolism. N Engl J Med 2013; 369: Connolly SJ, Ezekowitz MD, Yusuf S, Reilly PA, Wallentin L. Newly identified events in the re-ly trial. N Engl J Med 2010; 363: Clemens A, Fraessdorf M, Friedman J. Cardiovascular outcomes during treatment with dabigatran: Comprehensive analysis of individual subject data by treatment. Vasc Health Risk Manag 2013; 9: Larsen TB, Rasmussen LH, Skjøth F, et al. Efficacy and safety of dabigatran etexilate and warfarin in real world patients with atrial fibrillation: A prospective nationwide cohort study. J Am Coll Cardiol 2013; ((please complete reference)). 18. Lip GY, Lane DA. Does warfarin for stroke thromboprophylaxis protect against mi in atrial fibrillation patients? Am J Med 2010; 123: Ruff CT. Comprehensive meta-analysis comparing the efficacy and safety of new oral anticoagulants with warfarin in atrial fibrillation: An analysis including 71,683 patients from four large randomized clinical trials. Lancet 2013; 61: Hart RG, Pearce LA, Aguilar MI. Meta-analysis: Antithrombotic therapy to prevent stroke in patients who have nonvalvular atrial fibrillation. Ann Intern Med 2007; 146: Wan Y, Heneghan C, Perera R, et al. Anticoagulation control and prediction of adverse events in patients with atrial fibrillation: A systematic review. Circ Cardiovasc Qual Outcomes 2008; 1: Gallego P, Roldan V, Marin F, et al. Cessation of oral anticoagulation in relation to mortality and the risk of thrombotic events in patients with atrial fibrillation. Thromb Haemost 2013; 110: De Caterina R, Husted S, Wallentin L, et al. General mechanisms of coagulation and targets of anticoagulants (section i). Position paper of the ESC working group on thrombosis--task force on anticoagulants in heart disease. Thromb Haemost 2013; 109: De Caterina R, Husted S, Wallentin L, et al. Vitamin k antagonists in heart disease: Current status and perspectives (section iii). Position paper of the ESC working group on thrombosis task force on anticoagulants in heart disease. Thromb Haemost 2013; 110: Wallentin L, Yusuf S, Ezekowitz MD, et al. Efficacy and safety of dabigatran compared with warfarin at different levels of international normalised ratio control for stroke prevention in atrial fibrillation: An analysis of the RE-LY trial. Lancet 2010; 376: Antithrombotic agents for the prevention of stroke and systemic embolism in patients with atrial fibrillation. Ottawa ON: CADTH February 2013; O Neil WM, Welner SA, Lip GY. Do open label blinded outcome studies of novel anticoagulants versus warfarin have equivalent validity to those carried out under double-blind conditions? Thromb Haemost 2013; 109: Thrombosis and Haemostasis 111.5/2014 Schattauer 2014