A single institution’s experience with using dabigatran,rivaroxaban and warfarin for prevention of thromboembolism in atrial fibrillation

Abstract

Background:

Although non-vitamin K antagonist oral anticoagulants (NOACs) are approved for stroke prevention in atrial fibrillation (AF) patients, their use in the local clinical setting has not been well studied. This study aims to evaluate the clinical outcomes of dabigatran, rivaroxaban and warfarin in a tertiary hospital in Singapore.

Methods:

This is a retrospective cohort study with one-year follow-up. A total of 383 patients recruited between June 2011 and December 2014 were studied. Incidents of stroke, systemic embolism and clinically relevant bleeding events were compared between dabigatran, rivaroxaban and warfarin.

Results:

Stroke rates were 5.47% per year with warfarin, 7.27% per year with dabigatran (HR=1.32; 95% CI 0.48−3.64; p=0.591) and 2.76% per year with rivaroxaban (HR=0.49; 95% CI 0.14−1.69; p=0.261). The warfarin group had significantly higher incidence of minor bleeding (62.4% vs 3.64% for dabigatran vs 13.79% for rivaroxaban; p<0.001), major bleeding (3.91% for warfarin, 0.91% for dabigatran, 0% for rivaroxaban; p=0.028) and other adverse events (51.18% for warfarin, 3.64% for dabigatran, 8.28% for rivaroxaban; p<0.001). Incidence of dyspepsia was higher in both NOAC groups compared to warfarin (0% for warfarin, 7.27% for dabigatran, 5.52% for rivaroxaban; p=0.003).

Conclusion:

Stroke and venous thromboembolism rates after one year were comparable among dabigatran, rivaroxaban and warfarin. Warfarin was associated with more bleeding and adverse events while both NOACs were associated with higher rates of dyspepsia. Further study is needed to assess the clinical benefit of NOACs in the Singaporean population.

Keywords

Warfarin atrial fibrillation bleeding stroke non-vitamin K antagonist oral anticoagulant

Introduction

Atrial fibrillation (AF) is the most common form of cardiac arrhythmia, affecting about one in four, and it is associated with a three to fivefold increased risk of stroke.^1,2 This risk can be reduced with the use of anticoagulation.³ Traditionally, warfarin, a vitamin K antagonist, is the drug of choice for oral anticoagulation.⁴ However, its popularity is limited by its numerous interactions with drugs, herbs and foods, as well as the need for routine anticoagulation monitoring.⁵

Non-vitamin K oral anticoagulants (NOACs) were developed in an attempt to overcome the disadvantages of warfarin. Unlike warfarin, they have a more predictable dose response and do not require routine monitoring. Currently available NOACs fall into two classes: direct thrombin inhibitors (e.g. dabigatran) and factor Xa inhibitors (e.g. rivaroxaban, apixaban).^4,6 They have demonstrated better, or at least equivalent, efficacy compared to warfarin in major landmark trials without compromising on safety.^7,8

At present, there is a lack of post-marketing data on the use of dabigatran and rivaroxaban in Singapore. Thus, our study aimed to compare the clinical outcomes of dabigatran, rivaroxaban and warfarin in a tertiary hospital in Singapore.

Methods

Study design

A one-year retrospective cohort study was conducted at Tan Tock Seng Hospital (TTSH). Recruited patients were followed up for a year from the date of initiation of oral anticoagulant.

Study population

Patients who fulfilled the following criteria were recruited: (1) have their oral anticoagulation initiated at TTSH; (2) initiated on warfarin from June 2011 to July 2012 for AF; (3) started on dabigatran (between March 2011 and July 2012) or rivaroxaban (between June 2012 and December 2013) for AF; and (4) are intended for long-term anticoagulation. Patients whose anticoagulation therapy was neither started nor followed through at TTSH, intended for short-term use (less than one year) or used to treat venous thromboembolism (VTE) were excluded.

The list of eligible patients on dabigatran and rivaroxaban was generated using the pharmacy dispensing system iPharm, while patients who were newly prescribed with warfarin for AF were identified using TTSH Anticoagulation Clinic’s database. The patient list was then reviewed to exclude patients who failed to fulfill the inclusion criteria.

Data source

Data were retrieved from case notes and computerized patient support systems (CPSS).

Study drugs

The anticoagulant formulations available at TTSH are dabigatran 75, 110 and 150mg capsules, rivaroxaban 10, 15 and 20mg tablets, and warfarin 1, 3 and 5mg tablets.

Outcomes

The primary outcome was incidence of stroke (ischaemic and haemorrhagic) or systemic embolic events. Secondary outcomes included clinically relevant bleeding events as well as other adverse events. Major bleeding events were defined using International Society on Thrombosis and Haemostasis (ISTH) criteria [i.e. fatal bleeding; bleeding in a critical site (intracranial, intraspinal, intraocular, retroperitoneal, intra-articular or pericardial, or intramuscular with compartment syndrome); drop in haemoglobin concentration of more than or equal to 2.0g/dL; or transfusion of >2U of whole blood or packed red blood cells].¹² Minor bleeding events were defined as any bleeding event that did not meet the criteria for major bleeding.

Statistical analysis

Descriptive statistics were used to summarize the patient demographics. Continuous variables were presented as mean (standard deviation) and categorical variables were summarized using frequency (percentages). Chi-square and ANOVA tests were used to compare baseline characteristics as well as rates of stroke, venous thromboembolism and adverse events among the three treatment groups. A statistically significant difference was defined as p-value <0.05. Hazard ratios, confidence intervals and p values were calculated using Cox regression.

All analyses were based on the intention-to-treat principle, and conducted using Stata v13.1 (Stata Corp, College Station, Texas, USA).

Results

Patient characteristics

A total of 137 warfarin, 168 dabigatran and 279 rivaroxaban patients were identified. After applying the exclusion criteria, 128 warfarin, 110 dabigatran patients and 145 rivaroxaban patients were eligible. Baseline demographics are summarized in Table 1.

Table 1.

Baseline demographics of patients.

Variable	Treatment drug			p-value
Variable	Warfarin (n=128)	Dabigatran (n=110)	Rivaroxaban (n=145)	p-value
Gender
Male	69 (53.91%)	61 (55.45%)	74 (51.03%)	0.77
Race				0.148
Chinese	110 (85.94%)	96 (87.27%)	121 (83.45%)
Malay	13 (10.16%)	5 (4.55%)	14 (9.66%)
Indian	3 (2.34%)	1 (0.91%)	5 (3.45%)
Others	2 (1.56%)	8 (7.27%)	5 (3.45%)
Age, mean (years) (SD)	69.75 (10.39)	70.88 (9.47)	72.19 (9.66)	0.124
Diabetes mellitus	53 (41.41%)	44 (40.00%)	53 (36.55%)	0.698
Hypertension	104 (81.25%)	100 (90.91%)	110 (75.86%)	0.008
CHF or LV dysfunction	36 (28.13%)	19 (17.27%)	31 (21.38%)	0.125
Previous stroke or TIA	67 (52.34%)	50 (45.45%)	65 (44.83%)	0.367
Previous systemic embolic event	0 (0%)	9 (8.18%)	9 (6.21%)	0.001
Known vascular disease	9 (7.03%)	6 (5.45%)	10 (6.9%)	0.864
History of MI	20 (15.63%)	17 (15.45%)	19 (13.1%)	0.806
Mean CHADS₂ score (SD)	2.95 (1.32)	2.75 (1.21)	2.71 (1.56)	0.346
Mean CHA2DS2VASc score (SD)	4.30 (1.78)	4.20 (1.57)	4.06 (1.95)	0.5522
Mean CrCl (SD)	65.87 (29.97)	72.24 (25.24)	58.89 (24.82)	<0.001
CrCl Range				0.021
>50ml/min	87 (67.97%)	86 (78.18%)	82 (56.55%)
30−50ml/min	34 (26.56%)	21 (19.09%)	54 (37.24%)
15−29ml/min	6 (4.69%)	3 (2.73%)	7 (4.83%)
<15ml/min	1 (0.78%)	0 (0%)	0 (0%)
Missing	0 (0%)	0 (0%)	2 (1.38%)
Hepatic impairment	0 (0%)	1 (0.91%)	4 (2.76%)	0.133
Bleed history	52 (40.63%)	22 (20%)	46 (31.72%)	0.003
Previous antiplatelet use	92 (71.88%)	48 (43.64%)	90 (62.07%)	<0.001
On concurrent antiplatelet	34 (26.56%)	14 (12.73%)	27 (18.62%)	0.026
Mean HASBLED score (SD)	2.73 (1.28)	2.48 (1.04)	2.57 (1.25)	0.2659

CHF: congestive heart failure; CHADS₂: congestive heart failure, hypertension, age ⩾75 years, diabetes mellitus, stroke/transient ischaemic attack; CHA₂DS₂VASc: congestive heart failure, hypertension, age ⩾75 years, diabetes mellitus, stroke/transient ischaemic attack/thromboembolism, age 65−74, vascular disease; CrCl: creatinine clearance; LV: left ventricular dysfunction; SD: standard deviation; TIA: transient ischaemic attack.

Primary outcome

No systemic embolic event occurred in all treatment groups. Stroke occurred in seven patients receiving warfarin, eight patients receiving dabigatran, and four patients receiving rivaroxaban. Stroke rates were 5.47 per 100 patient years with warfarin, 7.27 per 100 patient years with dabigatran (HR=1.32; 95% CI = 0.48−3.64; p-value=0.591) and 2.76 per 100 patient years with rivaroxaban (HR=0.49; 95% CI = 0.14−1.69; p-value=0.261) (Table 2).

Table 2.

Cox regression model comparing stroke rates of NOACs vs warfarin.

Outcome	Warfarin			Dabigatran					Rivaroxaban
Outcome	No. of patients	No. of events	Event rate/100 patient years	No. of patients	No. of events	Event rate/100 patient years	Hazard Ratio(95% CI)	p-value	No. of patients	No. of events	Event rate/100 patient years	Hazard Ratio(95% CI)	p-value
Stroke after 1 year	128	7	5.47	110	8	7.27	1.32 (0.48−3.64)	0.591	145	4	2.76	0.49 (0.14−1.69)	0.261

The median follow-up time was 365 days.

Hazard ratios are for the dabigatran and rivaroxaban group as compared with the warfarin group.

Secondary outcomes

For safety outcomes, warfarin appeared to be associated with significantly higher rates of minor bleeding events (p<0.001), major bleeding events (p=0.028) and other adverse events (p<0.001) as compared to both NOACs. However, dyspepsia rates were significantly higher (p<0.001) for both NOACs compared with warfarin (Table 3).

Table 3.

Rates of bleeding and other adverse events.

Adverse event	Treatment drug			p-value
Adverse event	Warfarin (n=128)	Dabigatran (n=110)	Rivaroxaban (n=145)	p-value
Minor bleed	77 (61.60%)	4 (3.64%)	14 (9.66%)	<0.001
GI bleeding	3 (2.34%)	2 (1.82%)	4 (2.76%)	0.886
Intracranial bleeding	1 (0.78%)	0 (0%)	2 (1.38%)	0.465
Drop in Hb ⩾2g/dL	3 (2.36%)	3 (2.73%)	4 (2.76%)	1
Major bleed	5 (3.91%)	1 (0.91%)	0 (0.00%)	0.028
Dyspepsia	0 (0.00%)	8 (7.27%)	8 (5.52%)	0.003
Other adverse events	65 (51.18%)	4 (3.64%)	11 (7.59%)	<0.001
Death	0 (0.00%)	1 (0.91%)	7 (4.83%)	0.012
Treatment-related death	0 (0.00%)	1 (0.91%)	1 (0.69%)	0.7463

Comparison of dabigatran and rivaroxaban

The incidence of stroke and VTE was not significantly different between the dabigatran and rivaroxaban groups. Secondary outcomes like bleeding, death and other adverse events also did not differ significantly between the two groups (Table 4).

Table 4.

Rates of stroke, bleeding and other adverse events for dabigatran and rivaroxaban.

Event	Dabigatran (n=110)	Rivaroxaban (n=145)	p-value
Stroke	8 (7.27%)	4 (2.76%)	0.134
Minor bleed	4 (3.64%)	14 (9.66%)	0.084
Major bleed	1 (0.91%)	0 (0.00%)	0.431
Dyspepsia	8 (7.27%)	8 (5.52%)	0.567
Other adverse events	4 (3.64%)	11 (7.59%)	0.192
Death	1 (0.91%)	7 (4.83%)	0.143

Discussion

To our knowledge, this is the first local study to examine the clinical outcomes of dabigatran and rivaroxaban compared to warfarin in AF patients within TTSH. Incidence of stroke, systemic embolism, bleeding and other adverse events were compared.

Some differences in prescribing of oral anticoagulants were observed in this study. Firstly, although there was no significant difference in the baseline stroke risk (as determined by the CHA₂DS₂VASc scores), we observed that a higher proportion of patients on NOACs had previously experienced a systemic embolic event. This could be due to the high proportion of such patients in the clinical trial,⁸ which increased physicians’ confidence in prescribing NOACs for these patients. Since this study also included patients who were not anticoagulant-naïve, we speculate that some patients may have been switched from warfarin to a NOAC after they experienced a systemic embolic event while on warfarin.

Secondly, although there was no significant difference in baseline bleeding risk (as determined by the HASBLED score), we observed that there was a significantly higher proportion of warfarin patients who either had a history of bleed or were on concurrent antiplatelets. This could be due to the fact that combination antithrombotic therapy (e.g. oral anticoagulant with either one or two antiplatelet agents) is commonly indicated for patients with AF with acute coronary syndrome and/or percutaneous coronary intervention and due to limited evidence available for NOACs, warfarin is currently the preferred anticoagulant of choice in such cases.^18,19

We noticed that the stroke rates reported in this study appeared to be much higher than those reported in clinical trials such as the RE-LY [Randomized Evaluation of Long-Term Anticoagulation Therapy (RE-LY)] trial for dabigatran and ROCKET-AF (Rivaroxaban Once Daily Oral Direct Factor Xa Inhibition Compared with Vitamin K Antagonism for Prevention of Stroke and Embolism Trial in Atrial Fibrillation) trial for rivaroxaban; even the ROCKET-AF trial, which recruited patients with a higher baseline stroke risk (mean CHADS₂ score of 3.48±0.94 for the rivaroxaban group vs 3.46±0.95 for the warfarin group), showed overall lower stroke rates.^7,8 This discrepancy could be attributed to differences in design (retrospective cohort study versus randomized controlled trial), sample sizes, recruitment criteria and follow-up duration.

In this study, there was no significant difference in stroke rates between dabigatran, rivaroxaban and warfarin. These results are largely consistent with that from RE-LY and ROCKET-AF, which showed that both dabigatran dosed at 110mg twice daily and rivaroxaban had comparable stroke risk to warfarin.^7,8 Although the RE-LY trial showed that dabigatran dosed at 150mg twice daily was superior to warfarin in terms of stroke prevention, this difference was not detected in our study.⁷ A possible reason could be that this study did not distinguish between patients on different doses of dabigatran as the sample size was too small. Nonetheless, several other observational studies have reported similar effectiveness between various doses of dabigatran and warfarin in the “real-world” clinical setting, suggesting that dabigatran 150mg’s superior effect may be diminished in clinical practice.^20,21,22

Warfarin was associated with significantly higher rates of bleeding and adverse events compared to dabigatran and rivaroxaban. This was partly similar to the results of the RE-LY study, where warfarin had higher major and minor bleed rates than dabigatran dosed at 110mg twice daily, and higher minor bleed rates than dabigatran dosed at 150mg twice daily.⁷ One reason for the higher rates of bleeding and adverse events for warfarin in this study could be different monitoring frequency of patients in different treatment groups. In this study, pharmacists from the TTSH Anticoagulation Clinic team followed up with warfarin patients regularly at intervals ranging from two weeks to three months depending on the stability of their International Normalised Ratio (INR). At the follow-up sessions, patients were asked to report bleeding and adverse events that occurred during the treatment period. Patients on NOACs, on the other hand, were not followed up as frequently as those on warfarin. The longer follow-up duration, which ranged from three to six months, for these patients could have resulted in reporting bias.

We recognize, however, that the minor bleeding rates for the warfarin arm in this study could have been falsely elevated. This is because unlike other studies, our study did not limit the definition of “minor bleed” only to bleeding events that were clinically relevant.^7,8,23,24 As such, some nuisance bleeding events could have been included into the analysis even though they were not clinically significant. Nonetheless, our findings are consistent with that from other studies which showed lower bleed rates with NOACs when compared to warfarin.^22,24,26 A large prospective observational study involving 6784 patients from 311 centres across Europe, Israel and Canada also demonstrated that rates of major bleeding were low (2.1 events/100 patient years, compared to 3.6 events/100 patient years in ROCKET-AF) in patients on rivaroxaban in clinical practice.²⁵

In this study, both NOACs were associated with significantly higher rates of dyspepsia compared to warfarin. This was observed in RE-LY, where dabigatran dosed at both 110mg twice a day and 150mg twice a day was associated with significantly higher rates of dyspepsia compared to warfarin.⁷ Some studies have also cited dyspepsia as a common reason for treatment discontinuation.^27,28,30 It has been postulated that this gastrointestinal side effect is caused by tartaric acid, present in the core of dabigatran capsules, which is used to create an acidic environment for drug absorption.²⁹. In contrast, dyspepsia is not a commonly reported symptom for rivaroxaban in literature.³¹ It was also not evaluated in the ROCKET-AF study.⁸ As we did not collect information on factors that may predispose patients to dyspepsia, such as any history of dyspepsia, concomitant use of medications that may contribute to dyspepsia (e.g. non-steroidal anti-inflammatory medications, antibiotics, steroids), or any concomitant use of gastric protectants (e.g. proton pump inhibitors, H2 receptor antagonists), it is not known if the higher rates of dyspepsia observed with rivaroxaban in this study were due to these factors. More information would be needed before any conclusion can be made.

Although the death rate in the rivaroxaban group was significantly higher than the other groups (4.83% vs 0.91% for dabigatran and 0% for warfarin, p=0.012), only one case appeared to be treatment-related because the subject had died of spontaneous intracerebral haemorrhage. Among the other death cases, three were due to infections, one was due to non-ischaemic cardiomyopathy while reasons for the other deaths were unspecified. The single case of death in the dabigatran group was due to non-ST-segment elevation myocardial infarction (NSTEMI), and it could have been treatment-related since dabigatran was found to be associated with increased risk of myocardial infarction when compared with warfarin in several randomized controlled trials.^10,11

There were no significant differences in rates of stroke, VTE and adverse events between dabigatran and rivaroxaban. This suggests that the NOACs can be selected based on patient factors and preference.

Study limitations and future directions

Apart from the aforementioned reasons, other limitations of this study include its small sample size and short follow-up duration. Due to its small sample size, propensity matching could not be done to adjust for baseline differences and this limited the validity of our data. In addition, adherence to study medications, as well as the time in therapeutic INR range (TTR) for patients on warfarin were not measured.

Future studies should aim to have a longer follow-up duration to assess the long-term clinical outcomes of NOACs, since these drugs are usually intended for long-term use. Other future directions to consider include evaluating the individual doses of NOACs (e.g. dabigatran 110mg twice daily vs 150mg twice daily dosing), as well as adherence to treatment drugs.

Conclusion

In this study, stroke and venous thromboembolism rates appeared to be comparable among dabigatran, rivaroxaban and warfarin. Warfarin was associated with more bleeding and adverse events while the NOACs were associated with higher rates of dyspepsia. However, due to the observational nature and small sample size of this study, more studies with larger sample sizes are required to affirm the real-world benefit of NOACs, and to provide guidance for the choice of anticoagulation in AF patients.

Footnotes

Acknowledgements

The authors thank medical statistician Sun Bing for her help in statistical analysis.

Declaration of conflicting interests

The authors declare that there is no conflict of interest.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

References

Lloyd-Jones

Wang

Leip

et al . Lifetime risk for development of atrial fibrillation: The Framingham Heart Study. Circulation 2004; 110: 1042–1046.

Wolf

Abbott

Kannel

. Atrial fibrillation as an independent risk factor for stroke: The Framingham Study. Stroke 1991; 22: 983–988.

Hart

Pearce

Aguilar

. Meta-analysis: Antithrombotic therapy to prevent stroke in patients who have nonvalvular atrial fibrillation. Ann Intern Med 2007;146: 857–867.

Ahrens

Lip

Peter

. New oral anticoagulant drugs in cardiovascular disease. Thromb Haemost 2010; 104: 49–60.

January

Wann

Alpert

et al . 2014 AHA/ACC/HRS guideline for the management of patients with atrial fibrillation: A report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines and the Heart Rhythm Society. J Am Coll Cardiol 2014; 64: e1–e76.

Ageno

Gallus

Wittkowsky

et al . Oral anticoagulant therapy: Antithrombotic Therapy and Prevention of Thrombosis, 9th edition: American College of Chest Physicians Evidence-Based Clinical Practice Guidelines. Chest 2012; 141: e44S–e88S.

Connolly

Ezekowitz

Yusuf

et al . Dabigatran versus warfarin in patients with atrial fibrillation. N Engl J Med 2009; 361: 1139–1151. Erratum in: N Engl J Med 2010; 363: 1877.

Patel

Mahaffey

Garg

et al . Rivaroxaban versus warfarin in nonvalvular atrial fibrillation. N Engl J Med 2011; 365: 883–891.

Xarelto [package insert]. Titusville, NJ: Janssen Pharmaceuticals; 2011.

10.

Uchino

Hernandez

. Dabigatran association with higher risk of acute coronary events: Meta-analysis of noninferiority randomized controlled trials. Arch Intern Med 2012; 172: 397–402.

11.

Jonathan

Fanny

et al . Dabigatran etexilate and risk of myocardial infarction, other cardiovascular events, major bleeding, and all-cause mortality: A systematic review and meta-analysis of randomized controlled trials. J Am Heart Assoc 2014; 3: e000515.

12.

Schulman

Kearon

Subcommittee on Control of Anticoagulation of the Scientific and Standardization Committee of the International Society on Thrombosis and Haemostasis. Definition of major bleeding in clinical investigations of antihemostatic medicinal products in non-surgical patients. J Thromb Haemost 2005; 3: 692–694

13.

Mueck

Stampfuss

Kubitza

et al . Clinical pharmacokinetic and pharmacodynamic profile of rivaroxaban. Clin Pharmacokinet 2014; 53: 1–16.

14.

Heidbuchel

Verhamme

Alings

et al . Updated European Heart Rhythm Association Practical Guide on the use of non-vitamin K antagonist anticoagulants in patients with non-valvular atrial fibrillation. Europace 2015; 17: 1467–1507.

15.

Kubitza

Becka

Mueck

et al . Effects of renal impairment on the pharmacokinetics, pharmacodynamics and safety of rivaroxaban, an oral, direct Factor Xa inhibitor. Br J Clin Pharmacol 2010; 70: 703–712.

16.

Pradaxa [package insert]. Ridgefield, CT: Boehringer Ingelheim Pharmaceuticals; 2012.

17.

Ezekowitz

Connolly

Parekh

et al . Rationale and design of RE-LY: Randomized evaluation of long-term anticoagulant therapy, warfarin, compared with dabigatran. Am Heart J 2009; 157: 805–810.

18.

Capodanno

Angiolillo

. Management of antiplatelet and anticoagulant therapy in patients with atrial fibrillation in the setting of acute coronary syndromes or percutaneous coronary interventions. Circ Cardiovasc Interv 2014; 7: 113–124.

19.

Faxon

Eikelboom

Berger

et al . Consensus document: Antithrombotic therapy in patients with atrial fibrillation undergoing coronary stenting. A North-American perspective. J Thromb Haemost 2011; 106: 572–584.

20.

Seeger

Bykov

Bartels

et al . Safety and effectiveness of dabigatran and warfarin in routine care of patients with atrial fibrillation. Thromb Haemost 2015; 114: 1277–1289.

21.

Romanelli

Nolting

Dolginsky

et al . Dabigatran versus warfarin for atrial fibrillation in real-world clinical practice: A systematic review and meta-analysis. Circ Cardiovasc Qual Outcomes 2016; 9: 126–134.

22.

Korenstra

Wijtvliet

EPJ

Veeger

et al . Effectiveness and safety of dabigatran versus acenocoumarol in “real-world” patients with atrial fibrillation. Europace 2016; 18: 1319–1327.

23.

Beyer-Westendorf

Förster

Pannach

et al . Rates, management, and outcome of rivaroxaban bleeding in daily care: Results from the Dresden NOAC registry. Blood 2014; 124(6): 955–962.

24.

Larsen

Skjøth

Nielsen

et al . Comparative effectiveness and safety of non-vitamin K antagonist oral anticoagulants and warfarin in patients with atrial fibrillation: Propensity weighted nationwide cohort study. BMJ 2016; 353: i3189.

25.

Camm

Amarenco

Haas

et al . XANTUS: a real-world, prospective, observational study of patients treated with rivaroxaban for stroke prevention in atrial fibrillation. Eur Heart J 2016; 37: 1145–1153.

26.

Yao

Abraham

Sangaralingham

et al . Effectiveness and safety of dabigatran, rivaroxaban, and apixaban versus warfarin in nonvalvular atrial fibrillation. J Am Heart Assoc 2016; 5: e003725.

27.

Cheung

et al . Continuation of dabigatran therapy in “real-world” practice in Hong Kong. PLoS ONE 2014; 9: e101245.

28.

Chang

H-Y

Zhou

Tang

et al . Risk of gastrointestinal bleeding associated with oral anticoagulants: Population-based retrospective cohort study. BMJ 2015; 350: h1585.

29.

Ansell

. Warfarin versus new agents: Interpreting the data. Hematology Am Soc Hematol Educ Program 2010; 2010: 221–228.

30.

Coleman

Tangirala

Evers

. Treatment persistence and discontinuation with rivaroxaban, dabigatran, and warfarin for stroke prevention in patients with non-valvular atrial fibrillation in the United States. PLoS ONE 2016; 11: e0157769.

31.

Weitz

Gross

. New oral anticoagulants: Which one should my patient use? Hematology Am Soc Hematol Educ Program 2012; 2012: 536–540.