1

I
n the Rivaroxaban Once Daily Oral Direct Factor Xa Inhibitor
Compared With Vitamin K Antagonism for Prevention of
Stroke and Embolism Trial in Atrial Fibrillation (ROCKET AF)
trial, the efcacy and safety of rivaroxaban, a novel, oral, direct
factor Xa inhibitor, was assessed for the prevention of stroke
in patients with atrial brillation (AF) and at moderate to high
risk of stroke.
1
Results from this study showed that rivaroxaban
(20 mg once daily or 15 mg once daily in patients with a cre-
atinine clearance [CrCl] of 3049 mL/min) was noninferior to
dose-adjusted warfarin for the prevention of stroke and sys-
temic embolism in an intention-to-treat (ITT) analysis. Rates
of major or nonmajor clinically relevant bleeding were not sig-
nicantly different between the 2 study groups; however, intra-
cranial hemorrhage and fatal bleeding occurred signicantly
Background and PurposeIn Rivaroxaban Once Daily Oral Direct Factor Xa Inhibitor Compared With Vitamin K
Antagonism for Prevention of Stroke and Embolism Trial in Atrial Fibrillation (ROCKET AF) trial, rivaroxaban was
noninferior to dose-adjusted warfarin in preventing stroke or systemic embolism among patients with nonvalvular atrial
brillation at moderate to high stroke risk. Because of differences in patient demographics, epidemiology, and stroke
risk management in East Asia, outcomes and relative effects of rivaroxaban versus warfarin were assessed to determine
consistency among East Asians versus other ROCKET AF participants.
MethodsBaseline demographics and interaction of treatment effects of rivaroxaban and warfarin among patients within
East Asia and outside were assessed.
ResultsA total of 932 (6.5%) ROCKET AF participants resided in East Asia. At baseline, East Asians had lower weight,
creatinine clearance, and prior vitamin K antagonist use; higher prevalence of prior stroke; and less congestive heart
failure and prior myocardial infarction than other participants. Despite higher absolute event rates for efcacy and safety
outcomes in East Asians, the relative efcacy of rivaroxaban (20 mg once daily; 15 mg once daily for creatinine clearance
of 3049 mL/min) versus warfarin with respect to the primary efcacy end point (stroke/systemic embolism) was
consistent among East Asians and nonEast Asians (interaction P=0.666). Relative event rates for the major or nonmajor
clinically relevant bleeding in patients treated with rivaroxaban and warfarin were consistent among East Asians and
nonEast Asians (interaction P=0.867).
ConclusionsObserved relative efcacy and safety of rivaroxaban versus warfarin were similar among patients within and
outside East Asia. Rivaroxaban, 20 mg once daily, is an alternative to warfarin for stroke prevention in East Asians with
nonvalvular atrial brillation.
Clinical Trial RegistrationURL: http://www.clinicaltrials.gov. Unique identier: NCT00123456.
(Stroke. 2014;45:00-00.)
Key Words: atrial brillation

Far East

rivaroxaban

stroke
Rivaroxaban for Stroke Prevention in East Asian Patients
From the ROCKET AF Trial
Ka Sing Lawrence Wong, MD; Dai Yi Hu, MD; Abraham Oomman, MD; Ru-San Tan, MD;
Manesh R. Patel, MD; Daniel E. Singer, MD; Gnter Breithardt, MD;
Kenneth W. Mahaffey, MD; Richard C. Becker, MD; Robert Califf, MD; Keith A.A. Fox, MD;
Scott D. Berkowitz, MD; Werner Hacke, MD; Graeme J. Hankey, MD;
on behalf of The Executive Steering Committee and the ROCKET AF Study Investigators
Received July 25, 2013; nal revision received February 8, 2014; accepted February 27, 2014.
From the Department of Medicine and Therapeutics, Chinese University of Hong Kong, Prince of Wales Hospital, Hong Kong, China (K.S.L.W.); Heart
Center, Peking University Peoples Hospital, Beijing, China (D.Y.H.); Department of Cardiology, Apollo Hospital, Chennai, India (A.O.); National Heart
Centre, Singapore, Singapore (R.-S.T.); Department of Medicine, Division of Cardiology, Duke Clinical Research Institute (M.R.P., K.W.M., R.C.B.) and
Department of Medicine, Division of Cardiology, Duke Translational Medicine Institute (R.C.), Duke University Medical Center, Durham, NC; Department of
Epidemiology, Massachusetts General Hospital and Harvard Medical School, Boston (D.E.S.); Department of Cardiovascular Medicine, University Hospital
Mnster, Mnster, Germany (G.B.); Centre for Cardiovascular Science, University of Edinburgh and Royal Inrmary of Edinburgh, Edinburgh, Scotland
(K.A.A.F.); Bayer HealthCare Pharmaceuticals, Montville, NJ (S.D.B.); Department of Neurology, Ruprecht-Karls-University, Heidelberg, Germany (W.H.);
and Stroke Unit, Department of Neurology, Royal Perth Hospital, University of Western Australia, Perth, Australia (G.J.H.).
Guest Editor for this article was Kazunori Toyoda, MD.
The online-only Data Supplement is available with this article at http://stroke.ahajournals.org/lookup/suppl/doi:10.1161/STROKEAHA.
113.002968/-/DC1.
Correspondence to Ka Sing Lawrence Wong, MD, Department of Medicine and Therapeutics, Chinese University of Hong Kong, Prince of Wales
Hospital, Hong Kong, China. E-mail ks-wong@cuhk.edu.hk
2014 American Heart Association, Inc.
Stroke is available at http://stroke.ahajournals.org DOI: 10.1161/STROKEAHA.113.002968
Original Contribution
2 Stroke June 2014
less frequently in the rivaroxaban treatment group. There was
no signicant interaction in terms of the efcacy and safety of
rivaroxaban versus warfarin with respect to race or region.
1
Demographic differences exist between Asian populations
and other ethnic groups, which may affect the optimal dosing
of novel anticoagulants, including rivaroxaban. For example,
body weight and body mass index (BMI) values have been
shown to be lower in Chinese versus white populations.
2
The
proportion of strokes that are hemorrhagic has also been shown
to be higher in Asians than in whites.
3,4
Conversely, there is evi-
dence to suggest that cardioembolic strokes make up a greater
proportion of strokes in white versus Asian populations.
5
Data
from 2 studies have also shown East Asian populations to be
more sensitive to warfarin than Indian and white populations.
6,7

Warfarin may also interact with several food and herbal supple-
ments commonly used in Asia, thereby complicating its use.
8
There is a perception, and some supporting evidence,
that the incidence of intracranial hemorrhage is increased in
patients of Asian ethnicity (largely East Asian) who receive
warfarin compared with other ethnic groups.
9,10
As a result,
anticoagulants are underprescribed
11,12
or underdosed
13
in
patients within this region because of the fear that the risk of
bleeding offsets any benets.
As a result of the differences in patient demographics,
epidemiology, and stroke risk management in East Asia, out-
comes and the relative effects of rivaroxaban versus warfarin
were assessed as to whether they were consistent among East
Asians and nonEast Asians who participated in ROCKET
AF, particularly because warfarin was targeted to an interna-
tional normalized ratio (INR) range of 2.0 to 3.0.
Methods
Overview of the ROCKET AF Study
The ROCKET AF was a randomized, multicenter, double-blind,
double-dummy, event-driven study that was performed at 1178
participating sites in 45 countries.
1
The trial was supported by
Janssen Research and Development, LLC, and Bayer HealthCare
Pharmaceuticals. The Duke Clinical Research Institute coordinated
the trial, managed the database, and performed all primary analyses
independently of the sponsors; the trial itself was conducted by a
global consortium.
14
The protocol was approved by pertinent national
regulatory authorities and ethics committees at participating centers.
The objective of ROCKET AF was to compare once-daily rivaroxa-
ban with dose-adjusted warfarin for the prevention of stroke and systemic
embolism in patients with nonvalvular AF and at moderate to high risk
of stroke. Patients enrolled in ROCKET AF were randomly assigned to
xed-dose rivaroxaban (20 mg once daily or 15 mg once daily in patients
with a CrCl of 3049 mL/min) or adjusted-dose warfarin (target INR, 2.0
3.0). To maintain blinding, patients within each group were also given a
matching placebo. The primary efcacy end point was the composite of
stroke (ischemic or hemorrhagic) and noncentral nervous system (CNS)
systemic embolism. The principal safety outcome was the composite of
major and nonmajor clinically relevant bleeding. Major bleeding was de-
ned as clinically overt bleeding associated with any of the following:
fatal bleeding, critical organ bleeding, hemoglobin drop 2 g/dL, or trans-
fusions of 2 U of whole blood or packed red blood cells.
Analysis of the East Asian Cohort of ROCKET AF
A secondary analysis of results from randomized patients from 73 sites
in 4 geographic areas in East Asia (China, Korea, Taiwan, and Hong
Kong) who participated in the full ROCKET AF study was performed.
Efcacy and safety results for patients of different ethnicities within
a given geographic region are not available (eg, patients of Asian or
white ethnicity resident in East Asia). Therefore, for the purposes of
this secondary analysis, it was assumed that residence in East Asia is
a surrogate for ethnicity. Patients enrolled in ROCKET AF in the East
Asian region were randomized to the same study regimen as patients
in other regions. The same efcacy end points and safety outcomes as
dened in ROCKET AF (composite of stroke or systemic embolism,
major and nonmajor clinically relevant bleeding, etc) in patients from
the East Asian region were compared with patients from the remaining
ROCKET AF population (ie, excluding those at the 73 sites in the 4
geographic areas). These 4 areas were considered to be most represen-
tative of the East Asian region as a whole, although it should be noted
that no patients from Japan were enrolled in ROCKET AF. A separate
trial of rivaroxaban versus warfarin in patients with nonvalvular AF,
known as J-ROCKET AF, was conducted in Japan, in which a lower
15 mg once-daily dose was given based on pharmacokinetic modeling
and in line with specic Japanese clinical practice patterns and guide-
line recommendations for reduced target INR levels.
15,16
Patient Baseline Demographics
Baseline demographics were assessed for patients in the East Asian
region and remaining ROCKET AF population. Demographics analyzed
included age, weight, BMI, and CrCl; prior vitamin K antagonist (VKA)
and acetylsalicylic acid (ASA) use; baseline congestive heart failure,
diabetes mellitus, and hypertension; and prior stroke, transient ischemic
attack (TIA), non-CNS systemic embolism, and myocardial infarction.
Statistical Analyses
The ITT population included all patients who were randomized in
the study, whereas the safety population included all patients who
received 1 dose of a study drug. On treatment was dened as the
period from the date of the rst double-blind study medication to the
date of the last double-blind study medication administration plus 2
days. Site notication was dened as notication to the sites when
the required number of primary efcacy end point events had been
reached. In the current subanalysis, estimates and 2-sided 95% con-
dence intervals (CIs) for the hazard ratio (HR; rivaroxaban versus
warfarin) for patients residing within and outside East Asia are shown
in the ITT population until site notication for all efcacy end points
and in the safety population while on treatment for all safety out-
comes. HRs for patients residing within versus outside East Asia, ir-
respective of treatment assignment, have also been estimated for the
primary efcacy end point and for the principal safety outcome.
Cox proportional hazards regression models were used with treat-
ment, residence status (within or outside the East Asian region), and
the treatment by residence status interaction as covariates to test for
interaction between the differential effects of rivaroxaban and warfarin
on stroke or non-CNS systemic embolism (primary efcacy end point)
and major or nonmajor clinically relevant bleeding (principal safety
outcome) among patients residing within and outside East Asia. The
interaction of treatment group by residence status for predened sec-
ondary efcacy end points and safety outcomes was also assessed. For
selected signicant interactions, HRs for patients residing within ver-
sus outside East Asia were also assessed within each treatment group.
It should be noted that any signicant P values should be considered
nominal only because of the large number of comparisons that have
been performed on the database. There is also limited power to detect
signicant differences because of the small East Asian sample size.
The linear interpolation method of Rosendaal et al
17
was used to
calculate the overall time that INR values fell within the therapeutic
range of 2.0 to 3.0 for the warfarin group. A comparative analysis of
treatment efcacy according to quartiles of time that INR values fell
within the therapeutic range at participating study centers in the East
Asian region was also performed. For both East Asian and nonEast
Asian patient groups, center time in therapeutic range (cTTR) was
calculated using the total number of INR values in target range from
all subjects on warfarin within a center divided by the total number
of INR values from all subjects on warfarin within the same center.
Centers were categorized into quartiles with approximately equal
Wong et al Rivaroxaban for Stroke Prevention 3
numbers of subjects in each. Analyses of cTTR were calculated for
the safety population while on treatment.
Results
Patient Recruitment
A total of 932 patients (China [n=496], Korea [n=204], Taiwan
[n=159], Hong Kong [n=73]) formed the ITT population in the
ROCKET AF East Asian cohort. This equates to 6.5% of the
14 264 patients randomized in the full ROCKET AF study. In
total, 928 ITT patients from the ROCKET AF East Asian cohort
received 1 dose of study drug and therefore formed the safety
population. The patient disposition for the ROCKET AF East
Asian cohort is summarized in Figure I in the online-only Data
Supplement. The median duration of treatment exposure in the
East Asian subanalysis was 553 days, whereas the median fol-
low-up period was 668 days. This compares with 593 and 708
days, respectively, in the remaining study population.
Patient Baseline Demographics
Baseline demographics by treatment arm for patients in the East
Asian cohort of ROCKET AF and the ROCKET AF popula-
tion excluding East Asia are shown in Table 1. Baseline patient
Table 1. Baseline Characteristics of the Intention-to-Treat Population in the ROCKET AF East Asian Cohort and the Remaining
Study Population
East Asian Cohort (N=932) ROCKET AF Excluding East Asia (N=13 322)
Rivaroxaban (n=468) Warfarin (n=464) Rivaroxaban (n=6663) Warfarin (n=6669)
Mean age, y 69.6 69.7 71.3 71.3
Mean baseline weight, kg 67.3 66.4 83.1 82.7
Median BMI, kg/m
2
25.0 24.8 28.6 28.4
Baseline mean creatinine clearance, mL/min 64.0 66.1 73.6 73.0
Prior vitamin K antagonist use, % 48.5 49.6 63.3 63.4
Prior chronic ASA use, % 37.0 39.4 36.2 36.5
Baseline congestive heart failure, % 38.9 38.4 64.3 63.9
Baseline diabetes mellitus, % 38.3 35.6 40.5 39.8
Baseline hypertension, % 79.3 80.6 91.0 91.5
Prior stroke/TIA/non-CNS systemic embolism, % 65.0 65.1 54.2 53.9
Prior myocardial infarction 3.6 7.3 17.5 18.8
Mean baseline CHADS
2
score 3.2 3.2 3.5 3.5
CHA
2
DS
2
-VASc, %
1 0.0 0.0 0.0 <0.1
2 5.6 6.5 2.6 2.5
3 20.1 20.3 11.1 11.9
4 30.1 27.6 25.5 25.1
5 26.5 26.5 30.2 29.8
6 12.4 12.7 19.4 19.1
7 4.3 5.6 8.5 8.4
8 0.9 0.9 2.5 2.8
9 0.2 0.0 0.2 0.4
CHA
2
DS
2
-VASc, mean 4.4 4.4 4.9 4.9
HAS-BLED, %
0 0.9 1.1 0.6 0.6
1 6.2 6.0 7.0 7.0
2 29.9 26.9 30.5 29.6
3 38.7 37.7 41.5 42.2
4 18.8 23.1 18.0 18.0
5 5.1 4.7 2.4 2.6
6 0.4 0.4 0.1 0.1
HAS-BLED, mean 2.9 2.9 2.8 2.8
Time within 1.5<2.0, % 29.9 20.0
Time within 2.03.0, % 47.1 55.7
ASA indicates acetylsalicylic acid; BMI, body mass index; CHADS
2
, congestive heart failure, hypertension, age, diabetes, prior stroke/TIA; CHA
2
DS
2
-VASc, CHADS
2
,
vascular disease, age, sex category; CNS, central nervous system; HAS-BLED, hypertension, abnormal renal/liver function, stroke, bleeding history or predisposition,
labile international normalized ratio, elderly, drugs/alcohol; ROCKET AF, Rivaroxaban Once Daily Oral Direct Factor Xa Inhibitor Compared With Vitamin K Antagonism for
Prevention of Stroke and Embolism Trial in Atrial Fibrillation; and TIA, transient ischemic attack.
4 Stroke June 2014
characteristics were similar between the 2 treatment arms in
the East Asian cohort. Compared with patients residing outside
the East Asian region, patients in the East Asian cohort were
younger and had a lower body weight and BMI, and fewer
patients had congestive heart failure, hypertension, or previous
myocardial infarction at baseline. Mean baseline CrCl was also
lower for patients in the ROCKET AF East Asian cohort versus
patients outside this region. Conversely, levels of prior stroke/
TIA/non-CNS systemic embolism, as well as the percentage
of VKA-naive patients, were higher in the ROCKET AF East
Asian cohort than in the remaining study population.
Mean baseline congestive heart failure, hypertension, age,
diabetes, prior stroke/TIA (CHADS
2
) scores were lower for
patients in the East Asian cohort (3.2 in both the rivaroxa-
ban and warfarin treatment arms) than for patients outside
East Asia (3.5 in both treatment arms). A greater number of
East Asian patients in ROCKET AF had a baseline CHADS
2

score of 2 (rivaroxaban, 24.2%; warfarin, 23.9%) compared
with the nonEast Asian population (rivaroxaban, 12.2%;
warfarin, 12.3%). This is consistent with the observation of
higher prevalence of prior stroke or TIA observed among East
Asian patients, which would confer a CHADS
2
score of 2 in
the absence of other risk factors. Conversely, fewer East Asian
patients in ROCKET AF had a score 4 (rivaroxaban, 35.7%;
warfarin, 31.9%) compared with the rest of the study popula-
tion (rivaroxaban, 44.7%; warfarin, 43.3%).
Efcacy End Points
In the East Asian cohort, the observed annual event rates for
the primary efcacy end point of stroke or systemic embo-
lism were numerically higher (2.6% and 3.4% per year in
the rivaroxaban and warfarin arms, respectively [HR=0.78;
95% CI, 0.441.39]) compared with corresponding annual
rates in the remaining ROCKET AF population (2.1% and
2.4% per year [HR=0.89; 95% CI, 0.751.05]). However,
no signicant interaction between treatment and residen-
tial status was identied (P=0.666; Figure 1 and Table 2).
The efcacy of rivaroxaban versus warfarin for the primary
Figure 1. Intention-to-treat analyses of
efcacy end point events. This is a post
hoc analysis that has not been corrected
for multiple analyses. CNS indicates
central nervous system; MRS, modied
Rankin scale score; and ROCKET AF,
Rivaroxaban Once Daily Oral Direct Fac-
tor Xa Inhibitor Compared With Vitamin K
Antagonism for Prevention of Stroke and
Embolism Trial in Atrial Fibrillation. *Haz-
ard ratio (95% condence interval [CI])
from the Cox proportional hazard model
with treatment as a covariate. P value
for the interaction of treatment group and
residence location (East Asia vs nonEast
Asia) for the indicated outcome. Statisti-
cally signicant at nominal 0.05 (2-sided).
Wong et al Rivaroxaban for Stroke Prevention 5
efcacy end point was therefore consistent among patients
residing within and outside East Asia. The observed efcacy
of rivaroxaban versus warfarin for the primary efcacy end
point was also consistent across all prespecied subgroups in
the ROCKET AF East Asian cohort, including patients with
mild or moderate renal impairment (Figure II in the online-
only Data Supplement).
Observed rates of the 2 prespecied composite second-
ary efcacy end points (stroke, non-CNS systemic embo-
lism, and vascular death; and stroke, non-CNS systemic
embolism, myocardial infarction, and vascular death) were
numerically lower in the rivaroxaban treatment group than
in the warfarin treatment group of the East Asian cohort
(Figure 1). The observed rate of all-cause stroke was also
lower in the rivaroxaban treatment group. Overall, 4 East
Asian patients (0.5% per year) in the rivaroxaban arm had
a hemorrhagic stroke compared with 10 patients (1.2% per
year) in the warfarin arm (HR=0.40; 95% CI, 0.131.27).
Observed rates of all-cause mortality were also nonsignif-
icantly lower with rivaroxaban (2.6% per year) than with
warfarin (3.6% per year [HR=0.73; 95% CI, 0.411.27]),
although rates of myocardial infarction were the same in
both treatment arms (1.0% per year in each treatment arm
[HR=1.00; 95% CI, 0.382.66]).
When studying the 2 geographical areas, there was a strong
trend toward increased rates of stroke or systemic embolism
in the East Asian cohort versus the remaining study popu-
lation irrespective of treatment assignment (HR=1.34; 95%
CI, 1.001.80; P=0.053; Table 2). Observed annual rates of
all-cause stroke, hemorrhagic stroke, ischemic stroke, non-
disabling stroke, and disabling stroke were all higher in
the ROCKET AF East Asian cohort than in the remaining
study population (Figure 1). These differences may be more
marked if adjusted for the lower baseline CHADS
2
score in
East Asian patients. However, there was no signicant inter-
action between treatment and residential status, indicating
that the relative efcacy of rivaroxaban versus warfarin for
these stroke outcomes was consistent between the 2 study
populations. There was also no signicant interaction identi-
ed for the other secondary efcacy end points (non-CNS
systemic embolism; myocardial infarction; all-cause, vascu-
lar, and nonvascular death; Figure 1).
INR and cTTR Analyses
Among warfarin-treated patients in the ROCKET AF East
Asian cohort, patient INR values were within the therapeu-
tic range (2.03.0) for a mean of 47.1% of the time versus
55.7% for the ROCKET AF population excluding East Asia.
INR values were between 1.5 and <2.0 for a total mean of
29.9% and 20.0% of the time for warfarin-treated patients
in the ROCKET AF East Asian cohort and the ROCKET AF
population excluding East Asia, respectively (Table 1).
Table 2. Primary Efcacy End Point of Stroke or Systemic Embolism in the Intention-to-Treat Population Until Site Notication*
Analysis Method
Rivaroxaban Warfarin Rivaroxaban vs Warfarin
East Asia vs
NonEast Asia
Hazard Ratio (95% CI) n/N
Event Rate, %
per Year n/N
Event Rate, %
per Year Hazard Ratio (95% CI)
Interaction
P Value
ROCKET AF East Asian cohort 21/468 2.6 27/464 3.4 0.78 (0.441.39) 0.666 1.34 (1.001.80);
P=0.053
ROCKET AF (excluding East Asia) 248/6613 2.1 279/6626 2.4 0.89 (0.751.05)
CI indicates condence interval; and ROCKET AF, Rivaroxaban Once Daily Oral Direct Factor Xa Inhibitor Compared With Vitamin K Antagonism for Prevention of Stroke
and Embolism Trial in Atrial Fibrillation.
*The median follow-up period was 668 and 708 days for the intention-to-treat population in the ROCKET AF East Asian cohort and remaining ROCKET AF study
population, respectively.
Irrespective of treatment.
Number of events per 100 patient-years of follow-up.
Hazard ratios (95% CI) from the Cox proportional hazard model with treatment as a covariate.
P value for the interaction of treatment group and each baseline subgroup based on the Cox proportional hazard model including treatment group, baseline
subgroup, and their interaction. Statistically signicant at nominal 0.05 (2-sided).
Table 3. Primary Efcacy Treatment Effect by Quartiles of Center Time in Therapeutic Range in the ROCKET AF East Asian Cohort*
Rivaroxaban Warfarin
Rivaroxaban vs Warfarin
HR (95% CI) cTTR,% n/N (%) Event Rate, % per Year n/N (%) Event Rate, % per Year
0.0041.94 3/114 (2.6) 2.0 3/115 (2.6) 1.7 1.20 (0.245.95)
42.1947.82 2/109 (1.8) 1.2 5/107 (4.7) 3.4 0.36 (0.071.87)
49.8957.92 4/119 (3.4) 2.4 9/113 (8.0) 5.5 0.43 (0.131.41)
58.2173.13 5/112 (4.5) 2.9 6/127 (4.7) 3.0 0.97 (0.293.17)
Values presented as number of subjects with events/number of subjects in each cTTR quartile (%). Analysis based on the rst event in the safety population during
treatment: n=466 for the rivaroxaban arm and n=462 for the warfarin arm. CI indicates condence interval; cTTR, center time in therapeutic range; HR, hazard ratio; and
ROCKET AF, Rivaroxaban Once Daily Oral Direct Factor Xa Inhibitor Compared With Vitamin K Antagonism for Prevention of Stroke and Embolism Trial in Atrial Fibrillation.
*P value for interaction=0.585.
International normalized ratio, 2.03.0 inclusive.
Number of events per 100 patient-years of follow-up.
HR (95% CI) from the Cox proportional hazard model with treatment as a covariate.
6 Stroke June 2014
With regard to cTTR, no interaction with treatment effect
was apparent when assessing rivaroxaban for the primary ef-
cacy end point in the East Asian cohort (P=0.585 for interac-
tion; Table 3).
Safety Outcomes
In the ROCKET AF East Asian cohort, the annual event rates
for major or nonmajor clinically relevant bleeding (the prin-
cipal safety outcome) were 20.9% and 20.7% per year in the
rivaroxaban and warfarin arms, respectively (HR=1.01; 95%
CI, 0.791.30; Table 4). This compares with 14.5% and 14.1%,
respectively, per year in the ROCKET AF population exclud-
ing East Asia (HR=1.03; 95% CI, 0.961.11). However, no sig-
nicant interaction between treatment and residence status was
identied (P=0.867; Figure 2 and Table 4). With the exception
of patients with prior ASA use (patients who had not received
prior ASA and who were prescribed rivaroxaban had a lower
rate of the principal safety outcome; conversely, patients who
had received prior ASA and who were prescribed warfarin had
a lower rate of the principal safety outcome), there were no
Table 4. Principal Safety Outcome of Major or Nonmajor Clinically Relevant Bleeding in the Safety Population During Treatment
Cohort
Rivaroxaban Warfarin Rivaroxaban vs Warfarin
East Asia vs
NonEast Asia*
Hazard Ratio (95% CI) n/N
Event Rate, %
per Year n/N
Event Rate, %
per Year Hazard Ratio (95% CI) Interaction P Value
ROCKET AF East
Asian cohort
121/466 20.9 122/462 20.7 1.01 (0.791.30) 0.867 1.42 (1.251.62);
P<0.0001
ROCKET AF
(excluding East Asia)
1354/6645 14.5 1327/6663 14.1 1.03 (0.961.11)
CI indicates condence interval; and ROCKET AF, Rivaroxaban Once Daily Oral Direct Factor Xa Inhibitor Compared With Vitamin K Antagonism for Prevention of Stroke
and Embolism Trial in Atrial Fibrillation.
*Irrespective of treatment.
Number of events per 100 patient-years of follow-up. All analyses are based on the time to the rst event.
Hazard ratios (95% CI) from the Cox proportional hazard model with treatment as a covariate.
P value for the interaction of treatment group and each baseline subgroup based on the Cox proportional hazard model including treatment group, baseline subgroup,
and their interaction. Statistically signicant at nominal 0.05 (2-sided).
Figure 2. Safety population analyses of
safety outcome events. This is a post hoc
analysis that has not been corrected for
multiple analyses. ROCKET AF indicates
Rivaroxaban Once Daily Oral Direct Fac-
tor Xa Inhibitor Compared With Vitamin K
Antagonism for Prevention of Stroke and
Embolism Trial in Atrial Fibrillation. *Haz-
ard ratio (95% condence interval [CI])
from the Cox proportional hazard model
with treatment as a covariate. P value
for the interaction of treatment group and
residence location (East Asia vs nonEast
Asia) for the indicated outcome. Statisti-
cally signicant at nominal 0.05 (2-sided).
Irrespective of treatment group, the
difference between East Asians and
nonEast Asians with respect to major
bleeding was not signicant (hazard ratio
[HR]=1.23; 95% CI, 0.941.60; P=0.14).
Bleeding events were considered to
be critical if they occurred in intracra-
nial, intraspinal, intraocular, pericardial,
intra-articular, intramuscular (with com-
partment syndrome), or retroperitoneal
sites. See Table I in the online-only Data
Supplement.
Wong et al Rivaroxaban for Stroke Prevention 7
signicant interactions between treatment group and any pre-
specied subgroup in the ROCKET AF East Asian cohort for
the principal safety outcome, including patients with mild or
moderate renal impairment (Figures III and IV in the online-
only Data Supplement).
Rivaroxaban was also associated with reductions in critical
organ bleeding and intracranial hemorrhage relative to warfa-
rin in the East Asian cohort. Overall, there were 5 (0.7% per
year) and 18 (2.6% per year) critical organ bleeding events
in the rivaroxaban and warfarin treatment arms, respectively
(HR=0.28; 95% CI, 0.100.75). Rates of intracranial hemor-
rhage with rivaroxaban versus warfarin were 4 (0.6% per year)
and 17 (2.5% per year) events, respectively, in each treatment
group (HR=0.24; 95% CI, 0.080.71). Furthermore, there was a
signicant interaction between treatment and residence within
or outside East Asia for intracranial hemorrhage in ROCKET
AF (interaction P=0.044; Figure 2). Among patients receiv-
ing rivaroxaban, there was no signicant difference in the rate
of intracranial hemorrhage for East Asians versus nonEast
Asians (HR=1.20; 95% CI, 0.433.31; P=0.73; Table I in the
online-only Data Supplement). Conversely, among patients
receiving warfarin, the rate of intracranial hemorrhage was
signicantly greater in East Asian patients than in nonEast
Asian patients (HR=3.89; 95% CI, 2.296.63; P<0.0001;
Table I in the online-only Data Supplement).
Among East Asian patients, there were 20 events with hemo-
globin drop 2 g/dL in each of the rivaroxaban (3.0% per year)
and warfarin (2.9% per year) treatment groups. Rivaroxaban
was associated with fewer transfusion events than warfarin, with
8 (1.2% per year) and 15 events (2.2% per year), respectively,
in each group (HR=0.54; 95% CI, 0.231.28). There were also
fewer fatal bleeding events observed with rivaroxaban (0.2%
per year) than with warfarin (1.0% per year [HR=0.14; 95% CI,
0.021.16]).
The risk of major or nonmajor clinically relevant bleeding
was signicantly higher in the East Asian cohort than in the
remaining study population, irrespective of treatment assignment
(HR=1.42; 95% CI, 1.251.62; P<0.0001; Table 4). However,
there was no signicant difference in the risk of major bleeding
between patients residing within versus outside East Asia, irre-
spective of treatment regimen (HR=1.23; 95% CI, 0.941.60;
P=0.14; Figure 2). For most bleeding outcomes, the safety prole
of rivaroxaban versus warfarin was consistent among ROCKET
AF patients resident within and outside East Asia (Figure 2).
Overall, there were 108 (23.2%) and 144 (31.2%) non-
bleeding treatment-emergent serious adverse events in the
rivaroxaban and warfarin treatment groups, respectively, in the
ROCKET AF East Asian cohort (Table II in the online-only
Data Supplement). The most common adverse event was car-
diac failure, occurring in 9 (1.9%) patients in the rivaroxaban
group and 16 (3.5%) patients in the warfarin group. Dyspnea
was reported as a treatment-emergent nonserious adverse
event by 28 patients in each of the rivaroxaban (6.0%) and
warfarin (6.1%) treatment groups.
Discussion
In this regional analysis of ROCKET AF, we assessed the ef-
cacy and safety of rivaroxaban versus warfarin among East
Asian patients and compared results with those from ROCKET
AF patients recruited outside East Asia. Although event rates
were higher among East Asian patients, the relative efcacy
and safety of rivaroxaban versus warfarin was similar to that
of nonEast Asian patients. The signicant interaction between
treatment group and prior ASA use in the ROCKET AF East
Asian cohort for the principal safety outcome could be caused
by the risk prole of the East Asian population. However, the
observed interaction P value has not been corrected for multiple
analyses and could also be a chance nding. Therefore, although
patient numbers were relatively low in the East Asian cohort,
these results support extrapolation of the full trial results to the
East Asian patient population.
Analysis of the baseline demographics conrms much of
what was previously known about the population of East Asia.
The mean body weight and BMI of patients in the East Asian
cohort were lower than those of the other patients in ROCKET
AF; previous studies have also shown body weight and BMI to
be lower in Asian than in white populations.
2,6
This is an impor-
tant nding when analyzing the efcacy and safety results for
the ROCKET AF East Asian cohort and the remaining ROCKET
AF population because it is consistent with previous observa-
tions showing that rivaroxaban can be used without dose adjust-
ment for body weight.
18,19
Patients in the ROCKET AF East
Asian cohort had higher levels of prior stroke/TIA/non-CNS
systemic embolism compared with the remaining study popula-
tion, a nding supported by a recent publication that reported
a higher burden of stroke relative to ischemic heart disease in
East Asia.
20
Conversely, fewer patients in the East Asian cohort
had other risk factors for stroke (hypertension, congestive heart
failure, diabetes mellitus). There was also a higher number of
VKA-naive patients in the ROCKET AF East Asian cohort,
which is in line with other data from the region that suggest that
warfarin is underused.
11
Irrespective of treatment assignment,
patients in the East Asian cohort had signicantly higher lev-
els of major or nonmajor clinically relevant bleeding and there
was a strong trend toward increased rates of stroke or systemic
embolism versus the remaining ROCKET AF study population.
One possible explanation is that patients in the East Asian cohort
had higher levels of baseline prior stroke/TIA/nonsystemic
embolism compared with the remaining study population; it has
been shown that patients with AF and with prior stroke or TIA
have higher absolute rates of stroke or systemic embolism than
patients with AF without prior stroke or TIA.
21
However, the risk
of major bleeding was not signicantly higher in the East Asian
cohort than in the remaining study population, irrespective of
whether they received warfarin or rivaroxaban.
For the primary efcacy end point in the ROCKET AF East
Asian cohort, no interaction between the treatment effect of
rivaroxaban versus warfarin and cTTR quartiles was apparent
(Table 3), suggesting that the efcacy of rivaroxaban versus
warfarin is not affected by the quality of warfarin manage-
ment. However, care should be taken in interpreting these data
because of the low numbers of patients in the East Asian cohort
and the corresponding low number of efcacy events.
In analyzing the results for the secondary efcacy end
points in the ROCKET AF East Asian cohort (Figure 1), the
observed reduction in hemorrhagic stroke with rivaroxaban
8 Stroke June 2014
versus warfarin is an important nding given the concern over
the higher rates of hemorrhagic stroke and intracranial hemor-
rhage in Asian populations. Although statistical signicance
was not attained for the East Asian results, the results are con-
sistent with the nominally signicant results for patients out-
side East Asia and with the results of the overall trial.
1
There are several possible reasons why the uptake of VKAs
is relatively low in East Asia, one in particular being the fear of
intracranial hemorrhage. Rates of intracranial hemorrhage have
been shown to be higher in Asians than in other ethnic groups
receiving warfarin.
9,10
As a result of the concern around bleeding
risk in East Asia, patients in this region are underprescribed
11,12
or
underdosed
13
anticoagulants. Although there is limited evidence
to support a lower INR range for warfarin in East Asian patients,
a study in Chinese patients receiving warfarin therapy suggested
that a target range of 1.8 to 2.4 may be preferable.
22
The Chinese
Current Knowledge and Management Recommendations in AF
2010 guidelines recommend that warfarin should be adminis-
tered on a long-term basis at a target INR of 1.6 to 2.5 in patients
with nonmechanical valvular AF with high and intermediate
risk factors for stroke.
13
However, Taiwanese and Korean stroke
guidelines recommend an INR target of 2.0 to 3.0 when admin-
istering warfarin to patients with AF who are at risk of stroke.
23,24

These latter guidelines are in line with internationally used AF
guidelines such as those produced by the European Society of
Cardiology and the American College of Chest Physicians.
25,26
In
this secondary analysis, there was a signicant increase in intra-
cranial hemorrhage in East Asian patients receiving warfarin ver-
sus nonEast Asian patients. In contrast, there was no signicant
difference in the rate of intracranial hemorrhage in East Asian
versus nonEast Asian patients receiving rivaroxaban, and this
has important safety implications for the use of rivaroxaban in
this region. One possible reason for the increased risk of intra-
cranial hemorrhage in East Asian patients receiving warfarin
could be the greater number of patients in this group who were
VKA naive compared with the remaining study populationthe
risk of major hemorrhage is known to be higher in VKA-naive
patients when they rst start taking VKAs.
27
In addition, warfa-
rin has been shown to interact with some herbal products that
are commonly used in the Asia-Pacic region. The risk of bleed-
ing is increased when warfarin is combined with some of these
products, including ginkgo and garlic.
8
However, it is unknown
whether this was a contributory factor in this secondary analysis.
As well as the increased risk of intracranial hemorrhage in East
Asian patients receiving warfarin, the reduction in intracranial
hemorrhage with rivaroxaban compared with warfarin in this
cohort is also a signicant nding, especially considering the rel-
atively low number of patients (n=932) included in this analysis.
As well as the reduction in intracranial hemorrhage, rates
of critical organ bleeding were also lower with rivaroxaban
in the ROCKET AF East Asian cohort. It should be noted that
there was no apparent increase in major bleeding events for
rivaroxaban versus warfarin because of hemoglobin drops or
transfusions in the ROCKET AF East Asian cohort, in contrast
to the ROCKET AF population outside this region. Observed
rates of fatal bleeding were also lower with rivaroxaban ver-
sus warfarin in the East Asian patient population. These are
important ndings because patients in the Asian region are
prone to have more bleeding events (Table 4).
9,10
Broadly comparable results were observed in a recent analysis
of Asian patients with AF enrolled in the Randomized Evaluation
of Long-term Anticoagulation Therapy (RE-LY) trial of dabiga-
tran versus warfarin,
28
in that Asian patients experienced somewhat
greater rates of stroke or systemic embolism than non-Asians,
irrespective of treatment group. However, in contrast to this study,
rates of intracranial hemorrhage in Asian patients receiving warfa-
rin were only slightly greater than in non-Asian patients and there
was no signicant interaction between ethnic group (Asian versus
non-Asian) and the treatment effect of dabigatran versus warfarin.
In the recently published Effective Anticoagulation With Factor
Xa Next Generation in Atrial Fibrillation (ENGAGE AF) study,
the data from patients in the Asia region were consistent with the
overall trial results, that is, edoxaban was noninferior to warfarin
with respect to the prevention of stroke or systemic embolism and
was associated with signicantly lower rates of bleeding.
29
Study Limitations
When analyzing the results from this secondary analysis, it
should be noted that the number of patients in the East Asian
cohort was relatively low, which limits the power of the vari-
ous outcomes and can make extrapolation of the results dif-
cult. Other limitations of this analysis are that it involved a
subgroup from the main ROCKET AF study and that not all
countries in the East Asian region were represented. Although
it was assumed that residence in East Asia was a surrogate for
ethnicity for the purposes of this analysis, data generated are
for the population of the geographic area as a whole.
Conclusions
Several demographic and epidemiological differences exist for
the ROCKET AF East Asian cohort compared with participants
in ROCKET AF outside this region. Despite these differences,
efcacy and safety results were broadly consistent for patients
in ROCKET AF who resided within and outside the East Asian
region. Furthermore, intracranial hemorrhage and critical organ
bleeding were considerably reduced with rivaroxaban com-
pared with warfarin in the East Asian cohort. Thus, this analysis
supports the use of rivaroxaban at 20 mg once daily (15 mg
once daily for CrCl of 3049 mL/min) for the prevention of
stroke and systemic embolism as an alternative to warfarin in
adult patients with nonvalvular AF in the East Asian region.
Acknowledgments
We thank Michael Barton who provided medical writing services with
funding from Janssen Scientic Affairs, LLC, and Bayer HealthCare
Pharmaceuticals.
Sources of Funding
The ROCKET study was funded by Janssen Research and
Development, LLC, and Bayer HealthCare Pharmaceuticals.
Disclosures
Dr Wong has received honoraria as a member of a steering commit-
tee for Johnson & Johnson and Bayer; and also for participation in
clinical trials, contributions to advisory boards, or oral presentations
from Bayer, Sano-Aventis, Bristol-Myers Squibb, Boehringer
Ingelheim, and Pzer. Dr Oomman has received honoraria from
Bayer HealthCare. Dr Tan has received consultancy, advisory board,
Wong et al Rivaroxaban for Stroke Prevention 9
and lecture fees from Bayer HealthCare and Pzer; and consultancy
and advisory board fees from Boehringer Ingelheim, Bristol-Myers
Squibb, Johnson & Johnson, AstraZeneca, and Servier. Dr Patel has
received research grant from AstraZeneca and consulting fees from
Ortho-McNeil-Janssen and Bayer HealthCare and serves on the ad-
visory board for Genzyme. Dr Singer has served as consultant to
Bayer HealthCare, Boehringer Ingelheim, Bristol-Myers Squibb,
Daiichi Sankyo, Johnson & Johnson, Pzer, and CSL Behring;
has received research support from Johnson & Johnson; and owns
no shares in any company that would pose a conict of interest.
Dr Breithardt has received honoraria from Johnson & Johnson
and Bayer; and advisory board fees from Boehringer Ingelheim,
Bristol-Myers Squibb, Pzer, and Sano-Aventis. Dr Mahaffey
has received grant support (signicant) from AstraZeneca, Bayer,
Boehringer Ingelheim, Bristol-Myers Squibb, Daiichi Sankyo,
Eli Lilly, GlaxoSmithKline, Johnson & Johnson, Merck, Portola,
Regado Biotechnologies, Sano, Schering-Plough (now Merck),
and The Medicines Company; signicant consulting fees from
Bayer, Johnson & Johnson; and modest consulting fees from Bayer,
Boehringer Ingelheim, Bristol-Myers Squibb, Daiichi Sankyo,
Eli Lilly, GlaxoSmithKline, Johnson & Johnson, Merck, Ortho/
McNeill, Pzer, Polymedix, and Sano; all full disclosure can be
found at www.dcri.org. Dr Becker has received modest research
support from Bayer and Johnson & Johnson and serves on the
scientic advisory boards for Bayer. Dr Califf has received con-
sulting fees and other service from Bayer, Bristol-Myers Squibb
Foundation, CV Sight, LLC, DSI-Lilly, Gambro, Heart.org, Janssen
R&D, LLC, Kowa, Novartis; all other industry interactions are list-
ed at www.dcri.org. Dr Fox has received grant support and lecture
fees from Bayer, Janssen, Eli Lilly, AstraZeneca; and lecture fees
from Sano-Aventis. Dr Berkowitz has been employed as a clinical
research physician at Bayer HealthCare Pharmaceuticals; holds no
other consultancies, board memberships, stock, or stock options;
and does not receive grants, royalties, consulting fees or honoraria,
payment for lectures, or payment for manuscript preparation. Dr
Hacke has received honoraria for serving on an executive com-
mittee for Johnson & Johnson and Bayer and advisory board fees
from Boehringer Ingelheim. Dr Hankey has received consultancy
fees from advisory boards for Bayer Pharmaceuticals, Boehringer
Ingelheim, Bristol-Myers Squibb, and Pzer; has received hono-
raria for speaking at scientic symposia sponsored by Bayer
Pharmaceuticals and web-based education programs of heart.org; is
a member of the ROCKET AF Executive Steering Committee; and
has no shareholdings or funding grants. Dr Hu reports no conicts.
References
1. Patel MR, Mahaffey KW, Garg J, Pan G, Singer DE, Hacke W, et al;
ROCKET AF Investigators. Rivaroxaban versus warfarin in nonvalvular
atrial brillation. N Engl J Med. 2011;365:883891.
2. Wang D, Li Y, Lee SG, Wang L, Fan J, Zhang G, et al. Ethnic differences
in body composition and obesity related risk factors: study in Chinese
and white males living in China. PLoS One. 2011;6:e19835.
3. Thrift AG, Dewey HM, Macdonell RA, McNeil JJ, Donnan GA. Incidence
of the major stroke subtypes: initial ndings from the North East Melbourne
stroke incidence study (NEMESIS). Stroke. 2001;32:17321738.
4. Zhang LF, Yang J, Hong Z, Yuan GG, Zhou BF, Zhao LC, et al;
Collaborative Group of China Multicenter Study of Cardiovascular
Epidemiology. Proportion of different subtypes of stroke in China. Stroke.
2003;34:20912096.
5. Ng WK, Goh KJ, George J, Tan CT, Biard A, Donnan GA. A compara-
tive study of stroke subtypes between Asians and Caucasians in two
hospital-based stroke registries. Neurol J Southeast Asia. 1998;3:1926.
6. Zhao F, Loke C, Rankin SC, Guo JY, Lee HS, Wu TS, et al. Novel
CYP2C9 genetic variants in Asian subjects and their inuence on main-
tenance warfarin dose. Clin Pharmacol Ther. 2004;76:210219.
7. Yuen E, Gueorguieva I, Wise S, Soon D, Aarons L. Ethnic differences
in the population pharmacokinetics and pharmacodynamics of warfarin.
J Pharmacokinet Pharmacodyn. 2010;37:324.
8. Fugh-Berman A. Herb-drug interactions. Lancet. 2000;355:134138.
9. Shen AY, Yao JF, Brar SS, Jorgensen MB, Chen W. Racial/ethnic differ-
ences in the risk of intracranial hemorrhage among patients with atrial
brillation. J Am Coll Cardiol. 2007;50:309315.
10. Shen AY, Chen W, Yao JF, Brar SS, Wang X, Go AS. Effect of race/eth-
nicity on the efcacy of warfarin: potential implications for prevention of
stroke in patients with atrial brillation. CNS Drugs. 2008;22:815825.
11. Wen-Hang QI; Society of Cardiology, Chinese Medical Association.
Retrospective investigation of hospitalised patients with atrial brillation
in mainland China. Int J Cardiol. 2005;105:283287.
12. Gao Q, Fu X, Wei JW, Chen X, Huang Y, Wang J, et al; China QUEST
Study Investigators. Use of oral anticoagulation among stroke patients
with atrial brillation in China: the ChinaQUEST (Quality evaluation of
stroke care and treatment) registry study. Int J Stroke. 2013;8:150154.
13. Huang CX, Zhang S, Ma CS, Yang YZ, Huang DJ, Cao KJ, et al. Current
knowledge and management recommendations of atrial brillation. Chin
J Cardiac Arrhythmia. 2010;14:328369.
14. ROCKET AF Study Investigators. Rivaroxaban-once daily, oral, direct
factor Xa inhibition compared with vitamin K antagonism for prevention
of stroke and Embolism Trial in Atrial Fibrillation: rationale and design
of the ROCKET AF study. Am Heart J. 2010;159:340347.e1.
15. Atarashi H, Inoue H, Okumura K, Yamashita T, Kumagai N, Origasa
H; J-RHYTHM Registry Investigators. Present status of anticoagulation
treatment in Japanese patients with atrial brillation: a report from the
J-RHYTHM Registry. Circ J. 2011;75:13281333.
16. Hori M, Matsumoto M, Tanahashi N, Momomura S, Uchiyama S, Goto
S, et al. Rivaroxaban vs. warfarin in Japanese patients with atrial brilla-
tion the J-ROCKET AF study. Circ J. 2012;76:21042111.
17. Rosendaal FR, Cannegieter SC, van der Meer FJ, Brit E. A method to
determine the optimal intensity of oral anticoagulant therapy. Thromb
Haemost. 1993;69:236239.
18. Mueck W, Eriksson BI, Bauer KA, Borris L, Dahl OE, Fisher WD, et al.
Population pharmacokinetics and pharmacodynamics of rivaroxabanan
oral, direct factor Xa inhibitorin patients undergoing major orthopaedic
surgery. Clin Pharmacokinet. 2008;47:203216.
19. Kubitza D, Becka M, Zuehlsdorf M, Mueck W. Body weight has lim-
ited inuence on the safety, tolerability, pharmacokinetics, or pharma-
codynamics of rivaroxaban (BAY 59-7939) in healthy subjects. J Clin
Pharmacol. 2007;47:218226.
20. Kim AS, Johnston SC. Global variation in the relative burden of stroke
and ischemic heart disease. Circulation. 2011;124:314323.
21. Hankey GJ, Patel MR, Stevens SR, Becker RC, Breithardt G, Carolei
A, et al; ROCKET AF Steering Committee Investigators. Rivaroxaban
compared with warfarin in patients with atrial brillation and previous
stroke or transient ischaemic attack: a subgroup analysis of ROCKET
AF. Lancet Neurol. 2012;11:315322.
22. You JH, Chan FW, Wong RS, Cheng G. Is INR between 2.0 and 3.0 the opti-
mal level for Chinese patients on warfarin therapy for moderate-intensity
anticoagulation? Br J Clin Pharmacol. 2005;59:582587.
23. Chang YJ, Ryu SJ, Chen JR, Hu HH, Yip PK, Chiu TF; Consensus Group of
Taiwan Stroke Society. [Guidelines for the general management of patients
with acute ischemic stroke]. Acta Neurol Taiwan. 2008;17:275294.
24. Korean Clinical Research Center for Stroke. Clinical Practice Guidelines
for Stroke. 2010. http://www.stroke-crc.or.kr/cpgs%20for%20stroke%20
(english).pdf. Accessed January 1, 2011.
25. Camm AJ, Kirchhof P, Lip GYH, Schotten U, Savelieva I, Ernst S, et al.
Guidelines for the management of atrial brillation: The Task Force
for the Management of Atrial Fibrillation of the European Society of
Cardiology (ESC). Eur Heart J. 2010;31:23692429.
26. You JJ, Singer DE, Howard PA, Lane DA, Eckman MH, Fang MC, et al;
American College of Chest Physicians. Antithrombotic therapy for atrial
brillation: Antithrombotic Therapy and Prevention of Thrombosis, 9th
ed: American College of Chest Physicians Evidence-Based Clinical
Practice Guidelines. Chest. 2012;141(2 suppl):e531Se575S.
27. Garcia DA, Lopes RD, Hylek EM. New-onset atrial brillation and war-
farin initiation: high risk periods and implications for new antithrom-
botic drugs. Thromb Haemost. 2010;104:10991105.
28. Hori M, Connolly SJ, Zhu J, Liu LS, Lau CP, Pais P, et al; RE-LY
Investigators. Dabigatran versus warfarin: effects on ischemic and hem-
orrhagic strokes and bleeding in Asians and non-Asians with atrial bril-
lation. Stroke. 2013;44:18911896.
29. Giugliano RP, Ruff CT, Braunwald E, Murphy SA, Wiviott SD, Halperin
JL, et al; ENGAGE AF-TIMI 48 Investigators. Edoxaban versus warfarin
in patients with atrial brillation. N Engl J Med. 2013;369:20932104.