Oral Anticoagulation in Asian Patients With Atrial

Fibrillation and a History of Intracranial Hemorrhage

So-Ryoung Lee, MD; Eue-Keun Choi, MD; Soonil Kwon, MD; Jin-Hyung Jung, BSc;
Kyung-Do Han, PhD; Myung-Jin Cha, MD; Seil Oh, MD, PhD; Gregory Y.H. Lip, MD

Background and Purpose—Warfarin is associated with a better net clinical benefit compared with no treatment in patients
with nonvalvular atrial fibrillation (AF) and history of intracranial hemorrhage (ICH). There are limited data on nonvitamin
K antagonist oral anticoagulants (NOACs) in these patients, especially in the Asian population. We aimed to compare the
effectiveness and safety of NOACs to warfarin in a large-scale nationwide Asian population with AF and a history of ICH.
Methods—Using the Korean Health Insurance Review and Assessment database from January 2010 to April 2018, we
identified patients with oral anticoagulant naïve nonvalvular AF with a prior spontaneous ICH. For the comparisons,
warfarin and NOAC groups were balanced using propensity score weighting. Ischemic stroke, ICH, composite outcome
(ischemic stroke+ICH), fatal ischemic stroke, fatal ICH, death from composite outcome, and all-cause death were
evaluated as clinical outcomes.
Results—Among 5712 patients with AF with prior ICH, 2434 were treated with warfarin and 3278 were treated with NOAC.
Baseline characteristics were well-balanced after propensity score weighting (mean age 72.5 years and CHA2DS2-VASc
score 4.0). Compared with warfarin, NOAC was associated with lower risks of ischemic stroke (hazard ratio [HR], 0.77
[95% CI, 0.61–0.97]), ICH (HR, 0.66 [95% CI, 0.47–0.92]), and composite outcome (HR, 0.73 [95% CI, 0.60–0.88]).
NOAC was associated with lower risks of fatal stroke (HR, 0.54 [95% CI, 0.32–0.89]), death from composite outcome
(HR, 0.53 [95% CI, 0.34–0.81]), and all-cause death (HR, 0.83 [95% CI, 0.69–0.99]) than warfarin. NOAC showed
nonsignificant trends toward to reduce fatal ICH compared with warfarin (HR, 0.47 [95% CI, 0.20–1.03]).
Conclusions—NOAC was associated with a significant lower risk of ICH and ischemic stroke compared with warfarin.
NOAC might be a more effective and safer treatment option for Asian patients with nonvalvular AF and a prior history of
ICH.   (Stroke. 2020;51:416-423. DOI: 10.1161/STROKEAHA.119.028030.)
Downloaded from http://ahajournals.org by on September 9, 2020

Key Words: atrial fibrillation ◼ intracranial hemorrhages ◼ non-vitamin K antagonist oral anticoagulant
◼ oral anticoagulation ◼ stroke ◼ warfarin

O ral anticoagulation (OAC) therapy is essential for the

management of patients with atrial fibrillation (AF) to
prevent stroke and reduce all-cause mortality.1,2 Despite a
Although NOAC has become the first choice for general
patients with nonvalvular AF, the efficacy and safety of NOAC
compared with warfarin in patients with AF with a history of ICH
clear net clinical benefit of OAC, fears of bleeding compli- are unknown. The pivotal clinical trials of all 4 NOACs excluded
cations from OAC therapy remain, especially in the Asian patients with a prior history of ICH. Previous observational stud-
population, because the risk of warfarin-related intracra- ies have reported the clinical benefit of OAC treatment to lower
nial hemorrhage (ICH) is higher in the Asian population the risk for ischemic stroke in these populations7–10; however,
only warfarin was analyzed as an anticoagulation treatment
compared with non-Asians.3,4 With the availability of the
option in these studies. Ongoing randomized trials are awaited,
nonvitamin K antagonist oral anticoagulants (NOACs), the
but limited Asian data would be available from these trials.10–12
landscape has changed, with these agents showing com-
Therefore, there is no evidence whether NOAC shows bet-
parable efficacy and superior safety while reducing ICH ter performance with reducing the risk of recurrent ICH than
risk by 50%, compared with warfarin.5,6 Indeed, NOACs warfarin in patients with a history of ICH, especially in the
seem to have a greater relative risk reduction of ICH in the Asian population at high risk for ICH. In this study, we aimed
Asian population than in non-Asians.4 Therefore, current to compare the effectiveness and safety of NOAC to warfarin
guidelines recommend NOAC in preference to warfarin for in a large-scale nationwide Asian population with nonvalvular
stroke prevention in patients with nonvalvular AF.1,2 AF and a history of ICH.
Received August 12, 2019; final revision received October 13, 2019; accepted October 22, 2019.
From the Department of Internal Medicine, Seoul National University Hospital, Republic of Korea (S.-R.L., E.-K.C., S.K., M.-J.C., S.O., G.Y.H.L.);
Department of Medical Statistics, College of Medicine, Catholic University of Korea, Republic of Korea (J.-H.J., K.-D.H.); Liverpool Centre for
Cardiovascular Science, Liverpool Chest and Heart Hospital, University of Liverpool, United Kingdom (G.Y.H.L.); and Department of Clinical Medicine,
Aalborg University, Denmark (G.Y.H.L.).
The online-only Data Supplement is available with this article at https://www.ahajournals.org/doi/suppl/10.1161/STROKEAHA.119.028030.
Correspondence to Eue-Keun Choi, MD, PhD, Department of Internal Medicine, Seoul National University Hospital, 101 Daehak-ro, Jongno-gu, Seoul,
03080, Republic of Korea. Email choiek17@snu.ac.kr
© 2019 American Heart Association, Inc.
Stroke is available at https://www.ahajournals.org/journal/str DOI: 10.1161/STROKEAHA.119.028030

416
 Lee et al   NOAC in Asian Patients With AF and Previous ICH   417
Methods
Data Source
We conducted a retrospective observational cohort study using data
from the Korean Health Insurance Review and Assessment (HIRA)
database including patients with nonvalvular AF with a history of
spontaneous ICH with subsequent OAC treatment initiation. Detailed
information about data sources described in Methods in the online-
only Data Supplement. All data have been publicly available at
Korean HIRA Bigdata Hub (https://opendata.hira.or.kr). This study
was approved by the Institutional Review Board of Seoul National
University Hospital (E-1812-042-993). Informed consent was
exempted by the review board because the identification number of
each subject in the HIRA database is de-identified and encrypted to
protect the subject’s privacy.
Study Population
For the analysis, we first identified patients with AF who initiated
OAC treatment during the study period (from January 2010 to April
2018). Among these patients, we retrospectively identified a prior
claim of ICH (I60-62) as a primary diagnosis before index OAC
treatment. We included these patients with a history of sponta-
neous ICH, not anticoagulated at the time point of ICH. Detailed
study patient enrollment flow is presented in the online-only Data
Supplement and Figure 1. Finally, 5712 patients with nonvalvular
AF with a history of spontaneous ICH with subsequent OAC treat-
ment initiation were included.
Covariates
Patients’ demographic data, comorbidities, and concomitant medica-
tions were ascertained from the Korean HIRA database. Table I in the
online-only Data Supplement summarized the definitions of codes of
comorbidities. Detailed covariates were described in Methods in the
Downloaded from http://ahajournals.org by on September 9, 2020
online-only Data Supplement.
Outcomes and Follow-Up
To evaluate the effectiveness and safety of NOAC and warfarin, the
occurrence of ischemic stroke, ICH, and composite outcome (is-
chemic stroke+ICH) during the follow-up period were defined as
primary outcome events.13 The secondary outcome events were the
occurrence of fatal ischemic stroke, fatal ICH, death from composite
outcome, and all-cause death. Any event that led to death within the
index hospitalization period of its occurrence was considered a fatal
event. Detailed definitions of study outcomes are described in Table
I in the online-only Data Supplement. The index date was the first
date of NOAC or warfarin use. To assess clinical outcomes, patients
were censored at the occurrence of outcome events, discontinuation
of index treatment, or the end of the study period (April 30, 2018),
whichever came first. Discontinuation was defined as a 30-day gap of
index treatment from the last prescription.13
Statistical Analysis
For the comparisons, the inverse probability of treatment weighting
(IPTW) method was used to balance covariates between 2 treatment
groups (NOAC versus warfarin) regarding time-to-event analyses
using stabilized weights calculated from the propensity scores.14,15
Between 2 treatment groups, the propensity scores of being in each
treatment group were calculated using an ordinary logistic regression
based on all baseline covariates.14 We evaluated the balance between
2 groups using the absolute standardized difference (ASD) of all
baseline covariates.16 When ASDs were ≤0.1 (10%) in all covariates,
2 groups were well-balanced. After IPTW, incidence rates were cal-
culated using weighted event numbers during the follow-up period
divided by 100 person-years at risk in each clinical outcome. The risk
of each clinical outcome in the NOAC versus warfarin (reference)
was compared using survival analysis with the Kaplan-Meier method
(log-rank test) and weighted Cox proportional hazard regression
models with IPTW. For the comparisons among 4 NOACs and war-
farin, we also used IPTW of propensity scores method with a max-
imum ASDs ≤0.1 (10%) in all covariates among different treatment
groups.16 For all statistical analyses, statistical significance was de-
fined as Pvalue of <0.05. Statistical analyses were performing using
SAS 9.3 (SAS Institute, Inc, Cary, NC).
Subgroup and Sensitivity Analyses
We conducted subgroup analyses. Detailed about predefined
subgroups are summarized in Methods in the online-only Data
Supplement. For subgroup analyses, we used multivariable Cox
proportional hazards regression models using all variables in-
cluded propensity score calculation. The statistical significance
of the interaction between treatment and subgroup was defined as
P-interaction of <0.05.
To balance the timing of initiation of OAC after index sponta-
neous ICH, we conducted a sensitivity analysis for adjusting the time
from previous ICH to OAC initiation using the proportion of patients
with previous ICH within 1-year before OAC initiation. Also, to bal-
ance the follow-up period between NOAC and warfarin groups, a sen-
sitivity analysis was conducted with the restriction of the follow-up
period to 1 year.13,17,18
Results
Baseline Characteristics
Among a total of 5712 patients with nonvalvular AF and a
history of ICH with subsequent OAC treatment, 2434 patients
were treated with warfarin, and 3278 patients were treated
Figure 1. Flow chart of the study population.
AF indicates atrial fibrillation; ESRD, end-stage
renal disease; ICH, intracranial hemorrhage;
NOAC, nonvitamin K antagonist oral anticoagu-
lant; and OAC, oral anticoagulant.
 418  Stroke  February 2020
with NOACs (Figure 1). The Table presents the baseline char-
acteristics of the study population before and after IPTW. After
IPTW, all covariates were well balanced between 2 treatment
groups (ASDs for all covariates ≤0.1; Table; Figure I in the
online-only Data Supplement). Mean age was 72.5 years, and
mean CHA2DS2-VASc score was 4.0. Among patients treated
with NOAC, 65% of these patients received a reduced dose
of NOACs.
Primary Outcomes in NOAC and Warfarin Groups
During a median follow-up of 0.6-year (interquartile range,
0.2–1.7 year), ischemic stroke, ICH, and composite outcomes
occurred in 317 (5.5%), 149 (2.6%), and 451 (7.9%) patients,
respectively. Crude and weighted incidence rates of each treat-
ment group are presented in Table II in the online-only Data
Supplement. After IPTW, hazard ratios (HRs) for clinical out-
comes in warfarin and NOAC groups are presented in Figure 2.
Compared with warfarin, NOAC use was associated with
a 23% lower risk of ischemic stroke, a 34% lower risk of ICH,
and a 27% lower risk of composite outcome than warfarin. In
weighted Kaplan-Meier curves (Figure 3), NOAC use associ-
ated with significantly better event-free survival from ischemic
stroke, ICH, and composite outcome than warfarin use.
Secondary Outcomes in NOAC and Warfarin
Groups
NOAC use was also associated with lower risks of fatal is-
chemic stroke, death from the composite outcome, and all-
cause death. Although NOAC group had a lower weighted
Downloaded from http://ahajournals.org by on September 9, 2020
incidence rate of fatal ICH than warfarin group (0.3 versus 0.5
per 100 person-years), wider CI was observed due to the small
number of events (HR, 0.47 [95% CI, 0.20–1.03]; P=0.06;
Figure 2). Figure 3 shows weighted Kaplan-Meier curves for
fatal ischemic stroke, fatal ICH, death from composite out-
come, and all-cause death in NOAC and warfarin groups.
Subgroup Analyses
NOAC Types
Baseline characteristics of patients treated with 4 differ-
ent NOACs and warfarin were presented in Table III in the
online-only Data Supplement. After IPTW, all covariates
were well balanced between 5 treatment groups (maximum
ASDs for all covariates ≤0.1; Table IV in the online-only
Data Supplement). In each NOAC group, 63.3% of rivaroxa-
ban, 74.1% of dabigatran, 57.6% of apixaban, and 71.0% of
edoxaban groups received reduced dose NOAC. Crude and
weighted incidence rates of clinical outcomes in warfarin and
4 NOAC groups are presented in Table V in the online-only
Data Supplement. Figure 4 shows HRs for clinical outcomes
among warfarin and 4 NOAC groups. Compared with war-
farin, all 4 NOACs showed consistent results with the main
analysis. Figure II in the online-only Data Supplement shows
weighted Kaplan-Meier curves for all clinical outcomes in 4
NOACs and warfarin groups.
NOAC Dose Regimens
Among NOAC users, 65% of patients were prescribed
reduced dose NOACs. Baseline characteristics, according
to NOAC dose regimens, are presented in Table VI in the
online-only Data Supplement. Reduced dose NOAC users
were older, less likely to be men, less likely to have a history
of prior stroke, but had higher CHA2DS2-VASc score than
regular dose NOAC users and warfarin users. The crude in-
cidence rates of clinical outcomes are summarized in Table
VII in the online-only Data Supplement. The results for all
clinical outcomes were consistent in both regular and reduced
doses of NOACs (Figure 5).
Various Subgroups
The crude incidence rates of all clinical outcomes according
to treatment by NOAC or warfarin in various subgroups are
presented in Table VIII in the online-only Data Supplement.
NOAC use showed a trend for risk reduction across all sub-
groups compared with warfarin (Figures III and IV in the
online-only Data Supplement). Interaction with treatment
was significant for several outcomes in sex (ICH), Charlson
Comorbidity Index (ischemic stroke, composite outcome),
presence of heart failure (fatal ICH), presence of renal di-
sease (all-cause death), and CHA2DS2-VASc score (all-cause
death), but NOAC use was associated with a lower risk in
these outcomes consistently. Especially in subgroups strat-
ified by various cutoff of CHA2DS2-VASc scores, NOACs
generally showed clinical benefits across these subgroups
compared with warfarin, and no significant statistical inter-
action was found.
Sensitivity Analysis
Adjusting the Time From Previous ICH to OAC Initiation
The average time from previous ICH to index OAC start was
3.1±2.8 years. The NOAC group showed a longer time in-
terval from previous ICH to index OAC start than warfarin
group (3.6±3.0 versus 2.1±2.1 years; P<0.001). The propor-
tion of patients with previous ICH within 1-year before OAC
prescription was lower in the warfarin group than NOAC
group (15.2% versus 20.8%; P<0.001). After additional ad-
justment for the proportion of patients with previous ICH
within 1-year before OAC initiation, the HRs for clinical
outcomes were consistent with main results (Figure V in the
online-only Data Supplement).
Restricting Follow-Up Period to 1-Year
For a sensitivity analysis restricting the follow-up period to
1-year, HR trends for all clinical outcomes were similar to the
main analysis (Figure VI in the online-only Data Supplement).
Discussion
To the best of our knowledge, this was the first nationwide
comparison of the effectiveness and safety of NOACs and
warfarin in a large nationwide AF cohort focused on the Asian
population with a prior history of ICH. The main findings of
this study were as follows: (1) NOAC use was associated with
lower risks of ischemic stroke, ICH, and composite outcome
than warfarin in patients with a prior history of ICH; (2) NOAC
use was also associated with lower risks of fatal ischemic
stroke, death from composite outcome, and all-cause death,
and marginally associated with a lower risk of fatal ICH com-
pared with warfarin; and (3) a consistent trend was observed
 Lee et al   NOAC in Asian Patients With AF and Previous ICH   419

Table.  Baseline Characteristics of Patients Using Warfarin Versus NOAC

Pre IPTW Post IPTW

Total Warfarin NOAC Warfarin NOAC
(n=5712) (n=2434) (n=3278) ASD (n=2438) (n=3266) ASD
Age, y
 Mean±SD 72.4±10.0 70.8±10.4 73.7±9.6 0.291 72.5±10.2 72.5±9.9 <0.001
 Median (IQR) 73 (67–79) 72 (65–78) 74 (68–80) 73 (66–79) 74 (67–79)
 <65 19.2 24.3 15.5 19.5 18.9
 65–74 35.4 36.2 34.8 35.0 35.8
 ≥75 45.3 39.4 49.7 45.5 45.3
Men 56.9 58.5 55.8 0.053 57.2 57.2 <0.001
CHA2DS2-VASc score
 Mean±SD 4.0±1.6 3.9±1.7 4.0±1.5 0.059 4.0±1.6 4.0±1.6 0.001
 Median (IQR) 4 (3–5) 4 (3–5) 4 (3–5) 4 (3–5) 4 (3–5)
 0–1 5.4 7.0 4.3 5.8 4.8
 2–3 32.6 32.5 32.5 32.0 32.8
 ≥4 62.0 60.5 63.2 62.2 62.4
HAS-BLED score
 Mean±SD 4.4±1.2 4.5±1.2 4.4±1.2 0.096 4.4±1.1 4.4±1.2 0.006
 Median (IQR) 4 (4–5) 5 (4–5) 4 (4–5) 4 (4–5) 4 (4–5)
 ≥3 95.0 95.1 95.0 94.9 95.0
Charlson comorbidity index
 Mean±SD 5.1±2.6 5.3±2.6 5.0±2.6 0.141 5.2±2.5 5.2±2.6 0.006
Downloaded from http://ahajournals.org by on September 9, 2020

 Median (IQR) 5 (3–7) 5 (3–7) 5 (3–7) 5 (3–7) 5 (3–7)

 ≥3 84.4 86.1 83.1 85.4 84.8
Comorbidities
 Hypertension 90.1 90.6 89.8 0.316 90.1 90.2 0.002
 Diabetes mellitus 32.2 33.6 31.2 0.050 32.7 32.6 <0.001
 Dyslipidemia 57.4 57.0 57.8 0.014 58.0 58.2 0.003
 Heart failure 41.8 40.6 42.7 0.042 42.2 41.8 0.007
 Prior MI 5.4 5.8 5.0 0.035 5.6 5.4 0.007
 PAD 21.9 21.2 22.3 0.027 22.1 22.2 0.001
 Renal disease 16.8 18.3 15.8 0.066 17.3 16.9 0.009
 COPD 15.0 16.0 14.3 0.045 15.1 14.9 0.004
 Cancer 7.4 6.7 7.9 0.047 7.4 7.3 0.004
 Prior stroke 20.3 23.6 17.8 0.142 20.4 20.6 0.005
Concomitant medication
 Aspirin only 18.4 24.9 13.5 27.2 27.1
 Clopidogrel only 6.8 6.7 6.9 0.387* 15.7 15.5 0.005*
 Dual antiplatelet 9.0 13.1 6.0 8.9 8.8
 NSAID 38.1 40.2 36.5 0.076 37.2 37.5 0.006
 Proton pump inhibitor 30.9 28.4 32.7 0.093 31.2 31.1 0.001
NOAC dose
 Regular dose – – 34.3 – 35.0
 Reduced dose – – 65.7 – 65.0
ASD indicates absolute standardized difference; COPD, chronic obstructive pulmonary disease; IPTW, inverse probability of treatment weighting; IQR, interquartile
range; MI, myocardial infarction; NOAC, nonvitamin K antagonist oral anticoagulants; and NSAID, nonsteroidal anti-inflammatory drug.
*This ASD was for the use of any concomitant antiplatelet agent.
 420  Stroke  February 2020
in different types of NOACs, different dose regimens (regular
or reduced dose NOACs), and various subgroups.
In patients with AF who are needed to receive OAC
therapy for stroke prevention, a prior history of sponta-
neous ICH could lead physicians to become more hesitant
to start OACs. Especially in the Asian population, where
the incidence of warfarin-related ICH was much higher
than in the non-Asian population, prior ICH is one of the
associated factors with OAC underuse in the Asian popula-
tion.19 Although it was obvious that the risk of subsequent
thromboembolic events, including ischemic stroke might
increase without OAC therapy after ICH,20,21 the effective-
ness and safety of OAC therapy in these patients remain
uncertain.
Downloaded from http://ahajournals.org by on September 9, 2020
The net benefit of OAC therapy in patients with a prior
history of ICH has been reported from population-based ret-
rospective observational cohort studies, predominantly from
non-Asian cohorts; in general, OAC therapy in patients with
a prior ICH was associated with a lower risk of thromboem-
bolic events and all-cause mortality.8,21,22 Consistent with these
results, a recent meta-analysis concluded that OAC therapy
Figure 3. Weighted cumulative incidence curves of clinical outcomes in nonvitamin K antagonist oral anticoagulant (NOAC) vs warfarin groups.
Figure 2. Hazard ratios (HRs) of clinical out-
comes in nonvitamin K antagonist oral anti-
coagulant (NOAC) vs warfarin groups. ICH
indicates intracranial hemorrhage.
after ICH was associated with a lower risk of thromboembolic
complications and a similar risk of ICH recurrence.23
However, when focused on bleeding complications, pre-
vious studies have reported controversial results. While OAC
therapy showed comparable results in the risk of major bleed-
ing or recurrent ICH compared with non-OAC therapy,8,22
Nielsen et al21 reported that resumption of warfarin therapy
after spontaneous hemorrhagic stroke in patients with AF
was associated with a higher rate of recurrent ICH. In Asian
patients with AF and a prior history of ICH, warfarin therapy
significantly reduces the risk of ischemic stroke (HR, 0.66
[95% CI, 0.55–0.79]), but increases the risk of recurrent ICH
(HR, 1.60 [95% CI, 1.38–1.86]) compared with a no anti-
thrombotic therapy group.7 Furthermore, OAC therapy might
be beneficial only in patients who had a higher risk of stroke
(CHA2DS2-VASc score ≥6).7
NOACs have been introduced and widely prescribed in
patients with AF. According to the results of pivotal NOAC
clinical trials, NOAC reduced the risk of ICH by ≈50% com-
pared with warfarin, and this benefit seems to be greater in
the Asian population than in the non-Asian population.4,5
 Lee et al   NOAC in Asian Patients With AF and Previous ICH   421

Figure 4. Hazard ratios of clinical outcomes in 4 nonvitamin K antagonist oral anticoagulants (NOACs) vs warfarin groups.

NOACs might be expected to show better performance, es-
pecially with regard to the safety aspect, but patients with a
prior history of ICH were excluded in these pivotal clinical
trials. Also, NOACs were not included or used only in a mi-
nority of patients in previous population-based observational
Downloaded from http://ahajournals.org by on September 9, 2020
One recent nationwide observational study to evaluate
comparative effectiveness and safety of NOAC versus war-
farin in patients with AF and prior ICH.24 Among a total of
622 patients with AF with ICH history, 274 patients were
prescribed warfarin and 348 patients were prescribed NOAC:

studies for patients with a prior ICH.7,8,21–23 In summary, OAC
therapy is clearly needed to prevent thromboembolic events
and reduce all-cause mortality in patients with AF with a
prior ICH, but we need a safer treatment option instead of
warfarin to provide more confident usage of OAC therapy
in these population. While data from clinical trials in the
ICH population are needed, ongoing large randomized tri-
als are largely being conducted in non-Asian populations,11,12
and pending further definitive data in Asians, observational
cohorts provide some evidence although these are no substi-
tute for clinical trials.
NOAC use was associated with a nonsignificant lower risk of
ischemic stroke (HR, 0.52 [95% CI, 0.27–1.00]) and recur-
rent ICH (HR, 0.72 [95% CI, 0.38–1.38]) compared with war-
farin use. Because of the small numbers, the main results were
not statistically significant, and subgroup analyses by NOAC
types or dose regimens were not available.
Our study has included the largest number of patients with
AF with a prior history of ICH, focused on the Asian popula-
tion. We found that NOAC use was significantly associated
with lower risks of ischemic stroke and recurrent ICH with
warfarin, consistent with a previous study.24 NOAC group

Figure 5. Hazard ratios of clinical outcomes for reduced and regular dose nonvitamin K antagonist oral anticoagulants (NOACs) compared with warfarin.
 422  Stroke  February 2020
had a lower weighted incidence rate of fatal ICH than war-
farin group (0.3 versus 0.5 per 100 person-years) and showed
a trend for lower risk of fatal ICH (HR, 0.472 [95% CI,
0.201–1.034]; P=0.068). However, the number of events was
small, so the CIs were wide and statistically nonsignificant.
The lower incidence rate in the NOAC group could be related
to smaller number of ICH events in NOAC-associated with
ICH compared with warfarin-associated with ICH.25 Overall,
NOAC reduced the risk of composite outcome by 27%, death
from composite outcome by 47%, and all-cause death by 17%
compared with warfarin. Reducing fatal events of NOAC use
was generally consistent with the results in the general pop-
ulation.26 Also, these results were consistently observed in
the comparisons between 4 NOACs and warfarin. Although
statistical significances were not achieved due to smaller
numbers of each group and some clinical events, the general
trends were consistent for all the NOAC types. For the NOAC
dose regimen, both regular and reduced dose NOACs were
associated with lower risks of composite outcome compared
with warfarin.
In the warfarin era, CHA2DS2-VASc threshold for positive
net clinical benefit in patients with AF with a prior ICH was 6
in the Asian population.7 In our subgroup analyses stratified
various CHA2DS2-VASc score cutoff (<3 versus ≥3, <4 versus
≥4, <5 versus ≥5, and <6 versus ≥6), there was no interaction
between treatment (NOAC versus warfarin) and subgroups
stratified by CHA2DS2-VASc score; the benefit of NOAC com-
pared with warfarin was consistently observed irrespective of
CHA2DS2-VASc subgroups. Further study will be needed to
Downloaded from http://ahajournals.org by on September 9, 2020
demonstrate whether NOAC use can lower the threshold for
OAC treatment in Asian patients with AF with a prior ICH.
Study Limitations
First, the quality of anticoagulation control in the warfarin
group was not evaluated. The Korean HIRA database did
not include laboratory findings such as the international nor-
malized ratio of prothrombin time. Therefore, the quality of
warfarin treatment could not be evaluated. Also, actual drug
adherence for both warfarin and NOAC groups could not
be evaluated, which is an inherent limitation of claims data.
Second, we included patients who were naïve to OAC treat-
ment and had a history of spontaneous ICH, thus, patients
with anticoagulation treatment–related ICH and traumatic
ICH were not included in this study. Applying our study
results in these patients should be cautiously considered, and
we need further study to find the answer for these patients
in Asians. Also, the diagnosis of ICH is not homogeneous,
and some higher-risk subgroups on imaging data have been
proposed (eg, the presence and number of microbleeds or its
locations)27–29; our data are not able to address this aspect,
as imaging data were not available. The location of ICH is
another limitation that was not addressed in this study. The
results of ICH location on imaging studies are not available in
the claims database. Therefore, the extent and location of ICH
could not be evaluated in this study. Third, although the 2 treat-
ment groups seem to be balanced well by IPTW using given
17 covariates, there might be residual confounding factors,
such as measurable data not available in the claims database
(eg, body weight, laboratory findings, echocardiography, or
imaging studies), and unmeasurable confounding factors. To
minimize the difference between the 2 groups, we excluded
patients who had absolute contraindications for NOAC (val-
vular AF or ESRD). Also, we selected only patients who were
considered and finally received OAC therapy to minimize
the bias related to treatment choice. There might be a possi-
bility that the initial antithrombotic therapy would be decided
based on the patient’s probability of treatment selection, con-
ditional on observed baseline characteristics, thus it would
be more reasonable to compare within the patients who were
selected to use OAC. Indeed, previous studies had compared
no antithrombotic therapy, antiplatelet therapy, and warfarin
in patients with AF with prior ICH.7,8 However, both studies
did not show the benefit of NOAC compared with warfarin
in this high-risk population. Instead of repeating the same
comparisons as prior studies, we focused on the comparison
between warfarin and NOACs. Nonetheless, future random-
ized controlled studies are needed to guide the use of NOAC
in this high-risk population. Last, in the Korean HIRA data-
base, body weight and creatinine clearance were not included.
Therefore, we did not evaluate the appropriateness of NOAC
dosing for each patient.
Conclusions
In patients with a prior history of ICH, the use of NOACs was
associated with significantly lower risks of ischemic stroke,
ICH, and the composite outcome compared with warfarin.
Also, NOAC use was associated with significantly lower risks
of fatal ischemic stroke and death from the composite out-
come. NOAC might be a more effective and safer treatment
option for Asian patients with nonvalvular AF and a prior his-
tory of ICH, which needs to be confirmed with randomized
controlled trials.
Sources of Funding
This study was supported by grant no. 04-2019-3110 from the Seoul
National University Hospital Research Fund and was sponsored by
Daiichi Sankyo Co, Ltd (Tokyo, Japan).
Disclosures
SRL, SK, JHJ, KDH, MJC, SO: None
Dr Choi received research grant from Daiichi-Sankyo, BMS/
Pfizer, Biosense Webster, and Skylabs. Dr Lip was a consultant
for Bayer/Janssen, BMS/Pfizer, Medtronic, Boehringer Ingelheim,
Novartis, Verseon, and Daiichi-Sankyo. He was a speaker for Bayer,
BMS/Pfizer, Medtronic, Boehringer Ingelheim, and Daiichi-Sankyo.
He received no fees personally. The other authors report no conflicts.
References
1. January CT, Wann LS, Calkins H, Chen LY, Cigarroa JE, Cleveland JC Jr,
et al. 2019 AHA/ACC/HRS Focused Update of the 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 Clinical Practice Guidelines and the Heart Rhythm Society. J
Am Coll Cardiol. 2019;74:104–132. doi: 10.1016/j.jacc.2019.01.011
2. Kirchhof P, Benussi S, Kotecha D, Ahlsson A, Atar D, Casadei B, et al;
ESC Scientific Document Group. 2016 ESC Guidelines for the manage-
ment of atrial fibrillation developed in collaboration with EACTS. Eur
Heart J. 2016;37:2893–2962. doi: 10.1093/eurheartj/ehw210
3. Shen AY, Yao JF, Brar SS, Jorgensen MB, Chen W. Racial/ethnic
differences in the risk of intracranial hemorrhage among patients
 Lee et al   NOAC in Asian Patients With AF and Previous ICH   423
with atrial fibrillation. J Am Coll Cardiol. 2007;50:309–315. doi:
10.1016/j.jacc.2007.01.098
4. Wang KL, Lip GY, Lin SJ, Chiang CE. Non-Vitamin K antagonist oral
anticoagulants for stroke prevention in asian patients with nonvalvu-
lar atrial fibrillation: meta-analysis. Stroke. 2015;46:2555–2561. doi:
10.1161/STROKEAHA.115.009947
5. Ruff CT, Giugliano RP, Braunwald E, Hoffman EB, Deenadayalu N,
Ezekowitz MD, et al. Comparison of the efficacy and safety of new
oral anticoagulants with warfarin in patients with atrial fibrillation: a
meta-analysis of randomised trials. Lancet. 2014;383:955–962. doi:
10.1016/S0140-6736(13)62343-0
6. Lip G, Freedman B, De Caterina R, Potpara TS. Stroke prevention in
atrial fibrillation: past, present and future. comparing the guidelines and
practical decision-making. Thromb Haemost. 2017;117:1230–1239. doi:
10.1160/TH16-11-0876
7. Chao TF, Liu CJ, Liao JN, Wang KL, Lin YJ, Chang SL, et al. Use
of oral anticoagulants for stroke prevention in patients with atrial fi-
brillation who have a history of intracranial hemorrhage. Circulation.
2016;133:1540–1547. doi: 10.1161/CIRCULATIONAHA.115.019794
8. Nielsen PB, Larsen TB, Skjøth F, Gorst-Rasmussen A, Rasmussen LH,
Lip GY. Restarting anticoagulant treatment after intracranial hemorrhage
in patients with atrial fibrillation and the impact on recurrent stroke,
mortality, and bleeding: A Nationwide Cohort Study. Circulation.
2015;132:517–525. doi: 10.1161/CIRCULATIONAHA.115.015735
9. Pennlert J, Asplund K, Carlberg B, Wiklund PG, Wisten A, Åsberg S,
et al. Antithrombotic treatment following intracerebral hemorrhage in
patients with and without atrial fibrillation. Stroke. 2015;46:2094–2099.
doi: 10.1161/STROKEAHA.115.009087
10. Korompoki E, Filippidis FT, Nielsen PB, Del Giudice A, Lip GYH,
Kuramatsu JB, et al. Long-term antithrombotic treatment in intracranial
hemorrhage survivors with atrial fibrillation. Neurology. 2017;89:687–
696. doi: 10.1212/WNL.0000000000004235
11. van Nieuwenhuizen KM, van der Worp HB, Algra A, Kappelle LJ,
Rinkel GJ, van Gelder IC, et al; APACHE-AF Investigators. Apixaban
versus Antiplatelet drugs or no antithrombotic drugs after anticoagu-
lation-associated intraCerebral HaEmorrhage in patients with Atrial
Fibrillation (APACHE-AF): study protocol for a randomised controlled
Downloaded from http://ahajournals.org by on September 9, 2020
trial. Trials. 2015;16:393. doi: 10.1186/s13063-015-0898-4
12. Shoamanesh A. NASPAF-ICH: NOACs for Stroke Prevention in Patients
With Atrial Fibrillation and Previous ICH. Available at: https://clinical-
trials.gov/ct2/show/NCT02998905. Accessed July 28, 2019.
13. Lee SR, Choi EK, Kwon S, Han KD, Jung JH, Cha MJ, et al.
Effectiveness and safety of contemporary oral anticoagulants among
Asians with nonvalvular atrial fibrillation. Stroke. 2019;50:2245–2249.
doi: 10.1161/STROKEAHA.119.025536
14. Austin PC. The use of propensity score methods with survival or
time-to-event outcomes: reporting measures of effect similar to those
used in randomized experiments. Stat Med. 2014;33:1242–1258. doi:
10.1002/sim.5984
15. Austin PC, Stuart EA. Moving towards best practice when using inverse
probability of treatment weighting (IPTW) using the propensity score
to estimate causal treatment effects in observational studies. Stat Med.
2015;34:3661–3679. doi: 10.1002/sim.6607
16. Austin PC. Balance diagnostics for comparing the distribution of base-
line covariates between treatment groups in propensity-score matched
samples. Stat Med. 2009;28:3083–3107. doi: 10.1002/sim.3697
17. Lee SR, Choi EK, Han KD, Jung JH, Oh S, Lip GYH. Edoxaban in
asian patients with atrial fibrillation: effectiveness and safety. J Am Coll
Cardiol. 2018;72:838–853. doi: 10.1016/j.jacc.2018.05.066
18. Lip GYH, Keshishian A, Li X, Hamilton M, Masseria C,
Gupta K, et al. Effectiveness and safety of oral anticoagulants among
nonvalvular atrial fibrillation patients. Stroke. 2018;49:2933–2944. doi:
10.1161/STROKEAHA.118.020232
19. Lee SR, Choi EK, Han KD, Cha MJ, Oh S, Lip GYH. Temporal trends
of antithrombotic therapy for stroke prevention in Korean patients with
non-valvular atrial fibrillation in the era of non-vitamin K antagonist
oral anticoagulants: a nationwide population-based study. PLoS One.
2017;12:e0189495. doi: 10.1371/journal.pone.0189495
20. Brønnum Nielsen P, Larsen TB, Gorst-Rasmussen A, Skjøth F,
Rasmussen LH, Lip GYH. Intracranial hemorrhage and subsequent is-
chemic stroke in patients with atrial fibrillation: a nationwide cohort study.
Chest. 2015;147:1651–1658. doi: 10.1378/chest.14-2099
21. Nielsen PB, Larsen TB, Skjøth F, Lip GY. Outcomes associated with
resuming warfarin treatment after hemorrhagic stroke or traumatic intra-
cranial hemorrhage in patients with atrial fibrillation. JAMA Intern Med.
2017;177:563–570. doi: 10.1001/jamainternmed.2016.9369
22. Ottosen TP, Grijota M, Hansen ML, Brandes A, Damgaard D,
Husted SE, et al. Use of antithrombotic therapy and long-term clinical
outcome among patients surviving intracerebral hemorrhage. Stroke.
2016;47:1837–1843. doi: 10.1161/STROKEAHA.116.012945
23. Murthy SB, Gupta A, Merkler AE, Navi BB, Mandava P, Iadecola C,
et al. Restarting anticoagulant therapy after intracranial hemorrhage: a
systematic review and meta-analysis. Stroke. 2017;48:1594–1600. doi:
10.1161/STROKEAHA.116.016327
24. Nielsen PB, Skjøth F, Søgaard M, Kjældgaard JN, Lip GYH, Larsen TB.
Non-Vitamin K antagonist oral anticoagulants versus warfarin in atrial
fibrillation patients with intracerebral hemorrhage. Stroke. 2019;50:939–
946. doi: 10.1161/STROKEAHA.118.023797
25. Wilson D, Charidimou A, Shakeshaft C, Ambler G, White M, Cohen H,
et al; CROMIS-2 Collaborators. Volume and functional outcome of intra-
cerebral hemorrhage according to oral anticoagulant type. Neurology.
2016;86:360–366. doi: 10.1212/WNL.0000000000002310
26. Inohara T, Xian Y, Liang L, Matsouaka RA, Saver JL, Smith EE, et
al. Association of intracerebral hemorrhage among patients tak-
ing non-vitamin k antagonist vs Vitamin K antagonist oral antico-
agulants with in-hospital mortality. JAMA. 2018;319:463–473. doi:
10.1001/jama.2017.21917
27. Shoamanesh A, Charidimou A, Sharma M, Hart RG. Should patients
with ischemic stroke or transient ischemic attack with atrial fibrillation
and microbleeds be anticoagulated? Stroke. 2017;48:3408–3412. doi:
10.1161/STROKEAHA.117.018467
28. Wilson D, Ambler G, Shakeshaft C, Brown MM, Charidimou A,
Al-Shahi Salman R, et al; CROMIS-2 Collaborators. Cerebral micro-
bleeds and intracranial haemorrhage risk in patients anticoagulated for
atrial fibrillation after acute ischaemic stroke or transient ischaemic
attack (CROMIS-2): a multicentre observational cohort study. Lancet
Neurol. 2018;17:539–547. doi: 10.1016/S1474-4422(18)30145-5
29. Martí-Fàbregas J, Medrano-Martorell S, Merino E, Prats-Sánchez L,
Marín R, Delgado-Mederos R, et al; HERO Study Investigators. MRI
predicts intracranial hemorrhage in patients who receive long-term oral
anticoagulation. Neurology. 2019;92:e2432–e2443. doi: 10.1212/WNL.
0000000000007532