Professional Documents
Culture Documents
Oral Anticoagulan in Asian Patients With Atrial Fibrillation and A History of Intracranial Hemorrhage
Oral Anticoagulan in Asian Patients With Atrial Fibrillation and A History of Intracranial Hemorrhage
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
416
Lee et al NOAC in Asian Patients With AF and Previous ICH 417
with NOACs (Figure 1). The Table presents the baseline char- to NOAC dose regimens, are presented in Table VI in the
acteristics of the study population before and after IPTW. After online-only Data Supplement. Reduced dose NOAC users
IPTW, all covariates were well balanced between 2 treatment were older, less likely to be men, less likely to have a history
groups (ASDs for all covariates ≤0.1; Table; Figure I in the of prior stroke, but had higher CHA2DS2-VASc score than
online-only Data Supplement). Mean age was 72.5 years, and regular dose NOAC users and warfarin users. The crude in-
mean CHA2DS2-VASc score was 4.0. Among patients treated cidence rates of clinical outcomes are summarized in Table
with NOAC, 65% of these patients received a reduced dose VII in the online-only Data Supplement. The results for all
of NOACs. clinical outcomes were consistent in both regular and reduced
doses of NOACs (Figure 5).
Primary Outcomes in NOAC and Warfarin Groups Various Subgroups
During a median follow-up of 0.6-year (interquartile range, The crude incidence rates of all clinical outcomes according
0.2–1.7 year), ischemic stroke, ICH, and composite outcomes to treatment by NOAC or warfarin in various subgroups are
occurred in 317 (5.5%), 149 (2.6%), and 451 (7.9%) patients, presented in Table VIII in the online-only Data Supplement.
respectively. Crude and weighted incidence rates of each treat- NOAC use showed a trend for risk reduction across all sub-
ment group are presented in Table II in the online-only Data groups compared with warfarin (Figures III and IV in the
Supplement. After IPTW, hazard ratios (HRs) for clinical out- online-only Data Supplement). Interaction with treatment
comes in warfarin and NOAC groups are presented in Figure 2. was significant for several outcomes in sex (ICH), Charlson
Compared with warfarin, NOAC use was associated with Comorbidity Index (ischemic stroke, composite outcome),
a 23% lower risk of ischemic stroke, a 34% lower risk of ICH, presence of heart failure (fatal ICH), presence of renal di-
and a 27% lower risk of composite outcome than warfarin. In sease (all-cause death), and CHA2DS2-VASc score (all-cause
weighted Kaplan-Meier curves (Figure 3), NOAC use associ- death), but NOAC use was associated with a lower risk in
ated with significantly better event-free survival from ischemic these outcomes consistently. Especially in subgroups strat-
stroke, ICH, and composite outcome than warfarin use. ified by various cutoff of CHA2DS2-VASc scores, NOACs
generally showed clinical benefits across these subgroups
Secondary Outcomes in NOAC and Warfarin compared with warfarin, and no significant statistical inter-
Groups action was found.
NOAC use was also associated with lower risks of fatal is-
chemic stroke, death from the composite outcome, and all- Sensitivity Analysis
cause death. Although NOAC group had a lower weighted
Downloaded from http://ahajournals.org by on September 9, 2020
in different types of NOACs, different dose regimens (regular after ICH was associated with a lower risk of thromboembolic
or reduced dose NOACs), and various subgroups. complications and a similar risk of ICH recurrence.23
In patients with AF who are needed to receive OAC However, when focused on bleeding complications, pre-
therapy for stroke prevention, a prior history of sponta- vious studies have reported controversial results. While OAC
neous ICH could lead physicians to become more hesitant therapy showed comparable results in the risk of major bleed-
to start OACs. Especially in the Asian population, where ing or recurrent ICH compared with non-OAC therapy,8,22
the incidence of warfarin-related ICH was much higher Nielsen et al21 reported that resumption of warfarin therapy
than in the non-Asian population, prior ICH is one of the after spontaneous hemorrhagic stroke in patients with AF
associated factors with OAC underuse in the Asian popula- was associated with a higher rate of recurrent ICH. In Asian
tion.19 Although it was obvious that the risk of subsequent patients with AF and a prior history of ICH, warfarin therapy
thromboembolic events, including ischemic stroke might significantly reduces the risk of ischemic stroke (HR, 0.66
increase without OAC therapy after ICH,20,21 the effective- [95% CI, 0.55–0.79]), but increases the risk of recurrent ICH
ness and safety of OAC therapy in these patients remain (HR, 1.60 [95% CI, 1.38–1.86]) compared with a no anti-
uncertain. thrombotic therapy group.7 Furthermore, OAC therapy might
Downloaded from http://ahajournals.org by on September 9, 2020
The net benefit of OAC therapy in patients with a prior be beneficial only in patients who had a higher risk of stroke
history of ICH has been reported from population-based ret- (CHA2DS2-VASc score ≥6).7
rospective observational cohort studies, predominantly from NOACs have been introduced and widely prescribed in
non-Asian cohorts; in general, OAC therapy in patients with patients with AF. According to the results of pivotal NOAC
a prior ICH was associated with a lower risk of thromboem- clinical trials, NOAC reduced the risk of ICH by ≈50% com-
bolic events and all-cause mortality.8,21,22 Consistent with these pared with warfarin, and this benefit seems to be greater in
results, a recent meta-analysis concluded that OAC therapy the Asian population than in the non-Asian population.4,5
Figure 3. Weighted cumulative incidence curves of clinical outcomes in nonvitamin K antagonist oral anticoagulant (NOAC) vs warfarin groups.
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- One recent nationwide observational study to evaluate
pecially with regard to the safety aspect, but patients with a comparative effectiveness and safety of NOAC versus war-
prior history of ICH were excluded in these pivotal clinical farin in patients with AF and prior ICH.24 Among a total of
trials. Also, NOACs were not included or used only in a mi- 622 patients with AF with ICH history, 274 patients were
nority of patients in previous population-based observational prescribed warfarin and 348 patients were prescribed NOAC:
Downloaded from http://ahajournals.org by on September 9, 2020
studies for patients with a prior ICH.7,8,21–23 In summary, OAC NOAC use was associated with a nonsignificant lower risk of
therapy is clearly needed to prevent thromboembolic events ischemic stroke (HR, 0.52 [95% CI, 0.27–1.00]) and recur-
and reduce all-cause mortality in patients with AF with a rent ICH (HR, 0.72 [95% CI, 0.38–1.38]) compared with war-
prior ICH, but we need a safer treatment option instead of farin use. Because of the small numbers, the main results were
warfarin to provide more confident usage of OAC therapy not statistically significant, and subgroup analyses by NOAC
in these population. While data from clinical trials in the types or dose regimens were not available.
ICH population are needed, ongoing large randomized tri- Our study has included the largest number of patients with
als are largely being conducted in non-Asian populations,11,12 AF with a prior history of ICH, focused on the Asian popula-
and pending further definitive data in Asians, observational tion. We found that NOAC use was significantly associated
cohorts provide some evidence although these are no substi- with lower risks of ischemic stroke and recurrent ICH with
tute for clinical trials. 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- (eg, body weight, laboratory findings, echocardiography, or
farin group (0.3 versus 0.5 per 100 person-years) and showed imaging studies), and unmeasurable confounding factors. To
a trend for lower risk of fatal ICH (HR, 0.472 [95% CI, minimize the difference between the 2 groups, we excluded
0.201–1.034]; P=0.068). However, the number of events was patients who had absolute contraindications for NOAC (val-
small, so the CIs were wide and statistically nonsignificant. vular AF or ESRD). Also, we selected only patients who were
The lower incidence rate in the NOAC group could be related considered and finally received OAC therapy to minimize
to smaller number of ICH events in NOAC-associated with the bias related to treatment choice. There might be a possi-
ICH compared with warfarin-associated with ICH.25 Overall, bility that the initial antithrombotic therapy would be decided
NOAC reduced the risk of composite outcome by 27%, death based on the patient’s probability of treatment selection, con-
from composite outcome by 47%, and all-cause death by 17% ditional on observed baseline characteristics, thus it would
compared with warfarin. Reducing fatal events of NOAC use be more reasonable to compare within the patients who were
was generally consistent with the results in the general pop- selected to use OAC. Indeed, previous studies had compared
ulation.26 Also, these results were consistently observed in no antithrombotic therapy, antiplatelet therapy, and warfarin
the comparisons between 4 NOACs and warfarin. Although in patients with AF with prior ICH.7,8 However, both studies
statistical significances were not achieved due to smaller did not show the benefit of NOAC compared with warfarin
numbers of each group and some clinical events, the general in this high-risk population. Instead of repeating the same
trends were consistent for all the NOAC types. For the NOAC comparisons as prior studies, we focused on the comparison
dose regimen, both regular and reduced dose NOACs were between warfarin and NOACs. Nonetheless, future random-
associated with lower risks of composite outcome compared ized controlled studies are needed to guide the use of NOAC
with warfarin. in this high-risk population. Last, in the Korean HIRA data-
In the warfarin era, CHA2DS2-VASc threshold for positive base, body weight and creatinine clearance were not included.
net clinical benefit in patients with AF with a prior ICH was 6 Therefore, we did not evaluate the appropriateness of NOAC
in the Asian population.7 In our subgroup analyses stratified dosing for each patient.
various CHA2DS2-VASc score cutoff (<3 versus ≥3, <4 versus
≥4, <5 versus ≥5, and <6 versus ≥6), there was no interaction Conclusions
between treatment (NOAC versus warfarin) and subgroups In patients with a prior history of ICH, the use of NOACs was
stratified by CHA2DS2-VASc score; the benefit of NOAC com- associated with significantly lower risks of ischemic stroke,
pared with warfarin was consistently observed irrespective of ICH, and the composite outcome compared with warfarin.
CHA2DS2-VASc subgroups. Further study will be needed to Also, NOAC use was associated with significantly lower risks
Downloaded from http://ahajournals.org by on September 9, 2020
demonstrate whether NOAC use can lower the threshold for of fatal ischemic stroke and death from the composite out-
OAC treatment in Asian patients with AF with a prior ICH. come. NOAC might be a more effective and safer treatment
option for Asian patients with nonvalvular AF and a prior his-
Study Limitations tory of ICH, which needs to be confirmed with randomized
First, the quality of anticoagulation control in the warfarin controlled trials.
group was not evaluated. The Korean HIRA database did
not include laboratory findings such as the international nor- Sources of Funding
malized ratio of prothrombin time. Therefore, the quality of This study was supported by grant no. 04-2019-3110 from the Seoul
National University Hospital Research Fund and was sponsored by
warfarin treatment could not be evaluated. Also, actual drug
Daiichi Sankyo Co, Ltd (Tokyo, Japan).
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-
Disclosures
SRL, SK, JHJ, KDH, MJC, SO: None
ment and had a history of spontaneous ICH, thus, patients Dr Choi received research grant from Daiichi-Sankyo, BMS/
with anticoagulation treatment–related ICH and traumatic Pfizer, Biosense Webster, and Skylabs. Dr Lip was a consultant
ICH were not included in this study. Applying our study for Bayer/Janssen, BMS/Pfizer, Medtronic, Boehringer Ingelheim,
results in these patients should be cautiously considered, and Novartis, Verseon, and Daiichi-Sankyo. He was a speaker for Bayer,
BMS/Pfizer, Medtronic, Boehringer Ingelheim, and Daiichi-Sankyo.
we need further study to find the answer for these patients
He received no fees personally. The other authors report no conflicts.
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
References
1. January CT, Wann LS, Calkins H, Chen LY, Cigarroa JE, Cleveland JC Jr,
locations)27–29; our data are not able to address this aspect, et al. 2019 AHA/ACC/HRS Focused Update of the 2014 AHA/ACC/HRS
as imaging data were not available. The location of ICH is Guideline for the management of patients with atrial fibrillation: a report
another limitation that was not addressed in this study. The of the American College of Cardiology/American Heart Association Task
Force on Clinical Practice Guidelines and the Heart Rhythm Society. J
results of ICH location on imaging studies are not available in Am Coll Cardiol. 2019;74:104–132. doi: 10.1016/j.jacc.2019.01.011
the claims database. Therefore, the extent and location of ICH 2. Kirchhof P, Benussi S, Kotecha D, Ahlsson A, Atar D, Casadei B, et al;
could not be evaluated in this study. Third, although the 2 treat- ESC Scientific Document Group. 2016 ESC Guidelines for the manage-
ment of atrial fibrillation developed in collaboration with EACTS. Eur
ment groups seem to be balanced well by IPTW using given
Heart J. 2016;37:2893–2962. doi: 10.1093/eurheartj/ehw210
17 covariates, there might be residual confounding factors, 3. Shen AY, Yao JF, Brar SS, Jorgensen MB, Chen W. Racial/ethnic
such as measurable data not available in the claims database 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: 17. Lee SR, Choi EK, Han KD, Jung JH, Oh S, Lip GYH. Edoxaban in
10.1016/j.jacc.2007.01.098 asian patients with atrial fibrillation: effectiveness and safety. J Am Coll
4. Wang KL, Lip GY, Lin SJ, Chiang CE. Non-Vitamin K antagonist oral Cardiol. 2018;72:838–853. doi: 10.1016/j.jacc.2018.05.066
anticoagulants for stroke prevention in asian patients with nonvalvu- 18. Lip GYH, Keshishian A, Li X, Hamilton M, Masseria C,
lar atrial fibrillation: meta-analysis. Stroke. 2015;46:2555–2561. doi: Gupta K, et al. Effectiveness and safety of oral anticoagulants among
10.1161/STROKEAHA.115.009947 nonvalvular atrial fibrillation patients. Stroke. 2018;49:2933–2944. doi:
5. Ruff CT, Giugliano RP, Braunwald E, Hoffman EB, Deenadayalu N, 10.1161/STROKEAHA.118.020232
Ezekowitz MD, et al. Comparison of the efficacy and safety of new 19. Lee SR, Choi EK, Han KD, Cha MJ, Oh S, Lip GYH. Temporal trends
oral anticoagulants with warfarin in patients with atrial fibrillation: a of antithrombotic therapy for stroke prevention in Korean patients with
meta-analysis of randomised trials. Lancet. 2014;383:955–962. doi: non-valvular atrial fibrillation in the era of non-vitamin K antagonist
10.1016/S0140-6736(13)62343-0 oral anticoagulants: a nationwide population-based study. PLoS One.
6. Lip G, Freedman B, De Caterina R, Potpara TS. Stroke prevention in 2017;12:e0189495. doi: 10.1371/journal.pone.0189495
atrial fibrillation: past, present and future. comparing the guidelines and 20. Brønnum Nielsen P, Larsen TB, Gorst-Rasmussen A, Skjøth F,
practical decision-making. Thromb Haemost. 2017;117:1230–1239. doi: Rasmussen LH, Lip GYH. Intracranial hemorrhage and subsequent is-
10.1160/TH16-11-0876 chemic stroke in patients with atrial fibrillation: a nationwide cohort study.
7. Chao TF, Liu CJ, Liao JN, Wang KL, Lin YJ, Chang SL, et al. Use Chest. 2015;147:1651–1658. doi: 10.1378/chest.14-2099
of oral anticoagulants for stroke prevention in patients with atrial fi- 21. Nielsen PB, Larsen TB, Skjøth F, Lip GY. Outcomes associated with
brillation who have a history of intracranial hemorrhage. Circulation. resuming warfarin treatment after hemorrhagic stroke or traumatic intra-
2016;133:1540–1547. doi: 10.1161/CIRCULATIONAHA.115.019794 cranial hemorrhage in patients with atrial fibrillation. JAMA Intern Med.
8. Nielsen PB, Larsen TB, Skjøth F, Gorst-Rasmussen A, Rasmussen LH, 2017;177:563–570. doi: 10.1001/jamainternmed.2016.9369
Lip GY. Restarting anticoagulant treatment after intracranial hemorrhage 22. Ottosen TP, Grijota M, Hansen ML, Brandes A, Damgaard D,
in patients with atrial fibrillation and the impact on recurrent stroke, Husted SE, et al. Use of antithrombotic therapy and long-term clinical
mortality, and bleeding: A Nationwide Cohort Study. Circulation. outcome among patients surviving intracerebral hemorrhage. Stroke.
2015;132:517–525. doi: 10.1161/CIRCULATIONAHA.115.015735 2016;47:1837–1843. doi: 10.1161/STROKEAHA.116.012945
9. Pennlert J, Asplund K, Carlberg B, Wiklund PG, Wisten A, Åsberg S, 23. Murthy SB, Gupta A, Merkler AE, Navi BB, Mandava P, Iadecola C,
et al. Antithrombotic treatment following intracerebral hemorrhage in et al. Restarting anticoagulant therapy after intracranial hemorrhage: a
patients with and without atrial fibrillation. Stroke. 2015;46:2094–2099. systematic review and meta-analysis. Stroke. 2017;48:1594–1600. doi:
doi: 10.1161/STROKEAHA.115.009087 10.1161/STROKEAHA.116.016327
10. Korompoki E, Filippidis FT, Nielsen PB, Del Giudice A, Lip GYH, 24. Nielsen PB, Skjøth F, Søgaard M, Kjældgaard JN, Lip GYH, Larsen TB.
Kuramatsu JB, et al. Long-term antithrombotic treatment in intracranial Non-Vitamin K antagonist oral anticoagulants versus warfarin in atrial
hemorrhage survivors with atrial fibrillation. Neurology. 2017;89:687– fibrillation patients with intracerebral hemorrhage. Stroke. 2019;50:939–
696. doi: 10.1212/WNL.0000000000004235 946. doi: 10.1161/STROKEAHA.118.023797
11. van Nieuwenhuizen KM, van der Worp HB, Algra A, Kappelle LJ, 25. Wilson D, Charidimou A, Shakeshaft C, Ambler G, White M, Cohen H,
Rinkel GJ, van Gelder IC, et al; APACHE-AF Investigators. Apixaban et al; CROMIS-2 Collaborators. Volume and functional outcome of intra-
versus Antiplatelet drugs or no antithrombotic drugs after anticoagu- cerebral hemorrhage according to oral anticoagulant type. Neurology.
lation-associated intraCerebral HaEmorrhage in patients with Atrial 2016;86:360–366. doi: 10.1212/WNL.0000000000002310
Fibrillation (APACHE-AF): study protocol for a randomised controlled 26. Inohara T, Xian Y, Liang L, Matsouaka RA, Saver JL, Smith EE, et
Downloaded from http://ahajournals.org by on September 9, 2020
trial. Trials. 2015;16:393. doi: 10.1186/s13063-015-0898-4 al. Association of intracerebral hemorrhage among patients tak-
12. Shoamanesh A. NASPAF-ICH: NOACs for Stroke Prevention in Patients ing non-vitamin k antagonist vs Vitamin K antagonist oral antico-
With Atrial Fibrillation and Previous ICH. Available at: https://clinical- agulants with in-hospital mortality. JAMA. 2018;319:463–473. doi:
trials.gov/ct2/show/NCT02998905. Accessed July 28, 2019. 10.1001/jama.2017.21917
13. Lee SR, Choi EK, Kwon S, Han KD, Jung JH, Cha MJ, et al. 27. Shoamanesh A, Charidimou A, Sharma M, Hart RG. Should patients
Effectiveness and safety of contemporary oral anticoagulants among with ischemic stroke or transient ischemic attack with atrial fibrillation
Asians with nonvalvular atrial fibrillation. Stroke. 2019;50:2245–2249. and microbleeds be anticoagulated? Stroke. 2017;48:3408–3412. doi:
doi: 10.1161/STROKEAHA.119.025536 10.1161/STROKEAHA.117.018467
14. Austin PC. The use of propensity score methods with survival or 28. Wilson D, Ambler G, Shakeshaft C, Brown MM, Charidimou A,
time-to-event outcomes: reporting measures of effect similar to those Al-Shahi Salman R, et al; CROMIS-2 Collaborators. Cerebral micro-
used in randomized experiments. Stat Med. 2014;33:1242–1258. doi: bleeds and intracranial haemorrhage risk in patients anticoagulated for
10.1002/sim.5984 atrial fibrillation after acute ischaemic stroke or transient ischaemic
15. Austin PC, Stuart EA. Moving towards best practice when using inverse attack (CROMIS-2): a multicentre observational cohort study. Lancet
probability of treatment weighting (IPTW) using the propensity score Neurol. 2018;17:539–547. doi: 10.1016/S1474-4422(18)30145-5
to estimate causal treatment effects in observational studies. Stat Med. 29. Martí-Fàbregas J, Medrano-Martorell S, Merino E, Prats-Sánchez L,
2015;34:3661–3679. doi: 10.1002/sim.6607 Marín R, Delgado-Mederos R, et al; HERO Study Investigators. MRI
16. Austin PC. Balance diagnostics for comparing the distribution of base- predicts intracranial hemorrhage in patients who receive long-term oral
line covariates between treatment groups in propensity-score matched anticoagulation. Neurology. 2019;92:e2432–e2443. doi: 10.1212/WNL.
samples. Stat Med. 2009;28:3083–3107. doi: 10.1002/sim.3697 0000000000007532