DOI: 10.1161/CIRCULATIONAHA.115.

015735

Restarting Anticoagulant Treatment After Intracranial Haemorrhage in

Patients With Atrial Fibrillation and the Impact on Recurrent Stroke,
Mortality and Bleeding: A Nationwide Cohort Study

Running title: Nielsen et al.; Anticoagulant treatment after intracranial haemorrhage

Peter Brønnum Nielsen, MSc, PhD1,2; Torben Bjerregaard Larsen, MD, PhD1,2;

Flemming Skjøth, MSc, PhD1,2; Anders Gorst-Rasmussen, MSc, PhD1,2;

Lars Hvilsted Rasmussen, MD, PhD1,2; Gregory Y.H. Lip, MD2,33

1
Dept
De pt of
of Cardiology,
C rd
Ca dio
iollogy, Atrial Fibrillation Study
y Gr
Group, Aalborg U
University
niveers
rsity Hospital, Aalborg,

Denmark; 2Aa
Denmark; Aalborg
A lborg Thrombosis
Th
Thromb
h mb
mboosiss Research
Resseaarchh Unit,
Unit, Dept
Dept
ptt of
of C
Clinical
liiniccall Med
Medicine,
diccinee, Faculty
Facu
ultty off Health,
Heallth
l th ,

Aalborg
Aalb
Aa lbor
lborgg University,
or Univer
Un ersiity
er y, Aa
Aalborg,
Aalbor
rg,
g D rk;; 3Un
Denmark;
enmaark
en University
Univ
iver
iversi
r ity ooff Bi
Birm
Birmingham
rmin
rmi gh
ingham
am C
Centre
ent
n ree ffor
or C
Cardiovascular
ardioovascu
ar ulaar

Sciences, City Hospital, Birmingham, United Kingdom

Address for Correspondence:

Gregory Y.H. Lip, MD
City Hospital Birmingham
Dudley Road, B18 7QH
United Kingdom
Tel: +44 121 507 5080
Fax: +44 121 507 5097
E-mail: g.y.h.lip@bham.ac.uk

Journal Subject Code: Stroke treatment - medical:[70] Anticoagulants

1
Downloaded from http://circ.ahajournals.org/ by guest on September 23, 2015
 DOI: 10.1161/CIRCULATIONAHA.115.015735

Abstract

Background—Intracranial haemorrhage is the most feared complication of oral anticoagulant

treatment. The optimal treatment option for atrial fibrillation patients surviving an intracranial
haemorrhage remains unknown. We hypothesised that restarting oral anticoagulant treatment
was associated with a lower risk of stroke and mortality compared to not restarting.
Methods and Results—Linkage of three Danish nationwide registries in the period between
1997 and 2013 identified atrial fibrillation patients on oral anticoagulant treatment with incident
intracranial haemorrhage. Patients were stratified by treatment regimens (no treatment, oral
anticoagulant treatment or antiplatelet therapy) post the intracranial haemorrhage. Event rates
were assessed 6 weeks after hospital discharge and compared with Cox proportional hazard
models. In 1,752 patients (one year of follow-up) the rate of ischemic stroke/systemic embolism
and all-cause mortality (per 100 person-years) for oral anticoagulant treatment tre
eat
ated
ed ppatients
treated atie
at ient
ie ntss
nt
were 13.6, compared to 27.3 for non-treated patients and 25.7 for patients receivingg ant
tipllattellet
antiplatelet
herapy. For recurrent intracranial haemorrhage: 8.0 for oral anticoagulant treated, compared to
therapy.
8.66 for
for non-treated,
non-
n-tr
n- trreate
teed, and 5.3 for antiplatelet therapy.
therap
py. The adjusted hazard
h zard
ha rd
d ratio
ratio of ischemic
tro
oke/systemicc embolism
stroke/systemic e boolism
em sm and
and all-cause
all
ll-c
-cau
-c a se mortality
au mor
orrtaality
t was
was 0.55
0.55 (95%CI
(95
9 %C
%CII 0.39-0.78)
0.39--0.7
78) iin
n pa
patien
entss oon
en
patients n or
ora
oral
an
ntiico
c agulantt treatment
anticoagulant treeattmen
nt com
mpaared too nno
compared o trea
atm
mentt. F
treatment. orr is
For isch
chem
emic st
em
ischemic troke//syystem
emic emb
em
stroke/systemic mboliism
mb
embolismm an
nd
and
for all-cause
alll-ca
al l-ca
cauuse
us mortality
mo
ort
rtal
alit
al ityy HRs
it HR were
weere 0.59
0.59 (95%CI
(95%
(9 5%CI
5% CI 0.33-1.03)
0.333-1.
1 03
03)) and
an
nd 0.55
55 (95%CI
(955%C
%CII 0.37-0.82),
0.37
37-0
37-0
0.82)),
espectively
respectively.
Conclusions—Oral anticoagulant treatment was associated with a significant reduction in
ischemic stroke/all-cause mortality rates, supporting oral anticoagulant treatment reintroduction
post-intracranial haemorrhage as feasible. Future trials are encouraged to guide clinical practice
in these patients.

Key words: stroke, anticoagulant, hemorrhage, cerebral infarct

2
Downloaded from http://circ.ahajournals.org/ by guest on September 23, 2015
 DOI: 10.1161/CIRCULATIONAHA.115.015735

Introduction

Patients with atrial fibrillation (AF) are increasingly treated with oral anticoagulation (OAC),

which has been shown to reduce the risk of stroke/systemic embolism and all-cause mortality,

compared to control 1. Until recently, OAC usually meant Vitamin K Antagonists (VKA, e.g.

warfarin), and the most feared complication was major bleeding, particularly intracranial

haemorrhage (ICH). More recently, the Non-VKA Oral Anticoagulants (NOACs) have been

introduced into clinical practice and are associated with relatively lower – but not zero – risk of

ICH, when compared to warfarin 2,3. The acute development of ICH confers a poor prognosis

with a high rate of mortality and disability, whether or not a NOAC is used.

Patients with AF who survive an ICH are at an increased risk of subsequen

subsequent
entt ischaemic
en isch
ischae
ch aemi
ae micc
mi

troke 4,5. Thus, a major unanswered question is the efficacy and safety of restarting OAC
stroke

reatm
tmen
tment,
en t, relative
treatment, rel
elativ
ve to not restarting OAC, following
follow
win
ingg a presentation with ICH 6–9. Patients with
h IC

prevalent
prev
valent ICH were
werre excluded
exc
xclu
lude
lu dedd from
de from
m randomized
ran
ndoomiize
zed clinical
cllin
nical
al trials
tri
rial
a s off O
al OAC
AC ffor
or stro
stroke
ro
oke pprevention,
reve
re vent
ve ntio
nt on, aand
nd

only
onnly small
small ccase control
ase co
as contro cohort
ol or co
ohort
oho studies
or sttud v ble 10–
aavailable
u iees aree availab 10–13
. Currently,
Curr
Cu rren
entlly,
en y the
thee opt
optimal
p im
pt imal treatment
tre
reatme
meent

opti
op tion
ti
optionon iiss unkn
un
nkn
know
ownn an
unknown andd re
requ
quir
ires
ir
requireses bbalancing
alaanc
al ncin
ingg th
in thee co
comp
mpet
mp etin
et
competingingg ri
in risk
sk ooff is
isch
chem
ch emic
em
ischemicic sstroke
tro
roke
ke aand
nd rrecurrent
ecur
ec urre
rre
rent
nt

ICH in AF patients.

Using Danish nationwide registries cohort, we investigated the hypothesis that, amongst

AF patients presenting with an ICH, restarting OAC was associated with a lower risk of recurrent

stroke and/or mortality, but with a small increase in major bleeding (particularly ICH) compared

to not restarting.

Methods

We used the civil registration number assigned to all Danish residents to link three nationwide

3
Downloaded from http://circ.ahajournals.org/ by guest on September 23, 2015
 DOI: 10.1161/CIRCULATIONAHA.115.015735

databases 14–16 as follows: (i) discharge diagnoses classified by the International Classification of

Diseases (ICD), admission and discharge dates were obtained from the Danish National Patient

Registry; (ii) dispensed prescriptions identified by the Anatomical Therapeutic Chemical (ATC)

classification code, date of purchase, package size and volume of the medication since 1994 were

obtained from the Danish National Prescription Registry; and (iii) information on age, sex, date

of birth, emigration and vital status were attained from the Danish Civil Registration System.

Retrospective studies do not require ethical approval in Denmark.

Study population

All Danish citizens with an incident nonvalvular AF diagnosis between January 1997 and

December 2013 were identified. Patients with a subsequent incident ICH (ICD10:
0:: I60-I62;
I60
60-I
-I62
-I 62;;
62

S063C; S064-S066) requiring admission to a hospital were included. Exclusion criteria were

valv
vul
ular
ar A
valvular F de
AF ddefined
fiine
ned as presence of mitral stenosi
siss oorr mechanical heart vvalve
si
stenosis a ve (ICD10: I05 and
al

Z 9552-Z954). Pa
Z952-Z954). atie
tients
ts w
Patients ithh IC
it
with ICH
H orr ccomplications
ompplicat
om atio
at ons ffrom
room
om IICH
CH ((ICD10:
ICD
IC D10:
0: II690-692)
6900-6992) pprior
69 rior
ri or too th
thei
eirr AF
ei
their

di
iag
agnnosis we
diagnosis werre also
were also excluded.
ex
xcluded
ed
d. We
W fu
furt
the
herr re
further equuired
requiredd ppatients
attie
iennts to
o bbee in
nO AC tre
OAC eat
atm
mentt ddefined
treatment efin
ned ass a

clai
cl aime
ai medd VKA
me
claimed VKA or NOAC
NOA
OACC prescription
pres
pr escr
es crip
cr ipti
ip tion
ti on within
wit
ithi
it hinn 6 months
hi mont
mo nths
nt hs prior
pri
rior
or to
to the
the ICH
ICH event.
even
ev entt. This
en Thi
hiss was
waas done
done to
to

support the assumption that included patients were indeed in OAC treatment at the time of the

incident ICH event. Patients were followed in the National Patient Registry after a ‘quarantine

period’ defined as 6 weeks after hospital discharge. This period was chosen to ensure that events

could reasonably be attributed to treatment regimens rather than inadequate correction of the

initial coagulation deficit, i.e. reversal of the incident ICH (and adjacent) event 9,17–20 . Further,

the use of a relatively long quarantine period can reduce confounding by indication, as those

patients initially deemed at a very high risk of recurrent ICH would probably not be advised to

resume anticoagulation within this period (e.g. patients with lobar haemorrhage 21).

4
Downloaded from http://circ.ahajournals.org/ by guest on September 23, 2015
 DOI: 10.1161/CIRCULATIONAHA.115.015735

Exposure to antithrombotic treatment

We identified all prescriptions of antiplatelet therapy (aspirin/thienopyridines) and OAC

treatment (VKA/NOAC) to classify patients by treatment regimens as: ‘No antithrombotic

treatment’, ‘Antiplatelet therapy’ or ‘OAC treatment’. The following ATC codes were utilised to

account for OAC treatment: B01AA (coumarin derivatives), B01AE07 (dabigatran etexilate);

B01AF01-02 (rivaroxaban and apixaban). For antiplatelet therapy the following ATC codes were

utilised: B01AC06 (aspirin) and B01AC04, B01AC22 and B01AC24 (Thienopyridines). When

discharged from the hospital, all included patients were initially assigned to ‘no antithrombotic

treatment’ and if antithrombotic therapy was reinitiated, to either ‘Antiplatelet therapy’ or ‘OAC

treatment’
reatment’ categories, whichever prescription was claimed first [Figure 1]. Treatment
Treatme
ment
me nt exposure
exp
xpos
osur
os ure
ur

was consequently considered as a time-dependent covariate.

Outcome
Outc omee measures
tccom measur
me ures
ur e

The
T analysis
hee primary an
naly iss pplan
alysis laan us
lan ed tthe
used he pprincipal
rinc
ri al ooutcomes
ncipal mess ooff is
utcom
om ischemic
isch emicc sstroke/systemic
chem
ch em trookee/syyst
tr stem ic eembolism
emic
em mbollis
mb ismm

(SE),
E),, all-cause
SE) use mortality
all-caus
us mortallityy andd recurrent
mo reecu
c rr e t ICH.
rren CH.. Given
IC Givenn the
thee severity
ty ooff th
se erity
seve studied
tudieed ooutcomes,
thee st utcom wee oonly
omes,, w
om nlly

considered
cons
co nsid
ns ider
id ered
ered eevents
ntss if tthe
veent
vent he ppatient
atie
at ient
ie as aadmitted
nt was dmit
dm tedd to tthe
itte
it te tall 22; hence,
he hhospital
ospi
os pita
pi ta henc
he ncee, wee did
nc did not
not consider
cons
co nsid
ns ider
id er

ambulatory diagnosis. Additionally, emergency room coded diagnoses were not included in this

study due to poor validity 23. The primary study endpoint was a combined endpoint of ischemic

stroke/SE (ICD10 diagnosis: I63; I64; I74) and all-cause mortality, given the positive impact of

OAC on stroke and mortality 1 and recognising that in registry data, some deaths may be due to

undiagnosed stroke/SE as neither post-mortems nor cerebral imaging are mandated.

Secondary analyses were carried out for ischemic stroke/SE; for recurrent ICH; for major

extra cranial bleeding (ICD10: D62; J942; H113; H356; H431; N02; N95; R04; R31; R58; K250

K260 K270 K280 K290), and for all-cause mortality. If, in the primary analysis, antithrombotic

5
Downloaded from http://circ.ahajournals.org/ by guest on September 23, 2015
 DOI: 10.1161/CIRCULATIONAHA.115.015735

therapy did show benefit versus being left untreated (commonly seen in practice post-ICH 5,24), a

secondary analysis calculating net clinical benefit (NCB) of resuming antithrombotic therapy (vs

no antithrombotic therapy) was calculated by the methods previously proposed balancing

ischaemic stroke reduction versus the increased risk of ICH 25. NCB with 95% confidence

intervals (CI) were based on rate differences and standard errors estimated obtained using

Poisson regression. We applied a weight of 1 for ischemic stroke/SE and all-cause mortality and

a weight of 1.5 for recurrent ICH as proposed by Singer et al 25. Accompanying the NCB

analysis, we also investigated the event rate for the composite outcome of ischemic

stroke/SE/major bleeding (including recurrent ICH). Additionally, we carried out a sensitivity

analysis by altering the quarantine period when investigating the outcome of ische
hemi
he micc st
mi
ischemic stro
roke
ro ke/S
ke
stroke/SE/SE
/S

and all-cause mortality and recurrent ICH. This was done to assess if the choice of quarantine

peri
iod would
period wou
ould
ld differ
difffe
ferr from the main analysis (using
(usin
ng a 6 weeks quarantinee pperiod).
eriod).

P attient
ti
Patient characcte
terrist
sttic
characteristicsicss an
andd co
conc
nccom
mit
itant me
concomitant medica
ation
n
medication

Co
omo
morbidittie
iess w
Comorbidities ere as
were sceertain
in
ned fro
ascertained om ppreceding
from reeced
din
ng ho
ospittall dia
hospital agn
gnos
osess uuntil
os
diagnoses n il ddischarge
nt isch
chaarge ffrom
ch rom th
rom he
the

nci
cide
dent
de
incidentnt ICH
ICH event.
event
veent
nt. Filled
Fill
Fi lled
lled prescriptions
pre
resc
scri
sc ript
ri ptio
pt ions
io ns 1 year
yea
earr pprior
ea rior
ri or tto
o ba
base
seli
se line
li
baselinene ddefined
efin
efined
ined oother
ther
ther cconcomitant
onco
on comi
comita
mitant
tant

medication. The cardiovascular comorbidity and risk of stroke at baseline were assessed by the

CHA2DS2-VASc [congestive heart failure, hypertension, age >75 years, diabetes mellitus, stroke,

vascular disease, age 65–75 years, and female sex)] score 26,27. We calculated the HAS-BLED

[hypertension, abnormal renal/liver function, stroke/thromboembolism, bleeding history, labile

international normalized ratio (not included), elderly (age >65 years), drug consumption/alcohol

excess (aspirin was not included due to status of exposure)] score for each patient to assess the

bleeding risk 28. Supplementary Table 1 provides detailed description on outcomes and

concomitant medication.

6
Downloaded from http://circ.ahajournals.org/ by guest on September 23, 2015
 DOI: 10.1161/CIRCULATIONAHA.115.015735

Statistical analyses

Patients were followed until occurrence of the following censoring events: a claimed prescription

indicating dual treatment regimen (both OAC and antiplatelet treatment), occurrence of a study

endpoint, emigration, end-of-study, or death whichever came first [Figure 1]. For outcome

analyses events were ascertained from 6 weeks after hospital discharge. Exposure to

antithrombotic treatment was regarded as two time-dependent covariates (one for OAC treatment

and one for antiplatelet therapy) with one irreversible binary transition (from 0 to 1). We

calculated crude events rates at 1 and 5 years by dividing the number of events occurring during

follow-up with the person-years of follow-up for each treatment group. To compare event rates

associated under antithrombotic treatment regimens vs no antithrombotic treatmen

treatment,
t, ccrude
ent,
en rude
ru de aand
nd

adjusted hazard ratios (HR) were estimated by means of the Cox proportional hazards model.

adjusted
Thee ad
adju
just
ju sted
st analyses
ed ana
aly
lyses included information on ag spline);
agee (restricted cubic spli
linne); gender (binary); yea
li year

of iinclusion
of nclusion ((categories
caate
teggori
ries
ri e ooff ffive
es ve yyears);
ive ears
ears
rs); ssince
) time si
); last
ast cclaimed
nce la laim
la imed
im OAC
ed O AC pprescription
resccript
re ptio
pt before
ionn be
io befo
fore
fo re the
he

incident
nci dent ICH
c de H event
nt (restricted
eveent ed cubic
(reesttricted u ic spline);
cub ne)); and
spplin nd categories
an ri s off the
catteggories t e CHA
th CHA2D
DS
S2-V
-VAScc sscore
corre (0; 1;

•
•22 points)
poin
po ints
ints)) and
ts and HAS-BLED
HAS
HA S-BL
BLED
BL ED score
sco
core
re (0-2;
(0-2; •3 points).
2; •3 poin
po ints
ints)). W
ts Wee ad
addi
diti
di tion
ti onal
on
additionallyally
al ly use
sedd in
se
used info
form
fo rmat
rmatio
at ionn on iischemic
io
information sche
sc hemi
he micc
mi

stroke/SE and/or recurrent ICH events during the quarantine period (binary). The reasoning

behind this adjustment was that such events could affect the choice of antithrombotic treatment

before the observation time commenced. Additionally, a sensitivity analysis was conducted by

excluding patients who experienced an event within the quarantine period.

To depict the overall five year prognosis, we obtained the Kaplan-Meier estimates. The

patients were stratified according to purchase of OAC (respectively, antiplatelet drugs) during

the quarantine period (6 weeks landmark) and patients with dual treatment exposure were

excluded (prescription claim of both an OAC and antiplatelet drug). To assess the impact of

7
Downloaded from http://circ.ahajournals.org/ by guest on September 23, 2015
 DOI: 10.1161/CIRCULATIONAHA.115.015735

allowing more time for claiming a prescription used for treatment regimen allocation, we made a

sensitivity analysis by setting the landmark to 180 days.

We used a two-sided p-value threshold of p<0.05 for statistical significance. Statistical

analyses were performed by using SAS 9.3 (SAS Institute) and Stata version 13 (StataCorp LP).

Results

A total of 6,138 patients with non-valvular AF who suffered an incident ICH were identified in

the period from January 1st 1997 to December 31st 2013. Within the quarantine period of 6 weeks

1,652 patients died, while 32 patients were excluded due to insufficient follow-up time (i.e.

encountered study-end within the quarantine period). The study population was co
comp
comprised
mpri
mprise
risedd of
se

1,752 patients who fulfilled the inclusion criteria by having claimed an OAC prescription within

six
ix months
mont
mo nths
nt hs ooff thee in
iincident
cident ICH event and were st
stil
still
illl alive after the 6 we
il weeks
eek
eks quarantine period; see

Figure
F iggure
gu 1 for fflowchart
lo
owc
w haart aand
nd Ta
Table
Tabl
blee 1 ffor
bl or sstudy
tudy
dy ppopulation
opul
ullatio
on ch
char
characteristics.
araccte
ar teri
rist
stic
ticcs.
s TThe
hee pproportion
ropo
port
po rtio
rt on of

patients
pa
ati ents included
t en inclu d d in
lude
lu in the
thhe analysis
analy
ly siss with
ysi wiith prior
priior OAC
OA treating g receiving
treeatiing reece v ng VKA
eceivi
vi A was
VKA as 65%,
waas 65%,, and
and the
th
he

comb
co
combination
mbin
mb inat
in atio
at ionn of VKA
io VKA and
and antiplatelet
ant
ntip
ipla
ip late
la tele
te lett therapy
le ther
th eraapy
er py iin
n 33
33%
33%.
%. W
With
ithh NO
it NOAC
NOACs,
ACss, tthe
AC he pproportion
ropo
ro port
po rtio
rt ionn was
io waas 2%

while <1% received combination therapy with a NOAC and antiplatelet therapy. The mean

CHA2DS2-VASc score was 3.9 and the mean HAS-BLED score was 3.2, indicating a study

population at high risk of both thromboembolic events as well as high risk of bleeding.

Of the patients who resumed OAC treatment (n=621), 77% claimed an OAC prescription

within first 3 months after hospital discharge, with an overall median days of 34 to first claimed

prescription. For those who received antiplatelet therapy (n=759), 65% claimed a prescription

within first 3 months, with an overall median time of 24 days. During the first year of follow-up,

218 patients either shifted treatment or initiated dual antithrombotic treatment and were

8
Downloaded from http://circ.ahajournals.org/ by guest on September 23, 2015
 DOI: 10.1161/CIRCULATIONAHA.115.015735

accordingly censored.

Stroke/SE and all-cause mortality

A total of 39 patients suffered an ischemic stroke/SE during the quarantine period of 6 weeks; of

those, 5 patients had already claimed an OAC prescription, while 11 patients received

antiplatelet therapy. The overall event rates (expressed as per 100 person-years) using 1 year of

follow-up of the combined endpoint of ischemic stroke/SE and all-cause mortality in OAC

treated vs no antithrombotic treatment were 13.6 vs 27.3 (adjusted HR: 0.55, 95%CI 0.39-0.78),

and 25.7 for antiplatelet therapy (adjusted HR: 0.87, 95%CI 0.67-1.14) [Table 2 and Figure 2].

The event rates for ischemic stroke/SE in OAC treated vs no antithrombotic treatment were 5.3

vs 10.4 at one year of follow-up (adjusted HR: 0.59, 95%CI 0.33-1.03), whilst for
orr aantiplatelet
ntip
nt ipla
ip late
la tele
te lett
le

therapy,
herapy, 10.3 (adjusted HR: 0.98, 95%CI 0.65-1.49). For all-cause mortality, the event rates were

9.7 fo
forr OA
OAC ttreated
C tr ted vs 19.1 for no antithrombotic
eaate antithromboti treatment (adjusted HR
tiic tr HR: 0.55, 95%CI 0.37-

0.82),
0.822), and 19.5 for
for antiplatelet
an
ntipl
plat
pl therapy
ateeleet th
at the
eraapy (adjusted
apy (a
adjjustted HR:
te HR 0.90,
R: 0.
0.90 95%CI
90, 95
90 %CII 00.67-1.21).
95%C
%C .67
67-1
1.21)
1).. Th
1) ccrude
Thee cr ude an
ud andd

adjusted
ad
dju
just analyses
s ed ana
st lysees contrasting
aly conttraastin treatment
ng tr
trea
e tm nt rregimens
men egi
gime
mens w
me were
eree ccomparable.
om
mpar
arab
ar blee. Supplementary
ab S pp
Su mentarry Table
ppleme
me Taable 2

provides
prov
pr ovid
ides
ides eevent
veent rrates
vent es bbased
ates
at ed aand
ased
as nd adjusted
adj
djus tedd hazard
uste
ste haza
ha ard ratios
zard ratioss based
atio
io base
ba sedd on 5 years
se yea
ears
ea rs of
of follow-up,
foll
fo llow
ll ow-upp, sshowing
howi
ho wing
ing

consistency.

Figure 3 depicts the Kaplan-Meier estimates of 5 year survival in patients who resumed

OAC treatment, respectively received antiplatelet therapy, and those who did not. A landmark at

6 weeks (quarantine period) was used to allocate patients treatment regimens: 1,089 were

assigned to the no antithrombotic treatment group; 303 to OAC treatment; and 360 to the

antiplatelet therapy group.

Recurrent ICH and major bleeding

Of a total of 298 recurrent ICH events, 177 events (60%) occurred during the quarantine period,

9
Downloaded from http://circ.ahajournals.org/ by guest on September 23, 2015
 DOI: 10.1161/CIRCULATIONAHA.115.015735

i.e. before the observation time commenced; of these 27 patients had claimed an OAC

prescription, while 35 patients received antiplatelet therapy. For recurrent ICH using one year of

follow-up, the rates were 8.6 for OAC treated vs 8.0 for no antithrombotic treatment (adjusted

HR: 0.91, 95%CI 0.56-1.49), and 5.3 for antiplatelet therapy (adjusted HR: 0.60, 95%CI 0.37-

1.03). The event rates of major extra cranial bleedings in OAC treated vs no antithrombotic

treatment were 1.5 vs 1.5 (adjusted HR: 0.92, 95%CI 0.30-2.76), and 2.6 for antiplatelet therapy

(adjusted HR: 1.57, 95%CI 0.62-3.92) [Table 2 and Figure 2]. The crude and adjusted analyses

contrasting treatment regimens for bleeding outcomes were broadly comparable.

Secondary analysis

The one-year event rate for the combined endpoint of ischemic stroke/SE/major bl
bleeding
blee
e di
ee ding
ng

(including
including ICH) was 27.4 for no antithrombotic treatment, 17.1 for OAC treatment, and 23.2 for

patients
patien
en who
nts w ho receiving
o receiv ng antiplatelet therapy. As OAC
e vi
ei OAC did demonstrate a significant
si reduction in

stroke/SE
oke/SE andd al
tro all-cause
alll-ccaus mortality,
use mo
mort
rtal
rt alit
ality,
it wee ca
y, w ccalculated
lcul
ulat
ulatted the one
ne yyear
hee on
ne earr N
ea NCB
CB ooff OA
OAC
O AC or aantiplatelet
ntip
nt ipla
ipl te
tellett

therapy
apyy use versus
herap ve us no antithrombotic
an mbottic treatment.
ntithrrom tre
reat
a men The
ment. T hee NC
NCBB wa calculated
was ca
alccul
ulated weighted
ed ass a we
w ig
ght sum
um ooff
hted su

rate
ate ddifferences
iffe
if fere
fe renc
re nces
nc es ffor
or tthe
he ccombined
ombi
ombine
bi endpoint
nedd en
ne endp
dpoi
dp oint
oi nt ooff is
ischemic
isch
chem
ch ic sstroke/SE
emic
em trok
tr oke/
ok SE aand
e/SE
e/ nd aall-cause
ll-cau
cause mortality
se morta
or tali
ta ty aand
lity
li nd

recurrent ICH. The NCB for OAC vs versus no antithrombotic treatment was 14.6 (95%CI 6.4-

22.8), while the NCB was non-significant for antiplatelet therapy vs versus no antithrombotic

treatment, 6.5 (95%CI -2.1-15.2).

Sensitivity analysis

A sensitivity analysis of the Kaplan-Meier estimates stratified according to treatment regimens

using a landmark at 180 days (relative to hospital discharge) for treatment allocation is shown in

Supplementary Figure S1. A total of 1541 patients contributed to this analysis, and the survival

curves were alike to those presented in Figure 3 supporting OAC treatment.

10
Downloaded from http://circ.ahajournals.org/ by guest on September 23, 2015
 DOI: 10.1161/CIRCULATIONAHA.115.015735

Altering the quarantine period ranging from 2 weeks to 10 weeks did essentially not

affect the primary outcome of ischemic stroke/SE and all-cause mortality, see Figure 4. The total

number of patients contributing to these analyses was 1.838; 1.793; 1.752; 1.732; 1.712 using 2;

4; 6; 8; and 10 weeks quarantine period, respectively. Contrary, the outcome of recurrent ICH

was somewhat affected by the choice of quarantine period. The HR contrasting OAC treatment

vs no treatment was ranging from 0.41 (95%CI 0.27-0.63) for 2 weeks quarantine period, to HR

1.42 (95%CI 0.83-2.43) when applying a 10 weeks quarantine period. The obtained results on

HR contrasting treatment regimens were generally not affected by excluding those patients who

experienced an event (ischemic stroke/SE and/or recurrent ICH) during the quarantine period.

d aan
However, the outcome of recurrent ICH for patients on OAC treatment displayed adjusted
n ad
adju
just
ju sted
st HR
ed H R

of 1.30 (95%CI 0.75-2.27).

Discussion
D isscussion
sc

In
n tthis analysis
his analys
hi ysis
ys wee foun
i w found
und tha
un that AF
at A
at patients
F pa
patie who
entts w suffer
ffeer aan
hoo suff ICH
n IC
CH ar very hhigh
aree very ig
gh risk ischaemic
risk of is
iscchaaem
mic

stroke
tro ke and
roke mortality,
and m orta
or tali
ta lity
li ty, if tthey
hey are
he are not
not on antithrombotic
ant
ntit
ithr
ithrom
hrombo
omboti
boticc therapy.
ti ther
theraapy
er py. Patients
Pat
atie
ient
ientss who
nt weere nnot
whho were ot oon
n OA
OAC
C

treatment had the worse adverse outcomes, which were similar to rates seen with antiplatelet

therapy. Importantly, OAC treatment was associated with a significant reduction in subsequent

ischaemic stroke and all-cause mortality.

ICH is the most feared complication of OAC treatment, being associated with a high

mortality and morbidity, and survivors often have greater disability. Patients with ICH were not

included in randomised trials, and it is an open problem to decide the best time window to

reintroduce OAC treatment following a presentation with ICH. Our analysis does not define the

best time window to reintroduce OAC but makes the observation in this non-randomised

11
Downloaded from http://circ.ahajournals.org/ by guest on September 23, 2015
 DOI: 10.1161/CIRCULATIONAHA.115.015735

observational cohort that OAC use was associated with significantly lower risks of ischaemic

stroke and all-cause mortality.

The risk of recurrent ICH was also high in our patients, and was equally high amongst

non-OAC treated or OAC-treated, while slightly lower in patients who received antiplatelet

therapy. Factors contributing to recurrent ICH have been reported in different studies 8,9,29,30 and

include increasing age, concomitant use of aspirin or NSAIDs, uncontrolled hypertension, etc.

As previously shown in clinical trial and nationwide registries, the HAS-BLED score is the only

bleeding risk score that is predictive of major bleeding risk, and reflects the risk factors

commonly associated with this complication 4,28,31.

Common practice is to consider no antithrombotic

m treatment or antiplatele
antiplatelet
et th
ther
therapy
erap
er apyy
ap

(commonly
commonly aspirin) as a ‘safer’ alternative to OAC following a presentation with ICH, and the

AHA/ASA
AHA/
A/AS
A/ ASA
AS A gu
guidel
guideline
elin
eli e provides a Class IIb recom
in recommendation
mme
mendation of antiplate
antiplatelet
ele
lett therapy after all types

of ICH
of C (and “O
ICH “OAC
OAC might
AC m ht bbee cconsidered
ight
ig onsi
onsi
s de after
dereed af
fte nnon-lobar
ter no n-lo H”) 32. Our
ICH”)
obaar IC
ICH”
H” Our da
ddata
ata ddo
o no
nott su
supp
support
por
o t su
such
ch

an
n approach
app
pproachh with
pp w th
wi h a higher
hig
gher ischaemic
issch
chaaemi
micc stroke
sttro
oke and
an
nd substantially
suubstaanttiall
llly hi
hhigher
gheer mo
mort
mortality,
talit
ity,
it y, withh a no
non-
on-

significant
ign
gnif
ific
if ican
ic antt reduction
an redu
re duct
ctio
ct ionn in recurrent
io rec
ecur
urre
rre
rent
nt ICH
ICH amongst
amo
mong
ngst
ng st antiplatelet
ant
ntip
ipla
iplate
latele
telett therapy
le ther
therap
erapy users.
ap usser
user
erss. Ev
E
Even
ven
en iin
n ra
rand
randomised
ndom
nd omis
omised
is ed

trials of stroke prevention in AF, the risk of ICH (and major bleeding) is no different between

OAC and aspirin treated patients, especially in the elderly 33,34. Also, our secondary analyses

show that the NCB for OAC treatment was positive, vs no antithrombotic therapy – whilst the

NCB for aspirin vs no antithrombotic therapy was non-significant.

The present study has clinical implications, in that it supports OAC re-introduction as

soon as clinically feasible. However, given the non-randomized design and data limitations

caution on over-interpretation the results are warranted: we were not able to distinguish between

the severities of risk factors, which may reflect physician’s choices of OAC resumption.

12
Downloaded from http://circ.ahajournals.org/ by guest on September 23, 2015
 DOI: 10.1161/CIRCULATIONAHA.115.015735

Potentially correctable risk factors for ICH should be addressed including assessment of the

HASBLED score, correction of uncontrolled high blood pressure, concomitant use of aspirin or

NSAIDs, etc. ICH can sometimes be related to trauma or a vascular anomaly, and for the latter,

liaison with a neurosurgical service is important.

Limitations

We did not have data on INR values hence no estimate of OAC intensity, and we also recognise

the importance of the quality of anticoagulation control, as reflected by time in therapeutic range
35,36
. VKAs are associated with a relatively higher rate of ICH compared to NOACs, but the latter

were only used in a minority of patients in ourr study. We also did not ha
have
ve any cer
cerebral
ereb
er ebra
eb rall im
ra imag
imaging
agin
ag ing
in

data, to distinguish the subtypes of ICH, but ourr principal hypothesis pertained
pertained to the simple

ques
question
sti
tion
on of
of whether
w etthe
wh her or not restarting
r OAC after ICH
ICH was associatedd with
wit
ithh a beneficial effect on
it

stroke
tro mortality,
oke and mor
rta in ppatients
talitty, in atie
at ient
ie ntss wi
nt with AF.
th A Indeed,
F. In
Ind the
de , th
deed, presence
he pr
preesen
ence
ence ooff m
microbleeds
robbleeeds iiss an imp
icro
ic ro important
mpor
mp orta
or tant
ta n
nt

risk
isk for
o ICH,
for H, and
ICH a d many
an ny stroke
many strokke physicians
phys
ph ysic
ys ns would
i iaans
ic ould be
wou be cautious
cau
uti about
tiouus ab
abou restarting
ouut re
est OAC
s arttingg O patient
AC inn a pa
patien
nt

with
wi widespread
ith wid
ides
idespr
esprea
pr cerebral
eadd ce
ea cere
rebr
re bral
br microbleeds
al micro
icrobl
ro blee
bl ds 37,
eeds
ee 37,38
,38
.

There is a potential risk of misclassification of recurrent ICH events due to routine

procedures after hospital discharge: patients are likely to be re-admitted to a hospital for a

follow-up computerized tomography scan examining re-bleedings. This re-admission could

inherit the coding from the incident ICH with no evidence of any recurrent bleeding. This

conjecture was supported by the sensitivity analyses of altering the quarantine period, and also

by excluding those patients who experienced an event during the quarantine period. On the other

hand, with the lack of information on cause of death, the number of ischemic events or recurrent

ICH could be underestimated if they carried a terminal outcome. Our analysis on treatment

13
Downloaded from http://circ.ahajournals.org/ by guest on September 23, 2015
 DOI: 10.1161/CIRCULATIONAHA.115.015735

initiation was performed using an assumption of adherence to treatment (unless changed from

OAC to antiplatelet therapy or vice versa), recognising the limitations of a ‘real world’ cohort

design, where patients could change from treatment (and doses) to no treatment over time. We

also only included patients with an index hospitalisation for AF, and documented ICH – thus,

confounding by indication and selection bias could be evident, and our results may not be

generalizable to the ‘general’ non-hospitalised AF population. Nonetheless, most patients with

an ICH would require hospitalisation, and AF-related hospitalisations are common (and

increasing); thus, our analysis would reflect the likely burden of AF patients in healthcare

systems. Although this was a nationwide study, the ethnic background of the Danish population

iss presumably Caucasian, and thus the obtained results might not generalise to alll ethnic
ethn
et hnic
hnic

groups39.

In cconclusion,
oncluusio
on i n, AF patients who suffer an
io n IICH
C are at very high rrisk
CH iskk of ischaemic stroke
is

andd death, and

an d oour
u ddata
ur ata in
at indi
indicates
dica
di cate
ca tess po
te posi
positive
siiti
t ve cli
clinical
liniical be
li benefit
enef
efit
efit ffrom
rom an
rom anti
anticoagulant
ico
coag
agullantt tr
ag trea
treatment.
eatm
eatmen
tme t.
en t O
OAC
AC

treatment
reeatm
eatm
tment re
resulted
esu
sulteed in a significant
signiffic
ican
a t reduction
an redu
re ducttio
du i n inn subsequent
sub
bseequ
uen
ent ischaemic
iscchae
is chae
aem
micc stroke
stro
roke and
and all-cause
alll-c
-causse

mortality.
mort
mortal
rt alit
ality. As
it As such,
such
su ch, this
ch this study
stu
tudy
dy supports
suppo
up
ppo
port
rtss OAC
rt OAC reintroduction
rein
re intr
introd
tr oduc
oducti
cti
tion
on post-ICH,
pos
ostt-IC
ICH
ICH, but
butt due
duee to
to risk
risk of
of bias
bias by
by

unmeasured confounding, future randomized controlled trials investigating resumption of OAC

treatment post-ICH are encouraged to provide further evidence to guide clinical practice.

Funding Sources: The Obel Family Foundation partly funded this research by an unrestricted
grant. The sponsor had no role the design and conduct of the study; collection, management,
analysis, and interpretation of the data; and preparation, review, or approval of the manuscript.

Conflict of Interest Disclosures: Professor Lip has served as a consultant for Bayer, Astellas,
Merck, AstraZeneca, Sano¿, BMS/Pfizer and Boehringer Ingelheim, and has been on the speaker
bureaus for Bayer, BMS/Pfizer, Boehringer Ingelheim and Sano¿. Associate Professor Larsen

14
Downloaded from http://circ.ahajournals.org/ by guest on September 23, 2015
 DOI: 10.1161/CIRCULATIONAHA.115.015735

has served as an investigator for Janssen Scientific Affairs, LLC and Boehringer Ingelheim.
Associate Professor Larsen and Professor Rasmussen have been on the speaker bureaus for
Bayer, BMS/Pfizer, Roche Diagnostics, Boehringer Ingelheim and Takeda Pharma. Other
authors – none declared.

References:

1. Hart RG, Pearce LA, Aguilar MI. Meta-analysis: antithrombotic therapy to prevent stroke in
patients who have nonvalvular atrial fibrillation. Ann Intern Med. 2007;146:857–867.

2. Chan NC, Paikin JS, Hirsh J, Lauw MN, Eikelboom JW, Ginsberg JS. New oral
anticoagulants for stroke prevention in atrial fibrillation: impact of study design, double counting
and unexpected findings on interpretation of study results and conclusions. Thromb Haemost.
2014;111:798–807.

3. Ruff CT, Giugliano RP, Braunwald E, Hoffman EB, Deenadayalu N, Ezekowitzz MD MD,, Ca
Camm
mm
AJ, Weitz JI, Lewis BS, Parkhomenko A, Yamashita T, Antman EM. Compariso
Comparison on of
o tthe
he eefficacy
ffic
ff i acy
ic
and safety of new oral anticoagulants with warfarin in patients with atrial fibrillatio
fibrillation:
ionn: A m
meta-
eta-
t
analysis of randomised trials. Lancet. 2014;383:955–962.

4. FFriberg
ribe
riberg L
be L,, Rose
Rosenqvist
sennqvist M, Lip GYH. Evaluatio
se Evaluation
on ooff risk stratifica
stratification
caation sc
schemes for ischaemic
stroke
oke and bleeding
tro ble
leed
leed ing inn 182
din 182 6786788 patients
pat
atie ntss with
i nt
ie with atrial
atriall ffibrillation:
ibrill
llat
ll atio
at n the
ion:
io he Swedish
the Swe
weddis Atrial
i h At
tri
rial Fibrillation
al Fibriill
llat
a io
ationn
ccohort
co Eur
hort study. Euur HHeart
eaartt JJ.. 20
2012;33:1500–1510.
20112;3
33:
3:1150 00–15510.0

Nielsen
5. N iels
ielsen
ls PB,
en P Larsen
B, L arrse
s n TB Gorst-Rasmussen
TB, Go
Gorsrstt-Ra
Rasm
Ra mus
usse
senn A, S
se Skjøth
kkjø
jøth
jø th
hFF,, R
Rasmussen
asm
smuusse
sm usse
senn LH Lip
LH,, Li
ip GY H IIntracranial
GYH.
H. ntraacr
nt c an
ania
iall
ia
haemorrhage
haemorrhag ge and subsequent
subseqquent ischemic stroke in patients
p tients with atrial fibrillation: A nationwide
pa
cohort study. Chest. 2015;147:1651–1658
study Chest 2015;147:1651–1658.

6. Flynn RW V, MacDonald TM, Murray GD, Doney ASF. Systematic review of observational
research studying the long-term use of antithrombotic medicines following intracerebral
hemorrhage. Cardiovasc Ther. 2010;28:177–184.

7. Flynn R, Doney A. Antithrombotic medicines following intracerebral haemorrhage: where’s

the evidence? Ther Adv Drug Saf. 2011;2:205–211.

8. Majeed A, Kim Y-K, Roberts RS, Holmström M, Schulman S. Optimal timing of resumption
of warfarin after intracranial hemorrhage. Stroke. 2010;41:2860–2866.

9. Marsh EB, Gottesman RF. Brain hemorrhage: restarting anticoagulation after intracranial
hemorrhage. Nat Rev Neurol. 2011;7:130–132.

10. Romualdi E, Micieli E, Ageno W, Squizzato A. Oral anticoagulant therapy in patients with
mechanical heart valve and intracranial haemorrhage. A systematic review. Thromb Haemost.
2009;101:290–297.

15
Downloaded from http://circ.ahajournals.org/ by guest on September 23, 2015
 DOI: 10.1161/CIRCULATIONAHA.115.015735

11. Paciaroni M, Agnelli G. Should oral anticoagulants be restarted after warfarin-associated

cerebral haemorrhage in patients with atrial fibrillation? Thromb Haemost. 2014;111:14–18.

12. Goldstein JN, Greenberg SM. Should anticoagulation be resumed after intracerebral
hemorrhage? Cleve Clin J Med. 2010;77:791–799.

13. Horstmann S, Rizos T, Jenetzky E, Gumbinger C, Hacke W, Veltkamp R. Prevalence of

atrial fibrillation in intracerebral hemorrhage. Eur J Neurol. 2014;21:570–576.

14. Thygesen LC, Daasnes C, Thaulow I, Bronnum-Hansen H. Introduction to Danish

(nationwide) registers on health and social issues: Structure, access, legislation, and archiving.
Scand J Public Health. 2011;39:12–16.

15. Lynge E, Sandegaard JL, Rebolj M. The Danish National Patient Register. Scand J Public
Health. 2011;39:30–33.

16. Kildemoes HW, Sørensen HT, Hallas J. The Danish National Prescription Registry. Scand J
Public Health. 2011;39:38–41.
2011;39:38 41.

17. Bertram M, Bonsanto M, Hacke W, Schwab S. Managing the therapeutic dilemm

dilemma:
mmaa:
mm a: ppatients
atie
at ient
ie ntss
nt
with spontaneous intracerebral hemorrhage and urgent need for anticoagulation. J Neurol.
2000;247:209–214.

18.
18 Dee Vlees
8. D Vleeschouwer
esschouwer S, Van Calenbergh F, van Loon Loon J, Nuttin B, B, Goffin J, Plets C. Risk
analysis
anal
an thrombo-embolic
alysis of th
al hro
omb
mbo--em
embo
boli
bo licc an
li recurrent
andd re
recu
c rreentt bl
cu bbleeding
eedin events
ng evven
entstss iin the
n th management
he ma
mana
nageemeent off in
na intracranial
intr
t ac
acra
raani
nial
al
hhaemorrhage
ha emorrhage duee tto
em oral
o ora anticoagulation.
ral anti
ticcoagulatio
ti Acta
on. Ac cta Chir
a Ch Belg.
lg.. 105:268–274.
hirr Belg
lg 1055:26
10 68––274.

Cervera
19. Ce
Cerv
rver
rv eraa A, A
er Amaro
maaro S
maro S,, Cham
Chamorro
mor Oral
orrro A. Or
Oraal aanticoagulant-associated
al ntico
nticoag
agullan
antt-as
asssociiat
as ed iintracerebral
ated ntra
nt raace
c reebr al hhemorrhage.
bral em
mor
orrh
rhag
rh agee. J
Neurol.
N
Ne l 2012;259:212–224.
eurol
ol. 2012
20 12;2
12 ;259
;259:2
59 :212
:212–224
224
24.

20. Aguilar MI, Hart RG, Kase CS, Freeman WD, Hoeben BJ, García RC, Ansell JE, Mayer SA,
Norrving B, Rosand J, Steiner T, Wijdicks EFM, Yamaguchi T, Yasaka M. Treatment of
warfarin-associated intracerebral hemorrhage: literature review and expert opinion. Mayo Clin
Proc. 2007;82:82–92.

21. Eckman MH, Rosand J, Knudsen KA, Singer DE, Greenberg SM. Can patients be
anticoagulated after intracerebral hemorrhage? A decision analysis. Stroke. 2003;34:1710–1716.

22. Krarup L-H, Boysen G, Janjua H, Prescott E, Truelsen T. Validity of stroke diagnoses in a
National Register of Patients. Neuroepidemiology. 2007;28:150–154.

23. Johnsen SP, Overvad K, Sørensen HT, Tjønneland A, Husted SE. Predictive value of stroke
and transient ischemic attack discharge diagnoses in The Danish National Registry of Patients. J
Clin Epidemiol. 2002;55:602–607.

24. Maeda K, Koga M, Okada Y, Kimura K, Yamagami H, Okuda S, Hasegawa Y, Shiokawa Y,

16
Downloaded from http://circ.ahajournals.org/ by guest on September 23, 2015
 DOI: 10.1161/CIRCULATIONAHA.115.015735

Furui E, Nakagawara J, Kario K, Nezu T, Minematsu K, Toyoda K. Nationwide survey of neuro-

specialists’ opinions on anticoagulant therapy after intracerebral hemorrhage in patients with
atrial fibrillation. J Neurol Sci. 2012;312:82–85.

25. Singer DE, Chang Y, Fang MC, Borowsky LH, Pomernacki NK, Udaltsova N, Go AS. The
net clinical benefit of warfarin anticoagulation in atrial fibrillation. Ann Intern Med.
2009;151:297–305.

26. Lip GYH, Nieuwlaat R, Pisters R, Lane DA, Crijns HJGM. Refining clinical risk
stratification for predicting stroke and thromboembolism in atrial fibrillation using a novel risk
factor-based approach: the euro heart survey on atrial fibrillation. Chest. 2010;137:263–272.

27. Olesen JB, Torp-Pedersen C, Hansen ML, Lip GYH. The value of the CHA2DS2-VASc
score for refining stroke risk stratification in patients with atrial fibrillation with a CHADS2
score 0-1: a nationwide cohort study. Thromb Haemost. 2012;107:1172–1179.

28. Pisters R, Lane DA, Nieuwlaat R, de Vos CB, Crijns HJGM, Lip GYH. A novel user-
friendly score (HAS
(HAS-BLED)
BLED) to assess 11-year aatrial
year risk of major bleeding in patients with at rial
fibrillation: the Euro Heart Survey. Chest. 2010;138:1093–1100.

29. Poli D, Antonucci E, Dentali F, Erba N, Testa S, Tiraferri E, Palareti G. Recurrence of ICH
after resumption of anticoagulation with VK antagonists: CHIRONE study. Neurology.
2014;82:1020–1026.
20144;8
;82:
2:10
2:1020
10 20–102
20 026.
02

30.
30. Qureshi AI, Tuhrim
Tuhrrim S,
Tuhr S, Broderick
Brod
Br erricck JP,
oder
od J , Batjer
JP Batj
Bat er HHH,
H, HHondo
ondo
on H,, Ha
do H Hanley
Hanl
nley
nl DF.
ey D Spontaneous
F. S pont
po ntan
nt aneo
an eous
eo us
intracerebral
ntrrac hemorrhage.
m rrhaage. N E
a erebral hemo
mo Engl Med.
ngl J Med 2001;344:1450–1460.
ed. 2001
011;3444::14550–
0–14
14660.
14

31. Ap
Apostolakis
Apososto
os S,, L
tollakiss S
to Lane DA,, Gu
ane DA
ane D Y,, B
Guoo Y Buller
ull
lleer H,, Li
er H GYH.
Lipp GY
GYHH. PPerformance
erfform
er ance ooff th
mance tthee
HEMORR(2)HAGES,
HEMO
HE MORR
MO RR(2
RR (2)H
(2 )HAG
)H AGES
AG ES, ATRIA,
ES ATRI
AT A, and
RIA
RI and HAS-BLED
HAS-BLEBLELEDD bleeding
blee
bl eedi
eeding
ding risk-prediction
risk
ri sk-pre
pre
redi
dict
dictio
ctionn scores
io scor
sc es in
ores
or in patients
pati
pa tien
ti ts with
ents
en wit
ithh
it
atrial fibrillation undergoing anticoagulation: the AMADEUS (evaluating the use of SR34006
compared to warfarin or acenocoumarol in patients with atria. J Am Coll Cardiol. 2012;60:861–
867.

32. Morgenstern LB, Hemphill JC, Anderson C, Becker K, Broderick JP, Connolly ES,
Greenberg SM, Huang JN, MacDonald RL, Messé SR, Mitchell PH, Selim M, Tamargo RJ.
Guidelines for the management of spontaneous intracerebral hemorrhage: a guideline for
healthcare professionals from the American Heart Association/American Stroke Association.
Stroke. 2010;41:2108–2129.

33. Connolly SJ, Eikelboom J, Joyner C, Diener H-C, Hart R, Golitsyn S, Flaker G, Avezum A,
Hohnloser SH, Diaz R, Talajic M, Zhu J, Pais P, Budaj A, Parkhomenko A, Jansky P,
Commerford P, Tan RS, Sim K-H, Lewis BS, Van Mieghem W, Lip GYH, Kim JH, Lanas-
Zanetti F, Gonzalez-Hermosillo A, Dans AL, Munawar M, O’Donnell M, Lawrence J, Lewis G,
Afzal R, Yusuf S. Apixaban in patients with atrial fibrillation. N Engl J Med. 2011;364:806–817.

34. Mant J, Hobbs FDR, Fletcher K, Roalfe A, Fitzmaurice D, Lip GYH, Murray E. Warfarin

17
Downloaded from http://circ.ahajournals.org/ by guest on September 23, 2015
 DOI: 10.1161/CIRCULATIONAHA.115.015735

versus aspirin for stroke prevention in an elderly community population with atrial fibrillation
(the Birmingham Atrial Fibrillation Treatment of the Aged Study, BAFTA): a randomised
controlled trial. Lancet. 2007;370:493–503.

35. Gallego P, Roldan V, Marín F, Romera M, Valdés M, Vicente V, Lip GYH. Cessation of oral
anticoagulation in relation to mortality and the risk of thrombotic events in patients with atrial
fibrillation. Thromb Haemost. 2013;110:1189–1198.

36. De Caterina R, Husted S, Wallentin L, Andreotti F, Arnesen H, Bachmann F, Baigent C,

Huber K, Jespersen J, Kristensen SD, Lip GYH, Morais J, Rasmussen LH, Siegbahn A,
Verheugt FWA, Weitz JI. Vitamin K antagonists in heart disease: current status and perspectives
(Section III). Position paper of the ESC Working Group on Thrombosis--Task Force on
Anticoagulants in Heart Disease. Thromb Haemost. 2013;110:1087–1107.

37. Lovelock CE, Cordonnier C, Naka H, Al-Shahi Salman R, Sudlow CLM, Sorimachi T,
Werring DJ, Gregoire SM, Imaizumi T, Lee S-H, Briley D, Rothwell PM. Antithrombotic drug
use, cerebral microbleeds, and intracerebral hemorrhage: a systematic review of published and
2010;41:1222–1228.
unpublished studies. Stroke. 2010;41:1222 1228.

38. Yates PA, Villemagne VL, Ellis KA, Desmond PM, Masters CL, Rowe CC. Cerebral Cere
Ce rebr
re bral
br al
microbleeds: a review of clinical, genetic, and neuroimaging associations. Front Neurol.
2014;4:205.

39.
399. Shen
Sh AY-AY-J,
Y-J, Yao JF, Brar SS, Jorgensen MB, Ch
Y- Racial/ethnic
Chen W. Racial all/e
/ethnic differences in the risk of
intracranial
ntrracranial hemorrhage
hem
mor
o rh agee among
hag ongg patients
amon
amon paati nts with
tieent wiith atrial
atriaal fibrillation.
fibrril
illa
lati
lation
tion. J Am C
on Coll Cardiol.
olll Ca
ard iol. 2007;50:309–
rdiiol 2 07
20 07;5
50:
0:30
309–
30 9
315.
3 5.
31

18
Downloaded from http://circ.ahajournals.org/ by guest on September 23, 2015
 DOI: 10.1161/CIRCULATIONAHA.115.015735

Table 1. Baseline characteristics of the cohort.

Total Intracerebral Subdural Subarachnoid

population bleeding bleeding bleeding
Number of patients 1,752 812 755 185
Age, median (IQR) 78 (71-83) 78 (71-83) 78 (71-83) 75 (70-82)
Male gender No. (%) 1,081 (62) 471 (58) 502 (67) 109 (59)
Prior antithrombotic treatment no. (%)
VKA 1,145(65) 507 (62) 508 (67) 130 (70)
NOAC 28 (2) 11 (1) 12 (2) 5 (3)
VKA and APT 571 (33) 290 (36) 232 (31) 49 (27)
NOAC and APT 8 (<1) 4 (<1) 3 (<1) 1 (<1)
Comorbidity no. (%)
Congestive heart failure 485 (28) 198 (24) 220 (29) 67 (36)
Hypertension 1,112 (63) 494 (61) 490 (65) 128 (69)
Age 65-74 years 498 (28) 221 (27) 212 (28) 65 (35)
Age >75 years 1,068 (61) 500 (62) 470 (62) 988 (5
((53)
3)
Diabetes mellitus 323 (18) 134 (17) 149 (20) 440
0 (2
(22)
2)
Prior thromboembolism 656 (37) 373 (46) 226 (30) 57 (31)
Vascular disease 358 (20) 152 (19) 160 (21) 46 (25)
Concomitant
Conccom
omit itan
itantt medication
an medi
me d cation no. (%)
Beta-blockers
Beta
Be ta-block
ta kers 1,035 (59) 460 (57) 454 (60) 121 (65)
Calcium
Caalcium channel
chan nne
nel bl
blockers
loc
o ke
kers
rs 6613
133 ((35)
355) 2711 (3
27 (33)
3
3) 2280
80 (37
80 (37)
377) 62 ((34)
34))
34
Renin-angiotensin
Renin-angioten
Re nsiin system
syysttem 9914
14 (52)
(522) 425
425 (52)
(522) 393
3993 (52)
(5
52)) 966 ((52)
52)
inhibitors
inhi
in hibi
hi b tors
rss (ACEi/ARBs)
(AC
ACEii/ARB RBs)
RB
Loop diuretics
diuretics
iu 7033 (4
70 (40) 2299
99 (37
99 (37)
3 )
37 315 ((42)
42) 889
9 (48)
8
Statin 688 (39) 285 (35) 315 (42) 88 (48)
NSAID 381 (22) 172 (21) 172 (23) 37 (20)
Digoxin 853 (49) 376 (46) 386 (51) 91 (49)
Amiodarone 120 ( 7) 49 ( 6) 58 ( 8) 13 ( 7)
CHA2DS2-VASc no. (%)
0 31 (2) 14 (2) 13 (2) 4 (2)
1 123 (7) 44 (5) 62 (8) 17 (9)
2-9 1,598 (91) 754 (93) 680 (90) 164 (89)
HAS-BLED no. (%)
0-2 363 (21) 155 (19) 170 (23) 38 (21)
•3 1,389 (79) 657 (81) 585 (77) 147 (79)
ICH: Intracranial haemorrhage. IQR: Interquartile range. VKA: Vitamin K-antagonist. APT: Antiplatelet therapy.
NOAC: Non-vitamin K oral anticoagulant. NSAID: Non-steroidal anti-inflammatory drug. ACEi/ARBs:
Angiotensin-Converting Enzyme inhibitors/Angiotensin II Receptor Antagonists.

19
Downloaded from http://circ.ahajournals.org/ by guest on September 23, 2015
 DOI: 10.1161/CIRCULATIONAHA.115.015735

Table 2. Event rates of various outcomes according to stratification on treatment regimen using 1 year of follow-up.

Outcome No antithrombotic treatment OAC treatment Antiplatelet therapy

Ischemic stroke/SE/ and all-cause mortality
Events 179 43 83
Person-time (100 years) 655 316 335
Event rate (95% CI) 27.3 (23.6 - 31.6) 13.6 (10.1 - 18.3) 25.7 (20.7 - 31.9)
Ischemic stroke/SE
Events 69 17 34
Person-time
on-time (100 years) 666 322 329
329
Event
nt rate (95% CI) 10.4 (8.2 - 13.1) 5.3 (3.3 - 8.5) 10
10.3
0.3
3 ((7.4
7.44 - 14
7. 14.4
14.4)
.4))
.4
All-cause
ause mortality
nts
Events 130 32 69
Person-time
on-time (100 years) 682 330 353
Event
nt ra
rate
ate ((95%
95%
95 % CI
CI)) 19.1 (16.0 - 22.6) 9.7 (6.9 - 13.7) 19.5 (15.4 - 24.7)
rrren
e t IC
Recurrent IICHH
Events
nts 57
5 25 18
Person-time
o -t
on - ime (100 years) 662
6 3133 339
3 9
33
Event
nt rate
at (95% CI)) 8.6 (6.6
( .66 - 11.2))
(6 8.00 (5.4
( .4 - 11.8)
(5 5.
5.3
.3 (3
(3.3
3.3 - 88.4)
.4
4)
Majorr ex
extr
extracranial
xtr
trac
acra
ac r nial
ra all bbleeding
leeding
le n
nts
Events 10 5 9
on-titime (1
Person-time (100
00 years)) 6777
67 3299
32 3499
34
Event rate (95% CI) 1.5 (0.8 - 2.7) 1.5 (0.6 - 3.7) 2.6 (1.3 - 5.0)

20
Downloaded from http://circ.ahajournals.org/ by guest on September 23, 2015
 DOI: 10.1161/CIRCULATIONAHA.115.015735

Figure Legends:

Figure 1. Flowchart of the study design and population.

Figure 2. Forest plot of adjusted hazard ratios on treatment regimen of various outcomes using

one-year follow-up (adjusted for sex, age, year of inclusion, time since last claimed OAC

prescription, and occurrences of ischemic stroke/SE and recurrent ICH events during the

quarantine period).

Figure 3. Five years Kaplan-Meier survival curve for restarting OAC treatment, re
receiving
rece
ceiv
ce ivin
iv ingg
in

antiplatelet therapy, and for not receiving antithrombotic treatment using a landmark at 6 weeks

(relative
relat
ativ
ativee to ddischarge
iv i ch
is treatment
harrge from hospital) for treatmen regimens
nt re stratification.
egimens stratificatio
on..

Figure
Fi
igu
gure Sensitivity
r 4. Sen i ity aanalysis
ensiti
en tivi investigating
nalysiis in
inve
est
stigaatiing th outcomes
thee ou
utccomes ooff isch
mes ischemic
c em
ch emic stroke/SE
icc st
trooke/
e/SE
e/ nd all-cause
SE and all--caausee

mortality
mort
mo rtal
rt alit
al ity an
it recurrent
andd re
recu
curr
rren
rr ICH,
entt IC
en ICHH, sstratified
traati
tr tifi
fied
fi ed aaccording
ccor
cc ordi
or ding
di ng tto treatment
o tr
trea
eatm
ea tmen
tmentt re
en regimen
regi
gime
gi menn an
me andd to aapplied
ppli
pp ed qquarantine
lied
li uaara
uara
rant
ntin
nt inee
in

period.

21
Downloaded from http://circ.ahajournals.org/ by guest on September 23, 2015
 Downloaded from http://circ.ahajournals.org/ by guest on September 23, 2015
Figure 1
 Downloaded from http://circ.ahajournals.org/ by guest on September 23, 2015
Figure 2
 Downloaded from http://circ.ahajournals.org/ by guest on September 23, 2015
Figure 3
 Downloaded from http://circ.ahajournals.org/ by guest on September 23, 2015
Figure 4
SUPPLEMENTAL MATERIAL

Supplementary Table 1: Definitions of outcomes and medication

International Classification of Anatomical Therapeutic Chemical

Diseases 10th revision (ICD-10) (ATC) code

code

Condition

Congestive heart failure I11.0; I13.0; I13.2; I42.0; I50 CO3C

Left ventricular dysfunction I50.1; I50.9

Hypertension See specified definition*

Diabetes mellitus E10.0; E10.1; E10.9; E11.0; E11.1; A10

E11.9

Ischemic stroke I63; I64

Systemic embolism I74

Transient ischemic disease G45

Aortic plaque I70.0

Peripheral arterial disease I70.2-I70.9; I71; I73.9; I74

Myocardial infarction I21-I23

Abnormal renal function I12; I13; N00-N05; N07; N11;

N14; N17-N19; Q61

 Abnormal hepatic function B15.0; B16.0; B16.2; B19.0;

K70.4; K72; K76.6; I85

Prior Bleeding I60-I62; D62; J94.2; H11.3; H35.6;

H43.1; N02; N95; R04; R31; R58;

K25.0; K26.0; K27.0; K28.0;

K29.0; S06.3C; S06.4; S06.5;

S06.6

Alcohol intake E22.4; E52.9A; F10; G31.2;

G62.1; G72.1; I42.6; K29.2; K70;

K86.0; L27.8A; O35.4M; T51;

Z71.4; Z72.1

Atrial fibrillation I48

Major bleeding D62 J942 H113 H356 H431 N02

N95 R04 R31 R58

Intracranial bleeding I60 I61 I62

Traumatic intracranial bleeding S063C S064 S065 S066

Retinal bleeding H356

Sequelae of cerebrovascular disease I690; I691; I692

Medication

Dabigatran B01AE07

Rivaroxaban B01AE07
 Apixaban B01AF02

Coumarin derivatives B01AA

Aspirin B01AC06

Thienopyridines B01AC04; B01AC24; B01AC22

Beta-blockers C07

Calcium channel blockers C07F; C08; C09BB; C09DB

Renin-angiotensin system

inhibitors (ACEi/ARBs) C09

Loop diuretics C03C

Statin C10

Digoxin C01AA05

Amiodarone C01BD01

Non-steroidal anti-inflammatory drugs M01A

* We identified subjects with hypertension from combination treatment with at least two of the following classes of

antihypertensive Drugs:

I. Alpha adrenergic blockers (C02A, C02B, C02C)

II. Non-loop diuretics (C02DA, C02L, C03A, C03B, C03D, C03E, C03X, C07C, C07D, C08G, C09BA, C09DA,

C09XA52)

III. Vasodilators (C02DB, C02DD, C02DG, C04, C05)

IV. Beta blockers (C07)

V. Calcium channel blockers (C07F, C08, C09BB, C09DB)

VI. Renin-angiotensin system inhibitors (C09).
Supplementary Table 2: Event rates and adjusted hazard ratios of various outcomes according to stratification on treatment regimen using
5 years of follow-up.

Outcome No OAC therapy OAC therapy Antiplatelet therapy

Ischemic stroke/SE and
all-cause mortality
Events 326 106 240
Person-time (100 years) 1392 996 1204
Event rate (95% CI) 23.4 (21.0 - 26.1) 10.6 (8.8 - 12.9) 19.9 (17.6 - 22.6)
Adjusted hazard ratio (95%CI) Reference 0.54 (0.43 – 0.67) 0.85 (0.81 – 1.00)
Ischemic stroke/SE
Events 108 35 76
Person-time (100 years) 1413 1014 1217
Event rate (95% CI) 7.6 (6.3 - 9.2) 3.5 (2.5 - 4.8) 6.2 (5.0 - 7.8)
Adjusted hazard ratio (95%CI) Reference 0.55 (0.37 – 0.81) 0.87 (0.65 – 1.17)
All-cause mortality
Events 266 91 227
Person-time (100 years) 1470 1072 1420
Event rate (95% CI) 18.1 (16.0 - 20.4) 8.5 (6.9 - 10.4) 16.0 (14.0 - 18.2)
Adjusted hazard ratio (95%CI) Reference 0.56 (0.44 – 0.71) 0.87 (0.72 – 1.04)
Recurrent ICH
Events 72 36 34
Person-time (100 years) 1408 1012 1315
Event rate (95% CI) 5.1 (4.1 - 6.4) 3.6 (2.6 - 4.9) 2.6 (1.8 - 3.6)
Adjusted hazard ratio (95%CI) Reference 0.79 (0.52 – 1.19) 0.56 (0.37 – 0.84)
Major extra cranial bleeding
Events 20 14 26
Person-time (100 years) 1449 1052 1382
Event rate (95% CI) 1.4 (0.9 - 2.1) 1.3 (0.8 -2.2) 1.9 (1.3 - 2.8)
Adjusted hazard ratio (95%CI) Reference 0.90 (0.44 – 1.81) 1.31 (0.73 – 1.76)
Supplementary Figure S1: Five years Kaplan-Meier survival curve for restarting OAC treatment, receiving antiplatelet therapy, and for
not receiving antithrombotic treatment using a landmark at 180 days (relative to discharge from hospital) for treatment regimens
stratification. A total of 505 patients were allocated to ‘no antithrombotic treatment; 509 to OAC treatment; 527 patients to the antiplatelet
therapy group.
 Restarting Anticoagulant Treatment After Intracranial Haemorrhage in Patients With Atrial
Fibrillation and the Impact on Recurrent Stroke, Mortality and Bleeding: A Nationwide Cohort
Study
Peter Brønnum Nielsen, Torben Bjerregaard Larsen, Flemming Skjøth, Anders Gorst-Rasmussen,
Lars Hvilsted Rasmussen and Gregory Y.H. Lip

Circulation. published online June 9, 2015;

Circulation is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231
Copyright © 2015 American Heart Association, Inc. All rights reserved.
Print ISSN: 0009-7322. Online ISSN: 1524-4539

The online version of this article, along with updated information and services, is located on the
World Wide Web at:
http://circ.ahajournals.org/content/early/2015/06/09/CIRCULATIONAHA.115.015735

Data Supplement (unedited) at:

http://circ.ahajournals.org/content/suppl/2015/06/09/CIRCULATIONAHA.115.015735.DC1.html

Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally published in
Circulation can be obtained via RightsLink, a service of the Copyright Clearance Center, not the Editorial Office.
Once the online version of the published article for which permission is being requested is located, click Request
Permissions in the middle column of the Web page under Services. Further information about this process is
available in the Permissions and Rights Question and Answer document.

Reprints: Information about reprints can be found online at:

http://www.lww.com/reprints

Subscriptions: Information about subscribing to Circulation is online at:

http://circ.ahajournals.org//subscriptions/

Downloaded from http://circ.ahajournals.org/ by guest on September 23, 2015