/clinical/,/clinical/cckm-tools/,/clinical/cckm-tools/content/,/clinical/cckm-tools/content/questionnaires/,/clinical/cckm-tools/content/questionnaires/related/,

/clinical/cckm-tools/content/questionnaires/related/name-97240-en.cckm

201606168

page

100

UWHC,UWMF,

Clinical Hub,UW Health Clinical Tool Search,UW Health Clinical Tool Search,Questionnaires,Related

A Novel User-Friendly Score HAS-BLED To Assess 1-Year Risk of Major Bleeding in Patients with Atrial Fibrillation

A Novel User-Friendly Score HAS-BLED To Assess 1-Year Risk of Major Bleeding in Patients with Atrial Fibrillation - Clinical Hub, UW Health Clinical Tool Search, UW Health Clinical Tool Search, Questionnaires, Related


CHEST Original Research
ANTITHROMBOTIC THERAPY
www.chestpubs.org CHEST / 138 / 5 / NOVEMBER, 2010 1093
A
trial fi brillation (AF) is associated with a well-
known increase in ischemic stroke risk,
1
which is
further increased by individual conditions such as
heart failure, hypertension, diabetes, and prior throm-
boembolism.
2
Oral anticoagulation (OAC) dramati-
cally reduces this risk
3
and is therefore recommended
in patients with AF at moderate-high risk of stroke
and thromboembolism.
4
The increasing incidence
and prevalence of AF increases the likelihood of OAC
use in the AF population, which is usually elderly, and
comorbidities commonly coexist.
5-7
Indeed, clinical
decision making about whether OAC is justifi ed based
on stroke risk is supported by various practical stroke
risk classifi cation schema that incorporate known
clinical risk factors.
4,8

However, stroke risk is also closely related to
bleeding risk,
9
and OAC prescription needs to bal-
ance the benefi t from stroke prevention against the
risk of bleeding. Thus, there is often suboptimal
Objective: Despite extensive use of oral anticoagulation (OAC) in patients with atrial fi brillation
(AF) and the increased bleeding risk associated with such OAC use, no handy quantifi cation tool
for assessing this risk exists. We aimed to develop a practical risk score to estimate the 1-year risk
for major bleeding (intracranial, hospitalization, hemoglobin decrease . 2 g/L, and/or transfu-
sion) in a cohort of real-world patients with AF.
Methods: Based on 3,978 patients in the Euro Heart Survey on AF with complete follow-up, all
univariate bleeding risk factors in this cohort were used in a multivariate analysis along with his-
torical bleeding risk factors. A new bleeding risk score termed HAS-BLED (Hypertension,
Abnormal renal/liver function, Stroke, Bleeding history or predisposition, Labile international
normalized ratio, Elderly ( . 65 years), Drugs/alcohol concomitantly) was calculated, incorpo-
rating risk factors from the derivation cohort.
Results: Fifty-three (1.5%) major bleeds occurred during 1-year follow-up. The annual bleeding
rate increased with increasing risk factors. The predictive accuracy in the overall population
using signifi cant risk factors in the derivation cohort (C statistic 0.72) was consistent when applied
in several subgroups. Application of the new bleeding risk score (HAS-BLED) gave similar
C statistics except where patients were receiving antiplatelet agents alone or no antithrombotic
therapy, with C statistics of 0.91 and 0.85, respectively.
Conclusion: This simple, novel bleeding risk score (HAS-BLED) provides a practical tool to assess
the individual bleeding risk of real-world patients with AF, potentially supporting clinical decision
making regarding antithrombotic therapy in patients with AF. CHEST 2010; 138(5):1093–1100
Abbreviations: AF 5 atrial fi brillation; CHADS
2
5 congestive heart failure, hypertension, age . 75 years, diabetes
mellitus, previous stroke/transient ischemic attack (doubled); HAS-BLED 5 Hypertension, Abnormal renal/liver function,
Stroke, Bleeding history or predisposition, Labile international normalized ratio, Elderly ( . 65 years), Drugs/alcohol
concomitantly; HEMOR
2
RHAGES 5 Hepatic or renal disease, Ethanol abuse, Malignancy, Older age ( . 75 years),
Rebleeding, Reduced platelet count or function, Hypertension (uncontrolled), Anemia, Genetic factors, Excessive fall
risk, and Stroke; INR 5 international normalized ratio; OAC 5 oral anticoagulation
A Novel User-Friendly Score (HAS-BLED)
To Assess 1-Year Risk of Major Bleeding
in Patients With Atrial Fibrillation
The Euro Heart Survey
Ron Pisters , MD ; Deirdre A. Lane , PhD ; Robby Nieuwlaat , PhD ; Cees B. de Vos , MD ;
Harry J. G. M. Crijns , MD ; and Gregory Y. H. Lip , MD
For editorial comment see page 1032
Downloaded From: http://journal.publications.chestnet.org/ by a University of Wisconsin User on 11/02/2015

1094 Original Research
stroke. Hemorrhagic stroke was defi ned as a focal neurologic defi -
cit of sudden onset, diagnosed by a neurologist, lasting . 24 h and
caused by bleeding. Presence of chronic dialysis, renal transplan-
tation, or serum creatinine � 200 m mol/L was classifi ed as abnor-
mal kidney function. Abnormal liver function was defi ned as chronic
hepatic disease (eg, cirrhosis) or biochemical evidence of signifi cant
hepatic derangement (eg, bilirubin . 2 3 upper limit of normal,
in association with aspartate aminotransferase/alanine amin-
otransferase/alkaline phosphatase . 3 3 upper limit normal, and so
forth). Finally, valvular heart disease was defi ned as the presence
of any regurgitation or gradient over a valve with hemodynamic
signifi cance and/or related symptoms.
Statistical Analysis and Design of a New Bleeding Risk Score
Data analysis was performed with SPSS, version 16.0 (SPSS
Inc., Chicago, IL). The presence of any differences between the
groups with and without major bleeding during 1-year follow-up
was tested by Fisher exact test for categorical variables and by
independent samples t test for continuous variables.
All potential bleeding risk factors identifi ed from the uni-
variate analyses of the derivation cohort with a P value , .10
(age . 65 years, female sex, diabetes mellitus, heart failure, COPD,
valvular heart disease, kidney failure, prior major bleeding epi-
sode, and clopidogrel use), were used in the multivariate logis-
tic regression analyses along with more historical bleeding risk
factors: OAC, alcohol use, and hypertension. We disregarded
thyroid disease ( P 5 .039) because of diffi culties with interpre-
tation (only n 5 10 bleeding events), and this had not been iden-
tifi ed as a bleeding risk in prior systematic reviews.
11-13
Given
the persistent nature of the evidence of OAC, hypertension,
age . 65 years, renal failure, alcohol abuse, and prior major
bleed as bleeding risk factors, these variables were kept in the
model at all times. The other, less strongly linked variables were
removed stepwise from the model when the P value was . .10.
Variables with P , .05 in the fi nal model were considered to be
signifi cant contributors and were checked for interaction effects,
which did not exist.
In recognition of the limited number of major bleeds and rela-
tively short follow-up period in the Euro Heart Survey on AF, we
also added consistent risk factors for major bleeding (stroke, alco-
hol use, systolic BP . 160 mm Hg, and so forth) identifi ed in
recent systematic reviews
11,13
to use the data from the derivation
cohort to test a new bleeding risk score in a fi nal statistical model,
accepting the lack of statistical signifi cance of some variables in
our derivation cohort. For each of the variables in the fi nal model
the regression coeffi cient, net OR, and its 95% CI and P value
are reported ( Table 1 ). We then calculated the C statistic as a
measure of the predictive accuracy of the model incorporating
bleeding risk factors from the derivation cohort (that is, prior major
bleeding, age . 65 years, clopidogrel use, and kidney failure),
where, based on their respective multivariate regression coeffi -
cients, two points were awarded for prior major bleeding and one
point for each of the other bleeding risk factors. In addition, we report
the C statistics in a subgroup analysis of individuals discharged
with OAC monotherapy, OAC combined with an antiplate let
drug, an antiplatelet drug alone, or no antithrombotic therapy.
A new bleeding risk score (termed HAS-BLED) was calcu-
lated, incorporating risk factors from the derivation cohort as well
as signifi cant risk factors for major bleeding found in the literature
from systematic reviews.
11,13
HAS-BLED is an acronym for hyper-
tension (uncontrolled, . 160 mm Hg systolic), abnormal renal /
liver function (one point for presence of renal or liver impairment,
maximum two points), stroke (previous history, particularly lacunar),
bleeding history or predisposition (anemia), labile international
normalized ratio (INR) (ie, therapeutic time in range , 60%),
elderly ( . 65 years), drugs/alcohol concomitantly (antiplatelet
implementation of thromboprophylaxis among patients
with AF,
9,10
which may be partly due to the lack of a vali-
dated bleeding risk-stratifi cation schema that is user-
friendly.
11-13
This is further refl ected by the absence
of recommendations on bleeding risk assessment in
current antithrombotic guidelines for AF manage-
ment.
4,12
The available schemas estimating the risk of
bleeding associated with use of OAC either do not
focus on patients with AF in particular,
14,15
address a
(very) specifi c subgroup among patients with AF,
16
o r
incorporate routinely unavailable risk factors that
also overlap signifi cantly with stroke risk factors.
17

Furthermore, all published schema are based on
historical cohorts of patients and consequently may
not refl ect advancements in medical care over time
(for example, OAC monitoring) and treatment of
underlying heart disease.
14-17
Our aim was to develop
a practical risk score to estimate the 1-year risk of
major bleeding (intracranial, hospitalization, hemo-
globin decrease . 2 g/L, and/or transfusion) in a con-
temporary, real-world cohort of patients with AF.
Materials and Methods
We used the large population database of the prospective Euro
Heart Survey on AF, with data collected between 2003 and 2004.
A detailed study outline of the Euro Heart Survey on AF at base-
line
5
and follow-up assessment
18
has been previously described.
In summary, 5,333 ambulatory and hospitalized patients with AF
from 182 university, nonuniversity, and specialized hospitals
among 35 member countries of the European Society of Cardiol-
ogy were enrolled. Patients had to be � 18 years of age and have
an ECG or Holter-proven diagnosis of AF during the qualifying
admission or in the preceding year. A 1-year follow-up assessment
was performed to determine survival and major adverse cardio-
vascular events, such as major bleeding. Medical records and/or
medical information systems were used to populate the data set.
We defi ned major bleeding as any bleeding requiring hospital-
ization and/or causing a decrease in hemoglobin level of . 2 g/L
and/or requiring blood transfusion that was not a hemorrhagic
Manuscript received January 17, 2010; revision accepted March 1,
2010.
Affi liations: From the Department of Cardiology (Drs Pisters,
Nieuwlaat, de Vos, and Crijns), Maastricht University Medical
Centre, Maastricht, The Netherlands; and the University of Bir-
mingham Centre for Cardiovascular Sciences (Drs Lane and Lip),
City Hospital, Birmingham, England.
Funding/Support: The Euro Heart Survey has as industry spon-
sors the main sponsor, AstraZeneca; major sponsor, Sanofi -Aventis;
and sponsor, Eucomed. Funding institutions are the Austrian Heart
Foundation, Austrian Society of Cardiology, French Federa-
tion of Cardiology, Hellenic Cardiological Society, Netherlands
Heart Foundation, Portuguese Society of Cardiology, Spanish
Cardiac Society, Swedish Heart and Lung Foundation, and indi-
vidual centers.
Correspondence to: Gregory Y. H. Lip, MD, University of
Birmingham Centre for Cardiovascular Sciences, City Hospital,
Birmingham B18 7QH, England; e-mail: g.y.h.lip@bham.ac.uk
© 2010 American College of Chest Physicians. Reproduction
of this article is prohibited without written permission from the
American College of Chest Physicians ( http://www.chestpubs.org/
site/misc/reprints.xhtml ).
DOI: 10.1378/chest.10-0134
Downloaded From: http://journal.publications.chestnet.org/ by a University of Wisconsin User on 11/02/2015

www.chestpubs.org CHEST / 138 / 5 / NOVEMBER, 2010 1095
hemorrhage. The baseline demographic and clinical
characteristics of the 3,456 patients are presented in
Table 3 .
Of all discharged patients, 2,242 (64.8%) were on
OAC (286 [12.8%] of whom also received aspirin
and/or clopidogrel), 828 (24.0%) received antiplate-
let therapy alone (aspirin and/or clopidogrel), and
352 (10.2%) received no antithrombotic therapy.
The distribution of the risk factors for major bleeding
within 1 year among the different antithrombotic
treatment regimens in the derivation cohort is depicted
agents, nonsteroidal antiinfl ammatory drugs; one point for drugs
plus one point for alcohol excess, maximum two points) as shown
in Table 2 .
In order to compare the predictive accuracy of our novel
bleed risk score (HAS-BLED) with the previously proposed
HEMOR
2
RHAGES (Hepatic or renal disease, Ethanol abuse,
Malignancy, Older age [ . 75 years], Rebleeding, Reduced
platelet count or function, Hypertension [uncontrolled], Ane-
mia, Genetic factors, Excessive fall risk, and Stroke) scheme,
patients were classifi ed accordingly
17
; however, we considered
“uncontrolled hypertension” to be . 160 mm Hg systolic, a “his-
tory of malignancy” to be similar to “current malignancy,” and we
classifi ed � 8 units alcoholic consumption per week as “ethanol
abuse.” Relevant genetic and laboratory data (required for calcu-
lation of the HEMOR
2
RHAGES schema), apart from serum crea-
tinine, were not available for the Euro Heart Survey on AF cohort.
Results
Of the 5,272 patients with AF in the Euro Heart
Survey of AF who were discharged alive,
5
3,456 (66%)
patients without mitral valve stenosis or valvular
surgery had 1-year follow-up status regarding major
bleeding. The overall mean (SD) age was 66.8 (12.8)
years, and the majority were men (59%). Fifty-three
(1.5%) patients experienced a major bleed during the
fi rst year, including nine (17%) cases of intracerebral
Table 1— Clinical Risk Factors for Major Bleeding Within 1 Year in Patients With Atrial Fibrillation Enrolled in the
Euro Heart Survey
Risk Factor
MV Regression
Coeffi cient
Event Rate With
Risk Factor
Event Rate Without
Risk Factor
Univariate
P Value OR (95% CI) MV P Value
Systolic BP . 160 mm Hg - 0.52 4 (1.0) 11 (1.4) .515 0.60 (0.21-1.72) .337
Kidney failure 1.05 10 (5.4) 43 (1.3) , .001 2.86 (1.33-6.18) .007
Stroke - 0.44 4 (2.1) 48 (1.5) .532 0.94 (0.32-2.86) .940
Prior major bleeding 2.02 9 (14.8) 44 (1.3) , .001 7.51 (3.00-18.78) , .001
Age . 65 y 0.98 42 (2.3) 11 (0.7) , .001 2.66 (1.33-5.32) .007
Antiplatelet agent
a
- 0.22 5 (3.4) 46 (1.4) .066 0.81 (0.43-1.51) .504
Alcohol use
b
- 16.80 10 (1.6) 31 (1.9) 1.000 0.00 (0.00) .996
MV 5 multivariate.

a
Aspirin or clopidogrel.

b
. 8 U/wk.
Table 2— Clinical Characteristics Composing the
HAS-BLED Bleeding Risk Score
Letter Clinical Characteristic
a
Points Awarded
H Hypertension 1
A Abnormal renal and liver function
(1 point each)
1 or 2
S Stroke 1
B Bleeding 1
L Labile INRs 1
E Elderly 1
D Drugs or alcohol (1 point each) 1 or 2
HAS-BLED 5 Hypertension, Abnormal renal/liver function, Stroke,
Bleeding history or predisposition, Labile international normalized
ratio, Elderly ( . 65 years), Drugs/alcohol concomitantly; INR 5 inter-
national normalized ratio.

a
See “Materials and Methods” for definitions of the clinical
characteristics.
Table 3— Baseline Characteristics of Patients With
Atrial Fibrillation From the Euro Heart Survey With a
Known Follow-up Status Regarding Major Bleeding
Characteristic Bleed (n 5 53)
No Bleed
(n 5 3,910) P Value
Age, mean (SD), y 73 (10) 66 (13) , .001
Age . 65 y 42 (79) 1,748 (51) , .001
Female sex 28 (53) 1,329 (39) .046
BMI, mean (SD), kg/m
2
27 (5) 28 (8) .298
Atrial fi brillation type .827
First detected 9 (17) 628 (19)
Paroxysmal 15 (28) 1,029 (31)
Persistent/permanent 29 (56) 1,688 (51)
Medical history
Current smoker 6 (11) 450 (13) .716
Hypertension 39 (74) 2,228 (65) .245
Diabetes mellitus 12 (23) 610 (18) .367
Coronary artery disease 15 (29) 1,182 (35) .462
Heart failure 24 (45) 994 (24) .014
Valvular heart disease 18 (34) 607 (18) .006
COPD 12 (23) 427 (13) .037
Thyroid disease 10 (20) 326 (10) .059
Stroke/transient ischemic
attack
6 (12) 355 (11) .818
CHADS
2
score, mean (SD) 2.07 (1.16) 1.60 (1.27) .008
Bleeding risk factors
Prior major bleed 9 (17) 52 (2) , .001
Systolic BP, mean (SD) 137(20) 136 (22) .856
Malignancy 4 (8) 183 (5) .529
Renal failure 10 (19) 174 (5) , .001
Alcohol use � 8 U/wk 0 (0) 170 (5) .111
Values presented are No. (%) unless otherwise indicated . CHADS
2
5
congestive heart failure, hypertension, age . 75, diabetes mellitus,
and previous stroke/transient ischemic attack (doubled).
Downloaded From: http://journal.publications.chestnet.org/ by a University of Wisconsin User on 11/02/2015

1096 Original Research
21 patients with a CHADS
2
score � 1 discharged
without OAC who suffered a stroke within 1 year of
follow-up, only one had a HAS-BLED score outweigh-
ing the individual stroke risk. Of the patients with
CHADS
2
score � 1 who were discharged without
OAC with a higher HAS-BLED bleeding risk score,
all three patients suffered a major bleed.
Discussion
Using a derivation cohort based on the large, real-
world population of the Euro Heart Survey on AF, we
identifi ed four independent risk factors of major bleed-
ing within 1 year (prior major bleeding, age . 65 years,
clopidogrel use, and kidney failure). Incorporating
these risk factors with other established risk factors
from systematic reviews and multivariate analyses,
11,13,19

we developed and tested a novel, user-friendly bleed-
ing risk score, HAS-BLED, which demonstrated a
good predictive accuracy in the overall Euro Heart
survey cohort (C statistic 0.72) but performed par-
ticularly well in predicting bleeding risk where anti-
platelet therapy was used alone (C statistic 0.91), or
no antithrombotic therapy at all (C statistic 0.85).
Assessment of both stroke and bleed risk using the
CHADS
2
and HAS-BLED schemas, respectively, in
the Euro Heart Survey on AF population would have
resulted in withholding OAC therapy in 12% of the
patients who suffered a major bleed within 1 year and
the initiation of OAC in 95% of the patients at high
risk for stroke who were discharged without OAC
and had suffered a stroke within 1 year.
With the previously published HEMOR
2
RHAGES
schema
17
and others,
11,19
the concept of a risk score
for major bleeding in patients with AF is not new.
However, our novel, proposed HAS-BLED score has
several key advantages over the previously mentioned
bleeding risk stratifi cation method. First, the shorter
acronym means that physicians have fewer risk fac-
tors to memorize when using the HAS-BLED score,
thereby increasing the user-friendliness and sub-
sequent clinical application. Furthermore, in con-
trast to certain risk factors incorporated into the
HEMOR
2
RHAGES score that require laboratory
parameters or even genetic testing,
17
all risk factors of
the HAS-BLED score are either readily available
from the clinical medical history or routinely tested
in (new) patients with AF. This characteristic strongly
supports its use in all health-care settings and is
another signifi cant contributor to its superior user-
friendliness. Because less does not necessarily mean
more, it is important to note that the predictive accu-
racy of the HAS-BLED score is broadly similar when
compared with the HEMOR
2
RHAGES model in
the overall population (C statistic of 0.72 vs 0.66,
in Table 4 . The risk of major bleeding within 1 year in
patients with AF in the Euro Heart Survey determined
by the novel bleeding risk score, HAS-BLED, is shown
in Table 5 . The annual bleeding rate increased with
the addition of each risk factor from the derivation
cohort ( Table 5 ).
The corresponding unadjusted bleeding rates in
patients with OAC, antiplatelet therapy alone, or
no antithrombotic treatment were 1.75, 0.97, and
1.42 bleeds per 100 patient-years, respectively.
The predictive accuracy in the overall population
using signifi cant risk factors in the derivation cohort
(C statistic 0.72) was consistent when applied in
several subgroups, as shown in Table 6 . Application
of the new bleeding risk score (HAS-BLED) gave
similar C statistics to that derived in the derivation
cohort overall (0.72), or when patients were estab-
lished on OAC at baseline (0.69), or when patients were
on OAC plus antiplatelet therapy at baseline (0.78).
HAS-BLED substantially improved the predictive
accuracy of bleeding risk when patients with AF
were receiving antiplatelet therapy alone or in those
who were not on antithrombotic therapy at all (with
C statistics of 0.91 and 0.85, respectively). The
HEMOR
2
RHAGES bleeding scheme had a lower pre-
dictive accuracy compared with the new HAS-BLED
score, overall or in relation to antithrombotic therapy
use, except in the OAC plus antiplatelet therapy sub-
group ( Table 6 ).
Of all the 33 bleeding events in patients discharged
with OAC because of a CHADS
2
(congestive heart
failure, hypertension, age . 75 years, diabetes melli-
tus, and previous stroke/transient ischemic attack
[doubled]) score � 1, four (12%) patients had a
HAS-BLED-based bleeding risk that outweighed their
individual stroke risk. Conversely, of all 1,580 patients
discharged with OAC because of a CHADS
2
score � 1
who did not suffer a major bleed within 1 year, only
34 (2.2%) had a HAS-BLED-based bleeding risk
that outweighed their individual stroke risk. Of all
Table 4— Clinical Risk Factors for Major Bleeding
Within 1 Year in Patients With Atrial Fibrillation of
the Euro Heart Survey According to Antithrombotic
Treatment at Discharge
Bleeding Risk Factor
Oral
Anticoagulation
(n 5 2,115)
Aspirin/
Clopidogrel
(n 5 828)
Neither
(n 5 352)
Systolic BP . 160 mm Hg 10.0 13.4 13.7
Kidney failure 5.3 4.8 6.8
Stroke 5.9 5.0 3.2
Prior major bleeding 1.7 1.6 3.2
Age . 65 y 53.3 52.8 42.6
Antiplatelet agent 12.1 100 0.0
Alcohol use
a
5.7 4.8 5.0
Data shown as percentages.

a
� 8 U/wk.
Downloaded From: http://journal.publications.chestnet.org/ by a University of Wisconsin User on 11/02/2015

www.chestpubs.org CHEST / 138 / 5 / NOVEMBER, 2010 1097
of 0.85 vs 0.81, respectively), the HAS-BLED score
is simpler. This score would be particularly useful in
everyday clinical practice, when making decisions on
whether OAC can be initiated in a patient newly diag-
nosed with AF who is not taking any antithrombotic
therapy,
12
or when antiplatelet therapy (or Nonsteroi-
dal antiinfl ammatory drugs) use is being considered,
for example, in the setting of coronary artery disease.
20

As mentioned previously, balancing the individual
risk of bleeding and stroke is diffi cult
21
but of the utmost
importance to maximize appropriate antithrom-
botic therapy and minimize adverse events in patients
with AF, resulting in a net clinical benefi t for the
treated patient. In daily clinical practice, the CHADS
2

index
8
is a widely used tool to stratify stroke risk in
patients with AF. For now, the HEMOR
2
RHAGES
score is the only suitable counterpart available to assess
the risk of bleeding in AF patients.
17
Closer examina-
tion of the risk factors composing the CHADS
2
and
HEMOR
2
RHAGES schemas reveals an extensive
overlap between risk factors for bleeding and stroke,
which has obvious drawbacks.
8,17
Indeed, the patients
at highest stroke and thromboembolic risk are par-
adoxically more likely to sustain bleeding complica-
tions. This may lead to confusion when trying to decide
on the most appropriate antithrombotic regimen to
balance the risks of bleeding against the risk of stroke,
thereby limiting the applicability of such schemas.
The trade-off in terms of the benefi ts and risks
of OAC using the CHADS
2
index and HAS-BLED
score demonstrates that in the vast majority of patients
with AF who require OAC (CHADS
2
index � 2) the
risk of bleeding outweighs the potential benefi t of
OAC if the HAS-BLED bleed score exceeds the indi-
vidual CHADS
2
index. In the case of a CHADS
2
score
of 1, the HAS-BLED score must exceed 2 for the poten-
tial harm caused by OAC to offset its benefi cial effect
respectively).
17
However, the HAS-BLED score was
particularly useful when antiplatelet therapy was used
alone or no antithrombotic therapy was used at all
(C statistics of 0.91 and 0.85, respectively). Although
the HAS-BLED score and the HEMOR
2
RHAGES
model were broadly similar in subjects who were not
taking antithrombotic therapy at baseline (C statistics
Table 6— Predictive Power of the Bleeding Risk Scores
Used To Assess Risk of Major Bleeding Within 1 Year in
Patients With Atrial Fibrillation
Antithrombotic
Treatment Bleeding Risk Score No. C Statistic (CI)
Overall group Derivation cohort
a
3,381 0.72 (0.64-0.79)
HAS-BLED 3,071 0.72 (0.65-0.79)
HEMOR
2
RHAGES
b
3,040 0.66 (0.57-0.74)
OAC alone Derivation cohort 1,947 0.68 (0.58-0.78)
HAS-BLED 1,722 0.69 (0.59-0.80)
HEMOR
2
RHAGES 1,706 0.64 (0.53-0.75)
OAC 1 antiplatelet
therapy
Derivation cohort 240 0.80 (0.68-0.93)
HAS-BLED 239 0.78 (0.65-0.91)
HEMOR
2
RHAGES 235 0.83 (0.74-0.91)
Antiplatelet
therapy alone
Derivation cohort 788 0.74 (0.52-0.97)
HAS-BLED 753 0.91 (0.83-1.00)
HEMOR
2
RHAGES 728 0.83 (0.68-0.98)
No antithrombotic
therapy
Derivation cohort 348 0.75 (0.51-0.99)
HAS-BLED 315 0.85 (0.00-1.00)
HEMOR
2
RHAGES 311 0.81 (0.00-1.00)
HEMOR
2
RHAGES 5 Hepatic or renal disease, Ethanol abuse, Malig-
nancy, Older age ( . 75 years), Rebleeding, Reduced platelet count or
function, Hypertension (uncontrolled), Anemia, Genetic factors,
Excessive fall risk, and Stroke; OAC 5 oral anticoagulation. See Table
2 for expansion of other abbreviation.

a
Derivation cohort risk factors: bleeding history, age . 65 y, clopidogrel
use, and kidney failure.

b
Classifying bleeding risk by antithrombotic therapy use with the
HEMOR
2
RHAGES scheme resulted in mean scores of 1.17, 1.17,
1.31, 1.24, and 1.07, respectively.
Table 5— The Risk of Major Bleeding Within 1 Year in Patients With Atrial Fibrillation Enrolled
in the Euro Heart Survey
Derivation Cohort
a
HAS-BLED
Risk Factors/Score No. No. of Bleeds Bleeds Per 100 Patient-Years No. No. of Bleeds Bleeds Per 100 Patient-Years
0 1,517 9 0.59 798 9 1.13
1 1,589 24 1.51 1,286 13 1.02
2 219 7 3.20 744 14 1.88
3 41 8 19.51 187 7 3.74
4 14 3 21.43 46 4 8.70
5 1 0 … 8 1 12.50
6 .... .... .... 2 0 0.0
7 ... ... ... 0 ... ...
8 ... ... ... 0 ... ...
9 ... ... ... 0 ... ...
Any score 3,381 51 1.51 3,071 48 1.56
P value for trend , 0.001 0.007
See Table 2 for expansion of abbreviations.

a
Derivation cohort risk factors in multivariate analysis: bleeding history (given 2 points), age . 65 y, clopidogrel use, and kidney failure (maximum
score 5).
Downloaded From: http://journal.publications.chestnet.org/ by a University of Wisconsin User on 11/02/2015

1098 Original Research
with AF before it can be widely implemented into daily
practice. Potential selection bias might have occurred
because of 25% missing data regarding the occurrence
of major bleeding during the follow-up period. Patients
who were lost to follow-up were likely to have had
more comorbidities and transferred to nursing homes
or even have died, which might have led to underes-
timation of the overall bleeding rate. Also, we recog-
nize that the limited number of major bleeds and the
relatively short follow-up period make it possible
that other important risk factors for bleeding were
not identifi ed. Indeed, bleeding may occur following
changes to warfarin (eg, for surgery or interventions
such as pacemaker implant) with institution of bridg-
ing therapy with low-molecular heparin. We also did
not consider thyroid disease in our model, as this had
not been identifi ed as a bleeding risk in prior system-
atic reviews
11-13
; however, some pathophysiologic plau-
sibility is possible, since hypothyroidism has been
described to cause acquired von Willebrand disease
associated with low factor VIII levels and platelet
dysfunction.
26
Of note, there is an improved predictive
power of the HAS-BLED score over the HEMOR
2
-
RHAGES score in patients treated with antiplatelet
therapy (or nonsteroidal antiinfl ammatory drugs) or
no antithrombotic therapy at all ( Table 6 ).
The risk of major bleeding (especially intracranial
hemorrhage) is increased with advanced age.
26-28
Of
note, the risk of major hemorrhage can be similar
among elderly patients receiving warfarin and aspi-
rin.
29
The relatively small numbers of bleeding events
and the modest number of very elderly patients in our
cohort makes it rather diffi cult to draw too many fi rm
conclusions by introducing different weights to differ-
ent age categories (eg, one point for age 65-74 years,
two points for age 75-84 years, three points for
age � 85 years, and so forth), as well as introducing
additional complexity to our simple HAS-BLED
scoring system. Also, age is a continuous (rather than
categorized) risk for bleeding (as well as stroke) and
the biologic age of an elderly patient is probably more
relevant to bleeding risk than the chronologic age.
It must be stressed that in many instances, bleeding
risk among the elderly is multifactorial
30
and is often
the result of associated comorbidities, high anticoag-
ulation intensity, and labile INRs in this population.
11-13

The HAS-BLED score already takes some of these
aspects into account, allowing cumulative assessment
of risk factors for bleeding. In the present analysis,
the HAS-BLED score already outperforms the
HEMOR
2
RHAGES bleeding scheme, which was an
attempt by Gage et al
17
to have a simple method of
bleeding assessment. Future validation and refi ne-
ment of HAS-BLED among a huge elderly AF popu-
lation with prolonged follow-up may address the issue
of age as a continuous variable for bleeding risk.
on stroke risk reduction. Appropriate use of this prac-
tical rule in the Euro Heart Survey on AF population
could have prevented more than one out of every
10 (4/33) of the major bleeds. However, 34/1,580 (2.2%)
of the patients with a CHADS
2
score � 1 discharged
with OAC who did not suffer a major bleed within
1 year would have been denied OAC because of a
HAS-BLED bleed risk outweighing their stroke risk.
The potential impact on current clinical practice
of the novel HAS-BLED score is underlined by the
recently published Atrial Fibrillation Clopidogrel
Trial with Irbesartan for Prevention of Vascular
Events-A (ACTIVE-A) trial.
22
This large random-
ized clinical trial was designed to compare the pre-
ventive effect on all-cause vascular events of clopidogrel
plus aspirin vs aspirin alone in patients with AF deemed
unsuitable for OAC treatment. In half (n 5 � 3,500) of
these patients with AF with a high stroke risk, the
most common applied classifi cation was “unsuitable
for OAC,” which was solely based on physician clinical
judgment, without the presence of any predefi ned
risk factor of bleeding or other objective risk scoring.
Perhaps this refl ects physicians’ uncertainty about what
to consider as true risk factors of bleeding and their
fear of potential iatrogenic harm caused by OAC use.
Given the recent promising results of the Random-
ized Evaluation of Long-Term Anticoagulation Ther-
apy (RE-LY) trial,
23
patients assessed as being at higher
bleeding risk using the novel HAS-BLED score could
be prescribed the lower dose (110 mg bid) of the oral
direct thrombin inhibitor, dabigatran, which demon-
strated a signifi cant reduction in major bleeding
compared with warfarin, with a similar stroke risk
reduction to warfarin, whereas those at lower bleed-
ing risk could be prescribed dabigatran 150 mg bid,
which offers superior effi cacy but a similar major
bleeding risk to warfarin.
24
Further, the HAS-BLED
score could also be used to identify patients who may
benefi t from a left atrial appendage occlusion device
25

(ie, patients at high risk of ischemic stroke who have
such an increased risk of bleeding that OAC is con-
traindicated). Thus, physicians initiating OAC (whether
with the vitamin K antagonist or new oral anticoagu-
lants, such as dabigatran) in a patient with AF could
use the HAS-BLED score to assess the potential
bleeding risk and feel more confi dent in prescribing
OAC as appropriate or refer the patient for implantation
of a left atrial appendage occlusion device. Indeed,
bleeding risk scores should also be validated in dabig-
atran-treated patients, as well as those being con-
sidered for left atrial appendage occlusion devices.
Limitations
The HAS-BLED score needs to be validated in at
least one other large contemporary cohort of patients
Downloaded From: http://journal.publications.chestnet.org/ by a University of Wisconsin User on 11/02/2015

www.chestpubs.org CHEST / 138 / 5 / NOVEMBER, 2010 1099
2 . Risk factors for stroke and effi cacy of antithrombotic therapy
in atrial fi brillation. Analysis of pooled data from fi ve ran-
domized controlled trials . Arch Intern Med . 1994 ; 154 ( 13 ):
1449 - 1457 .
3 . Hart RG , Pearce LA , Aguilar MI . Meta-analysis: antithrom-
botic therapy to prevent stroke in patients who have nonvalvu-
lar atrial fi brillation . Ann Intern Med . 2007 ; 146 ( 12 ): 857 - 867 .
4 . Fuster V , Rydén LE , Cannom DS , et al ; European Heart
Rhythm Association ; Heart Rhythm Society ; American College
of Cardiology ; American Heart Association Task Force on
Practice Guidelines ; European Society of Cardiology Com-
mittee for Practice Guidelines ; Writing Committee to Revise
the 2001 Guidelines for the Management of Patients With
Atrial Fibrillation . ACC/AHA/ESC 2006 guidelines for the
management of patients with atrial fi brillation—executive
summary: a report of the American College of Cardiology/
American Heart Association Task Force on Practice Guidelines
and the European Society of Cardiology Committee for
Practice Guidelines (Writing Committee to Revise the 2001
Guidelines for the Management of Patients With Atrial
Fibrillation) . J Am Coll Cardiol . 2006 ; 48 ( 4 ): 854 - 906 .
5 . Nieuwlaat R , Capucci A , Camm AJ , et al ; European Heart
Survey Investigators . Atrial fi brillation management: a prospec-
tive survey in ESC member countries: the Euro Heart Survey
on Atrial Fibrillation . Eur Heart J . 2005 ; 26 ( 22 ): 2422 - 2434 .
6 . Heeringa J , van der Kuip DA , Hofman A , et al . Prevalence,
incidence and lifetime risk of atrial fi brillation: the Rotterdam
study . Eur Heart J . 2006 ; 27 ( 8 ): 949 - 953 .
7 . Miyasaka Y , Barnes ME , Gersh BJ , et al . Secular trends in
incidence of atrial fi brillation in Olmsted County, Minnesota,
1980 to 2000, and implications on the projections for future
prevalence . Circulation . 2006 ; 114 ( 2 ): 119 - 125 .
8 . Gage BF , Waterman AD , Shannon W , Boechler M , Rich
MW , Radford MJ . Validation of clinical classifi cation schemes
for predicting stroke: results from the National Registry of
Atrial Fibrillation . JAMA . 2001 ; 285 ( 22 ): 2864 - 2870 .
9 . Nieuwlaat R , Capucci A , Lip GY , et al ; Euro Heart Survey
Investigators . Antithrombotic treatment in real-life atrial
fi brillation patients: a report from the Euro Heart Survey on
Atrial Fibrillation . Eur Heart J . 2006 ; 27 ( 24 ): 3018 - 3026 .
10 . Waldo AL , Becker RC , Tapson VF , Colgan KJ ; NABOR
Steering Committee . Hospitalized patients with atrial
fibrillation and a high risk of stroke are not being pro-
vided with adequate anticoagulation . J Am Coll Cardiol .
2005 ; 46 ( 9 ): 1729 - 1736 .
11 . Palareti G , Cosmi B . Bleeding with anticoagulation therapy—
who is at risk, and how best to identify such patients . Thromb
Haemost . 2009 ; 102 ( 2 ): 268 - 278 .
12. National Collaborating Centre for Chronic Conditions . Atrial
Fibrillation: National Clinical Guideline for Management
in Primary and Secondary Care. London, England: Royal
College of Physicians.
13 . Hughes M , Lip GY ; Guideline Development Group for the
NICE national clinical guideline for management of atrial fi bril-
lation in primary and secondary care . Risk factors for antico-
agulation-related bleeding complications in patients with atrial
fi brillation: a systematic review . QJM . 2007 ; 100 ( 10 ): 599 - 607 .
14 . Beyth RJ , Quinn LM , Landefeld CS . Prospective evaluation
of an index for predicting the risk of major bleeding in outpa-
tients treated with warfarin . Am J Med . 1998 ; 105 ( 2 ): 91 - 99 .
15 . Kuijer PM , Hutten BA , Prins MH , Büller HR . Prediction of
the risk of bleeding during anticoagulant treatment for venous
thromboembolism . Arch Intern Med . 1999 ; 159 ( 5 ): 457 - 460 .
16 . Shireman TI , Howard PA , Kresowik TF , Ellerbeck EF .
Combined anticoagulant-antiplatelet use and major bleed-
ing events in elderly atrial fi brillation patients . Stroke . 2004 ;
35 ( 10 ): 2362 - 2367 .
Finally, data about INR control are obviously not
available when having to decide on starting OAC
for the fi rst time in a patient. When on OAC, the
INR is often elevated at the time of admission for a
bleeding event, but it is unknown which measure
of INR control best predicts bleeding in such a
manner that clinical action could prevent the bleed-
ing. We did not include actual INR values during
follow-up but acknowledge its importance as risk
factor of bleeding.
19
Of note, the current alternative
(HEMOR
2
RHAGES model) was also developed
without the availability of INR values.
17

Conclusion
We propose a novel bleeding risk score, HAS-BLED,
that provides an easy, practical tool to assess the indi-
vidual bleeding risk of patients with AF. The use of
this simple score may potentially support clinical deci-
sion making regarding antithrombotic therapy for
stroke prevention.
Acknowledgments
Author contributions: Dr Pisters: contributed to statistical anal-
yses, data interpretation, and drafting of the manuscript.
Dr Lane: contributed to drafting and revision of the manuscript.
Dr Nieuwlaat: contributed to statistical analyses, data interpreta-
tion, and drafting of the manuscript.
Dr de Vos: contributed to drafting and revision of the manuscript.
Dr Crijns: contributed to drafting and revision of the manu-
script.
Dr Lip: contributed to study design and hypothesis, concept
of the HAS-BLED score (the Birmingham Atrial Fibrillation
Bleeding Risk schema), data interpretation, drafting of manu-
script, and revisions.
Financial/nonfi nancial disclosures: The authors have reported
to CHEST the following confl icts of interest: Dr Pisters has con-
sulting fees from Bayer and Boehringel Ingelheim and lecture
fees from Boehringer Ingelheim. Dr Lane is the recipient of an
investigator-initiated educational grant from Bayer Healthcare
and has received sponsorship to attend the European Society of
Cardiology Congress 2009 from AstraZeneca. Dr Crijn has received
consulting fees from Boehringer Ingelheim, Sanofi -Aventis, and
AstraZeneca; grant support from St. Jude Medical, Boston Scien-
tifi c, Boehringer Ingelheim, Sanofi -Aventis, Medapharma, and
Merck; and honoraria from Medtronic, Sanofi -Aventis, Medapharma,
Merck, Boehringer Ingelheim, and Biosense Webster. Dr Lip has
served as a consultant for Bayer, Astellas, Merck, AstraZeneca,
Sanofi -Aventis, Aryx, Portola, Biotronic, and Boehringher Ingelheim,
and has been on the speakers bureau for Bayer, Boehringher
Ingelheim, and Sanofi -Aventis. Drs Nieuwlaat and de Vos have
reported that no potential confl icts of interest exist with any com-
panies/organizations whose products or services may be discussed
in this article.
Other contributions: We thank the Euro Heart Survey team,
national coordinators, investigators, and data collection offi cers
for performing the survey.
References
1 . Wolf PA , Abbott RD , Kannel WB . Atrial fi brillation as an
independent risk factor for stroke: the Framingham Study .
Stroke . 1991 ; 22 ( 8 ): 983 - 988 .
Downloaded From: http://journal.publications.chestnet.org/ by a University of Wisconsin User on 11/02/2015

1100 Original Research
17 . Gage BF , Yan Y , Milligan PE , et al . Clinical classifi ca-
tion schemes for predicting hemorrhage: results from the
National Registry of Atrial Fibrillation (NRAF) . Am Heart J .
2006 ; 151 ( 3 ): 713 - 719 .
18 . Nieuwlaat R , Prins MH , Le Heuzey JY , et al . Prognosis, dis-
ease progression, and treatment of atrial fi brillation patients
during 1 year: follow-up of the Euro Heart Survey on atrial
fi brillation . Eur Heart J . 2008 ; 29 ( 9 ): 1181 - 1189 .
19 . Tay KH , Lane DA , Lip GY . Bleeding risks with combination
of oral anticoagulation plus antiplatelet therapy: is clopidogrel
any safer than aspirin when combined with warfarin? Thromb
Haemost . 2008 ; 100 ( 6 ): 955 - 957 .
20 . Lip GY , Huber K , Andreotti F , et al . Management of antithrom-
botic therapy in atrial fi brillation patients presenting with acute
coronary syndrome and/or undergoing percutaneous coronary
intervention/ stenting . Thromb Haemost . 2010 ; 103 ( 1 ): 13 - 28 .
21 . Lip GY , Zarifi s J , Watson RD , Beevers DG . Physician varia-
tion in the management of patients with atrial fi brillation .
Heart . 1996 ; 75 ( 2 ): 200 - 205 .
22 . Connolly SJ , Pogue J , Hart RG , et al ; ACTIVE Investigators .
Effect of clopidogrel added to aspirin in patients with atrial
fi brillation . N Engl J Med . 2009 ; 360 ( 20 ): 2066 - 2078 .
23 . Connolly SJ , Ezekowitz MD , Yusuf S , et al ; RE-LY Steering
Committee and Investigators . Dabigatran versus warfarin in
patients with atrial fi brillation . N Engl J Med . 2009 ; 361 ( 12 ):
1139 - 1151 .
24 . Fihn SD , Callahan CM , Martin DC , McDonell MB , Henikoff
JG , White RH ; The National Consortium of Anticoagulation
Clinics . The risk for and severity of bleeding complications
in elderly patients treated with warfarin . Ann Intern Med .
1996 ; 124 ( 11 ): 970 - 979 .
25 . Holmes DR , Reddy VY , Turi ZG , et al ; PROTECT AF
Investigators . Percutaneous closure of the left atrial appendage
versus warfarin therapy for prevention of stroke in patients with
atrial fi brillation: a randomised non-inferiority trial . Lancet .
2009 ; 374 ( 9689 ): 534 - 542 .
26 . Homoncik M , Gessl A , Ferlitsch A , Jilma B , Vierhapper H .
Altered platelet plug formation in hyperthyroidism and
hypothyroidism . J Clin Endocrinol Metab . 2007 ; 92 ( 8 ):
3006 - 3012 .
27 . Fang MC , Chang Y , Hylek EM , et al . Advanced age, anticoag-
ulation intensity, and risk for intracranial hemorrhage among
patients taking warfarin for atrial fi brillation . Ann Intern Med .
2004 ; 141 ( 10 ): 745 - 752 .
28 . van Walraven C , Hart RG , Connolly S , et al . Effect of age
on stroke prevention therapy in patients with atrial fi brilla-
tion: the atrial fi brillation investigators . Stroke . 2009 ; 40 ( 4 ):
1410 - 1416 .
29 . Mant J , Hobbs FDR , Fletcher K , et al ; BAFTA investigators ;
Midland Research Practices Network (MidReC) . Warfarin
versus aspirin for stroke prevention in an elderly community
population with atrial fi brillation (the Birminigham Atrial
Fibrillation Treatment of the Aged Study, BAFTA): a ran-
domised controlled trial . Lancet . 2007 ; 370 (9586): 493 - 503 .
30 . Arboix A , Vall-Llosera A , García-Eroles L , Massons J ,
Oliveres M , Targa C . Clinical features and functional outcome
of intracerebral hemorrhage in patients aged 85 and older .
J Am Geriatr Soc . 2002 ; 50 ( 3 ): 449 - 454 .
Downloaded From: http://journal.publications.chestnet.org/ by a University of Wisconsin User on 11/02/2015