DOI: 10.1161/CIRCULATIONAHA.113.008370
Differences in Short-Term Versus Long-term Outcomes of Older Black versus
White Patients with Myocardial Infarction: Findings from CRUSADE
Running title: Mathews et al.; MI Outcomes by Race
Robin Mathews, MD1; Anita Y. Chen, MS1; Laine Thomas, PhD1; Tracy Y. Wang, MD, MHS,
Downloaded from http://circ.ahajournals.org/ by guest on June 14, 2017
MSc1; Chee Tang Chin, MD2; Kevin L. Thomas, MD1; Matthew T. Roe, MD, MHS1;
Eric D. Peterson, MD, MPH1
1
Duke
Du
ke C
Clinical
lini
li
n ca
ni
c l Research Institute, Duke Un
University Medic
Medical
ical C
Center,
ente
en
t r, Durham, NC;
te
2
National
Nati
Na
tion
ti
onnal H
Heart
eaart C
Centre
en
ntre S
Singapore,
inggap
por
oree, S
Singapore
inga
in
gapo
pore
re
Address
Add
Ad
dress for
forr Correspondence:
Corr
Co
rres
espo
pond
nden
ence
c :
Eric
MD,
E
i D.
D Peterson,
P
MD MPH
Duke Clinical Research Institute
Duke University Medical Center
2400 Pratt Street
Durham, NC 27705
Tel: 919-668-8947
Fax: 919-668-7061
E-mail: [email protected]
Journal Subject Code: Ethics and policy:[100] Health policy and outcome research
1
DOI: 10.1161/CIRCULATIONAHA.113.008370
Abstract
Background—Blacks are less likely than whites to receive coronary revascularization and
evidence-based therapies following acute myocardial infarction (MI), yet the impact of these
differences on long-term outcomes is unknown.
Methods and Results—We linked CRUSADE registry data to national Medicare claims, creating
a longitudinal record of care and outcomes among 40,500 non–ST-segment elevation MI
(NSTEMI) patients treated at 446 hospitals to examine mortality and readmission rates (mean
Downloaded from http://circ.ahajournals.org/ by guest on June 14, 2017
follow-up 2.4 years) among black and white patients. Relative to whites (n=37,384), blacks
(n=3,116)
renal
n=3,116) were more often younger and female; more often had diabetes and ren
nal
al ffailure;
aiilu
ure
re;; an
aand
d
received
eceived less aggressive interventions including cardiac catheterization (60.7% vs. 54.0%,
p<0.001),
p<
<0.
0.00
001)
00
1), percutaneous
1)
p rcut
pe
uttan
aneo
e us coronary intervention (32.1%
(332.11% vs. 23.8%, p<0.001),
p 0.
p<
0.00
001)
00
1 , or coronary bypass
surgery
vs.
p<0.001).
Though
lower
30-day
urg
ger
e y (9.2%
% vs
s. 55.7%,
.7%
7%,, p<
7%
p<0.
0.00
0.
001)
1)). Thou
T
hou
ugh bblacks
laackss hhad
add lo
oweer 30
30-day
-da
ay mortality
morrta
tallit
ityy (9.1%
(9.1
(9
.1%
.1
% vs.
vs 9.9%,
9.9%
9.
9%,,
9%
adjusted
HR
CI]:
had
higher
mortality
(27.9%
ad
dju
just
sted
st
e H
ed
R [9
[95%
5%
%C
I] 0.
I]:
00.80
80 [[0.71,
0.71
0.
71,, 0.92
71
00.92]),
.92
2])), th
they
y ha
ad hi
high
gher
gh
er oobserved
bser
bs
e ve
er
vedd mo
m
ort
rttal
alit
ityy at
it
a 1 yyear
eaar (2
27.
7.9%
9%
was
vs. 24.5%, pp<0.001),
<0.0
<0
.001
01)),
01
), aalthough
ltthooug
ughh th
this
i w
is
ass nnot
o ssignificant
ot
igni
ig
n fi
ni
f ca
cant
n aafter
nt
fter
ft
er aadjustment
djus
dj
ustm
us
men
entt on llong-term
ongon
g te
gterm
rm ffollow-up
ollow-up
(HR [95% CI]:1.00 [0.94, 1.07]). Black patients also had higher 30-day (23.6% vs. 20.0%,
p<0.001) and 1-year (62.0% vs. 54.6%, p<0.001) all-cause readmission, but these differences
were no longer significant after risk adjustment on 30-day (HR [95% CI]: 1.02 [0.92, 1.13]) and
long-term (HR [95% CI]: 1.05[1.00, 1.11]) follow-up.
Conclusions—While older blacks with an acute MI had lower initial mortality rates than whites,
this early survival advantage did not persist during long-term follow-up. The reasons for this are
multifactorial, but may include differences in comorbidities and post-discharge care.
Key words: non–ST-segment elevation myocardial infarction, race, disparities, outcomes
2
DOI: 10.1161/CIRCULATIONAHA.113.008370
Racial disparities in cardiac care have been widely reported, including lower use of evidencebased medications and revascularization procedures following myocardial infarction (MI)1-5
among minority populations. These disparities are further exacerbated by lower rates of
treatment in patients of older age.6-8 To date, studies have found that these differences in care
processes do not appear to adversely affect acute (in-hospital) outcomes. Nevertheless, there has
been limited study as to whether longitudinal survival and readmission risk are adversely
affected by differential cardiac care.
We linked Can Rapid Risk Stratification of Unstable Angina Patients Suppress Adverse
Downloaded from http://circ.ahajournals.org/ by guest on June 14, 2017
Outcomes with Early Implementation of American College of Cardiology (ACC)/American
Heart Association (AHA) Guidelines (CRUSADE) registry data to Centers for M
ed
dic
icar
aree &
ar
Medicare
Medicaid Services (CMS) administrative data to compare long-term outcomes between older
bl
laccks and
and whites
whitees in a nationwide sample of MII patients.
patients. Our objectives
objjec
e tive
vees were
w re to compare both
we
blacks
hort
orrt- and long-term
lonngg-tterrm
rm outcomes
outtco
ome
m s off all-cause
all
l -ccau
use death
death and
an
nd readmission
readm
read
missio
miss
io
on (for
(for all-cause,
alll-c
-ccau
ause
se,, acute
a ut
ac
u e MI,
MI and
and
shorthear
he
artt fa
ar
fail
ilur
urre [HF
[[HF])
HF])) of nnon–ST-segment
on–STon–S
T seegm
Tgmen
entt el
elev
evat
ev
atio
ionn myocardial
io
myoocar
my
ardi
diaal infarction
di
inf
nfar
a ct
ar
ctio
i n (NSTEMI)
io
(N
NST
STE
EMI) patients
EMI)
pattiennts
nts
heart
failure
elevation
according to
o race.
rac
ace.
e. Furthermore,
Fur
urth
ther
th
e mo
more
r , we
w sought
sou
ugh
ghtt to evaluate
eva
vallua
uate
t if
te
if differences
diff
di
ffer
ff
erren
ence
cess in outcomes
outc
ou
tccom
mes persisted
per
ersi
s sted after
adjustment for patient baseline characteristics and treatment differences.
Methods
Data Sources
The CRUSADE registry was a voluntary observational data collection and quality improvement
initiative that collected prospective data on patients with non–ST-segment elevation acute
coronary syndrome from January 2001 to December 2006. CRUSADE was designed to track
guideline adherence, provide feedback about performance, and develop quality improvement
3
DOI: 10.1161/CIRCULATIONAHA.113.008370
tools to improve adherence to ACC/AHA recommendations for the treatment of acute coronary
syndrome patients.9 Inclusion and exclusion criteria, data collection, and variables have been
previously described.9 The individual institutional review board of each reporting hospital
approved participation in CRUSADE. All data were abstracted retrospectively and anonymously;
therefore, informed consent was not required.
The process of linking clinical registry data to CMS administrative data has been
previously described.10 Briefly, we used indirect identifiers (i.e., site, age, admission and
discharge date, and gender) to link CRUSADE registry patients to a unique record in the CMS
Downloaded from http://circ.ahajournals.org/ by guest on June 14, 2017
administrative database.
From the original sample of 101,464 patients with NSTEMI •65 years old
ld
d iincluded
n lu
nc
lude
dedd in
de
CRUSADE from 2001 to 2006, 73,660 patients (73 %) with NSTEMI from 514 United States
hoosp
spit
ital
it
alss we
al
w
re m
atch
at
c ed to CMS longitudinal adm
minnistrative data th
throug
uggh the end of 2008.
hospitals
were
matched
administrative
through
Study Population
Stu
Popula
atiion
Study
lim
imit
ited
ed
d our
our population
pop
pul
ulaatiion too the
he CRUSADE
CRU
RUSA
SA
ADE enrollment
enr
n olllmeentt period
perio
io
od between
betw
be
twee
tw
eeen Fe
ebr
b ua
u ry 115,
5, 20
003 an
003
nd
We limited
February
2003
and
1, 22006
0006 du
ue to
o tthe
he aavailability
vail
va
illab
a illit
i y of
o m
orre ccomplete
o pl
om
plet
etee da
et
data
ta dduring
urin
ur
ingg th
in
that
a ttime,
at
ime,
im
e yyielding
e,
i lding a
ie
December 31,
due
more
study population of 45,860 patients. From this sample, we sequentially excluded: patients who
were neither black nor white (n=3,043), patients who did not match on gender (n=501), patients
who died but were listed as being discharged on a later date in CMS data (n=28), and non-index
admissions (i.e., patients from the originally linked sample who had repeat admissions in the
CRUSADE registry and only their index admission was analyzed; n=1,788). The final mortality
analysis population consisted of 40,500 patients from 446 sites participating in CRUSADE. For
the readmission analyses, an additional three exclusion criteria were applied (sequentially):
patients who died during the index admission (n=2,522), patients who were not eligible for
4
DOI: 10.1161/CIRCULATIONAHA.113.008370
Medicare Part A and B fee-for-service plans during the index admission (n=1,781), and patients
with an invalid time to readmission (n=7), yielding a final population of 36,190 patients across
446 CRUSADE sites.
Study Variables and Definitions
Participating hospitals used a standardized set of data elements and definitions to collect detailed
information on baseline demographic and clinical characteristics, processes of care, and inhospital outcomes. Data were screened upon entry and only those that met predetermined criteria
for completeness and accuracy were entered into the database. Acute care and discharge
Downloaded from http://circ.ahajournals.org/ by guest on June 14, 2017
treatments were defined as the use of medications within the first 24 hours of index admission
respectively.
Alll ot
other
and those on medications discharged from the index hospitalization, respectively
y. Al
A
othe
herr
he
variables have been previously described.9 Since cause-specific mortality was not available from
he CR
CRUS
USAD
US
ADE an
AD
andd CMS databases, the outcome
mess ooff interest in oour
me
u stu
ur
tu
udy were all-cause
the
CRUSADE
outcomes
study
m
mortality.
orrtality.
rt
Furthermore,
Furrth
herm
more, we studied
mor
stu
tudi
died
ed
d all-cause
all
ll-c
-caausee readmission,
readm
dm
missiionn, acute
acuute
ute MI
MI readmission,
reaadmisssi
sion
on,, and
and HF
readmission,
ead
admi
miss
mi
ssio
io
on, where
wheree using
using
ng International
Intter
e na
nati
tion
ti
onal
all Classification
Cla
lasssiifi
fica
cattioon of
ca
of Diseases,
Dise
Di
seeas
a es, Ninth
N nth
Ni
nth Revision
Reevi
v si
s onn primary
prim
mary
mary
r
discharge diagnosis
diag
agno
ag
n si
no
siss code
co
ode
d 410.x
410
10.x
.x was
wass used
useed to classify
clas
cl
a si
as
s fy acute
acu
ute MI
MI readmission
read
re
a mi
m ss
s io
ionn an
aand
d 42
428.
428.x,
8.x,
8.
x, 402.x1,
404.x1, 404.x3, and 398.91 were used to classify HF readmission. Transfers to or from another
hospital and admission for rehabilitation were not considered readmissions.
Statistical Analysis
Patient baseline characteristics, as well as in-hospital treatment patterns and outcomes were
compared between black and white patients using the Mantel-Haenszel Chi-square test for
categorical variables and the Wilcoxon rank-sum test for continuous variables; frequencies and
percentages were used for categorical variables and medians (with 25th and 75th percentiles) were
used for continuous variables.
5
DOI: 10.1161/CIRCULATIONAHA.113.008370
The primary outcomes of interest were time-to-first event for: 1) all-cause mortality; and
2) readmission (for all-cause, acute MI, and HF). Kaplan-Meier curves were generated to
estimate the probability of mortality by race and the log-rank test was used to assess whether the
differences between the mortality curves were statistically significant at p<0.05. The incidence of
all-cause, acute MI, and HF readmission is directly influenced by the competing risk for
mortality; therefore, we estimated the cumulative incidence of readmission at various time points
following the index admission.11 The incidence for all-cause, acute MI, and HF readmission was
compared between black and white patients using the Gray test.
Downloaded from http://circ.ahajournals.org/ by guest on June 14, 2017
Cox proportional hazards modeling was performed to examine the association between
ace and short- and long-term outcomes. Robust standard errors were used to acc
cou
ounnt ffor
or
race
account
clustering of patients within hospitals.12 Since historical data on survival of black versus white
pa
atiien
ents
ts have
hav
ave sh
how
ownn different short- and long-te
erm m
ortality according
accor
ording
or
ngg to
to race,13-15 separate
patients
shown
long-term
mortality
an
analyses
nallys
y es weree pe
performed
erfforrmedd in eeach
ach ti
ach
ttime
me pperiod.
me
eriodd. Thee ffollow-up
ollow
ow-u
ow
-u
up pe
per
period
riod ffor
rio
orr sshort-term
horrtho
rt-ter
-term
m mor
m
mortality
ortal
aliity
al
i ty w
was
as
300 days
day
ayss from
from the
thee date
dat
atee off index
ind
ndeex
ex admission
adm
dmiissi
issiionn (n=40,500).
(n=40
40,5
40
,5500). Short-term
Sho
horrt-t
rt ter
e m readmission
read
re
a mi
ad
miss
ssio
ionn was
io
waas assessed
asssesssed
among thosee wh
whoo survived
surv
su
rv
viv
ved
d the
the
h initial
ini
niti
tial
a hospitalization
al
hos
o pi
p ta
tali
liza
li
z ti
za
t on and
andd those
tho
hose
see eligible
eli
l gib
ible
le ffor
or M
Medicare
edic
ed
icar
ic
aree Pa
ar
P
Part
rt A and
B fee-for-service plans during the index admission, with follow-up from discharge to 30 days
afterwards (n=36,190).
Analysis of long-term mortality included only those patients that survived the initial
hospitalization and the first 30 days after discharge (n=36,046) and continued for a mean of 2.6
years. Long-term readmission was evaluated starting 30 days after discharge and included only
survivors who had not been readmitted within the first 30 days (n=27,618). Maximum follow-up
was 4 years and subjects were censored beyond this point.
The models were adjusted sequentially in order to see the contribution of different types
6
DOI: 10.1161/CIRCULATIONAHA.113.008370
of covariates. The primary mortality analysis was adjusted for variables based on the CRUSADE
mortality model16: age, weight, sex, prior stroke, diabetes, peripheral artery disease,
hypertension, dyslipidemia, prior percutaneous coronary intervention (PCI), prior MI, prior
coronary bypass grafting (CABG), current or recent smoker, prior history of HF, signs of HF,
family history of coronary artery disease, systolic blood pressure and heart rate on admission,
initial serum creatinine, initial hematocrit, initial troponin ratio, and electrocardiogram (ECG)
findings (Model 1).
For the primary readmission analyses, acute episodes of care within the past year and
Downloaded from http://circ.ahajournals.org/ by guest on June 14, 2017
hospital transfer-in status were additionally included in the list of covariates; we referred to this
model as Model 1 for all readmission analyses. Including the covariates in Model
Mode
deel 11,, we
w then
the
henn
equentially adjusted for insurance coverage beyond Medicare (Model 2), income level (income
sequentially
eveel was
was determined
dete
de
t rm
te
min
ined
e using the 2006 area resourc
rcee ffile
rc
ile17 merged wi
w
with
th
h tthe
he facility zip code; high
level
resource
and
an
nd lo
llow
w income
me w
was
as ddetermined
etter
ermi
mine
mi
nedd usin
ne
uusing
sin
i gm
median
ediian
n inc
income;
come;; M
Model
odel 3),
odel
3), ddischarge
isch
is
c arge
ch
gee medications
med
dic
i atio
ationss
(aspirin,
inhibitor,
asp
spir
irrin
in,, beta
beeta blocker,
blo
ock
ker
er,, any
an
ny lipid
liipi
pidd lowering
low
ower
erin
in
ng agent,
agen
ag
en
nt, aangiotensin
ngio
ng
i te
t ns
nsin
in cconverting
o vert
on
vert
rtin
i g eenzymes
in
nzy
yme
mess in
inh
hibi
biito
or, aand
nd
clopidogrel;; Mo
Model
Mode
deel 44),
) and
),
andd revascularization
rev
e as
ascu
c la
cu
lariiza
z ti
tion
on through
thr
h ou
ough
gh PCI
PCI
C or
or CABG
CABG (Model
(Mo
Mode
dell 55).
de
). A
Also,
lso,
ls
o, we
investigated the impact of hospital region (Northeast, Midwest, West, and South) and academic
versus non-academic (defined as Membership in the Council of Teaching Hospitals). Finally, we
evaluated whether the impact of specific prognostic subgroups varied by race: gender, age group
(•75 years old, <75 years old), diabetes, and income level by testing for interaction between race
and subgroups. For statistically significant interactions, the hazard ratio (HR; 95% confidence
interval [CI]) for outcomes by race were reported among subgroups.
To assess the differential impact between large MIs caused by acute plaque rupture and
smaller MIs resulting from complications of hypertension, renal failure, and other comorbidities,
7
DOI: 10.1161/CIRCULATIONAHA.113.008370
we performed sensitivity analyses restricted to patients having large MIs as defined by peak
troponin values greater than 5-fold the upper limit of normal (ULN; n=27,196).
All analyses were performed using the SAS software package (SAS 9.3; SAS Institute, Cary,
NC).
Results
Baseline Clinical Characteristics
In our final population, 7.7% (n=3,116) of patients were black and 92.3% (n=37,384) were
Downloaded from http://circ.ahajournals.org/ by guest on June 14, 2017
white. Although all patients were older than 65 years of age, black patients were younger than
prior
white patients (median age of 76 years vs. 78 years), but had a higher prevalencee off pr
rio
iorr HF
HF,,
stroke,
troke, diabetes, and hypertension, and a much higher prevalence of both renal insufficiency and
ongoing
dialysis
HF
presentation,
on
ngo
goin
ingg di
in
dial
alyysiss ((Table
al
Ta
Table
1). Blacks more often hhad
ad H
ad
F on present
tat
a ion,
n, bbut
u less often exhibited
ut
dynamic
ST changes
additional
dynamic
dyn
na
cha
hang
ngess on
ng
on ann ECG.
ECG
CG. Compared
Com
Co
mpared
mpa
d with
withh whites,
whittess, black
black patients
bla
patien
pati
entts less
les
esss ofte
ooften
ftenn ha
hadd ad
ddi
diti
tionnal
ti
private
insurance
priv
pr
ivat
iv
atee in
at
insu
suura
ranc
ncee cove
ccoverage.
oveeraage
g .
Patterns of Treatment
Trea
Tr
e tm
ea
tmen
en
nt
Black patients were less often treated with acute antiplatelet therapies such as glycoprotein
IIb/IIIa inhibitors (34.0% vs. 41.0%) and clopidogrel (45.0% vs. 52.5%) while rates of aspirin,
heparin, or beta blocker use were not significantly different between the groups (Table 2). Black
patients less often underwent diagnostic cardiac catheterization, as well as revascularization
through PCI or CABG. At discharge, black patients less often received clopidogrel compared
with whites, but there was no difference in the receipt of aspirin, beta blockers, and lipid
lowering agents.
Outcomes
8
DOI: 10.1161/CIRCULATIONAHA.113.008370
Mortality
At 30-days post-MI, black patients had a similar unadjusted mortality rate relative to whites
(9.1% vs. 9.9%, p=0.13), yet by six months these curves crossed (20.1% vs. 18.9%, p=0.17), and
by 1-year, mortality was higher among blacks than whites and remained parallel thereafter: 1
year (27.9% vs. 24.5%, p<0.001), 2 years (37.7% vs. 33.3%, p<0.001) and 3 years (44.6% vs.
40.5%, p<0.001; Figure 1). After adjustment for patient baseline clinical risk factors, the
difference in 30-day mortality widened in favor of black patients (hazard ratio [HR]; 95%
confidence interval [CI]: 0.80 [0.71, 0.92] of Model 1; Figure 2). In contrast, beyond 30 days,
Downloaded from http://circ.ahajournals.org/ by guest on June 14, 2017
long-term adjusted mortality was not significantly different between groups (HR [95% CI]: 1.00
[0.94,
0.94, 1.07] of Model 1) and similar results were found with models adjusting for
for additional
addditi
ad
diti
tion
onal
on
al
variables of other insurance coverage beyond Medicare, income level, discharge medications,
and
in-hospital
revascularization
an
nd in
in-h
-hos
hosspi
pita
t l re
ta
evascularization
va
(Figure 2). When
Whe
henn hospital characteristics
he
charaact
c errissti
ticcs were included in the
last
astt model,
model, wee also
als
l o noted
nooteed no additional
addit
ddittio
onaal impact
impaact off hospital
hospi
pittal
pi
tal region
regi
g onn or
gi
or teaching
teeachi
hing
ng hospital
hos
ospi
p ta
pi
tall on
on sshortho
hort
o t- oor
long-term
mortality
(data
presented).
Furthermore,
factors
such
ong
ng-t
-tter
e m mo
mort
rtaalitty (d
datta no
nott pr
pres
esen
es
ente
en
ted)
d . Fu
d)
Furt
the
herm
rm
more,
ore,, tthe
he eeffect
ffec
ff
e t of pprognostic
ec
ro
ognnossti
t c fa
fact
ctoorss su
ct
uch aass
sex,
income
was
according
short-term
mortality
ex, age, diabetes,
diab
abet
etes
et
e , an
es
andd in
inco
ome w
a ssimilar
as
im
mil
i ar ac
acco
cord
co
rddin
ng to
o rrace
acee fo
ac
forr sh
shor
o tor
t te
term
rm m
orta
or
tali
ta
lity
li
ty
(pinteraction>0.05; Supplement 1).
Readmission
Observed cumulative incidence rates of all-cause readmission were higher among black patients
compared with white patients at 30 days (23.6% vs. 20.0%, p< 0.001), 6 months (50.5% vs.
43.2%, p<0.001), 1 year (62.0% vs. 54.6%, p<0.001), 2 years (73.1% vs. 66.7%, p<0.001), and 3
years (78.9% vs. 73.5%, p<0.001) (Figure 3A). Nevertheless, after adjusting for patient casemix and treatment strategies, there was no significant difference in short- and long-term all-cause
readmission between black and white patients (Figure 4). We also found no difference in
9
DOI: 10.1161/CIRCULATIONAHA.113.008370
readmission after adjustment for hospital region and teaching hospital status (data not presented).
Finally, the interaction between subgroups (according to gender, age group, diabetes, and income
level) and race was not statistically significant for short- and long-term all-cause readmission
(Supplement 2).
We also evaluated cause-specific readmission due to acute MI or HF. As seen in the
cumulative incidence curves, black patients had an increased incidence of acute MI and HF
readmission over time (Figures 3B and 3C). After adjustment, there was no difference in shortand long-term readmission for acute MI between the race groups, but black patients remained
Downloaded from http://circ.ahajournals.org/ by guest on June 14, 2017
more likely to be readmitted long-term with HF (HR [95% CI]: 1.13 [1.03, 1.24] of Model 1,
between
Figure 4). After adjusting for patient case-mix (Model 1), the interaction betwee
eeen se
ssex
x and
an
nd ra
race
c
for short-term readmission for HF was significant (HR [95% CI]: 1.38 [1.10, 1.74] among males;
males
CI]:
HR
R [[95%
95%
95
% CI
CI]
]: 00.83
.883 [0.65, 1.06] among females)) ((Supplement
Suupplement 2).
Sensitivity
Analysis
S
en
ns
nsitivity
An
nal
alyssiss
size
outcomes,
limiting
To assess
ass
sses
e s the
es
the effect
efffec
ef
fect
ct ooff MI
MI siz
izze on cclinical
liniicaal ou
lini
outc
com
omess, we
we pperformed
erfoorm
erfo
rmed
d ssensitivity
en
nsiiti
tivi
viity aanalyses
naaly
lysees li
imi
mitting
ng
those
with
greater
our population
onn tto
o th
thos
osee pa
os
ppatients
tien
ti
entss w
en
ith
h peak
pea
eakk troponin
trop
tr
opon
op
onin
on
in
n values
val
a ue
uess gr
grea
eate
ea
terr th
te
than
an 55-fold
-ffol
o d th
thee UL
ULN.
N The
results of sensitivity analyses were similar to the main analysis: black patients continued to have
lower 30-day adjusted mortality (HR [95% CI]: 0.80 [0.68, 0.94] of Model 1), but similar
adjusted long-term mortality compared to white patients (HR [95% CI]: 0.98 [0.90, 1.07] of
Model 1; Supplement 1). The increased risk of 30-day and long-term all-cause readmission that
blacks maintained over whites was again attenuated after adjusting for patient case-mix (HR
[95% CI]: 1.08 [0.96, 1.21] and HR [95% CI]: 1.03 [0.95, 1.11] of Model 1, respectively;
Supplement 2). Although black patients remained more likely to be readmitted for HF over longterm follow-up in our main analysis, this no longer reached statistical significance among
10
DOI: 10.1161/CIRCULATIONAHA.113.008370
patients with peak troponin >5x ULN (HR [95% CI]: 1.13 [0.98, 1.30] of Model 1).
Discussion
Despite receiving less guideline-based acute medical treatments and coronary revascularization,
blacks had lower 30-day mortality than whites after adjustment of patient baseline differences;
however, among survivors to 30 days, this early advantage of blacks was lost over the long-term.
Additionally, black patients were more often readmitted for HF than white patients in long-term
follow-up.
Downloaded from http://circ.ahajournals.org/ by guest on June 14, 2017
Differences in Care
Similar to previous reports, we note similar use of most secondary prevention m
medications,
ediccattio
edic
ions
ns,, bu
ns
bbut
utt
found poorer uptake of newer and more potent anti-platelet therapies,5 as well as less use of
nvaasi
sivve
ve pprocedures
roccedu
ro
ure
ress among blacks in the treatmen
ent ooff MI.18 The reasons
en
reeason
on
ns for
for this are multifactoria
multifactorial
invasive
treatment
and
an
nd li
llikely
kely inclu
include
ludde a pphysician’s
hysi
hy
sici
si
cian
ci
an’’s
an
’s ddecision
ecis
ec
isiion to aavoid
void
dm
more
oree aaggressive
gggreess
ssiive tr
trea
treatments
eaatm
men
nts ddue
uee tto
o pa
patient
atiient
e nt
comorbidities.
are
among
certain
co
omo
morb
rbid
rb
idit
ittie
ies.
s. Invasive
Invas
nvas
a iv
ve procedures
p oc
pr
oced
edur
ed
ures
ur
es carry
carrry
y upfront
upfro
fro
ont
n rrisks
isskss tthat
hatt ar
ha
re increased
in
ncrrea
ease
seed am
mon
ongg ce
cer
rt i n
rtai
subgroups.
ubgroups. These
The
hese
s risks
se
ris
isks
kss may
may lead
lea
e d to
t renal
ren
nal
a failure
fai
ailu
lu
ure po
post
post-contrast
stt-c
- on
ontr
tras
tr
ast dy
as
dyee in
inje
injection
ject
je
ctio
ct
ionn orr iincreased
io
ncre
nc
reeas
a ed
morbidity among those who undergo CABG19; however, the lack of an invasive evaluation may
deprive these higher risk patients of the long-term benefits from revascularization such as
improved downstream morbidity or mortality.7 Other factors that have been cited include lack of
financial resources and possible provider bias.20 Additionally, it is possible that age and culturalrelated issues such as patient-provider communication,21 patient self-health awareness,22 and
overall social support structures23 may contribute to disparities in receipt of diagnostic and
therapeutic treatments.
Short-term Mortality
11
DOI: 10.1161/CIRCULATIONAHA.113.008370
We noted significantly lower adjusted 30-day mortality in blacks than in whites, despite
differential treatments with less guideline-based therapies. Several prior analyses have found
similar results of a short-term survival advantage seen among blacks more so than whites for
cardiac and non-cardiac conditions.13-15,24,25 One possibility for these findings in black patients
may be the higher prevalence of Type 2 or secondary MIs seen in the setting of renal failure or
hypertensive disease. These MIs are caused by increased oxygen demand or decreased oxygen
supply rather than acute plaque rupture26 and are often associated with smaller infarct size and
less myocardial damage as evidenced by lower troponin elevations. These secondary MIs may
Downloaded from http://circ.ahajournals.org/ by guest on June 14, 2017
have less of a detrimental effect on clinical outcomes than larger MIs. Nevertheless, our
ensitivity analyses did not support this hypothesis; rather, we found that blacks sstill
til
illl ha
hadd lo
lowe
w r
we
sensitivity
lower
30-day adjusted mortality even with smaller MIs excluded. Alternatively, the racial differences
n sshort-term
hort
ho
rt-ter
ter
erm
m surv
rviival
rv
iv may also reflect a “survivor
“survivo
voor eeffect,”
ffect,” wheree blacks
b accks who
bl
who survive to age 65,
in
survival
13-15
13
15
m
ayy representt a healthier
heeallth
thieer cohort
coho
co
hort
ho
rt of
of patients
patieents than
pat
th
han white
whitee ppatients
atiients
at
nts of
of tthe
he ssame
am
me ag
ge.13IItt is also
als
lsoo
may
age.
unkn
un
unknown
know
kn
ownn iff black
ow
bla
lack
ck
k patients
pat
atiients
nts with
wiith larger
lar
arge
gerr MIs
MIs may
mayy bee more
ma
more
r likely
lik
ikel
elyy too die
die before
bef
efoore
ore re
rreaching
a hi
ac
hing
ng tthe
he hhospital;
he
osspiita
tal
therefore,
herefore, tho
those
ose bblacks
lack
la
c s who
ck
w o are
wh
ar admitted
admi
ad
mitt
mi
t ed may
may have
hav
ve less
less severe
sev
ever
eree disease
er
dise
di
seas
a e than
as
than whites.
whi
hite
tes.
te
s. In
In addition too
age and gender, there are likely additional unmeasured differences between these groups that
confer a protective benefit toward black patients that we have not fully captured or adjusted for.
Long-term Mortality
The early survival advantage noted at 30 days where blacks had lower adjusted mortality than
whites is not sustained over the long-term where risk-adjusted mortality among blacks was
similar to whites. We found no incremental difference in the magnitude of the HR in long-term
mortality among the groups after additional adjustment for socioeconomic status, discharge
medications, and in-hospital revascularization.
12
DOI: 10.1161/CIRCULATIONAHA.113.008370
Our results support previous findings that black patients are less likely to be referred for
coronary angiography, as well as less likely to undergo percutaneous or surgical
revascularization after an acute MI, even when accounting for comorbid illnesses and disease
severity.1,5,7,8,24,27-30 Although this did not impact their short-term mortality where they had a
survival advantage, it is clear that blacks did not maintain that same trajectory of survival over
whites. This convergence of long-term survival may be due to a combination of less
revascularization, as well as accelerated progression of complications from their comorbidities
and post-discharge influences that were not adequately captured in this analysis. Although we
Downloaded from http://circ.ahajournals.org/ by guest on June 14, 2017
did not find that differential rates of revascularization modified the relative risk for death among
blacks over time, it is possible thatt the receipt of revascularization
revascularization itself may inst
sttea
eadd be a m
a ker
ar
instead
marker
for better follow-up or long-term care such as more aggressive management of comorbidities and
ca
ard
dia
iacc risk
riisk factors.
facto
to
ors
rs.
cardiac
ead
dm
dmissions
eadmissions
Prev
Pr
Previous
evio
ev
ious
io
u studies
us
stu
tudi
dies
es hhave
ave fo
ave
found
oun
undd th
tthat
at bblack
lack
la
ck
k rrace
acee ma
ac
mayy be
be aan
n in
independent
nde
depe
pend
pe
n en
nd
entt pr
ppredictor
edicto
edi
icto
or of
o 330-day
0-da
0dayy
eadmission am
amon
ongg Me
on
M
d caaree bbeneficiaries.
di
enef
en
effic
i iaari
r es
e .31 De
Despite
D
spit
sp
itte th
the
he fa
fact
c tthat
ct
hat we ffound
ha
ound
ou
nd cconsistently
onsi
on
sist
si
sten
st
ently higher
en
readmission
among
Medicare
crude rates of all-cause and HF readmission within 30 days in blacks than in whites, these
differences diminished after adjustment. Similarly, when examining long-term follow-up, blacks
continued to have higher unadjusted cumulative incidence rates of all-cause and acute MI
readmission that again were attenuated after adjustment. This attenuation of readmission risk
with adjustment of patient characteristics reflects the increased burden of comorbid disease in
blacks more so than in whites at the time of their event. Interestingly, even after adjustment,
blacks remained more likely to be readmitted for HF over long-term follow-up. In addition to
differences in chronic conditions between blacks and whites, it is possible that reasons other than
13
DOI: 10.1161/CIRCULATIONAHA.113.008370
the covariates included in this multivariate model contribute to the likelihood of readmission. For
example, poor medication adherence post-discharge, limited access to follow-up care, or poor
quality outpatient management,32-34 may also be implicated in readmission.
Clearly, early survival advantage of black patients erodes as time from the index
admission passes, implicating unmeasured factors once these patients return home. Recent
attention on readmissions deemed “preventable” has placed focus on care that occurs postdischarge, such as availability of social support, quality of ambulatory care, and adherence to
prescribed medications. All of these post-discharge factors may influence long-term outcomes
Downloaded from http://circ.ahajournals.org/ by guest on June 14, 2017
and may, in part, explain some of the differential long-term findings in our study. Therefore,
moving forward, quality improvement efforts should be multifaceted and addres
ss di
isppar
arit
itie
it
iess no
ie
not
address
disparities
only in the acute or short-term setting, but along the entire continuum of care.
Limi
Li
Limitations
mita
mi
tati
ta
tion
ti
onss
on
O
Our
urr st
sstudy
udy shou
should
ould
ld bbee cconsidered
onsid
nsid
ider
ered
er
ed iin
n li
ligh
light
gh
ht ooff sev
several
ev
verall llimitations.
imiitaati
tioons.
s. F
First,
irsst, ou
our
ur lo
long
long-term
ng--ter
ng
-term
m an
anal
analysis
a yssis
al
is of
of
mo
mortality
ort
r al
alit
i y and
it
and readmission
read
re
admi
ad
m ssio
mi
on outcomes
ouutc
tcom
omes
om
es is
is limited
liimi
mittedd to
t ann elderly
eld
l er
ld
erly
ly population
pop
o ulat
ulat
atio
i n enrolled
io
enro
oll
l ed in
in fee-for-service
fee--fo
feeorr-se
seervvic
i e
Medicare; ass a rresult,
essul
ult,
t ou
t,
our analysis
anal
an
a yssis may
al
may
y not
not be
be generalizable
gene
ge
nera
ne
rali
ra
l za
zabble to
to younger
youn
yo
unnge
gerr patients
pati
pa
tien
ti
en
nts or
or those
thos
th
o e with
alternate primary insurance coverage. Second, despite the fact that the association between race,
care rendered, and outcomes was adjusted for patient case-mix, unmeasured confounding or bias
in this observational data set is possible. For example, the effect of race on cardiovascular
outcomes is often intertwined with measures of socioeconomic status including insurance
coverage, income and educational level, and access to health care. Despite the fact that we
attempted to control for some of these factors through the use of aggregate census data, the
methodology is imperfect. Also, the use of revascularization during the index admission is not
randomized and is influenced by differences in baseline comorbidities between the race groups,
14
DOI: 10.1161/CIRCULATIONAHA.113.008370
as well as differences in coronary anatomy which are not collected in our data set. Third, we do
not have data on post-discharge factors that may influence downstream outcomes, such as
adherence to medications prescribed or physician follow-up after discharge. Finally, hospitals
participate in CRUSADE voluntarily and urban centers that treat a larger proportion of black and
minority patients may not be well represented in this group.
Conclusions
Despite efforts aimed at mitigating racial disparities in the use of evidence-based treatments and
Downloaded from http://circ.ahajournals.org/ by guest on June 14, 2017
adverse clinical outcomes among patients with acute MI, our study found persistent differences,
among
even among older patients. In addition to a higher burden of chronic illness amo
ong
ng bblacks,
laack
cks,
s,
differences in MI care, as well as post-discharge management, may contribute to the
discrepancies
di
isccre
reppanc
panccie
iess in
n sshortho and long-term outcomes am
hortaamong
ong blacks and
ndd whites.
whi
hiite
tess. The factors underlying
racial
disparities
may
differences
ranging
presentation
aciial dispariti
ties
es aare
re ccomplex
om
mpl
plex
ex
x aand
nd
dm
ay
y rreflect
efleect diff
ferren
nce
cess ra
angin
ngin
ng fr
from
om
m cclinical
liini
nica
call pr
ressen
enta
t ti
ta
tion
onn aand
nd
medical
decision-making
access
care
once
me
edi
dica
call de
ca
eci
cisi
sion
onn-m
mak
kin
ingg to
t environmental
envvir
iron
onme
on
meent
ntal
al barriers
bar
arri
r ers llimiting
ri
im
mittin
ingg ac
acce
ceesss tto
o qual
qquality
ualit
itty he
hhealth
allth ca
are on
are
onc
ce
ce
transitioned
Further
work
patients havee tr
tran
a si
an
siti
tion
ti
onned
e bback
ack in
ac
into
t tthe
to
hee ccommunity.
ommu
om
munnit
mu
i y.
y F
urrth
ther
e w
er
orkk iss nneeded
or
eede
ee
dedd to
de
o uunderstand
nder
nd
erst
er
s and and
st
prevent the erosion of benefit in long-term outcomes in blacks.
Acknowledgments: The authors would like to thank Erin Hanley, MS for her editorial
contributions to this manuscript. Ms. Hanley did not receive compensation for her assistance,
apart from her employment at the institution where this study was conducted.
Funding Sources: The Can Rapid Risk Stratification of Unstable Angina Patients Suppress
Adverse Outcomes with Early Implementation of American College of Cardiology
(ACC)/American Heart Association (AHA) Guidelines (CRUSADE) registry was funded by
Millennium Pharmaceuticals, Schering-Plough Corporation, and the Bristol-Myers
Squibb/Sanofi Pharmaceuticals Partnership.
15
DOI: 10.1161/CIRCULATIONAHA.113.008370
Conflict of Interest Disclosures: Dr. Wang reports consulting for the American College of
Cardiology (significant) and Medco (modest); research funding for AstraZeneca, Bristol Myers
Squibb, Gilead, Heartscape Technologies, Lilly, Sanofi-Aventis, Schering-Plough, and The
Medicines Company (all significant); educational activities of lectures for AstraZeneca (modest).
Dr. Roe reports research funding from Eli Lilly & Company, KAI Pharmaceuticals, and SanofiAventis (all significant); educational activities or lectures for AstraZeneca and Janssen
Pharmaceuticals (both modest); consulting for Bristol Myers Squibb, Eli Lilly & Company,
Glaxo Smith Kline, and Regeneron (all modest), Merck & Co., Janssen Pharmaceuticals, and
Daiichi-Sankyo (all significant). Dr. Peterson reports research funding from Eli Lilly &
Company, Ortho-McNeil-Janssen Pharmaceuticals, Inc., Society of Thoracic Surgeons,
Downloaded from http://circ.ahajournals.org/ by guest on June 14, 2017
American Heart Association, American College of Cardiology (all significant); consulting for
g Ingelheim,
g
, Genentech,, Johnson & Johnson,, Ortho-McNeil-Janssen
AstraZeneca,, Boehringer
Pharmaceuticals, Inc., Pfizer, Sanofi-Aventis, and WebMD (all modest). The rem
main
inin
in
ng au
auth
thor
th
ors
remaining
authors
have no disclosures to report.
Re
efe
fere
renc
n es
es::
References:
1. Peterson
1.
Peterson ED, Wr
W
Wright
ighht SM,
M,, D
Daley
aley
al
y JJ,, Th
Thibau
Thibault
ultt GE
GE.
E. Ra
Raci
Racial
cial
ial vvariation
ariattion in ca
cardiac
ardi
diiacc proce
procedure
c du
ce
dure
re usee aand
n
nd
survival
urv
r iv
i al follo
following
lowi
lo
w ng aacute
wi
cu
utee myo
myocardial
yo
ocaard
rdia
iaal in
infarction
nfaarc
rcti
tion
n iin
n th
the
he De
Depa
Department
parrtm
ment ooff V
men
Veterans
eter
et
eran
anss Af
an
Affa
Affairs.
faairrs. JA
JAMA.
AMA
A.
1994;271:1175-1180.
1994
19
94;2
94
;271
;2
71:1
71
:11175:1
175--11
1180
80..
80
2. Bradley
yE
EH,
H H
H,
Herrin
e ri
er
rinn J,
J W
Wang
angg Y,
an
Y, M
McNamara
cN
Nam
a ar
a a RL
RL, We
W
Webster
bstter
bs
ter TR
TR, Ma
Magi
Magid
gidd DJ
gi
DJ, Bl
Blaney
lan
aney
ey M
M,, Peterson
ED, Canto JG, Pollack CV Jr, Krumholz HM. Racial and ethnic differences in time to acute
reperfusion therapy for patients hospitalized with myocardial infarction. JAMA. 2004;292:15631572.
3. Whittle J, Conigliaro J, Good CB, Lofgren RP. Racial differences in the use of invasive
cardiovascular procedures in the Department of Veterans Affairs medical system. N Engl J Med.
1993;329:621-627.
4. Sheifer SE, Escarce JJ, Schulman KA. Race and sex differences in the management of
coronary artery disease. Am Heart J. 2000;139:848-857.
5. Sonel AF, Good CB, Mulgund J, Roe MT, Gibler WB, Smith SC Jr, Cohen MG, Pollack CV
Jr, Ohman EM, Peterson ED; CRUSADE Investigators. Racial variations in treatment and
outcomes of black and white patients with high-risk non-ST-elevation acute coronary
syndromes: insights from CRUSADE (Can Rapid Risk Stratification of Unstable Angina Patients
Suppress Adverse Outcomes With Early Implementation of the ACC/AHA Guidelines?).
Circulation. 2005;111:1225-1232.
16
DOI: 10.1161/CIRCULATIONAHA.113.008370
6. Alexander KP, Roe MT, Chen AY, Lytle BL, Pollack CV Jr, Foody JM, Boden WE, Smith SC
Jr, Gibler WB, Ohman EM, Peterson ED; CRUSADE Investigators. Evolution in cardiovascular
care for elderly patients with non–ST-segment elevation acute coronary syndromes: results from
the CRUSADE National Quality Improvement Initiative. J Am Coll Cardiol. 2005;46:14791487.
7. Peterson ED, Shaw LK, DeLong ER, Pryor DB, Califf RM, Mark DB. Racial variation in the
use of coronary-revascularization procedures. Are the differences real? Do they matter? New
Engl J Med. 1997;336:480-486.
8. Chen J, Rathore SS, Radford MJ, Wang Y, Krumholz HM. Racial differences in the use of
cardiac catheterization after acute myocardial infarction. New Engl J Med. 2001;344:1443-1449.
Downloaded from http://circ.ahajournals.org/ by guest on June 14, 2017
9. Hoekstra JW, Pollack CV Jr, Roe MT, Peterson ED, Brindis R, Harrington RA, Christenson
RH, Smith SC, Ohman EM, Gibler WB. Improving the care of patients with non-ST-elevation
acute coronary syndromes in the emergency department: the CRUSADE initiative. Acad Emerg
Med. 2002;9:1146-1155.
10. Hammill BG, Hernandez AF, Peterson ED, Fonarow GC, Schulman KA, Cu
Curtis
urtis
iss L
LH.
H. L
H.
Linking
inki
in
king
g
inpatient
npatient clinical registry data to Medicare claims data using indirect identifiers. Am H
Heart
eart
ea
rtt JJ..
2009;157:995-1000.
11
11.
1. Gr
Gray
Gra
ay R
ay
RJ.
J. A class
cla
lass
ss of K-sample tests for comp
comparing
par
a ing the cumula
cumulative
ati
t ve iincidence
ncid
nc
i ence of a competing
risk.
isk
sk.. Ann
A n St
An
SStatist.
atisst. 19
at
11988;16:1141-1154.
88;1
88
;16:
;1
6::11
1 41
41-1
-1
115
1 4.
12.
Lin
Goodness-of-fit
Cox
regression
model
based
parameter
12
2. L
in DY.. Go
G
odnesss-o
odn
of-fi
fitt anal
aanalysis
nalyssis
i ffor
or thee C
ox re
egr
gres
essi
es
sion
si
on m
odel bas
sedd oon
n a cl
cclass
asss pa
araametterr
estimators.
es
sti
tima
mato
ma
t rs
to
rs.. J Am SStatist
tati
ta
tisst Assoc.
Assoc
occ. 1991;86:725-728.
1991
19
91;8
91
;86:
6 72
6:
7255-72
5-7228.
13. Polsky D
D,, Jh
Jhaa AK
AK,, La
L
Lave
ve JJ,, Pa
Paul
Pauly
u y MV
ul
MV, Ce
Cenn L, Kl
Klus
Klusaritz
u ar
arit
itzz H, C
it
Chen
henn Z,
he
Z Vo
Volp
Volpp
lppp KG
lp
KG. Sh
S
Shortort- and
long-term
ong-ter
term
m mortality
mort
mo
rtal
alit
ityy after
afte
af
terr an acute
acu
cute
te illness
ill
llne
ness
ss for
for elderly
eld
lder
erly
ly whites
whi
hite
tess and
and blacks.
blac
bl
acks
ks He
Health
Heal
alth
th Serv
Ser
ervv Res.
Ress
Re
2008;43:1388-1402.
14. Volpp KG, Stone R, Lave JR, Jha AK, Pauly M, Klusaritz H, Chen H, Cen L, Brucker N,
Polsky D. Is thirty-day hospital mortality really lower for black veterans compared with white
veterans? Health Serv Res. 2007;42:1613-1631.
15. Rathore SS, Foody JM, Wang Y, Smith GL, Herrin J, Masoudi FA, Wolfe P, Havranek EP,
Ordin DL, Krumholz HM. Race, quality of care, and outcomes of elderly patients hospitalized
with heart failure. JAMA. 2003;289:2517-2524.
16. Roe MT, Chen AY, Thomas L, Wang TY, Alexander KP, Hammill BG, Gibler WB, Ohman
EM, Peterson ED. Predicting long-term mortality in older patients after non–ST-segment
elevation myocardial infarction: the CRUSADE long-term mortality model and risk score. Am
Heart J. 2011;162:875-883.e1.
17. Technical documentation with field numbers for the area resource file (ARF) 2006 release.
17
DOI: 10.1161/CIRCULATIONAHA.113.008370
Penn State Simple Online Data Archive for POPulation Studies.
http://sodapop.pop.psu.edu/codebooks/arf/arf2006.txt. Accessed December 5, 2013.
18. Epstein AJ, Polsky D, Yang F, Yang L, Groeneveld PW. Coronary revascularization trends
in the United States, 2001-2008. JAMA. 2011;305:1769-1776.
19. Serruys PW, Farooq V, Vranckx P, Girasis C, Brugaletta S, Garcia-Garcia HM, Holmes DR
Jr, Kappetein AP, Mack MJ, Feldman T, Morice MC, Ståhle E, James S, Colombo A, Pereda P,
Huang J, Morel MA, Van Es GA, Dawkins KD, Mohr FW, Steyerberg EW. A global risk
approach to identify patients with left main or 3-vessel disease who could safely and
efficaciously be treated with percutaneous coronary intervention: the SYNTAX Trial at 3 years.
JACC Cardiovasc Interv. 2012;5:606-617.
20. Sabin JA, Rivera FP, Greenwald AG. Physician implicit attitudes and stereotypes about race
and quality of medical care. Med Care. 2008;46:678-685.
Downloaded from http://circ.ahajournals.org/ by guest on June 14, 2017
21. Hajjar I, Miller K, Hirth V. Age-related bias in the management of hypertension: a national
survey
urvey of physicians' opinions on hypertension in elderly adults. J Gerontol A Bio
Biol
ol Sc
Scii Me
Medd Sc
Sci.
2002;57:M487-M491.
22. Ostchega Y, Dillon CF, Hughes JP, Carroll M, Yoon S. Trends in hypertension prevalence,
awareness, treatment, and control in older U.S. adults: data from the National Health and
Nutrition
Nutr
Nu
trit
tr
itio
it
ionn Ex
io
Exa
Examination
amin
in
natio
ati n Survey 1988 to 2004. J Am Geriatr
Geriatr Soc. 2007;55:1056-1065.
20
007;5
;5
55:1056-1065.
5:
223.
3. C
Chen
hen B, Co
Covi
Covinsky
v nsky
ky K
KE,
E, S
Stijacic
tijjaci
ti
jaciic Ce
C
Cenzer
nzzer II,, A
Adler
dleer N, W
Williams
illliams BA
il
BA. Su
S
Subjective
bjec
eccti
tive
ve ssocial
ocia
oc
iall st
ia
status
tat
atus
uss aand
nd
functional
fu
unccti
t onal ddecline
eclinne in oolder
ec
ldder adu
aadults.
dultss. J Ge
Gen
en Inte
Intern
e rn M
Med.
ed
d. 20
2012;27:693-699.
012
12;2
; 7:69
;2
93--699.
24. Popescu
Popesc
Po
scuu I,, Vaughan-Sarrazin
Vau
ugh
ghan
a -S
Sar
a ra
razi
zinn MS,
M , Rosenthal
MS
Rose
Ro
sent
ntha
hall GE.
GE. Differences
Diff
Di
ffer
eren
ence
c s in mortality
morta
tali
l ty and
and use
se off
evascularizat
attio
ionn in bblack
l ck
la
c aand
nd w
h tee ppatients
hi
atie
at
ieent
ntss wi
with
th
h aacute
c tee M
cu
admi
mitt
mi
t ed tto
o ho
hosp
spit
sp
ittal
alss wi
w
th and
revascularization
white
MII ad
admitted
hospitals
with
with
wi
thou
outt re
reva
vasc
scul
ular
ariz
izat
atio
ionn se
serv
rvic
ices
es JAMA.
JAMA
JA
MA 2007;297:2489-2495.
2007
20
07;2
;297
97:2
:248
4899-24
2495
95
without
revascularization
services.
25. Barnato AE, Lucas FL, Staiger D, Wennberg DE, Chandra A. Hospital-level racial disparities
in acute myocardial infarction treatment and outcomes. Med Care. 2005;43:308-319.
26. Bassand JP, Hamm CW, Ardissino D, Boersma E, Budaj A, Fernández-Avilés F, Fox KA,
Hasdai D, Ohman EM, Wallentin L, Wijns W; Task Force for Diagnosis and Treatment of NonST-Segment Elevation Acute Coronary Syndromes of European Society of Cardiology.
Guidelines for the diagnosis and treatment of non-ST-segment elevation acute coronary
syndromes. Eur Heart J. 2007;28:1598-1660.
27. Vaccarino V, Rathore SS, Wenger NK, Frederick PD, Abramson JL, Barron HV, Manhapra
A, Mallik S, Krumholz HM; National Registry of Myocardial Infarction Investigators. Sex and
racial differences in the management of acute myocardial infarction, 1994 through 2002. N Engl
J Med. 2005;353:671-682.
28. Thomas KL, Honeycutt E, Shaw LK, Peterson ED. Racial differences in long-term survival
18
DOI: 10.1161/CIRCULATIONAHA.113.008370
among patients with coronary artery disease. Am Heart J. 2010;160:744-751.
29. Cohen MG, Fonarow GC, Peterson ED, Moscucci M, Dai D, Hernandez AF, Bonow RO,
Smith SC Jr. Racial and ethnic differences in the treatment of acute myocardial infarction:
findings from the Get With The Guidelines-Coronary Artery Disease program. Circulation.
2010;121:2294-2301.
30. Hochman JS, Lamas GA, Buller CE, Dzavik V, Reynolds HR, Abramsky SJ, Forman S,
Ruzyllo W, Maggioni AP, White H, Sadowski Z, Carvalho AC, Rankin JM, Renkin JP, Steg PG,
Mascette AM, Sopko G, Pfisterer ME, Leor J, Fridrich V, Mark DB, Knatterud GL; Occluded
Artery Trial Investigators. Coronary intervention for persistent occlusion after myocardial
infarction. N Engl J Med. 2006;355:2395-2407.
31. Jencks SF, Williams MV, Coleman EA. Rehospitalizations among patients in the Medicare
fee-for-service program. N Engl J Med. 2009;360:1418-1428.
Downloaded from http://circ.ahajournals.org/ by guest on June 14, 2017
32. Hernandez AF, Greiner MA, Fonarow GC, Hammill BG, Heidenreich PA, Yancy CW,
Peterson ED, Curtis LH. Relationship between early physician follow-up and 30-d
30-day
day
readmission
eadmission among Medicare beneficiaries hospitalized for heart failure. JAMA.. 20
22010;303:1716010
1 ;3
10
303
03:1
:171
:1
7161722.
33. Benatar D, Bondmass M, Ghitelman J, Avitall B. Outcomes of chronic heart failure. Arch
Intern
In
nteern M
Med.
ed.. 20
ed
22003;163:347-352.
03
3;1
;163
6 :347-352.
334.
4. S
Sisk
isk JE, H
Hebert
eb
berrt PL,
PL Horowitz
Horo
Ho
ro
owi
witz
tz C
CR,
R, McLaughlin
McLau
aughliin MA
au
MA,
A, Wang
Wanng
ng JJJ,
J, C
Chassin
hasssin
ha
nM
MR.
R. E
Effects
ffec
ff
e ts ooff nu
ec
nnurse
urrse
management
m
annagement
na
nt oon
n th
the qu
quality
ualityy of hea
heart
e rtt ffailure
ailu
ure ccare
arre inn mino
m
minority
ino
nori
rity
ri
ty ccommunities:
om
mmunitiiess: a ra
randomized
and
n om
mized ttrial.
riall.
Annn In
An
Inte
Intern
tern
rn Med.
Med
ed.. 2006;145:273-283.
20
0066;1
;145
45:2
273
7 -2
-283
83.
83
19
DOI: 10.1161/CIRCULATIONAHA.113.008370
Table 1. Baseline Characteristics by White and Black Race
Downloaded from http://circ.ahajournals.org/ by guest on June 14, 2017
Demographics
Age (yrs)*
Female sex
BMI (kg/m2)*
Medical history
Diabetes mellitus
Hypertension
Prior CABG
Prior PCI
Prior HF
Currently on dialysis
Prior PAD
Current/recent smoker
Dyslipidemia
Prior MI
Prior stroke
No admissions within past 1 year#
Presentation characteristics
Signs
Sign
gns of H
HF
F
ECG
EC
CG changes
chan
ch
han
ngees†
Sysstolic blood
Sy
blo
ood
o pressure
pre
r ss
ssur
uree (mmHg)
ur
(mmH
(m
mH
Hg)*
Systolic
Heart
Heear
a t rate on
n presentation
prresen
enttati
tion
ti
on (bpm)
(bp
bpm)
m))*
‡
Creatinine
Crea
Cr
e tinine clearance
cle
l arancee (mL/min)
(m
mL/m
mL/m
min)
in)‡*
Initial
In
nit
i ia
iall serum
seeru
rum
m creatinine
crreaati
tini
niinee (mg/dL)
(mg
g/d
/dL)
L)‡*
Peak
Pe
eak troponin
tro
opo
poniin (xULN)
( UL
(x
ULN)
N)
nsurance st
tattus§
Insurance
status
Medicare
Medi
Me
dica
care
re
HMO/PVT
Medicaid
Self/none
Hospital region
West
Northeast
Midwest
South
Academic hospitalŒ
White (n=37,384)
Black (n=3,116)
p-value
78 (72,84)
17,515, (46.9%)
26.7 (23.5,30.5)
76 (70,82)
1,838 (59.0%)
27.1 (23.5,31.6)
<0.001
<0.001
<0.001
12,654 (33.8%)
28,215 (75.5%)
8,836 (23.6)
7,786 (20.8%)
8,348 (22.3%)
731 (2.0%)
5,625 (15%)
4646 (12.4%)
19,442 (52.0%)
11,362 (30.4%)
5,009 (13.4%)
23,716 (63.4%)
1,421 (45.6%)
2,703 (86.7%)
480 (15.4%)
496 (15.9%)
940 (30.2%)
254 (8.2%)
423 (13.6%)
490 (15.7%)
1,385 (44.4%)
971 (31.2%)
544 (17.5%)
1,639 (52.6%)
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
0.06
<0.001
<0.001
00.25
.255
.2
<0
0.001
.0001
<0.001
<
<0
<0.001
0.001
001
11,5655 (3
((30.9%)
0.9%)
11,6566 ((31.2%)
31.2%)
1143
43 ((122,164)
12
22,16
64)
8844 (70
(70,101)
0,1101)
1)
4411 (2
(29,55)
29,55
55))
11.2
.22 (0.9,
(0..9,, 1.
(0
1.5)
.5)
15.55 ((4.6,57.1)
15
4 6,,57.1)
4.
1)
1,078 (3
((34.6%)
4 6%)
4.
8880
80 ((28.3%)
28.3%)
28
1149
49 ((124,174)
12
24,
4 17
74)
87
87 (73,103)
(7
73,,1003)
3)
336
6 (2
(25,50)
25,500)
25,5
11.3
.3
3 ((1.0,1.7)
1.0,
0,,1.
1.7)
7)
10.77 ((3.4,37.4)
10
3 4,
3.
4,37
3 .44)
<0.001
0.006
<
<0.001
0.00
0.
0011
00
<
<0.001
0.00
0011
00
<
<0.001
0.0001
<0
<0.001
0.0001
0 01
<0.0
<0
<0.001
001
<
<0
<0.001
.001
27
27,949
949 ((74.8
74 8 %
%))
9,044 (24.2%)
77 (0.2%)
179 (0.5%)
22,506
5066 (8
50
(80
(80.4%)
0 4%
4%))
527 (16.9%)
39 (1.3%)
33 (1.1%)
<0.001
2850 (7.6%)
8862 (23.7%)
14574 (39.0%)
10914 (29.2%)
9907 (26.5%)
124 (4.0%)
404 (13.0%)
1366 (43.8%)
1209 (38.8%)
1606 (51.5%)
<0.001
BMI indicates body mass index; bpm, beats per minute; CABG, coronary artery bypass graft surgery; ECG,
electrocardiogram; HF, heart failure; HMO/PVT, health maintenance organization/private; MI, myocardial
infarction; PAD, peripheral artery disease; PCI, percutaneous coronary intervention; xULN, times the upper limit of
normal. *Continuous variables displayed as medians with 25th, 75th percentiles; †Transient ST elevations or ST
depressions; ‡Excludes dialysis patients; §Patients may have multiple insurance plans; ŒMembership in the Council
of Teaching Hospitals; #From the index admission
20
DOI: 10.1161/CIRCULATIONAHA.113.008370
Table 2. Medications and In-hospital Procedures Among Eligible Patients by White and Black
Race
Downloaded from http://circ.ahajournals.org/ by guest on June 14, 2017
Acute medications*
Aspirin
Any heparin
Beta-blocker
Clopidogrel
GP IIb/IIIa
Invasive procedures
Diagnostic catheterization
PCI
CABG
Discharge medications
Aspirin
Beta-blocker
Lipid lowering agent†
ACEI
ACEI‡
Clopidogrel
White (n=37,384)
Black (n=3,116)
p-value
33,096 (95.1%)
29,533 (87.9%)
29,042 (90.0%)
16,487 (52.5%)
11,720 (41.0%)
2,791 (94.9%)
2,499 (87.8%)
2,453 (88.8%)
1,213 (45.0%)
792 (34.0%)
0.60
0.78
0.043
<0.001
<0.001
22,687 (60.7%)
12,018 (32.1%)
3,422 (9.2%)
1,684 (54.0%)
743 (23.8%)
178 (5.7%)
<0.001
<0.001
<0.001
27,049 (94.1%)
26,463 (92.2%)
17,206 (86.3%)
16,114 (61.6%)
14,641 (63.5%)
18,742 (69.4%)
2,367 (93.9%)
2,319 (92.1%)
1,358 (86.4%)
1,463 (64.1%)
1,404 (65.0%)
1,452 (61.4%)
0.47
0.71
00.93
.9
93
00.011
.01
01
11
00.14
.1
14
<0.001
ACEI indicates ang
angiotensin-converting
giotensin-converting enzymes inhibitor,, GP, glycoprotein; All other abbreviations can be found in
Table
Ta
abl
blee 1.
*
Administered
Admi
Ad
mini
mi
nist
s ered
ed within
withi
hinn 24 hours
hou
ours
rs of
of ho
hospital
osp
spittal
a arrival
arriv
i all
†
Among
Amon
Am
ong patients
patien
nts w
with
ithh do
it
documented
ocume
cume
ment
nteed
ed hhyperlipidemia
yper
yp
e li
er
lipi
pide
pi
d mi
de
m a an
and/
and/or
d/or
/or m
measured
eaasure
redd low-density
re
lowlo
w-de
wdens
nsityy lipoprotein
ns
lipo
li
popr
po
prooteiin >1
>100
00 m
mg/dL
g//dL
‡
Among
Am
mon
o g patients wi
with
ith ejec
ejection
ctioon ffraction
ract
ra
ctiion
io <4
<40
<40%,
0% hheart
0%,
eart ffailure,
aiilure,, diabetes
diabeete
tess mellitus,
me littuss, or
mell
or hypertension
hypper
erttens
nssio
ionn
Figure
Figu
Fi
gure
gu
ree Legends:
Leg
egen
e ds
en
ds::
Figure 1. Mortality Estimates During Follow-up Period From Index Admission by Race.
Cumulative Kaplan-Meier mortality estimates during the follow-up period from the index
admission by black and white race.
Figure 2. Unadjusted and Adjusted Mortality by Race. A forest plot of unadjusted and adjusted
mortality by black and white race. Unadjusted model: black versus white race. Model 1: age,
weight, sex, prior stroke, diabetes, peripheral artery disease, hypertension, dyslipidemia, prior
percutaneous coronary intervention; prior MI, prior coronary artery bypass graft surgery, current
21
DOI: 10.1161/CIRCULATIONAHA.113.008370
or recent smoker, prior HF, signs of HF, family history of coronary artery disease, systolic blood
pressure and heart rate on admission, initial serum creatinine, initial hematocrit, initial troponin
ratio, and electrocardiogram findings. Model 2: Model 1 + additional insurance coverage. Model
3: Model 2 + income level. Model 4: Model 3 + discharge medications. Model 5: Model 4 + inhospital revascularization. *30 days from index admission; data is displayed as hazard ratio (95%
confidence interval). †Long-term among survivors who had not been readmitted within the first
30 days post-discharge; data is displayed as hazard ratio (95% confidence interval). HF indicates
heart failure; MI, myocardial infarction.
Downloaded from http://circ.ahajournals.org/ by guest on June 14, 2017
Figure 3. Readmission Estimates During Follow-up Period From Post-dischargee bby
y Ra
Race
ce.
e.
Race.
Readmission estimates during the follow-up period from post-discharge by black and white race
byy cumulative
cumul
ulat
lat
ativ
ivee in
iv
nci
ciddence of: (A) all-cause readmi
miss
mi
ssiion; (B) MI re
ead
a mi
miss
ssiion;
ss
io and (C) HF
incidence
readmission;
readmission;
ead
dmi
m ssion. HF iindicates
nddic
i attess hheart
eart
ea
rt ffailure;
ailu
ai
l re;; MI
lu
I, m
yoocardiiall iinfarction.
nfarct
nfa
ctiion.
readmission.
MI,
myocardial
Figure 4. Un
nad
adju
just
ju
sted
st
e and
ed
and
n A
djus
dj
usste
tedd Re
Read
admi
ad
miss
mi
sssio
on by R
ace.
ac
e. A fforest
ores
or
e t pl
plot
ot ooff un
uunadjusted
adju
ad
just
ju
sted
st
e and
ed
Unadjusted
Adjusted
Readmission
Race.
adjusted all-cause, MI, and HF readmission by black and white race. Unadjusted model: black
versus white race. Model 1: age, weight, sex, prior stroke, diabetes, peripheral artery disease,
hypertension, dyslipidemia, prior percutaneous coronary intervention, prior MI, prior coronary
artery bypass graft surgery, current or recent smoker, prior HF, signs of HF, family history of
coronary artery disease, systolic blood pressure and heart rate on admission, initial serum
creatinine, initial hematocrit, electrocardiogram findings, number of prior admissions within 1
year from the index admission, and hospital transfer-in status. Model 2: Model 1 + additional
insurance coverage. Model 3: Model 2 + income level. Model 4: Model 3 + discharge
22
DOI: 10.1161/CIRCULATIONAHA.113.008370
medications. Model 5: Model 4 + in-hospital revascularization. *30 days from post-discharge;
data is displayed as hazard ratio (95% confidence interval). †Long-term among survivors who
had not been readmitted within the first 30 days post-discharge; data is displayed as hazard ratio
(95% confidence interval). HF indicates heart failure; MI, myocardial infarction.
Downloaded from http://circ.ahajournals.org/ by guest on June 14, 2017
23
Downloaded from http://circ.ahajournals.org/ by guest on June 14, 2017
Figure 1
Downloaded from http://circ.ahajournals.org/ by guest on June 14, 2017
Figure 2
Downloaded from http://circ.ahajournals.org/ by guest on June 14, 2017
Figure 3
Downloaded from http://circ.ahajournals.org/ by guest on June 14, 2017
Figure 4
Differences in Short-Term Versus Long-term Outcomes of Older Black versus White Patients
with Myocardial Infarction: Findings from CRUSADE
Robin Mathews, Anita Y. Chen, Laine Thomas, Tracy Y. Wang, Chee Tang Chin, Kevin L. Thomas,
Matthew T. Roe and Eric D. Peterson
Downloaded from http://circ.ahajournals.org/ by guest on June 14, 2017
Circulation. published online July 7, 2014;
Circulation is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231
Copyright © 2014 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/2014/07/07/CIRCULATIONAHA.113.008370
Data Supplement (unedited) at:
http://circ.ahajournals.org/content/suppl/2014/07/07/CIRCULATIONAHA.113.008370.DC1
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/
SUPPLEMENTAL MATERIAL
Supplemental Table 1. Subgroup Analysis for Short- and Long-term Adjusted Mortality by
Mortality (Black vs. White)
Short-term
HR
95% CI
Long-term
p-value
HR
95% CI
(interaction)
≥75 years old
0.82
0.70,0.95
<75 years old
0.70
0.55,0.90
Diabetes
0.79
0.65,0.97
No diabetes
0.81
0.69,0.96
Females
0.85
0.72,1.00
Males
0.74
0.60,0.91
High income
0.74
0.59,0.92
Low income
0.84
0.72,0.98
Race
0.30
0.87
0.28
0.34
p-value
(interaction)
0.91
0.84,0.98
1.11
0.99,1.24
0.94
0.85,1.03
1.07
0.99,1.16
1.04
0.96,1.13
0.95
0.87,1.04
0.97
0.88,1.07
1.03
0.94,1.12
0.004
0.030
0.12
0.40
CI indicates confidence interval; HR, hazard ratio
Supplemental Table 2. Subgroup Analysis of Short- and Long-term Adjusted Readmission
by Race
Readmission (Black vs. White)
Short-term
HR
Long-term
95% CI
p-value
HR
95% CI
(interaction)
p-value
(interaction)
All-cause
≥75 years old
1.04
0.92,1.16
<75 years old
0.99
0.85,1.16
Diabetes
1.00
0.87,1.14
No diabetes
1.05
0.93,1.18
Females
1.02
0.92,1.13
Males
1.03
0.88,1.21
High income
1.09
0.96,1.24
Low income
0.97
0.83,1.13
≥75 years old
1.08
0.76,1.52
<75 years old
1.21
0.75,1.94
Diabetes
1.05
0.77,1.44
No diabetes
1.23
0.72, 2.10
Females
1.13
0.84,1.51
Males
1.17
0.66,2.08
High income
0.92
0.61,1.38
Low income
1.27
0.86,1.89
≥75 years old
0.98
0.77,1.27
<75 years old
1.03
0.81,1.32
Diabetes
0.89
0.68,1.16
No diabetes
1.20
0.96,1.50
Females
0.83
0.65,1.06
0.62
0.49
0.85
0.25
1.05
0.98,1.12
1.03
0.94,1.12
1.04
0.95,1.13
1.06
0.98,1.15
1.05
0.97,1.14
1.04
0.97,1.12
1.11
1.02,1.22
1.00
0.94,1.07
1.07
0.92,1.24
1.10
0.83,1.46
1.08
0.91,1.29
1.12
0.90,1.40
1.06
0.90,1.25
1.17
0.92,1.48
0.98
0.81,1.18
1.19
0.95,1.48
1.08
0.96,1.21
1.18
1.00,1.39
0.99
0.87,1.12
1.35
1.16,1.57
1.07
0.94,1.21
0.75
0.75
0.90
0.08
Myocardial infarction
0.59
0.59
0.89
0.23
0.82
0.73
0.39
0.18
Heart failure
0.77
0.08
<0.001
0.40
0.005
0.09
Males
1.38
1.10,1.74
High income
1.11
0.85,1.44
Low income
0.94
0.73,1.22
CI indicates confidence interval; HR, hazard ratio
0.40
1.25
1.09,1.42
1.16
1.02,1.33
1.12
0.99,1.27
0.70