Download Effect of anemia on 1-year mortality in patients with acute myocardial

Effect of anemia on 1-year mortality in patients
with acute myocardial infarction
Nezar Al Falluji, MD, Janet Lawrence-Nelson, PhD, John B. Kostis, MD, Clifton R. Lacy, MD, Rajiv Ranjan, MD,
and Alan C. Wilson, PhD, for the Myocardial Infarction Data Acquisition system (MIDAS #8) Study Group New
Brunswick, NJ
Background Limited data are available on the effect of anemia on mortality in patients with acute myocardial infarction (MI).
Methods We examined the association of anemia with mortality at 1 year among 30,341 patients hospitalized with
acute MI in 1986 (prethrombolytic era, n ⫽ 15,584) and 1996 (thrombolytic era, n ⫽ 14,757). The records were obtained from the Myocardial Infarction Data Acquisition System, a database of all patients with MI admitted to nonfederal
hospitals in New Jersey.
Results Anemia was present in 996 patients (6.4%) in 1986 and 1510 patients (10.2%, P ⬍.0001) in 1996. In both
years, patients with anemia were older, more frequently female and nonwhite, and more likely to have left ventricular dysfunction, non-Q MI and coronary artery bypass graft. In addition, in 1996, patients with anemia were more likely to undergo percutaneous transluminal coronary angioplasty and less likely to have a history of MI. One-year mortality was
lower overall in 1996 compared with 1986 (1996 23.6%, 95% CI 22.9-24.3 vs 1986 24.9%, 95% CI 24.2-25.6, P ⫽
.0001). In both years, patients with anemia had significantly higher unadjusted risk for 1-year mortality (RR ⫽ 1.40, P ⫽
.0001 in both years). However, after controlling for demographics, left ventricular dysfunction, arrhythmias, Q versus
non-Q MI, comorbid conditions, and revascularization procedures in a multivariable regression model, 1-year mortality in
the anemia group was similar to the nonanemia group in both years.
Conclusion In the Myocardial Infarction Data Acquisition System database, anemia appears to have no significant
direct effect on 1-year mortality. The higher unadjusted mortality observed among patients with acute MI and anemia is
probably the result of older age, higher comorbidity, and more left ventricular dysfunction. (Am Heart J 2002;144:
636-41.)
The association of anemia with short- and long-term
mortality has been studied in patients with end-stage
renal diseases, post-coronary artery bypass graft
(CABG), in critically ill patients, and perioperatively,
but not in patients with acute coronary syndromes.
One recent study of Medicare beneficiaries with acute
myocardial infarction (MI) found that transfusing patients with hematocrit values below 30% was associated with reduced 30-day mortality.1 The prognostic
importance of anemia in the setting of acute MI is not
well defined, and the impact of anemia on long-term
survival after MI has not been studied.
A prospective observational study of 2202 patients
undergoing CABG in the United States found higher
From the University of Medicine and Dentistry of New Jersey-Robert Wood Johnson
Medical School, New Brunswick, NJ.
Submitted July 27, 2001; accepted February 20, 2002.
Reprint requests: John B. Kostis, MD, UMDNJ-Robert Wood Johnson Medical School,
One Robert Wood Johnson Pl, PO Box 19, New Brunswick, NJ 08903-0019.
E-mail: kostis @ umdnj.edu
© 2002, Mosby, Inc. All rights reserved.
0002-8703/2002/$35.00 ⫹ 0 4/1/124351
doi:10.1067/mhj.2002.124351
postoperative hematocrit values (⬎34%) were associated with higher rates of postoperative MI, left ventricular dysfunction (LVD), and mortality.2 In patients
with end-stage renal disease, anemia is associated with
the development of left ventricular hypertrophy.2 Correction to near-normal hemoglobin values3 or to values
above 10 g/dL4 was associated with fewer cardiovascular events, longer duration to first MI, regression of left
ventricular hypertrophy, and decreased resting and
exercise induced angina. However in a randomized,
open-label prospective trial with 29-month follow-up
of 1233 patients with end-stage renal disease and history of heart disease, administration of erythropoietin
to maintain a normal hematocrit (⬎42%) was associated with higher rates of acute nonfatal MI and a
higher mortality rate compared to low hematocrit of
30%.5 The above studies assessed the association of
anemia with short-term mortality and were not conducted in the setting of acute MI.
This retrospective cohort study was conducted to
evaluate the effect of anemia on 1-year mortality in
patients with acute MI admitted to nonfederal hospitals in New Jersey. Two patient cohorts were investi-
American Heart Journal
Volume 144, Number 4
gated: from 1986 (prethrombolytic era) and 1996
(thrombolytic era), as this database lacks information
about the use of thrombolytic therapy. This is a major
confounder because thrombolytic use is associated
with improved survival as well as with a greater
chance of developing anemia during hospitalization.
Methods
Data source and study population
The study was conducted with the use of the Myocardial
Infarction Data Acquisition System (MIDAS).6 This statewide
database provided discharge abstract information on index
admissions of all patients (n ⫽ 30,341) who were hospitalized with the diagnosis of acute MI in nonfederal hospitals in
New Jersey in the years 1986 (n ⫽ 15,584) and 1996 (n ⫽
14,757). Patient discharge abstract data (UB-82 and UB-92)
were obtained from the New Jersey hospital discharge data
system. Hospital discharge records from administrative files
were included in the MIDAS database if an International Classification of Diseases, 9th Revision, diagnosis code in the
range 410.0-410.9 (acute MI) was either the principal diagnosis or one of the secondary diagnoses. The MIDAS records
included codes for procedures and comorbid conditions, and
other relevant patient information such as age, sex, race, insurance status, admission, and discharge dates. Composite
indices of complications were constructed for LVD and arrhythmia.6 Anemia was defined as the presence of the disease codes 280.00-281.30, 281.90, 282.00, and 283.00-285.99.
The confidentiality of hospital patient records was maintained and the State and Robert Wood Johnson Institutional
Review Boards approved MIDAS.
Survival follow-up was performed by matching the MIDAS
records with New Jersey death registration files, through the
use of specifically designed automated record linkage software (Automatch, Matchware Technologies, Inc, Burtonsville,
Md).6,7
Patients with the diagnosis of cancer (International Classification of Diseases-9 codes 140.00-208.00) and those hospitalized in another state or federal hospital were excluded from
the study. Only the first recorded (index) infarction that occurred in this database was considered.
Statistical analysis
Key baseline demographic and clinical characteristics
among patients with and without anemia were compared by
the ␹2 test for proportions and the t test for means for each
year. Strength of association of the presence of anemia with
categorical variables (eg, sex, race, age group, etc) was analyzed with the use of the Mantel-Haenszel ␹2 relative risk. A
probability level ⬍.05 was considered significant. Comparison of the effect of anemia on survival at 1-year was performed for each year with the use of the LIFETEST procedure (SAS, Cary, NC) and the log-rank statistic. Univariate
Cox regression analyses were used to identify significant explanatory variables for inclusion in a multivariable Cox regression model. If a variable was significantly associated with
1-year mortality in either year, it was included. To avoid colinearity problems, prior MI was not included because it is
used in the composite LVD index.
Al Falluji et al 637
Data auditing
A random sample of the MIDAS database underwent chart
audit for accuracy of MI diagnosis and other variables as reported elsewhere. The diagnosis of MI was verified for 91%
of the women and 95% of the men audited. The database
information was correct in 99.3% of the records audited for
sex, 99.7% for age, 98.8% for race, 99.3% for length of stay,
98.8% for vital status at discharge and 98.8% for procedures.6
To validate the diagnosis of anemia in MIDAS, the 94
charts with the diagnosis of anemia (approximately 10% of
the audited charts) and an equal number of control charts
(selected as the chart after each anemia chart in the order of
filing) were audited. The mean hemoglobin level, on admission, of the patients in the anemia group was 10.5 g/dL and
of the control group on admission, 14.3 g/dL. In the control
group, anemia on admission (hemoglobin below normal laboratory reference values) was noted in 7 charts. The remaining
87 (93%) patients were found to have no anemia on admission. In the anemia group, 81 charts of the total 94 audited
charts (86.2%) were found to have had anemia by the same
criteria. The sensitivity and specificity of the diagnosis of anemia on admission in MIDAS were 92% and 87%, respectively.
The sensitivity and specificity of anemia diagnosis in MIDAS
in detecting anemia at any time during the admission were
86% and 99%, respectively.
Results
Anemia was included among the discharge diagnoses
of 996 of the 15,584 MI patients (6.4%) in 1986 and
1510 of 14,757 patients (10.2%) in 1996 (Table I). The
mean length of hospital stay was longer in the anemia
group (13.0 ⫾ 6.2 vs 10.3 ⫾ 5.5 days, P ⫽ .001 in
1986 and 9.5 ⫾ 5.8 vs 6.4 ⫾ 4.7 days, P ⫽ .001 in
1996) (Table I). Patients with anemia were older, less
likely to be white, and more likely to be female and to
have LVD, non-Q MI and CABG in both years (Table
II). In addition, in 1996, patients with anemia were
less likely to have prior MI and more likely to have
undergone percutaneous transluminal coronary angioplasty (PTCA).
One-year mortality was lower overall in the 1996
cohort compared with 1986 (1996 23.6%, 95% CI 22.924.3, vs 1986 24.9%, 95% CI 24.2-25.6, P ⫽ .0001).
One-year mortality in the nonanemia group was significantly lower in 1996 (22.7%, 95% CI 22.0 —23.5) compared with 1986 (24.3%, 95% CI 23.6-25.0, P ⫽ .0002).
There was a similar mortality decrease in the anemia
group in 1996, but it did not reach statistical significance (30.9%, 95% CI 28.5-33.2, in 1996 vs 33.5%, 95%
CI 30.6-36.5, in 1986, P ⫽ .16) (Figure 1).
Univariate Cox analysis revealed that age, female sex,
LVD, arrhythmia, diabetes, anterior wall MI, chronic
obstructive pulmonary disease (COPD), and anemia
were all associated significantly with higher 1-year
mortality in both 1986 and 1996 (Table III). The magnitude of the increase in 1-year mortality in patients
with anemia was similar in the 2 years studied (1986
American Heart Journal
October 2002
638 Al Falluji et al
Table I. Baseline characteristics of the cohort used to study anemia
1986
Variable
All (n)
White (%)
Female (%)
Age (mean ⫾ SD) (%)
⬍65
65-75
⬎75
LVD* (%)
Arrhythmia* (%)
Anterior MI (%)
Non-Q MI (%)
Prior MI (%)
Diabetes (%)
COPD (%)
Chronic liver disease (%)
Hypertension (%)
CATH (%)
PTCA (%)
CABG (%)
LOS (mean ⫾ SD)
1996
Anemia
No anemia
Delta
P
996
851 (85.4)
552 (55.4)
72.7 ⫾ 12.1
232 (23.3)
318 (31.9)
446 (44.8)
518 (52.0)
447 (44.9)
322 (32.3)
235 (23.6)
84 (8.4)
219 (22.0)
113 (11.4)
9 (0.9)
311 (31.2)
88 (8.8)
17 (1.7)
19 (1.9)
13.0 ⫾ 6.2
14,588
12,972 (88.9)
5450 (37.4)
66.2 ⫾ 12.9
6301 (43.2)
4682 (32.1)
3604 (24.7)
5955 (40.8)
6771 (46.4)
5080 (34.8)
2626 (18.0)
1431 (9.8)
3332 (22.8)
1521 (10.4)
38 (0.3)
5096 (34.9)
1345 (9.2)
183 (1.3)
104 (0.7)
10.3 ⫾ 5.5
⫺3.5
18.0
6.5
⫺19.9
⫺0.2
20.1
11.2
⫺1.5
⫺2.5
5.6
⫺1.4
⫺0.8
1.0
0.6
⫺3.7
⫺0.4
0.4
1.2
2.7
.001
.001
.001
.001
.001
.34
.11
.001
.16
.54
.36
.001
.02
.68
.22
.001
.001
Anemia
No anemia
Delta
P
1510
1231 (83.1)
825 (54.6)
73.4 ⫾ 12.5
349 (23.1)
421 (27.9)
740 (49.0)
751 (49.7)
571 (37.8)
382 (25.3)
604 (40.0)
66 (4.4)
402 (26.6)
234 (15.5)
9 (0.6)
738 (48.9)
457 (30.3)
173 (11.5)
189 (12.5)
9.5 ⫾ 5.8
13,247
11,272 (87.6)
5137 (38.8)
67.9 ⫾ 13.6
5057 (38.2)
3929 (29.7)
4261 (32.2)
4616 (34.8)
5080 (38.4)
3547 (26.8)
4835 (36.5)
935 (7.1)
3585 (27.1)
1815 (13.7)
41 (0.3)
6370 (48.1)
3588 (27.1)
1293 (9.8)
509 (3.8)
6.4 ⫾ 4.7
⫺4.5
15.8
5.5
⫺15.1
⫺1.8
16.8
⫺14.9
⫺0.6
⫺1.5
3.5
⫺2.7
⫺0.5
1.8
0.3
0.8
3.2
1.7
8.7
3.1
.001
.001
.001
.001
.001
.69
.22
.008
.001
.72
.06
.07
.56
.001
.04
.001
.001
*LVD and Arrhythmia denotes composite indices of mechanical and electrical complications. The first, left ventricular dysfunction, included presence of any of the following:
old MI (412.00), congestive heart failure (428.00), left heart failure (428.10), cardiomegaly (429.30), alcoholic cardiomyopathy (425.50), rupture of papillary muscle
(429.60), rupture of chordae tendineae (429.50), acquired cardiac septal defects (429.71), hypertensive heart disease with CHF (402.91), cardiogenic shock (785.51), or
ventricular aneurysm (414.10). The second, electrical instability (arrhythmia), included presence of cardiac dysrhythmias (427.00), complete atrioventricular block (426.00),
unspecified atrioventricular block (426.10), Mobitz (type II) atrioventricular block (426.12), left bundle branch block (426.20), other left bundle branch block (426.30), right
bundle branch block (426.40), other and unspecified bundle branch block (426.50), other heart block (426.6), or other specified and unspecified conduction disorders
(426.90).
RR ⫽ 1.395, 95% CI 1.247-1.560, and 1996 RR ⫽
1.398, 95% CI 1.268-1.541, P ⫽ .0001 both years).
However, in a muitivariable Cox proportional hazard
regression model controlling for demographics, LVD,
arrhythmias, Q versus non-Q MI, comorbid conditions,
and revascularization procedures, the 1-year mortality
in the anemia groups was not significantly higher than
that of the nonanemia groups in either year (RR ⫽
1.005, P ⫽ .93 in 1986 and RR ⫽ 1.084, P ⫽ .11 in
1996) (Table IV). A significant interaction between
anemia and the LVD index was observed, but inclusion
of this term in the multivariate model did not improve
the model score significantly. Age, female sex, LVD,
arrhythmia, diabetes, and chronic liver disease remained significantly associated with a higher 1-year
mortality after adjustment in both years.
Discussion
The impact of anemia in patients with acute MI has
not been well characterized, as most previous investigations were small, had short duration follow-up, or
focused on a specific group of patients.1-5,8,9 In this
study, anemia was present in a higher percentage of
patients with MI in 1996 than 1986, probably because
of higher rates of CABG, use of primary PTCA, throm-
bolytic therapy, and older age of patients with acute
MI.
Anemia was associated with higher unadjusted mortality rate, but this effect was lost in analyses controlling for demographics, comorbidities, complications,
and revascularization. Thus, anemia predicted a negative outcome in patients through its association with
other factors conferring high risk. Age, LVD, female
sex, nonwhite race, and Q wave MI were associated
with higher mortality in this and previous studies.6,10-12
Anemia was also associated with revascularization
procedures, which may have given patients who developed anemia a survival advantage. However, only
14.4% of the anemia cases in MIDAS-audited charts had
anemia develop during the hospitalization. Thus, the
majority of the patients with anemia had anemia on
admission.
Anemia may affect prognosis in acute MI in opposite
directions. High hematocrit values may improve the
short-term survival by improving the oxygen carrying
capacity and counteracting free radical stress.13-16 In
animal models, anemic dogs showed ischemic ST-segment changes and locally depressed cardiac function at
higher hemoglobin levels with experimentally created
coronary stenosis varying from 50% to 80%.17 In a ret-
American Heart Journal
Volume 144, Number 4
Al Falluji et al 639
Table II. Association of the presence of anemia with various patient characteristics
1986 (n ⴝ 15, 584)
n (%)
Males
Females
White
Nonwhite
Age (y)
⬍65
ⱖ65
LVD*
No LVD
Arrhythmia*
No arrhythmia
Anterior MI
No anterior MI
Non-Q MI
Q-Wave MI
Prior MI
No prior MI
Diabetes
No diabetes
COPD
No COPD
Chronic liver disease
No chronic liver disease
Hypertension
No hypertension
CATH
No CATH
PTCA
No PTCA
CABG
No CABG
1996 (n ⴝ 14, 757)
P
RR
95% CI
n (%)
P
RR
95% CI
444 (4.6)
552 (9.2)
851 (6.2)
145 (8.2)
.001
0.504
0.448-0.567
.001
0.563
0.512-0.619
.001
0.748
0.631-0.886
685 (7.8)
825 (13.8)
1231 (9.9)
251 (13.6)
.001
0.725
0.638-0.824
232 (3.6)
764 (8.4)
518 (8.0)
478 (5.3)
447 (6.2)
549 (6.6)
322 (6.0)
674 (6.6)
235 (8.2)
761 (6.0)
84 (5.5)
912 (6.5)
219 (6.2)
777 (6.5)
113 (6.9)
883 (6.3)
9 (19.2)
987 (6.4)
311 (5.8)
685 (6.7)
88 (6.1)
908 (6.4)
17 (8.5)
979 (6.4)
19 (15.5)
977 (6.3)
.001
0.421
0.367-0.483
.001
0.520
0.465-0.581
.001
1.525
1.354-1.719
.001
1.731
1.575-1.904
.347
0.944
0.836-1.065
.686
0.980
0.888-1.081
.11
0.900
0.792-1.024
.218
0.933
0.836-1.042
.001
1.373
1.193-1.581
.008
1.142
1.036-1.259
.158
0.855
0.689-1.062
.001
0.630
0.494-0.802
.535
0.955
0.826-1.104
.715
1.023
0.907-1.154
.36
1.093
0.904-1.320
.056
1.137
0.997-1.298
.001
3.014
1.648-5.514
.069
1.764
0.956-3.254
.017
0.855
0.751-0.973
.562
1.029
0.935-1.132
.684
0.957
0.774-1.183
.009
1.149
1.036-1.275
.22
1.336
0.841-2.121
.037
1.173
1.010-1.363
.001
2.445
1.598-3.739
349 (6.5)
116 (12.4)
751 (13.9)
759 (8.1)
571 (10.1)
939 (10.3)
382 (9.7)
1128 (10.4)
604 (11.1)
906 (9.7)
60 (6.6)
1450 (10.5)
402 (10.1)
1108 (10.3)
234 (11.4)
1276 (10.1)
9 (18.0)
1501 (10.2)
738 (10.4)
772 (10.1)
457 (11.3)
1053 (9.8)
173 (11.8)
1337 (10.1)
189 (27.1)
1321 (9.4)
.001
2.882
2.511-3.308
*Composite indices
Figure 1
One-year mortality rates for patients with and without anemia after admission to the hospital.
rospective analysis of 1958 surgical patients who refused blood transfusion for religious reasons, patients
with cardiovascular disease had a much greater risk of
death at 30 days than patients without cardiovascular
disease when the preoperative hemoglobin level was
ⱕ10 g/dL.8 These results strongly suggest that patients
with underlying cardiovascular disease are less tolerant
of anemia than patients without cardiovascular disease.
Conversely, high viscosity associated with a high hematocrit value may worsen prognosis.9,14-20 In a retrospective analysis of 8787 patients aged ⱖ60 years undergoing hip fracture surgical repair, hemoglobin
American Heart Journal
October 2002
640 Al Falluji et al
Table III. Univariate Cox proportional hazards regression for one-year mortality
Total
Age
Sex
Race
LVD
Arrhythmia
Diabetes
Non-Q MI
Anterior MI
Prior MI
Hypertension
Anemia
COPD
Chronic liver disease
CATH
PTCA
CABG
1986
1996
P
RR
P
RR
P
RR
.0001
.0001
.0002
.0001
.0001
.0001
.0001
.0001
.075
.0001
.0001
.0001
.0499
.0001
.0001
.0001
1.056
0.613
0.879
3.388
2.083
1.297
0.754
1.168
0.925
0.751
1.380
1.234
1.415
0.336
0.335
0.417
.0001
.0001
.0001
.0001
.0001
.0001
.0001
.0001
.612
.0001
.0001
.0078
.4868
.0001
.0001
.0185
1.057
0.631
0.776
3.104
2.200
1.261
0.586
1.179
1.028
0.686
1.395
1.141
1.205
0.307
0.232
0.582
.0001
.0001
.2676
.0001
.0001
.0001
.0001
.0001
.0001
.0001
.0001
.0001
.0342
.0001
.0001
.0001
1.057
0.591
0.952
3.720
1.955
1.347
0.890
1.136
0.738
0.827
1.398
1.337
1.649
0.335
0.351
0.395
Table IV. Multivariate proportional hazards regression for one-year mortality for 1986 and 1996
1986
Demographics
and comorbidities
Age
Sex
Race
Diabetes
Anterior MI
Non-Q MI
Hypertension
COPD
Chronic liver disease
LVD
Arrhythmia
CATH
PTCA
CABG
Anemia
1996
Adding
severity
Adding
procedures
Demographics
and comorbidities
Adding
severity
Adding
procedures
P
RR
P
RR
P
RR
P
RR
P
RR
P
RR
⬍.0001
.0009
.6436
⬍.0001
.1315
⬍.0001
⬍.0001
.1193
.0594
1.057
0.894
0.974
1.281
1.053
0.512
0.683
1.081
1.658
⬍.0001
⬍.0001
.7206
.0001
.4477
⬍.0001
⬍.0001
.8043
.1315
⬍.0001
⬍.0001
1.047
0.870
0.980
1.223
0.974
0.521
0.711
1.012
1.499
2.172
1.748
1.058
0.883
1.099
1.422
1.020
0.713
0.796
1.201
1.838
⬍.0001
.0017
.0326
⬍.0001
.2859
⬍.0001
⬍.0001
.0253
.0063
⬍.0001
⬍.0001
1.044
0.895
1.257
1.276
0.956
0.673
0.798
1.105
1.909
2.425
1.462
0.996
1.0000
1.000
1.045
0.879
0.981
1.214
0.977
0.525
0.710
1.005
1.458
2.161
1.741
0.516
0.650
1.489
1.005
⬍.0001
⬍.0005
.0489
⬍.0001
.6325
⬍.0001
⬍.0001
⬍.0001
.01
.9404
⬍.0001
⬍.0002
.7288
⬍.0001
.4905
⬍.0001
⬍.0001
.922
.1604
⬍.0001
⬍.0001
⬍.0001
.1329
.0953
.9333
.2951
1.054
.6265
1.025
⬍.0001
.0215
.0097
⬍.0001
.6044
⬍.0001
⬍.0001
.1829
.0218
⬍.0001
⬍.0001
⬍.0001
.0289
.0025
.11
1.038
0.922
1.134
1.236
0.978
0.665
0.798
1.061
1.722
2.411
1.454
0.518
0.817
0.686
1.084
values ⱖ8.0 g/dL did not appear to influence the risk
of 30- and 90-day mortality in that elderly population.9
In a prospective, randomized clinical trial of 838 critically ill patients in Canada, maintaining hemoglobin
values of 10 to 12 g/dL through blood transfusion was
not associated with a survival advantage when compared
with patients with hemoglobin values of 7 to 9 g/dL.13
This study is limited in that it is a retrospective observational analysis, hemoglobin and hematocrit values
are available only in the audited charts, data on the
onset, treatment, duration, and severity of anemia are
not available, and it is based on administrative data.
However, the audit of a random sample of charts
showed good accuracy, and the size of the sample is
large and includes all MIs in the state of New Jersey
(except federal hospitals). Another important limitation in this study was the lack of information about
thrombolytic therapy and the timing of PTCA (primary
vs elective). The presence of anemia on admission may
have influenced the decision with regard to utilization
of either modality in 1996. Neither primary PTCA nor
thrombolytic therapy was used commonly in 1986.
American Heart Journal
Volume 144, Number 4
Thus, it appears that anemia in the degree of severity that is encountered in the community does not
have a significant direct effect on 1-year mortality in
patients with acute MI. The higher unadjusted mortality in patients with anemia is probably caused by association with other factors (LVD, older age, etc) that
confer worse prognosis.
References
1. Wu WC, Rathore SS, Radford M J, et al. Blood transfusion in elderly patients with acute myocardial infarction. New Engl J Med
2001;345:1230-6.
2. Spiess BD, Ley C, Body SC, et al. Hematocrit value on intensive
care unit entry influences the frequency of Q-wave myocardial infarction after coronary artery bypass grafting. The Institutions of
the Multicenter Study of Perioperative Ischemia (McSPI) Research
Group. J Thorac Cardiovasc Surg 1998;116:460-7.
3. Berweck S, Hennig L, Sternberg C, et al. Cardiac mortality prevention in uremic patients. Therapeutic strategies with particular attention to complete correction of renal anemia. Clin Nephrol 2000;
53:S80-5.
4. Wizemann V, Kaufmann J, Kramer W. Effect of erythropoietin on
ischemia tolerance in anemic hemodialysis patients with confirmed
coronary artery disease. Nephron 1992;62:161-5.
5. Besarab A, Bolton WK, Bowne JK, et al. The effects of normal as
compared with low hematocrit values in patients with cardiac disease who are receiving hemodialysis and epoetin. N Engl J Med
1998;339:584-90.
6. Kostis JB, Wilson AC, O’Dowd K, et al. Sex differences in the
management and long-term outcome of acute myocardial infarction. A statewide study. Circulation 1994;90:1715-29.
7. Jaro MA. Advances in record-linkage methodology as applied to
matching the 1985 census of Tampa, Florida. J Am Stat Assoc
1989;84:414-20.
8. Carson JL, Duff A, Poses RM, et al. Effect of anemia and cardio-
Al Falluji et al 641
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
vascular disease on surgical mortality and morbidity. Lancet 1996;
348:1055-60.
Carson JL, Duff A, Berlin JA, et al. Perioperative blood transfusion
and postoperative mortality JAMA 1998;279:199-205.
Herlitz J, Karlson BW, Sjolin M, et al. Prognosis during one year
of follow-up after acute myocardial infarction with emphasis on
mortality. Clin Cardiol 1994;17:15-20.
Maynard C, Every NR, Douglas WW. Factors associated with rehospitalization in patients with acute myocardial infarction. Am J
Cardiol 1997;80:777-9.
Maynard C, for the CASCADE Investigators. Rehospitalization in
surviving patients of out-of-hospital ventricular fibrillation (the
CASCADE study). Am J Cardiol 1993;72:1295-300.
Hebert PC, Wells G, Blajchman MA, et al. A multicenter, randomized, controlled clinical trial of transfusion requirement in critical
care. N Engl J Med 1999;340:409-17.
Rostand SG, Rutsky EA. Cardiac disease in dialysis patients. In:
Nissenson AR, Fine RN, Gentile DE, editors. Clinical dialysis. 3rd
ed. Norwalk, Conn: Appleton and Lange; 1990. p. 652-98.
Macdougall IC, Lewis NP, Saunders MJ, et al. Long-term cardiorespiratory effects of amelioration of renal anemia by erythropoietin.
Lancet 1990;335:489-93.
Parfrey PS, Harnett JD, Barre PE, et al. The natural history of myocardial disease in dialysis patients. J Am Soc Nephrol 1991;2:212.
Yoshikawa H, Powell WJ Jr, Bland JH, et al. The effect of acute
anemia on experimental myocardial ischemia. Am J Cardiol
1973;32:670-8.
Hagl S, Heimisch W, Meisner H, et al. The effect of hemodilution
on regional myocardial function in the presence of coronary stenosis. Basic Res Cardiol 1977;72:344-64.
Geha AS, Baue AE. Graded coronary stenosis and coronary flow
during acute normovolemic anemia. World J Surg 1978;2:64552.
Anderson HT, Kessinger JM, McFarland WJ, et al. Response of the
hypertrophied heart to acute anemia and coronary stenosis. Surgery 1978;84:8-15.