International Journal of Cardiology 168 (2013) 5221–5228

Contents lists available at ScienceDirect

International Journal of Cardiology

journal homepage: www.elsevier.com/locate/ijcard

Prognostic implications of anemia with or without chronic kidney disease

in patients undergoing elective percutaneous coronary intervention☆
Yuichiro Kitai a, Neiko Ozasa b,⁎, Takeshi Morimoto c, Bingyuan Bao a, Yutaka Furukawa d, Yoshihisa Nakagawa e,
Kazushige Kadota f, Motoko Yanagita a, Satoshi Shizuta b, Takeshi Kimura b
on behalf of the CREDO-Kyoto registry investigators
a
Department of Nephrology, Graduate School of Medicine, Kyoto University, Japan
b
Department of Cardiovascular Medicine, Graduate School of Medicine, Kyoto University, Japan
c
Division of General Medicine, Hyogo Medical College, Japan
d
Department of Cardiovascular Medicine, Kobe City Medical Center General Hospital, Japan
e
Division of Cardiology, Tenri Hospital, Japan
f
Division of Cardiology, Kurashiki Central Hospital, Japan

a r t i c l e i n f o a b s t r a c t

Article history: Background: Little is known about the prognostic implications of anemia in patients undergoing elective percu-
Received 26 June 2013 taneous coronary intervention (PCI), especially when they have coexisting chronic kidney disease (CKD).
Accepted 3 August 2013 Methods: We identified 7299 patients who underwent elective PCI from the CREDO-Kyoto registry cohort-2. The
Available online 16 August 2013 primary outcome was 3-year major adverse cardiac events (MACE); composite of all cause death, heart failure
hospitalization, and myocardial infarction.
Keywords:
Results: In total, 1466 patients (20.0%) had mild anemia (hemoglobin = 11.0–11.9 g/dL for women and 11.0–
Anemia
Chronic kidney disease
12.9 g/dL for men), and 740 patients (10.1%) had moderate-to-severe anemia (hemoglobin b 11.0 g/dL both
Coronary artery disease for women and for men). Compared to the no-anemia group, cumulative incidence of MACE was significantly
Prognosis higher in the mild and moderate-to-severe anemia groups (7.9%, 20.1%, and 34.2%, respectively). The adjusted
hazard ratios of mild and moderate-to-severe anemia versus no-anemia for MACE were 1.77 (95% confidence
interval: 1.47–2.15) and 2.53 (95% confidence interval: 2.03–3.14), respectively. In a subgroup analysis, signifi-
cantly higher risk for MACE was consistently observed with mild and moderate-to-severe anemia both in
patients with and without CKD. The risk for MACE showed an accretive increment with exacerbation in either
the renal function or anemia (interaction p b 0.001).
Conclusions: Even mild anemia was associated with significantly worse 3-year clinical outcomes in patients who
underwent elective PCI. Coexisting CKD additively increased the risk for MACE in these patients.
© 2013 Elsevier Ireland Ltd. All rights reserved.

1. Introduction undergoing percutaneous coronary intervention (PCI) [6–8]. However,

the hemoglobin (Hb) level that may affect cardiovascular outcomes
Anemia is a common comorbidity in patients with coronary artery and whether the presence of CKD influences the outcomes in these
disease (CAD) [1]. Its prevalence increases with disease severity as a patients remain obscure. Moreover, little data is available regarding
consequence of associated disorders such as chronic kidney disease the impact of anemia on prognosis in patients undergoing elective PCI.
(CKD), iron deficiency, malnutrition and cancer-related blood losses In the present study, we investigated the relationship between severity
[2]. Anemia may increase the risk of cardiovascular events by inducing of anemia, CKD, and clinical outcomes in patients who underwent elec-
myocardial ischemia [3], congestive heart failure [4], or by contributing tive PCI, analyzing data from a multicenter registry in Japan.
to high residual platelet activity on clopidogrel [5]. Recent cohort stud-
ies have reported worse clinical outcomes among anemic patients
2. Methods

2.1. Study design and patient population

☆ This study was supported by the Pharmaceuticals and Medical Devices Agency (PMDA)
in Japan. We included the study patients from the CREDO-Kyoto registry cohort-2. The design
⁎ Corresponding author at: Department of Cardiovascular Medicine, Kyoto University and patient enrollment of the CREDO-Kyoto registry cohort-2 has been previously
Graduate School of Medicine, 54 Shogoin, Kawahara-cho, Sakyo-ku, Kyoto, 606-8507, described [9]. In brief, the CREDO-Kyoto registry cohort-2 is a physician-initiated non-
Japan. Tel.: +81 75 751 4255; fax: +81 75 751 3299. company sponsored multi-center registry enrolling consecutive patients who underwent
E-mail address: nei126@kuhp.kyoto-u.ac.jp (N. Ozasa). first coronary revascularization in 26 centers in Japan between January 2005 and

0167-5273/$ – see front matter © 2013 Elsevier Ireland Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.ijcard.2013.08.029
5222 Y. Kitai et al. / International Journal of Cardiology 168 (2013) 5221–5228

December 2007 (Supplemental Appendix A). Among 15,263 patients enrolled in the regis- complications such as prior myocardial infarction (MI), prior stroke,
try, we excluded 4900 patients with myocardial infarction (MI), 2005 patients who
peripheral vascular disease, heart failure, multivessel disease, moderate
underwent coronary artery bypass grafting (CABG), 609 patients who underwent emer-
gent PCI, 74 patients with unknown blood Hb level or estimated glomerular filtration or severe mitral regurgitation, and target of unprotected left main coro-
rate (eGFR), and 376 patients with dialysis. Therefore, the present study population nary artery were more frequently observed in the anemia groups than
consisted of 7299 patients who underwent first elective PCI (Fig. 1). the no-anemia group. The frequency of DES use was not significantly
Demographics, clinical factors, angiographic data, and discharge medications were different between the groups. The prescription of antiplatelet therapy or
collected from hospital charts or hospital databases according to pre-specified definitions
by experienced clinical research coordinators (Supplemental Appendix B). Blood samples
warfarin at discharge was also not significantly different. Beta-blockers,
for Hb and eGFR were collected before PCI. Follow-up data were obtained from hospital ACE-I/ARB, nicorandil, and proton pump inhibitors were more often
charts or by contacting patients or referring physicians. Because final data collection for used at discharge in the anemia groups, whereas statins were more
follow-up events was initiated on July 1, 2009, follow-up events were censored on this often used in the no-anemia group.
date. Median follow-up duration was 953 (interquartile range 701 to 1230) days. Com-
plete 1-year follow-up information was obtained in 7077 patients (96.9%) out of 7299
patients. 3.2. Clinical outcomes
The relevant review boards or ethics committees in all 26 participating centers ap-
proved the study protocol. Because of retrospective enrollment, written informed consent In the entire study population, crude 3-year incidence of MACE was
from the patients was waived; however, patients who refused to participate in the study significantly higher in the anemia groups (Table 2A, Fig. 2A). After
when contacted for follow-up were excluded. This strategy is in agreement with guide-
lines for epidemiologic studies issued by the Ministry of Health, Labor and Welfare of
adjusting for confounders, the risk for MACE remained significant with
Japan. HR of 1.77 (95% CI: 1.47–2.15, p b 0.0001) for the mild anemia group
compared to the no-anemia group and with HR of 2.53 (95% CI: 2.03–
2.2. Definitions and endpoints 3.14, p b 0.0001) for the moderate-to-severe anemia group compared
to the no-anemia group (Table 2A and Supplemental Table 3). For the
Definitions of baseline characteristics were described previously [9]. According to the secondary endpoints, crude 3-year incidences of all cause death, heart
definition of World Health Organization (WHO), anemia was defined as an Hb level
failure hospitalization, MI, and stroke were significantly higher in the
b12.0 g/dL for women and b13.0 g/dL for men [10]. In this study, therefore, mild anemia
was defined as an Hb level of 11.0–11.9 g/dL for women and 11.0–12.9 g/dL for men, and anemia groups (Table 2A). In addition, mild anemia was associated
moderate-to-severe anemia was defined as an Hb level b11.0 g/dL both for women and with significantly higher adjusted HRs for all cause death and heart fail-
for men [11]. The eGFR calculated by the Modification of Diet in Renal Disease formula ure hospitalization (Table 2A). Significantly higher adjusted HRs of
for Japanese patients was used as a surrogate of renal function [12]. CKD was defined as moderate-to-severe anemia for all cause death, heart failure hospitaliza-
eGFR b60 ml/min/1.73 m2, according to the National Kidney Foundation KDOQI guide-
lines [13].
tion and MI were also observed (Table 2A).
The primary outcome was major adverse cardiac events (MACE) defined as composite
of all cause death, heart failure hospitalization, and MI. The secondary outcome included 3.3. Subgroup analyses in patients with and without CKD
the individual components of the primary outcome and stroke. Heart failure hospitaliza-
tion was defined as hospitalization due to worsening of heart failure requiring intravenous
In subgroup analyses in patients with and without CKD (CKD group
drug therapy. MI was defined according to the definition in the Arterial Revascularization
Therapy Study [14]. Stroke was defined as ischemic or hemorrhagic stroke requiring hos- and no-CKD group), the differences in baseline characteristics between
pitalization with symptoms lasting N24 h. All cause death, heart failure hospitalization, MI the groups were similar to the differences in the entire study population
and stroke were adjudicated by a clinical event committee (Supplemental Appendix C). (Supplemental Table 1 and 2). In patients with CKD, crude 3-year in-
cidence of MACE was significantly higher in the anemia groups
2.3. Statistical analysis (Table 2B, Fig. 2B). The risk of moderate-to-severe anemia for MACE is
particularly pronounced in the presence of CKD, with 3-year incidence
We divided the study patients into 3 groups according to the definitions of anemia:
no-anemia group, mild anemia group, and moderate-to-severe anemia group. Differences of MACE up to 37.4% (Table 2B, Fig. 2B). After adjusting for confounders,
in baseline clinical characteristics between the 3 groups were compared using the chi- the risk for MACE remained significant with HR of 1.46 (95% CI: 1.13–
square test for categorical variables, and the ANOVA or Kruskal–Wallis tests for continuous 1.90, p = 0.003) for the mild anemia group and with HR of 1.76 (95%
variables. Continuous variables are presented as mean ± SD, and categorical variables are CI: 1.33–2.34, p b 0.0001) for the moderate-to-severe anemia group
expressed as number and percentages. Cumulative incidence was estimated by the
compared to the no-anemia group (Table 2B). For the secondary end-
Kaplan–Meier method and differences were assessed with the log-rank test. We used
Cox proportional-hazards model stratified by centers to estimate the hazard ratios (HR) points, crude 3-year incidences of all cause death and heart failure
of mild and moderate-to-severe anemia by incorporating severity of anemia as dummy hospitalization were also significantly higher in the anemia groups
variables with 36 clinically relevant risk-adjusting variables listed in Table 1. Results of (Table 2B). In addition, moderate-to-severe anemia was associated
the multivariable analysis are expressed as HRs with 95% confidence interval.
with significantly higher adjusted HRs for all cause death and heart failure
As subgroup analyses, the impact of anemia on clinical outcomes was assessed in
patients with CKD (n = 2792) and without CKD (n = 4507) in the same way. While
eGFR b60 ml/min/1.73 m2 was chosen as a risk-adjusting variable in the entire study
population, eGFR b45 ml/min/1.73 m2 was used in the presence of CKD, and eGFR of
60–74 ml/min/1.73 m2 was used in the absence of CKD (Table 1 and Supplemental
Table 1 and 2). To estimate the adjusted HRs for interactions between anemia and reduced
renal function, we developed the same Cox proportional-hazards model for the interaction
variables with the risk adjustment variables. All statistical analyses were conducted by
physicians (Y.K. and N.O.) and statistician (T.M.) using JMP 10 and SAS 9.2 (SAS Institute
Inc., Cary, NC, USA), and all the reported p values were 2 sided. p values b0.05 were con-
sidered statistically significant.

3. Results

3.1. Baseline characteristics

Anemia was observed in 2206 patients (30.2% of the study popula-

tion) composed of mild anemia (n = 1466, 20.1%) and moderate-to-
severe anemia (n = 740, 10.1%) (Table 1). Compared to patients in
the no-anemia group, patients in the anemia groups were older, likely Fig. 1. Study flow chart. AMI = acute myocardial infarction, CABG = coronary artery
to be female, and had more comorbidities such as CKD, hypertension, bypass surgery, eGFR = estimated glomerular filtration rate, Hb = hemoglobin, and
diabetes mellitus, thrombocytopenia, and malignancy. Cardiovascular PCI = percutaneous coronary intervention.
 Y. Kitai et al. / International Journal of Cardiology 168 (2013) 5221–5228 5223

Table 1
Baseline clinical characteristics, procedural characteristics and discharge medications in the entire study population.

No anemia Mild anemia Moderate-to-severe anemia p value

Number of patients n = 5093 n = 1466 n = 740

(A) Clinical characteristics

Age (years) 66.6 ± 10.0 72.9 ± 8.6 74.8 ± 9.1 b0.0001
Age ≥ 75 yearsa 1184 (23%) 700 (47%) 428 (57%) b0.0001
Malea 3779 (74%) 1069 (72%) 369 (49%) b0.0001
BMI 24.2 22.7 22.0 b0.0001b
(22.3–26.3) (20.7–24.8) (19.9–24.5)
BMI b 25.0a 3092 (60%) 1119 (76%) 589 (79%) b0.0001
Hypertensiona 4238 (83%) 1253 (85%) 650 (87%) 0.001
Diabetes mellitus 1928 (37%) 588 (40%) 354 (47%) b0.0001
On insulin therapya 359 (7.0%) 147 (10%) 123 (16%) b0.0001
Current smokinga 1514 (29%) 288 (19%) 122 (16%) b0.0001
Heart failurea 494 (9.7%) 249 (16%) 244 (32%) b0.0001
Multivessel diseasea 2817 (55%) 916 (62%) 493 (66%) b0.0001
Mitral regurgitation grade 3/4a 168 (4.0%) 81 (6.8%) 82 (13%) b0.0001
LVEF (%) 61.9 ± 11.9 60.7 ± 13.0 58.0 ± 14.3 b0.0001
Prior MIa 695 (13%) 240 (16%) 174 (23%) b0.0001
Prior strokea 478 (9.3%) 213 (14%) 139 (18%) b0.0001
Peripheral vascular diseasea 392 (7.6%) 221 (15%) 103 (13%) b0.0001
eGFR (ml/min/1.73 m2) 68.8 ± 24.8 60.2 ± 19.8 49.2 ± 22.7 b0.0001
eGFR b 60 ml/min/1.73 m2a 1554 (30.5%) 738 (50.3%) 500 (67.5%) b0.0001
Atrial fibrillationa 398 (7.8%) 131 (8.9%) 64 (8.6%) 0.33
Platelet b 100 ∗ 109/La 41 (0.8%) 20 (1.3%) 22 (2.9%) b0.0001
COPDa 193 (3.7%) 71 (4.8%) 27 (3.6%) 0.18
Liver cirrhosisa 120 (2.3%) 42 (2.8%) 25 (3.3%) 0.20
Malignancya 374 (7.3%) 225 (15%) 140 (18%) b0.0001

(B) Procedural characteristics

Number of target lesions 1.48 ± 0.77 1.51 ± 0.77 1.49 ± 0.74 0.42
Target of proximal LADa 3020 (59%) 856 (58%) 444 (60%) 0.73
Target of unprotected LMCAa 167 (3.2%) 50 (3.4%) 43 (5.8%) 0.005
Target of CTOa 869 (17%) 223 (15%) 122 (16%) 0.23
Target of bifurcationa 1820 (35%) 504 (34%) 266 (35%) 0.60
Side-branch stentinga 282 (5.5%) 84 (5.7%) 43 (5.8%) 0.92
Total number of stents 1.92 ± 1.28 1.98 ± 1.28 2.02 ± 1.30 0.06
Total stent length (mm) 41.3 ± 30.9c 42.8 ± 30.8c 43.2 ± 30.7c 0.11
Total stent length N 28 mma 2461 (51%)c 754 (54%)c 398 (56%)c 0.006
Minimum stent size (mm) 2.87 ± 0.42c 2.83 ± 0.42c 2.81 ± 0.41c 0.0002
Minimum stent size b 3.0 mma 2303 (47%)c 736 (52%)c 379 (54%)c 0.0002
DES usea 3359 (69%) 995 (71%) 471 (67%) 0.17

(C) Discharge Medications

Antiplatelet therapy
Thienopyridine 5009 (98.3%) 1444 (98.4%) 724 (97.8%) 0.52
Ticlopidine 4501 (90.1%) 1286 (89.3%) 646 (89.2%) 0.56
Clopidogrel 494 (9.8%) 154 (10.6%) 78 (10.7%) 0.56
Aspirin 5022 (98.6%) 1444 (98.4%) 728 (98.3%) 0.86
Cilostazola 440 (8.6%) 144 (9.8%) 76 (10.2%) 0.18
Other medications
Statinsa 2846 (55%) 650 (44%) 282 (38%) b0.0001
Beta-blockersa 1245 (24%) 400 (27%) 210 (28%) 0.01
ACE-I/ARBa 2480 (48%) 798 (54%) 415 (56%) b0.0001
Nitratesa 1942 (38%) 599 (40%) 305 (41%) 0.07
Calcium channel blockersa 2655 (52%) 782 (53%) 379 (51%) 0.59
Nicorandila 1076 (21%) 349 (23%) 185 (25%) 0.01
Warfarina 355 (6.9%) 128 (8.7%) 59 (7.9%) 0.06
Proton pump inhibitorsa 864 (16%) 319 (21%) 218 (29%) b0.0001
H2-blockersa 1103 (21%) 330 (22%) 156 (21%) 0.69

Variables are mean ± SD, n (%), and median (IQR).

ACE-I = angiotensin converting enzyme inhibitor, ARB = angiotensin receptor blocker, BMI = body mass index, COPD = chronic obstructive pulmonary disease, CTO = chronic total
occlusion, DES = drug-eluting stent, eGFR = estimated glomerular filtration rate, H2-blocker = histamine type2 receptor blocker, LAD = left anterior descending artery, LMCA = left
main coronary artery, LVEF = left ventricular ejection fraction, and MI = myocardial infarction.
a
Clinically relevant risk adjusting variables selected for the Cox proportional-hazards model.
b
Wilcoxon p value.
c
Exclude 390 patients without stent implantation, 280 in the no-anemia group, 70 in the mild anemia group, and 40 in the moderate-to-severe anemia group.

hospitalization (Table 2B). Although not statistically significant, there 3.4. Adjusted HRs for interactions between anemia and reduced
remained a trend toward a higher adjusted HR for heart failure hospitali- renal function
zation with mild anemia (Table 2B). In patients without CKD, the direc-
tion of the impact of anemia on clinical outcomes was similar to the Relative hazards of MACE based on anemia and reduced renal func-
entire study population; however, the adjusted HRs were higher than tion are shown in Fig. 3. For all eGFR levels, adjusted HRs for MACE in-
those observed in the entire study population (Table 2C, Fig. 2C). creased as anemia worsens. The risk for MACE showed an accretive
5224 Y. Kitai et al. / International Journal of Cardiology 168 (2013) 5221–5228

Table 2
Unadjusted and adjusted relationship between severity of anemia and 3-year clinical outcomes.

Crude event ratesa Adjusted HR [95% CI]b

No anemia Mild anemia Moderate -to-severe anemia Log-rank p value Mild anemia p value Moderate -to-severe anemia p value

(A) The entire study population n = 5093 n = 1466 n = 740

MACE 348 249 217 b0.0001 1.77 b0.0001 2.53 b0.0001
(7.9%) (20.1%) (34.2%) [1.47–2.15] [2.03–3.14]
All cause death 150 132 142 b0.0001 1.86 b0.0001 3.35 b0.0001
(3.5%) (11.2%) (22.9%) [1.42–2.42] [2.52–4.47]
Heart failure hospitalization 139 114 102 b0.0001 1.87 b0.0001 2.25 b0.0001
(3.1%) (9.5%) (16.9%) [1.40–2.51] [1.61–3.15]
MI 112 45 32 0.0008 1.14 0.54 1.87 0.01
(2.5%) (3.3%) (5.8%) [0.74–1.75] [1.13–3.09]
Stroke 149 62 36 0.0001 1.15 0.42 0.90 0.68
(3.4%) (5.1%) (6.0%) [0.81–1.62] [0.54–1.48]

(B) With CKD n = 1554 n = 738 n = 500

MACE 164 154 161 b0.0001 1.46 0.003 1.76 b0.0001
(12.2%) (23.5%) (37.4%) [1.13–1.90] [1.33–2.34]
All cause death 76 76 109 b0.0001 1.29 0.17 2.19 b0.0001
(5.8%) (12.1%) (26.1%) [0.89–1.88] [1.50–3.18]
Heart failure hospitalization 85 76 76 b0.0001 1.43 0.05 1.56 0.03
(6.2%) (12.0%) (19.0%) [0.98–2.07] [1.03–2.36]
MI 36 28 19 0.06 1.07 0.80 1.30 0.44
(2.8%) (4.1%) (5.2%) [0.60–1.93] [0.65–2.59]
Stroke 64 33 26 0.31 1.04 0.86 0.62 0.16
(4.7%) (5.6%) (6.4%) [0.64–1.69] [0.32–1.20]

(C) Without CKD n = 3539 n = 728 n = 240

MACE 184 95 56 b0.0001 2.05 b0.0001 3.54 b0.0001
(6.0%) (16.7%) (27.3%) [1.52–2.75] [2.42–5.17]
All cause death 74 56 33 b0.0001 2.66 b0.0001 5.29 b0.0001
(2.6%) (10.3%) (16.1%) [1.78–3.97] [3.16–8.87]
Heart failure hospitalization 54 38 22 b0.0001 2.60 0.0002 2.43 0.01
(1.7%) (6.9%) (12.4%) [1.56–4.34] [1.23–4.82]
MI 76 17 13 0.006 1.10 0.77 3.27 0.003
(2.3%) (2.6%) (7.0%) [0.56–2.15] [1.49–7.15]
Stroke 85 29 10 0.004 1.29 0.31 1.27 0.56
(2.8%) (4.5%) (5.0%) [0.78–2.13] [0.55–2.97]

CI = confidence interval, CKD = chronic kidney disease, MACE = major adverse cardiac events, and MI = myocardial infarction.
a
Incidences at 3 years were estimated by Kaplan–Meier method.
b
The no-anemia group served as a reference group in the Cox proportional-hazards model.

increment with exacerbation in either the renal function or anemia [17]. These hemodynamic responses induce additional stress to left ven-
(interaction p b 0.001). tricular function which may lead to progressive left ventricular hyper-
trophy and further myocardial cell death [18,19]. On the other hand,
4. Discussion anemic patients had more comorbidities such as CKD. CKD, which
frequently occurs in CAD patients [20], can deteriorate anemia and
In this large cohort of 7299 patients undergoing elective PCI, (1) ane- may complicate the association between anemia and clinical outcomes
mia was observed in as many as 2206 patients (30.2% of the entire study through various causal mechanisms such as impaired erythropoie-
population), composed of 1466 patients with mild anemia (20.1%) and tin production [21], and bone suppression from increased levels of
740 patients with moderate-to-severe anemia (10.1%), (2) anemic pa- proinflammatory cytokines [22]. Recently, Silverberg DS et al. proposed
tients had more comorbidities such as CKD, advanced age, hyperten- a novel clinical entity they named Cardio-Renal-Anemia syndrome in
sion, diabetes mellitus, malignancy, and cardiovascular complications, which each one of its three components, cardiac performance, anemia,
(3) both mild and moderate-to-severe anemia were associated with and CKD could negatively influence each other [23].
worse long-term outcomes independent of the presence of CKD, Although the present study demonstrated that anemia was an im-
(4) anemia and reduced renal function independently and cumula- portant prognostic factor for cardiovascular events in patients undergo-
tively increased the risk for MACE in these patients. ing elective PCI, it is unclear whether treatment for anemia improves
There are previous studies reporting increased risk of mortality the outcomes of these patients. No randomized controlled trial (RCT)
and cardiovascular events in anemic patients undergoing PCI [6–8]; has been conducted to assess the clinical efficacy of treatment for ane-
however, few studies investigated the prognostic impact of mild anemia mia in patients with CAD. In CKD patients, two recent large RCTs, the
[15,16], and data on the impact of anemia on clinical outcomes in pa- Trial to Reduce Cardiovascular Events with Aranesp Therapy (TREAT)
tients undergoing elective PCI is limited. Our study demonstrated that [24] and the Correction of Hemoglobin and Outcomes in Renal insuffi-
in patients who underwent elective PCI, even mild anemia was asso- ciency (CHOIR) trial [25], have reported that complete correction
ciated with significantly higher risk for MACE even after adjusting for of moderate-to-severe anemia (Hb b 11.0 g/dL) with erythropoiesis-
confounders. stimulating agent (ESA) therapy was associated with increased risk for
The mechanisms for worse long-term outcomes in anemic patients cardiovascular events compared to partial correction. A recent large
undergoing elective PCI are supposed to be multifactorial. Anemia RCT, the Cardiovascular Risk Reduction by Early Anemia Treatment
leads to impaired oxygen delivery to coronary artery and myocardial with Epoetin Beta (CREATE) trial, has shown that complete correction
cell ischemia [3]. Anemia also increases myocardial oxygen demand of mild anemia (Hb = 11.0–12.5 g/dL) with ESA therapy had no signif-
through increments in heart rate, cardiac index, and stroke volume icant effect on the risk for cardiovascular events [26]. In the present
 Y. Kitai et al. / International Journal of Cardiology 168 (2013) 5221–5228 5225

study, mild anemia (Hb = 11.0–11.9 g/dL for women and 11.0– stimulating agent, or iron supplementation. Second, we had limited
12.9 g/dL for men) was also an independent predictor for MACE in data on details about the etiology of anemia. In many cases we were un-
patients with CKD. Current guidelines recommend the target Hb level able to define the cause of anemia such as occult gastrointestinal bleed-
should generally be in the range of 11.0–12.0 g/dL and should not be ing that independently may have worsen the prognosis of these patients.
greater than 13.0 g/dL in CKD patients with anemia [27], however, it In addition, we were not able to rule out the possibility that other
is unknown whether this recommendation is applicable in patients unmeasured confounding factors might have influenced the outcomes.
with CAD. The post-hoc analysis of the CHOIR trial revealed no in-
creased risk associated with complete anemia correction among CKD 5. Conclusions
patients with heart failure [28]. The target Hb level may be higher in
CKD patients with obvious cardiac disease compared to those without Anemia was significantly associated with worse 3-year clinical out-
it. Further investigations may clarify the effect of treatment for anemia comes in patients undergoing elective PCI. According to the WHO defi-
and the recommended target Hb level in patients with CAD. nition of anemia, even mild anemia was significantly associated with
the risk for MACE. Coexisting CKD additively increased the risk for
4.1. Study limitations MACE in these patients.

There are several important limitations in this study. First, we ana- Acknowledgments
lyzed only Hb levels at baseline. In addition to the absence of analyses
of serial Hb levels, we have not established whether patients received This study was supported by the Pharmaceuticals and Medical
any specific treatment for anemia such as transfusion, erythropoiesis- Devices Agency (PMDA) in Japan. We appreciate the support and

A)

Innterrvall 0 dayy 30 dayys 1 yeaar 2 year

y rs 3 year
y rs
N
No aanemiaa
N oof evennts 61
6 2003 287 348
8
N oof ppatieentss at risk
k 5093
3 5 01
500 447447 3 3
366 1920
1 0
Inccidencee 1 %
1.2% 44.0%
% 5.8%
5 % 7.9%
7 %

M
Mild
d an
nem
mia
N oof evennts 31
3 137 203 249
9
N oof ppatieentss at risk
k 1466
6 1418 112882 936 480
0
Inccidencee 22.1%
% 99.5%
% 14
4.6%
% 20
0.1%
%

M deraate--to--sev
Mod veree an
nem
mia
N oof evennts 24
2 134 194 217
7
N oof ppatieentss at risk
k 740
0 70
04 5773 383 184
4
Inccidencee 33.2%
% 188.6%
% 8.1%
28 % 4.2%
34 %

Fig. 2. Crude cumulative incidence curves of MACE in the entire study population (A), patients with CKD (B), and those without CKD (C). CKD = chronic kidney disease,
PCI = percutaneous coronary intervention.
5226 Y. Kitai et al. / International Journal of Cardiology 168 (2013) 5221–5228

B)

Interval 0 dayy 30 dayys 1 yeaar 2 year

y rs 3 year
y rs
N
No aanemiaa
N of events 19 992 137 164
4
N of patients at risk 1554 152
27 114228 1 9
108 588
8
Incidence 1.2%
1 % 5.9%
5 % 9.2%
9 % 122.2%
%

M
Mild
d an
nem
mia
N of events 23
2 90 135 154
4
N of ppatientss at riskk 738 709 627 457 244
4
Inccidence 33.1%
% 122.4%
% 19
9.3%
% 233.5%
%

M
Mod
deraate--to--sev
veree an
nem
mia
N oof evennts 19 100 144 161
N oof ppatieentss at riskk 500 471 379 248 115
Inccidencee 3.8% 20.6% 31.0% 37.4%

Fig. 2 (continued).

collaboration of the co-investigators participating in the CREDO- [4] Anand IS. Heart failure and anemia: mechanisms and pathophysiology. Heart Fail Rev
Kyoto PCI/CABG registry cohort-2. We are indebted to the outstanding 2008;13:379–86.
[5] Toma C, Zahr F, Moguilanski D, et al. Impact of anemia on platelet response to
effort of the clinical research coordinators for data collection. We clopidogrel in patients undergoing percutaneous coronary stenting. Am J Cardiol
also thank Tomás Romero, MD, Associate Clinical Professor, Universi- 2012;109:1148–53.
ty of California, San Diego, for his advice in the review and editing [6] McKechnie RS, Smith D, Montoye C, et al. Prognostic implication of anemia
on in-hospital outcomes after percutaneous coronary intervention. Circulation
the manuscript. 2004;110:271–7.
[7] Nikolsky E, Mehran R, Aymong ED, et al. Impact of anemia on outcomes of pa-
tients undergoing percutaneous coronary interventions. Am J Cardiol 2004;94:
Appendix A. Supplementary data 1023–7.
[8] Pilgrim T, Vetterli F, Kalesan B, et al. The impact of anemia on long-term clinical
outcome in patients undergoing revascularization with the unrestricted use of
Supplementary data to this article can be found online at http://dx.
drug-eluting stents. Circ Cardiovasc Interv 2012;5:202–10.
doi.org/10.1016/j.ijcard.2013.08.029. [9] Kimura T, Morimoto T, Furukawa Y, et al. Long-term safety and efficacy of sirolimus-
eluting stents versus bare-metal stents in real world clinical practice in Japan.
Cardiovasc Intervent Ther 2011;26:234–45.
References [10] Nutritional anaemias. Report of a WHO scientific group. World Health Organ Tech
Rep Ser 1968;405:5–37.
[1] Reinecke H, Trey T, Wellmann J, et al. Haemoglobin-related mortality in patients [11] WHO. Haemoglobin concentrations for the diagnosis ofanaemia and assessment of se-
undergoing percutaneous coronary interventions. Eur Heart J 2003;24:2142–50. verity. Vitamin and Mineral Nutrition Information System. Geneva: World Health Orga-
[2] Spivak JL. The blood in systemic disorders. Lancet 2000;355:1707–12. nization; 2011 (WHO/NMH/NHD/MNM/11.1)(http://www.who.int/vmnis/indicators/
[3] Levy PS, Quigley RL, Gould SA. Acute dilutional anemia and critical left anterior haemoglobin.pdf, accessed [date]).
descending coronary artery stenosis impairs end organ oxygen delivery. J Trauma [12] Matsuo S, Imai E, Horio M, et al. Revised equations for estimated GFR from serum
1996;41:416–23. creatinine in Japan. Am J Kidney Dis 2009;53:982–92.
 Y. Kitai et al. / International Journal of Cardiology 168 (2013) 5221–5228 5227

C)

Innterrvall 0 dayy 30 dayys 1 yeaar 2 year

y rs 3 year
y rs
N
No aanemiaa
N oof evennts 43
4 111 151 184
4
N oof ppatieentss at riskk 3539
9 3 74
347 33319 2 4
257 1332
1 2
Inccidencee 1 %
1.2% 33.1%
% 4.4%
4 % 6.0%
6 %

M
Mild
d an
nem
mia
N oof evennts 9 48 69 95
5
N oof patieentss at riskk 728 709 655 479 236
Inccidencee 1.2% 6.7% 9.9% 16.7%

M
Mod
deraate--to--sev
veree an
nem
mia
N oof evennts 5 34 50 56
N oof ppatieentss at riskk 240 233 194 135 69
Inccidencee 2.0% 14.4% 22.2% 27.3%

Fig. 2 (continued).

[13] Foundation NK. K/DOQI clinical practice guidelines for chronic kidney disease: eval- [21] Eckardt KU, Kurtz A. Regulation of erythropoietin production. Eur J Clin Invest
uation, classification, and stratification. Am J Kidney Dis 2002;39:S1–S266. 2005;35:13–9.
[14] Serruys PW, Ong AT, van Herwerden LA, et al. Five-year outcomes after coronary [22] de Francisco AL, Stenvinkel P, Vaulont S. Inflammation and its impact on anaemia in
stenting versus bypass surgery for the treatment of multivessel disease: the final chronic kidney disease: from haemoglobin variability to hyporesponsiveness. NDT
analysis of the Arterial Revascularization Therapies Study (ARTS) randomized trial. Plus 2009;2:i18–26.
J Am Coll Cardiol 2005;46:575–81. [23] Silverberg DS, Wexler D, Blum M, et al. The interaction between heart failure,
[15] Lee PC, Kini AS, Ahsan C, Fisher E, Sharma SK. Anemia is an independent predictor renal failure and anemia — the cardio-renal anemia syndrome. Blood Purif
of mortality after percutaneous coronary intervention. J Am Coll Cardiol 2004;44:541–6. 2004;22:277–84.
[16] Poludasu S, Marmur JD, Weedon J, Khan W, Cavusoglu E. Effect of hemoglobin level [24] Pfeffer MA, Burdmann EA, Chen CY, et al. A trial of darbepoetin alfa in type 2 diabetes
on long-term all-cause mortality after percutaneous coronary intervention in and chronic kidney disease. N Engl J Med 2009;361:2019–32.
African-Americans. Am J Cardiol 2009;103:1078–82. [25] Singh AK, Szczech L, Tang KL, et al. Correction of anemia with epoetin alfa in chronic
[17] Duke M, Abelmann WH. The hemodynamic response to chronic anemia. Circulation kidney disease. N Engl J Med 2006;355:2085–98.
1969;39:503–15. [26] Drüeke TB, Locatelli F, Clyne N, et al. Normalization of hemoglobin level in patients
[18] Metivier F, Marchais SJ, Guerin AP, Pannier B, London GM. Pathophysiology of anaemia: with chronic kidney disease and anemia. N Engl J Med 2006;355:2071–84.
focus on the heart and blood vessels. Nephrol Dial Transplant 2000;15:14–8. [27] KDOQI. KDOQI clinical practice guideline and clinical practice recommendations for
[19] Katz AM. The cardiomyopathy of overload: an unnatural growth response. Eur Heart anemia in chronic kidney disease: 2007 update of hemoglobin target. Am J Kidney
J 1995;16:110–4. Dis 2007;50:471–530.
[20] Reddan DN, Szczech LA, Tuttle RH, et al. Chronic kidney disease, mortality, and treat- [28] Szczech LA, Barnhart HX, Sapp S, et al. A secondary analysis of the CHOIR trial shows
ment strategies among patients with clinically significant coronary artery disease. J that comorbid conditions differentially affect outcomes during anemia treatment.
Am Soc Nephrol 2003;14:2373–80. Kidney Int 2010;77:239–46.
5228 Y. Kitai et al. / International Journal of Cardiology 168 (2013) 5221–5228

Fig. 3. Relative hazard of MACE based on anemia and reduced renal function. The group
with no anemia and eGFR ≥60 ml/min/1.73 m2 served as a reference group in the Cox
proportional-hazards model. eGFR = estimated glomerular filtration rate.