G Model

JJCC-1554; No. of Pages 8

Journal of Cardiology xxx (2017) xxx–xxx

Contents lists available at ScienceDirect

Journal of Cardiology
journal homepage: www.elsevier.com/locate/jjcc

Original article

In-hospital mortality analysis of Japanese patients with acute coronary

syndrome using the Tokyo CCU Network database: Applicability of the
GRACE risk score
Kota Komiyama (MD, PhD)a,b *, Masato Nakamura (MD, PhD, FJCC)a,
Kengo Tanabe (MD, PhD)b, Hiroki Niikura (MD)a, Hajime Fujimoto (MD, FJCC)a,
Keiko Oikawa (MD)a, Hiroyuki Daida (MD, PhD, FJCC)a, Takeshi Yamamoto (MD)a,
Ken Nagao (MD, PhD, FJCC)a, Morimasa Takayama (MD, FJCC)a Tokyo CCU Network
Scientific Committee
a
Tokyo CCU Network Scientific Committee, Tokyo, Japan
b
Department of Cardiology, Mitsui Memorial Hospital, Tokyo, Japan

A R T I C L E I N F O A B S T R A C T

Article history: Background: The GRACE risk score was developed to predict in-hospital mortality for acute coronary
Received 5 June 2017 syndrome (ACS) using multinational registries, but did not include Japanese data. Therefore, GRACE risk
Received in revised form 23 August 2017 scores are not extensively used in Japan. The present study aimed to evaluate the relationship between
Accepted 12 September 2017
the GRACE risk score and in-hospital mortality among Japanese patients with ACS using the Tokyo CCU
Available online xxx
(cardiovascular care unit) Network Database.
Methods and results: A total of 9460 patients with ACS hospitalized at 67 Tokyo CCUs between January
Keywords:
2011 and December 2013 were retrospectively reviewed and GRACE risk scores were calculated. Patients
GRACE risk score
in the Tokyo CCU Network database had more severe conditions compared to those of the original GRACE
Japanese
In-hospital mortality study. There was a strong correlation between the GRACE risk score and in-hospital mortality for patients
Acute coronary syndrome with ST-segment elevation myocardial infarction (STEMI) or non ST-segment elevation myocardial
infarction (NSTEMI) (r = 0.99, p < 0.001); however, the correlation was not significant for patients with
unstable angina (r = 0.35, p = 0.126). For STEMI + NSTEMI patients, the discrimination ability of the
GRACE risk score was excellent, with a c statistic of 0.87 (95% confidence interval, 0.86–0.89).
Conclusions: The GRACE risk score is a good predictor of in-hospital mortality for Japanese patients with
STEMI or NSTEMI, and can help clinicians stratify patients by risk for optimal patient triage and early
treatment management.
© 2017 Published by Elsevier Ltd on behalf of Japanese College of Cardiology.

Introduction
Acute coronary syndrome (ACS) represents a major cause of
morbidity and mortality throughout the world [1,2]. ACS comprises
a wide spectrum of conditions including ST-segment elevation
myocardial infarction (STEMI), non-ST-segment elevation myocar-
dial infarction (NSTEMI), and unstable angina (UA). The presence of
ACS guides triage and key management decisions. Prognostic
scoring systems, including the GRACE risk score [3], TIMI risk score
* Corresponding author at: Department of Cardiology, Mitsui Memorial Hospital,
Kanda-Izumi-cho 1, Chiyoda-ku, Tokyo 101-8643, Japan.
E-mail address: yang3291@hotmail.com (K. Komiyama).
[4], and PURSUIT risk score [5], have been reported to be useful in
predicting not only long-term mortality, but also in-hospital
mortality among patients with ACS [5–7]. Among these scoring
systems, the GRACE risk score has the highest validity [8,9]. It was
designed to reflect an unbiased population of patients with ACS
from 94 hospitals in 14 countries (Australia, Austria, Argentina,
Belgium, Brazil, Canada, France, Germany, Italy, New Zealand,
Poland, United Kingdom, USA, and Spain). All cases were assigned
to one of the following categories: STEMI, NSTEMI, or UA. All eight
factors used to calculate the GRACE risk score can be obtained at
the initial examination. These predictive factors include age, heart
rate, systolic blood pressure, initial serum creatinine level, Killip
class [10], cardiac arrest at hospital admission, elevated cardiac
markers, and ST-segment deviation, which are weighted to

https://doi.org/10.1016/j.jjcc.2017.09.006
0914-5087/© 2017 Published by Elsevier Ltd on behalf of Japanese College of Cardiology.

Please cite this article in press as: Komiyama K, et al. In-hospital mortality analysis of Japanese patients with acute coronary syndrome
using the Tokyo CCU Network database: Applicability of the GRACE risk score. J Cardiol (2017), https://doi.org/10.1016/j.jjcc.2017.09.006
G Model
JJCC-1554; No. of Pages 8

2 K. Komiyama et al. / Journal of Cardiology xxx (2017) xxx–xxx

calculate the risk of in-hospital mortality [3]. In the guidelines of
the European Society of Cardiology (ESC), GRACE risk scores >140
are defined as a high risk of in-hospital mortality for patients with
ACS [11].
In Japan, the GRACE risk scoring system is not widely used.
Japanese facilities did not participate in the GRACE study; thus, the
guidelines for the management of patients with STEMI issued by
the Japanese Circulation Society in 2013 stated that it is necessary
to pay attention to its use in Japan [12]. Moreover, the guidelines
for the management of ACS without persistent ST-segment
elevation issued by the Japanese Circulation Society in 2012 iden-
tified limitations in the use of the GRACE risk score in Japan owing
to a higher prevalence of vasospastic angina in the Japanese
[13]. Therefore, the aim of the present study was to examine the
relationship between the GRACE risk score and the in-hospital
mortality among Japanese patients with ACS using data from the
Tokyo CCU (cardiovascular care unit) Network Database.
Methods
Patients
A retrospective review of data from the Tokyo CCU Network
database was conducted. This network was established in 1978 by
the Metropolitan Tokyo CCU Communication Society with help
from the ambulance units and control room of the Tokyo Fire
Department [14–16].
In 2012, the Tokyo CCU Network included 67 hospitals.
Hospitals belonging to the Tokyo CCU Network routinely record
and submit the details of all patients treated in their CCUs using
dedicated survey forms. In the GRACE study, the enrolled patients
who were similar to those in the MONICA Project [17] had to be at
least 18 years of age, be admitted to participating hospitals with
symptoms consistent with acute coronary ischemia, and have at
least 1 of the following items: electrocardiographic changes
consistent with ACS, serial increases in serum markers of cardiac
necrosis, and/or documentation of coronary artery disease
[3,18]. The present study used the same inclusion criteria, with
the exception that patients had to be admitted to participating
CCUs of the Tokyo CCU Network. In the present study, 13,325
patients with ACS, treated at participating institutions of the Tokyo
CCU Network between January 2011 and December 2013, met
inclusion criteria. Among them, 3865 patients who lost their medical
history and the 8 factors needed for calculating the GRACE risk score
were excluded, and 9460 patients were analyzed in this study.
Clinical definitions and end points
“STEMI” was defined as ACS with persistent ST-segment
elevation in the electrocardiogram, “NSTEMI” was defined as
ACS without persistent ST-segment elevation in the electrocardio-
gram, and “UA” was defined as myocardial ischemia at rest or
minimal exertion in the absence of cardiac necrosis. Death was
defined as all-cause mortality during hospitalization. Vital signs,
Killip class, and biochemical and electrocardiographic findings
were collected at the time of hospital admission. Electrocardio-
grams were interpreted locally in terms of ST-segment depressions
of at least 1 mm in the anterior, inferior, or lateral lead groups, and
Q waves measuring either one-third the height of the R waves or
lasting longer than 0.04 s. Initial cardiac markers were defined as
positive when troponin T/I or other cardiac enzymes were elevated
according to the GRACE scoring system.
Calculation of the GRACE risk score
The GRACE prediction score card and nomogram were used to
calculate the prognostic score and estimate the risk of mortality for
individual patients (Fig. 1). The calculation of the GRACE risk score
was performed in 3 steps. Step 1 involved identifying the score for
each individual predictive factor (age, heart rate, systolic blood
pressure, initial serum creatinine level, Killip class, cardiac arrest at
hospital admission, elevated cardiac markers, and ST-segment
deviation). Step 2 involved summing the individual factor scores to
obtain a total score. Step 3 involved determining the risk of in-
hospital mortality using a nomogram. The GRACE risk score can also
be calculated online (www.outcomes.org/grace). In addition, smart-
phone applications are available as free downloads; thus, the GRACE
risk score can be easy to use in emergency medical situations.

Fig. 1. GRACE risk nomogram. The GRACE risk score is calculated in 3 steps. In step 1, the points for each predictive factor are determined. In step 2, the individual factors points
are summed. In step 3, the risk corresponding to the total points is identified [3]. SBP, systolic blood pressure.

Please cite this article in press as: Komiyama K, et al. In-hospital mortality analysis of Japanese patients with acute coronary syndrome
using the Tokyo CCU Network database: Applicability of the GRACE risk score. J Cardiol (2017), https://doi.org/10.1016/j.jjcc.2017.09.006
G Model
JJCC-1554; No. of Pages 8

K. Komiyama et al. / Journal of Cardiology xxx (2017) xxx–xxx 3

Statistical analysis
Consecutive data are expressed as median and 25th and 75th
percentiles, in a manner similar to that in the original GRACE study.
One-sample Wilcoxon signed-rank and Fisher‘s exact tests were
used to compare the distribution of the present study to the
median values of the GRACE study. The total calculated GRACE risk
scores were rounded to the nearest whole number. The relation-
ship between the GRACE risk score and in-hospital mortality was
evaluated using the Spearman's rank correlation coefficient. A
receiver-operating characteristics curve for in-hospital mortality
was evaluated and the c-statistic was calculated. The calibration of
adaptation for in-hospital mortality was investigated using the
Hosmer–Lemeshow test. In addition, univariate analyses were
performed to determine the factors associated with the incidence
of in-hospital mortality. Factors associated with the incidence of
in-hospital mortality in the univariate analyses were further
analyzed in a multivariable ordinal logistic regression model to
identify a set of predictors for in-hospital mortality. Factors whose
Table 1
Baseline characteristics in the present (Tokyo CCU Network) and GRACE registry
studies.
GRACE overall Tokyo CCU
population Network over-
(n = 11,389) all population
(n = 9460)
odds ratio was over 2.00 in univariate analysis were analyzed by
multivariate analysis. Statistical analyses were performed using
SPSS software version 23 (IBM SPSS Statistics, IBM Corporation,
Armonk, NY, USA). Construction of the calibration plot was
performed using STATA version 13.0 (LightStone Corp, Tokyo,
Japan). A p-value < 0.05 was considered statistically significant.
Results
Patient demographics and clinical characteristics in comparison with
the GRACE registry
A total of 9460 patients with ACS (median age, 69.0 years;
range, 60–78 years; sex, 25.3% women; STEMI, 63.0%; NSTEMI,
15.5%; UA, 21.5%; mortality, 5.0%) from the Tokyo CCU Network
database were retrospectively analyzed. The original GRACE study
included 11,389 patients (median age, 66.3 years; range, 56–75
years; sex, 33.5% females; STEMI, 35.3%; NSTEMI, not described;
UA, not described; mortality, 4.6%) with ACS. The baseline
characteristics for the overall study population of the present
study and the GRACE study are shown in Table 1, and those for the
patients who died in the hospital are shown in Table 2. The overall
study population in the GRACE study had a higher prevalence of
p value
previous coronary artery bypass grafting (CABG), congestive heart
failure, myocardial infarction, percutaneous coronary intervention,

Demographics Table 2
Age, y 66.3 69.0 <0.05 Baseline characteristics of patients experiencing in-hospital death in the present
(56.0–75.0) (60.0–78.0) (Tokyo CCU Network) and GRACE registry studies.
Female, % 33.5 25.3 <0.05
GRACE in- CCU Network p value
Weight, kg 76 (67–86) 62 (53–70) N1 <0.05
hospital deaths in-hospital
Height, cm 168 (160–175) 163 (156–169) <0.05
(n = 509) deaths (n = 470)
N2
Medical history, % Demographics
Angina, % 68.1 No data Age, y 76.2 79.0 (70.0–86.0) <0.05
Atrial fibrillation, % 8.0 No data (67.8–82.6)
CABG, % 12.6 2.2 <0.05 Female, % 41.7 35.7 0.07
Congestive heart 11.0 1.8 <0.05 Weight, kg 71 (62–80) 55 (49–65) N1 <0.05
failure, % Height, cm 165 (159–172) 160 (150–165) <0.05
Diabetes mellitus, % 23.3 31.8 <0.05 N2
Hyperlipidemia, % 43.6 44.1 0.48 Medical history, %
Hypertension, % 57.8 61.8 <0.05 Angina 62 No data –
Myocardial infarction, % 32 9.9 <0.05 Atrial fibrillation 15.5 No data –
PCI, % 14 12.9 <0.05 CABG 8 3.0 <0.05
Peripheral vascular 10.3 1.9 <0.05 Congestive heart failure 23.1 4.9 <0.05
disease, % Diabetes mellitus 30.1 34.9 0.12
Renal dysfunction, % 7.2 3.4 <0.05 Hyperlipidemia 27.5 27.9 0.94
Smoking, % 56.7 41.9 <0.05 Hypertension 62.8 62.3 0.89
Stroke, % 8.3 6.1 <0.05 Myocardial infarction 29.3 11.3 <0.05
Presentation characteristics PCI 7.7 10.4 0.15
Pulse, beats/min 76 (65–90) 79 (65–90) <0.05 Peripheral vascular 15.1 4.5 <0.05
DBP, mmHg 80 (70–90) 79 (67–92) N3 <0.05 disease
SBP, mmHg 140 (120–160) 136 (116–157) <0.05 Renal dysfunction 12.6 7.0 <0.05
Killip class I, % 82.7 78 <0.05 Smoking 43.7 28.9 <0.05
Killip class II, % 13.2 10.7 Stroke 13.4 12.1 0.63
Killip class III, % 3.1 4.7 Presentation characteristics
Killip class IV, % 1 6.6 Pulse, beats/min 87 (70–100) 87 (67–106) 0.61
Cardiac arrest, % 1.5 2.1 <0.05 DBP, mmHg 70 (60–87) 67 (52–80) N3 <0.05
Initial cardiac markers 31.6 38.2 <0.05 SBP, mmHg 126 (100–148) 110 (87–135) <0.05
positive, % Killip class I, % 49.2 26.2 <0.05
Initial serum 1.0 (0.9–1.2) 0.9 (0.7–1.1) <0.05 Killip class II, % 26.3 14.0
creatinine, mg/dl Killip class III, % 11.8 16.0
Conditions Killip class IV, % 12.7 43.8
STEMI, % 35.3 63.0 <0.05 Cardiac arrest, % 9.2 10.2 0.67
NSTEMI, % No data 15.5 Initial cardiac markers 55.3 76.0 <0.05
UA, % No data 21.5 positive, %
Mortality Initial serum 1.3 (1.0–1.7) 1.2 (0.9–1.8) 0.18
In-hospital death, % 4.6 5.0 0.22 creatinine, mg/dl
CABG, coronary artery bypass grafting; PCI, percutaneous coronary interven- CABG, coronary artery bypass grafting; PCI, percutaneous coronary interven-
tion; DBP, diastolic blood pressure; SBP, systolic blood pressure; N1, not tion; DBP, diastolic blood pressure; SBP, systolic blood pressure; N1, not
available 852 cases; N2, not available 970 cases; N3, not available 283 cases. available 110 cases; N2, not available 110 cases; N3, not available 58 cases.

Please cite this article in press as: Komiyama K, et al. In-hospital mortality analysis of Japanese patients with acute coronary syndrome
using the Tokyo CCU Network database: Applicability of the GRACE risk score. J Cardiol (2017), https://doi.org/10.1016/j.jjcc.2017.09.006
G Model
JJCC-1554; No. of Pages 8

4 K. Komiyama et al. / Journal of Cardiology xxx (2017) xxx–xxx

Table 3 population was linear and comparable to that of the GRACE study
Cause of death in the Tokyo CCU Network study.
(Fig. 3A). Among the specific conditions in the Tokyo CCU
Cause of death STEMI (n, 376) NSTEMI (n, 73) UA (n, 21) Network data, significant relationships between the GRACE risk
Cardiac shock, % (n) 27.4 (103) 31.5 (23) 23.8 (5)
score and in-hospital mortality were found for patients with
Congestive heart failure, % (n) 12.5 (47) 20.5 (15) 9.5 (2) STEMI (r = 0.99, p < 0.001), patients with NSTEMI (r = 0.89,
Cardiac rupture, % (n) 10.9 (41) 2.7 (2) 0.0 (0) p < 0.001), and the combined population of patients with STEMI
Arrhythmia, % (n) 8.5 (32) 5.5 (4) 4.8 (1) or NSTEMI (STEMI + NSTEMI; r = 0.99, p < 0.001) (Fig. 3B, C, and
Stent thrombosis, % (n) 0.5 (2) 1.4 (1) 4.8 (1)
E). However, for patients with UA, the correlation between
Multiple organ failure, % (n) 15.4 (58) 12.3 (9) 23.8 (5)
Infection, % (n) 0.8 (3) 0.0 (0) 0.0 (0) GRACE risk score and in-hospital mortality was low and non-
Hemorrhage, % (n) 0.0 (0) 0.0 (0) 0.0 (0) significant (r = 0.35, p = 0.126) (Fig. 3D).
Unknown, % (n) 23.9 (90) 26.2 (19) 33.3 (7) The receiver-operating characteristics curve and calibration
STEMI, ST-segment elevation myocardial infarction; NSTEMI, non-ST-segment plot for in-hospital mortality are shown in Fig. 4 (Tokyo CCU
elevation myocardial infarction; UA, unstable angina. network data). The c statistic for STEMI + NSTEMI was 0.87 (95%
confidence interval: 0.86–0.89), 0.84 for the GRACE model [3]. The
calibration plot was suggestive of a strong correlation between the
GRACE risk score and in-hospital mortality for patients with STEMI
peripheral vascular disease, renal dysfunction, and stroke com- or NSTEMI. Moreover, a GRACE risk score of 140 had the best
pared to that in the present study, implying that the GRACE combined sensitivity (97.0%) and specificity (58.2%) for the
population included more complex patients compared to that in development of in-hospital mortality. Of note, only 17 patients
the present study. In addition, there was a higher prevalence of (STEMI, 7 patients; NSTEMI, 7; UA, 3) with in-hospital death had a
Killip class III/IV, cardiac arrest, positive cardiac markers on GRACE score 2140 in the present study (Fig. 2).
admission, and STEMI among the present study population
compared to that in the GRACE population, indicating that the Predictors of mortality
patients in the present study had more severe presentation
characteristics (Table 1). This was further reinforced in the data The predictive values for potential risk factors of in-hospital
pertaining to in-hospital deaths; the prevalence of Killip class III/IV mortality are shown in Table 4. In the univariate analyses, a GRACE
and positive cardiac markers were also higher in the patients with score > 140, Killip class per higher class, and cardiac arrest on
in-hospital death in the present study compared to that in the admission were associated with the incidence of in-hospital
original GRACE study (Table 2). The distribution of the cause of mortality. In multivariate modeling, a GRACE score > 140 (odds
death for the present study is shown in Table 3 (that for the GRACE ratio, 8.82; 95% confidence interval, 5.28–14.74; p < 0.001)
study was not described in the original paper). Cardiac shock was remained independently associated with the incidence of in-
the most common cause of death for each of the three conditions hospital mortality.
(STEMI, NSTEMI, and UA).
Discussion
Relationship between the GRACE risk score and in-hospital mortality
The present study was the first to examine the suitability of the
The distribution of total GRACE risk score in Tokyo CCU GRACE risk scoring system for Japanese patients using a large
network is described in Fig. 2. The relationship between the sample. The results demonstrated three important findings with
GRACE risk score and in-hospital mortality of Tokyo CCU clinical implications: (1) a strong correlation between the GRACE
network are shown in Fig. 3. The GRACE risk score had a high, risk score and in-hospital deaths was found despite large
significant correlation with the overall in-hospital mortality differences in the patient backgrounds in the present study
(r = 0.99, p < 0.001). The relationship between the GRACE risk compared to those in the original GRACE study; (2) a GRACE
score and in-hospital mortality in the overall Tokyo CCU Network score > 140 indicated a high risk of in-hospital mortality for

300

250

200

150

100

50

0
STEMI STEMI NSTEMI NSTEMI UA UA
alive death alive death alive death

Fig. 2. The distribution of total GRACE risk score in Tokyo CCU network. STEMI, ST-segment elevation myocardial infarction; NSTEMI, non-ST-segment elevation myocardial
infarction; UA, unstable angina.

Please cite this article in press as: Komiyama K, et al. In-hospital mortality analysis of Japanese patients with acute coronary syndrome
using the Tokyo CCU Network database: Applicability of the GRACE risk score. J Cardiol (2017), https://doi.org/10.1016/j.jjcc.2017.09.006
G Model
JJCC-1554; No. of Pages 8

K. Komiyama et al. / Journal of Cardiology xxx (2017) xxx–xxx 5

Fig. 3. The relationship between the GRACE risk score and in-hospital mortality rate for Tokyo CCU Network data and GRACE registry data are shown for comparison. (A)
Overall population (STEMI + NSTEMI + UA); (B) STEMI alone; (C) NSTEMI alone; (D) UA alone; (E) STEMI + NSTEMI.
STEMI, ST-segment elevation myocardial infarction; NSTEMI, non-ST-segment elevation myocardial infarction; UA, unstable angina.

patients with STEMI or NSTEMI, which is the same as that in the
ESC guidelines [11]; and (3) a strong correlation between the
GRACE risk score and in-hospital mortality was found despite a
higher prevalence of vasospastic angina in Japan compared to that
in western countries.
The strong correlation between the GRACE risk score and in-
hospital mortality despite differences in patient backgrounds
between the present study and the global registry, confirms the
reliability and validity of the GRACE risk score as claimed by
previous studies. However, for patients with UA in Tokyo CCU
Network database, in-hospital mortality was not correlated with
the GRACE risk score. The mortality rate of patients with UA in the
present study was 1%. In contrast to the patients with STEMI or
NSTEMI, patients with UA did not experience myocardial necrosis
and had a substantially lower risk of death.
In the present study, a GRACE score > 140 was derived as the
cut-off value indicating a high risk of in-hospital morality for
patients with STEMI or NSTEMI. This is the same value as that in
the ESC guidelines [11], despite a higher prevalence of severe
diseases in the present study. The Tokyo CCU Network data
included more patients with severe pathophysiological condi-
tions requiring treatment in a care unit compared to that in the
GRACE population, comprising a much wider disease spectrum.
Thus, the present results further demonstrate that the GRACE risk
scoring system is a statistically well-considered scoring system;
even with differences in the patient characteristics, its use could
lead to superior results.
Furthermore, the present study demonstrated a strong correla-
tion between the GRACE risk score and in-hospital mortality
among Japanese patients with STEMI or NSTEMI, allaying the
doubts that “that there are limits to the use of GRACE risk score in
Japan where the prevalence of vasospastic angina is higher, and
suggest that a specifically Japanese risk score may be required”
stated by the Japanese Circulation Society [12,13] and others
regarding the suitability of the GRACE risk scoring system for
Japanese patients. The main basis for this reservation in accepting
a globally validated score was the higher prevalence of
vasospastic angina in Japan. Coronary artery spasm is a known
factor contributing to ACS [19]. However, there are racial
differences in the rate of coronary artery spasm, with reported

Please cite this article in press as: Komiyama K, et al. In-hospital mortality analysis of Japanese patients with acute coronary syndrome
using the Tokyo CCU Network database: Applicability of the GRACE risk score. J Cardiol (2017), https://doi.org/10.1016/j.jjcc.2017.09.006
G Model
JJCC-1554; No. of Pages 8

6 K. Komiyama et al. / Journal of Cardiology xxx (2017) xxx–xxx

Fig. 3. (Continued ).

rates of 11–21% in the western population [20,21] compared to a Study limitations

high rate of 69% among the Japanese population [22]. While the
present study did not examine the distribution of coronary
artery spasm, the Tokyo CCU Network database included mostly
Japanese patients, suggesting the inclusion of patients with
coronary artery spasm. However, the present results demon-
strate that the GRACE risk score is useful for optimal triage
despite differences in the pathophysiological background of
patients (e.g. an involvement of coronary spasm) between
Japanese patients with STEMI/NSTEMI and those enrolled in the
GRACE study.
Finally, in the present study, there was an unexpected dip in the
in-hospital mortality for GRACE scores of 220 and 3250 in patients
with NSTEMI (Fig. 3C). Two reasons will be considered. First, the
total calculated GRACE risk scores in Tokyo CCU network were
rounded to the nearest whole number. Second, Fig. 3C was
analyzed with only NSTEMI in Tokyo CCU network, so the number
of the population has decreased. There were only 24 patients for
220 points and 16 patients for 3250 points. For these reasons,
NSTEMI of Tokyo CCU network could not draw a sharp curve, such
as GRACE study.
The results of the present study should be interpreted in the
context of several potential limitations. First, differences exist in
the indications for admission and initial treatment strategy among
Tokyo CCUs, which may have resulted in some selection bias.
Second, there is a possibility that the definition of medical histories
differs slightly between GRACE registry and Tokyo CCU network. In
GRACE registry, the definition of medical histories conforms to
ICD-10 [18]. On the other hand, in Tokyo CCU network, medical
histories conform to the contents that each institution has
diagnosed based on domestic guidelines. Third, the GRACE risk
score has been validated not only for the estimation of in-hospital
mortality, but also for 6-month mortality in patients with ACS.
However, the Tokyo CCU Network database was a prospective
cohort database designed to record the patient information from
hospitalization to discharge in a pre-determined format. It was
impossible to trace patients following their discharge from the
hospitals. Therefore, the results of the present study were confined
to in-hospital mortality alone, and do not provide any data
concerning long-term mortality and/or the difference of prognosis

Please cite this article in press as: Komiyama K, et al. In-hospital mortality analysis of Japanese patients with acute coronary syndrome
using the Tokyo CCU Network database: Applicability of the GRACE risk score. J Cardiol (2017), https://doi.org/10.1016/j.jjcc.2017.09.006
G Model
JJCC-1554; No. of Pages 8

K. Komiyama et al. / Journal of Cardiology xxx (2017) xxx–xxx 7

Fig. 4. Receiver-operating characteristics (ROC) curve and calibration plot for in-hospital mortality of patients with STEMI or NSTEMI (Tokyo CCU network data). (A) The ROC
curve calculated using the total GRACE risk score. The c statistic was 0.84 (95% confidence interval: 0.77–0.90). A GRACE score cut-off of 140 had the best combined sensitivity
(97.0%) and specificity (58.2%) for the development of in-hospital mortality. (B) The calibration plot is shown. The diagonal line indicates perfect calibration.

Table 4
Predictive value of potential risk factors for in-hospital mortality (Tokyo CCU network data).

Univariate analysis Multivariate analysis

Odds ratio 95% CI p value Odds ratio 95% CI p value

Age, per 10-year increase 1.84 1.69–2.00 <0.001

Female 1.69 1.39–2.6 <0.001
CABG 1.42 0.82–2.47 0.21
Congestive heart failure 2.99 1.91–4.67 <0.001 1.77 1.07–2.94 0.028
Diabetes mellitus 1.16 0.95–1.41 0.14
Hypertension 1.02 0.85–1.24 0.82
Hyperlipidemia 0.47 0.39–0.58 <0.001
Myocardial infarction 1.17 0.87–1.57 0.30
Previous PCI 0.78 0.58–1.05 0.10
Peripheral vascular disease 2.65 1.66–4.22 <0.001 1.75 1.03–2.95 0.038
Smoking 0.55 0.49–0.67 <0.001
Stroke 2.24 1.68–3.00 <0.001 1.53 1.11–2.11 0.01
Pulse, per 20-beats/min increase 1.34 1.23–1.45 0.001
SBP, per 20 mmHg decrease 1.81 1.69–1.94 0.01
Killip class, per higher class 3.06 2.83–3.31 <0.001 2.32 2.13–2.52 <0.001
Cardiac arrest 6.80 4.84–9.55 <0.001 1.79 1.22–2.63 0.003
ST deviation 2.96 2.30–3.81 <0.001 1.28 0.97–1.69 0.086
Initial cardiac marker positive 2.01 1.62–2.50 <0.001 1.50 1.18–1.91 0.001
Initial serum creatinine level, per 0.4 mg/dl increase 1.24 1.20–1.27 <0.001
GRACE risk score > 140 28.11 17.29–45.70 <0.001 8.71 5.21–14.55 <0.001

CI, confidence interval; CABG, coronary artery bypass grafting; PCI, percutaneous coronary intervention; SBP, systolic blood pressure.

due to the difference in the treatment content [23,24]. Finally, the GRACE risk score can help clinicians stratify patients by risk,
distribution of coronary artery spasm in the present study, as well optimizing patient triage and management.
as in the GRACE study, has not been elucidated.

Conclusions
This present study demonstrated that there is a strong
correlation between the GRACE risk score and in-hospital
mortality for Japanese patients with STEMI or NSTEMI; however,
the GRACE risk score was not correlated with in-hospital mortality
for Japanese patients with UA. Furthermore, a GRACE risk score
cut-off value of 140 had the best combined sensitivity (97.0%) and
specificity (58.2%) for indicating the development of in-hospital
mortality. Thus, in Japanese patients with STEMI or NSTEMI, the
Funding
This research was not funded by any funding agency in the
public, commercial, or not-for-profit sectors.
Conflict of interest
The authors have no financial conflicts of interest to disclose
concerning the manuscript. The Tokyo CCU network data registry is
financially supported by the Tokyo Metropolitan Government.

Please cite this article in press as: Komiyama K, et al. In-hospital mortality analysis of Japanese patients with acute coronary syndrome
using the Tokyo CCU Network database: Applicability of the GRACE risk score. J Cardiol (2017), https://doi.org/10.1016/j.jjcc.2017.09.006
G Model
JJCC-1554; No. of Pages 8
8 K. Komiyama et al. / Journal of Cardiology xxx (2017) xxx–xxx
Acknowledgments
The authors gratefully thank Ms Nobuko Yoshida of the Tokyo
CCU Network for assistance with the writing of the manuscript.
References
[1] World Health Organization. Cardiovascular diseases: fact sheet number
310. Available at: http://www.who.int/mediacentre/factsheets/fs310/en/
[accessed 03.02.17].
[2] Ohira T, Iso H. Cardiovascular disease epidemiology in Asia: an overview. Circ J
2013;77:1646–52.
[3] Granger CB, Goldberg RJ, Dabbous O, Pieper KS, Eagle KA, Cannon CP, Van De
Werf F, Avezum A, Goodman SG, Flather MD, Fox KA. Predictors of hospital
mortality in the global registry of acute coronary events. Arch Intern Med
2003;163:2345–53.
[4] Sabatine MS, Morrow DA, Giugliano RP, Murphy SA, Demopoulos LA, DiBattiste
PM, Weintraub WS, McCabe CH, Antman EM, Cannon CP, Braunwald E. Impli-
cations of upstream glycoprotein IIb/IIIa inhibition and coronary artery stent-
ing in the invasive management of unstable angina/non-ST-elevation
myocardial infarction: a comparison of the Thrombolysis In Myocardial In-
farction (TIMI) IIIB trial and the Treat angina with Aggrastat and determine
Cost of Therapy with Invasive or Conservative Strategy (TACTICS)-TIMI 18 trial.
Circulation 2004;109:874–80.
[5] Boersma E, Pieper KS, Steyerberg EW, Wilcox RG, Chang WC, Lee KL, Akkerhuis
KM, Harrington RA, Deckers JW, Armstrong PW, Lincoff AM, Califf RM, Topol EJ,
Simoons ML. Predictors of outcome in patients with acute coronary syndromes
without persistent ST-segment elevation. Results from an international trial of
9461 patients. The PURSUIT Investigators. Circulation 2000;101:2557–67.
[6] Fox KA, Dabbous OH, Goldberg RJ, Pieper KS, Eagle KA, Van de Werf F, Avezum
A, Goodman SG, Flather MD, Anderson Jr FA, Granger CB. Prediction of risk of
death and myocardial infarction in the six months after presentation with
acute coronary syndrome: prospective multinational observational study
(GRACE). BMJ 2006;333:1091.
[7] Pollack Jr CV, Sites FD, Shofer FS, Sease KL, Hollander JE. Application of the TIMI
risk score for unstable angina and non-ST elevation acute coronary syndrome
to an unselected emergency department chest pain population. Acad Emerg
Med 2006;13:13–8.
[8] Abu-Assi E, Ferreira-Gonzalez I, Ribera A, Marsal JR, Cascant P, Heras M, Bueno
H, Sánchez PL, Arós F, Marrugat J, García-Dorado D, Peña-Gil C, González-
Juanatey JR, Permanyer-Miralda G. Do GRACE (Global Registry of Acute Coro-
nary events) risk scores still maintain their performance for predicting mor-
tality in the era of contemporary management of acute coronary syndromes?
Am Heart J 2010;160:826–34. e821–823.
[9] Eagle KA, Lim MJ, Dabbous OH, Pieper KS, Goldberg RJ, Van de Werf F, Goodman
SG, Granger CB, Steg PG, Gore JM, Budaj A, Avezum A, Flather MD, Fox KA,
GRACE Investigators. A validated prediction model for all forms of acute
coronary syndrome: estimating the risk of 6-month postdischarge death in
an international registry. JAMA 2004;291:2727–33.
[10] Killip 3rd T, Kimball JT. Treatment of myocardial infarction in a coronary care
unit. A two year experience with 250 patients. Am J Cardiol 1967;20:457–64.
[11] Roffi M, Patrono C, Collet JP, Mueller C, Valgimigli M, Andreotti F, Bax JJ, Borger
MA, Brotons C, Chew DP, Gencer B, Hasenfuss G, Kjeldsen K, Lancellotti P,
Landmesser U, et al. 2015 ESC Guidelines for the management of acute
coronary syndromes in patients presenting without persistent ST-segment
Please cite this article in press as: Komiyama K, et al. In-hospital mortality analysis of Japanese patients with acute coronary syndrome
using the Tokyo CCU Network database: Applicability of the GRACE risk score. J Cardiol (2017), https://doi.org/10.1016/j.jjcc.2017.09.006
elevation: Task Force for the Management of Acute Coronary Syndromes in
Patients Presenting without Persistent ST-Segment Elevation of the European
Society of Cardiology (ESC). Eur Heart J 2016;37:267–315.
[12] Kimura K, Asai T, Ogawa H, Okumura K, Kimura T, Goto Y, Sumiyoshi T, Daida H,
Tanaka T, Nagao K, Hirayama A, Mizuno K, Miyazaki S, Yamashina A, Yokoyama
S, et al., The Japanese Circulation Society. Guidelines for the management of
patients with ST-elevation acute myocardial infarction (JCS 2013); 2013 [in
Japanese].
[13] Kimura T, Isshiki T, Oono T, Ogawa H, Kimura K, Sakata R, Sumiyoshi T,
Takanashi S, Kayano M, Tsutsui H, Nakao K, Nakagawa Y, Nakamura M, Nonoki
H, Hirayama H, et al., The Japanese Circulation Society. Guidelines for man-
agement of acute coronary syndrome without persistent ST segment elevation
(JCS 2012); 2012 [in Japanese].
[14] Takano T, Tanaka K, Endo T, Nagamine K, Katoh T, Hayakawa H, Hirosawa K.
CCU network as primary care of acute myocardial infarction. Circ J
1984;48:690–7.
[15] Tanabe Y, Obayashi T, Yamamoto T, Nakata J, Yagi H, Takayama M, Nagao K.
Current status of the use of inferior vena cava filters in cases of pulmonary
embolism in CCUs: from the Tokyo CCU Network. J Cardiol 2014;63:385–9.
[16] Tokyo CCU Network Scientific Committee. Latest management and outcomes
of major pulmonary embolism in the cardiovascular disease early transport
system: Tokyo CCU Network. Circ J 2010;74:289–93.
[17] Tunstall-Pedoe H, Kuulasmaa K, Amouyel P, Arveiler D, Rajakangas AM, Pajak A.
Myocardial infarction and coronary deaths in the World Health Organization
MONICA Project. Registration procedures, event rates, and case-fatality rates in
38 populations from 21 countries in four continents. Circulation 1994;90:
583–612.
[18] Investigators GRACE. Rationale and design of the GRACE (Global Registry of
Acute Coronary Events) Project: a multinational registry of patients hospital-
ized with acute coronary syndromes. Am Heart J 2001;141:190–9.
[19] Lin CS, Penha PD, Zak FG, Lin JC. Morphodynamic interpretation of acute
coronary thrombosis, with special reference to volcano-like eruption of ath-
eromatous plaque caused by coronary artery spasm. Angiology 1988;39:
535–47.
[20] Bertrand ME, Lablanche JM, Tilmant PY, Thieuleux FA, Delforge MG, Chahine
RA. The provocation of coronary arterial spasm in patients with recent
transmural myocardial infarction. Eur Heart J 1983;4:532–5.
[21] Mongiardo R, Finocchiaro ML, Beltrame J, Pristipino C, Lombardo A, Cianflone
D, Mazzari MA, Maseri A. Low incidence of serotonin-induced occlusive
coronary artery spasm in patients with recent myocardial infarction. Am J
Cardiol 1996;78:84–7.
[22] Okumura K, Yasue H, Matsuyama K, Ogawa H, Morikami Y, Obata K,
Sakaino N. Effect of acetylcholine on the highly stenotic coronary artery:
difference between the constrictor response of the infarct-related coro-
nary artery and that of the noninfarct-related artery. J Am Coll Cardiol
1992;19:752–8.
[23] Raposeiras-Roubín S, Abu-Assi E, López-López A, Bouzas-Cruz N, Castiñeira-
Busto M, Cambeiro-González C, Álvarez-Álvarez B, Virgós-Lamela A, Varela-
Román A, García-Acuña JM, González-Juanatey JR. Risk stratification for the
development of heart failure after acute coronary syndrome at the time of
hospital discharge: predictive ability of GRACE risk score. J Cardiol
2015;66:224–31.
[24] Liu HL, Jin ZG, Yang SL, Han W, Jing QM, Zhang L, Luo JP, Ma DX, Liu Y, Yang LX,
Jiang TM, Qu P, Li WM, Li SM, Xu B, et al. Five-year outcomes of ST-elevation
myocardial infarction versus non-ST-elevation acute coronary syndrome trea-
ted with biodegradable polymer-coated sirolimus-eluting stents: Insights
from the CREATE trial. J Cardiol 2017;69:149–55.