International Journal of Cardiology 301 (2020) 108–113

Contents lists available at ScienceDirect

International Journal of Cardiology

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

Characteristics and in-hospital mortality of patients with myocardial

infarction in the absence of obstructive coronary artery disease in super-
aging society
Masanobu Ishii a,b,c, Koichi Kaikita a, ∗, Kenji Sakamoto a, Tomotsugu Seki b, Koji Kawakami b, Michikazu Nakai c,
Yoko Sumita c, Kunihiro Nishimura c, Yoshihiro Miyamoto c, Teruo Noguchi c, Satoshi Yasuda c,
Hiroyuki Tsutsui e, Issei Komuro f, Yoshihiko Saito d, Hisao Ogawa d, Kenichi Tsujita a, on the behalf of
JROAD Investigators
a
Graduate School of Medical Sciences, Kumamoto University, Kumamoto, Japan
b
Graduate School of Medicine and Public Health, Kyoto University, Kyoto, Japan
c
National Cerebral and Cardiovascular Center, Suita, Japan
d
School of Medicine, Nara Medical University, Kashihara, Japan
e
Faculty of Medical Sciences, Kyusyu University, Fukuoka, Japan
f
Graduate School of Medicine and Faculty of Medicine, The University of Tokyo, Tokyo, Japan

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

Article history: Background: Myocardial infarction with nonobstructive coronary arteries (MINOCA) is a common presentation of
Received 14 August 2019 acute myocardial infarction (AMI) and has a better prognosis. However, there were few reports on large-scale,
Received in revised form 11 September 2019 high aged population. The aim of this study was to determine the differences in the clinical characteristics and
Accepted 16 September 2019 short-term prognosis between MINOCA and myocardial infarction with obstructive coronary artery disease
Available online 25 October 2019
(MI-CAD) using a nationwide administrative database in the super-aging society, Japan.
Methods: This was an observational study using data of 137,678 AMI patients who underwent angiography be-
tween April 2012 through March 2016. Using the international classification of diseases 10th revision, AMI pa-
tients were divided into two groups based on the presence or absence of revascularization and coronary
atherosclerosis, identifying 123,633 MI-CAD and 14,045 working diagnosis of MINOCA patients. The true
MINOCA (n = 13,022) was defined as the MINOCA excluding non-ischemic causes. We assessed in-hospital mor-
tality within 30 days.
Results: Both MINOCA groups were typically found in non-obese, non-smoker young females, with a low grade on
Killip classification, and non-low ADL status. Compared to MI-CAD, chronic pulmonary diseases, peripheral vas-
cular diseases, liver diseases, renal diseases, and cerebrovascular diseases were more common, whereas diabetes
was less common in the MINOCA groups. In-hospital mortality within 30 days was higher in both MINOCA
groups than in MI-CAD. Multivariate frailty model identified both MINOCA groups as an independent risk factor
for in-hospital mortality.
Conclusions: Our large-population study demonstrated that MINOCA was associated with a high risk of in-
hospital mortality compared with MI-CAD in the super-aging society.
© 2019 Elsevier B.V. All rights reserved.

1. Introduction atherosclerosis [1]. However, myocardial infarction with nonobstructive

Acute myocardial infarction (AMI) is one of the major causes of heart
failure and sudden death, and its pathology is related to sudden plaque
disruption and acute coronary occlusion in association with
∗ Corresponding author. Department of Cardiovascular Medicine, Graduate School of
Medical Sciences, Kumamoto University, 1-1-1, Honjo, Chuo-ku, Kumamoto, 860-8556,
Japan.
E-mail address: kaikitak@kumamoto-u.ac.jp (K. Kaikita).
coronary arteries (MINOCA), defined as no angiographic coronary ste-
nosis (≥50%) is not an uncommon form of AMI [2–7]. The etiology of
MINOCA includes coronary spasm, microvascular spasm, microvascular
dysfunction, coronary thromboembolism, coronary dissection, other
forms of type 2 myocardial infarction, myocarditis, and Takotsubo car-
diomyopathy [4].
It is often difficult to evaluate the prognosis of MINOCA based on the
wide-range pathology and there is little or no consensus on prognosis. A
systematic review showed better in-hospital prognosis and 12-months

https://doi.org/10.1016/j.ijcard.2019.09.037
0167-5273/© 2019 Elsevier B.V. All rights reserved.
 M. Ishii et al. / International Journal of Cardiology 301 (2020) 108–113 109

mortality rate of patients with MINOCA compared to those of myocar-
dial infarction with coronary artery disease (MI-CAD) [3]. On the other
hand, a recent prospective observational study concluded there was
no difference in the prognosis of MINOCA and MI-CAD patients [8].
The discrepancy in the reported prognosis could be due to differences
in the diagnostic criteria, sample size, and patient population. In addi-
tion, because of the most super-aging society of Japan in the world, pre-
viously reported results for young population may differ from those for
the elderly.
The present large-scale study was designed to determine the differ-
ence in clinical characteristics and short-term prognosis of patients with
MINOCA and MI-CAD. Subjects of the study were registered in the na-
tionwide administrative Diagnostic Procedure Combination (DPC) data-
base, the Japanese Of All cardiac and vascular Diseases (JROAD)-DPC
[9–11], which is large enough to allow an accurate estimate of the
cause-specific risk of MINOCA.
2. Methods
2.1. Study design and population
This retrospective observational study was conducted using the nationwide JROAD-
DPC discharge database [9–11], which included data of 3,626,656 health records from
953 Japanese Circulation Society-certificated hospitals between April 1, 2012, through
March 31, 2016.
We enrolled 163,471 AMI patients in this study, aged 20 years or older, who were di-
agnosed with AMI, defined as the International Classification of Disease, Tenth Revision
(ICD-10) codes of I21.0, I21.1, I21.2, I21.3, I21.4, and I21.9, which was recorded in the
“main diagnosis”, “admission precipitating diagnosis”, “most resource-consuming diagno-
sis”, or “second most resource-consuming diagnosis” (Fig. 1). ICD-10 code for AMI (I21)
was validated in another study using the Japanese administrative database [12]. Out of
the 163,471 patients, we excluded 25,793 patients because of the following reasons; re-
hospitalization for AMI (n = 2865), diagnosis of old myocardial infarction (ICD-10
codes; I22.0, I22.1, I22.8, I22.9, I25.2, I25.5, I25.8, and I25.9, n = 3510), and neither coro-
nary angiography nor re-vascularization procedure were performed (n = 19,418). Finally,
137,678 patients with first admission to the hospital for AMI during the study period were
included as subjects of the study. They were divided into three groups; The first, the work-
ing diagnosis of MINOCA(5) group, included individuals with neither codes for percutane-
ous or surgical revascularization as procedure, nor for coronary artery stenosis (I25.1) as
comorbidity, included in the records. The second group, the MI-CAD group, included pa-
tients with either of the above two codes were recorded. The current universal definition
recommended that non-ischemic causes such as myocarditis or Takotsubo cardiomyopa-
thy were excluded(1). Therefore, we analyzed in the population excluding the non-
ischemic causes [Takotsubo cardiomyopathy (ICD-10 code; I518, n = 777), myocarditis,
(ICD-10 codes; I400, I401, I408, I409, I410, I411, I412, I418, I423, I514, I012, and I090, n
= 167) and pulmonary emboli (ICD-10 code; I269, n = 79)] from the working diagnosis
of MINOCA group, identifying 13,022 true MINOCA patients.
For the prohibition of linkage of the administrative database with medical records in
Japan, we investigated the accuracy of the above algorithm identifying MINOCA patients
by an independent validation study with nested case-control design in Kumamoto univer-
sity hospital, as previously described [13]. Details in the validation were provided in the
Supplementary Materials. The validation study indicated that the algorithm for MINOCA,
which is combined with ICD-10, diagnostic, and procedure codes, has a positive predictive
value of 100%.
2.2. Data collection
The following data were obtained from the JROAD-DPC database; age, sex, body mass
index, duration of hospitalization, smoking status (Brinkman index; 0 or N0), Killip

Fig. 1. Study Flow Chart. This chart shows enrollment criteria and the flow of patients who were divided into two groups according to the presence or absence of revascularization and
coronary atherosclerosis. Patients of the working diagnosis of MINOCA group did not undergo percutaneous or surgical revascularization, and had no coronary artery stenosis (code
I25.1), whereas those of the MI-CAD group had at least one of the above. OMI = old myocardial infarction, JROAD = Japanese Registry Of All cardiac and vascular Diseases, DPC = Diagnosis
Procedure Combination, MI-CAD = myocardial infarction with coronary artery disease, MINOCA = myocardial infarction with nonobstructive coronary arteries.
110 M. Ishii et al. / International Journal of Cardiology 301 (2020) 108–113

classification for AMI, activities of daily living (ADL) score (Barthel index; 100 or b100), co- Table 1
morbidities (chronic pulmonary disease, peripheral vascular disease, cerebrovascular dis- Baseline characteristics of patients with AMI who were admitted for the first time in the
ease, liver disease, diabetes with or without chronic complications, renal disease, and JROAD- DPC Institutes from April 2012 through March 2016.
malignancy) defined as Charlson comorbidity index [14,15], in-hospital use of medica-
tions (angiotensin-converting enzyme inhibitors, angiotensin Ⅱ receptor blockers, β- AMI, n = 137,678
blockers, calcium channel blockers, statins, anti-platelet drugs, and oral anticoagulants), MI-CAD n = Working diagnosis of
number of hospital beds, and hospital teaching status. The cause of MINOCA was defined 123,633 (89.8%) MINOCA
as follows; ICD10 code for coronary spasm; I201, coronary dissection; I248, coronary em- n = 14,045 (10.2%)
boli; I240, Takotsubo cardiomyopathy; I518, myocarditis; I400, I401, I408, I409, I410, I411,
I412, I418, I423, I514, I012, I090, pulmonary emboli; I269, aortic dissection; and I710, I711, Age, median [IQR] 69 [60–78] 71 [61–80]
I712, I713, I714, I715, I718, which were recorded in “main diagnosis”, “admission precip- ≥20 to b50 years, n (%) 10,954 (8.9) 1446 (10.3)
itating diagnosis”, “most resource-consuming diagnosis”, “second most resource- ≥50 to b65 33,333 (27.0) 3227 (23.0)
consuming diagnosis”, and “comorbidity”. “Unidentified” cause was defined as MINOCA ≥65 to b75 36,219 (29.3) 3803 (27.1)
that was not associated with coronary spasm, coronary dissection, coronary emboli, ≥75 43,127 (34.9) 5569 (39.7)
Takotsubo cardiomyopathy, myocarditis, pulmonary emboli, or aortic dissection. Males, n (%) 93,470 (75.6) 8871 (63.2)
Body mass index, kg/m2 23.6 [21.4–26.0] 22.9 [20.4–25.5]
2.3. Outcome Body mass index, not recorded, n (%) 10,129 (8.2) 1438 (10.2)
In-hospital days, median [IQR] 14 [10–20] 10 [4–18]
The primary outcome was in-hospital mortality, defined as all-cause death within 30 Smoking, n (%) 60,322 (55.4) 5553 (45.3)
days. The secondary outcome was all-cause death within 24 h and within 30 days not in- Smoking, not recorded, n (%) 14,700 (11.9) 1778 (12.7)
cluding deaths within 24 h. Killip classification, n (%)
Killip 1 57,625 (47.7) 5636 (51.4)
Killip 2 34,145 (28.3) 2559 (23.3)
2.4. Ethical considerations
Killip 3 9833 (8.1) 800 (7.3)
Killip 4 15,536 (12.9) 998 (9.1)
This study protocol was approved by the ethics committees of the Japanese Circula-
Unknown 6494 (5.3) 4052 (28.9)
tion Society and each participating clinical facility. No informed consent was needed be-
Killip ≥3 25,369 (25.8) 1798 (18.0)
cause the anonymized data did not include information that identified the participating
Chronic pulmonary disease, n (%) 2795 (2.3) 487 (3.5)
individuals. After the original DPC data on patient information were anonymized at each
Peripheral vascular disease, n (%) 4935 (4.0) 712 (5.1)
hospital using the code change equations, the anonymized data were sent to the Ministry
Cerebrovascular disease, n (%) 5674 (4.6) 860 (6.1)
of Health, Labor and Welfare. All patients had the opportunity to opt-out, as advertised
Liver disease, n (%) 1330 (1.1) 299 (2.1)
through posters or internet websites that notified them on the inclusion of their informa-
Diabetes, n (%) 37,627 (30.4) 3141 (22.4)
tion in this study.
Renal disease, n (%) 5575 (4.5) 814 (5.8)
Malignancy, n (%) 3250 (2.6) 612 (4.4)
2.5. Statistical analysis
Low ADL status on admission, n (%) 86,810/118,828 8732/13,616 (64.1)
(73.1)
Data were expressed as the median and interquartile range (IQR) for continuous var-
Hospital with the number of hospital 56,407 (45.6) 6514 (46.4)
iables, and numbers (percentages) for categorical variables. To estimate the odds ratios
beds ≥500
(ORs) and 95% confidence intervals (95% CIs) for factors associated with MINOCA, a mul-
Hospital teaching status 122,802 (99.3) 13,857 (98.7)
tivariable generalized linear mixed model was applied with random intercepts to account
Medications
for clinical institution-related variation [16]. To estimate the hazard ratios (HRs) and 95%
ACE inhibitors, n (%) 60,849 (49.2) 3377 (24.0)
CIs for risk factors of in-hospital mortality, a Cox frailty model was performed with random
ARB, n (%) 33,034 (26.7) 2960 (21.1)
intercepts to account for clinical institution-related variation [17]. Since there was no in-
Beta blockers, n (%) 79,190 (64.1) 5399 (38.4)
formation about death of the patients after discharge from the hospital, the date of dis-
CCBs, n (%) 25,976 (21.0) 5545 (39.5)
charge was used as the censor date. Landmark analysis for in-hospital mortality was
Statins, n (%) 100,402 (81.2) 6350 (45.2)
conducted at the time of 24 h because we considered that the diagnosis and cause-
Anti-platelet drugs, n (%) 120,914 (97.8) 10,651 (75.8)
related treatment of patients with MINOCA require some time, although patients with
Aspirin, n (%) 118,546 (95.9) 10,065 (71.7)
MI-CAD received revascularization to improve the prognosis. Age, sex, obesity, smoking
Clopidogrel, n (%) 91,663 (74.1) 5538 (39.4)
status, Killip classification, comorbidities, ADL status, hospital features, and AMI subtype
Oral anti-coagulants, n (%) 15,471 (12.5) 2940 (20.9)
were entered into the forced entry method in the multivariate models. Two multivariate
models were performed; 1) model 1 included age, sex, obesity, smoking status, Killip clas- AMI = acute myocardial infarction, JROAD = Japanese Registry Of All cardiac and vascular
sification, comorbidities, ADL status, hospital features, and AMI subtype, 2) model 2 in- Diseases, MI-CAD = myocardial infarction with coronary artery disease, MINOCA = myo-
cluded the variables listed above for model 1 but excluding comorbidities since patients cardial infarction with nonobstructive coronary arteries, ADL = activities of daily living,
who died earlier lost the opportunity for diagnosis of comorbidities. The rates of missing ACE = angiotensin converting enzyme, ARB = angiotensin Ⅱ receptor blocker, CCB = cal-
data in the above two models were as follows; 8.2%, 10.2% for body mass index; 11.9%, cium channel blocker, IQR = interquartile range.
12.7% for smoking status; and 5.3%, 28.9% for Killip classification for MI-CAD and working
diagnosis of MINOCA patients, respectively (Table 1). Multiple imputation was used to
handle missing data and was performed with 20 imputed datasets generated by the 3.2. AMI subtypes and in-hospital mortality
fully conditional specification method. The results across 20 imputed datasets were com-
bined using Rubin's rules [18]. A two-sided p value of b0.05 was considered to denote the The median duration of hospitalization was 10 [IQR; 4–18] and 14 [IQR; 10–20] days
presence of a statistically significant difference. All statistical analyses were conducted for working diagnosis of MINOCA and MI-CAD, respectively (Table 1). During the study pe-
using SAS 9.4 (SAS Institute, Cary, NC). riod, 894 working diagnosis of MINOCA (6.4%, No. of event/No. of patient; 894/14,045) and
7644 MI-CAD (6.2%, No. of event/No. of patient; 7644/123,633) patients developed in-
hospital death and the in-hospital mortality rates were 4.48, 4.57 and 3.46/1000 person-
3. Results
days for working diagnosis of MINOCA, true MINOCA, and MI-CAD, respectively
(Table 3). Univariate Cox frailty model showed that working diagnosis of MINOCA and
3.1. Patient characteristics
true MINOCA were significantly associated with in-hospital mortality, deaths within
24 h, and deaths within 30 days (excluding deaths within 24 h). Furthermore, in each
Working diagnosis of MINOCA (median age 71 years old, men 63.2%) and MI-CAD
model, multivariate Cox frailty model identified working diagnosis of MINOCA and true
(median age 69 years old, men 75.6%) were diagnosed in 14,045 (10.2%) and 123,633
MINOCA as a significant predictor of in-hospital mortality in AMI patients (Table 3).
(89.8%) of 137,678 patients with the first admission for AMI included in the present anal-
ysis, respectively (Fig. 1). In terms of clinical characteristics (Table 1), those MINOCA pa-
tients tended to be female non-smokers who were less likely to have diabetes and low 3.3. Potential causes and predictors of in-hospital mortality in MINOCA
ADL status on admission, and were more likely to have lower body mass index, lower
Killip classification, chronic pulmonary disease, peripheral vascular disease, cerebrovascu- Sub-analysis of the working diagnosis of MINOCA group showed that coronary spasm
lar disease, liver disease, renal disease, and malignancy, compared with MI-CAD patients. was the most common background factor in working diagnosis of MINOCA (n = 1,647,
According to the generalized linear mixed model (Table 2), multivariable analysis 11.7%), followed by Takotsubo cardiomyopathy (n = 777, 5.5%) (Supplementary
identified male sex, obesity, smoking, Killip ≥3, diabetes, and low ADL status on admission Table 1). Cox frailty models identified predictors of in-hospital mortality in working diag-
as significant negative predictors, whereas younger (≥20 age b 50 years), chronic pulmo- nosis of MINOCA patients (Supplementary Table 2). Multivariate Cox frailty model identi-
nary disease, peripheral vascular disease, cerebrovascular disease, liver disease, renal dis- fied coronary spam and Takotsubo cardiomyopathy as significant negative predictors of
ease, and malignancy, as significant positive predictors of working diagnosis of MINOCA, in in-hospital mortality of working diagnosis of MINOCA patients in each model, and myo-
consistent with the results of true MINOCA group. carditis and aortic dissection as significant positive predictors.
 M. Ishii et al. / International Journal of Cardiology 301 (2020) 108–113 111

Table 2 SWEDEHEART (the Swedish Web-system for Enhancement and Devel-

Results of a multivariable generalized linear mixed model for clinical characteristics of opment of Evidence-based care in Heart disease Evaluated According
working diagnosis of MINOCA and true MINOCA.
to Recommended Therapy) registry reported that 8.0% of 118,260 AMI
working diagnosis of true MINOCA patients who underwent angiography in Sweden were diagnosed with
MINOCA MINOCA when the discharge diagnosis was AMI (ICD-10 code: I21–
Odds ratio 95% CI Odds ratio 95% CI I22) and that these patients had b50% stenosis on angiography during
Age the index hospitalization [19]. Furthermore, a systematic review of
≥20 to b50 years, n (%) 1.45 1.36, 1.56 1.45 1.33, 1.58 176,502 angiographically diagnosed AMI patients that were included
≥50 to b65 1.00 0.95, 1.05 1.01 0.95, 1.08 in 27 studies with a reported prevalence of MINOCA ranging from 1%
≥65 to b75 0.95 0.91, 0.99 0.98 0.92, 1.04 to 14%, showed the average prevalence of MINOCA was 6% (95%CI,
≥75 Ref Ref
Male 0.59 0.56, 0.62 0.65 0.62, 0.69
5–7%) [3].
Obesity (BMI ≥25 kg/m2) 0.83 0.80, 0.87 0.89 0.84, 0.93 The present study demonstrated that the prevalence of coronary
Smoking 0.79 0.75, 0.83 0.81 0.77, 0.86 spasm in the working diagnosis of MINOCA was 11.7%, which was ex-
Killip classification ≥3 0.75 0.71, 0.80 0.79 0.74, 0.84 tremely lower than that reported in previous Japanese studies [20,21].
Chronic pulmonary disease 1.57 1.42, 1.74 1.36 1.18, 1.56
Deyama et al. [20] reported that the prevalence of coronary spasm
Peripheral vascular disease 1.27 1.16, 1.39 1.22 1.08, 1.37
Cerebrovascular disease 1.40 1.30, 1.52 1.46 1.32, 1.62 was 71.4% in MINOCA among 437 AMI patients who underwent
Liver disease 2.03 1.78, 2.32 2.05 1.73, 2.42 acetylcholine-provocation test. Furthermore, Nakayama et al. [21] re-
Diabetes 0.68 0.65, 0.71 0.70 0.67, 0.74 ported that the prevalence of coronary spasm was 71.3% in 449 patients
Renal disease 1.41 1.31, 1.53 1.54 1.40, 1.70 who presented with non-ST elevation acute coronary syndrome with-
Malignancy 1.59 1.45, 1.75 1.64 1.45, 1.85
Low ADL status on admission 0.61 0.58, 0.64 0.62 0.58, 0.65
out culprit lesion. On the other hand, one review study estimated the
prevalence of coronary spasm at 28% based on 14 studies that included
Abbreviations as in Table 1.
402 MINOCA patients who underwent spasm provocation test, includ-
ing one Japanese study that reported a high prevalence (81%) [3], and
4. Discussion another (the VIRGO study) showed extremely low prevalence of coro-
nary spasm (3.7%) in 299 MINOCA patients [8]. Thus, the prevalence
The major findings of this study were as follows: 1) among AMI pa- of coronary spasm in MINOCA was quite different among the studies al-
tients, working diagnosis of MINOCA and true MINOCA correlated neg- though the prevalence of MINOCA was almost identical. These differ-
atively with traditional coronary risk factors (old age, obesity, smoking, ences could be due to the population samples studied, including race,
and diabetes), but correlated positively with comorbidities (chronic criteria used for diagnosis, and protocol of the spasm provocation test-
pulmonary disease, peripheral vascular disease, cerebrovascular dis- ing, including the dose used, type of agent applied in the test (acetylcho-
ease, liver disease, renal disease, and malignancy); 2) true MINOCA line or ergonovine), and/or mode of administration (intra-coronary
was significantly associated with high risk of in-hospital mortality, infusion or intra-venous infusion). To investigate the epidemiology of
even if working diagnosis of MINOCA; and 3) coronary spasm and coronary spasm-induced MINOCA, international standardized diagnos-
Takotsubo cardiomyopathy were significant negative predictors of in- tic criteria [22] should be used in clinical settings. Furthermore, the rea-
hospital MINOCA-related mortality, whereas myocarditis and aortic dis- son why the prevalence of coronary spasm was small might be an
section were significant positive predictors. To the best of our knowl- underestimation due to not actively performance of diagnostic tests
edge, this is the first large population study using a nationwide such as provocation testing for coronary spasm at all participating hos-
administrative data to investigate the clinical characteristics and pitals in this study. In addition, the previous studies were reported by a
short-term prognosis of patients with MINOCA compared to those single or several centers where the diagnostic test for coronary spasm
with MI-CAD in the super-aging society. was actively performed, which might cause biases publication and se-
The prevalence of MINOCA (working diagnosis of MINOCA) in the lection. To diagnose coronary spasm is very important for improving
present study (10.2%) was almost identical to that reported in previous the prognosis of patients with MINOCA because calcium channel
studies [8,19]. The VIRGO (Variation in Recovery: Role of Gender on blocker is optimal medical therapy for suppression of coronary spasm
Outcomes of Young AMI Patients) study reported that 11.1% of 2690 in patients with vasospastic angina. Thus, appropriate treatment can
AMI patients aged 18–55 years who underwent coronary angiography be provided by accurately differential diagnosing the cause of MINOCA.
in 103 hospitals were diagnosed with MINOCA [8]. Furthermore, the The proposed standardized diagnostic algorithm for MINOCA [23] must

Table 3
Results of univariate and multivariate Cox frailty model for risk factors of in-hospital mortality.

Number of events/1000 person-days Univariate analysis Multivariate model 1 Multivariate model 2

HR (95%CI) HR (95%CI) HR (95%CI)

In-hospital mortality
MI-CAD 3.46 Ref Ref Ref
working diagnosis of MINOCA 4.48 1.23 (1.15, 1.32) 1.32 (1.23, 1.42) 1.34 (1.25, 1.44)
true MINOCA 4.57 1.26 (1.17, 1.35) 1.31 (1.19, 1.45) 1.34 (1.21, 1.48)
Deaths within 24 h
MI-CAD 0.96 Ref Ref Ref
working diagnosis of MINOCA 1.75 1.46 (1.30, 1.64) 1.70 (1.52, 1.91) 1.70 (1.51, 1.91)
true MINOCA 1.83 1.52 (1.36, 1.71) 1.54 (1.29, 1.86) 1.54 (1.29, 1.85)
Deaths within 30 days
MI-CAD 2.50 Ref Ref Ref
working diagnosis of MINOCA 2.74 1.12 (1.02, 1.22) 1.16 (1.06, 1.26) 1.19 (1.09, 1.30)
true MINOCA 2.74 1.12 (1.02, 1.23) 1.24 (1.10, 1.39) 1.27 (1.13, 1.43)

In model 1, the hazard ratios and 95% CIs were adjusted for age (b50, ≥50 to b64, ≥65 to b75, ≥75), sex, obesity, smoking, Killip ≥3, chronic pulmonary disease, peripheral vascular disease,
cerebrovascular disease, liver disease, diabetes, renal disease, malignancy, low ADL status on admission, hospital with number of hospital beds ≥500, hospital teaching status. In model 2,
the hazard ratios and 95% CIs were adjusted for age (b50, ≥50 to b65, ≥65 to b75, ≥75), sex, obesity, smoking, Killip ≥3, low ADL status on admission, hospital with the number of hospital
beds ≥500, hospital teaching status.
112 M. Ishii et al. / International Journal of Cardiology 301 (2020) 108–113
be fully understood by all cardiologists and the algorithm should be per-
formed in clinical practice at all institutions.
The multivariable mixed model identified old age, male sex, obesity,
smoking, and diabetes as significant negative predictors of MINOCA,
whereas peripheral vascular disease, cerebrovascular disease, liver dis-
ease, renal disease, and malignancy were significant positive predictors
of MINOCA, compared to MI-CAD. These findings indicate that patients
with MINOCA were more likely to have comorbidities although they
were less likely to have traditional coronary risk factors. The clinical
characteristics of MINOCA described in the present study are similar
to those reported in previous studies. One previous single-center retro-
spective study described MINOCA patients as non-smoker elderly fe-
males who tended to present with atrial fibrillation, no history of
previous MI, and showed non-ST segment elevation on the ECG [24].
The VIRGO study also found a negative association between traditional
coronary risk factors and MINOCA, whereas predisposition to
hypercoagulation correlated positively with MINOCA compared to MI-
CAD [8]. In the systematic review study [3], patients with MINOCA
were more likely to be a younger female who had no dyslipidemia, al-
though they exhibited other cardiovascular factors similar to those
with MI-CAD. Based on these findings, it seems MINOCA tends to affect
young females with fewer or similar cardiovascular risk factors com-
pared to MI-CAD.
There is little or no information on the short-term prognosis of
MINOCA patients. Surprisingly, our study showed worse prognosis
with both working diagnosis and true MINOCA (Table 3) compared
with MI-CAD. In this regard, the VIRGO study [8] showed similar, 1-
month mortality with MINOCA and MI-CAD (1.1%, 1.7%, respectively).
On the other hand, the systematic review study [3] showed better in-
hospital mortality with MINOCA, compared to MI-CAD (1.1%, 3.2%, re-
spectively). This difference in short-term prognosis of MINOCA between
the present study and the previous studies could be related to the het-
erogeneity of the patient characteristics and the inclusion criteria of
the studies. Otherwise, for the large sample size, the present study
showed a statistical significance in the mortality between MINOCA
and MI-CAD even though there was no clinical difference. First, the dis-
crepancy for the results of prognosis might be due to the difference in
age of study population. The median age of our study population was
higher than the previous studies. The median age of our study was 69
years old in MI-CAD and 71 in working diagnosis of MINOCA, whereas
those were 48 in MI-CAD and 46 in MINOCA in the VIRGO study [8],
and 61.3 in MI-CAD and 58.8 in MINOCA in the systematic review [3].
The results of this study suggested that the short-term prognosis of
MINOCA worsens in an aging society. Second, the diagnostic algorithm
for MINOCA might vary between the studies because of nature of retro-
spective analysis. The ESC working group stated that MINOCA should be
diagnosed using the AMI criteria, non-obstructive coronary artery on
angiography, and no clinical presentation of a specific cause at the
time of angiography (expressed as working diagnosis of MINOCA in
the present study) [5], whereas another paper suggested that non-
coronary cause- and myocardial cause-related disorders, such as myo-
carditis, Takotsubo cardiomyopathy, and other cardiomyopathies
should be excluded from the MINOCA (expressed as true MINOCA in
the present study) [6]. This double interpretation could cause heteroge-
neity and might explain the difference in the clinical outcome. However,
it is important to note that the timing of diagnosis is different between
those diagnostic criteria; the former is diagnosed at the time of angiog-
raphy with no available relevant cause for acute presentation, whereas
the latter is diagnosed when a clear cause for acute presentation is avail-
able. A standardized diagnostic approach needs to be established to
identify the cause and improve the prognosis since the present study
showed a good prognosis of coronary spasm, which can be treated
with calcium channel blockers. Recently, the American Heart Associa-
tion published a scientific statement regarding the diagnosis and man-
agement of MINOCA(23), so that the differential diagnosis of the cause
of MINOCA can be increased in accuracy, resulting that cardiologist
can provide an appropriate treatment according to the etiology and
the prognosis of MINOCA patients will be improved in the future.
Since this registry was conducted between 2012 and 2016, the stan-
dardized diagnosis and management of MINOCA were not fully per-
formed among the participating hospitals, resulting that some
MINOCA patients might not have received appropriate treatment and
might have a poor prognosis. A prospective study with the universal
clinical algorithm for the diagnosis and management of MINOCA is
needed to clarify the characteristics, etiology, and prognosis of MINOCA.
The present study has several limitations. First, several confounding
and potential prognostic factors were not measured in this study, such
as cardiac biomarkers, myocardial viability, and socio-economic status
because the observational study used only administrative database. Sec-
ond, we could not investigate cause-specific deaths, such as cardiovas-
cular deaths or non-cardiac deaths because no such data (i.e., cause of
death) were available. Also, we had no information for the major in-
hospital complications because this study collected administrative
data from the participating hospitals retrospectively. To collect data ret-
rospectively may miss data for in-hospital complications, and the study
using such data would underestimate the prevalence of the complica-
tions. A prospective study to investigate predefined cause-specific
death and in-hospital complications is needs to be conducted in the fu-
ture. Third, the long-term prognosis of MINOCA and MI-CAD could not
be assessed from the data available in the JROAD-DPC database because
the database did not include information on patients after discharge
from the hospital. Fourth, the cause of MINOCA could not be identified
in all patients. The validity of diagnosis did not provide a clue to the eti-
ology of MINOCA and no validation study for MINOCA had been con-
ducted. Therefore, we performed a validation study of the algorithm
for the diagnosis of MINOCA (Supplementary Materials). The validation
study showed that the algorithm for the diagnosis of working diagnosis
of MINOCA used in this study had a positive predictive value of 100%.
Fifth, because of the nature of the administrative database used in this
study, we could not obtain the data of the degree of coronary stenosis
evaluated by coronary angiography and the results of intravascular im-
aging such as IVUS or OCT. Also, the standardized diagnosis for MINOCA
might not have been performed among the participating hospitals.
These might cause misclassification of the MINOCA group, resulting
that the result of this study might be limited. Finally, AMI patients
who did not receive coronary angiography or revascularization were ex-
cluded from the study population, which may have caused a section
bias.
In conclusion, the present study compared the clinical characteris-
tics, predictive factors, and in-hospital mortality of patients with
MINOCA and MI-CAD. The results indicated that the short-term progno-
sis is worse in patients with MINOCA compared to those with MI-CAD in
the super-aging society. However, the prognosis of MINOCA was het-
erogeneous due to the diversity of the cause of MINOCA. The findings
emphasize the need for a standardized diagnostic approach in order to
help determine the etiology of MINOCA in individual cases so as to pro-
vide effective tailored treatment.
Sources of funding
This study was supported by Scholarship from the Minister of Educa-
tion, Science, Sports and Culture, who entrusts the head with the man-
agement of accounts, and JSPS KAKENHI Grant-in-Aid for Early-Career
Scientists Grand Number JP19K17531.
Declaration of competing interest
All authors declare no potential conflict of interest in connection
with this paper. For that not related to this paper, the potential conflict
of interest was denoted in supplementary material.
 M. Ishii et al. / International Journal of Cardiology 301 (2020) 108–113
Appendix A. Supplementary data
Supplementary data to this article can be found online at https://doi.
org/10.1016/j.ijcard.2019.09.037.
References
[1] K. Thygesen, J.S. Alpert, A.S. Jaffe, B.R. Chaitman, J.J. Bax, D.A. Morrow, et al., Fourth
universal definition of myocardial infarction, J. Am. Coll. Cardiol. 72 (18) (2018)
2231–22642018.
[2] J.F. Beltrame, Assessing patients with myocardial infarction and nonobstructed cor-
onary arteries (MINOCA), J. Intern. Med. 273 (2) (2013) 182–185.
[3] S. Pasupathy, T. Air, R.P. Dreyer, R. Tavella, J.F. Beltrame, Systematic review of pa-
tients presenting with suspected myocardial infarction and nonobstructive coro-
nary arteries, Circulation 131 (10) (2015) 861–870.
[4] S. Pasupathy, R. Tavella, J.F. Beltrame, The what, when, who, why, how and where of
myocardial infarction with non-obstructive coronary arteries (MINOCA), Circ. J. 80
(1) (2016) 11–16.
[5] S. Agewall, J.F. Beltrame, H.R. Reynolds, A. Niessner, G. Rosano, A.L. Caforio, et al., ESC
working group position paper on myocardial infarction with non-obstructive coro-
nary arteries, Eur. Heart J. 38 (3) (2017) 143–153.
[6] S. Pasupathy, R. Tavella, J.F. Beltrame, Myocardial infarction with nonobstructive
coronary arteries (MINOCA): the past, present, and future management, Circulation
135 (16) (2017) 1490–1493.
[7] J.E. Tamis-Holland, H. Jneid, Myocardial infarction with nonobstructive coronary ar-
teries (MINOCA): it's time to face reality! J. Am. Heart Assoc. 7 (13) (2018).
[8] B. Safdar, E.S. Spatz, R.P. Dreyer, J.F. Beltrame, J.H. Lichtman, J.A. Spertus, et al., Pre-
sentation, clinical profile, and prognosis of young patients with myocardial infarc-
tion with nonobstructive coronary arteries (MINOCA): results from the VIRGO
study, J. Am. Heart Assoc. 7 (13) (2018).
[9] S. Yasuda, K. Nakao, K. Nishimura, Y. Miyamoto, Y. Sumita, T. Shishido, et al., The cur-
rent status of cardiovascular medicine in Japan- analysis of a large number of health
records from a nationwide claim-based database, JROAD-DPC, Circ. J. 80 (11) (2016)
2327–2335.
[10] S. Yasuda, Y. Miyamoto, H. Ogawa, Current status of cardiovascular medicine in the
aging society of Japan, Circulation 138 (10) (2018) 965–967, https://doi.org/10.
1161/CIRCULATIONAHA.118.035858.
[11] K. Kanaoka, S. Okayama, K. Yoneyama, M. Nakai, K. Nishimura, H. Kawata, et al.,
Number of board-certified cardiologists and acute myocardial infarction-related
mortality in Japan- JROAD and JROAD-DPC registry analysis, Circ. J. (2018)
2845–2851.
113
[12] H. Yamana, M. Moriwaki, H. Horiguchi, M. Kodan, K. Fushimi, H. Yasunaga, Validity
of diagnoses, procedures, and laboratory data in Japanese administrative data, J.
Epidemiol. 27 (10) (2017) 476–482.
[13] C.J. Biesheuvel, Y. Vergouwe, R. Oudega, A.W. Hoes, D.E. Grobbee, K.G. Moons, Ad-
vantages of the nested case-control design in diagnostic research, BMC Med. Res.
Methodol. 8 (2008) 48.
[14] H. Quan, V. Sundararajan, P. Halfon, A. Fong, B. Burnand, J.C. Luthi, et al., Coding al-
gorithms for defining comorbidities in ICD-9-CM and ICD-10 administrative data,
Med. Care 43 (11) (2005) 1130–1139.
[15] H. Quan, B. Li, C.M. Couris, K. Fushimi, P. Graham, P. Hider, et al., Updating and val-
idating the Charlson comorbidity index and score for risk adjustment in hospital dis-
charge abstracts using data from 6 countries, Am. J. Epidemiol. 173 (6) (2011)
676–682.
[16] M.A. Jaffa, M. Gebregziabher, D.K. Luttrell, L.M. Luttrell, A.A. Jaffa, Multivariate gener-
alized linear mixed models with random intercepts to analyze cardiovascular risk
markers in type-1 diabetic patients, J. Appl. Stat. 43 (8) (2016) 1447–1464.
[17] P.C. Austin, A tutorial on multilevel survival analysis: methods, models and applica-
tions, Int. Stat. Rev. 85 (2) (2017) 185–203.
[18] A. Marshall, D.G. Altman, R.L. Holder, P. Royston, Combining estimates of interest in
prognostic modelling studies after multiple imputation: current practice and guide-
lines, BMC Med. Res. Methodol. 9 (2009) 57.
[19] B. Lindahl, T. Baron, D. Erlinge, N. Hadziosmanovic, A. Nordenskjold, A. Gard, et al.,
Medical therapy for secondary prevention and long-term outcome in patients
with myocardial infarction with nonobstructive coronary artery disease, Circulation
135 (16) (2017) 1481–1489.
[20] J. Deyama, T. Nakamura, Y. Saito, J.E. Obata, D. Fujioka, K. Nakamura, et al., Effect of
coronary artery spasm on long-term outcomes in survivors of acute myocardial in-
farction, Int. J. Cardiol. 257 (2018) 7–11.
[21] N. Nakayama, K. Kaikita, T. Fukunaga, Y. Matsuzawa, K. Sato, E. Horio, et al., Clinical
features and prognosis of patients with coronary spasm-induced non-ST-segment
elevation acute coronary syndrome, J. Am. Heart Assoc. 3 (3) (2014), e000795.
[22] J.F. Beltrame, F. Crea, J.C. Kaski, H. Ogawa, P. Ong, U. Sechtem, et al., International
standardization of diagnostic criteria for vasospastic angina, Eur. Heart J. 38 (33)
(2017) 2565–2568.
[23] J.E. Tamis-Holland, H. Jneid, H.R. Reynolds, S. Agewall, E.S. Brilakis, T.M. Brown, et al.,
Contemporary diagnosis and management of patients with myocardial infarction in
the absence of obstructive coronary artery disease: a scientific statement from the
American heart association, Circulation 139 (18) (2019) e891–e908.
[24] F. Montenegro Sa, C. Ruivo, L.G. Santos, A. Antunes, F. Saraiva, F. Soares, et al., Myo-
cardial infarction with nonobstructive coronary arteries: a single-center retrospec-
tive study, Coron. Artery Dis. 29 (6) (2018) 511–515.