Chest symptoms, such as chest pain and chest discomfort, are chief complaints leading to hospital visits, particularly visits to emergency departments.^1,2 Around 25% of individuals experiencing chest pain reportedly present to a hospital.³ Non-cardiac chest pain (NCCP) is defined as recurring angina-like retrosternal chest pain of non-cardiac origin.⁴ In England and Wales, 30-60% of patients admitted to emergency departments with chest pain receive non-cardiac diagnoses.¹ NCCP consists of musculoskeletal, pulmonary, infectious, drug-related, psychological, and gastrointestinal disorders, including functional chest pain as defined by the Rome IV criteria.^5,6

Elevated levels of anxiety, reduced quality of life (QOL), further episodes of chest pain, and frequent use of healthcare services have been reported in patients with NCCP.⁷ Patients with NCCP and patients with cardiac chest pain (CCP) use outpatient healthcare services to similar extents.⁸ Furthermore, the histories and characteristics of patients do not reliably distinguish between cardiac and esophageal causes of chest pain.⁹ Patients with NCCP experience similar or higher levels of psychological distress compared to patients with CCP.¹⁰ Previous studies have confirmed that psychological distress has negative impacts on health-related QOL in patients with NCCP.^11,12 NCCP therefore places a substantial burden on healthcare costs and resources.¹³

Only a few studies have evaluated the consultation behavior of individuals with NCCP in the general population.^14,15 Furthermore, due to cultural, ethnic and socio-economic differences, the consultation behavior and impact on QOL in Asian NCCP patients may differ from those observed in western countries.

NCCP has been reported to affect 19-33% of the population in the United States, Australia, Argentina, and China,^14,16-18 while the epidemiology and characteristics of NCCP in other countries, including Japan, remain unclear. We therefore aim to clarify the prevalence of NCCP in a sample of the Japanese general population, to identify factors associated with consultation behavior and QOL.

Internet-based Questionnaire Survey

Two internet surveys were conducted in 2017, with the cooperation of an internet research company (Macromill Inc, Tokyo, Japan). This study was approved by the Ethics Committee of Hyogo College of Medicine (No. 2401). The first survey was a screening survey to determine the prevalence of NCCP in the Japanese population. The second survey was a more detailed survey to determine the characteristics of NCCP subjects, including severity, frequency, duration and location of chest pain, along with specific scales to evaluate QOL, anxiety, and depression.

The first survey included 10 000 subjects from the Japanese general population between 20 and 69 years old. People aged 70 years or older were not included because of the potential difficulties in responding the online survey.¹⁹ The survey was conducted between March 9 and 10, 2017. Male and female subjects in age strata of 20 to < 30 years old (20s), 30 to < 40 years old (30s), 40 to < 50 years old (40s), 50 to < 60 years old (50s), and 60 to < 70 years old (60s) were randomly selected from the registered panelists based on Japanese demographics (Table 1). Subjects were recruited until the targeted number was reached.

Table 1 . Demographic Characteristics of Subjects in the First Survey of 10 000 Members of the General Population in Japan

Demographic characteristics	n = 10 000
Age (yr)	46.0 ± 13.7
20-29 yr (M/F)	789/760
30-39 yr (M/F)	987/966
40-49 yr (M/F)	1159/1142
50-59 yr (M/F)	963/969
60-69 yr (M/F)	1103/1162
Sex
Male	5001 (50)
Female	4999 (50)
Marital status
Singlea	3735 (37)
Married	6265 (63)
Presence of children
No	4403 (44)
Yes	5597 (56)
Occupation
Civil servant	321 (3)
Manager/Executive	160 (2)
Company employee	3774 (37)
Self-employed	572 (6)
Freelance	180 (2)
Housewife/househusband	2031 (20)
Part-time job	1356 (14)
Students	311 (3)
Others	313 (3)
Without work	982 (10)

^aIncluding divorced/separated/widowed.

M, male; F, female.

Data are presented as mean ± SD, n, or n (%).

The second internet survey was conducted from March 17 to 21, 2017, again targeting the Japanese general population between 20 and 69 years old. Respondents in the survey were evenly distributed in terms of both age and sex. The median age of subjects was 45 years old. To include a sufficient number of subjects with chest pain, equal numbers of subjects with chest symptoms (chest pain or chest discomfort) and without chest symptoms were recruited. A total of 2480 subjects were thus included in the second survey.

Quality of Life Assessment

QOL was measured using the validated Short Form-8.²⁰ Short Form-8 comprises 8 subscales (general health, physical functioning, role physical, body pain, vitality, social functioning, mental health, and role emotional). In the present study, scores for each of the 8 subscales, the physical component summary score (PCS), and the mental component summary score (MCS) were determined using a norm-based scoring method, based on a large-scale population study conducted in Japan and validated by previous studies.^21,22

Hospital Anxiety and Depression Scale

Hospital anxiety and depression scale (HADS) was employed to measure anxiety and depression.²³ This scale includes 14 items, with 7 items related to the anxiety dimension (HADS-A) and the remaining 7 items related to the depression dimension (HADS-D). The total score of these anxiety and depression subscales can range from 0 to 21. The validated Japanese version of HADS was used.²⁴ Anxiety or depression was defined as a score ≥ 8 for HADS-A or HADS-D, respectively.²⁵

Gastroesophageal Reflux Disease Questionnaire

Gastroesophageal reflux disease questionnaire (GerdQ) is a self-administered 6-item questionnaire for diagnosis and the evaluation of treatment response in patients with gastroesophageal reflux disease (GERD).²⁶ The validated Japanese version of GerdQ was used, and a score ≥ 8 is considered predictive of GERD as stated before.²⁷

Non-cardiac Chest Pain and Cardiac Chest Pain

CCP was defined as present in subjects who consulted physicians for chest pain and were diagnosed with angina pectoris, myocardial infarction, arrhythmia, or heart failure. NCCP was defined as present in subjects who consulted physicians for chest pain and received diagnoses other than angina pectoris, myocardial infarction, arrhythmia, or heart failure.

Statistical Methods

Continuous variables are presented as mean ± standard deviation and categorical variables are presented as the number and percentage. Data were analyzed using the χ² test and Student’s t test, as appropriate. To determine factors associated with consultation behavior among subjects with chest pain, we estimated odds ratios (ORs) and 95% confidence intervals (95% CIs). Univariate analyses and unrestricted multivariate logistic regression analysis were performed to test the influence of several factors in association with consultation behavior. All tests were two-sided with a significance level of P < 0.05. Data were analyzed using SPSS version 23.0 (IBM Corp, Armonk, NY, USA).

Prevalence of Chest Pain in the Japanese General Population

Among the 10 000 subjects, 63% were married and 56% had children. Subjects were predominantly company employees (37%) and housewives/househusbands (20%) (Table 1). Chest pain was reported by 524 (5%) respondents, who were more likely to be female and younger than those did not report chest pain (Table 2). Among subjects with chest pain, 155 (29%) consulted a physician. Females were less likely to seek consultations for chest pain than males (Table 3). Of the 155 consulters for chest pain, 108 (70%) consulters were diagnosed with NCCP.

Table 2 . Characteristics of Subjects With Chest Pain in the First Survey

Characteristics	Total		Subjects without chest pain		Subjects with chest pain	P-valuea
	N = 10 000		n = 9476		n = 524
Age (yr)	46.0 ± 13.7		46.1 ± 13.7		44.2 ± 14.1	0.003
Female	4999 (50)		4688 (50)		311 (60)	< 0.001
Married	6265 (63)		5955 (63)		310 (59)	0.090
Presence of children	5597 (56)		5298 (56)		299 (57)	0.605

^aSubjects with chest pain compared with subjects without chest pain.

Data are presented as mean ± SD or n (%).

Table 3 . Consultation Behavior of Subjects With Chest Pain in the First Survey

Characteristics	Non-consultors		Consultors	P-value
	n = 369		n = 155
Age (yr)	43.6 ± 14.0		45.7 ± 14.2	0.115
Female	232 (63)		79 (51)	0.015
Married	210 (57)		100 (65)	0.129
Presence of children	202 (55)		97 (63)	0.119

Data are presented as mean ± SD or n (%).

Demographic Characteristics

To better determine the characteristics of NCCP, the second survey was then performed. A total of 2480 subjects (1240 without chest symptoms, 780 with chest pain and 460 with chest discomfort) were included in this survey, and 632 subjects (25%) were overlapped with the first survey. When subjects with chest pain and without chest symptoms were compared, no differences in age, sex, marital status, or presence of children were identified. PCS and MCS were lower in subjects with chest pain than in those without chest symptoms. Frequencies of anxiety, depression, and GERD were higher in subjects with chest pain than in subjects without chest symptoms (Table 4).

Table 4 . Characteristics of Subjects in the Second Survey

Characteristics	Subjects without chest symptoms		Subjects with chest pain	P-value
	(n = 1240)		(n = 780)
Age (yr)	44.8 ± 13.9		43.7 ± 14.0	0.084
Female	620 (50)		417 (53)	0.142
Married	683 (55)		453 (58)	0.202
Presence of children	634 (51)		407 (52)	0.679
SF-8
PCS	49.3 ± 6.9		47.0 ± 6.7	< 0.001
MCS	48.1 ± 8.0		44.7 ± 8.7	< 0.001
GerdQ (–/+)	1110/130		601/179	< 0.001
HADS-A (–/+)	899/341		386/394	< 0.001
HADS-D (–/+)	576/664		279/501	< 0.001

SF-8, Short Form-8; PCS, physical component summary score; MCS, mental component summary score; GerdQ, gastroesophageal reflux disease questionnaire; HADS-A, hospital anxiety and depression scale related to anxiety dimension; HADS-D, Hospital anxiety and depression scale related to depression dimension.

For GerdQ, HADS-A, and HADS-D, sores ≥ 8 were defined as positive.

Data are presented as mean ± SD, n (%), or n.

Consultation Behavior in Subjects With Chest Pain

To better define NCCP, subjects with chest pain were extracted for further analysis. Among all 780 subjects with chest pain, 233 (30%) consulted physicians. Females were again less likely to seek consultations for chest pain. Consulters showed a higher severity, higher frequency, and longer duration of chest pain. Further, the frequencies of anxiety, depression, and GERD were higher among consulters for chest pain than among non-consulters. Multivariate logistic regression analysis was performed to reveal factors independently associated with consultation behavior. Male sex, greater severity, or greater frequency of chest pain, lower PCS, and greater frequency of GERD were factors independently associated with consultation behavior among subjects with chest pain (Table 5).

Table 5 . Factors Associated With Consultation Behavior in Subject With Chest Pain

Factors	Non-consulters		Consulters		Univariate			Multivariate
	(n = 547)		(n = 233)		OR (95% CI)	P-value		OR (95% CI)	P-value
Age (yr)	43.2 ± 14.0		45.0 ± 13.8		1.0 (1.0, 1.0)	0.101
Female	319 (58)		98 (42)		0.5 (0.4, 0.7)	< 0.001		0.6 (0.4, 0.9)	0.006
Married	315 (58)		138 (59)		1.1 (0.8, 1.5)	0.671
Presence of children	285 (52)		122 (52)		1.0 (0.7, 1.4)	0.947
Severity
Very mild	155 (28)		46 (20)		1 (reference)	-
Mild	246 (45)		89 (38)		1.2 (0.8, 1.8)	0.342		0.9 (0.5, 1.3)	0.482
Moderate	128 (23)		75 (32)		2.0 (1.3, 3.1)	0.002		1.3 (0.8, 2.1)	0.327
Severe	17 (3)		16 (7)		3.2 (1.5, 6.8)	0.003		1.6 (0.7, 3.6)	0.312
Very severe	1 (0)		7 (3)		23.6 (2.8, 196.7)	0.003		12.1 (1.3, 110.4)	0.027
Frequency
1-2/wk	435 (80)		145 (62)		1 (reference)	-
3-4/wk	82 (15)		51 (22)		1.9 (1.3, 2.8)	0.002		1.5 (1.0, 2.3)	0.062
5-6/wk	13 (2)		14 (6)		3.2 (1.5, 7.0)	0.003		2.3 (1.0, 5.3)	0.055
7/wk	17 (3)		23 (10)		4.1 (2.1, 7.8)	< 0.001		2.9 (1.4, 5.9)	0.004
Duration
≤ 5 sec	110 (20)		30 (13)		1 (reference)	-
≤ 30 sec	118 (22)		49 (21)		1.5 (0.9, 2.6)	0.115		1.3 (0.8, 2.3)	0.319
≤ 1 min	83 (15)		35 (15)		1.5 (0.9, 2.7)	0.130		1.2 (0.7, 2.2)	0.557
≤ 5 min	130 (24)		54 (23)		1.5 (0.9, 2.5)	0.108		1.1 (0.6, 1.9)	0.833
≤ 30 min	51 (9)		23 (10)		1.7 (0.9, 3.1)	0.122		1.0 (0.5, 2.1)	0.936
> 30 min	55 (10)		42 (18)		2.8 (1.6, 4.9)	< 0.001		1.4 (0.7, 2.6)	0.370
SF-8
PCS	47.9 ± 6.3		44.7 ± 7.2		0.9 (0.9, 1.0)	< 0.001		1.0 (0.9, 1.0)	< 0.001
MCS	45.1 ± 8.6		43.7 ± 8.7		1.0 (1.0, 1.0)	0.050
GerdQ (–/+)	444/103		157/76		2.1 (1.5, 3)	< 0.001		1.6 (1.1, 2.3)	0.021
HADS-A (–/+)	288/259		98/135		1.5 (1.1, 2.1)	0.007		1.2 (0.8, 1.7)	0.435
HADS-D (–/+)	210/337		69/164		1.5 (1.1, 2.1)	0.020		1.0 (0.7, 1.5)	0.879

SF-8, Short Form-8; PCS, physical component summary score; MCS, mental component summary score; GerdQ, gastroesophageal reflux disease questionnaire; HADS-A, hospital anxiety and depression scale related to anxiety; HADS-D, hospital anxiety and depression scale related to depression.

For GerdQ, HADS-A and HADS-D, sores ≥ 8 were defined as positive.

Data are presented as mean ± SD, n (%), or n.

Among female subjects with chest pain, older age (≥ 45 years old) was associated with a greater likelihood of consulting a physician than younger age (< 45 years old; OR, 1.9, 95% CI, 1.1-3.1, P = 0.020). However, among males, age was not independently associated with consultation behavior for chest pain (OR, 1.1; 95% CI, 0.6-1.8; P = 0.818).

Impacts of Non-cardiac Chest Pain and Cardiac Chest Pain on Quality of Life

Among all 233 consulters, 174 (75%) were diagnosed with NCCP and 59 (25%) were diagnosed with CCP. Among consulters, 138 subjects with NCCP (79%) and 51 subjects with CCP (86%) similarly located chest pain in the left chest or sternum (P = 0.227, Figure). No age or sex differences were identified between subjects with NCCP and those without chest symptoms (Table 6). Subjects with NCCP showed lower PCS and MCS than those without chest symptoms, whereas subjects with NCCP more frequently experienced GERD, anxiety, and depression than those without chest symptoms (Table 6). Subjects with NCCP were younger than those with CCP. No difference in PCS, MCS, or frequencies of anxiety, depression, or GERD were identified between CCP and NCCP subjects (Table 6).

Figure 1. Location and prevalence of chest pain in consultors. Non-cardiac chest pain (top); cardiac chest pain (bottom).

Table 6 . Characteristics in the Control, Cardiac Chest Pain, and Non-cardiac Chest Pain Groups

Characteristics	Control n = 1240	CCP n = 59	P-valuea	NCCP n = 174	P-valuea	P-valueb
Age (yr)	44.8 ± 13.9	49.0 ± 13.4	0.025	43.6 ± 13.7	0.286	0.009
Female	620 (50)	36 (61)	0.098	75 (43)	0.088	0.580
SF-8
PCS	49.3 ± 6.9	44.3 ± 8.7	< 0.001	44.9 ± 6.6	< 0.001	0.570
MCS	48.1 ± 8.0	44.7 ± 9.1	0.002	43.4 ± 8.6	< 0.001	0.345
GerdQ (–/+)	1110/130	40/19	< 0.001	117/57	< 0.001	0.937
HADS-A (–/+)	899/341	28/31	< 0.001	70/104	< 0.001	0.331
HADS-D (–/+)	576/664	22/37	0.213	47/127	< 0.001	0.135

^aCompared with control subjects.

^bCompared with subjects with cardiac chest pain (CCP).

NCCP, non-cardiac chest pain; SF-8, Short Form-8; PCS, physical component summary score; MCS, mental component summary score; GerdQ, gastroesophageal reflux disease questionnaire; HADS-A, hospital anxiety and depression scale related to anxiety; HADS-D, hospital anxiety and depression scale related to depression.

For GerdQ, HADS-A and HADS-D, sores ≥ 8 were defined as positive.

Data are presented as mean ± SD, n (%), or n.

Examination and Diagnosis of Subjects With Non-cardiac Chest Pain

Among subjects with NCCP (subjects with chest pain not diagnosed with CCP) (n = 174), 82% visited a primary-care physician for consultation. The proportions of those who visited an emergency physician, cardiologist, and gastroenterologist were similar (8%, 11%, and 8%, respectively). The most frequently performed examinations were chest radiography, electrocardiography, and blood tests (51%, 47%, and 36%, respectively) (Table 7). Among all subjects visited emergency physicians (n = 30), 14 subjects (47%) were diagnosed with NCCP. Among all subjects visited cardiologists (n = 41), 20 subjects (49%) were diagnosed with NCCP.

Table 7 . Examinations and Diagnosis of Subjects With Non-cardiac Chest Pain

Variables	n = 174
Physician visits
Primary care physician	143 (82)
Emergency physician	14 (8)
Cardiologist	20 (11)
Gastroenterologist	14 (8)
Others	10 (6)
Examination
Blood test	63 (36)
Chest radiography	88 (51)
Electrocardiography	82 (47)
Cardiovascular angiography	11 (6)
CT scanning	27 (16)
MRI scanning	15 (9)
Esophagogastroduodenoscopy	20 (11)
Colonoscopy	5 (3)
Echocardiography	31 (18)
Abdominal ultrasound	14 (8)
Other	10 (6)
No test	26 (15)
Diagnosis
NCCP (diagnosed by a physician)	5 (3)
RE	26 (15)
Functional chest pain	4 (2)
Abnormal esophageal movement	3 (2)
Achalasia	4 (2)
Gastritis or peptic ulcer	6 (3)
FD/IBS	2 (1)
Anxiety	19 (11)
Unknown	96 (55)
Others	29 (17)

NCCP, non-cardiac chest pain; RE, reflux esophagitis; FD, functional dyspepsia; IBS, irritable bowel syndrome.

Data are presented as n (%).

Only 3% were diagnosed with NCCP by a physician. In comparison, 15% were diagnosed with reflux esophagitis, 11% were diagnosed with anxiety, and the diagnosis was unknown in 55% (Table 7).

The prevalence of chest pain was 5% in the Japanese general population. Approximately 30% of participants with chest pain consulted a physician. Among subjects who sought consultations for chest pain, 70% were diagnosed with NCCP. Furthermore, females were less likely to seek consultation for chest pain than males. To the best of our knowledge, the present study is the first study to describe the epidemiology of NCCP in Japan. Sex and both the severity and frequency of chest pain were associated with consultation behavior.

Previous studies revealed the prevalence of NCCP as approximately 25%.^14,16-18 However, those studies differ in many respects, such as in the definition of NCCP, sample size, and ethnic disparities. The prevalence of chest pain in the present study was 5% in the Japanese general population, lower than in previous studies of other populations. One possible reason might be that different people suffering the same symptom of chest pain might describe the experience differently due to cultural differences.²⁸

Less than one-third of subjects with chest pain consulted a physician regarding their symptoms in the present study. This is comparable to a previous study in Australia, in which approximately 25% of individuals who suffered chest pain consulted physicians.¹⁴ Approximately 80% of the subjects with NCCP visited a primary-care physician in the present study, consistent with previous studies from the United States and Australia.^29,30 Unlike a previous study conducted in Australia, in which respondents with NCCP were more likely to consult a cardiologist, but less likely to consult a gastroenterologist,³⁰ subjects with NCCP in the present study consulted cardiologists, gastroenterologists, and emergency physicians at similar levels. This difference between studies might be due to differences in the medical service systems among countries.

Previous studies showed that the severity and frequency of NCCP were factors independently associated with consultation behavior.^14,31 Another study showed that patients with GERD consulted physicians due to frequency and severity of symptoms, impaired health-related QOL, and concern of symptoms.³² In the present study, subjects with severe or frequent chest pain or lower PCS were more likely to consult a physician.

Females were less likely to seek consultations for chest pain than males in the present study. Similarly, a survey from Australia found that male patients were more likely to seek consultations for NCCP than female patients, perhaps due to an awareness that male sex represents a higher risk factor for possible acute coronary syndrome.¹⁴ In addition, females were less sensitized to the association between chest pain and heart attack and were less likely to seek consultations or more likely to wait longer to consult a physician, as they considered the symptoms attributable to anxiety or stress.^33,34 Besides, even females felt chest pain, they might try to avoid the recognition of herself as a sick person, which might lead to less seeking medical care.³⁵

Among females with chest pain, older subjects were more likely to consult a physician than younger subjects, while age was not a factor associated with consultation behavior in males. Older females might be more aware of being at higher risk of acute coronary syndrome than younger females. A previous report demonstrated that younger females with acute myocardial infarction had not recognized their personal risk of heart disease, which led to longer delays in seeking prompt care than males.³⁶

GERD-related symptoms represented a factor independently associated with consultation behavior among subjects with chest pain in the present study. A previous study similarly showed that GERD was predictive of an increased number of patients with coronary artery disease experiencing chest pain, as well as emergency department visits. At the same time, patients with coronary artery disease utilizing proton pump inhibitors were less likely to report chest pain or emergency visits than those not using proton pump inhibitors.³⁷ GERD-related symptoms might thus be associated with consultation behavior among subjects with chest pain.

Whether anxiety or depression could affect the consultation behavior of subjects with chest pain remains controversial. Anxiety and depression have been reported to be associated with increased healthcare use among patients with NCCP.^8,31,38 However, in the present study, neither anxiety nor depression was associated with consultation behavior in subjects with chest pain, consistent with several previous studies.^14,17

Subjects with NCCP were younger than those with CCP in the present study, consistent with previous findings.³⁹ However, subjects with NCCP displayed similar QOL (PCS and MCS scores), as well as similar frequencies of anxiety and depression to subjects with CCP, but lower QOL (PCS and MCS scores) and more frequent anxiety and depression than subjects without chest symptoms. The location of chest pain in subjects with NCCP or CCP were mainly in left chest or sternum in the present study. These findings were consistent with previous studies,^11,14,40 and NCCP is not easily distinguished from CCP by symptoms alone.

Functional chest pain, as a subset of functional gastrointestinal disorders, has been estimated to account for 32-35% among NCCP.^6,41 However, functional chest pain was diagnosed in 2% of NCCP in the present study. Since the diagnosis of functional chest pain requires exclusion of CCP, GERD, and other esophageal motor disorders, the diagnosis of functional chest pain is usually made by gastroenterologists.⁶ In the present study, however, a majority of subjects with NCCP consulted primary-care physicians, who might not be able to diagnose functional chest pain. In addition, the knowledge of functional chest pain may be scarce among subjects and physicians in Japan.

The present study has several limitations that need to be kept in mind when interpreting the results. First, response rate could not be calculated as the recruitment was stopped when the target number was reached, which might have caused a selection bias.⁴² Subjects with chest pain might be more interested in the survey than those without chest pain in the first survey. Since the prevalence of chest pain in present study was lower than previous studies,^14,16-18 this bias might be less important. Second, the present study including subjects aged 20 to 69 years might lead to low prevalence of CCP, since most patients with CCP are elderly. Third, subjects with NCCP were not proven as the format was an online survey and no validated self-assessment questionnaire exist. Although a large proportion of participants had an unknown diagnosis, CCP was at least excluded from the present survey. Forth, the frequency of depression was higher compared to previous studies, albeit similar to that from other internet-based surveys of general populations.^43-46 Fifth, the study did not ask subsequent methods and post-treatment course of chest pain in NCCP subjects.

In summary, the prevalence of chest pain in a Japanese general population was 5%. Among all subjects with chest pain, approximately 30% sought consultations for their chest pain, 70% of whom were diagnosed with NCCP. Sex and both the severity and frequency of chest pain were associated with consultation behavior.

This study was supported by a Grant-in Aid for Scientific Research from the Japan Society for the Promotion of Science (JSPS) Grant Number JP16K09298 to Hiroto Miwa and by a research fund from Ono pharmaceutical Co, Ltd.

This study was supported by a Grant-in Aid for Scientific Research from the JSPS (Grant No. JP16K09298) and by a research fund from Ono pharmaceutical Co, Ltd.

None.

Junji Chen and Tadayuki Oshima analyzed and interpreted the data and wrote the manuscript; Toshihiko Tomita and Hirokazu Fukui contributed to interpretation of data; Shinichiro Shinzaki was the supervisor in this study and revised the paper critically for important intellectual content; and Takashi Kondo and Hiroto Miwa designed the research study and contributed to interpretation of data. All authors approved the final version of the manuscript.