Association between Metabolic Syndrome and Cardiovascular Events in a Japanese Population with and without Obesity: The Shizuoka Kokuho Database Study

Abstract

Aim: The accumulation of metabolic risk factors, namely high blood pressure, hyperlipidemia, and hyperglycemia, has been associated with cardiovascular diseases. However, little evidence is available on the prognostic significance of metabolic risk factor accumulation in nonobese individuals. This study investigated this issue by analyzing prefecture-wide health checkup and health insurance data in Japan.

Methods: We analyzed data from 366,881 adults aged 40–74 years who were enrolled in the National Health Insurance, excluding those who experienced a stroke or coronary artery diseases or required long-term care. Baseline clinical information was obtained from annual health checkup data. Incidences of stroke and coronary artery diseases were obtained from insurance data.

Results: In the nonobese population, the hazard ratio for stroke increased linearly with the number of accumulated metabolic risk factors, particularly among those aged ＜65 years men (one factor: 2.21, two factors: 2.60; three factors: 3.93) and women (one factor: 1.49, two factors: 1.57; three factors: 2.27). Similar results were observed in the analysis for coronary artery diseases. After excluding participants receiving medications, the association of metabolic risk factor with stroke remained significant, although its association with coronary artery disease became less significant. In the analysis for each metabolic risk factors, high blood pressure (men: hazard ratio = 2.85; women: hazard ratio = 2.17; P＜0.001), but not hyperlipidemia and hyperglycemia, was associated with stroke in the nonobese population.

Conclusion: The accumulation of metabolic risk factors needs to be considered a risk factor for cardiovascular diseases even in individuals without obesity.

Introduction

Evidence has shown that the accumulation of metabolic risk factors, otherwise referred to as metabolic syndrome, is a strong risk factor for cardiovascular diseases (CVDs) among not only Europeans¹⁾ but also the Japanese^2-8). Indeed, several previous studies on community residents throughout Japan have reported that metabolic syndrome was associated with the incidence^2-4) and mortality^5-8) of CVDs. Although several definitions of metabolic syndrome exist, they commonly characterize it as the accumulation of visceral obesity, high blood pressure (BP), dyslipidemia, and hyperglycemia⁶⁾. Both the International Diabetes Federation⁹⁾ and Japanese Society of Internal Medicine¹⁰⁾ consider visceral obesity an essential criterion. When visceral obesity is mandatory in the definition of metabolic syndrome, non-obese individuals are not considered to have metabolic syndrome even if they have accumulated a number of metabolic risk factors. However, several studies have reported that the CVD risks for individuals with and without obesity are equivalent when the number of metabolic risk factors is the same^5-7).

To reduce the incidence of CVD, a nationwide health screening and intervention program specifically targeting metabolic syndrome had been commenced across Japan in April 2008 ¹¹⁾. This program aims to (1) identify individuals who are likely to develop noncommunicable diseases via the Specific Health Checkup (Tokutei Kensin) and (2) provide early intervention (Specific Health Guidance) (Tokutei Hokenshidou) to prevent CVD development. In this program, based on the number of metabolic risk factors assessed during the health checkups, individuals with obesity are classified into three intervention groups, namely intensive support intervention (ISI), motivation-support intervention (MSI), and information supply only (ISO) groups, and individuals in the ISI or MSI groups who are not taking any medication receive lifestyle modification interventions. Because visceral obesity or body mass index (BMI) ≥ 25 kg/m² is also an essential criterion in the risk assessment in the Specific Health Checkup¹¹⁾, individuals without obesity are excluded from the Specific Health Guidance despite having metabolic risk factors. The higher CVD risk in nonobese individuals were observed when the Specific Health Checkup criteria was employed for risk stratification¹²⁾, although the number of participants in that study was limited because the study was a meta-analysis of existing cohort studies.

Given these backgrounds, we aimed to clarify the CVD risk of metabolic syndrome among individuals with and without obesity classified according to the risk stratification method used in the Specific Health Checkup by analyzing Shizuoka Prefecture-wide health insurance data and clinical information obtained from the Specific Health Checkup to identify target populations for CVD prevention interventions.

Methods

Data Source

This study analyzed the Shizuoka Prefecture-wide Kokuho Database (SKDB), which contained data on health insurance claims by individuals enrolled in the National Health Insurance and the Latter-Stage Elderly Medical Care System. The National Health Insurance is designed to partially cover the medical expenditure of individuals aged ＜75 years who are not eligible to be members of any employee-based health insurance. Older individuals aged ≥ 75 years, except those who have an occupation, are required to enroll in the Latter-Stage Elderly Medical Care System. Individuals enrolled in the National Health Insurance have the opportunity to receive the Specific Health Checkup (Tokuei Kenshin, Kouki Koureisha Kenshin) held annually, with the clinical information obtained at these health checkups being included in the SKDB.

Moreover, the SKDB includes claims data on the Long-Term Care Insurance System, which covers the daily care expenses for older people. The long-term care approval board in each municipality determines the certified care level by assigning a support level (levels 1–2) or care level (levels 1–5) based on the applicant’s mental and physical condition as well as their primary physician’s opinion. Insured individuals who require long-term care are eligible for in-home or facility-based services according to their certified care level. The care requirement certification is designed to be applied uniformly on a nationwide basis.

The current version of the SKDB covers the period from April 2012 to September 2022 (SKDB version 2024.1 with the analysis data generation system version 4.0)^13-15).

Study Setting

This longitudinal observational study included Shizuoka Prefecture residents aged ＜75 years who were enrolled in the National Health Insurance. During our analysis, the earliest day of participation in the Specific Health Checkup after 2013 was set as the index day, with the baseline period spanning 12 months prior to the index day (Supplementary Fig.1). The end of the follow-up period was defined as the day on which any of the following occurred: 1) incidence of a stroke or coronary artery diseases (CAD), 2) all-cause death, 3) insurance withdrawal for another reason, or 4) end of the current follow-up period (i.e., September 2022). The follow-up duration was calculated as the number of days from the index day until the end of follow-up. Death cases and participants who withdrew their insurance were treated as a censored case.

Supplementary Fig.1. Illustration of the study design

Index day was defined as day of earliest participation to the Specific Health Checkup where one year baseline period was available (after April 2013). The end of the follow-up was defined as the day of any of the following; 1) incidence of a stroke or MI, 2) all-cause death, 3) insurance withdrawal for another reason, or 4) end of the current follow-up period (i.e., September 2022).

In the analysis of the earlier version of the SKDB, readmission cases with the same insurance ID were considered continuous enrollees by disregarding the intermediate period during which they were unsubscribed. However, the method of calculating the follow-up period was changed to exclude the intermediate period by setting the baseline to fall within the consecutive insurance enrollment period. Those who died due to any reason were also censored. All-cause mortality was identified by referring to the withdrawal reason described in the health insurance data.

Study Population

Among the 2,654,305 residents in the Shizuoka Prefecture who were included in the current version of the SKDB, those who participated in the health checkup after April 2013 (n = 696,064) were extracted. A total of 366,881 individuals were ultimately included in the analysis after excluding those who met the following criteria: ≥ 75 years of age (n = 197,906); had a history of stroke or CAD (n = 92,149) or had been hospitalized during the 12 months prior to the index day (baseline period) (n = 21,299); certified to require long-term care (n = 1,324); or had clinical measurements that were missing or deviated substantially from their distribution (i.e., out of the reference range; Supplementary Table 1) or were ≥ 99.99% or ≤ 0.01% of their distribution (n = 16,505).

Supplementary Table 1.Reference range used in the Specific Health Checkup

	Minimum	Maximum
Height, cm	100	250
Weight, kg	20	250
Body mass index, kg/m2	10	99.9
Systolic blood pressure, mmHg	60	300
Diastolic blood pressure, mmHg	30	150
High-density lipoprotein cholesterol, mg/dL	10	500
Low-density lipoprotein cholesterol, mg/dL	20	1000
Hemoglobin A1c, %	3.0	20.0
Creatinine, mg/dL	0.10	20.00
Estimated glomerular filtration rate, ml/min/1.73m2	1.0	500.0

The reference range is defined by the Ministry of Health, Labour and Welfare, Japan and used nationwide. This information is available on the website (Standardized protocol for Specific Health Checkup and Specific Health Guidance 2023, PP 189. https://www.mhlw.go.jp/ content/10900000/001093926.pdf)

Ethical Considerations

Our study protocol involving SKDB analysis was approved by the ethics committee of the Shizuoka Graduate University of Public Health (SGUPH_2021_001). Approval for using their insurance data in medical studies was also obtained from the review board of each municipality prior to receiving the data. Before the receipt of the SKDB data, all personal details were anonymized by the Shizuoka Federation of National Health Insurance Organizations. To ensure that the participants can refuse the use of their data, information related to this study was disclosed on the websites of the Shizuoka Prefectural Government Office and Shizuoka Graduate University of Public Health.

Outcome Definition

Stroke cases were identified using the International Statistical Classification of Diseases and Related Health Problems, 10th edition. Stroke cases were defined as individuals who were hospitalized for over 3 days under specific codes I60 (subarachnoid hemorrhage), I61 (hemorrhage stroke), or I63 (ischemic stroke) as either the major disease or the disease that requiring the most hospital resources, as well as those who underwent computed tomography (medical procedure codes: 170011710, 170011810, 170028610,170033410, or 170015410) (Supplementary Table 2)¹⁶⁾. Information on the medical procedure codes were available for hospitals where Diagnosis Procedure Combination, a method for inpatient classification in the acute phase illness, were not introduced¹⁷⁾. CAD was defined as being hospitalized for 3 days or more with the I21 (acute myocardial infarction) code and having records indicating that they received percutaneous coronary intervention (medical procedure codes: 150260350, 150284310, 150318310, 150359310, 150374910, 150375010, 150375110, 150375210, 150375310, and 150375410) (Supplementary Table 2). When identifying the incidences of stroke or CAD cases who died within 7 days of diagnosis were considered incident cases regardless of the duration of hospitalization.

Supplementary Table 2.Procedure codes used for identification of incidence of stroke or myocardial infarction

Disease	Procedure code	Procedure
Stroke	170011710	Multi-slice computed tomography (another device)
	170011810	Multi-slice computed tomography (≥ 16-slice, ＜64-slice)
	170028610	Multi-slice computed tomography (≥ 4-slice, ＜16-slice)
	170033410	Multi-slice computed tomography (≥ 64-slice)
	170015410	Image diagnosis (computed tomography)
Myocardial infarction	150260350	Percutaneous transluminal coronary athelectomy
	150284310	Percutaneous transluminal coronary angioplasty (atherectomy catheter)
	150318310	Percutaneous transluminal coronary thrombectomy
	150359310	Percutaneous transluminal coronary angioplasty (excimer laser coronary angioplasty)
	150374910	Percutaneous transluminal coronary angioplasty (acute myocardial infarction)
	150375010	Percutaneous transluminal coronary angioplasty (unstable angina pectoris)
	150375110	Percutaneous transluminal coronary angioplasty (other diseases)
	150375210	Percutaneous stenting (acute myocardial infarction)
	150375310	Percutaneous stenting (unstable angina pectoris)
	150375410	Percutaneous stenting (other diseases)

Metabolic Risk Factors

Metabolic risk factors (i.e., visceral obesity, high BP, hyperlipidemia, and hyperglycemia) were assessed using clinical information obtained at the Specific Health Checkup. The following are the definitions of each metabolic risk factor used in this study: visceral obesity, waist circumference ≥ 85 and ≥ 90 cm in men and women, respectively; obesity, BMI ≥ 25 kg/m²; high BP, systolic BP ≥ 130 mmHg and diastolic BP ≥ 85 mmHg or taking antihypertensive medication; dyslipidemia, triglyceride ≥ 150 mg/dL, high-density lipoprotein cholesterol ＜40 mg/dL, or taking lipid-lowering medications; hyperglycemia, fasting blood glucose ≥ 100 mg/dL, hemoglobin A1c ≥ 5.6%, or taking antihyperglycemic medications. The cutoff value for blood glucose is different between the Japanese Society of Internal Medicine criteria (≥ 110 mg/dL)⁸⁾ and the Specific Health Checkup criteria (≥ 100 mg/dL)⁵⁾, with this study adopting the later criterion. According to the risk stratification algorithm used in the Specific Health Checkup, participants were classified into the ISI, MSI, and ISO groups based on the number of accumulated metabolic risk factors and smoking habits (Fig.1). Individuals aged ≥ 65 years who were classified into the ISI group were placed in the MSI group.

Fig.1. Risk stratification and lifestyle intervention grouping in the nationwide Specific Health Checkup program in Japan

High blood pressure: systolic blood pressure ≥ 130 mmHg, diastolic blood pressure ≥ 85 mmHg, or taking antihypertensive medications; dyslipidemia: fasting triglyceride ≥ 150 mg/dL, high-density lipoprotein cholesterol ＜40 mg/dL, or taking lipid-lowering medications; hyperglycemia: fasting blood glucose ≥ 100 mmHg, hemoglobin A1c ≥ 5.6 %, or taking antihyperglycemic medications. Individuals aged ≥ 65 years who were classified into the ISI group were placed in the MSI group. ISO, information supply only; MSI, motivation-support intervention; ISI, intensive support intervention.

Clinical Parameters

Data on histories of stroke and CAD, medication use, and smoking habit were obtained using a structured questionnaire during the Specific Health Checkup. The Charlson Comorbidity Index¹⁸⁾ was calculated using the health insurance claims data requested during the baseline period and used as an index of severe comorbidity. Nonvalvular and valvular atrial fibrillation were identified using any of the following disease codes: nonvalvular atrial fibrillation: 8847814, 8846942, 8847818, 8846694, 8850690, 8847735, 4273014, 8846503, 8846608, 4273006, 8844493, 8846906, 4273011, and 8847815; and valvular atrial fibrillation: 8846941. History of hospitalization during the baseline period was obtained from the health insurance claims data. Information on the participants’ latest certified care level during the baseline period was obtained from the Long-Term Care Insurance System.

Statistical Analysis

Values are expressed as the mean±standard deviation or frequency. The Cox proportional hazards model was used to determine factors associated with stroke and CAD, using nonobese individuals without any metabolic risk factors as the reference group. The assumption for proportionality was assessed using the Schoenfeld residuals test. The population attributable fraction was calculated using the following formula: pd_i (1−1/HR_i), where pd_i is the proportion among total cases arising from the ith exposure category and HR_i is the multivariable HR for the ith exposure category¹⁹⁾.

Statistical analyses were performed using JMP version 17.2.0 (SAS Institute, Cary, NC) or STATA version 18.5 (StataCorp LLC, College Station, TX). P＜0.05 indicated statistical significance.

Results

The clinical characteristics of the study participants are summarized in Table 1. After classifying participants aged ＜65 years according to the risk stratification method used in the Specific Health Checkup (Fig.1), the number of participants with obesity categorized into ISO, MSI, and ISI was 8,121, 15,897, and 32,778, respectively. Among the nonobese participants (n = 113,744), 40,437, 20,684, and 5,881 had one, two, and three metabolic risk factors, respectively. The detailed clinical characteristics of the study population according to the number of accumulated metabolic risk factors are presented in Supplementary Tables 3 (men) and 4 (women). The same information on the participants aged ≥ 65 and ＜75 years is shown in Supplementary Tables 5 (men) and 6 (women). Given that individuals with obesity who had three risk factors were classified as MSI in this age group, none of them were categorized into ISI.

Table 1.Clinical characteristics of the study participants

	＜65 years (170,540)	≥ 65, ＜75 years (196,341)
Age, years	55.9±8.2	69.3±2.8
Sex, men%	41.3	37.8
Current smoking, %	19.3	10.9
Body mass index, kg/m2	22.8±3.7	22.7±3.3
Body mass index ≥ 25kg/m2, %	24.1	21.2
Waist circumference, cm	81.8±10.1	82.5±9.3
Abdominal obesity, %	28.4	29.2
Systolic blood pressure, mmHg	124±17	131±17
Diastolic blood pressure, mmHg	75±12	76±11
Antihypertensive medication, %	17.9	36.7
High blood pressure, %	45.0	67.5
Triglyceride, mg/dL	118±92	116±71
HDL cholesterol, mg/dL	65±18	64±17
Lipid-lowering medication, %	13.6	26.8
Dyslipidemia, %	32.5	42.2
Fasting blood glucose, mg/dL	96±20	99±20
Hemoglobin A1c, %	5.6±0.7	5.8±0.7
Antihyperglycemic medication, %	4.1	7.8
Hyperglycemia, %	34.3	45.7
Atrial fibrillation, %	0.5	1.2
Charlson comorbidity index	0.6±1.1	1.0±1.4

Values are mean±standard deviation or frequency. The number of participants is shown in parentheses. Abdominal obesity: waist circumferences ≥ 85 and ≥ 90 cm for men and women, respectively; high blood pressure: systolic blood pressure ≥ 130 mmHg, diastolic blood pressure ≥ 85 mmHg, or taking antihypertensive medications; dyslipidemia: fasting triglyceride ≥ 150 mg/dL, high-density lipoprotein cholesterol ＜40 mg/dL, or taking lipid-lowering medications; hyperglycemia: fasting blood glucose ≥ 100 mmHg, hemoglobin A1c ≥ 5.6%, or taking antihyperglycemic medications. Fasting triglyceride was available for 66,951 and 80,456 participants aged ＜65 and ≥ 65 years, respectively.

Supplementary Table 3.Clinical characteristics of study participants by the accumulation of risk factors and intervention groups

	Men: ＜65 years
	None-obesity Number of risk factors				Obesity Intervention groups
	0	1	2	3	ISO	MSI	ISI
N	11,661	13,368	7,663	2,230	4,435	7,777	23,310
Age, years	51.7±8.9	55.6±8.4	57.9±7.3	59.3±6.4	51.1±8.7	54.8±8.5	56.0±8.0
BMI, kg/m2	21.0±1.9	21.4±1.8	21.8±1.8	22.2±1.6	25.3±2.3	25.8±2.6	26.6±3.1
BMI ≥ 25kg/m2, %	0	0	0	0	52.3	61.7	67.3
Waist circumference, cm	76.4±5.2	77.9±4.8	79.2±4.4	80.3±3.5	89.8±5.3	90.5±6.4	93.2±7.3
Abdominal obesity, %	0	0	0	0	93.5	88.7	98.6
Systolic blood pressure, mmHg	113±10	125±16	133±16	137±15	116±9	127±16	134±16
Diastolic blood pressure, mmHg	70±8	77±11	81±11	83±11	72±8	79±11	83±12
Antihypertensive medication, %	0	11.8	26.8	46.8	0	18.7	36.3
High blood pressure, %	0	48.2	78.3	100.0	0	50.8	79.1
Triglyceride, mg/dL	81±29	111±82	155±125	211±151	96±29	132±89	196±139
HDL cholesterol, mg/dL	65±15	64±17	60±17	55±14	57±12	56±13	51±13
Lipid lowering medication, %	0	3.8	14.8	37.2	0	5.0	20.8
Dyslipidemia, %	0	22.5	56	100.0	0	29.9	72.7
Fasting blood glucose, mg/dL	89±6	96±17	107±27	119±30	90±6	96±16	110±32
Hemoglobin A1c, %	5.4±0.3	5.5±0.6	5.8±0.9	6.2±1.1	5.4±0.3	5.6±0.6	6.0±1.1
Antihyperglycemic medication, %	0	2.3	7.6	17.1	0	2.3	11.3
Hyperglycemia, %	0	29.3	65.8	100.0	0	23.5	66.0

Values are mean±standard deviation or frequency. Obesity was defined as abdominal obesity (waist circumference, men ≥ 85 cm, women ≥ 90 cm) or BMI ≥ 25 kg/m². The number of risk factors was calculated as the cumulative number of high blood pressure, dyslipidemia, and hyperglycemia. ISO: information supply only, MSI: motivation-support intervention, ISI: intensive support intervention.

Supplementary Table 4.Clinical characteristics of study participants by the accumulation of risk factors and intervention groups

	Women: ＜65 years
	None-obesity Number of risk factors				Obesity Intervention groups
	0	1	2	3	ISO	MSI	ISI
N	35,081	27,069	13,021	3,651	3,686	8,120	9,468
Age, years	53.2±8.6	58.0±6.9	60.5±4.9	61.5±3.7	52.5±8.5	57.2±7.3	58.6±6.5
BMI, kg/m2	20.2±2.1	20.8±2.2	21.4±2.1	22.1±1.9	26.6±2.6	26.9±2.7	28.3±3.7
BMI ≥ 25kg/m2, %	0	0	0	0	82.8	88.1	85.7
Waist circumference, cm	74.3±6.8	76.3±6.8	78.4±6.5	80.3±5.8	90.7±6.9	90.6±7.1	96.0±7.7
Abdominal obesity, %	0	0	0	0	57.5	48.9	87.2
Systolic blood pressure, mmHg	110±11	125±17	132±16	136±14	115±9	129±17	135±16
Diastolic blood pressure, mmHg	67±8	75±11	78±11	79±10	70±8	78±11	80±11
Antihypertensive medication, %	0	12.8	31.1	54.3	0	23.5	48.2
High blood pressure, %	0	48.2	77.7	100.0	0	58.6	86.2
Triglyceride, mg/dL	73±27	93±53	123±80	156±101	88±29	116±65	159±93
HDL cholesterol, mg/dL	75±16	73±17	68±17	63±16	65±14	63±14	58±13
Lipid lowering medication, %	0	11.4	36.7	65.9	0	15.1	41.8
Dyslipidemia, %	0	20.7	59.7	100.0	0	30.9	75.3
Fasting blood glucose, mg/dL	87±6	92±13	100±19	115±27	89±6	95±16	111±31
Hemoglobin A1c, %	5.4±0.3	5.5±0.4	5.8±0.6	6.1±0.9	5.4±0.3	5.7±0.6	6.1±1.0
Antihyperglycemic medication, %	0	1.0	4.4	14.4	0	2.8	14.6
Hyperglycemia, %	0	31.1	62.6	100.0	0	33.1	75.5

Values are mean±standard deviation or frequency. Obesity was defined as abdominal obesity (waist circumference, men ≥ 85 cm, women ≥ 90 cm) or BMI ≥ 25 kg/m². The number of risk factors was calculated as the cumulative number of high blood pressure, dyslipidemia, and hyperglycemia. ISO: information supply only, MSI: motivation-support intervention, ISI: intensive support intervention.

Supplementary Table 5.Clinical characteristics of study participants by the accumulation of risk factors and intervention groups

	Men: ≥ 65, ＜75 years
	None-obesity Number of risk factors				Obesity Intervention groups
	0	1	2	3	ISO	MSI	ISI
N	6,367	14,746	12,243	4,495	2,086	34,306	0
Age, years	68.5±2.5	68.8±2.6	68.9±2.6	69.0±2.6	68.5±2.5	68.8±2.6
BMI, kg/m2	20.6±2.0	21.1±1.9	21.5±1.8	22.0±1.7	24.3±1.9	25.4±2.4
BMI ≥ 25kg/m2, %	0	0	0	0	33.3	51.8
Waist circumference, cm	76.5±5.4	77.8±5.1	78.9±4.6	80.2±3.9	89.1±4.1	91.3±5.8
Abdominal obesity, %	0	0	0	0	96.7	97.3
Systolic blood pressure, mmHg	116±10	130±17	136±16	138±16	118±8	136±16
Diastolic blood pressure, mmHg	70±8	77±11	79±11	79±11	71±7	80±11
Antihypertensive medication, %	0	25.4	42.0	59.2	0	53.2
High blood pressure, %	0	60.9	86.0	100.0	0	85.0
Triglyceride, mg/dL	81±27	93±48	122±78	169±103	95±28	150±94
HDL cholesterol, mg/dL	66±15	65±16	61±17	55±15	58±12	54±14
Lipid lowering medication, %	0	4.9	19.6	50.5	0	24.6
Dyslipidemia, %	0	11.6	43.4	100.0	0	56.2
Fasting blood glucose, mg/dL	90±6	96±17	109±26	119±27	91±6	107±25
Hemoglobin A1c, %	5.4±0.3	5.6±0.6	5.9±0.8	6.2±1.0	5.5±0.3	6.0±0.8
Antihyperglycemic medication, %	0	4.3	12.6	24.0	0	13.6
Hyperglycemia, %	0	27.5	70.6	100.0	0	61.5

Values are mean±standard deviation or frequency. Obesity was defined as abdominal obesity (waist circumference, men ≥ 85 cm, women ≥ 90 cm) or BMI ≥ 25 kg/m². The number of risk factors was calculated as the cumulative number of high blood pressure, dyslipidemia, and hyperglycemia. ISO: information supply only, MSI: motivation-support intervention, ISI: intensive support intervention.

Supplementary Table 6.Clinical characteristics of study participants by the accumulation of risk factors and intervention groups

	Women: ≥ 65, ＜75 years
	None-obesity Number of risk factors				Obesity Intervention groups
	0	1	2	3	ISO	MSI	ISI
N	17,533	34,810	28,335	11,424	1,867	28,129	0
Age, years	69.0±2.8	69.5±2.9	69.8±2.9	70.0±2.8	69.3±2.9	69.8±2.9
BMI, kg/m2	20.2±2.2	20.7±2.2	21.2±2.1	21.8±2.0	25.5±2.2	26.7±2.7
BMI ≥ 25kg/m2, %	0	0	0	0	64.5	78.2
Waist circumference, cm	75.4±7.2	77.0±7.0	78.5±6.8	80.1±6.0	91.6±5.8	93.0±6.7
Abdominal obesity, %	0	0	0	0	71.9	73.5
Systolic blood pressure, mmHg	115±10	129±17	135±16	138±15	118±8	136±16
Diastolic blood pressure, mmHg	68±8	74±11	76±10	76±10	70±8	77±10
Antihypertensive medication, %	0	22.2	42.1	62.9	0	54.9
High blood pressure, %	0	58.1	84.0	100.0	0	85.3
Triglyceride, mg/dL	80±27	92±41	112±62	141±81	92±27	129±69
HDL cholesterol, mg/dL	74±16	71±16	67±16	63±16	66±14	61±14
Lipid lowering medication, %	0	14.7	44.7	75.4	0	43.9
Dyslipidemia, %	0	19.8	59.4	100.0	0	60.4
Fasting blood glucose, mg/dL	88±6	92±12	99±19	114±24	90±6	102±22
Hemoglobin A1c, %	5.5±0.3	5.6±0.4	5.8±0.6	6.1±0.7	5.5±0.3	5.9±0.7
Antihyperglycemic medication, %	0	1.4	6.0	17.6	0	11.2
Hyperglycemia, %	0	22	56.6	100.0	0	58.1

Values are mean±standard deviation or frequency. Obesity was defined as abdominal obesity (waist circumference, men ≥ 85 cm, women ≥ 90 cm) or BMI ≥ 25 kg/m². The number of risk factors was calculated as the cumulative number of high blood pressure, dyslipidemia, and hyperglycemia. ISO: information supply only, MSI: motivation-support intervention, ISI: intensive support intervention.

Fig.2 depicts the hazard ratios and 95% confidence intervals for stroke. In the younger age group (＜65 years), the hazard ratio increased linearly with the accumulation of metabolic risk factors in both obese and nonobese groups, though the hazard ratio of the ISO group did not reach statistical significance. Similar results were observed in the older age group (≥ 65, ＜75 years) with somewhat smaller hazard ratios than those in the younger group. The Kaplan-Meier curves for stroke are shown in Supplementary Fig.2. Hazard ratios for CAD are depicted in Fig.3. Similar to the results of the analysis for stroke, a linear association was observed with the cumulative number of metabolic risk factors, though the trend was less clear than that with stroke possibly due to the smaller number of incident cases of CAD. The Kaplan-Meier curves for stroke are depicted in Supplementary Fig.3. Detailed calculation of the hazard ratio and population attributable fraction are summarized in Supplementary Tables 7, 8, 9, 10. When the hazard ratio of visceral obesity for stroke was calculated using a conventional Cox proportional hazard model adjusting of age, sex, high BP, dyslipidemia, and hyperglycemia, Charlson comorbidity index, and atrial fibrillation, the hazard ratio was significant only in younger population (＜65 years, hazard ratio = 1.17, P = 0.007; ≥ 65 years, hazard ratio = 1.07, P = 0.050). Similar results were observed in the analysis for CAD (＜65 years, hazard ratio = 1.65, P＜0.001; ≥ 65 years, hazard ratio = 1.14, P = 0.064), indicating that stratification by visceral obesity was less important in the elderly population.

Fig.2. Hazard ratios and population attributable fractions for stroke

Values are hazard ratio and 95% confidence interval (CI). Adjusted factors in the calculation of the hazard ratios were age, the Charlson Comorbidity Index, and atrial fibrillation. Obesity was defined as abdominal obesity (waist circumferences ≥ 85 and ≥ 90 cm for men and women, respectively) or body mass index ≥ 25 kg/m². The number of risk factors was calculated as the cumulative sum of high blood pressure, dyslipidemia, and hyperglycemia. Individuals aged ≥ 65 years who were classified into the ISI group were placed in the MSI group. The population attributable fractions of statistically significant groups are shown in the graph. ISO, information supply only; MSI, motivation-support intervention; ISI, intensive support intervention; N/A, not applicable.

Supplementary Fig.2. Kaplan-Meier curve for incidence of stroke

ISO: information supply only, MSI: motivation-support intervention, ISI: intensive support intervention, N/A: not applicable.

Fig.3. Hazard ratios and population attributable fractions for coronary artery diseases

Values are hazard ratio and 95% confidence interval (CI). Adjusted factors in the calculation of the hazard ratios were age, the Charlson Comorbidity Index, and atrial fibrillation. Obesity was defined as abdominal obesity (waist circumferences ≥ 85 and ≥ 90 cm for men and women, respectively) or body mass index ≥ 25 kg/m². The number of risk factors was calculated as the cumulative sum of high blood pressure, dyslipidemia, and hyperglycemia. Individuals aged ≥ 65 years who were classified into the ISI group were placed in the MSI group. The population attributable fractions of statistically significant groups are displayed in the graph. ISO, information supply only; MSI, motivation-support intervention; ISI, intensive support intervention; N/A, not applicable.

Supplementary Fig.3. Kaplan-Meier curve for incidence of coronary artery diseases

ISO: information supply only, MSI: motivation-support intervention, ISI: intensive support intervention.

Supplementary Table 7.Hazard ratio and population attributable fraction by the accumulation of risk factors and intervention groups

		Men: ＜65 years
		None-obesity Number of risk factors				Obesity Intervention groups
		0	1	2	3	ISO	MSI	ISI
		11,661	13,368	7,663	2,230	4,435	7,777	23,310
Stroke	Follow-up period, person-years	24,177,822	28,507,927	16,710,705	4,909,788	8,537,369	16,031,339	46,539,177
	Incident cases, n	53	164	124	58	22	86	399
	Adjusted HR (95% CI)	ref.	2.21	2.60	3.93	1.22	2.12	3.24
			(1.62−3.01)	(1.88−3.59)	(2.70−5.72)	(0.74−2.00)	(1.51−2.99)	(2.43−4.33)
	PAF, %		9.9	8.4	4.8	−	5.0	30.4
CAD	Follow-up period, person-years	24,201,003	28,638,156	16,804,603	4,969,469	8,537,421	16,091,962	46,737,816
	Incident cases, n	26	49	36	13	11	27	213
	Adjusted HR (95% CI)	ref.	1.44	1.72	2.05	1.23	1.46	3.91
			(0.89−2.32)	(1.03−2.86)	(1.05−4.00)	(0.61−2.49)	(0.85−2.50)	(2.59−5.89)
	PAF, %		−	4.0	1.8	−	−	42.3

Obesity was defined as abdominal obesity (waist circumference, men ≥ 85 cm, women ≥ 90 cm) or BMI ≥ 25 kg/m². The number of risk factors was calculated as the cumulative number of high blood pressure, dyslipidemia, and hyperglycemia. Adjusted factors in the calculation of hazard ratio (HR) were age, Charlson comorbidity index, and atrial fibrillation. P values for proportional hazard assumption was 0.894 (stroke) and 0.963 (coronary artery diseases). ISO: information supply only, MSI: motivation-support intervention, ISI: intensive support intervention, CAD: coronary artery disease, PAF: population attributable fraction.

Supplementary Table 8.Hazard ratio and population attributable fraction by the accumulation of risk factors and intervention groups

		Women: ＜65 years
		None-obesity Number of risk factors				Obesity Intervention groups
		0	1	2	3	ISO	MSI	ISI
		35,081	27,069	13,021	3,651	3,686	8,120	9,468
Stroke	Follow-up period, person-years	70,673,086	59,185,015	29,833,824	8,473,839	7,113,361	16,956,889	19,929,426
	Incident cases, n	122	178	103	44	15	65	102
	Adjusted HR (95% CI)	ref.	1.49	1.57	2.27	1.27	1.94	2.44
			(1.18−1.88)	(1.20−2.06)	(1.60−3.23)	(0.74−2.17)	(1.43−2.62)	(1.87−3.20)
	PAF, %		9.3	5.9	3.9	−	5.0	9.6
CAD	Follow-up period, person-years	70,774,386	59,374,640	29,943,227	8,521,229	7,128,566	17,010,262	20,009,362
	Incident cases, n	12	23	17	9	1	8	17
	Adjusted HR (95% CI)	ref.	1.81	2.35	4.13	0.87	2.26	3.73
			(0.90−3.67)	(1.10−5.00)	(1.71−9.98)	(0.11−6.72)	(0.92−5.55)	(1.76−7.91)
	PAF, %		−	11.2	7.8	−	−	14.3

Obesity was defined as abdominal obesity (waist circumference, men ≥ 85 cm, women ≥ 90 cm) or BMI ≥ 25 kg/m². The number of risk factors was calculated as the cumulative number of high blood pressure, dyslipidemia, and hyperglycemia. Adjusted factors in the calculation of hazard ratio (HR) were age, Charlson comorbidity index, and atrial fibrillation. P values for proportional hazard assumption was 0.677 (stroke) and 0.034 (coronary artery diseases). ISO: information supply only, MSI: motivation-support intervention, ISI: intensive support intervention, CAD: coronary artery disease, PAF: population attributable fraction.

Supplementary Table 9.Hazard ratio and population attributable fraction by the accumulation of risk factors and intervention groups

		Men: ≥ 65, ＜75 years
		None-obesity Number of risk factors				Obesity Intervention groups
		0	1	2	3	ISO	MSI	ISI
		6,367	14,746	12,243	4,495	2,086	34,306	0
Stroke	Follow-up period, person-years	12,756,628	28,222,414	22,720,611	8,093,470	4,003,498	61,363,736
	Incident cases, n	80	319	346	113	35	875
	Adjusted HR (95% CI)	ref.	1.76	2.34	2.12	1.41	2.18
			(1.38−2.25)	(1.84−2.99)	(1.59−2.82)	(0.95−2.09)	(1.74−2.75)
	PAF, %		7.8	11.2	3.4	−	26.8
CAD	Follow-up period, person-years	12,816,145	28,395,922	22,934,063	8,157,399	4,026,629	61,886,225
	Incident cases, n	18	92	107	55	13	318
	Adjusted HR (95% CI)	ref.	2.31	3.35	4.85	2.31	3.71
			(1.39−3.83)	(2.03−5.52)	(2.85−8.27)	(1.13−4.71)	(2.31−5.97)
	PAF, %		8.7	12.4	7.2	1.2	38.5

Obesity was defined as abdominal obesity (waist circumference, men ≥ 85 cm, women ≥ 90 cm) or BMI ≥ 25 kg/m². The number of risk factors was calculated as the cumulative number of high blood pressure, dyslipidemia, and hyperglycemia. Adjusted factors in the calculation of hazard ratio (HR) were age, Charlson comorbidity index, and atrial fibrillation. P values for proportional hazard assumption was 0.533 (stroke) and 0.708 (coronary artery diseases). ISO: information supply only, MSI: motivation-support intervention, ISI: intensive support intervention, CAD: coronary artery disease, PAF: population attributable fraction.

Supplementary Table 10.Hazard ratio and population attributable fraction by the accumulation of risk factors and intervention groups

		Women: ≥ 65, ＜75 years
		None-obesity Number of risk factors				Obesity Intervention groups
		0	1	2	3	ISO	MSI	ISI
		17,533	34,810	28,335	11,424	1,867	28,129	0
Stroke	Follow-up period, person-years	42,803,949	83,245,127	66,795,354	26,426,068	4,514,720	63,960,263
	Incident cases, n	197	619	559	252	27	617
	Adjusted HR (95% CI)	ref.	1.54	1.68	1.86	1.27	1.92
			(1.32−1.81)	(1.43−1.98)	(1.55−2.25)	(0.85−1.90)	(1.63−2.25)
	PAF, %		9.6	10.0	5.1	−	13.0
CAD	Follow-up period, person-years	42,981,483	83,834,933	67,280,027	26,644,755	4,530,337	64,545,705
	Incident cases, n	27	70	94	41	7	88
	Adjusted HR (95% CI)	ref.	1.28	2.10	2.29	2.41	2.06
			(0.82−2.00)	(1.37−3.23)	(1.40−3.72)	(1.05−5.52)	(1.34−3.18)
	PAF, %		−	15.1	7.1	1.3	13.8

Obesity was defined as abdominal obesity (waist circumference, men ≥ 85 cm, women ≥ 90 cm) or BMI ≥ 25 kg/m². The number of risk factors was calculated as the cumulative number of high blood pressure, dyslipidemia, and hyperglycemia. Adjusted factors in the calculation of hazard ratio (HR) were age, Charlson comorbidity index, and atrial fibrillation. P values for proportional hazard assumption was 0.168 (stroke) and 0.655 (coronary artery diseases). ISO: information supply only, MSI: motivation-support intervention, ISI: intensive support intervention, CAD: coronary artery disease, PAF: population attributable fraction.

Considering that individuals receiving a regular treatment are excluded from the Specific Health Guidance, the hazard ratio for stroke and CAD were recalculated in the subpopulation who did not receive treatment. Among individuals aged ＜65 years, the hazard ratio for stroke increased linearly with the accumulation of metabolic risk factors in both obese and nonobese individuals (Supplementary Fig.4). In the nonobese population, high BP, but not hyperlipidemia and hyperglycemia, was associated with stroke when the analysis was performed for each metabolic risk factors individually (Table 2). The linear association was less pronounced among those aged ≥ 65 years (Supplementary Fig.4). In contrast, no clear association was observed between the number of metabolic risk factors and CAD in the subpopulation that did not take medications (Supplementary Fig.5).

Supplementary Fig.4. Hazard ratio and population attributable fraction for stroke in participants without taking medications

Hazard ratio was calculated using the nonobese group without any risk factors as a reference. Adjusted factors in the calculation of hazard ratio (HR) were age, Charlson comorbidity index, and atrial fibrillation. Population attributable fractions of the groups that showed significant associations are shown in the graph. ISO: information supply only, MSI: motivation-support intervention, ISI: intensive support intervention, N/A: not applicable.

Table 2.Hazard ratios for stroke in nonobese participants aged ＜65 years not taking any medication

	Men			Women
	n	HR (95% CI)	P	n	HR (95% CI)	P
Participants	28,253			62,110
Incidence of stroke	268			301
High blood pressure	9,345	2.85 (2.21–3.67)	＜0.001	15,187	2.17 (1.72–2.74)	＜0.001
Hyperlipidemia	5,951	1.37 (1.05–1.80)	0.022	5,793	1.13 (0.80–1.59)	0.490
Hyperglycemia	7,802	1.01 (0.78–1.30)	0.948	13,636	0.88 (0.67–1.15)	0.349

Nonobese individuals were defined as those having a waist circumference of ＜85 or ＜90 cm for men and women, respectively, and a body mass index of ＜25 kg/m². The follow-up duration was 59,162,847 and 128,718,481 person-years for men and women, respectively. High blood pressure: systolic blood pressure ≥ 130 mmHg or diastolic blood pressure ≥ 85 mmHg; dyslipidemia: fasting triglyceride ≥ 150 mg/dL or high- density lipoprotein (HDL) cholesterol ＜40 mg/dL; hyperglycemia: fasting blood glucose ≥ 100 mmHg or hemoglobin A1c ≥ 5.6%. HR, hazard ratio; CI, confidence interval.

Supplementary Fig.5. Hazard ratio and population attributable fraction for coronary artery diseases in participants without taking medications

Hazard ratio was calculated using the nonobese group without any risk factors as a reference. Adjusted factors in the calculation of hazard ratio (HR) were age, Charlson comorbidity index, and atrial fibrillation. Population attributable fractions of the groups that showed significant associations are shown in the graph. ISO: information supply only, MSI: motivation-support intervention, ISI: intensive support intervention, N/A: not applicable.

Discussion

Using prefecture-wide health insurance data, the current longitudinal study revealed that the accumulation of metabolic risk factors linearly increased the risk for stroke and CAD to the same degree, regardless of the presence or absence of obesity. A similar association was observed in the analysis for stroke in a subpopulation not taking any medications, which was the target population of the Specific Health Guidance.

Several studies on the Japanese general population^5-7) have reported that the accumulation of metabolic risk factors linearly increased the risk of CVD mortality even among individuals without obesity. Given that the hazard ratio was equivalent between obese and nonobese populations with the same number of metabolic risk factors, one can question the significance of visceral obesity as a criterion for metabolic syndrome. Irie et al.⁶⁾ reported that the criteria for metabolic syndrome established by the International Diabetic Federation definition, which adopts visceral obesity as an essential criterion, was not associated with CVD mortality, whereas those of the American Heart Association/National Heart, Lung, and Blood Institute, which consider obesity as one of metabolic risk factors, was associated with CVD mortality. Although these previous studies performed the analysis with CVD mortality as an outcome, the results of the current study and another study that combined general population-based cohorts¹²⁾ showed that the accumulation of metabolic risk factors in nonobese individuals was also associated with incidence of CVD. These findings suggest that visceral obesity is not an essential criterion for defining metabolic syndrome. Although changing the criteria for visceral obesity might yield somewhat different results, we did not try it because the aim of this study was to examine the prognostic significance of metabolic syndrome determined according to the risk stratification method used in the Specific Health Checkup.

East Asians have a smaller body size than European populations²⁰⁾. The results of the analysis of NIPPON DATA 2010, which contains data from the nationwide health and nutrition survey throughout Japan, indicated that 35.8% of men and 20.1% of women aged 20–64 years had a BMI of ≥ 25 kg/m^{2 21)}, which was somewhat larger than that observed in our study population. In contrast, the analysis of a nationally representative sample of civilian noninstitutionalized individuals from the United States from 2009 to 2010 reported that approximately 70% adults had a BMI of ≥ 25 kg/m^{2 22)}. The large frequency difference indicates that the need for lifestyle interventions among nonobese individuals varies by population. Japanese individuals may benefit from interventions aimed at preventing CVD among nonobese individuals.

In the nationwide health screening system in Japan, obese individuals classified into the ISI group receive intensive health intervention with the goal of reducing waist circumference by 2 cm and/or body weight by 2 kg. However, our study results showed that the hazard ratio for stroke was significant, especially among those aged ＜65 years, in nonobese individuals even after excluding individuals receiving treatment from the analysis. A similar population attributable fraction was observed between nonobese and obese populations, further supporting the importance of lifestyle interventions in the Specific Health Guidance for the nonobese population. In the nonobese population, only high BP showed a significant association with stoke, suggesting that interventions aimed at managing high BP would be efficient in this group.

A major strength of this study was its large sample size, which allowed detailed subgroup analysis of the accumulated number of metabolic risk factors. However, this study also has several limitations. First, the outcomes may be misclassified given our focus on the incidence of stroke and coronary artery diseases based on health insurance claim information. However, any misclassification would be unrelated to the clinical characteristics of the incident cases and is unlikely to cause serious bias in the results. Second, the study participants were all enrolled in the National Health Insurance. Given our lack of data on individuals enrolled in other insurance systems, the study population was not representative of the entire Japanese population. Caution should be exercised when interpreting our findings regarding the population attributable fraction.

In conclusion, our longitudinal analysis using prefecture-wide National Health Insurance data revealed that the accumulation of metabolic risk factors is associated with the incidence of CVD even among individuals without obesity. This suggests the importance of closely monitoring nonobese individuals with metabolic risk factors, who may be overlooked by metabolic syndrome criteria that consider visceral obesity an essential criterion.

Acknowledgements

We thank the Shizuoka Prefectural Government, 35 city and town offices in the Shizuoka Prefecture and the Shizuoka Federation of National Health Insurance Organizations for their help in constructing the SKDB dataset. We also thank the editors of Crimson Interactive Pvt. Ltd. for their assistance in the preparation of this manuscript.

Financial Support

No financial support was received for this study.

Conflict of Interest

The authors have no conflict of interest to disclose.

References

• 1)  Mottillo S, Filion KB, Genest J, Joseph L, Pilote L, Poirier P, Rinfret S, Schiffrin EL, Eisenberg MJ. The metabolic syndrome and cardiovascular risk a systematic review and meta-analysis. J Am Coll Cardiol, 2010; 56: 1113-1132
• 2)  Iso H, Sato S, Kitamura A, Imano H, Kiyama M, Yamagishi K, Cui R, Tanigawa T, Shimamoto T. Metabolic syndrome and the risk of ischemic heart disease and stroke among Japanese men and women. Stroke, 2007; 38: 1744-1751
• 3)  Ninomiya T, Kubo M, Doi Y, Yonemoto K, Tanizaki Y, Rahman M, Arima H, Tsuryuya K, Iida M, Kiyohara Y. Impact of metabolic syndrome on the development of cardiovascular disease in a general Japanese population: the Hisayama study. Stroke, 2007; 38: 2063-2069
• 4)  Kokubo Y, Okamura T, Yoshimasa Y, Miyamoto Y, Kawanishi K, Kotani Y, Okayama A, Tomoike H. Impact of metabolic syndrome components on the incidence of cardiovascular disease in a general urban Japanese population: the Suita study. Hypertens Res, 2008; 31: 2027-2035
• 5)  Kadota A, Hozawa A, Okamura T, Kadowak T, Nakmaura K, Murakami Y, Hayakawa T, Kita Y, Okayama A, Nakamura Y, Kashiwagi A, Ueshima H; NIPPON DATA Research Group. Relationship between metabolic risk factor clustering and cardiovascular mortality stratified by high blood glucose and obesity: NIPPON DATA90, 1990-2000. Diabetes Care, 2007; 30: 1533-1538
• 6)  Irie F, Iso H, Noda H, Sairenchi T, Otaka E, Yamagishi K, Doi M, Izumi Y, Ota H. Associations between metabolic syndrome and mortality from cardiovascular disease in Japanese general population, findings on overweight and non-overweight individuals. Ibaraki Prefectural Health Study. Circ J, 2009; 73: 1635-1642
• 7)  Saito I, Iso H, Kokubo Y, Inoue M, Tsugane S. Metabolic syndrome and all-cause and cardiovascular disease mortality: Japan Public Health Center-based Prospective (JPHC) Study. Circ J, 2009; 73: 878-884
• 8)  Kadota A, Miura K, Okamura T, Hozawa A, Murakami Y, Fujiyoshi A, Takashima N, Hayakawa T, Kita Y, Okayama A, Nakamura Y, Ueshima H; NIPPON DATA90 Research Group. Relationship of moderate metabolic risk factor clustering to cardiovascular disease mortality in non-lean Japanese: a 15-year follow-up of NIPPON DATA90. Atherosclerosis, 2011; 215: 209-213
• 9)  Alberti KG, Zimmet P, Shaw J. Metabolic syndrome--a new world-wide definition. A Consensus Statement from the International Diabetes Federation. Diabet Med, 2006; 23: 469-480
• 10)  Definition and the diagnostic standard for metabolic syndrome--Committee to Evaluate Diagnostic Standards for Metabolic Syndrome. Nihon Naika Gakkai Zasshi, 2005; 94: 794-809. Japanese
• 11)  Kohro T, Furui Y, Mitsutake N, Fujii R, Morita H, Oku S, Ohe K, Nagai R. The Japanese national health screening and intervention program aimed at preventing worsening of the metabolic syndrome. Int Heart J, 2008; 49: 193-203
• 12)  Iso H, Cui R, Takamoto I, Kiyama M, Saito I, Okamura T, Miyamoto Y, Higashiyama A, Kiyohara Y, Ninomiya T, Yamada M, Nakagawa H, Sakurai M, Shimabukuro M, Higa M, Shimamoto K, Saito S, Daimon M, Kayama T, Noda M, Ito S, Yokote K, Ito C, Nakao K, Yamauchi T, Kadowaki T. Risk classification for metabolic syndrome and the incidence of cardiovascular disease in Japan with low prevalence of obesity: a pooled analysis of 10 prospective cohort studies. J Am Heart Assoc, 2021; 10: e020760
• 13)  Nakatani E, Tabara Y, Sato Y, Tsuchiya A, Miyachi Y. Data Resource Profile of Shizuoka Kokuho Database (SKDB) Using Integrated Health- and Care-insurance Claims and Health Checkups: The Shizuoka Study. J Epidemiol, 2022; 32: 391-400
• 14)  Tabara Y, Shoji-Asahina A, Ogawa A, Sato Y. Additive association of blood pressure and short stature with stroke incidence in 450,000 Japanese adults: the Shizuoka study. Hypertens Res, 2024; 47: 2075-2085
• 15)  Tabara Y, Shoji-Asahina A, Ogawa A, Sato Y. Metabolically healthy obesity and risks of cardiovascular disease and all-cause mortality, a matched cohort study: the Shizuoka study. Int J Obes (Lond), 2024; 48: 1164-1169
• 16)  Park J, Kwon S, Choi EK, Choi YJ, Lee E, Choe W, Lee SR, Cha MJ, Lim WH, Oh S. Validation of diagnostic codes of major clinical outcomes in a National Health Insurance database. Int J Arrhythm, 2019; 20: 5
• 17)  Hayashida K, Murakami G, Matsuda S, Fushimi K. History and profile of Diagnosis Procedure Combination (DPC): Development of a real data collection system for acute inpatient care in Japan. J Epidemiol, 2021; 31: 1-11
• 18)  Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis, 1987; 40: 373-383
• 19)  Rockhill B, Newman B, Weinberg C. Use and misuse of population attributable fractions. Am J Public Health, 1998; 88: 15-19
• 20)  NCD Risk Factor Collaboration (NCD-RisC). Trends in adult body-mass index in 200 countries from 1975 to 2014: a pooled analysis of 1698 population-based measurement studies with 19·2 million participants. Lancet, 2016; 387: 1377-1396
• 21)  Nakamura T, Nakamura Y, Saitoh S, Okamura T, Yanagita M, Yoshita K, Kita Y, Murakami Y, Yokomichi H, Nishi N, Okuda N, Kadota A, Ohkubo T, Ueshima H, Okayama A, Miura K. Relationship Between Socioeconomic Status and the Prevalence of Underweight, Overweight or Obesity in a General Japanese Population: NIPPON DATA2010. J Epidemiol, 2018; 28: S10-S16
• 22)  Flegal KM, Carroll MD, Kit BK, Ogden CL. Prevalence of obesity and trends in the distribution of body mass index among US adults, 1999-2010. JAMA, 2012; 307: 491-497