Effects of the Relative Distribution of Body Mass Index on the Incidence of Hypertension and Diabetes Mellitus,Using the Korean Obesity Index

Abstract

Background:

The Korean Obesity Index, which contains standard reference data (SRD) of obesity, was established in 2016 and revised in 2017 based on national health screening data to provide the distribution of the body mass index (BMI) of the whole population of Korea as a reference. This study aimed to investigate the effect of the SRD of obesity on the incidence of hypertension (HTN) and diabetes mellitus (DM).

Methods:

The percentile of BMI was calculated for each of the 864 subgroups by defined by gender, region, and age group according to the groupings in the SRD. Incident cases were defined as the presence of HTN and DM and medication prescription in the health care utilization database for a given individual in 2017, but not in 2015–2016. Logistic regression for the incidence of HTN and DM according to the relative distribution of BMI was performed. Gender, age, insurance type, insurance contribution, smoking, drinking, physical activity, blood pressure, waist circumference, fasting glucose, triglycerides, high-density lipoprotein cholesterol, the Charlson comorbidity index (2012–2014), and diagnosis and medication for HTN and DM (2015–2017) were adjusted in the analysis.

Results:

The C-statistics of the fully adjusted model for HTN and DM were 0.799 and 0.852, respectively. The risks of HTN and DM increased by 1.007 and 1.011 times, respectively, for each 1-percentile increase in BMI.

Conclusion:

The results showed that BMI was associated with the incidence of HTN and DM according to the SRD. The relative distribution of BMI can be used to motivate self-care through providing more detailed information to individuals.

Introduction

It is well known that obesity is a cause of various diseases and mortality.^1,2 Management of obesity by exercise and nutrition has become essential for preventing complications.³ Providing knowledge and information is essential for effective management.^4,5 The “Know your number” campaign was a good example of self-care through information provision.⁶ Body mass index (BMI) is the most important piece of information in this regard because it is used as a diagnostic criterion for obesity. However, the diagnostic criteria for obesity vary by race and region. In many countries, a BMI ≥30 kg/m² is the diagnostic criterion for obesity, whereas 25 kg/m² is used as the diagnostic criterion in the Asia-Pacific region.⁷ This means that an analysis based on local data is important for obesity management. In the Republic of Korea (hereafter “Korea”), the Korean Obesity Index, which contains standard reference data (SRD) of obesity, was established in 2016 and revised in 2017 based on national health screening data to provide the distribution of the BMI of the whole population of Korea as a reference.⁸

In the SRD of obesity, the mean BMI, relative distribution, and uncertainty of measurements were presented for 864 subgroups of the Korean population by region, gender, and age. The SRD was created based on measurement data of roughly 18 million people from the National Health Insurance Service (NHIS). The Korea Research Institute of Standards and Science and the Ministry of Trade, Industry and Energy participated in the process of verification, review, and registration of the SRD (Supplementary Table S1). The Ministry of Health and Welfare provides biennially (annually for manual workers) a general health screening program for the adult insured population in Korea. In 2016, the screening participation rate was 77.7% among the eligible population, and the number of participants reached about 13 million.⁹ The screening data have been used to compare regional health levels in Korea and internationally to provide statistics for projects such as the noncommunicable disease-risk factor collaboration.¹⁰

In the SRD of obesity, the relative values are given within populations with similar characteristics (gender, age, and region) rather than as absolute values of BMI. Previous research is not sufficient to determine whether relative values of BMI are associated with chronic disease. In the study of associations between BMI and disease, many studies have focused on cutoff points for the absolute value of BMI.^11,12 It is widely known that the higher the BMI, the higher the risk of complications, and the lower the BMI, the lower the incidence of complications. However, even if the BMI is not very high, the risk is much higher for younger people with a relatively high BMI. In the elderly, even for individuals with the same BMI, the risk of disease is higher if that BMI value is relatively high among similar groups. In other words, information about the relative distribution of BMI can be helpful in disease management and prevention.¹³

This study aimed to investigate the effect of the SRD of obesity, which provides the relative position of BMI compared with similar groups, on the incidence of hypertension (HTN) and diabetes mellitus (DM). Through the results of this study, we hope to support self-care among obese people and to contribute to health promotion.

Methods

Data sources and variables

We used data from the National Health Information Database (NHID) of the NHIS (NHIS-2018-1-202). The NHID includes (1) an eligibility database containing socioeconomic status indicators, such as gender, age, and income-based insurance contribution; (2) a national health screening database containing a questionnaire on health behavior, such as smoking and alcohol use, and anthropometric information such as BMI and waist circumference; (3) a health care utilization database containing information on inpatient and outpatient medical services (diagnosis, length of stay, services provided, and treatment costs) and prescription records (medication codes, days prescribed, and daily dosage); (4) a long-term care insurance database containing information on applications for long-term care service; and (5) a health care provider database containing information on health care facilities, equipment, and manpower.¹⁴ Since the national health insurance covers the entire population as the single public insurer of Korea, this information includes the eligibility of the entire population and all medical services claimed by health care facilities in Korea.

In this study, we used the data from the eligibility database in 2015, the national health screening database in 2015–2016, and the health care utilization database in 2012–2017. The variables of gender, age, insurance type, and income-based insurance contribution in the eligibility database were included, and the valuables drawn from the health screening database included BMI, smoking, drinking (frequency per week and drinking amount in terms of standard drinks), physical activity (weekly frequency of high-intensity activity, moderate-intensity activity, and walking), systolic and diastolic blood pressure, waist circumference, fasting glucose, triglycerides, and high-density lipoprotein (HDL) cholesterol levels. The Charlson comorbidity index (CCI) (2012–2014) and diagnosis and medication codes for HTN and DM (2015–2017) were drawn from the health care utilization database. The CCI was calculated using the disease score and classification outlined in a previous study.¹⁵ The percentile of BMI was calculated for each of the 864 subgroups defined by gender (2 groups: men and women), regions (16 groups defined by metropolitan cities and provinces: Seoul, Busan, Daegu, Incheon, Gwangju, Daejeon, Ulsan, Gyeonggi-do, Gangwon-do, Chungcheongbuk-do, Chungcheongnam-do, Jeollabuk-do, Jeollanam-do, Gyeongsangbuk-do, Gyeongsangnam-do, and Jeju-do), and age group (27 groups: 20–24, 25–26, 27–28, 29–30, …, 73–74, 75+ years) according to the grouping in the SRD of obesity.

The research protocol was reviewed and approved by the Institutional Review Board of NHIS (IRB No. Sa-2018-HR-04-001).

Statistical methods

A descriptive analysis of the general characteristics of the study participants and a logistic regression analysis of the incidence of HTN and DM were performed. Incident cases were defined as the presence of a HTN or DM diagnosis and medication prescription codes in the health care utilization database for an individual in 2017, but not in 2015–2016. To clearly identify the incidence of HTN and DM even if there was no diagnosis and prescription history, we excluded cases with an elevated systolic (140 mmHg or more) or diastolic blood pressure (90 mmHg or more) or fasting blood glucose (126 mg/dL or more) in the 2015–2016 national health screening database. The diagnosis code was described in accordance with the Korean Standard Classification of Disease. The HTN codes were I10–I15 and the DM codes were E10–E14. Logistic regression analysis of the incidence of HTN and DM was performed, and we analyzed four models according to the variables that were adjusted. In model 1, only the percentile of the BMI was used as an independent variable. In model 2, gender and age were added as variables that were adjusted. In model 3, we additionally adjusted for the following variables, as well as those in model 2: smoking, alcohol drinking, physical activity, insurance type, income-based insurance contribution, and CCI score. In model 4, measured values of waist circumference, fasting blood glucose, triglycerides, HDL cholesterol levels, and systolic and diastolic blood pressure were added to the variables that were adjusted in model 3. In addition, a stratification analysis was performed according to gender, age, and BMI to investigate differences in the results according to the characteristics of the study participants.

Results

General characteristics of the study population

The general characteristics of the study population are shown in Table 1. The largest age group was 40–49 years old, and there were more nonsmokers than smokers or ex-smokers. There were more women than men, and 56.6% of participants were employed insured. In 2017, the incidence rates of HTN and DM were 1.6% and 0.3%, respectively. The most common CCI score was 0 points (61.7%). Alcohol was consumed 0.9 times per week on average with 2.6 standard drinks per time. The mean systolic and diastolic blood pressure was 117.2 and 72.9 mmHg, respectively. The mean frequency of high-intensity activity, moderate-intensity activity, and walking were 1.1, 1.3, and 3.0 times per week, respectively. The mean waist circumference, fasting glucose, triglyceride, HDL cholesterol, and income-based insurance contribution were 78.9 cm, 93.8 mg/dL, 117.4 mg/dL, 57.8 mg/dL, and 103,893.7 Korean won, respectively.

Table 1.

General Characteristics of the Study Participants (n = 12,208,507)

Variables	n	%
Total	12,208,507	100.0
Age (years)
≤39	3,257,661	26.7
40–49	3,777,334	30.9
50–59	2,995,358	24.5
60–69	1,558,289	12.8
70+	619,865	5.1
Smoking
Current smoker (20 pack-years or more)	731,922	6.0
Current smoker (under 20 pack-years)	1,602,436	13.1
Past smoker (20 pack-years or more)	461,063	3.8
Past smoker (under 20 pack-years)	1,310,379	10.7
Nonsmoking	8,102,707	66.4
Gender
Men	5,371,833	44.0
Women	6,836,674	56.0
Insurance type
Self-employed insured	1,475,168	12.1
Dependents of the self-employed insured	1,058,457	8.7
Employed insured	6,909,948	56.6
Dependents of the employed insured	2,620,153	21.5
Medical aid beneficiaries	120,863	1.0
Dependents of medical aid beneficiaries	23,918	0.2
Incident cases of HTN	192,492	1.6
Incident cases of DM	37,389	0.3
CCI
0	7,535,794	61.7
1	3,247,874	26.6
2+	1,424,839	11.7

	Mean	SD
Weekly frequency of alcohol drinking	0.9	1.4
Amount of alcohol drinking (standard drinks)	2.6	4.1
Systolic blood pressure (mmHg)	117.2	11.5
Diastolic blood pressure (mmHg)	72.9	8.2
Weekly frequency of high-intensity exercise	1.1	1.6
Weekly frequency of moderate-intensity exercise	1.3	1.8
Weekly frequency of walking	3.0	2.4
Waist circumference (cm)	78.9	13.2
Fasting blood glucose (mg/dL)	93.8	10.7
Triglycerides (mg/dL)	117.4	86.2
HDL cholesterol (mg/dL)	57.8	19.6
Monthly insurance contribution (Korean won)	103,893.7	91,300.0

CCI, Charlson comorbidity index; DM, diabetes mellitus; HDL, high-density lipoprotein; HTN, hypertension.

Logistic regression results for the incidence of HTN and DM

The results of the logistic regression analysis of the incidence of HTN and DM are presented in Tables 2 and 3. As a result of the analysis for HTN and DM, the C-statistics of model 4 (adjusted for all variables in Table 2) were 0.799 and 0.852, respectively. The risks of HTN and DM increased by 1.007 and 1.011 times, respectively, for each 1-percentile increase in BMI. Nonetheless, the absolute BMI value was more predictive of disease than the relative distribution (Supplementary Tables S2 and S3; C-statistics of model 4 for HTN and DM: 0.800 and 0.854, respectively).

Table 2.

Logistic Regression Results for the Incidence of Hypertension According to the Percentile of Body Mass Index (n = 12,208,507)

	Model 1			Model 2			Model 3			Model 4
	OR	95% CI		OR	95% CI		OR	95% CI		OR	95% CI
Percentile of BMI	1.010	1.010	1.010	1.013	1.013	1.013	1.013	1.013	1.013	1.007	1.007	1.007
Gender
Men (ref: women)				1.283	1.271	1.295	1.093	1.079	1.107	0.869	0.857	0.881
Age				1.067	1.067	1.067	1.067	1.066	1.067	1.059	1.058	1.059
Smoking (ref: nonsmoker)
Current smoker (20 pack-years or more)							1.305	1.280	1.330	1.345	1.319	1.371
Current smoker (under 20 pack-years)							1.168	1.147	1.189	1.145	1.124	1.166
Past smoker (20 pack-years or more)							1.101	1.079	1.123	1.154	1.131	1.177
Past smoker (under 20 pack-years)							1.077	1.058	1.095	1.081	1.062	1.100
Weekly frequency of alcohol drinking							1.039	1.035	1.043	1.028	1.024	1.032
Amount of alcohol drinking (standard drinks)							1.018	1.017	1.020	1.008	1.007	1.010
Insurance type (ref: employed insured)
Self-employed insured							1.041	1.026	1.056	1.044	1.030	1.059
Dependents of the self-employed insured							1.100	1.081	1.119	1.059	1.040	1.078
Dependents of the employed insured							0.997	0.984	1.010	0.975	0.962	0.988
Medical aid beneficiaries							1.493	1.439	1.550	1.547	1.490	1.607
Dependents of the medical aid beneficiary							1.520	1.401	1.649	1.497	1.379	1.625
Weekly frequency of high-intensity exercise							0.997	0.994	1.001	0.998	0.994	1.001
Weekly frequency of moderate-intensity exercise							0.995	0.992	0.999	0.996	0.993	0.999
Weekly frequency of walking							0.992	0.99	0.994	0.991	0.989	0.993
CCI							1.134	1.129	1.139	1.150	1.145	1.154
Monthly insurance contribution (Korean won)							1.000	1.000	1.000	1.000	1.000	1.000
Waist circumference (cm)										1.003	1.003	1.004
Fasting blood glucose (mg/dL)										1.004	1.003	1.004
Triglycerides (mg/dL)										1.000	1.000	1.000
HDL cholesterol (mg/dL)										0.996	0.996	0.997
Systolic blood pressure (mmHg)										1.043	1.042	1.043
Diastolic blood pressure (mmHg)										1.032	1.031	1.033
C-statistics	0.581			0.752			0.758			0.799
AIC (intercept only: 1,979,554.5)	1,963,914.4			1,827,076.0			1,818,866.1			1,753,405.0

AIC, Akaike information criterion; BMI, body mass index; CI, confidence interval; OR, odds ratio.

Table 3.

Logistic Regression Results for the Incidence of Diabetes Mellitus According to the Percentile of Body Mass Index (n = 12,208,507)

	Model 1			Model 2			Model 3			Model 4
	OR	95% CI		OR	95% CI		OR	95% CI		OR	95% CI
Percentile of BMI	1.018	1.017	1.018	1.020	1.020	1.021	1.020	1.020	1.020	1.011	1.011	1.012
Gender
Men (ref: women)				1.412	1.383	1.441	1.187	1.153	1.223	0.845	0.819	0.872
Age				1.061	1.060	1.062	1.056	1.055	1.057	1.041	1.040	1.042
Smoking (ref: nonsmoker)
Current smoker (20 pack-years or more)							1.993	1.915	2.074	1.692	1.624	1.762
Current smoker (under 20 pack-years)							1.455	1.399	1.514	1.285	1.235	1.337
Past smoker (20 pack-years or more)							1.424	1.364	1.487	1.336	1.279	1.395
Past smoker (under 20 pack-years)							1.112	1.069	1.156	1.071	1.030	1.115
Weekly frequency of alcohol drinking							0.951	0.942	0.960	0.931	0.922	0.941
Amount of alcohol drinking (standard drinks)							1.002	0.999	1.006	0.989	0.985	0.993
Insurance type (ref: employed insured)
Self-employed insured							1.075	1.042	1.109	1.048	1.016	1.081
Dependents of the self-employed insured							1.090	1.047	1.134	1.037	0.996	1.079
Dependents of the employed insured							0.998	0.969	1.027	0.983	0.954	1.012
Medical aid beneficiaries							1.795	1.668	1.932	1.631	1.514	1.756
Dependents of the medical aid beneficiary							1.757	1.492	2.069	1.587	1.345	1.872
Weekly frequency of high-intensity exercise							0.989	0.982	0.997	0.992	0.985	1.000
Weekly frequency of moderate-intensity exercise							0.999	0.992	1.006	1.003	0.996	1.010
Weekly frequency of walking							0.992	0.988	0.996	0.997	0.993	1.002
CCI							1.235	1.224	1.245	1.149	1.139	1.159
Monthly insurance contribution (Korean won)							1.000	1.000	1.000	1.000	1.000	1.000
Waist circumference (cm)										1.005	1.004	1.005
Fasting blood glucose (mg/dL)										1.101	1.100	1.102
Triglycerides (mg/dL)										1.001	1.001	1.001
HDL cholesterol (mg/dL)										0.989	0.988	0.989
Systolic blood pressure (mmHg)										1.001	0.999	1.002
Diastolic blood pressure (mmHg)										1.006	1.005	1.008
C-statistics	0.627			0.746			0.760			0.852
AIC (intercept only: 507,518.7)	498,508.3			475,036.2			470,893.0			424,976.4

The stratification analysis (Table 4) showed that the C-statistic in model 4 was higher in women (0.824) than men (0.777) for HTN and higher in men (0.889) than women (0.860) for DM. According to the age group, the C-statistics of HTN (0.815) and DM (0.894) were higher in individuals in their 30s or under than in other age groups. According to the BMI classification, underweight participants had the highest C-statistics for both HTN (0.873) and DM (0.897). In this subgroup, decreases in the percentile of BMI were associated with a higher risk of HTN and DM.

Table 4.

Stratified Logistic Regression Results for the Incidence of Hypertension and Diabetes Mellitus

Target disease	Model 1				Model 2				Model 3				Model 4
Stratification variables	Percentile of BMI			C-statistics	Percentile of BMI			C-statistics	Percentile of BMI			C-statistics	Percentile of BMI			C-statistics
Stratification variables	OR	95% CI		C-statistics	OR	95% CI		C-statistics	OR	95% CI		C-statistics	OR	95% CI		C-statistics
Hypertension
Gender
Men	1.024	1.023	1.025	0.682	1.024	1.023	1.025	0.718	1.024	1.023	1.025	0.726	1.013	1.012	1.015	0.777
Women	1.010	1.010	1.010	0.578	1.013	1.013	1.013	0.774	1.013	1.013	1.013	0.779	1.007	1.007	1.007	0.824
Age (years)
≤39	1.023	1.023	1.024	0.675	1.024	1.023	1.025	0.760	1.024	1.023	1.024	0.767	1.014	1.013	1.015	0.815
40–49	1.018	1.018	1.019	0.642	1.018	1.018	1.019	0.680	1.018	1.018	1.018	0.692	1.009	1.009	1.010	0.775
50–59	1.013	1.013	1.013	0.602	1.013	1.013	1.013	0.621	1.013	1.013	1.013	0.637	1.006	1.006	1.007	0.725
60–69	1.010	1.009	1.010	0.578	1.010	1.009	1.010	0.592	1.010	1.010	1.010	0.608	1.005	1.004	1.005	0.675
70+	1.007	1.007	1.008	0.560	1.008	1.007	1.008	0.579	1.008	1.007	1.008	0.589	1.005	1.004	1.005	0.626
BMI (kg/m²)
≤18.5	0.814	0.802	0.827	0.629	0.945	0.929	0.962	0.841	0.956	0.939	0.973	0.847	0.959	0.942	0.976	0.873
18.5–24.9	1.000	1.000	1.001	0.501	1.008	1.008	1.009	0.759	1.009	1.008	1.009	0.765	1.005	1.004	1.005	0.812
25–29.9	1.009	1.008	1.010	0.524	1.012	1.011	1.012	0.699	1.011	1.010	1.012	0.705	1.008	1.007	1.009	0.747
30+	1.132	1.119	1.145	0.553	1.071	1.058	1.085	0.683	1.070	1.056	1.083	0.687	1.054	1.041	1.068	0.715
Diabetes mellitus
Gender
Men	1.046	1.043	1.048	0.786	1.046	1.043	1.048	0.804	1.045	1.042	1.047	0.816	1.026	1.023	1.030	0.889
Women	1.019	1.018	1.019	0.629	1.022	1.021	1.022	0.757	1.021	1.021	1.022	0.769	1.012	1.012	1.013	0.860
Age (years)
≤39	1.049	1.046	1.051	0.796	1.050	1.047	1.052	0.825	1.048	1.046	1.051	0.834	1.032	1.029	1.034	0.894
40–49	1.030	1.029	1.031	0.714	1.030	1.029	1.031	0.749	1.029	1.028	1.030	0.768	1.016	1.015	1.017	0.877
50–59	1.020	1.019	1.021	0.655	1.020	1.020	1.021	0.670	1.020	1.020	1.021	0.696	1.010	1.009	1.011	0.840
60–69	1.014	1.013	1.015	0.611	1.014	1.013	1.015	0.619	1.014	1.014	1.015	0.646	1.007	1.006	1.008	0.798
70+	1.012	1.011	1.013	0.600	1.012	1.011	1.013	0.601	1.012	1.011	1.013	0.625	1.007	1.006	1.008	0.759
BMI (kg/m²)
≤18.5	0.741	0.703	0.780	0.675	0.849	0.801	0.900	0.854	0.861	0.813	0.913	0.870	0.868	0.819	0.921	0.897
18.5–24.9	1.004	1.003	1.004	0.520	1.012	1.012	1.013	0.732	1.013	1.012	1.014	0.745	1.007	1.006	1.008	0.850
25–29.9	1.018	1.016	1.019	0.551	1.021	1.019	1.023	0.697	1.020	1.018	1.022	0.713	1.013	1.011	1.015	0.833
30+	1.220	1.194	1.246	0.588	1.190	1.164	1.217	0.656	1.183	1.158	1.210	0.674	1.136	1.111	1.162	0.790

Discussion

The results showed that BMI was associated with the incidence of HTN and DM according to the SRD of obesity, which refers to the gender, age, and regional distribution of BMI. In the stratification analysis, the distribution of BMI was especially strongly associated with disease in the young and underweight groups. This means that the relative distribution of BMI was more informative in younger and underweight individuals. However, the absolute value of BMI was more relevant to disease than the relative distribution. Therefore, it is not useful to establish the diagnostic criteria for obesity using the percentile of BMI.

Some studies have presented descriptive analytics of BMI percentiles. Must et al.¹⁶ reported the 85th and 95th percentiles of BMI in the United States in 1991, and Flegal et al.¹⁷ reported the distribution of BMI in the United States in 1999–2010. Comparing these results, the 85th percentile in the U.S. population in 1991 was higher than the 90th percentile in our results, and the median BMI of the United States in 1999–2010 was higher than the median BMI in our results. This indicates that Koreans have a lower BMI than Americans. However, in the men in their 20s in this study, the results were similar to those of the United States in 1991, indicating that obesity among younger Koreans is a major problem. Compared with the results in 1990–1994 in Japanese adults,¹⁸ the BMI in the population analyzed in this study was higher in both men and women than in the Japanese population in 1990–1994. Considering the worldwide trend for BMI to increase, these findings can be regarded as reflecting the current trends in BMI.¹⁰

It is widely known that increases in BMI are related to HTN and DM, as established in various studies.^19,20 In this study, a decrease in BMI percentile in the underweight group was found to increase the incidence of HTN and DM. A previous study has noted that underweight was associated with the development of DM in the Japanese population.²¹ That study suggested that magnesium deficiency and malnutrition may cause a decline of insulin secretion. Salahudeen et al.²² found that underweight hemodialysis patients had a higher prevalence of HTN than their obese counterparts. Although the exact mechanism has not been elucidated, malnutrition and trace element deficiencies have been suggested as risk factors for the incidence of HTN.

The limitations of this analysis are as follows. First, a long follow-up was not possible, because the registration of SRD of obesity was carried out in 2017 using data from 2015 to 2016. Owing to the short follow-up period, this study did not observe severe disease and death. Second, there was a possibility of selection bias due to the characteristics of health screening program participants. Since health screenings were not mandatory, voluntary participants may have been more concerned with their health. However, since there was no co-payment (i.e., it was free for participants) and the participation rate was high (77.7%), this impact is expected to have been minimal. Third, there may be inaccuracies in the disease codes of claims data. Since claims data are used for the purpose of payment of medical expenses, there may be problems such as upcoding of disease codes. However, prescriptions of medications were additionally used to overcome this limitation. Fourth, information on smoking, alcohol drinking, and physical activity was obtained from a questionnaire, and the results may have been affected by the response tendencies of study participants.

Conclusion

The relative distribution of BMI was associated with the incidence of HTN and DM. Although the absolute value of BMI is used as a diagnostic criterion for obesity, the relative distribution can be used to motivate self-care through providing more detailed information to individuals.

Footnotes

Acknowledgment

This work was supported by the NHIS.

Author Disclosure Statement

No conflicting financial interests exist.

Supplementary Material

Supplementary Table S1

Supplementary Table S2

Supplementary Table S3

References

Meigs

, Wilson

, Fox

, et al. Body mass index, metabolic syndrome, and risk of type 2 diabetes or cardiovascular disease. J Clin Endocrinol Metab, 2006; 91:2906–2912.

Calle

, Thun

, Petrelli

, et al. Body-mass index and mortality in a prospective cohort of US adults. N Engl J Med, 1999; 341:1097–1105.

Jensen

, Ryan

, Apovian

, et al. 2013 AHA/ACC/TOS guideline for the management of overweight and obesity in adults: A report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines and The Obesity Society. J Am Coll Cardiol, 2014; 63(25 Part B):2985–3023.

Hill

, Wyatt

. Outpatient management of obesity: A primary care perspective. Obes Res, 2002; 10(S12):124S–130S.

Jung

, Chung

. Knowledge-based dietary nutrition recommendation for obese management. Inf Technol Manag, 2016; 17:29–42.

Cadihac

, Kilkenny

, Johnson

, et al. The Know Your Numbers (KYN) program 2008 to 2010: Impact on knowledge and health promotion behavior among participants. Int J Stroke, 2015; 10:110–116.

WHO Expert Consultation. Appropriate body-mass index for Asian populations and its implications for policy and intervention strategies. Lancet, 2004; 363:157–163.

Shin

. “Am I obese? National standard reference development for comparison of obesity status.” Yonhap New Agency. Available at: www.yonhapnews.co.kr/bulletin/2016/12/20/0200000000AKR20161220062200017.HTML?input=1179m (last accessed December 20, 2016 ).

National Health Insurance Service (NHIS). National Health Screening Statistical Yearbook. Seoul: National Health Insurance Service.

10.

NCD Risk Factor Collaboration. Trends in adult body-mass index in 200 countries from 1975 to 2014: A pooled analysis of 1,698 population-based measurement studies with 19.2 million participants. Lancet, 2016; 387:1377–1396.

11.

Okorodudu

, Jumean

, Montori

, et al. Diagnostic performance of body mass index to identify obesity as defined by body adiposity: A systematic review and meta-analysis. Int J Obes, 2010; 34:791.

12.

Lin

, Lee

, Chen

, et al. Optimal cut-off values for obesity: Using simple anthropometric indices to predict cardiovascular risk factors in Taiwan. Int J Obes, 2002; 26:1232.

13.

Penman

, Johnson

. The changing shape of the body mass index distribution curve in the population: Implications for public health policy to reduce prevalence of adult obesity. Prev Chron Dis, 2006; 3:A74.

14.

Seong

, Kim

, Khang

, et al. Data resource profile: The National Health Information Database of the National Health Insurance Service in South Korea. Int J Epidemiol, 2017; 46:799–800.

15.

Christensen

, Johansen

, Christiansen

, et al. Comparison of Charlson comorbidity index with SAPS and APACHE scores for prediction of mortality following intensive care. Clin Epidemiol, 2011; 3:203–211.

16.

Must

, Dallal

, Dietz

. Reference data for obesity: 85th and 95th percentiles of body mass index (wt/ht2) and triceps skinfold thickness. Am J Clin Nutr, 1991; 53:839–846.

17.

Flegal

, Carroll

, Kit

, et al. Prevalence of obesity and trends in the distribution of body mass index among US adults, 1999–2010. JAMA, 2012; 307:491–497.

18.

Yoshiike

, Matsumura

, Zaman

, et al. Descriptive epidemiology of body mass index in Japanese adults in a representative sample from the National Nutrition Survey 1990–1994. Int J Obes, 1998; 22:684.

19.

Brown

, Higgins

, Donato

, et al. Body mass index and the prevalence of hypertension and dyslipidemia. Obes Res, 2000; 8:605–619.

20.

Bays

, Chapman

, Grandy

, et al. The relationship of body mass index to diabetes mellitus, hypertension and dyslipidaemia: Comparison of data from two national surveys. Int J Clin Pract, 2007; 61:737–747.

21.

Sairenchi

, Iso

, Irie

, et al. Underweight as a predictor of diabetes in older adults: A large cohort study. Diabetes Care, 2008; 31:583–584.

22.

Salahudeen

, Fleischmann

, Bower

, et al. Underweight rather than overweight is associated with higher prevalence of hypertension: BP vs BMI in haemodialysis population. Nephrol Dial Transplant, 2004; 19:427–432.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.09 MB

0.13 MB

0.12 MB