Cut-off values for visceral fat area (VFA) measured by computed tomography (CT) for identifying individuals at risk of metabolic syndrome (MetS) have not been clearly established in Korean adults, particularly for large populations. We aimed to identify optimal VFA and waist circumference (WC) cut-off values and compare the ability of VFA and WC to predict the presence of ≥2 metabolic risk factors.
We included 36,783 subjects aged 19–79 years undergoing abdominal fat CT during regular health checkups between January 2007 and February 2015 in Seoul. The risk factors for MetS except WC were based on the International Diabetes Federation criteria. Receiver operating characteristic curve analyses were used to determine the appropriate VFA and WC cut-off values for MetS.
VFA was a more significant predictor of metabolic risk factors than WC and body mass index (BMI). The optimal cut-off values for VFA and WC were 134.6 cm2 and 88 cm for men and 91.1 cm2 and 81 cm for women, respectively. We estimated age-specific cut-off values for VFA, WC, and BMI. VFA cut-off values increased with age, particularly among women.
This large population study proposed the cut-off values for VFA and WC for identifying subjects at risk of MetS among Korean adults. For more accurate diagnosis, different age-specific cut-off values for VFA and WC may be considered.
Metabolic syndrome (MetS) is defined as a cluster of metabolic abnormalities characterized by central obesity, hypertension, high triglyceride (TG) levels, low high-density lipoprotein cholesterol (HDL-C) levels, and high fasting plasma glucose (FPG) levels. Individuals with MetS have been known to have a higher risk of type 2 diabetes mellitus and cardiovascular disease, and a subsequent increase in disease-related mortality or morbidity [
In Korea, the incidence of MetS has increased markedly over the past decades, as in some Western countries [
WC is widely used to diagnose abdominal obesity because of its ease of measurement. However, it has been reported that insulin resistance is more strongly associated with VFA directly measured by computed tomography (CT) than with WC or body mass index (BMI) [
In this study, we investigated the appropriate sex- and age-specific cut-off values for VFA as reference values for predicting metabolic risk factors in Korean adults.
This cross-sectional study involved 39,181 subjects (23,728 men and 15,432 women) aged between 19 and 79 years who underwent health check-up tests, including abdominal fat CT, between January 2007 and February 2015 at Seoul National University Hospital Gangnam Center, Seoul, Republic of Korea. For participants undertaking health examinations on more than two occasions, we included only the result of the first examination. Our study excluded participants with a past medical history of any cancer (n=1,429), stroke (n=267), and cardiovascular event or stent insertion (n=702). Finally, 36,783 participants (22,365 men and 14,418 women) were enrolled in this study. Lifestyle factors, such as smoking, alcohol consumption, degree of exercise, marital status, and degree of education, were assessed using a self-report questionnaire. Alcohol consumption was calculated as the amount of drinking per week. Exercise was classified as regular exercise where subjects exercised ≥3 times per week. This study protocol was approved by the Institutional Review Board of Seoul National University Hospital (IRB approval no., H-1608-018-781).
Anthropometric measurements of individuals were performed by well-trained nurses. Height and weight were determined using a fixed wall-scale measuring device that measured to the nearest 0.1 cm or 0.1 kg, respectively, using an electronic scale that was calibrated before each measurement. BMI was calculated by dividing weight (kg) by squared height (m2). WC was measured at the end of normal expiration, measuring the minimum circumference at the level of the umbilicus to the nearest 0.5 cm. All subjects had their blood pressure (BP) checked twice after at least 5 minutes of rest in sitting position and were asked to answer questionnaires under the supervision of a welltrained interviewer. Blood samples were collected in the morning after an overnight fast. FPG, total cholesterol, low-density lipoprotein cholesterol, HDL-C, and TG levels were measured enzymatically using the Hitachi 747 chemical analyzer (Hitachi, Tokyo, Japan).
The adipose tissue areas were measured at the level of the umbilicus using a 16-detector row CT scanner (Somatom Sensation 16; Siemens Medical Solutions, Forchheim, Germany) with the subjects in a supine position. A 5-mm-thick umbilical-level abdominal section was obtained using a 16-detector row CT scanner. The cross-sectional area (cm2) of abdominal fat was calculated using Rapidia 2.8 CT software (INFINITT, Seoul, Korea) with attenuation values for the region of interest within a range of −250 to −50 Hounsfield units. The technique used for adipose tissue area measurements with CT cross-sectional images has been previously standardized and validated. The visceral adipose tissue area was defined as intra-peritoneal fat bound by parietal peritoneum or transversalis fascia, and the subcutaneous adipose tissue area was defined as the fat area external to the abdomen and back muscles [
In this study, we used the following criteria for the definition of MetS proposed by the IDF, excluding WC as a risk factor [
Statistical analyses were performed separately for men and women because of known differences in abdominal fat distribution between sexes. Continuous variables were expressed as means and standard deviation, and discrete variables were expressed as proportions. Differences between sexes were determined by the Student t-test. A receiver operating characteristic (ROC) curve analysis was used to develop a cut-off for each anthropometric value associated with the presence of two or more risk factors of MetS, except WC, and to compare the predictive value of VFA, WC, and BMI. We performed subgroup analysis to obtain the age-specific cut-off values. The cut-off values for VFA, WC, and BMI that maximized the Youden index (sensitivity+specificity-1) were defined as optimal. The Youden index is an integrative indicator of sensitivity and specificity [
A total of 36,783 participants (22,365 men and 14,418 women) were included in the analysis. The characteristics of our study population are presented in
The optimal cut-off values for VFA, WC, and BMI to identify ≥2 risk factors for MetS, except WC, were 134.6 cm2, 88.4 cm, and 24.7 kg/m2 and for men and 91.1 cm2, 81.0 cm, and 22.4 kg/m2 for women, respectively (
Our results demonstrated that the optimal VFA cut-off values (134.6 cm2 for men and 91.1 cm2 for women) yielded maximum sensitivity and specificity. Considering our findings, the WC cut-off values were comparable to the IDF criteria for subjects of South Asian ethnicity [
To date, few studies have reported Korean-specific VFA cut-off values. Han et al.[
Kim et al. [
Comparable results have also been reported in Japan, another Northeast Asian country. Since the Japan Society for the study of Obesity proposed a VFA cut-off value of 100 cm2 in 2002, it has been used as the gold standard for the definition of central obesity among Japanese adults [
In recent years, studies have been conducted in older populations. Seo et al. [
Because previous studies have been performed in small populations and studies including the elderly have reported different cut-off values, we conducted an additional analysis to obtain cut-off values depending on age (
In our study, the AUC for women was larger than that for men across all age groups. This finding can be explained by the fact that men tend to have other MetS risk factors, which are not used as diagnostic criteria for MetS, such as smoking, alcohol consumption, and stress. Therefore, the contribution of VFA to the multiple risk factors of MetS may be relatively smaller in men than in women [
The present study has some limitations. First, this study involved subjects regularly undergoing health check-ups, who are relatively more concerned with health and are likely to be managing their health well on their own. Further, the study subjects had a relatively high socioeconomic status and most of them lived in an urban area. Thus, there are limitations regarding the representativeness and generalizability of the research findings. Second, due to the cross-sectional nature of our study, our analysis could not reflect changes over time. Further studies are needed to prospectively examine the association between the accumulation of visceral fat and the presence of risk factors. Third, subjects who were diagnosed and treated with hypertension and diabetes were identified using a self-report questionnaire. Thus, it was difficult to confirm the presence of medical conditions as these were largely self-reported. To reduce such disadvantages, we reviewed the questionnaires for the use of medicines and checked the drug names provided by subjects in the questionnaires. Fourth, thyroid hormone abnormalities and menstrual status that could affect the study result were not included owing to the limitation of a self-report questionnaire.
To the best of our knowledge, this is the largest study to propose VFA cut-off values for the Korean population. We suggest that VFA values higher than the cut-off value of 134.6 cm2 for men and 91.1 cm2 for women are risk factors for predicting MetS that could result in cardiovascular disease or type 2 diabetes mellitus. Future prospective studies are needed to validate the association between the optimal WC/VFA cut-off values and the incidence of or mortality resulting from cardiovascular diseases.
No potential conflict of interest relevant to this article was reported.
This work was supported by the Technology Innovation Program (10050154, Business Model Development for Personalized Medicine Based on Integrated Genome and Clinical Information) funded by the Ministry of Trade, Industry and Energy (MI, Korea).
Age-adjusted ROC curves to identify ≥2 metabolic risk factors other than WC. In both men (A) and women (B), the area under the ROC curve values for VFA were shown to be higher values than those for WC and BMI. ROC, receiver operating characteristic; VFA, visceral fat area; WC, waist circumference; BMI, body mass index.
Baseline characteristics of the study population
Characteristic | Total (n=36,783) |
|
---|---|---|
Men (n=22,365) | Women (n=14,418) | |
Age (y) | 50.8±9.5 | 51.7±9.5 |
Body mass index (kg/m2) | 24.6±2.8 | 22.3±2.9 |
WC (cm) | 88.0±8.0 | 80.6±8.0 |
Total fat area (cm2) | 277.2±94.7 | 259.9±97.3 |
Visceral fat area (cm2) | 136.0±53.1 | 83.0±42.7 |
Systolic BP (mm Hg) | 119.5±13.1 | 113.4±14.9 |
Diastolic BP (mm Hg) | 78.6±10.3 | 71.3±10.7 |
Total cholesterol (mg/dL) | 195.4±34.6 | 198.5±35.3 |
TG (mg/dL) | 129.5±89.9 | 87.4±54.7 |
HDL-C (mg/dL) | 49.3±13.0 | 58.4±13.0 |
Fasting glucose (mg/dL) | 100.3±21.2 | 91.8±15.8 |
Alcohol consumption (g/wk) |
222.0±190.1 | 127.2±151.1 |
Current smoker (%) |
36.1 | 4.4 |
Regular exercise (≥3 times/wk) |
28.5 | 27.1 |
Married (%) |
93.7 | 87.0 |
Educational status |
||
Middle school or less (%) | 4.0 | 9.6 |
High school (%) | 16.6 | 24.5 |
College or more (%) | 83.5 | 65.9 |
Features of metabolic syndrome except WC | ||
BP ≥130/85 mm Hg or Med | 9,462 (42.3) | 3,661 (25.4) |
Fasting blood sugar ≥100 mg/dL or Med | 8,641 (38.6) | 2,727 (18.9) |
TG ≥150 mg/dL | 6,210 (27.8) | 1,321 (9.2) |
Low HDL-C (<40 for men, <50 for women) | 4,560 (20.4) | 4,211 (29.2) |
≥2 metabolic risk factors of the IDF criteria, except WC | 8,868 (39.7) | 3,090 (21.4) |
Values are presented as mean±standard deviation, %, or number (%). Gender differences in all variables were statistically significant (P<0.001).
WC, waist circumference; BP, blood pressure; TG, triglyceride; HDL-C, high-density lipoprotein cholesterol; Med, medication; IDF, International Diabetes Federation.
Missing values are included in this data.
AUC for VFA, WC, and BMI to identify the presence of metabolic risk factors other than WC
Variable | Men (n=22,365) | Women (n=14,418) |
---|---|---|
≥2 metabolic risk factors of the IDF criteria other than WC | ||
VFA (cm2) | 0.703 (0.696–0.710) | 0.801 (0.793–0.809) |
WC (cm) | 0.682 (0.675–0.689) | 0.779 (0.770–0.784) |
BMI | 0.683 (0.676–0.690) | 0.770 (0.761–0.779) |
Fasting blood sugar ≥100 mg/dL or diabetes mellitus medications | ||
VFA (cm2) | 0.654 (0.641–0.668) | 0.693 (0.672–0.714) |
WC (cm) | 0.638 (0.625–0.651) | 0.659 (0.638–0.681) |
BMI | 0.640 (0.627–0.653) | 0.657 (0.635–0.679) |
High blood pressure or hypertension medications | ||
VFA (cm2) | 0.660 (0.649–0.672) | 0.711 (0.694–0.728) |
WC (cm) | 0.643 (0.631–0.655) | 0.650 (0.632–0.668) |
BMI | 0.637 (0.626–0.649) | 0.654 (0.636–0.672) |
Hypertriglycemia | ||
VFA (cm2) | 0.628 (0.609–0.646) | 0.671 (0.628–0.713) |
WC (cm) | 0.615 (0.596–0.634) | 0.655 (0.612–0.697) |
BMI | 0.609 (0.590–0.628) | 0.652 (0.609–0.695) |
Low HDL-C | ||
VFA (cm2) | 0.672 (0.653–0.690) | 0.732 (0.717–0.747) |
WC (cm) | 0.641 (0.622–0.660) | 0.681 (0.665–0.697) |
BMI | 0.645 (0.626–0.663) | 0.680 (0.664–0.696) |
Values are presented as AUC (95% confidence interval).
AUC, areas under the receiver operating characteristic curve; VFA, visceral fat area; WC, waist circumference; BMI, body mass index; IDF, International Diabetes Federation; HDL-C, high density lipoprotein cholesterol.
Comparison of the cut-off values for VFA, WC and BMI for identifying metabolic risk factors in the present and previous studies
Variable | Cut-off | Sensitivity (%) | Specificity (%) | PPV (%) | NPV (%) |
---|---|---|---|---|---|
Men | |||||
VFA (cm2) | |||||
Cut-off in present study |
134.6 | 67.3 | 62.3 | 63.5 | 66.1 |
Han et al. [ |
100.0 | 67.1 | 51.9 | 80.3 | 35.0 |
Kim et al. [ |
136.0 | 74.3 | 78.3 | 72.3 | 80.0 |
Seo et al. [ |
92.6 | 70.4 | 70.5 | 79.7 | 59.1 |
Lim S et al. [ |
140.0 | 49.1 | 74.1 | 46.6 | 76.0 |
JIM [ |
100.0 | 83.4 | 34.8 | 78.9 | 41.7 |
Oka et al. [ |
132.6 | 63.3 | 64.3 | 50.0 | 75.6 |
WC (cm) | |||||
Cut-off in present study |
88.4 | 61.3 | 64.7 | 59.7 | 66.2 |
KOSSO’s cut value [ |
90.0 | 54.2 | 81.9 | 62.0 | 76.6 |
IDF criteria for South Asians [ |
90.0 | 44.6 | 76.1 | 50.4 | 71.6 |
BMI (kg/m2) | |||||
Cut-off in present study |
24.7 | 60.8 | 62.9 | 57.3 | 66.2 |
WHO Asia Pacific criteria for obesity21) | 25.0 | 45.8 | 84.8 | 67.4 | 69.5 |
Women | |||||
VFA (cm2) | |||||
Cut-off in present study |
91.1 | 71.4 | 72.2 | 60.2 | 81.1 |
Han et al. [ |
70.0 | 80.3 | 55.2 | 69.2 | 69.1 |
Kim et al. [ |
95.0 | 79.1 | 77.4 | 64.5 | 87.7 |
Seo et al. [ |
88.9 | 72.5 | 59.4 | 53.3 | 77.3 |
Lim S et al. [ |
100.0 | 65.2 | 56.5 | 39.7 | 78.7 |
JIM [ |
100.0 | 519.0 | 81.4 | 55.1 | 79.4 |
Oka et al. [ |
91.5 | 67.0 | 74.4 | 60.4 | 79.4 |
WC (cm) | |||||
Cut-off in present study |
81.0 | 74.1 | 61.1 | 59.6 | 75.3 |
KOSSO’s cut value [ |
85.0 | 76.4 | 51.5 | 35.2 | 86.4 |
IDF criteria for South Asians [ |
80.0 | 82.1 | 42.5 | 57.2 | 71.7 |
BMI (kg/m2) | |||||
Cut-off in present study |
22.4 | 71.4 | 63.3 | 59.2 | 74.8 |
WHO Asia Pacific criteria for obesity [ |
25.0 | 54.6 | 64.4 | 34.6 | 80.5 |
VFA, visceral fat area; WC, waist circumference; BMI, body mass index; PPV, positive predictive value; NPV, negative predictive value; JIM, Japanese Society of Internal Medicine; KOSSO, Korean Society for the Study of Obesity; IDF, International Diabetes Federation; WHO, World Health Organization.
The optimal cut value was obtained from Youden index as {maximum (sensitivity+specificity-1)}.
Age-specific cut-off values of VFA, WC, and BMI for predicting two or more risk factors for metabolic syndrome
Age (y) | VFA |
WC |
BMI |
|||
---|---|---|---|---|---|---|
Cut-off (cm2) | AUC | Cut-off (cm) | AUC | Cut-off (kg/m2) | AUC | |
Men (n=22,365) | ||||||
19–39 (n=2,652) | 120.5 | 0.760 | 89.3 | 0.751 | 24.7 | 0.742 |
40–59 (n=15,645) | 134.4 | 0.699 | 88.5 | 0.673 | 25.0 | 0.671 |
60–79 (n=4,068) | 131.7 | 0.656 | 87.8 | 0.633 | 24.3 | 0.627 |
Women (n=14,418) | ||||||
19–39 (n=1,428) | 62.6 | 0.838 | 77.2 | 0.816 | 22.4 | 0.821 |
40–59 (n=10,093) | 85.9 | 0.775 | 80.8 | 0.717 | 23.1 | 0.729 |
60–79 (n=2,897) | 115.6 | 0.679 | 82.5 | 0.630 | 24.0 | 0.635 |
VFA, visceral abdominal fat area; WC, waist circumference; BMI, body mass index; AUC, area under the receiver operating characteristic curve.
Comparison of the mean age, BMI, WC, and VFA, and cut-off values of VFA in the present and previous study subjects
Variable | Age (y) | BMI (kg/m2) | WC (cm) | VFA (cm2) | VFA cut-off (cm2) |
---|---|---|---|---|---|
Men | |||||
The present study (n=22,365) | 50.8±9.5 | 24.6±2.8 | 88.0±8.0 | 136.0±8.0 | 134.6 |
Han et al. [ |
45.5±9.4 | 25.4±2.5 | 89.6±7.5 | 114.8±47 | 100.0 |
Kim et al. [ |
54.5±8.7 | 24.6±2.9 | 88.1±7.7 | 129±59 | 136.0 |
Seo et al. [ |
71.2±5.2 | 24.0±2.8 | 88.0±7.6 | 103.6±49.1 | 92.6 |
Lim et al. [ |
77.0±9.0 | 23.6±3.3 | 88.4±8.3 | 121.3±63.0 | 140.0 |
Oka et al. [ |
50.6±7.8 | 24.9±2.7 | 86.9±7.1 | 128 (97, 159) | 132.6 |
Women | |||||
The present study (n=14,418) | 51.7±9.5 | 22.3±2.9 | 80.6±8.0 | 83.0±42.7 | 91.1 |
Han et al. [ |
40.7±12.2 | 25.4±2.7 | 85.1±7.5 | 80.5±39.7 | 70.0 |
Kim et al. [ |
53.0±6.7 | 24.2±3.2 | 81.9±8.1 | 91±43 | 95.0 |
Seo et al. [ |
70.5±4.7 | 25.0±3.1 | 88.5±8.3 | 108.9±45.4 | 88.9 |
Lim et al. [ |
77.2±9.1 | 23.8±3.5 | 86.3±9.2 | 109.3±48.4 | 100.0 |
Oka et al. [ |
51.2±7.2 | 23.5±3.1 | 81.9±8.2 | 69 (48, 98) | 91.5 |
Values are presented as mean±standard deviation and median (interquartile range).
BMI, body weight index; WC, waist circumference; VFA, visceral abdominal fat area.