The authors have declared that no competing interests exist.
This study examined a range of anthropometric indices and their relationships with metabolic syndrome (MetS). Despite recommendations that central obesity assessment should be employed as a marker of metabolic health, there is no consensus regarding the protocol for measurement. The present study included 720 men aged 71 ± 8 years and 919 women aged 71 ± 7 years from a rural village. We examined the relationship between anthropometric indices {e.g., body mass index (BMI), waist-to-height ratio (WHtR), waist-to-hip ratio (WHpR)}, and MetS based on the modified criteria of the National Cholesterol Education Program’s Adult Treatment Panel (NCEP-ATP) III report in a cross-sectional (N = 1,639) and cohort (N = 377) data. A receiver operating curve (ROC) analysis was performed to determine the optimal cut-off value and best discriminatory value of each of these anthropometric indices to predict MetS. In the cross-sectional study, WHtR as well as BMI and WHpR showed significantly predictive abilities for MetS in both genders; and WHtR showed the strongest predictive ability for the presence of MetS. Also in the cohort study, WHtR as well as BMI and WHpR showed significantly predictive abilities for incident MetS in both genders, and in men WHtR showed the strongest predictive ability for incident MetS, but in women BMI showed the strongest predictive ability. In the cross-sectional study, the optimal WHtR cutoff values were 0.52 (sensitivity, 71.0%; specificity, 77.9%) for men and 0.53 (sensitivity, 79.8%; specificity, 75.7%) for women. In the cohort study, the optimal WHtR values were 0.50 (sensitivity, 60.7%; specificity, 73.2%) for men and 0.50 (sensitivity, 75.0%; specificity, 56.1%) for women. Increased WHtR was significantly and independently associated with prevalence of MetS in both genders. These results suggest that WHtR is a useful screening tool for determining metabolic risk in Japanese elderly community dwelling individuals.
The underlying mechanism of metabolic syndrome (MetS), or a clustering of cardiovascular risk factors, such as hypertension, glucose intolerance, hypertriglyceridemia, and low high-density lipoprotein cholesterol (HDL-C) levels, is insulin resistance, which is also known as a pre-disease state that leads to an increased risk of cardiovascular disease (CVD) [
To address this hypothesis, we investigated the relationship between baseline visceral obesity indices and potential risk factors such as age, smoking status, drinking status, exercise habits, presence of CVD, low-density lipoprotein cholesterol (LDL-C), serum uric acid (SUA), estimated glomerular filtration ratio (eGFR), and incident MetS using prospective cohort data from community-dwelling elderly individuals.
The subjects of this study population were recruited from the Nomura Health and Welfare Center in a rural town in Ehime prefecture of Japan through annual health checkup process closely related to the area (17). This study was started in 2014, and included 1639 community-dwelling participants aged 55–95 years. Follow-up assessment cycles are performed every three years.
In the present study, we included data from the assessment cycles of 2014 and 2017. Blood samples were only obtained from respondents who participated in the medical interview at baseline. For the cross-sectional analyses, data of the 2014 cycle (n = 1639) were used as all five components of MetS were measured in this cycle. For the longitudinal analyses, a sub-cohort of the 2014 cycle was used including only participants in whom MetS was not prevalent at baseline in 2014 (n = 377).
For the cross-sectional analyses, data of the 2014 cycle (n = 1,639) that were used in this cycle were measured. For the longitudinal analyses, only participants in whom MetS was not prevalent at baseline in 2014 were included in the longitudinal analyses (n = 377).
This study complies with the Declaration of Helsinki, written informed consent was obtained from each subject, and the study was approved by the Ehime University Medical School Ethics Committee. All procedures performed in the study involving human participants were in accordance with the ethical standards of the institutional research committee in which the study was conducted. (IRB Approval number: 1402009).
Information on demographic characteristics and risk factors was collected using clinical files. Body mass index (BMI) was calculated by dividing weight (kilograms) by the square of height (meters). WHtR was calculated as WC (cm)/height (cm). WHpR was calculated as WC (cm)/hip circumference (cm). Other characteristics such as exercise, smoking habit, alcohol consumption, and medication, were investigated by individual interviews conducted using a structured questionnaire. Smoking habit was defined as the number of cigarette packs per day multiplied by the pack years (pack year), and participants were classified into never smokers, past smokers, light smokers (<30 pack year), and heavy smokers (≥30 pack year) [
For all these individuals, triglycerides (TG), HDL-C, low-density lipoprotein cholesterol (LDL-C), hemoglobin A1c (HbA1c), serum uric acid (SUA), and creatinine (Cr) were measured during an overnight fast of over 11 hours. eGFR was calculated using CKD-EPI equations modified by the coefficient of Japan (eGFRCKDEPI): Male, Cr ≤0.9 mg/dl, 141 × (
Based on the modified criteria of the National Cholesterol Education Program’s Adult Treatment Panel (NCEP-ATP) III report [
Unless otherwise specified, data are presented as the mean ± standard deviation (SD) and for parameters with non-normal distributions (i.e., TG, HbA1c) data are shown as median (interquartile range) values. For all analyses, parameters with non-normal distributions were used after log-transformation. Statistical analysis was performed using IBM SPSS Statistics Version 21 (Statistical Package for Social Science Japan, Inc., Tokyo, Japan). Subjects were divided into two groups based on gender and differences among the groups were analyzed by Student’s t-test for continuous variables or the χ2 -test for categorical variables. Multiple logistic linear regression analysis was used to evaluate the contribution of the baseline WHtR and confounding factors (i.e., gender, age, exercise habit, smoking habits, alcohol consumption, and prevalence of CVD, LDL-C, SUA, and eGFR) for prevalence of MetS in the cross-sectional study and incidence of MetS in the cohort study. In addition, areas under the receiver operating characteristic (ROC) curves were determined for each variable to identify the predictors of MetS. An ROC curve is a plot of sensitivity (true positive) versus 1–specificity (false positive) for each potential marker tested. Areas under the ROC curves are provided with standard errors. The area under the ROC curve is a summary of the overall diagnostic accuracy of the test. The best marker has an ROC curve shifted to the left with area under the curve close to unity. Predictive values were calculated as
Baseline characteristics of the subjects categorized by gender are illustrated in
Baseline Characteristics N = 1,639 | Men N = 720 | Women N = 919 | |
---|---|---|---|
Age (years) | 71 ± 8 | 71 ± 7 | 0.602 |
Body mass index (kg/m2) | 23.1 ± 2.9 | 22.6 ± 3.2 | |
Waist circumference (cm) | 82.4 ± 8.1 | 80.5 ± 9.0 | |
Waist/height ratio | 0.51 ± 0.05 | 0.54 ± 0.06 | |
Waist/hip ratio | 0.90 ± 0.06 | 0.89 ± 0.06 | |
Smoking habit (never/past/light/heavy (%)) | 41.7/40.6/4.3/13.5 | 96.8/2.1/0.7/0.4 | |
Drinking Status (never/occasional/light/heavy (%)) | 24.7/22.1/16.8/36.4 | 71.6/22.1/4.6/1.7 | |
Exercise habits (%) | 36.9 | 38.6 | 0.505 |
Cardiovascular disease (%) | 10.3 | 4.4 | |
Systolic blood pressure (mmHg) | 137 ± 17 | 137 ± 18 | 0.714 |
Diastolic blood pressure (mmHg) | 80 ± 10 | 77 ± 10 | |
Antihypertensive medication (%) | 48.1 | 45.3 | 0.272 |
Triglycerides (mg/dl) | 90 (68–131) | 87 (65–117) | |
HDL cholesterol (mg/dl) | 62 ± 16 | 68 ± 17 | |
LDL cholesterol (mg/dl) | 114 ± 28 | 124 ± 29 | |
Antidyslipidemic medication (%) | 14.0 | 30.0 | |
Hemoglobin A 1c (%) | 5.7 (5.4–6.0) | 5.7 (5.5–5.9) | |
Antidiabetic medication (%) | 13.6 | 5.5 | |
Serum uric acid (mg/dL) | 6.0 ± 1.3 | 4.7 ± 1.1 | |
Estimated GFR (ml/min/1.73 m2/year) | 69.4 ± 12.1 | 71.8 ± 10.8 | |
Number of metabolic syndrome component | 2.3 ± 1.1 | 2.6 ± 1.2 | |
Metabolic syndrome (%) | 37.8 | 51.1 |
HDL, high-density lipoprotein; LDL, low-density lipoprotein; GFR glomerular filtration ratio. Data presented are mean ± standard deviation. Data for triglycerides and HemoglobinA1c is skewed, and presented as median (interquartile range) values.
*
The univariate effect of WHtR on MetS and its components is presented in
Cross-sectional study N = 1,639 | Men N = 720 | Women N = 919 | ||||
---|---|---|---|---|---|---|
Yes/No | Odds ratio (95% CI) | Yes/No | Odds ratio (95% CI) | |||
Metabolic syndrome | 272/448 | 3.62 (2.95–4.43) | 470/449 | 3.46 (2.93–4.08) | ||
Central obesity | 254/466 | 15.8 (10.6–23.3) | 471/448 | 19.9 (13.4–29.5) | ||
Elevated blood pressure | 572/148 | 1.62 (1.34–1.97) | 716/203 | 1.66 (1.43–1.91) | ||
Elevated triglycerides | 128/592 | 1.62 (1.35–1.95) | 112/807 | 1.87 (1.50–2.33) | ||
Lowering HDL cholesterolemia | 139/581 | 1.62(1.35–1.95) | 349/570 | 1.54 (1.35–1.75) | ||
Elevated hemoglobin A 1c | 527/193 | 1.24 (1.05–1.47) | 704/215 | 1.17 (1.02–1.34) | ||
Metabolic syndrome | 28/149 | 1.99 (1.21–3.28) | 36/164 | 2.02 (1.34–3.06) | ||
Central obesity | 41/136 | 5.38 (2.96–9.77) | 54/146 | 5.51 (3.29–9.23) | ||
Elevated blood pressure | 125/52 | 1.18 (0.76–1.82) | 0.459 | 117/83 | 1.01 (0.74–1.39) | 0.956 |
Elevated triglycerides | 21/156 | 1.03 (0.57–1.86) | 0.930 | 13/187 | 1.03 (0.55–1.93) | 0.933 |
Lowering HDL cholesterolemia | 16/161 | 0.75 (0.36–1.57) | 0.450 | 31/169 | 1.50 (0.98–2.28) | 0.061 |
Elevated hemoglobin A 1c | 107/70 | 0.89 (0.60–1.31) | 0.551 | 129/71 | 0.82 (0.59–1.13) | 0.219 |
CI, confidence interval.
*Bold values indicate significance (
To further investigate whether WHtR can explain MetS and its components independently of other confounding factors, a multiple logistic regression analysis using MetS and its components as dependent variables and various confounding factors (e.g., age, smoking status, drinking status, exercise habits, presence of CVD, LDL-C, SUA, and eGFR) as explanatory variables was performed with subjects categorized by gender (
Cross-sectional study N = 1,639 | Men N = 720 | Women N = 919 | ||||
---|---|---|---|---|---|---|
Yes/No | Odds ratio (95% CI) | Yes/No | Odds ratio (95% CI) | |||
Metabolic syndrome | 272/448 | 3.82 (3.08–4.72) | 470/449 | 3.33 (2.81–3.96) | ||
Central obesity | 254/466 | 19.5 (12.6–30.1) | 471/448 | 31.0 (19.5–49.2) | ||
Elevated blood pressure | 572/148 | 1.57 (1.28–1.93) | 716/203 | 1.43 (1.23–1.68) | ||
Elevated triglycerides | 128/592 | 1.59 (1.30–1.95) | 112/807 | 1.75 (1.38–2.21) | ||
Lowering HDL cholesterolemia | 139/581 | 1.70 (1.40–2.07) | 349/570 | 1.48 (1.28–1.70) | ||
Elevated hemoglobin A1c | 527/193 | 1.24 (1.04–1.47) | 704/215 | 1.10 (0.95–1.28) | 0.200 | |
Metabolic syndrome | 28/149 | 1.94 (1.14–3.32) | 36/164 | 1.93 (1.23–3.03) | ||
Central obesity | 41/136 | 5.75 (3.05–10.8) | 54/146 | 8.28 (4.28–16.0) | ||
Elevated blood pressure | 125/52 | 1.07 (0.68–1.69) | 0.769 | 117/83 | 0.88 (0.62–1.24) | 0.467 |
Elevated triglycerides | 21/156 | 1.07 (0.55–1.90) | 0.960 | 13/187 | 1.09 (0.55–2.19) | 0.803 |
Lowering HDL cholesterolemia | 16/161 | 1.07 (0.48–2.37) | 0.869 | 31/169 | 1.27 (0.82–1.96) | 0.292 |
Elevated hemoglobin A 1c | 107/70 | 0.90 (0.59–1.40) | 0.648 | 129/71 | 0.83 (0.58–1.17) | 0.280 |
*Multivariate-adjusted for age, smoking status, drinking status, exercise habits, presence of cardiovascular disease, low-density lipoprotein cholesterol, serum uric acid, and estimated GFR. Bold values indicate significance (
In the cross-sectional study, the optimal WHtR cutoff values for predicting MetS according to WHtR were 0.52 (sensitivity, 71.0%; specificity, 77.9%) for men and 0.53 (sensitivity, 79.8%; specificity, 75.7%) for women (
AUC (95% CI) | Cut off value | Sensitivity | specificity | PPV | NPV | ||
---|---|---|---|---|---|---|---|
Men N = 720 | 0.812 (0.780–0.844) | 0.5185 | 71.0% | 77.9% | 76.3% | 72.9% | |
Women N = 919 | 0.832 (0.806–0.859) | 0.5349 | 79.8% | 75.7% | 76.7% | 78.9% | |
Men N = 177 | 0.698 (0.600–0.797) | 0.4991 | 60.7% | 73.2% | 69.4% | 63.1% | |
Women N = 200 | 0.681 (0.591–0.772) | 0.4957 | 75.0% | 56.1% | 62.2% | 69.2% |
AUR, Area under the receiver operating curve; PPV: positive predictive value; NPV: negative predictive value. Bold values indicate significance (
In this study where data from the Nomura study of 2014 and 2017 was used, the AUC analyses indicated that WHtR as well as BMI and WHpR had significant predictive ability for MetS in both genders, and that WHtR was significantly and independently associated with the prevalence of MetS in this cross-sectional study as well as the incidence of MetS in this cohort study. The usefulness of this cutoff value as a screening tool for the prediction of MetS was superior to those of BMI and WHpR, which are conventional obesity indices among both genders. This study showed that WHtR might be an appropriate definition from the point of view of knowing the presence and incidence of MetS. To the best of our knowledge, few epidemiologic studies have quantified the relevance between WHtR and incident MetS in Japanese elderly community-dwelling individuals.
This cross-sectional study showed that WHtR as well as BMI and WHpR was useful for predicting MetS, which is consist with Gu et al.’s research [
The mechanisms that lead to increased incidence of MetS in individuals with increased WHtR remain to be clarified. BMI is strongly related to body fat but is not necessarily related to abdominal obesity. WC may accurately reflect the degree of visceral fat, but WC can overestimate or underestimate the risk of CVD as WC does not take into account differences in height [
Several limitations should be considered in this study. First, our cross-sectional study design does not eliminate the cause and effect on conventional obesity indices and MetS. Second, the measurement of WHtR is based on a single evaluation of the equation, which may introduce a misclassification bias. Third, we could not eliminate the influence that taking medications for hypertension, dyslipidemia, and hyperglycemia has on the present findings. Fourth, as the WC component is included in the MetS definition, the AUC estimate for the WHtR can be disturbed. Fifth, the longitudinal analyses were limited by a smaller sample size and discrepancies in the sequential measurements of the components of MetS in 2014 and 2017. The cohort was slightly younger and healthier compared to participants not included in the longitudinal analyses, this might have caused an underestimation of incident MetS after three years of follow-up. Therefore the demographics and referral source may limit generalizability of the study findings.
The present study showed that anthropometric indices such as WHtR, BMI, and WHpR are strongly associated with incident MetS among Japanese community-dwelling individuals. The underlying mechanism behind this relationship is unknown, but it seems to be independent of confounding factors such as age, exercise habits, smoking habits, drinking status, prevalence of CVD, LDL-C, SUA or eGFR. Thus, WHtR might be an important marker for the assessment of risk and become a therapeutic target for MetS. For healthy community residents, prospective population-based studies are necessary to investigate mechanisms such as effective lifestyle improvement and other interventions to control WHtR in adults.
(ODS)