The prevalence of obesity among children and adolescents is increasing in many countries, accompanied by an increase in associated complications.^1,2 This phenomenon has also been observed in Korea.^3-5 Body mass index (BMI) has been traditionally used to diagnose obesity,⁶ with reference values for each population group.^7-9 However, BMI has the drawback of not adequately reflecting the proportions of body fat and muscle mass because it is calculated by simply dividing body weight by the square of height.¹⁰

Several studies have recently identified pediatric abdominal obesity as a strong predictor of cardiometabolic risk factors and cardiovascular disease in adulthood.^11,12 Waist circumference (WC) is the most commonly used method of evaluating abdominal obesity because of its simplicity. For diagnosis of metabolic syndrome (MS) in children and adolescents, as well as in adults, WC rather than BMI is used as a diagnostic criterion.^13-15 However, WC is positively associated with height, limiting its accuracy as an indicator of abdominal obesity. Alternatively, the waist–height ratio (WHtR) has been used in recent years because it is consistent according to age and sex and is a good indicator for assessing abdominal obesity.

Because children and adolescents continue to grow physically, the normal values of WC and WHtR differ according to sex and age, necessitating sex- and age-specific reference values.^16,17 Furthermore, children and adolescents with abdominal obesity are more likely to be classified as normal if the reference values are derived from data in which the prevalence of obesity is elevated. The currently used WC reference, which was published in the 2007 Korean National Growth Charts, has several drawbacks; specifically, its development did not adopt standardized measurements of WC, only included several percentile values, presented no graphical charts, and did not exclude children and adolescents with severe obesity.^18,19

Therefore, a new WC reference needs to be established that adequately considers the increase in obesity using large-scale standardized anthropometric measurement data, and WHtR reference values also need to be developed. The purposes of this study were to present these values in the form of tables and charts and to compare the new reference values to BMI for predicting cardiometabolic risk factors using nationally representative survey data.

Study population

Data were obtained from the Korea National Health and Nutrition Examination Survey (KNHANES) conducted between 2007 and 2019. The KNHANES is an ongoing cross-sectional survey that collects nationally representative data. Although it has been conducted since 1998 by the Korean Centers for Disease Control and Prevention, data collected between 2007–2019 were included in the present study because the survey design and detailed methods were updated in 2007. The sampling of KNHANES data among non-institutionalized citizens in Korea employs a stratified, multistage, clustered probability method. Each KNHANES consists of a health interview, an examination, and a nutrition survey. The data collection procedures of KNHANES have been described in detail elsewhere.²⁰ Among the 113,724 individuals enrolled in the KNHANES during 2007−2019, 22,495 (11,613 boys and 10,882 girls) aged ≤20 years with anthropometric data, including height, weight, and WC, were considered as candidates for WC and WHtR reference development in this study. Among them, 20,033 participants aged 2−18 years were finally included in analysis of reference curves and tables. For the reference evaluation, participants aged 10−18 years with blood pressure (BP) or laboratory data (12,158 individuals) were included (Supplementary Fig. 1).

Anthropometric measurements

Height, weight, and WC were measured by trained medical personnel. Height was measured to the nearest 0.1 cm using a stadiometer (Seca 225; Seca, Hamburg, Germany). Weight was measured to the nearest 0.1 kg using an electronic balance (GL-6000-20; Gtech, Seoul, Korea). BMI was calculated as weight (kg) divided by height squared (m²). Height, weight, and BMI values were transformed into z-scores using the 2017 Korean National Growth Chart.⁹ WC was measured using a flexible tape (Seca 220) to the nearest 0.1 cm at the midpoint between the lowest margin of the rib and the highest margin of the iliac crest during expiration. WHtR was calculated as WC (cm) divided by height (cm). BP was measured three times on the right arm using a mercury sphygmomanometer with a cuff appropriate for arm circumference after the participant had rested for ≥5 minutes in a sitting position (Baumanometer Desk model 0320 in 2007–2012 and Baumanometer Wall Unit 33[0850] in 2013– 2019; W.A. Baum, Copiague, NY, USA). The mean values of the second and third systolic and diastolic BP measurements were analyzed in this study.

Laboratory tests

Blood samples were collected by trained medical personnel. Fasting blood samples obtained from venipuncture were transported to a central laboratory for analysis within 24 hours. Plasma glucose, total cholesterol (TC), high-density lipoprotein cholesterol (HDL-C), low-density lipoprotein cholesterol (LDL-C), and triglyceride levels were measured using a Hitachi Automatic Analyzer 7600 (Hitachi, Tokyo, Japan). The glycated hemoglobin (HbA1c) level was measured via high-performance liquid chromatography (HPLC-723G7; Tosoh, Tokyo, Japan), which is the method certified by the National Glycohemoglobin Standardization Program.

Definitions of abnormal cardiometabolic risk factors

High BP was defined as systolic or diastolic BP ≥90th percentile for corresponding sex, age, and height. High fasting glucose and elevated HbA1c were defined as fasting glucose ≥100 mg/dL and HbA1c ≥5.7%, respectively. Lipid abnormalities were defined as follows: triglycerides ≥150 mg/dL, LDL-C ≥130 mg/dL, non– HDL-C ≥145 mg/dL (obtained by subtracting HDL-C from TC), and HDL-C <40 mg/dL.

Generation of reference curves and LMS table

To derive reference values for WC and WHtR using data from the KNHANES 2007−2019, growth charts for WC and WHtR were generated through a method similar to that applied for development of the 2017 Korean National Growth Charts.⁹ Among KNHANES participants 1–20 years of age, those with weight for height <−3 standard deviation (SD) or >+2 SD were excluded. These cutoff values for subject selection were used in previous Korean Growth Charts and are considered appropriate for preventing skewness.⁹ Using the selected data, reference values for WC and WHtR were generated. The empirically used percentile values of WC and WHtR, including the 3rd, 5th, 10th, 25th, 50th, 75th, 90th, 95th, and 97th percentiles, were calculated by each age and sex. Locally estimated scatterplot smoothing (LOESS) regression was performed to smooth each percentile curve in Stata 16.1 (StataCorp., College Station, TX, USA); “LOESS” defines a process of locally weighted regression for estimating smoothed curves from scattered plots.²¹

The z-scores and percentile values for WC and WHtR were calculated using Box-Cox transformation and the LMS method, where L, M, and S stand for the power of Box-Cox transformation, the median, and the coefficient of variation, respectively.^22,23 A non-linear least-squares analysis was used for estimation of the L, M, and S values. Tables containing L, M, and S values and each percentile value of WC and WHtR according to sex and age were generated. The z-score was calculated using L, M, and S values as z=[(x/M(t)) L(t)−1]/[L(t)S(t)], where x stands for the WC or WHtR value and t stands for age at measurement.

WC and WHtR curves for children and adolescents aged 1–20 years were produced, but only the curves for those aged 2–18 years were presented to avoid an edge effect at both age extremes.²⁴ Each curve demonstrated the 3rd, 5th, 10th, 25th, 50th, 75th, 90th, 95th, and 97th percentiles of WC and WHtR by sex.

Statistical analysis

All statistical analyses were performed using Stata 16.1. Simple linear regression analysis was applied to evaluate the correlation between WC z-scores and WHtR z-scores. Logistic regression analysis was performed to evaluate the associations between cardiometabolic risk factors and the z-scores of WC, WHtR, and BMI, which were presented as odds ratios (ORs) and 95% confidence intervals (CIs) after adjusting for age and sex. The P-value <0.05 was considered statistically significant.

Ethics statement

The present study protocol was reviewed and approved by the Institutional Review Board of Seoul National University Bundang Hospital (No. X-1909-565-904). All subjects submitted informed consent upon enrollment.

New references for WC and WHtR

The newly developed curves for WC and WHtR for Korean children and adolescents are presented in Figs. 1 and 2. In Tables 1 and 2, the LMS values for z-score calculation and the 3rd, 5th, 10th, 25th, 50th, 75th, 90th, 95th, and 97th percentiles by sex and age are demonstrated. Compared with the WC curves presented in the 2007 growth chart, every line in the new chart was lower than the corresponding line (Fig. 3). The WHtR curves showed sex-specific differences, although they demonstrated a relative plateau among those aged ≥10 years regardless of sex.

Correlation between WC and WHtR

Using the newly developed charts for WC and WHtR, the zscores of WC and WHtR were significantly correlated (r²=0.846, P<0.001 for boys; r²=0.834, P<0.001 for girls) (Supplementary Fig. 2).

ORs for predicting cardiometabolic risk factors by z-scores of WC, WHtR, and BMI

In logistic regression analysis, the WC z-score and WHtR z-score using the newly developed charts revealed higher adjusted ORs for all cardiometabolic risk factors than the BMI z-score (Table 3).

In the present study, new reference values for WC and WHtR in Korean children and adolescents were developed using up-to-date statistical methodology. Age- and sex-specific tables and charts of WC and WHtR were derived. In particular, the LMS values for WC and WHtR were added to calculate the z-score or percentile for the corresponding sex and age, which can help to diagnose abdominal obesity and determine cardiometabolic risk in Korean children and adolescents.

Abdominal obesity is associated with abundant visceral fat, which is an important metabolic factor with a close link to cardiovascular disease. Additionally, abdominal obesity is one of the diagnostic criteria for MS in both children and adults.^13-15 WC, which varies according to population, is the most widely used method for diagnosing abdominal obesity. For adults, WC ≥90 cm for men and ≥85 cm for women is the diagnostic criterion for abdominal obesity in Korea.²⁵ In children and adolescents, WC ≥90th percentile for sex and age has been defined as abdominal obesity.^13,15,17 In Korea, the 2007 Korean National Growth Charts have been used as a reference for WC.¹⁸ However, updates to these charts to address their limitations, including lack of standardized measurements for WC, inclusion of few percentile values, absence of graphical charts, and non-exclusion of children and adolescents of extreme weight, have been overdue.¹⁹

Compared with the WC curves presented in the 2007 growth charts, every line in the newly developed chart was lower than the corresponding line (Fig. 3). This result mainly originates from inclusion of extreme values in the 2007 reference. Abdominal obesity might be underestimated if the reference values are derived using raw data containing extreme values. To overcome the drawbacks of previous reference charts, the newly developed WC reference adopted up-to-date statistical methodology, such as the removal of subjects with extreme weight, smoothing of percentile lines, and the LMS estimation method, which was applied in the 2017 Korean National Growth Charts.⁹ These methods have been used to develop growth charts in other countries.^8,26

In a recent study, the prevalence of abdominal obesity in Korean children and adolescents aged 10–18 years was estimated as 9.66% using the 2007 reference values.²⁷ However, this might be an underestimation. Moreover, further research is required because underestimating abdominal obesity might lead to an underestimation of MS, since abdominal obesity is one of the major diagnostic criteria for MS. WHtR, which is calculated as WC divided by height, has been used for diagnosis of abdominal obesity.²⁸ WHtR could be useful for defining abdominal obesity because it is easy to calculate and has a simple cutoff value of ≥0.5.²⁹

However, there are several concerns regarding the use of WHtR of 0.5 as a criterion for abdominal obesity in children and adolescents.³⁰ According to a recent study of Korean adolescents 13−18 years of age, the cutoff value of WHtR for diagnosing abdominal obesity was 0.48, which is lower than 0.5.¹⁹ For Japanese children aged 9−11 years, WHtR values of 0.519 for boys and 0.499 for girls were determined to be the most appropriate cutoffs for diagnosing abdominal obesity.³¹ International studies have also shown that the WHtR cutoff of 0.5, which is mainly used in adults, is not applicable to children or adolescents, who are still growing, and that it would be better to use WHtR percentiles that are specific for age and sex.^16,32 As shown in the newly created graph in this study, it is difficult to apply this criterion because of the high WHtR values in young ages, especially those younger than 6 years. Therefore, children and adolescents with WHtR ≥0.5 are highly likely to have abdominal obesity, so further evaluation may be necessary. However, in children and adolescents, it seems more appropriate to evaluate WHtR using percentiles rather than a single value.

A recent study using KNHANES data to analyze WHtR in children and adolescents aged 10−18 years showed LMS data and percentile values, but it was limited by the lack of data from subjects under the age of 10, that the analysis was conducted without removing extreme values, and the possibility of WHtR variability at both ends of the age spectrum.³³ In the present study, reference data and charts were presented to compensate for those shortcomings.

WC is the most widely used indicator of abdominal obesity; however, it has a positive correlation with height. Therefore, it may be more reasonable to define abdominal obesity using WHtR than WC, but further research is needed to determine the optimal cutoff point. WC and WHtR are important abdominal obesity-related indicators associated with cardiometabolic risk factors and MS.¹² It is not clear which of these indicators is superior, although a study has shown that WHtR reflects body fat better than WC in children and adolescents.³⁴ WHtR may be widely used as a screening risk factor for cardiometabolic risk due to its simplicity of calculation, even if it is not superior to WC.¹¹ In this study, WC and WHtR had a similar ability to predict cardiometabolic risk factors (Table 3).

There were several limitations of the present study. First, the charts were derived from cross-sectional data. Second, the WC and WHtR values from the KNHANES data were presented according to ages in years, not months, limiting the degree to which the reference values could be precisely applied for more narrowly defined age groups.

In conclusion, WC and WHtR reference values were developed using up-to-date methodology and the latest nationally representative anthropometric data. Sex-specific tables and charts for children and adolescents aged 2–18 years were presented with LMS values for calculating z-scores or percentiles. The newly developed WC and WHtR reference values are expected to be widely used in clinical practice to monitor and track abdominal obesity, cardiometabolic risk, and MS among at-risk children and adolescents in Korea.

The authors declare no conflict of interest.

This research was supported by grant no. 09-2019-0008 from the Seoul National University Bundang Hospital Research Fund.

Study concept and design: JK; acquisition of data: JL and JK; analysis and interpretation of data: SCK and SSH; drafting of the manuscript: JL, OK, and JK; critical revision of the manuscript; JL, SCK, OK, SSH, and JSM; statistical analysis: SCCK and SSH; funding acquisition: JK; administrative, technical, or material support: JK; and study supervision: JK and JSM.

Supplementary materials can be found online at https://doi.org/10.7570/jomes22033.