This study was based on data from the Korea National Health and Nutrition Examination Survey (KNHANES) 2011–2016. This cross-sectional nationwide survey, conducted by the Korea Centers for Disease Control and Prevention (KCDC), used a multistage, stratified, and clustered probability sampling method to select a representative sample of the Korean civilian population. Of the total 47,164 participants, children and adolescents aged between 10 and 18 years were selected (n=5,043). Exclusion criteria were as follows: subjects who fasted less than 8 hours before blood sampling (n=553); absence of results of fasting glucose or elevated fasting glucose values of ≥126 mg/dL (n=1,050); absence of results of HbA1_c or elevated HbA1_c of ≥6.5% (n=1,074); previously diagnosed DM (n=3); pregnancy (n=0); absence of any component of MS (n=1,105), and absence of anthropometric measurements (n=382). Finally, 3,885 participants (2,074 young males and 1,811 young females) were eligible for the analysis (Fig. 1).

Anthropometric measurements were performed on all participants by trained personnel. Height was measured to the nearest 0.1 cm by using a stadiometer (Seca 225, Seca, Hamburg, Germany). Weight was measured by using an electronic balance (GL-600020, G-tech, Seoul, Korea) to the nearest 0.1 kg, and body mass index (BMI) was calculated as weight (kg) divided by height squared (m²). Height and BMI were transformed into z-scores by using the 2017 Korean National Growth Charts [22]. Waist circumference (WC) was measured to the nearest 0.1 cm by using a flexible tape measure, from the midpoint between the lowest border of the rib cage and the uppermost border of the iliac crest at the end of expiration (Seca 220, Seca). Blood pressure (BP) was measured by using a mercury sphygmomanometer (Baumanometer Desk model 0320 in 2011–2012 and Baumanometer Wall Unit 33[0850] in 2013–2016, Baum, NY, USA) on the right arm after the subject had rested for at least 5 minutes in a seated position. BP was measured 3 times, and the mean of the last 2 values were used for the analyses.

Fasting blood samples were collected from each participant by trained medical personnel. Plasma glucose, total cholesterol, high-density lipoprotein cholesterol (HDL-C), triglyceride (TG), aspartate aminotransferase, and ALT were measured by using a Hitachi Automatic Analyzer 7600 (Hitachi, Tokyo, Japan). The level of HbA1_c was measured by using high performance liquid chromatography (HLC-723G7, Tosoh, Tokyo, Japan), which was the certified method by the National Glycohemoglobin Standardization Program. All laboratory tests were performed in the Central Laboratory.

HGI was defined as the difference between measured HbA1_c and predicted HbA1_c (HGI=measured HbA1_c–predicted HbA1_c). Predicted HbA1_c was calculated by using the following formula through individual measurements of fasting plasma glucose (FPG), which were derived from linear regression analysis of measured HbA1_c and fasting glucose (HbA1_c (%)=0.0089×FPG [mg/dL]+4.6173 [R-squared=0.0466, P<0.001]) (Fig. 2). Study subjects were classified into 3 groups according to the tertile of HGI.

MS was defined by using the International Diabetes Federation criteria [23]. MS was confirmed, when a subject had abdominal obesity with 2 or more abnormalities in the following 4 measurements: TG, HDL-C, BP, and FPG. Abdominal obesity was defined when WC was ≥90th percentile in those aged 10–15 years or WC was ≥90 cm in males and ≥85 cm in females aged 16–18 years. Abnormal results were defined as follows: systolic or diastolic BP≥130/85 mmHg, FPG≥100 mg/dL, TG≥150 mg/dL, HDL-C<40 mg/dL in males aged 10–18 years and females aged 10–15 years, and <50 mg/dL in females aged between 16 and 18 years. CMRF clustering was defined as the presence of ≥2 abnormal components of MS [8,24]. ALT was used as a marker for NAFLD in a pediatric population [25]. ALT elevation was defined as ≥30 IU/L for males and ≥19 IU/L for females [26,27].

Data were analyzed using Stata 14.2 (StataCorp LP, College Station, TX, USA) with svy commands appropriate for the KNHANES sampling design. Data were expressed as the means±standard error for continuous variables or number of subjects (weighted percent) for categorical variables. Linear regression analysis was used to calculate predicted HGI values. Clinical characteristics were compared between tertiles of HGI by using the Student t-test or regression analysis for continuous variables, or logistic regression analysis and chi-square test for categorical variables. P-value for comparison between tertiles and P for trend from the lower to the upper HGI tertiles were obtained. Multiple logistic regression analysis was performed to determine the odds ratio (OR) with 95% confidence interval (CI) for association between HGI and components of MS and elevated ALT with adjustment for hemoglobin and BMI z-score. Further multiple logistic regression analyses were performed to identify association between HGI and CMRF clustering, MS and ALT elevation, after stratification by sex, age group (10–14 years and 15–18 years) and BMI category (normal weight and overweight/obese). Association between HGI and CMRFs were analyzed using simple and multiple linear regression analysis adjusting hemoglobin and BMI z-score. Due to the multicollinearity, HbA1_c and HGI was not used in the same analysis. A P value of <0.05 was considered statistically significant.

The KNHANES was approved by the Institutional Review Board of the KCDC (2011-02CON-06-C, 2012-01EXP-01-2C, 2013-07CON-03-4C, 2013-12EXP-03-5C). Written informed consent was obtained from all participants prior to the survey. The present study was approved by the Institutional Review Board of the Seoul National University Bundang Hospital (approval number: X-1807/483-902).

The mean of HGI (%) was 0.000±0.006, with a range of between -1.887 and 1.002. The mean and range of HGI by lower (HGI-T1), mid (HGI-T2), and upper (HGI-T3) tertiles were -0.277±0.006 (-1.887 to -0.107), 0.000±0.005 (-0.105 to 0.111), and 0.280±0.005 (0.113 to 1.002), respectively. The cumulative proportion of HGI according to measured HbA1_c is presented in Fig. 3.

The anthropometric and metabolic characteristics of study participants according to tertiles of HGI values are shown in Table 1. The mean age was 14.3±0.1 years, and 53.6% of participants were males. Overweight or obese subjects were 19.9% (n=768). Participants were divided into 2 age groups; 10–14 years (n=2,253, 48.8%) and 15–18 years (n=1,632, 51.2%). No significant differences between the groups were observed for age, sex, height z-scores, and WC. BMI z-scores and proportion of obesity increased with respect to HGI tertile. Subjects in higher HGI tertiles had higher values of HbA1_c, without differences in FPG. Serum level of total cholesterol and ALT statistically significantly increased with increasing HGI (Table 1). Proportion of abdominal obesity, elevated TG, and CMRF clustering showed an increasing tendency with respect to HGI tertile (Fig. 4). However, proportion of MS, hyperglycemia, high BP, low HDL-C, and elevated ALT did not reveal differences among the groups (Fig. 4).

There was no significant difference in components of MS and elevated ALT between HGI-T1 and HGI-T2 (Table 2). However, adjusted OR (95% CI) of elevated TG in HGI-T3 as compared with that in HGI-T1 was 1.65 (1.18–2.30). Other components, including MS and elevated ALT, did not show statistical significance.

In linear regression analysis, BMI z-score, HbA1_c, TG, and ALT showed a significant positive association with HGI after adjusting BMI z-score and hemoglobin (Table 3). WC, BP, FPG, and HDL-C were not associated with HGI.

When stratified according to sex, age group, and BMI category, those overweight and obese showed increased OR for predicting CMRF clustering and ALT elevation with respect to the HGI (Table 4). OR (95% CI) for predicting CMRF clustering and ALT elevation in overweight/obese participants was 2.92 (1.47–5.80) and 2.72 (1.24–5.98), respectively. However, HGI was not significant for predicting MS.