Age-related insulin resistance changes in children and adolescents and its impact on the accuracy of diagnosis of metabolic syndrome

Insung Kim; Kyu-Na Lee; Juyoung Sung; Yoon Ji Ahn; Minji Im; Kyungdo Han; Sung Yoon Cho

doi:10.6065/apem.2448180.090

Ann Pediatr Endocrinol Metab > Volume 30(4); 2025 > Article

Kim, Lee, Sung, Ahn, Im, Han, and Cho: Age-related insulin resistance changes in children and adolescents and its impact on the accuracy of diagnosis of metabolic syndrome

Original Article

Annals of Pediatric Endocrinology & Metabolism 2025;30(4): 213-222.

Published online: August 31, 2025

DOI: https://doi.org/10.6065/apem.2448180.090

Age-related insulin resistance changes in children and adolescents and its impact on the accuracy of diagnosis of metabolic syndrome

Insung Kim¹

, Kyu-Na Lee²

, Juyoung Sung¹

, Yoon Ji Ahn¹

, Minji Im¹

, Kyungdo Han³

, Sung Yoon Cho¹

¹Department of Pediatrics, Samsung Medical Center, Sungkyunkwan University School of Medicine, Seoul, Korea

²Department of Public Health, Catholic University, Seoul, Korea

³Department of Statistics and Actuarial Science, Soongsil University, Seoul, Korea

Address for correspondence: Sung Yoon Cho Department of Pediatrics, Samsung Medical Center, Sungkyunkwan University School of Medicine, 81 Irwon-ro, Gangnam-gu, Seoul 06351, Korea Email: nadri1217@naver.com

Address for co-correspondence: Kyungdo Han Department of Statistics and Actuarial Science, Soongsil University, Seoul, Republic of Korea, 369 Sangdo-ro, Dongjak-gu, Seoul 06978, Korea Email: hkd917@naver.com

Received August 6, 2024 Revised November 26, 2024 Accepted December 10, 2024

This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/4.0/) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

Abstract

Purpose

Although insulin resistance (IR) varies with age and puberty in children and adolescents, most previous attempts to determine cutoff values for IR indices overlook factor. This study assesses age-related differences in IR index values and evaluates how diagnostic performance varies by age when using a uniform cutoff for diagnosing metabolic syndrome (MetS) without considering age.

Methods

We analyzed age-related differences in IR indices (the homeostatic model assessment of insulin resistance [HOMA-IR], triglyceride-glucose [TyG] index, and triglyceride/high-density lipoprotein cholesterol [TG/HDL-C] ratio) among 1,641 participants in the 2019–2021 Korea National Health and Nutrition Examination Survey. We also examined IR index values for diagnosing MetS in 1,574 participants.

Results

IR indices showed significant age-related variations in group-comparison tests, with a peak at ages 12–13 years in males and 11–13 years in females (P<0.001 for the HOMA-IR, P<0.005 for the TG/HDL-C ratio in both males and females, and P=0.003 for the TyG index in females). Applying a uniform cutoff derived from receiver operating characteristic curve analysis for diagnosing MetS showed substantial age-related variation in diagnostic accuracy, with standard deviation-to-mean ratios of age-specific accuracy of >10% for the HOMA-IR and >5% for the TyG index, while showing minor variation (<5%) for the TG/HDL-C ratio. Using age-specific percentiles for the HOMA-IR (80th of the general population) and TyG index (80th of those without MetS) reduced these variations to <5% while maintaining similar diagnostic performance.

Conclusions

This study highlights the importance of age-related variation in IR in children and adolescents.

Keywords: Insulin resistance, Metabolic syndrome, Child, Adolescent

Highlights

· Insulin resistance (IR) in children and adolescents shows significant age-related variation, particularly during puberty.

· Age and pubertal development should be considered when applying IR indices in clinical practice.

· This study demonstrated that using age-specific percentiles of IR indices is likely more appropriate than applying a uniform cutoff in children and adolescents, as shown in their application to the diagnosis of metabolic syndrome.

Introduction

The prevalence of metabolic syndrome (MetS) and its subcriteria, such as obesity, hypertension, hyperglycemia, and dyslipidemia, is increasing in children and adolescents [1-3]. MetS in children and adolescents increases the risk of complications, including type 2 diabetes, in adulthood [4], with increased insulin resistance (IR) playing a key role as an important mechanism [5]. It is therefore important to monitor IR in children and adolescents with MetS.

The gold standard for measuring IR is the euglycemic hyperinsulinemic clamp (EHC) method [6]. However, because this method is too complex and cumbersome for routine use in large numbers of patients or the general population, various simpler indices of IR have been studied as alternatives. Among these, several surrogate markers that are easier to calculate and correlate closely with the use of EHCs have been introduced recently to predict the occurrence and prognosis of conditions such as diabetes, cardiovascular disease, and MetS.

Several studies have used the homeostatic model assessment for insulin resistance (HOMA-IR), which has a high correlation with EHC use, as a surrogate marker for IR [7-9]. However, the concentration of blood insulin required to calculate HOMA-IR values is not routinely measured in the blood tests of patients without diabetes. Alternative IR indices using lipid parameters, such as the triglyceride-glucose (TyG) index and the triglyceride/high-density lipoprotein cholesterol (TG/HDL-C) ratio are therefore more appropriate for individuals without measured blood insulin levels [10-13].

Studies of the EHC technique have found that IR in children and adolescents changes with the progression of puberty [14], and insulin requirements in children and adolescents with type 1 diabetes vary with age [15]. However, most studies of surrogate IR markers in children and adolescents did not take age and sexual development into account and instead set cutoff values based on data that combined the entire age range of children and adolescents [16-19].

This study examines the age-specific distribution of IR values to determine whether significant differences exist across age groups, based on data from children and adolescents aged 10 to 18 years in the Korea National Health and Nutrition Examination Survey (KNHANES). Additionally, this study evaluates how diagnostic accuracy varies by age when applying a uniform IR index cutoff derived from the entire age group in diagnosing MetS in children and adolescents, and explores whether using this cutoff or age-specific percentiles would be more appropriate in clinical practice.

Materials and methods

1. Study-data selection and variables

This study was conducted using data from the 8th KNHANES, which was conducted from 2019 to 2021. The 22,559 participants in the 8th KNHANES included 1,919 children and adolescents aged 10 to 18 years. Among these, 1,641 met the criteria of a fasting time of 8 hours or more and had available blood test results that could be used to analyze the age-specific distribution of IR index values. Additionally, 1,574 entries contained information on the presence of MetS for analysis of the relationship between MetS and the IR index.

Age, sex, weight, height, waist circumference (WC), systolic blood pressure (SBP) and diastolic blood pressure (DBP), and fasting blood test results (for glucose, hemoglobin A1c, insulin, total cholesterol, triglycerides [TG], and high-density lipoprotein cholesterol [HDL-C]) were collected. From this, the IR index values for each participant were calculated and the presence of MetS was determined as explained in section 2 of the Materials and Methods.

2. Definition of MetS and value of insulin resistance markers

The diagnostic criteria for MetS in children and adolescents are not clearly defined by a single standard. Standards set by the International Diabetes Federation (IDF) [20] and National Cholesterol Education Program (NCEP) Adult Treatment Panel III criteria tailored for children and adolescents [21] are commonly used. In this study, each diagnostic criteria was used separately for analysis.

In the IDF criteria, MetS is defined by a WC ≥ the 90th percentile for sex and age [22] for those under the age of 16, ≥90 cm for males aged 16 and older, or ≥80 cm for females aged 16 and older [23], and at least 2 of the following: SBP ≥ 130 mmHg or DBP ≥ 85 mmHg, fasting glucose ≥ 100 mg/dL, TG ≥ 150 mg/dL, and HDL-C < 40 mg/dL for all males and females under the age of 16 or ≤50 mg/dL for females aged 16 and older. In the NCEP criteria, a MetS diagnoses requires least 3 of the following 5 criteria to be met: WC ≥ the 90th percentile for sex and age; SBP or DBP ≥ the 90th percentile for sex, age, and height [24], fasting glucose ≥ 110 mg/dL, TG ≥ 110 mg/dL, and HDL-C ≤ 40 mg/dL.

Recent studies of Korean children and adolescents with MetS have used both WC reference values published in 2007 [22] and those from 2022 [25]. In this study, MetS was assessed using the 2007 WC values, while results derived using the 2022 WC reference values are provided in the Supplementary materials. We chose this approach because the 2007 reference values were used in the most recent studies on the prevalence of MetS [2] and because most prior studies examining the relationship between IR indices and MetS used the 2007 WC values, and using the same criteria is necessary to compare our results with those of previous studies.

The IR indices were calculated and analyzed using 3 methods: the HOMA-IR, TyG index, and TG/HDL-C ratio. The IR values were calculated using the following formulas: HOMA-IR=[fasting glucose (mg/dL) × fasting insulin concentration (μIU/mL)]/405; TyG index=Ln (TG (mg/dL) × fasting glucose (mg/dL)/2); and TG/HDL-C ratio=TG (mg/dL)/HDL-C (mg/dL).

3. Statistical analysis

Study data were analyzed using IBM SPSS Statistics ver. 29.0.2.0 (20) (IBM Co., USA) and R ver. 4.3.2 (R Foundation for Statistical Computing, Austria). As KNHANES data comprise complex samples with stratification, clustering, and weighting to represent the South Korean population, all analyses were performed using complex sample-analysis methods.

The references for each percentile of the IR index are presented as weighted percentiles, divided by the total population, sex, and age group. For the analysis of differences in IR index values by age, if the IR index value followed a normal distribution, a mean-differences test was conducted using analysis of variance; if the values did not follow a normal distribution, a Kruskal-Wallis median-difference test was used. To analyze the differences in IR index values based on the presence or absence of MetS, a t-test was used for normally distributed values, and the Mann-Whitney U-test was used for nonnormally distributed values.

The ability of the IR index to diagnose MetS was evaluated through a weighted receiver operating characteristic (ROC) curve analysis using the bootstrap method repeated 10,000 times, and the area under the curve (AUC) value and 95% confidence intervals were calculated. The statistical significance of differences in AUC values between different IR indices was also tested. The optimal IR index cutoff value for diagnosing MetS was the point with the highest Youden J statistic (sensitivity + specificity − 1), and the sensitivity, specificity, and accuracy at this point are presented.

We further analyzed age-specific accuracy (aCV) by applying the overall cutoff to individual age groups and compared sensitivity, specificity, Youden J, and accuracy, as well as the variation in aCV between a uniform cutoff and age-specific percentiles that showed comparable diagnostic performance. In this analysis, the diagnostic performance of age-specific percentiles was considered comparable if the Youden J statistic and sensitivity reached at least 90% and 95%, respectively, of those achieved with the uniform cutoff. To quantitatively present the variation in age-specific accuracy, the coefficient of variation of in aCV was defined as:

Age-specific accuracy standard deviation=∑i=1018mean of age ‐ specific accuracy − accuracy for age i29 age groups − 1

Age-specific accuracy coefficient of variation (%)=age ‐ specific accuracy standard deviationmean of age ‐ specific accuracy×100

The significance level for all analyses <0.05. A flow chart of the data selection and analysis process, population estimates, and distributions of variables for Korean children and adolescents aged 10 to 18 years derived from the 8th KNHANES data, as well as the measurement equipment and methods for each variable, are detailed in Supplementary material 1.

4. Ethical statement

This research was conducted using raw data from the 8th KNHANES. Approval for data access and analysis for academic purposes was granted by the Korea Disease Control and Prevention Agency (Institutional Review Board approval number: 2018-01-03-C-A, 2018-01-03-2C-A, 2018-01-03-5C-A). This study was reviewed and approved by the Institutional Review Board of Samsung Medical Center (approval No. 2024-10-071).

Results

The distribution of IR indices values for Korean children and adolescents aged 10–18 years, based on the 8th KNHANES data, is presented in Table 1 and Supplementary material 2A and B. IR indices by age followed a pattern in which, for males, the values were higher around the ages of 12–13 years, followed by a subsequent decrease, while in females this trend was evident around the ages of 11–13 years. In the analysis of whether there were significant differences in IR index values by age, the median values of HOMA-IR showed statistically significant differences by age in both males (P<0.001) and females (P<0.001), as did the TG/HDL-C ratio in males (P=0.004) and females (P=0.002). For the TyG index, the mean values showed statistically significant differences by age in females (P=0.003), while in males, the pattern of change was similar across ages (P=0.892) (Fig. 1). The analysis of IR indices values derived exclusively from data for patients without MetS (based on both the IDF and NCEP criteria) and data for patients with normal weight and normal glucose levels can be found in Supplementary material 2C–N. When the analysis was conducted using these data, the previously mentioned patterns in the changes in IR indices values by age were also observed.

Before determining the cutoff for diagnosing MetS, we analyzed whether IR index values were significantly higher in the MetS group compared with those in the non-MetS group. Results showed that IR values were significantly higher in the MetS group across all sexes, criteria, and IR indices (all P<0.001) (Table 2). In the ROC curve analysis to evaluate the performance of IR indices in diagnosing MetS, the TG/HDL-C ratio had the highest AUC. The difference in AUC values between TG/HDL-C ratio and HOMA-IR was statistically significant in males but not in females. Similar results can be observed in the analysis based on the 2022 WC criteria for classifying MetS, as detailed in Supplementary material 3. When applying the cutoffs derived from the ROC curve analysis, the age-specific diagnostic accuracy of MetS using the HOMA-IR and TyG showed a notable decrease in males aged 12–13 years and females aged 11–12 years, followed by a general increase in accuracy with advancing age (Fig. 2).

The aCV values for the HOMA-IR were >10% and the highest among the 3 indices, except for females as defined by NCEP criteria. The aCV values of TyG were >5%, while the TG/HDL-C ratio had the lowest aCV values at <5%, regardless of sex or MetS diagnostic criteria. When diagnosing MetS using age-specific percentiles for the HOMA-IR and TyG index derived from general population data, which had a diagnostic performance similar to the uniform cutoff, the aCV decreased to <5%. This reduction was more pronounced for the HOMA-IR. However, there was little difference in aCV between applying a uniform cutoff and age-specific percentiles for the TG/HDL-C ratio (Fig. 3). Detailed data on diagnostic performance and results from applying age-specific percentiles derived from those without MetS and those with normal weight and glucose levels are presented in Supplementary material 4.

Discussion

Previous studies using data from the KNHANES found variations in HOMA-IR values by age in children and adolescents [26]. Comparisons of data from 2007–2010 with those from 2019–2020 revealed that HOMA-IR values increased in overweight and obese children and adolescents, but confirmed that HOMA-IR values vary according to age, particularly during puberty. We built on these findings by statistically testing whether IR differs by age through an age-based comparison and increases the reliability of the findings by analyzing lipid profile-based IR indices. Moreover, we identified the clinical issues that can arise when cutoff values for IR indices are determined without considering age-related changes in IR, and showed how applying age-specific values is superior in terms of predicting MetS.

In the previous study mentioned earlier [26], HOMA-IR values were significantly higher in the present day compared to 2007–2010 period. It showed that levels of fasting glucose, insulin, and HDL-C differed significantly between the past and present, whereas triglyceride levels did not differ significantly. Compared with studies conducted in the past, the TyG index and TG/HDL-C ratio, both of which are IR indices related to TG, did not exhibit any notable differences. When compared with data from KNHANES conducted from 2007 to 2010 [17], the interquartile ranges changed little, with values remaining similar or slightly increasing at all percentiles (median: males 8.08 to 8.16, females 8.14 to 8.13) for the TyG index. When compared with data from KNHANES from 2007 to 2013 [27], TG/HDL-C ratio values were similar or slightly increased at all percentiles, similar to the TyG index (median: males 1.44 to 1.47, females 1.40 to 1.42). Given the well-established link between TG and overweight/obesity [28], and the role that the accumulation of free fatty acids in the body plays in the development of MetS and IR [29,30], these findings suggest that TG may be a less-sensitive marker for IR or overweight/obesity compared with glucose, insulin, and HDL-C. Further research is warranted on this topic.

Despite the aforementioned findings, the predictive performance achieved in this study for MetS was generally superior or similar those produced by the TG/HDL-C ratio and TyG index compared with the HOMA-IR. This may be because TG/HDL-C ratio and TyG index factors are already part of MetS criteria. The superior performance of lipid profile-based IR indices in predicting MetS compared with the HOMA-IR has been confirmed in other studies [31-33]. The AUC of the TG/HDL-C ratio in this study was close to or greater than 0.95, regardless of sex or the diagnostic criteria for MetS, which is a relatively high value. This is higher than the results of a previous study of the performance of the TyG index and TG/HDL-C ratio in diagnosing MetS using data from the 2013–2016 KNHANES [34]. Additionally, the cutoff in this study was determined using the point at which Youden J statistic was maximized, and the diagnostic performance was similar or superior to the previous method, in which the cutoff was set at the point where the Euclidean index was minimized.

Clinically, IR indices related to MetS are used primarily in 2 situations. First, they can be used to indirectly measure the severity of IR in patients already diagnosed with MetS. However, there is no consistent criterion for defining IR, and various cutoffs are used in clinical practice. Some commonly used HOMA-IR cutoffs include a value of 4.39, which corresponds to +2 standard deviations in a population of normal-weight individuals [35], and a value of 3.16 derived from a study, where IR was defined as a total insulin concentration of greater than 300 μU/mL during an oral glucose tolerance test, with HOMA-IR and ROC curve analyses used to determine the cutoff based on Youden J statistic [16]. Other methods include using the 95th percentile of a normal-weight population with normal fasting glucose [36], the 95th percentile of the overall population [19], or the 90th percentile of the overall population [37]. However, none of these studies proposed age- or puberty-specific cutoff values for children and adolescents. According to our findings, IR index values fluctuated significantly depending on age and pubertal stage. Age-specific reference values should therefore be used when applying these indices to assessments of IR in children and adolescents. The most appropriate standard for establishing age-specific cutoff values for IR would be the values associated with an increased risk of developing type 2 diabetes or other cardiovascular diseases. Further research is needed to explore the relationship between elevated IR index values during childhood and adolescence, and the long-term risk of developing cardiometabolic diseases.

A secondary use of IR indices is as a screening tool to identify individuals who may need further confirmation and treatment for MetS based on blood test results. In this case, the appropriate cutoff depends on which aspect of diagnostic performance is considered most important. In this study, we prioritized high sensitivity when setting the cutoff, even at the expense of slightly lower specificity, as confirming the diagnosis of MetS through additional measurements of WC and blood pressure after blood tests does not incur significant costs. We therefore set a stricter threshold for acceptable sensitivity when determining that the diagnostic performance of age-specific percentiles was comparable to that of a uniform cutoff. If the diagnostic performance is similar, a cutoff with a lower aCV is preferable. Furthermore, a unified and straightforward standard should be established for clinical utility. From this perspective, the 80th percentile values of the HOMA-IR in the general population were considered the most appropriate cutoffs across all sexes and MetS criteria in this study (Supplementary material 4B). For the TyG index, the 80th percentiles of the reference data from study participants without MetS were the appropriate cutoffs (Supplementary material 4C-1). In the case of the TG/HDL-C ratio, no notable differences in aCV were evident using age-specific percentiles and a uniform cutoff, making the uniform cutoff a more straightforward choice. Using data from the non-MetS population as a reference for age-specific percentiles aimed at MetS screening could help identify a unified percentile applicable across sexes and MetS criteria. However, data from individuals with normal weight and normal glucose levels did not effectively reduce the aCV or help find a unified percentile (Supplementary material 4D-1). This is likely because the distribution of IR values in this dataset does not accurately reflect that of the actual population.

This study has several limitations. First, as with other studies, our results strongly suggest that IR index values differ according to the progression of puberty. However, because the study data do not include information on sexual maturity, this cannot be definitively proven. To establish the most appropriate cutoff for IR indices in diagnosing MetS in children and adolescents, further research is needed to quantitatively confirm whether IR index values differ by pubertal stage within the same age group or whether the same IR cutoff can be applied to different ages if the pubertal stage is similar. However, large-scale research data that includes both information on IR indices and sexual maturity is lacking. Additionally, the observed increase in IR among boys aged 12–13 years and girls aged 11–13 years in this study appears to coincide with the onset of puberty, as suggested by studies on the standard sexual development stages of Korean children and adolescents conducted in 1994 and 2006 [38,39]. However, these studies are outdated, and given the recent increase in precocious puberty in Korea [40], the reference may not reflect the changes in the timing of the sexual development of children and adolescents. If a specific age group and sexual maturity arbitrarily linked without considering this trend, our results may not be applicable to individuals whose sexual maturity is significantly delayed or accelerated in clinical practice. Further research and data on updated standards of sexual maturity in children and adolescents are necessary.

Second, while this study tested whether the values of IR markers differed by age and obtained significant results, insufficient data for each subgroup were included to conduct a post hoc analysis of which groups showed significant differences, as the participants were divided into detailed groups by sex and age. Moreover, it was not possible to perform ROC analysis with sufficient sample sizes to derive AUC values with confidence intervals appropriate for determining the optimal IR index cutoff values for diagnosing MetS within each subgroup. To address this limitation, we compared diagnostic performance and the CV for accuracy across different age percentiles, thereby suggesting alternative age-specific cutoffs.

In conclusion, this study confirmed that IR indices in children and adolescents vary significantly with age. Applying a uniform cutoff for IR indices resulted in considerable diagnostic inaccuracies across different age groups when diagnosing MetS. The use of age-specific percentiles for IR indices can reduce variation of diagnostic accuracy, highlighting the necessity of considering age-related variations in clinical practice. Further research is needed to explore changes in IR indices according to sexual maturity and the relationship between IR indices and long-term health outcomes.

Supplementary material

Supplementary materials 1-4 are available at https://doi.org/10.6065/apem.2448180.090.

Notes

Conflicts of interest

No potential conflict of interest relevant to this article was reported.

Funding

This study received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Data availability

The data that support the findings of this study can be provided by the corresponding author upon reasonable request.

Acknowledgments

We thank the Korea Centers for Disease Control and Prevention for providing the data.

Author contribution

Conceptualization: IK, JS, KH, SYC; Data curation: IK, KNL; Formal analysis: IK, KNL, KH; Methodology: IK, KNL, JS, YJA, MI, KH, SYC; Project administration: SYC; Visualization: IK; Writing - original draft: IK, SYC; Writing - review & editing: IK, KNL, JS, YJA, MI, KH, SYC

Fig. 1.

Changes in insulin resistance indices by age. HOMA-IR and TG/HDL-C are expressed as the median and interquartile range, and analyzed using the Kruskal-Wallis test. TyG is expressed as mean and 95% confidence interval, and analyzed using analysis of variance. HOMA-IR, homeostatic model assessment for insulin resistance; TyG, triglyceride-glucose index; TG/HDL-C, triglyceride/high-density lipoprotein cholesterol ratio.

Fig. 2.

Receiver operating characteristic curves and age-specific accuracy variability when applying uniform cutoffs. (A) IDF criteria for males. (B) NCEP criteria for males. (C) IDF criteria for females. (D) NCEP criteria for females. The dashed lines show the accuracy from entire age group when applying cutoff values. HOMA-IR, homeostatic model assessment for insulin resistance; TyG, triglyceride-glucose index; TG/HDL-C, triglyceride/high-density lipoprotein cholesterol ratio; IDF, International Diabetes Federation; NCEP, modified National Cholesterol Education Program Adult Treatment Panel III; AUC, area under curve; CI, confidence interval; Sen, sensitivity; Spe, specificity; Acc, accuracy.

Fig. 3.

Diagnostic performance and age-specific accuracy coefficient of variation for uniform cutoff and age-specific percentiles. (A) IDF criteria for males. (B). NCEP criteria for males. (C) IDF criteria for females. (D) NCEP criteria for females. The age-specific percentiles used are based on general population data (Supplementary material 2A-1 and 2B-1). The dashed lines represent the percentile from the entire age group corresponding to the uniform cutoff. HOMA-IR, homeostatic model assessment for insulin resistance; TyG, triglyceride-glucose index; TG/HDL-C, triglyceride/high-density lipoprotein cholesterol ratio; IDF, International Diabetes Federation; NCEP, modified National Cholesterol Education Program Adult Treatment Panel III.

Table 1.

Distribution of insulin resistance indices for general Korean children and adolescents aged 10–18 years

Sex	IR index	5p	10p	15p	20p	25p	30p	40p	50p	60p	70p	75p	80p	85p	90p	95p
Total	HOMA-IR	1.05	1.33	1.51	1.71	1.85	2.00	2.38	2.71	3.19	3.83	4.20	4.63	5.33	6.20	8.07
	TyG	7.38	7.55	7.67	7.76	7.81	7.89	8.01	8.15	8.27	8.40	8.50	8.59	8.69	8.87	9.09
	TG/HDL-C	0.60	0.75	0.84	0.92	1.00	1.08	1.26	1.45	1.70	2.04	2.25	2.50	2.93	3.39	4.41
Male	HOMA-IR	1.03	1.27	1.42	1.66	1.82	1.98	2.38	2.74	3.34	4.00	4.29	4.88	5.47	6.61	8.53
	TyG	7.34	7.52	7.64	7.73	7.79	7.87	8.03	8.16	8.29	8.40	8.51	8.59	8.69	8.89	9.12
	TG/HDL-C	0.57	0.71	0.84	0.93	1.00	1.10	1.28	1.47	1.77	2.17	2.36	2.66	3.02	3.49	4.55
Female	HOMA-IR	1.14	1.43	1.60	1.76	1.89	2.05	2.39	2.70	3.07	3.68	4.00	4.41	5.04	5.86	7.66
	TyG	7.44	7.58	7.70	7.78	7.84	7.90	8.01	8.13	8.26	8.41	8.50	8.60	8.69	8.84	9.07
	TG/HDL-C	0.64	0.77	0.84	0.92	0.98	1.07	1.24	1.42	1.62	1.92	2.15	2.33	2.63	3.21	4.02

IR, insulin resistance; p, percentile; HOMA-IR, homeostatic model assessment for insulin resistance; TyG, triglyceride-glucose index; TG/HDL-C, triglyceride/high-density lipoprotein cholesterol ratio.

Table 2.

Differences in IR index values between groups with and without metabolic syndrome

HOMA-IR		MetS	Median (IQR)	P-value
Male	IDF	(-)	2.68 (1.80–4.19)	<0.001
		(+)	8.04 (4.91–10.46)
	NCEP	(-)	2.63 (1.76–4.17)	<0.001
		(+)	7.14 (4.05–9.14)
Female	IDF	(-)	2.67 (1.87–3.85)	<0.001
		(+)	7.66 (5.39–12.31)
	NCEP	(-)	2.63 (1.87–3.83)	<0.001
		(+)	7.81 (5.96–11.75)
TyG		MetS	Mean±SE
Male	IDF	(-)	8.13±0.02	<0.001
		(+)	9.13±0.05
	NCEP	(-)	8.11±0.02	< 0.001
		(+)	8.97±0.06
Female	IDF	(-)	8.15±0.02	<0.001
		(+)	9.19±0.12
	NCEP	(-)	8.14±0.02	<0.001
		(+)	9.11±0.10
TG/HDL-C		MetS	Median (IQR)
Male	IDF	(-)	1.44 (1.00–2.24)	<0.001
		(+)	4.71 (3.68–6.30)
	NCEP	(-)	1.39 (0.98–2.17)	<0.001
		(+)	4.28 (3.30–5.31)
Female	IDF	(-)	1.40 (0.97–2.09)	<0.001
		(+)	4.59 (3.70–5.83)
	NCEP	(-)	1.36 (0.97–2.00)	<0.001
		(+)	3.98 (3.37–5.47)

HOMA-IR, homeostatic model assessment for insulin resistance; MetS, metabolic syndrome; IQR, interquartile range, IDF, International Diabetes Federation criteria; NCEP, modified National Cholesterol Education Program Adult Treatment Panel III criteria; TyG, triglyceride-glucose index; TG/HDL-C, triglyceride/high-density lipoprotein cholesterol ratio; SE, standard error.

HOMA-IR and TG/HDL-C ratio were tested using the Mann-Whitney U-test, while the TyG index was tested using the t-test.

References

1. Chae J, Seo MY, Kim SH, Park MJ. Trends and risk factors of metabolic syndrome among Korean Adolescents, 2007 to 2018. Diabetes Metab J 2021;45:880-9.

2. Jeong SM, Jung JH, Yang YS, Kim W, Cho IY, Lee YB, et al. 2023 Obesity fact sheet: prevalence of obesity and abdominal obesity in adults, adolescents, and children in Korea from 2012 to 2021. J Obes Metab Syndr 2024;33:27-35.

3. Park SI, Suh J, Lee HS, Song K, Choi Y, Oh JS, et al. Ten-year trends of metabolic syndrome prevalence and nutrient intake among Korean children and adolescents: a population-based study. Yonsei Med J 2021;62:344-51.

4. Morrison JA, Friedman LA, Wang P, Glueck CJ. Metabolic syndrome in childhood predicts adult metabolic syndrome and type 2 diabetes mellitus 25 to 30 years later. J Pediatr 2008;152:201-6.

5. Fahed G, Aoun L, Bou Zerdan M, Allam S, Bou Zerdan M, Bouferraa Y, et al. Metabolic syndrome: updates on pathophysiology and management in 2021. Int J Mol Sci 2022;23:786.

6. Gastaldelli A. Measuring and estimating insulin resistance in clinical and research settings. Obesity (Silver Spring) 2022;30:1549-63.

7. Stern SE, Williams K, Ferrannini E, DeFronzo RA, Bogardus C, Stern MP. Identification of individuals with insulin resistance using routine clinical measurements. Diabetes 2005;54:333-9.

8. Gayoso-Diz P, Otero-González A, Rodriguez-Alvarez MX, Gude F, García F, De Francisco A, et al. Insulin resistance (HOMA-IR) cut-off values and the metabolic syndrome in a general adult population: effect of gender and age: EPIRCE cross-sectional study. BMC Endocr Disord 2013;13:47.

9. Isokuortti E, Zhou Y, Peltonen M, Bugianesi E, Clement K, Bonnefont-Rousselot D, et al. Use of HOMA-IR to diagnose non-alcoholic fatty liver disease: a population-based and inter-laboratory study. Diabetologia 2017;60:1873-82.

10. McLaughlin T, Reaven G, Abbasi F, Lamendola C, Saad M, Waters D, et al. Is there a simple way to identify insulin-resistant individuals at increased risk of cardiovascular disease? Am J Cardiol 2005;96:399-404.

11. Guerrero-Romero F, Simental-Mendía LE, González-Ortiz M, Martínez-Abundis E, Ramos-Zavala MG, Hernández-González SO, et al. The product of triglycerides and glucose, a simple measure of insulin sensitivity. Comparison with the euglycemic-hyperinsulinemic clamp. J Clin Endocrinol Metab 2010;95:3347-51.

12. Jeppesen J, Hein HO, Suadicani P, Gyntelberg F. Relation of high TG-low HDL cholesterol and LDL cholesterol to the incidence of ischemic heart disease. An 8-year follow-up in the Copenhagen Male Study. Arterioscler Thromb Vasc Biol 1997;17:1114-20.

13. Fiorentino TV, Marini MA, Succurro E, Andreozzi F, Sesti G. Relationships of surrogate indexes of insulin resistance with insulin sensitivity assessed by euglycemic hyperinsulinemic clamp and subclinical vascular damage. BMJ Open Diabetes Res Care 2019;7:e000911.

14. Moran A, Jacobs DR Jr, Steinberger J, Hong CP, Prineas R, Luepker R, et al. Insulin resistance during puberty: results from clamp studies in 357 children. Diabetes 1999;48:2039-44.

15. Wiegand S, Raile K, Reinehr T, Hofer S, Näke A, Rabl W, et al. Daily insulin requirement of children and adolescents with type 1 diabetes: effect of age, gender, body mass index and mode of therapy. Eur J Endocrinol 2008;158:543-9.

16. Keskin M, Kurtoglu S, Kendirci M, Atabek ME, Yazici C. Homeostasis model assessment is more reliable than the fasting glucose/insulin ratio and quantitative insulin sensitivity check index for assessing insulin resistance among obese children and adolescents. Pediatrics 2005;115:e500-3.

17. Kim JW, Park SH, Kim Y, Im M, Han HS. The cutoff values of indirect indices for measuring insulin resistance for metabolic syndrome in Korean children and adolescents. Ann Pediatr Endocrinol Metab 2016;21:143-8.

18. Song K, Park G, Lee HS, Choi Y, Oh JS, Choi HS, et al. Prediction of insulin resistance by modified triglyceride glucose indices in Youth. Life (Basel) 2021;11:286.

19. Vanlancker T, Schaubroeck E, Vyncke K, Cadenas-Sanchez C, Breidenassel C, González-Gross M, et al. Comparison of definitions for the metabolic syndrome in adolescents. The HELENA study. Eur J Pediatr 2017;176:241-52.

20. Zimmet P, Alberti KG, Kaufman F, Tajima N, Silink M, Arslanian S, et al. The metabolic syndrome in children and adolescents - an IDF consensus report. Pediatr Diabetes 2007;8:299-306.

21. Cook S, Auinger P, Li C, Ford ES. Metabolic syndrome rates in United States adolescents, from the National Health and Nutrition Examination Survey, 1999-2002. J Pediatr 2008;152:165-70.

22. Moon JS, Lee SY, Nam CM, Choi JM, Choe BK, Seo JW, et al. 2007 Korean National Growth Charts: review of developmental process and an outlook. Korean J Pediatr 2008;51:1-25.

23. Lee J, Kang SC, Kwon O, Hwang SS, Moon JS, Chae HW, et al. Temporal trends of the prevalence of abdominal obesity and metabolic syndrome in Korean children and adolescents between 2007 and 2020. J Obes Metab Syndr 2023;32:170-8.

24. Kim SH, Park Y, Song YH, An HS, Shin JI, Oh JH, et al. Blood pressure reference values for normal weight Korean children and adolescents: data from The Korea National Health and Nutrition Examination Survey 1998-2016: the Korean Working Group of Pediatric Hypertension. Korean Circ J 2019;49:1167-80.

25. Lee J, Kang SC, Kwon O, Hwang SS, Moon JS, Kim J. Reference values for waist circumference and waist-height ratio in Korean children and adolescents. J Obes Metab Syndr 2022;31:263-71.

26. Kim S, Song K, Lee M, Suh J, Chae HW, Kim HS, et al. Trends in HOMA-IR values among South Korean adolescents from 2007-2010 to 2019-2020: a sex-, age-, and weight status-specific analysis. Int J Obes (Lond) 2023;47:865-72.

27. Shim YS, Baek JW, Kang MJ, Oh YJ, Yang S, Hwang IT. Reference values for the triglyceride to high-density lipoprotein cholesterol ratio and non-high-density lipoprotein cholesterol in Korean children and adolescents: the Korean National Health and Nutrition Examination Surveys 2007-2013. J Atheroscler Thromb 2016;23:1334-44.

28. Jung MK, Yoo EG. Hypertriglyceridemia in obese children and adolescents. J Obes Metab Syndr 2018;27:143-9.

29. Rachek LI. Free fatty acids and skeletal muscle insulin resistance. Prog Mol Biol Transl Sci 2014;121:267-92.

30. Tagi VM, Chiarelli F. Obesity and insulin resistance in children. Curr Opin Pediatr 2020;32:582-8.

31. Dundar C, Terzi O, Arslan HN. Comparison of the ability of HOMA-IR, VAI, and TyG indexes to predict metabolic syndrome in children with obesity: a cross-sectional study. BMC Pediatr 2023;23:74.

32. Chu SY, Jung JH, Park MJ, Kim SH. Risk assessment of metabolic syndrome in adolescents using the triglyceride/high-density lipoprotein cholesterol ratio and the total cholesterol/high-density lipoprotein cholesterol ratio. Ann Pediatr Endocrinol Metab 2019;24:41-8.

33. Zhou F, Sun X, Liu J, Li L, Li L, Li P. Triglyceride to high-density lipoprotein cholesterol ratio in adolescence as a predictive marker of metabolic syndrome and obesity in early adulthood in China. Endocrine 2022;76:331-40.

34. Lee J, Ah Lee Y, Yong Lee S, Ho Shin C, Hyun Kim J. Comparison of lipid-derived markers for metabolic syndrome in youth: triglyceride/HDL cholesterol ratio, triglyceride-glucose index, and non-HDL cholesterol. Tohoku J Exp Med 2022;256:53-62.

35. Lee JM, Okumura MJ, Davis MM, Herman WH, Gurney JG. Prevalence and determinants of insulin resistance among U.S. adolescents: a population-based study. Diabetes Care 2006;29:2427-32.

36. Yi KH, Hwang JS, Kim EY, Lee SH, Kim DH, Lim JS. Prevalence of insulin resistance and cardiometabolic risk in Korean children and adolescents: a population-based study. Diabetes Res Clin Pract 2014;103:106-13.

37. Ascaso JF, Romero P, Real JT, Lorente RI, Martínez-Valls J, Carmena R. Abdominal obesity, insulin resistance, and metabolic syndrome in a southern European population. Eur J Intern Med 2003;14:101-6.

38. Hong CH, Rho HO, Song SH. The sexual maturity rating of adolescent boys and girls in Korea. J Korean Pediatr Soc 1994;37:193-8.

39. Park MJ, Lee IS, Shin EK, Joung H, Cho SI. The timing of sexual maturation and secular trends of menarchial age in Korean adolescents. Korean J Pediatr 2006;49:610-6.

40. Kang S, Park MJ, Kim JM, Yuk JS, Kim SH. Ongoing increasing trends in central precocious puberty incidence among Korean boys and girls from 2008 to 2020. PLoS One 2023;18:e0283510.