Association between Triglyceride Glucose Index and Insulin Resistance among Thai Obese Adolescents

Globally, obesity is one of the most important chronic medical conditions which affects cell metabolism and there increase in adipose tissue, thereby increasing the risk of metabolic syndrome such as Cardiovascular Diseases, hypertension and Type 2 Diabetes Mellitus (T2DM) [1-4]. The prevalence of obesity among children and adolescents is a risk factor for chronic diseases in adulthood, and consequently affects morbidity/mortality in future [5,6]. Obesity is an excess of body fat or adiposity, whether measured by the Body Mass Index (BMI) or body fat [7,8]. Excess adiposity is related to IR, as indexed by elevated HOMA-IR levels [7].

Insulin resistance is a key factor of the metabolic syndrome, and correlates positively with obesity [9,10]. As a consequence of IR in obesity, there is impairment of glucose metabolism [11]. Accordingly, IR plays an important roles in the development of T2DM [12]. In a recent research, Guerrero-Romero et al., proposed that a new fasting index for the early recognition of IR, the product of fasting TyG index. HOMA-IR is a gold standard technique for assessing IR, however this method is difficult and expensive [13]. In addition, the advantage of TyG index is its easily applicability in the clinical setting and it does not require measurement of insulin levels [14].

Recently, the TyG index was evaluated in some diseases (T2DM, CVD), and ethnic groups such as Mexican-Americans and Caucasians from the San Antonio, Koreans, Argentine, Italian, and Brazilian adults [15-18], however it has not yet been determined in the ethnic groups in Southeast Asia. Because, some studies reported that TyG index was different depending upon ethnicities [15-18], hence this study aimed to investigate the association between the TyG index, clinical parameters, and IR for pathogenesis risk of surrogate IR in Thai obese adolescents.

Materials and Methods

The present cross-sectional study was carried out in 300 adolescents between the age group of 13-18 years selected from the Obesity Outpatient Clinic of Mahasarakham Provincial Public Health Office from 2009 to 2013. The Ethical Committee of Mahasarakham University approved this research (number: 0128/2008).

Sample Size Calculation

The sample size was calculated by Cohen [19] for multiple regression analysis with the outcome of IR at a significance level of 5% (α=0.05), statistical power of 0.9, and assuming 20%

The lambda (λ) value depended on the degree of freedom of the denominator of the v value. v=N-u-1 and v value was taken from Cohen [19] to obtain the lambda value; N is sample size from previous study [20] and u is independent variables. While R²_Y,A is the observed multiple correlation coefficient of the model from previous study.

For a trail value of v=71, λ=20.5, substituting λ into the sample size formula, gives R²_Y,A=0.1 multiple correlation coefficient of the model from Vasques AC et al., [20].

Calculating for a 20% drop-out by formula: N_d=N/(1-R)² where N is the sample size calculated no dropouts, N_d that required with dropouts and R dropout rate is expected from Lachin JM et al., and approximately 300 [21].

Inclusion criteria were age 13-18 years and obesity following BMI cut-off points by Cole TJ et al., [22]. Adolescents whose BMI was between 20.90-33.63 kg/m² were included and those whose BMI were not in this range were excluded from the study. Informed consent was obtained from the subjects and their parents before the study.

Demographic information including age, sex, family history of diabetes, and family history of obesity were collected from all the participants using a detailed proforma. For all the subjects, family history of diabetes and obesity was obtained from their parents, sibling and grandparents. The proforma was adapted from Ethical Committee of Mahasarakham University.

Anthropometry

Weight and height (without shoes) were measured by carefully calibrated beam balance (Detecto^@) for weight and a vertical measuring rod for height. These measurements were used for the calculation of BMI for each subject. Skin fold thickness was measured at the biceps and triceps with a calliper according to standard procedures; the average of three readings was recorded in millimeter.

Laboratory Determinations

A 10 mL of fasting blood samples were obtained from all the subjects by nurses and medical technologists. They were immediately processed and divided into aliquots and stored at -80°C for further analyses. The levels of TG were determined using commercially available test kits from Siemens Healthcare Diagnostic Inc®. Fasting Blood Sugar (FBS) levels were calculated using enzymatic test kits from Dade Behring Inc®. In addition, fasting serum insulin was measured by radioimmunoassay test from Linco Research, Inc®. TG and FBS were used to calculate TyG index by multiplying TG (mg/dL) with FBS (mg/dL) and dividing the product by 2 [18]. The TyG index was categorised into three groups based on tertiles of their natural distributions.

Insulin Resistance (IR)

Evaluation of IR was done through the HOMA-IR, which is the product of fasting serum insulin (μU/mL) and FBS (mmol/L) and divided by 22.5 [23]. IR was defined as HOMA-IR value with higher than 90^th percentile of the subjects [17].

Statistical Analysis

The TyG index groups were divided by the concept of self-cutting point as tertiles mode score group ranking of existed data by STATA software. The descriptive statistics as mean and standard deviation was performed for clinical variables, and the differences of those variables in TyG groups were assessed by using ANOVA. Bivariate odds ratio (OR) to examine the association of each factor and IR was conducted. From bivariate analysis, factors with a value where p<0.05 were selected for multiple logistic regression. The adjusted OR estimated by multiple logistic regression indicated the association between the IR and TyG index after adjustments were made for others factors. The reference group of IR was lower than 90^th percentile. The statistical significance level was set as a p-value <0.05, and STATA version 13.0 software was implied for all data analyses.

Results

The clinical laboratory parameter for all subjects according to TyG index groups is illustrated in [Table/Fig-1].

[Table/Fig-1]:

Clinical parameters for all subjects according to TyG index groups.

Parameters	Q1	Q2	Q3	p-value
Number of subjects (n)	100	100	100
Age (years)	15.51±1.74	15.14±1.63	14.55±1.48	<0.001**
Sex (M/F)	25/75	28/72	28/72	0.535
BMI (kg/m2)	25.05±4.15	26.59±4.35	28.91±4.72	<0.001**
FBS (mg/dL)	85.62±6.67	86.40±5.97	90.28±14.45	<0.001**
TG (mg/dL)	50.84±13.27	89.89±12.21	173.76±63.68	<0.001**
Bicep (mm)	16.65±6.75	18.65±7.87	20.04±7.09	0.012*
Tricep (mm)	21.34±7.92	22.32±8.56	24.32±7.28	0.022*
TyG index	2.16±0.26	3.87±0.53	7.88±3.35	<0.001**
HOMA-IR	0.43±0.24	0.47±0.24	0.75±0.83	<0.001**
Family history of diabetes (%)	51 (51.00)	42 (42.00)	47 (47.00)	0.848
Family history of obesity (%)	32 (32.00)	35 (35.00)	26 (26.00)	0.808

Data are the mean±SD and number (percentage)

p<0.05* statistically significant (ANOVA)

p<0.001** statistically highly significant (ANOVA)

The subjects were stratified into tertiles based on their TyG index levels and dividing the data by STATA software. Few factors including age, BMI, FBS, and TG were statistically highly significant (p<0.001) among the three groups. The tertile (Q3) with the highest mean TyG index also had higher BMI, FBS, TG, bicep, tricep HOMA-IR than (Q2 and Q1). However, sex, family history of diabetes and obesity were not significantly different among the tertile groups. In the same way TyG and HOMA-IR were significantly different among the tertiles (p<0.001).

The association between the TyG index and IR was further explored by categorising the HOMA-IR into high and low. In the bivariate logistic regression analysis, it was observed that HOMA-IR and TyG index were significantly highly associated (3.19; 95% CI 1.782, 5.369; p<0.001), whereas BMI and tricep thickness were significantly associated (p<0.05) with HOMA-IR [Table/Fig-2].

[Table/Fig-2]:

Association of factors and IR with adolescent obesity using Bivariate logistic regression.

Factor	Crude OR	95% CI	p-value
Age	-1.78	0.186, 1.08	0.06
Sex	0.04	0.434, 2.390	0.965
BMI	2.28	1.143, 4.857	0.018*
Bicep	1.89	0.972, 3.773	0.052
Tricep	1.96	1.020, 3.917	0.043*
TyG index	3.19	1.782, 5.369	<0.001**
Family history of diabetic
No	1
Yes	0.39	0.542, 2.484	0.699
Family history of obesity
No	1
Yes	0.54	0.541, 2.953	0.583

OR: Odds ratio, CI: Confidence interval, p<0.05* statistically significant, p<0.001** statistically highly significant

In the multiple logistic regression analysis in fully adjusted OR, the association between TyG index and HOMA-IR was adjusted for other factors including BMI and tricep (3.06; 95% CI 1.780, 5.170; p<0.001) [Table/Fig-3].

[Table/Fig-3]:

Odds ratios for TyG index on IR based on multiple logistic regression.

Factor	Crude OR	Adjusted OR	95% CI	p-value
TyG index	3.19	3.06	1.780, 5.170	<0.001**

p<0.001** statistically highly significant

The multiple logistic regression model included variables with p-value <0.05.

Discussion

In the present study, the association between the BMI, tricep and TyG index with HOMA-IR in Thailand obese adolescents in a fully adjusted model was identified. Furthermore, no significant association between sex, age, bicep, family history of diabetes, and family history of obesity with HOMA-IR was found. To the best of our knowledge, this is the first study on ethnic groups in Southeast Asia in obese adolescents to investigate the relationship between increasing levels of TyG index and HOMA-IR. And, the findings were in agreement with other researches that HOMA-IR is not only associated with TyG index, but also other body adiposity indicators such as percentages of total and central body fat, and waist-to-height ratio [17].

A study by Moon S et al., had similar findings, they proposed that TyG index was a good surrogate for insulin sensitivity in Korean adolescents, and cutoff point with sensitivity (89.8%) and specificity (83.1%) had a value of 8.55 [24]. As, IR is characterised with a decrease in cell sensitivity to insulin, and it is one of the leading factors for causing metabolic disease especially T2DM [14,25,26]. Some researchers have explained that TyG index is correlated with IR and TyG index has been considered to be a surrogate marker of IR [14,15,27]. In addition, Kodama K et al., proposed that ethnic differences should be considered for IR risk [28]. Moon S et al., also suggested that studies based on ethnic and gender-specific characteristics of TyG index in adolescents are lacking [24]. Therefore, the present study showed that TyG index was associated with IR in Thailand obese adolescents, when the association was adjusted with factors of obesity including BMI and tricep thickness. This result corroborates the previous finding that TyG index and HOMA-IR were positively correlated (p<0.001) in Caucasian children [29], and with Dikaiakou E et al., who studied Greek children and adolescents and found that TyG index showed a positive correlation with IR [30]. Few researchers in Thailand, studied about TyG index, Chamroonkiadtikun P et al., reported that TyG index was associated with risk of T2DM and could be used as a biomarker of developing T2DM in adults [31].

The TyG index level was likewise significantly associated with HOMA-IR in bivariate and multiple logistic regression analysis. Our results correspond to Kang B et al., who reported that the association was found to be (r=0.41) between the TyG index and HOMA-IR in Korean adolescents [32]. Meanwhile, Locateli JC et al., showed that HOMA-IR had a positive correlation with TyG index (r=0.46, p<0.001) and TG/HDL index (r=0.36, p<0.001) in South American overweight and obese children and adolescents [33]. Additionally, association between TyG index and HOMA-IR from bivariate analysis (p<0.001) was found and also found a significant association between the TyG index and HOMA-IR after adjusted with BMI, and tricep (p<0.001). Thus, HOMA-IR is not only associated with TyG index, but also other obesity risk factors. In a recent study, Vieira-Ribeiro SA et al., reported that body adiposity is positively associated with IR, evaluated with the TyG index by multiple linear regression (β=0.010, 95% CI=0.003-0.013; p<0.001) [17]. The triceps skinfold thickness is a simple anthropometric measurement while screening for adiposity in male children and adolescents [34,35]. Chao YP et al., showed that Mid-Arm Muscle Circumference (MAMC) can be used as a surrogate marker in predicting IR in non obese elderly, since MAMC had a significant association with HOMA-IR (r=0.213, p<0.05). And they mentioned that MAMC may have substantial additional value in screening for IR [35]. While Abe Y et al., showed that increased abdominal adiposity as assessed by waist-to-height ratio was related with an increased prevalence of IR, as evaluated by the HOMA-IR [36].

Limitation(s)

The limitations of the study included the cross-sectional design which limited the ability to infer a causal association between TyG index and IR. The pubertal stage was not evaluated. Because obesity in adolescents was associated with higher HOMA-IR values and IR increased at the end of pubertal maturation, adolescents obesity can have a possible influence on the TyG index. Further experiments designed to investigate the role of pubertal stage in association with TyG index will help clarify the role of puberty in obesity adolescent.

Conclusion(s)

In conclusion, the results of TyG index was significantly associated with IR, in Thai obese adolescents, hence confirmed that TyG index can be used as a surrogate marker for IR. However, further experiments are needed to clarify the mechanisms of this association.

Data are the mean±SD and number (percentage)p<0.05* statistically significant (ANOVA)p<0.001** statistically highly significant (ANOVA)OR: Odds ratio, CI: Confidence interval, p<0.05* statistically significant, p<0.001** statistically highly significantp<0.001** statistically highly significant

[1]. Cardenas-Vargas E, Nava JA, Garza-Veloz I, Torres-Castañeda MC, Galván-Tejada CE, Cid-Baez MA, The influence of obesity on puberty and insulin resistance in Mexican children Int J Endocrinol 2018 2018:706729210.1155/2018/706729230254673  [Google Scholar]  [CrossRef]  [PubMed]

[2]. Kostovski M, Simeonovski V, Mironska K, Tasic V, Gucev Z, Metabolic profiles in obese children and adolescents with insulin resistance OA Maced J Med Sci 2018 6(3):511-18.10.3889/oamjms.2018.09729610610  [Google Scholar]  [CrossRef]  [PubMed]

[3]. McMorrow AM, Connaughton RM, Lithander FE, Roche HM, Adipose tissue dysregulation and metabolic consequences in childhood and adolescent obesity: Potential impact of dietary fat quality Proc Nutr Soc 2015 74(1):67-82.10.1017/S002966511400158X25497038  [Google Scholar]  [CrossRef]  [PubMed]

[4]. Morandi A, Maffeis C, Predictors of metabolic risk in childhood obesity Horm Res Paediatr 2014 82(1):03-11.10.1159/00036223724923289  [Google Scholar]  [CrossRef]  [PubMed]

[5]. Weiss R, Caprio S, The metabolic consequences of childhood obesity Best Pract Res Clin Endocrinol Metab 2005 19(3):405-19.10.1016/j.beem.2005.04.00916150383  [Google Scholar]  [CrossRef]  [PubMed]

[6]. Cho WK, Kim H, Lee HY, Han KD, Jeon YJ, Jung IA, Insulin resistance of normal weight central obese adolescents in Korea stratified by waist to height ratio: Results from the Korea National Health and Nutrition Examination Surveys 2008-2010 Int J Endocrinol 2015 2015:15875810.1155/2015/15875826257779  [Google Scholar]  [CrossRef]  [PubMed]

[7]. Lim SM, Choi DP, Rhee Y, Kim PC, Association between obesity indices and insulin resistance among healthy Korean adolescents the JS high school study PLoS One 2015 10:e012523810.1371/journal.pone.012523825970186  [Google Scholar]  [CrossRef]  [PubMed]

[8]. Güngör NK, Overweight and obesity in children and adolescents J Clin Res Pediatr Endocrinol 2014 6(3):129-43.10.4274/jcrpe.147125241606  [Google Scholar]  [CrossRef]  [PubMed]

[9]. Cunningham SA, Kramer MR, Narayan KM, Incidence of childhood obesity in the United States N Engl J Med 2014 370(17):1660-61.10.1056/NEJMoa1309753  [Google Scholar]  [CrossRef]

[10]. Hamdy O, Porramatikul S, Al-Ozairi E, Metabolicobesity: The paradox between visceral and subcutaneous fat Curr Diabetes Rev 2006 2(4):367-73.10.2174/157339981060204036718220642  [Google Scholar]  [CrossRef]  [PubMed]

[11]. Weiss R, Kaufman FR, Metabolic complications of childhood obesity: Identifying and mitigating the risk Diabetes Care 2008 31(Suppl 2):S310-16.10.2337/dc08-s27318227502  [Google Scholar]  [CrossRef]  [PubMed]

[12]. Lastra G, Manrique C, The expanding role of oxidative stress, renin angiotensin system, and beta-cell dysfunction in the cardiometabolic syndrome and type 2 diabetes mellitus Antioxid Redox Sign 2007 9(7):943-54.10.1089/ars.2007.161517508916  [Google Scholar]  [CrossRef]  [PubMed]

[13]. Guerrero-Romero F, Simental-Mendía LE, González-Ortiz M, Martínez-Abundis E, Ramos-Zavala MG, Hernández-González SO, The product of triglycerides and glucose, a simple measure of insulin sensitivity. Comparison with the euglycemic-hyperinsulinemic clamp J Clin Endocrinol Metab 2010 95(7):3347-51.10.1210/jc.2010-028820484475  [Google Scholar]  [CrossRef]  [PubMed]

[14]. Navarro-González D, Sánchez-Íñigo L, Fernández-Montero A, Pastrana-Delgado J, Martinez JA, TyG Index change is more determinant for forecasting type 2 diabetes onset than weight gain Medicine 2016 95(19):e364610.1097/MD.000000000000364627175686  [Google Scholar]  [CrossRef]  [PubMed]

[15]. Kim MK, Ahn CW, Kang S, Nam JS, Kim KR, Park JS, Relationship between the triglyceride glucose index and coronary artery calcifcation in Korean adults Cardiovasc Diabetol 2017 16(1):10810.1186/s12933-017-0589-4  [Google Scholar]  [CrossRef]

[16]. Guerrero-Romero F, Villalobos-Molina R, Jimenez-Flores JR, Simental-Men-dia LE, Mendez-Cruz R, Murguia-Romero M, Fasting triglycerides and glucose index as a diagnostic test for insulin resistance in young adults Arch Med Res 2016 47(5):382-87.10.1016/j.arcmed.2016.08.01227751372  [Google Scholar]  [CrossRef]  [PubMed]

[17]. Vieira-Ribeiro SA, Fonseca PC, Andreoli CS, Ribeiro AQ, Hermsdorff HH, Pereira PF, The TyG index cut off point and its association with body adiposity and lifestyle in children J Pediatr(Rio J) 2019 95(2):217-23.10.1016/j.jped.2017.12.01229457996  [Google Scholar]  [CrossRef]  [PubMed]

[18]. Simental-Mendıa LE, Rodrıguez-Moraan M, Guerrero-Romero F, The product off a sting glucose and triglycerides as surrogate for identifying insulin resistance in apparently healthy subjects Metab Syndr Relat Disord 2008 6(4):299-304.10.1089/met.2008.003419067533  [Google Scholar]  [CrossRef]  [PubMed]

[19]. Cohen J, Statistical Power Analysis for the Behavioral Sciences 1988 2nd EdNew YorkLawrence Erlbaum:452  [Google Scholar]

[20]. Vasques AC, Novaes FS, de Oliveira Mda S, Souza JR, Yamanaka A, Pareja JC, TyG index performs better than HOMA in a Brazilian population: A hyperglycemic clamp validated study Diabetes Res Clin Pract 2011 93(3):e98-100.10.1016/j.diabres.2011.05.03021665314  [Google Scholar]  [CrossRef]  [PubMed]

[21]. Lachin JM, Marks JW, Schoenfield LJ, the NCGS Protocol Committee, The National Cooperative Gallstone Study groupDesign and methodological considerations in the National Cooperative Gallstone Study. A multicenter clinical trial Control Clin Trials 1981 2(3):177-229.10.1016/0197-2456(81)90012-X  [Google Scholar]  [CrossRef]

[22]. Cole TJ, Bellizzi MC, Flegal KM, Dietz WH, Establishing a standard definition for child overweight and obesity worldwide: International survey BMJ 2000 320:124010.1136/bmj.320.7244.124010797032  [Google Scholar]  [CrossRef]  [PubMed]

[23]. 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(4):e500-03.10.1542/peds.2004-192115741351  [Google Scholar]  [CrossRef]  [PubMed]

[24]. Moon S, Park JS, Ahn Y, The cut-off values of triglycerides and glucose index for metabolic syndrome in American and Korean adolescents J Korean Med Sci 2017 32(3):427-33.10.3346/jkms.2017.32.3.42728145645  [Google Scholar]  [CrossRef]  [PubMed]

[25]. Targher G, Alberiche M, Zenere MB, Bonadonna RC, Muggeo M, Bonora E, Cigarette smoking and insulin resistance in patients with noninsulin-dependent diabetes mellitus J Clin Endocrinol Metab 1997 82(17):3619-24.10.1210/jc.82.11.3619  [Google Scholar]  [CrossRef]

[26]. Bonora E, Targher G, Alberiche M, Bonadonna RC, Zenere MB, Saggiani F, Muggeo M, Intracellular partition of plasma glucose disposal in hypertensive and normotensive subjects with type 2 diabetes mellitus J Clin Endocrinol Metab 2001 86(5):2073-79.10.1210/jc.86.5.207311344209  [Google Scholar]  [CrossRef]  [PubMed]

[27]. Er LK, Wu S, Chou HH, Hsu LA, Teng MS, Sun YC, Ko YL, Triglyceride glucose-body mass index is a simple and clinically useful surrogate marker for insulin resistance in nondiabetic individuals PLoS ONE 2016 11(3):e014973110.1371/journal.pone.014973126930652  [Google Scholar]  [CrossRef]  [PubMed]

[28]. Kodama K, Tojjar D, Yamada S, Toda K, Patel CJ, Butte AJ, Ethnic differences in the relationship between insulin sensitivity and insulin response: A systematic review and meta-analysis Diabetes Care 2013 36(6):1789-96.10.2337/dc12-123523704681  [Google Scholar]  [CrossRef]  [PubMed]

[29]. Calcaterra V, Montalbano C, de Silvestri A, Pelizzo G, Regalbuto C, Paganelli V, Triglyceride Glucose Index as a Surrogate Measure of Insulin Sensitivity in a Caucasian Pediatric Population J Clin Res Pediatr Endocrinol 2019 2019:31088046  [Google Scholar]

[30]. Dikaiakou E, Vlachopapadopoulou EA, Paschou SA, Athanasouli F, Panagiotopoulos I, Kafetzi M, Triglycerides-glucose (TyG) index is a sensitive marker of insulin resistance in Greek children and adolescent Endocrine 2020 2020:01-07.10.1007/s12020-020-02374-632557329  [Google Scholar]  [CrossRef]  [PubMed]

[31]. Chamroonkiadtikun P, Ananchaisarp T, Wanichanon W, The triglyceride-glucose index, a predictor of type 2 diabetes development: A retrospective cohort study Prim Care Diabetes 2020 14(2):161-67.10.1016/j.pcd.2019.08.00431466834  [Google Scholar]  [CrossRef]  [PubMed]

[32]. Kang B, Yang Y, Lee EY, Yang HK, Kim HS, Triglycerides/glucose index is a useful surrogate marker of insulin resistance among adolescents Int J Obes 2017 41(5):789-92.10.1038/ijo.2017.1428104918  [Google Scholar]  [CrossRef]  [PubMed]

[33]. Locateli JC, Lopes WA, Simões CF, De Oliveira GH, Oltramar K, Bim RH, Triglyceride/glucose index is a reliable alternative marker for insulin resistance in South American overweight & obese children & adolescents J Pediatr Endocrinol Metab 2019 32(10):1163-70.10.1515/jpem-2019-003731472061  [Google Scholar]  [CrossRef]  [PubMed]

[34]. Sarria A, Moreno LA, Garcia-Llop LA, Fleta J, Morellon MP, Bueno M, Body mass index, triceps skinfold and waist circumference in screening for adiposity in male children and adolescents Acta Paediatr 2001 90(4):387-92.10.1080/08035250175012619511332928  [Google Scholar]  [CrossRef]  [PubMed]

[35]. Chao YP, Lai YF, Kao TW, Peng TC, Lin YY, Shih MT, Mid-arm muscle circumference as a surrogate in predicting insulin resistance in non-obese elderly individuals Oncotarget 2017 8(45):79775-84.10.18632/oncotarget.1934029108358  [Google Scholar]  [CrossRef]  [PubMed]

[36]. Abe Y, Okada T, Okuma H, Kazama M, Yonezama R, Saito E, Abdominal Obesity, Insulin Resistance, and Very Low-Density Lipoprotein Subclass Profile in Japanese School Children J Child Obes 2016 1(3):01-06.10.21767/2572-5394.100003  [Google Scholar]  [CrossRef]