Triglyceride to HDL-C Ratio is Associated with Insulin Resistance in Overweight and Obese Children

The purpose of this study was to investigate the usefulness of triglyceride to hdl-c ratio (TG:HDL-C) as an insulin resistance (IR) marker for overweight and obese children. A total of 271 blood samples of obese and overweight children aged 9–16 years were analysed for fasting glucose, lipids and insulin. Children were divided into IR and non-insulin resistance, using homeostasis model assessment (HOMA). The children were then stratified by tertiles of TG: HDL-C ratio. The strength between TG:HDL-C ratio and other parameters of IR were quantified using Pearson correlation coefficient (r). Odds ratio was estimated using multiple logistic regression adjusted for age, gender, pubertal stages and IR potential risk factors. Children with IR had significantly higher TG:HDL-C ratio (2.48) (p = 0.01). TG:HDL-C ratio was significantly correlated with HOMA-IR (r = 0.104, p < 0.005) and waist circumference (r = 0.134, p < 0.001). Increasing tertiles of TG:HDL-C ratio showed significant increase in mean insulin level (p = 0.03), HOMA-IR (p = 0.04) and significantly higher number of children with acanthosis nigricans and metabolic syndrome. The odds of having IR was about 2.5 times higher (OR = 2.47; 95% CI 1.23, 4.95; p = 0.01) for those in the highest tertiles of TG:HDL-C ratio. Hence, TG:HDL-C may be a useful tool to identify high risk individuals.

Scientific RepoRts | 7:40055 | DOI: 10.1038/srep40055 In individuals with insulin resistance, TG levels increased while HDL-C levels decreased 11 . Triglyceride to high-density lipoprotein cholesterol (TG:HDL-C) ratio has been proposed to be an alternative tool for gauging insulin resistance 11 . A higher ratio would represent a poorer health status because there is a large amount of circulating fats in the blood stream and/or a low amount of healthy cholesterol. A TG:HDL-C ratio of ≥ 3 has been shown to be closely correlated to insulin resistance 11 . The TG:HDL-C ratio has also been shown to be an independent risk factor for coronary heart disease (CHD) among Iranian men 12 .
In Malaysia, it was estimated that 3.9% of children under the age of 18 years were obese 13 . A lifestyle school-based intervention programme was therefore developed specifically for overweight and obese children to address this problem. Named as "My Body is Fit and Fabulous (MyBFF@school), the programme was piloted between January and December 2014 in selected public schools in Putrajaya, the Administrative Capital City of Malaysia. The objective of this study was to evaluate the association between TG:HDL-C ratio and IR among the overweight and obese school children who participated in MyBFF@school programme.

Results
From a total of 425 children who participated, blood samples were obtained from 274; 86 children aged 9-12 years old and 188 adolescents aged 13-16 years old. However, only 271 children had complete data set of lipid values and only their data (n = 271) were used for the analysis (Fig. 1). As shown in Table 1 median age was 14 years old with girls were slightly older than boys (13 years versus 14 years). Gender was equally distributed with 49.1% boys and 50.9% girls. Based on BMI z-score, 22.5% were overweight and 77.5% were obese. A total of 135 (53.1%) children had acanthosis nigricans while 28 (10.3%) children had metabolic syndrome. Table 2 described the general anthropometric and biochemical measurements of the 271 children. There were 126 (49%) children found to have IR as defined by HOMA-IR ≥ 4.0 ( Table 3). The mean TG:HDL-C ratio was significantly higher among children with IR compared to the non-IR (2.48 versus 1.73, p = 0.01). Likewise, BMI z-score, waist circumference and body fat percentage were significantly higher in the IR group. Table 4 summarised the Pearson correlations coefficient (r) between TG:HDL-C ratio and other parameters of IR. TG:HDL-C was significantly correlated with HOMA-IR (r = 0.104, p < 0.005) and waist circumference (r = 0.134, p < 0.001). In addition, HOMA-IR was significantly correlated with BMI z-score (r = 0.202, p < 0.001) and waist circumference (r = 0.258, p < 0.001). Table 5 summarised the anthropometric, clinical and metabolic parameters when stratified according to tertiles of TG:HDL-C ratio. There was no significant difference in age, gender and pubertal status across the tertiles. Highest number of children with acanthosis nigricans (n = 58, 43%) and metabolic syndrome (n = 18, 64.2%) were seen at the third tertiles of TG:HDL-C ratio. However, there was no significant difference in blood pressure status across the tertiles. The BMI z-score and waist circumference rose significantly across TG:HDL-C ratio tertiles (p = 0.03 and p < 0.01 respectively). Similarly, there was significant increase in mean insulin level (p = 0.03) and HOMA-IR (p = 0.04) with increasing tertiles, implying worsening of IR. When based on multiple linear regression analysis using HOMA-IR as the dependent variable and adjusting for age, gender, pubertal stages, waist circumference and BMI-z score, children in the third tertiles of TG:HDL-C ratio posed 2.5 times higher risk of developing IR compared to those in the lower tertiles (OR = 2.47; 95% CI 1.23, 4.95; p = 0.01) ( Table 6).

Discussions
McLaughlin et al. 11 was the first to demonstrate the clinical utility of TG:HDL-C ratio in identifying generally healthy Caucasians with IR. It was shown that the TG:HDL-C ratio was as closely associated with specific measure of insulin-mediated glucose disposal as was the fasting plasma insulin concentration, a surrogate estimate of insulin action that has been widely used to study the relation between IR and various clinical syndromes 11 . However, several studies have shown that the association may be ethnicity dependent. While Hirscheler et al. 6 observed significant association between TG:HDL-C ratio and IR among their indigenous Argentinean children, Gianini et al. 7 showed that the association was only significant among the white obese children but not in the Hispanic or African American children. Similarly, poor association between TG:HDL-C ratio and IR was also reported among adult African-Americans 8 . On the other hand, TG:HDL-C ratio was found to be strongly associated with IR among Korean population 9 and obese South East Asian Immigrant youths 9 .
In the present study involving overweight and obese children of 9-16 years old, TG:HDL-C ratio was found to be significantly correlated with HOMA-IR (p < 0.05) and was significantly higher (p = 0.01) among children with HOMA-IR ≥ 4. In addition, TG:HDL-C ratio was also found to be significantly correlated with waist circumference (p < 0.001). A number of studies have reported significant predictive power of waist circumference as anthropometric surrogate marker in predicting insulin resistance among children [14][15][16][17] . Hence, indicating the practicality of introducing TG:HDL-C ratio as a relatively simple biochemical marker for IR.
In a study using a large and nationally representative sample (n = 2652) among non-diabetic adults of three major racial/ethnic subpopulations in the United States, Li et al. 18 found that TG/HDL-C ratio was the best predictor of hyperinsulinimia than that of triglyceride or HDL-C alone for the prediction of hyperinsulinemia. In addition, triglycerides and HDL-cholesterol are the two important lipid measures in the diagnosis of metabolic syndrome. A combined lipid ratio may better reflect the overall interaction between lipid/lipoprotein fractions, and therefore associations with insulin resistance 19 .
When stratified according to TG:HDL-C tertiles, we found that mean insulin level and HOMA-IR significantly increased across tertiles. In addition, we also found that significantly higher number of children at the third tertile of TG:HDL-C ratio had acanthosis nigricans and metabolic syndrome. This indicates that besides having significant correlation to HOMA-IR index, TG:HDL-C ratio also have significant association with physical findings (which is very practical) related to the clinical risk of IR. Age, median (p25, p75) 13 (11,14) 14 (13,14) 0.004 14 (11,14) Pubertal status Moreover, children in the highest tertile of TG:HDL-C ratio posed 2.5 times higher risk of developing IR compared to those in the two lower tertiles. Similar trend was also reported in a study involving indigenous Argentinian children 6 . Indeed, Pacifico et al. 20 also demonstrated that in a study of Caucasian children highest TG:HDL-C ratio showed a 1.8-to 3.8-fold increased risk of central obesity, insulin resistance, high-sensitivity c-reactive protein, non alcoholic fatty liver disease, metabolic syndrome, and increased carotid artery intima-media thickness.
As shown in many studies, hypertriglyceridemia and decreased HDL-C are the hallmarks of dyslipidemia which is a characteristic of insulin resistance and T2D. The Treatment Options for Type 2 Diabetes in Adolescents and Youth (TODAY) study showed that 79.8% of T2D youth had a low HDL-C and 10.2% had high triglycerides within a few months of diagnosis 21 and the SEARCH for Diabetes in Youth study found that 73% of 2096 US youth with T2D of longer duration had lower HDL and 60-65% had hypertriglyceridemia 22 . Therefore, the clinical utility of measuring TG and HDL-C extended beyond identifying patients with IR. Not surprising, we also found that BMI z-scores and waist circumference rose significantly across TG:HDL-C ratio tertiles. Association between obesity and morbidity, independent of insulin resistance and diabetes has been shown in many studies [23][24][25][26][27] . In a prospective study involving general population in China, logistic regression model showed that TG:HDL-C ratio could independently predict future diabetes mellitus 28 .
In conclusion, this study has shown that TG:HDL-C ratio is significantly associated with IR in overweight and obese Malay children. The TG:HDL-C ratio is an inexpensive predictor of IR and may be a useful tool to identify high risk individuals for early intervention and thereby prevent or delay the development of IR-associated diseases such as T2D and hypertension. A larger cohort study involving children from different ethnic groups should be carried out to confirm our current findings.

Methods
Study design and population. A total of 425 obese and overweight children aged 9 to 16 years old with body mass index (BMI) z-score exceeded 2 or 3 standard deviation according to WHO BMI chart were recruited for the study (Fig. 1). Children came from a total of 6 randomly selected public schools in Putrajaya, the Administrative Capital City of Malaysia. Exclusion criteria included physical or mental disability that would prevent children from participating in moderate-to-vigorous intensity physical activity and children with known co-morbidities such as T2D, hypertension and cardiovascular disease. Ethical approval was obtained from the Medical Research and Ethics Committee (MREC) Ministry of Health Malaysia. Written informed consent was obtained from parent or guardian and all participating children were also required to sign an assent form. All testing was performed in accordance with the approved guidelines.  Health and physical examination. Prior to the study visit, these children were asked to fast overnight for at least 8 hours. All anthropometric measurements were performed by trained personnel, and health examinations were performed by pediatricians. Standing height was measured without shoes to the nearest 0.1 cm using calibrated stadiometer (Seca 217, Germany). Body weight and body fat mass were measured in light clothing without shoes and socks to the nearest 0.1 kg using a pre-calibrated body impedance analyzer (InBody 720, Korea). Waist circumference was measured twice to the nearest 0.1 cm over the skin midway between the tenth rib and the iliac crest at the end of normal expiration, using a non-extensible tape (Seca 201, Germany) and the mean was recorded. Two readings of blood pressure was measured after 5 minutes of resting using a mercury sphygmomanometer (Accoson, UK) in a seated position with the arm supported at heart level, and the mean was recorded. Pubertal status was assessed (self-administered) using Tanner staging scale 29,30 and participants were also examined for the presence of acanthosis nigricans over the neck 4 .

Biochemical parameters. Venipuncture was performed by experienced nurses or doctors. Blood samples
were transported cold to the central laboratory at the Institute for Medical Research within 2 hours of collection and processed on the same day. Aliquots of serum/plasma samples were kept at − 20 °C prior to analysis. HbA1c level was determined by cationic exchanged high performance liquid chromatography (Adams A1c HA-8160, Arkray Inc, Japan) and followed the National Glycohemoglobin Standardization Programme Guidelines. Fasting plasma glucose, triglycerides, total cholesterol, HDL and LDL were analyzed using an automated analyzer (Dirui CS-400, China) with reagents purchased from Randox Laboratories (Antrim, UK). Fasting insulin concentration was measured using an automated enzyme immunoassay analyzer (TOSOH AIA-360, Japan). Interassay coefficient of variability (CV) for insulin at 8.7, 44.4 and 143.2 μ U/ml was 2.5%, 2.6% and 2.4% respectively.

Definition of Measures.
Overweight or obese was defined as BMI z-score exceeded 2 or 3 standard deviation respectively for their age and sex, according to WHO BMI chart 31 . Tanner staging was assessed by showing a standardized Tanner staging pictures to the child where stage 1 external genitalia development and breast development for boys and girls respectively was classified as pre-pubertal, while stage 2 and above was defined as pubertal. Determination of acanthosis nigricans (AN) was based on Burke's quantitative dichotomous score 4 .  Comparison of means among groups was conducted using independent group t-tests or analysis of variance (ANOVA) where applicable and the Mann-Whitney test was used for not normally distributed variables. Categorical comparisons were made using Chi-square test. Multiple logistic regression analyses were performed to examine the relationship between insulin resistance and TG:HDL-C adjusted for age, gender, pubertal stages, waist circumference and BMI-zscore. All statistical tests were performed at 5% significance level.  Table 6. Associated factors of IR by simple and multiple logistic regression among obese and overweight students. b = regression coefficients, OR = odds ratio, CI = confidence interval. Multivariate regression analysis was applied. Hosmer Lemeshow Test (chi-square = 5.56, p-value = 0.696) and classification table (overall correctly classified percentage = 67.5%).