Comparison of Ten Surrogate Insulin Resistance and Obesity Markers to Identify Metabolic Syndrome in Mexican Adults

Metabolic syndrome (MetS) is a group of clinical traits directly linked to type 2 diabetes mellitus and cardiovascular diseases, whose prevalence has been rising nationally and internationally. We aimed to evaluate ten known and novel surrogate markers of insulin resistance and obesity to identify MetS in Mexican adults. The present cross-sectional study analyzed 10575 participants from ENSANUT-2018. The diagnosis of MetS was based on the Adult Treatment Panel III (ATP III) criteria and International Diabetes Federation (IDF) criteria, stratified by sex and age group. According to ATP III, the best biomarker was the metabolic score for insulin resistance (METS-IR) in men aged 20–39 and 40–59 years and lipid accumulation product (LAP) in those aged ≥60 years. The best biomarker was LAP in women aged 20–39 and triglyceride–glucose index (TyG) in those aged 40–59 and ≥60 years. Using the IDF criteria, the best biomarker was LAP in men of all ages. TyG gave the best results in women of all ages. The best biomarker for diagnosis of MetS in Mexican adults depends on the criteria, including sex and age group. LAP and TyG are easy to obtain, inexpensive, and especially useful at the primary care level.


Introduction
Metabolic syndrome (MetS) is a set of related clinical characteristics that include central obesity, hypertension, hyperglycemia, and atherogenic dyslipidemia (low HDL cholesterol and hypertriglyceridemia).It is a clinical entity directly related to chronic diseases that cause an increase in morbidity, mainly type 2 diabetes mellitus (DMT2) and cardiovascular diseases (CVDs) [1][2][3].Cardiovascular diseases are one of the leading causes of death worldwide [4], and it is estimated that just over 10.5% of the world's population suffers from DMT2 [5].
In Mexico, for years, heart diseases, especially ischemic diseases, have been among the leading causes of death [6].In fact, deaths related to these diseases increased from 9.2 in 2011 to 17.9 per 10,000 inhabitants in 2021 [7].According to data from 2018, the prevalence of MetS has increased by 16.75% (IC95%, according to Adult Treatment Panel III criteria (ATP III)) and 41.88% (IC95%, under the International Diabetes Federation (IDF) criteria) [8].
The two conditions that have been confirmed as closely related to MetS are obesity and insulin resistance (IR).The relationship between increased subcutaneous adipose tissue and the development of a proinflammatory state has been studied.In addition, the contribution of increased fat tissue to the development of IR is known [9][10][11][12].
There are various sets of criteria for the definition of MetS.The best accepted and used are the IDF [13] and ATP III criteria [14].These criteria are applicable at the secondary and tertiary care levels.For this reason, other indices or markers have been explored to diagnose MetS at the primary care level, where some fasting biochemical parameters are not available to make the diagnosis.The traditional indicators of obesity, such as body mass index (BMI) and waist circumference (WC), have been used to diagnose MetS but have shown poor diagnostic performance with sensitivity or specificity below 75% [15], and they are also of limited utility for identifying subclinical conditions [16].Recently, different surrogate markers for a simple, affordable, and inexpensive MetS diagnosis have been studied.Some of these markers have been analyzed in Asian countries.Studies in China [17] and Bangladesh [18] found TyG to be one of the best surrogate markers for determining MetS.However, they also found LAP and the visceral adiposity index (VAI) to be efficacious in middle-aged and elderly Chinese [18].The results were similar to those obtained in the Korean population [19].TyG was a better marker of MetS in India [20]; however, in Taiwan, it was determined that the VAI as a marker was more accurate but difficult to access [15].
One US study proposed a new measurement index, the body roundness index (BRI), to assess individual body fat distribution and overall adiposity.It was introduced as an alternative to traditional measures of obesity, such as BMI, which may not capture variations in body fat distribution.BRI takes into account both WC and hip circumference (HC) to provide a more comprehensive assessment of body shape, and it is a good predictor of body fat percentage [21].On the other hand, a Colombian study that evaluated diabetes predictors such as BMI, WC, waist-to-height ratio (WtHR), triglyceride-to-glucose fasting related to BMI (TyG-BMI), triglyceride-to-glucose fasting related to WC (TyG-WC), and triglyceride-to-glucose fasting related to WtHR (TyG-WtHR) concluded that TyG was the best indicator in the adult population [22].
A cross-sectional study conducted on adults from the center of Mexico considered only the triglyceride/high-density cholesterol (TG/HDL) ratio [23].No other studies have been conducted in Mexico to identify the best markers of insulin resistance and obesity to improve the diagnosis of metabolic syndrome in adults, considering fasting blood chemistry parameters at the primary care level.Therefore, the present project aims to evaluate (known and novel) surrogate markers of insulin resistance and obesity to diagnose MetS in Mexican adults.

Study Population
This analysis was carried out based on the information obtained from the National Health and Nutrition Survey (ENSANUT, for its acronym in Spanish) 2018 and was designed to quantify the frequency and distribution of health and nutrition conditions of the Mexican population.The ENSANUT 2018 had a transversal, probabilistic, multi-stage, and cluster sampling design, with regional representation (north, center, Mexico City, and south) and considered urban (population ≥ 2500 inhabitants) and rural (population < 2500 inhabitants) locality.Information was obtained from 50654 households distributed in the 32 states of the country.When considering selecting individuals (one for each age group per household), 43,078 adults aged 20 or older were interviewed.A detailed description of the sampling procedures and survey methodology has already been published [24].
For this study, the inclusion criteria were Mexican men and women over 20 years of age.The exclusion criteria were serum glucose concentrations of less than 70 mg/dL or greater than 500 mg/dL, serum triglycerides greater than 1200 mg/dL, and a height of less than 1.30 m, while the elimination criteria were missing information on body composition and biochemical indices.

Definition of Metabolic Syndrome
The diagnosis of MetS was based on the IDF [13] and ATP III [14] definitions.The IDF diagnoses MetS when patients present a waist circumference of ≥90 cm in men or ≥80 cm in women, plus two of the following conditions: triglycerides ≥ 150 mg/dL or medication treatment to control triglycerides; high-density cholesterol < 40 mg/dL in men and < 50 mg/dL in women; blood pressure ≥ 130/85 mmHg or previous diagnosis of hypertension; fasting glucose ≥100 mg/dL or previous diagnosis of diabetes.ATP III establishes the presence of MetS when three or more of the following findings occur: waist circumference ≥ 102 cm in men or ≥88 cm in women; fasting glucose ≥ 110 mg/dL or previous diagnosis of diabetes; triglycerides ≥ 150 mg/dL; blood pressure ≥ 130/85 mmHg or previous diagnosis of hypertension; high-density lipoprotein cholesterol < 40 mg/dL in men and <50 mg/dL in women.The diagnostic criteria of ATP III and the IDF are shown in Table A1.

Surrogate Markers of Insulin Resistance and Obesity
Surrogate markers of insulin resistance and obesity include body shape index (ABSI), BRI, LAP, the metabolic score for insulin resistance (METS-IR), the metabolic score for visceral fat (METS-VF), single-point insulin sensitivity estimator (SPISE), TG/HDL, TyG, -1 being the maximum value reached within the ROC curve.Confidence intervals for sensitivity, specificity, and positive and negative predictive values were computed with the "exact" Clopper-Pearson confidence intervals.Using the Bootstrap method, we implemented a simple code in the R software 4.4.1 to calculate the confidence intervals of the Youden index.
To compare the areas under the curve (AUC) between the indices, we used the bootstrap test for two correlated ROC curves implemented in the "pROC" library of the R software 4.4.1 [32].All statistical analyses were performed by using IBM SPSS Statistics 23.0 and R statistical software 4.4.1, whereas ROC curves were elaborated in IBM SPSS Statistics 23.0.Results were considered significant at p < 0.05 (2-sided).

Characteristics of the Study Population
The ENSANUT-2018 database reported two databases, the anthropometric parameters database with 33,818 and the biochemical database with 13,220.To carry out the calculations of the different markers, these bases were joined; applying the elimination criteria, we have 13,101; of these, when applying the elimination criteria, there were 10,575.
Two databases were created from the 10,575 database by sex: one with 6041 women and another with 4534 men.Each database was subdivided within it by age group (Figure 1).The database contained the following variables: age, weight, height, systolic and diastolic blood pressure, WC, total cholesterol, HDL cholesterol, triglycerides, glucose, and the dichotomous variables of treatment and diagnosis for diabetes, hypercholesterolemia, hypertriglyceridemia, and hypertension, respectively.The ATP III and IDF criteria were calculated using the variables from each database.The variables from the database were used to calculate the ten markers.Finally, each biomarker was evaluated for its discriminatory ability for each metabolic syndrome criterion.The sociodemographic, anthropometric, and biochemical characteristics of the study population are presented in Table 1.The study population consisted of 10,575 participants, most of whom (57%) were women.The highest proportion of the study population was in the 40-59-year age group (38.7%).Concerning nutritional status measured by body mass index, the majority were obese (52.7%), and women had more significant central obesity measured by waist circumference in relation to the ATP III and IDF criteria, respectively (73.6% and 88.9%).Systolic blood pressure was above 135 mmHg in 34.6% of the population, while diastolic blood pressure was elevated in 20.7%.In terms of biochemical parameters, 33.6% of the participants had elevated cholesterol levels above 200 mg/dL; 40.9% and 76.5% of men and women had HDL cholesterol below 40 and 50 mg/dL in men and women, respectively; and 58.6% had triglycerides above 150 mg/dL.Based on the IDF or ATP III criteria, blood glucose concentrations above 100 mg/dL and 110 mg/dL were obtained in 30.4% and 19.1% of participants, respectively.
Of the total population of women, 52.6% and 60.0% presented MetS according to the ATP III and IDF criteria, respectively, while of the total population of men, it was 39.9% and 53.6%, respectively.The sociodemographic, anthropometric, and biochemical characteristics of the study population are presented in Table 1.The study population consisted of 10,575 participants, most of whom (57%) were women.The highest proportion of the study population was in the 40-59-year age group (38.7%).Concerning nutritional status measured by body mass index, the majority were obese (52.7%), and women had more significant central obesity measured by waist circumference in relation to the ATP III and IDF criteria, respectively (73.6% and 88.9%).Systolic blood pressure was above 135 mmHg in 34.6% of the population, while diastolic blood pressure was elevated in 20.7%.In terms of biochemical parameters, 33.6% of the participants had elevated cholesterol levels above 200 mg/dL; 40.9% and 76.5% of men and women had HDL cholesterol below 40 and 50 mg/dL in men and women, respectively; and 58.6% had triglycerides above 150 mg/dL.Based on the IDF or ATP III criteria, blood glucose concentrations above 100 mg/dL and 110 mg/dL were obtained in 30.4% and 19.1% of participants, respectively.Of the total population of women, 52.6% and 60.0% presented MetS according to the ATP III and IDF criteria, respectively, while of the total population of men, it was 39.9% and 53.6%, respectively.In men aged 20-39, we found that METS-IR had a higher Youden index (AUC = 0.884 (95% CI 0.868-0.900);Se = 88.5 (95% CI 85.4-91.1);Sp = 71.3(95% CI 68.6-73.4);YI = 59.8 (95% CI 56.0-63.6);and cut-off value ≥ 45.29).The following markers had lower diagnostic accuracy than METS-IR: SPISE (AUC = 0.875 (95% CI 0.858-0.892);Se = 67.8(95% CI 63.5-71.8);Sp = 87.5 (95% CI 85.8-89.5);YI = 55.3 (95% CI 51.0-59.9);and cut-off value ≤4.13) and LAP (AUC = 0.855 (95% CI 0.837-0.873);Se = 85.5 (95% CI 82.2-88.5);Sp = 69.1 (95% CI 66.4-71.7);YI = 54.6 (95% CI 50.6-58.6);and cut-off value ≥ 65.22).See Figure 2 and Table 2.We found significant differences in the AUC for the diagnosis of MetS according to the ATP III criteria between the METS-IR and SPISE (p < 0.001) and between SPISE and LAP (p = 0.001).We chose the METS-IR to diagnose MetS according to the ATP III criteria in men aged 20-39 because it has a greater discriminating capacity.

Indices for the Diagnosis of MetS According to the IDF Criteria
Figures 7-12 compare ROC curves, and Tables 8-13 report the accuracy measures for the diagnostic of MetS according to the IDF criteria using the same markers as for the ATP-III criteria.

Discussion
The present study investigated the value of SPISE, LAP, VAI, TyG, the TG/HDL ratio, METS-IR, METS -VF, and VAT, known and novel surrogate markers of insulin resistance and obesity, in identifying MetS according to different criteria in three age groups.Our results show that LAP and TyG have reliable predictive accuracy for diagnosing MetS in the ATP-III and IDF criteria.METS-IR had higher accuracy than LAP for diagnosing MetS according to the ATP criteria in men aged 20-39 and 40-59.Still, LAP could be used at the primary care level because it has acceptable accuracy.With the IDF criteria, LAP and TyG were better for diagnosing MetS in men and women.This study is the first report based on the analysis and comparison of four novel markers (METS-IR, METS-VF, VAT, and SPISE) for predicting MetS using different criteria in adult Mexicans in particular.
In an Indian adult population, LAP, BMI, and WC were studied as predictors of MetS according to the ATP-III criteria modified for the Asian Population [20].They reported that LAP was the best predictor (AUC = 0.901 (95% CI 0.85-0.95),Se = 76.4,Sp = 91.1,YI = 67.0,and cut-off value ≥ 38.05), but they did not stratify by age or sex.A study in middle-aged and elderly Chinese analyzed LAP, VAI, and TyG to predict MetS [34], and they found that LAP was the best predictor (AUC = 0.855 (95% CI 0.831-0.878),Se = 73.9,Sp = 79.7,YI = 53.6, and cut-off value ≥ 31.46).They also did not stratify by age or sex.A study of the Peruvian adult population examined BMI, LAP, VAI, and TyG for the diagnosis of MetS [35].They reported that LAP was the best for diagnosing MetS in men (AUC = 0.929 (95% CI 0.907-0.952),Se = 91.6,Sp = 84.5,YI = 71.3, and cut-off value ≥ 59.85) and women (AUC = 0.950 (95% CI 0.940-0.960),Se = 92.4,Sp = 86.4,YI = 78.8, and cut-off value ≥ 53.06), and they did not stratify by age.In a population of Spanish adults, they analyzed different anthropometric indices (BMI, the Ponderal index, AVI, BAI, VAI, BRI, CI, Cholindex, WHR, and WHtR) and atherogenic indices (LAP; CT/HDL, LDL/HDL, and TG/HDL ratios; and non-HDL/HDL) [36].They found that LAP was one of the best predictors of MetS in men (AUC = 0.946 (95% CI 0.943-0.950),Se = 96.0;Sp = 70.1,YI = 66.6, and cut-off value ≥ 18.40) and women (AUC = 0.942 (95% CI 0.935-0.950),Se = 95.0,Sp = 75.7,YI = 71.0,and cut-off value ≥ 36.04).Our results for the diagnostic capacity of LAP in men were not better than those reported in the populations of India, China, Peru, and Spain.By contrast, LAP in women had a better discriminative capacity than in the populations of India, China, and Spain but worse than in the population of Peru.
In the adult Mexican population, only the TG/HDL ratio (AUC = 0.853(95% CI 0.831-0.872),Se = 79.6,Sp = 76.4,YI = 55.95, and cut-off value ≥ 3.46) was studied to identify subjects with MetS in the Mexican population.Still, they did not stratify by sex or age [23].In a population of Spanish adults [36], they also found that the TG/HDL ratio was one of the best predictors of MetS in men (AUC = 0.949 (95% CI 0.946-0.952),Se = 95.5, Sp = 81.8,YI = 77.7,and cut-off value ≥ 2.96) and women (AUC = 0.923 (95% CI 0.914-0.932),Se = 92.2,Sp = 70.6,YI = 63.0, and cut-off value ≥ 1.75).Our results showed that the discriminant capacity of the TG/HDL ratio using AUC and the Youden index was lower in men and higher in women, compared with those of Baez-Duarte et al. [36].On the other hand, in the population of Spanish adults [36], the discriminant capacity of the TG/HDL ratio was better in men and women than in our study.However, the TG/HDL ratio could not be used at the primary care level, in contrast to LAP and TyG.
Studies using the ABSI index in the populations of Taiwan [15] and Spain [36] found very poor diagnostic performance, similar results to those found in men and women of different age groups in our study.
The ABSI in the studies consulted [15,36] has not been shown to be a good predictive indicator of MetS; similarly, in our study, despite the AUC of around 0.600, globally, it was the worst-performing indicator.
We have yet to find a study assessing the relationship between the METS-IR, SPISE, METS-VF, and VAT indices with MetS according to the ATP-III criteria.Therefore, we cannot make any comparisons with our results, which were acceptable to very good.
Several other studies analyzed the indices using the IDF criteria [34,37,38].In a population of middle-aged and elderly Chinese [39], Li et al. studied LAP, VAI, and TyG, and they found that LAP (AUC = 0.865 (95% CI 0.841-0.889),Se = 73.2;Sp = 84.5,YI = 56.8, and cut-off value ≥ 37.99) was superior to VAI and TyG.They did not stratify by age or sex.In our results, LAP showed better discrimination capability in men and women aged 20-39 than in the study of Li et al.; in the other age groups, the results were similar.Our results on the diagnostic capacity of TyG in men and women of any age group showed that it was better than that in Li et al. (AUC = 0.746 (95% CI 0.712-0.779);Se = 70.2;Sp = 71.8,YI = 42.1; and cut-off value ≥ 8.697).
In a Chinese elderly population, they studied the BMI, WHtR, TG/HDL-C ratio, LAP, and VAI [39] and found that LAP was a better index in men (AUC = 0.897 (95% CI 0.885-0.907),Se = 85.09,Sp = 79.31,YI = 64.4,and cut-off value ≥ 26.35) and women (AUC = 0.875 (95% CI 0.864-0.886),Se = 79.17,Sp = 80.69, YI = 59.8; and cut-off value ≥ 31.4).LAP had a better diagnostic capacity in women and a similar one in men compared to the previous study.Our results on the diagnostic capacity of the TG/HDL index in women of any age group and men aged 20-39 showed that it was better than that in Gu. et  In an adult population of Iranis analyzed BMI, ABSI, BRI, and VAI [38], it was reported that VAI was the best index in men (AUC = 0.824 (95% CI 0.812-0.836),Se = 80.08, Sp = 70.0,YI = 51.20, and cut-off value ≥ 4.12) and women (AUC = 0.866 (95% CI 0.855-0.877),Se = 83.1,Sp = 70.0,YI = 58.00,and cut-off value ≥ 4.28).In this work, they did not analyze LAP and TyG.Our results for men and women aged 20-39 years showed that VAI had a better diagnostic capacity than the results of Baveicy et al.; for men and women over 40 years, the discrimination capacity of VAI was similar.Our results, as well as those of Baveicy et al., showed that ABSI performed poorly in MetS diagnosis.
LAP and TyG have shown good accuracy in various populations but have also shown variability in the selection of cut-off values for the diagnosis of MetS, even when the same criteria are used.As we have been able to review, in most LAP and TyG studies, they are useful for predicting MetS.Considering that these indices are easy, fast to calculate, and inexpensive, they could be an alternative to MetS screening at the first level of attention, even more so in countries where access to certain biochemical tests is limited.
The LAP is a marker that combines measures of abdominal obesity and triglyceride levels to provide insights into lipid overaccumulation and metabolic dysfunction [26].This is explained by the fact that visceral fat has a high lipolytic capacity, so it is able to release fatty acids into the circulation, which in turn can be taken up by the liver, which uses them to form glucose.This means that the greater the waist circumference, the higher the triglyceridemia, hyperglycemia, and insulin resistance.TyG is an indicator of insulin resistance because when you have a high concentration of glucose, insulin secretion is stimulated, which is a physiological mechanism, but if this hyperglycemia is sustained, the pancreas loses the ability to produce insulin in the necessary amounts [40], which makes it especially useful in the diagnosis of diabetes mellitus.
Lipid metabolism is influenced by the circadian clock, which is modulated by light exposure and 24-h dietary patterns.Research on circadian rhythms of triglycerides consistently reveals a peak during the night [41], often aligning closely with the melatonin phase, although the strength of this rhythm varies among individuals [41].However, nocturnal eating habits can alter nocturnal triglyceride levels, potentially shifting toward a daytime pattern similar to that of HDL and LDL cholesterol, which are less affected by nocturnal meals [42].Additionally, administering melatonin before the evening meal may enhance postprandial triglyceride levels [43].
Triglycerides' sensitivity as a biomarker can be attributed to circadian factors (note that metabolic syndrome can even be considered circadian syndrome [44,45]).Nocturnal eating habits and compromised metabolic health are closely associated with an evening chronotype [46] and exposure to evening light [47], which may also contribute to the development of metabolic syndrome [48].Furthermore, lipid measurements such as triglycerides or HDL sampled at a fixed morning time may vary depending on individual circadian phase differences, such as melatonin secretion or light exposure (e.g., [49]).This variability can influence the interpretation of results obtained from a single morning sample.Conversely, it suggests that indicators of metabolic syndrome are intricately linked with disrupted circadian rhythms, evident in delayed light exposure and/or meal timing.
One of the strengths of our study is that a sample with national representation was obtained from ENSANUT-2018.However, this study has several limitations.The crosssectional design cannot evaluate longitudinal relationships between these surrogate insulin resistance and obesity markers and MetS.The participants in our study were limited to Mexicans; therefore, our results might not be generalizable to other populations.

Metabolites 2024 , 26 Figure 1 .
Figure 1.Flow diagram for obtaining the database for the diagnosis of MetS.

Figure 1 .
Figure 1.Flow diagram for obtaining the database for the diagnosis of MetS.

Figure 7 .
Figure 7. ROC curves for women over 60 years according to the ATP III criteria, evaluating the following surrogate markers with the reference line (a) SPISE; (b) METS-IR, LAP, METS-VF, VAT, BRI, VAI, TG/HDL-C ratio, TyG, and ABSI.

Figure 7 .
Figure 7. ROC curves for women over 60 years according to the ATP III criteria, evaluating the following surrogate markers with the reference line (a) SPISE; (b) METS-IR, LAP, METS-VF, VAT, BRI, VAI, TG/HDL-C ratio, TyG, and ABSI.

Table 1 .
Sociodemographic, anthropometric, and biochemical characteristics of 10575 Mexican adults with suspected metabolic syndrome according to ENSANUT 2018.

Table 2 .
Accuracy measures of surrogate insulin resistance and obesity markers to identify metabolic syndrome in Mexican adult men aged 20-39 according to the ATP III criteria.The sample included 506 subjects with MetS and 1216 without MetS.

Table 3 .
Accuracy measures of surrogate insulin resistance and obesity markers for the identification of metabolic syndrome in Mexican adult men aged 40-59 according to the ATP III criteria.The sample included 766 subjects with MetS and 941 without MetS.

Table 4 .
Accuracy measures of surrogate insulin resistance and obesity markers to identify metabolic syndrome in Mexican adult men over 60 according to the ATP III criteria.The sample included 536 subjects with MetS and 569 without MetS.

Table 4 .
Accuracy measures of surrogate insulin resistance and obesity markers to identify metabolic syndrome in Mexican adult men over 60 according to the ATP III criteria.The sample included 536 subjects with MetS and 569 without MetS.

Table 5 .
Accuracy measures of surrogate markers of insulin resistance and obesity to identify metabolic syndrome in Mexican adult women aged 20-39 according to the ATP III criteria.The sample consisted of 825 subjects with MetS and 1459 without MetS.

Table 5 .
Accuracy measures of surrogate markers of insulin resistance and obesity to identify metabolic syndrome in Mexican adult women aged 20-39 according to the ATP III criteria.The sample consisted of 825 subjects with MetS and 1459 without MetS.

Table 6 .
Accuracy measures of surrogate insulin resistance and obesity markers to identify metabolic syndrome in Mexican adult women aged 40-59 according to the ATP III criteria.The sample included 1471 subjects with MetS and 917 without MetS.

Table 6 .
Accuracy measures of surrogate insulin resistance and obesity markers to identify metabolic syndrome in Mexican adult women aged 40-59 according to the ATP III criteria.The sample included 1471 subjects with MetS and 917 without MetS.ATP III: Adult Treatment Panel III of the National Cholesterol Education Program; AUC: area under the curve; Se: sensitivity; Sp: specificity; PPV: positive predictive value; NPV: negative predictive value; LRP: positive likelihood ratio; LRN: negative likelihood ratio; SPISE: single-point insulin sensitivity estimator; TG/HDL: triglyceride-HDL ratio; TyG: triglyceride-glucose index; LAP: lipid accumulation product; VAI: visceral adiposity index; METS-IR: metabolic score for insulin resistance; METS-VF: metabolic score for visceral fat; VAT: visceral adipose tissue; ABSI: body shape index; BRI: body rounding index.

Table 7 .
Accuracy measures of surrogate insulin resistance and obesity markers to identify metabolic syndrome in Mexican adult women over 60 according to ATP III criteria.The sample consisted of 884 subjects with MetS and 485 without MetS.

Table 8 .
Accuracy measures of surrogate insulin resistance and obesity markers to identify metabolic syndrome in Mexican adult men aged 20-39 according to the IDF criteria.The sample consisted of 738 subjects with MetS and 984 without MetS.

Table 8 .
Accuracy measures of surrogate insulin resistance and obesity markers to identify metabolic syndrome in Mexican adult men aged 20-39 according to the IDF criteria.The sample consisted of 738 subjects with MetS and 984 without MetS.