Tri-Ponderal Mass Index Reference Values for Screening Metabolic Syndrome in Children and Adolescents: Results From Two National-Representative Cross-Sectional Studies in China and America

Introduction To ascertain the possible cut point of tri-ponderal mass index (TMI) in discriminating metabolic syndrome (MetS) and related cardio-metabolic risk factors in Chinese and American children and adolescents. Methods A total of 57,201 Chinese children aged 7-18 recruited in 2012 and and 10,441 American children aged 12-18 from National Health and Nutrition Examination Survey (NHANES 2001-2014) were included to fit TMI percentiles. Participants were randomly assigned to a derivation set (75%) and validation set (25%). The cut points of TMI with the lowest misclassification rate under the premise of the highest area under curves (AUC) were selected for each sex, which were additionally examined in the validation set. All of data analysis was conducted between September and December in 2019. Results TMI showed good capacity on discriminating MetS, with AUC of 0.7658 (95% CI: 0.7544-0.7770) to 0.8445 (95% CI: 0.8349-0.8537) in Chinese and 0.8871 (95% CI: 0.8663-0.9056) to 0.9329 (95% CI: 0.9166-0.9469) in American children. The optimal cut points were 14.46 kg/m3 and 13.91 kg/m3 for Chinese boys and girls, and 17.08 kg/m3 and 18.89 kg/m3 for American boys and girls, respectively. The corresponding misclassification rates were 17.1% (95% CI: 16.4-17.8) and 11.2% (95% CI: 9.9-12.6), respectively. Performance of these cut points were also examined in the validation set (sensitivity 67.7%, specificity 82.4% in Chinese; sensitivity 84.4%, specificity 88.7% in American children). Conclusions A sex- and ethnicity- specific single cut point of TMI could be used to distinguish MetS and elevated risk of cardio-metabolic factors in children and adolescents.


INTRODUCTION
Metabolic syndrome (MetS), specified by the National Cholesterol Education Program Adult Treatment Panel III (ATP III), is a strong and independent predictor of all-cause and cardiovascular disease (CVD) mortality (1,2). Its prevalence has increased from approximately 2% in the mid-1990s to a current estimate of 10% in American children and adolescents (3)(4)(5). It was also estimated that nearly two thirds (63.4%) of American adolescents had at least 1 metabolic abnormality (5). In Chinese children aged 10-16, the prevalence of MetS ranged from 5.5% to 10.9%, depending on the definition used in the study (6,7). Results from The China Health and Nutrition Survey suggested an overall prevalence of 3.37% in children aged 7-18, using ATP III definition (8). Since childhood MetS is likely to track into adulthood, early identification could help to prevent definitive lesions (9) and target interventions to improve future cardiovascular health (5,10).
In consideration of the great importance of early identification, an accessible screening tool is needed for daily public health practice (11), especially in the less developed areas where professional pediatricians are deficient. Body mass index (BMI), calculated as weight(kg)/height 2 (m 2 ) has been the most widely used indicator in screening obesity and related diseases, including MetS (12). In recent decades, given the fact that central obesity is the biggest risk factor of MetS (13), waist circumference has been accepted as another efficient screening tool for MetS in young population (14)(15)(16). However, as children's body shape changes dramatically during puberty, both measures required multiple cut-off values to fit age and sex subgroups of children.
Tri-ponderal mass index (TMI), calculated as weight(kg)/ height 3 (m 3 ), has been reported to be a satisfactory adiposity indicator with good age-stability during adolescence (17,18). Studies have also proven it to be an efficient indicator in screening obesity related cardio-metabolic risks including dyslipidemia, insulin resistance and other cardiovascular diseases (19)(20)(21). However, few studies have evaluated TMI cut point for screening of MetS and cardio-metabolic risks in pediatric populations. Although one study conducted in Colombian children and adolescents suggested that TMI could be a satisfying indicator to discriminate MetS, the authors gave age-and sex-specific screening thresholds (22), which didn't take the best advantage of its age-stability.
Using cross-sectional baseline data from the Chinese schoolbased Health Life Intervention and United States National Health and Nutrition Examination Survey (NHANES 2001(NHANES -2014, the present study aimed to ascertain the optimal TMI cut point for identifying MetS and cardio-metabolic risk factors in Chinese and American children and adolescents.

Study Population
Data for Chinese participants were obtained from baseline data of a Chinese school-based Health Life Intervention study, which was a national school-based multi-centered cluster randomized controlled trial against obesity in Chinese children and adolescents conducted in 2012. The sampling procedure of this study has been published elsewhere in detail (23). Briefly, more than 60,000 children and adolescents from 7 provinces, including Liaoning, Tianjin, Ningxia, Shanghai, Chongqing, Hunan, and Guangdong, participated in the study. In each province, 12-16 primary and secondary schools totaling around 10,000 participants were randomly selected. Of each selected school, two classes were randomly selected from each grade and students in those classes were invited for blood sample collection. Of the 59,916 participants aged 7 to 18 years, 2,715 were excluded due to missing sex, height or weight data. The remaining sample size for TMI derivation and analysis was 57,201, including 15,045 participants with blood samples. The original study was approved by Ethical Committee of the university. All participants and their parents provided signed informed consent.
Data for American participants were obtained from 7 cycles of National Health and Nutrition Examination Survey (NHANES 2001-2014) public release data, for which collection methods had been described in detail elsewhere (24). NHANES is an American national representative and continuous crosssectional survey conducted every 2 years. In this study, only adolescents aged between 12 and 18 were included, as serum glucose and lipid profiles were unavailable for those younger than 12. Among those aged 12 to 18 years old, 667 participants with missing data of height and weight were excluded. Thus, the sample size for TMI analysis in this study was 10,441, and the sample size for the following analysis was 2,910.
Within each population, participants were randomly allocated to derivation set (75%) and validation set (25%) by a random number table in Stata. Participants with full records of height, weight and age were recruited in fitting TMI percentiles, while only those with complete records of all 5 risk components of MetS were recruited in the following analysis. Basically, the optimal TMI cut points were developed from derivation set and were validated in validation set.

Measurements
All Chinese participants underwent physical examination according to a standard protocol (19). Height was measured using the portable stadiometer (model TZG, China) to the nearest 0.1 cm, with participants standing straight barefoot.  (21). Mercury sphygmomanometers (model XJ11D, China), stethoscopes (model TZ-1, China), and appropriate cuffs were used for blood pressure measurement. Participants were asked to sit quietly for at least 5 minutes prior to the first reading. Systolic blood pressure (SBP) was determined by onset of the first Korotkoff sound and diastolic blood pressure (DBP) was determined by the fifth Korotkoff sound. Blood pressure was measured twice with five minutes' gap between two measurements, and the average of SBP and DBP values were separately calculated. Blood samples were collected after a 12-hour fasting. Fasting plasma glucose (FPG), triglyceride (TG), and high-density lipoprotein cholesterol (HDL-C) were measured using an automatic biochemical analyzer (Roche Modular P800 ISE900; Hoffmann-La Roche Ltd) by a qualified biological testing company. Rigid quality control was enforced in this study. All measurement instruments were calibrated before use, and all examiners were required to pass a standard training course before commencement.
For American participants, anthropometric measurements were collected according to protocols previously published in detail (24). Sex and age were self-reported during home interviews. FPG, TG and HDL-C levels were measured enzymatically at Johns Hopkins University Lipoprotein Analytic Laboratory with the use of a Hitachi 704 Analyzer (Boehringer Mannheim Diagnostics, Indianapolis, IN). Lipid collection and analyses were standardized to Centers for Disease Control and Prevention criteria (25).
TMI for each participant was calculated as weight (kg)/ height 3 (m 3 ).
As reference values for waist circumference vary across ethnic groups, "High Waist Circumference Screening Threshold Among Chinese Children and Adolescents Aged 7-18 years" (27) and "Anthropometric Reference Data for Children and Adults: United States" (28) were used for Chinese and American populations, respectively. Blood pressure was evaluated with the 2017 version of the Clinical practice guideline for screening and management of high blood pressure in children and adolescents (29).

Statistical Analysis
All of data analysis was conducted between September and December in 2019. The characteristics of participants were given as mean (SD) or number (%). Independent two-tailed t-tests for continuous variables and chi-square (c 2 ) tests for categorical variables were used to compare the differences between derivation set and validation set in Chinese and American participants, respectively. LMS curves were generated with LMScharmaker (version 2.5, developed by HARLOW PRINTING LIMITED) to estimate TMI changes with age in both sexes in Chinese and American participants, respectively. Since TMI showed good stability with increasing age, especially within the range from P 5 to P 90 (results displayed in Supplementary Figure 1), the following analyses were carried in the absence of age factors.
The relationships between TMI percentile and prevalence of MetS components were estimated with logistic regression and fractional polynomial regression models by sex. Receiver operating characteristic (ROC) analysis was used to find the possible range of TMI cut-off percentiles, and the alternative points were identified as those with the highest Youden Index (30). Five optimal cut-off percentiles were selected for further analysis based on their capacity to distinguish MetS. The cut-off percentile with the lowest misclassification rate were selected under the premise of the largest area under ROC curves (AUC) (31). The cut-off values obtained from the derivation set were then validated in the validation set, by calculating their AUC and misclassification rate. Sensitivity analyses were conducted between sex groups for participants in both countries, while urban-rural disparities and ethical disparities were estimated for Chinese and American populations, respectively. All analyses were performed using Stata 14.0 (College Station, TX, USA) and associations were considered significant when performed at levels of P < 0.05 (two sides).

RESULTS
A total of 57,201 Chinese participants and 10,441 American participants were included in the present study, while 15,045 Chinese and 2,910 American participants had complete records of MetS risk components. The summary of their characteristics, including height, weight, waist circumference and MetS risk factors, are shown in Table 1. The prevalence of MetS is 6.6% to 7.0% for Chinese participants and 3.7% to 4.3% for American participants depending on derivation or validation set. Generally, no significant difference in MetS prevalence or other variables was observed between the two sets.
As TMI percentile increased, estimated prevalence of all cardiometabolic risk factors, including MetS, increased considerably. Notably, between P 70 and P 80 in Chinese participants, the estimated prevalence of MetS increased rapidly in both sexes. A similar pattern was also detected in American participants, notably between P 80 and P 90 (Figures 1A-D).  Table 2), therefore P 70 , P 75 , P 80 , P 85 and P 90 were selected for further evaluation. The corresponding TMI values of these percentiles are listed in Supplementary Table 3.
The results of AUC and misclassification rate analyses are shown in Figures 2A, B, demonstrating a reduction in the misclassification rate and AUC as higher percentiles were tested. In Chinese participants, AUC of P 70 , P 75 and P 80 varied from 0.7486 (95% CI: 0.7313-0.7660) to 0.7570 (95% CI: 0.7405-0.7735) with no statistical differences, while the AUC decreased significantly from P 80 (0.7570, 95% CI: 0.7405-0.7735) to P 85   Table 2).
The results for sensitivity analyses were displayed in Supplementary Tables 4, 5 with derivation population. In Chinese population, the capacity of the cut points were generally consistent with those from all population, although 13.93 kg/m 3 (P 75 ) may be optimal for Chinese boys. However, as the sample size of derivations set was rather small, the results should be test in further studies with bigger population. In the American population, P 85 showed good consistency in all participants except for Mexican American, for whom P 90 may be more appropriate cut-off percentile, as the AUC was significantly higher than other percentiles.

DISCUSSION
MetS affects approximately 35% of adults in the United States of America and is also a great threat in developing countries, including China (32,33). Let alone the MetS and related cardiovascular risks originated from childhood has significant influence on future diseases (34), it is of consensus that regular monitoring/management is important for children, especially for those with high MetS risk (7).
The present study found that TMI could be a satisfying index in identifying MetS and other cardio-metabolic risks in young Chinese and American populations. The thresholds established to identify elevated MetS risk were 14.46 kg/m 3 for boys vs. 13.91 kg/m 3 for girls in Chinese participants aged between 7 to 18, and 17.08 kg/m 3 for boys vs. 18.89 kg/m 3 for girls in American participants aged between 12 to 18. These cut points showed favorable capacity in discriminating MetS and related risks and performed good consistency in subgroup analysis.
TMI as an alternative option for BMI has been paid increasing attention in recent years, especially in pediatric populations where BMI changes with age. Previous studies found that TMI remained steady throughout childhood and showed good consistency to the percent of body fat (17,18,20). These  qualities made TMI a satisfactory indicator in daily screening of excessive weight and related cardiometabolic diseases. The present study verified its age-stability and ulterior discussed the possible TMI thresholds to discriminate MetS in both Chinese and American children. Compared to the thresholds derived from a previous study in a Columbian population (22), the present study found higher TMI cut-points in Chinese children, and even higher cut-points in American children.
The distinction between ethnic groups had been discovered previously when researchers attempted to set country-specific BMI thresholds, as they found that people from Asia and Pacific regions had a higher percent of body fat and waist-to-hip ratio than Caucasians at a given BMI (35). Further studies also found that even if their anthropometric measurements were within a safe range, people from eastern countries would have higher disease risk than their western counterparts (36,37). Besides, various cut-off points among different sex, race and ethic subgroups were found, indicating that even if TMI presented good age-stability during childhood, sex and ethic disparities should be considered in setting country-specific cut-points. For more validated TMI cutoff points, more standardized research under global cooperation should be conducted, in order to eliminate any possible deviations resulted from operators and testing examinations. For many years, researchers have been seeking an appropriate screening tool to identify adiposity, as well as obesity-related disease risk in children. Apart from ageand sex-specific BMI used for screening general adiposity, waist circumference and waist-to-height ratio are also becoming popular for central obesity (38). However, as these indicators were normally recommended for further assessment in children with risk of developing other long-term health problems than obesity, their applications in general younger populations were still under debate (39). Although, in our previous research, further combination with waist circumference did not improve the ability of discriminating high risk population of TMI in children and adolescents (20). The present study, with two national-representative surveys, provided a possible choice with both age-stability and good capacity in recognizing cardio-metabolic abnormalities. From a practical point of view, TMI bases only on height and weight measurements, and requires very simple calculation. The number of TMI thresholds for screening adiposity is also much less than those of BMI, and doesn't depend on age-specific percentiles. These qualities make TMI an accessible, easy, and accurate indicator for long-term monitoring of childhood cardio-metabolic abnormalities in primary care settings. And it could also be useful for school-based and community surveillance efforts in early prevention of cardio-metabolic risks (40). The current study had its limitations. Cross-sectional data was obtained from Chinese and American populations. Although majority of adolescent years were included in the present study, TMI variance at individual levels was still undetected. Meanwhile, it remained undetected that whether the time lag between two populations would make any difference to the primary findings of the present study. In addition, ethnic differences must be taken into consideration in application of TMI. Although this study conducted ethnic-stratified sensitivity analyses in American participants and found good consistency between non-Hispanic white, non-Hispanic black and other ethnic groups, it may be insufficient for making a general cutpoint. Ethnic-specific studies may need further consideration, while any differences resulted from operator, testing equipment or definition on specific metabolic disorders should also be taken into consideration. Moreover, the method of waist circumference in Chinese children was not standard. Although it was proved to be similar to the standard method in analyzing abdominal fat (data published in Chinese only), underestimation of actual waist circumference was also possible. This emphasizes the importance of collaborative and standardized cooperation again.
TMI could be an accurate and convenient population-based screening tool for MetS and related cardio-metabolic risk factors in Chinese and American children and adolescents. Corresponding TMI values for sex-specific P 80 in Chinese and P 85 in American populations could be utilized as appropriate thresholds in clinical practice.

DATA AVAILABILITY STATEMENT
The raw data supporting the conclusions of this article will be made available by the authors, upon reasonable request.

ETHICS STATEMENT
The studies involving human participants were reviewed and approved by Ethical Committee of Peking University. Written informed consent to participate in this study was provided by the participants' legal guardian/next of kin.

AUTHOR CONTRIBUTIONS
XW performed all statistical analyses and wrote the manuscript. JM is the principle investigator of the original study. BD and YD are co-investigators and reviewed and edited the manuscript. ZZ and YM are co-investigator of the original study. LA, YC, and WL reviewed and edited the manuscript. XW takes full responsibility for the contents of the article. All authors contributed to the article and approved the submitted version.