Hypertriglyceridemic-waist phenotype is strongly associated with cardiovascular risk factor clustering in Chinese adolescents

The early identification of predictors related to cardiovascular risk factor clustering (CVRFC) in adolescents can help prevent Cardiovascular disease. The hypertriglyceridemic-waist circumference (HTW) phenotype is considered a simple and useful indicator to identify cardiovascular disease. However, there is limited research on the relationship between the HTW phenotype and (CVRFC) in adolescents. It is unclear whether the HTW phenotype can identify early the risk of developing CVRFC in adolescents. The study aimed to examine the association of the HTW phenotype with CVRFC in adolescents. A total of 1478 adolescents aged 12–18 years were classified into normal waist circumference (WC) and normal triglyceride (TG) (NWNT, 66.4%), normal WC and high TG (HTG, 5.5%), enlarged WC and normal TG (EW, 22.2%) and enlarged WC and high TG (HTW, 5.8%). High TG was defined as TG ≥ 1.47 mmol/L and enlarged WC ≥ 90th percentile by gender and age. CVRFs in this study included elevated blood pressure (BP), impaired fasting glucose (IFG), high total cholesterol (TC), low high-density lipoprotein cholesterol (HDL-C), and high low-density lipoprotein cholesterol (LDL-C). CVRFC ≥ 2 or CVRFC ≥ 3 were defined as the presence of at least two or three cardiovascular risk factors. After adjustment for BMI, gender and age, the HTW phenotype increased the risk of CVRFC ≥ 2 and CVRFC ≥ 3 compared to the NTNW phenotype, OR and 95%CI were 2.40 (1.23–4.58) and 3.63 (1.49–8.86), respectively. After stratification by gender, similar results were found in boys, however, girls with the EW phenotype had a lower risk of CVRFC ≥ 2 and CVRFC ≥ 3 compared with the NTNW phenotype after adjustment for BMI and age. The area under the ROC curve was 0.698 (0.661–0.736) and 0.782 (0.725–0.840) when TG was combined with WC to detect cardiovascular risk factors clustering, which was better than BMI, WHtR, TG or WC alone. And similar results were obtained for both boys and girls when stratified by gender. These results revealed that different combinations of TG and WC levels are closely associated with cardiovascular risk factors clustering in both boys and girls, and TG combining WC performed better than BMI, WHtR, TG or WC alone in detecting cardiovascular risk factor clustering in adolescents.

Physical measurement. Height and weight were measured using a mechanical stadiometer (Model: ZH7082) and an electronic scale (Model: RGT-140), with the subject removing shoes and heavy clothing, both measured twice and averaged for inclusion in the final analysis, to an accuracy of 0.1 cm and 0.1 kg for height and weight respectively. Waist circumference (WC) was measured using a nylon tape measure and the measurements were averaged twice to an accuracy of 0.1 cm. BMI was calculated as weight divided by the squared height (kg/m 2 ). Blood pressure (BP) was measured by using a calibrated electronic sphygmomanometer (Model: HEM-7012, Omron, Japan) according to the standard method by the "American Hypertension Education Project Working Group" 14 . A suitable cuff was chosen for the measurement (7 cm, 9 cm, 12 cm, etc. for BP measurement in adolescents) and the subject was seated facing the measurer and BP was measured on the right upper arm with the elbow at the same level as the sphygmomanometer and the heart. Systolic blood pressure (SBP) and diastolic blood pressure (DBP) were measured three times at 1-min intervals, and the average of the last two readings was recorded for the final analysis (a third measurement was taken if the difference between the first two blood pressure values exceeded 10 mm Hg (1 mm Hg = 0.133 kPa)).
Biochemical analysis. Venous blood samples were collected after at least 12 h of overnight fasting. Fasting plasma glucose (FPG), triglyceride (TG), total cholesterol (TC), high-density lipoprotein cholesterol (HDL-C) and low-density lipoprotein cholesterol (LDL-C) were measured by using an automated biochemistry analyzer (Model: AU480, American). FPG, TG and TC were detected by enzymatic methods, HDL-C and LDL-C were measured by the direct method-peroxidase method.
Definitions. Cardiovascular risk factors (CVRFs) were defined as follows, elevated WC, elevated BP and impaired fasting glucose (IFG) were defined using the harmonized criteria of the International Diabetes Federation (IDF) 15 : enlarged WC, WC ≥ 90th percentile by gender and age; elevated BP, SBP or DBP ≥ 90th percentile by gender and age; IFG, FPG ≥ 5.6 mmol/L. Adverse lipid concentrations were defined according to the National Heart, Lung, and Blood Institute expert panel on integrated guidelines for cardiovascular health and risk reduction in children and adolescents 16 : high TG, TG ≥ 1.47 mmol/L; high TC, TC ≥ 5.18 mmol/L; high LDL-C, LDL-C ≥ 3.37 mmol/L; low HDL-C, HDL-C ≤ 1.03 mmol/L. Cardiovascular disease risk factor clustering (CVRFC) refers to the number of five factors as follows: elevated BP, IFG, high TC, high LDL-C and low HDL-C (High TG was not included in the definition of CVRFC in this study to avoid spurious associations). CVRFC ≥ 2 was defined as the presence of at least two cardiovascular risk factors and CVRFC ≥ 3 was defined as the presence of at least three cardiovascular risk factors.
The study subjects were divided into four groups according to WC and TG level:(1) normal TG and normal WC (NWNT): TG < 1.47 mmol/L 16
Prevalence of cardiovascular risk factors and cardiovascular risk factors clustering in the participants across different HTW phenotype groups. Figure 1 shows that the prevalence of elevated BP, IFG, high TC, low HDL-C, high LDL-C, CVRFC ≥ 2 and CVRFC ≥ 3 differed by different HTW phenotypes in the total population and both boys and girls (all P < 0.05). Table 4 presents the results of binary logistic regression analyses for different HTW phenotypes with specific cardiovascular risk factors adjusted for BMI, gender and age. After adjustment for covariates, the HTW  27-5.71). After gender stratification, the results showed that compared with the NTNW phenotype, the HTW phenotype was a risk factor for Elevated BP, High TC, Low HDL-C and High LDL-C in boys, respectively, the HTG phenotype was a risk factor for IFG, High TC and Low HDL-C in girls, while EW phenotype was negatively associated with IFG (all P < 0.05).  www.nature.com/scientificreports/ Table 5 indicates that before adjusting for variables, the HTW phenotype increased the risk of CVRFC ≥ 2 and CVRFC ≥ 3 compared to the NTNW phenotype. This risk persisted after adjusting for BMI, gender and age, OR and 95% CI were 2.40 (1.23-4.58) and 3.63 (1.49-8.86), respectively. After stratification by gender, similar results were found in boys, girls with the HTG phenotype had a higher risk of CVRFC ≥ 2, however, those with the EW phenotype had a lower risk of CVRFC ≥ 2 and CVRFC ≥ 3 compared with the NTNW phenotype (all P < 0.05). Table 6 shows the results of the ROC curve analysis. TG combining WC performed better than BMI, WHtR, TG or WC alone in detecting cardiovascular risk factor clustering. The AUC was 0.698 (0.661-0.736) and 0.782 (0.725-0.840) in predicting CVRFC ≥ 2 and CVRFC ≥ 3 after adjustment for BMI, gender and age. Similar results were found after gender stratification (all P < 0.01).

Discussion
In this study, we found that after adjusting for BMI, gender and age, the HTW phenotype had a higher risk of CVRFs compared to the NTNW phenotype. A one-year cohort study of children and adolescents showed that the HTW phenotype was a risk factor for longitudinal changes in SBP during follow-up 17 . Another study showed that the HTW phenotype was a strong predictor of incident hypertension, those with HTW phenotype were 2.3 times more likely to develop hypertension than those with NTNW phenotype after adjusting for gender and age 18 . The results of other studies also suggest that the HTW phenotype with a higher prevalence of hypertension compared to the NTNW phenotype 11 . Those are consistent with the results of the present study. Esmaillzadeh et al. suggest that adolescents with the HTW phenotype are not significantly associated with the development of IFG compared to adolescents with the NTNW phenotype 19 . However, another study of children and adolescents identified that after adjusting for confounding variables, an increase in fasting glucose means of 3.87 mg/ dl (95%CI: 1.68-6.05) at one-year follow-up in those with the HTW phenotype 17 . Several studies in adults have also shown that the HTW phenotype is associated with IFG and even with the incident of type 2 diabetes 9,20,21 . Our study showed that compared with the NTNW phenotype, the HTW phenotype was not associated with IFG in both boys and girls, but the EW phenotype was negatively associated with IFG in girls after adjusting for BMI  www.nature.com/scientificreports/ and age. We also found that compared with the NTNW phenotype, the girls with the EW phenotype had lower FPG and the prevalence of IFG. This difference may be related to differences in study populations and regions, as well as the lower prevalence of IFG in adolescents in this population. Further explanation of the association between the HTW phenotype and glucose in adolescents is still needed in more studies.
The previous study illustrated that adolescents with the HTW phenotype were significantly more likely to have high TC (OR = 2.9; 95%CI:2.0-4.2), high LDL-C (OR = 1. 8; 95% CI 1.3-2.7) and low HDL-C (OR = 1.6; 95%CI:1.3-2.0) after controlling for potential confounding variables 19 . Another study on adolescents thought that after controlling for age and gender, adolescents with the high TG and high WHtR were more likely to have high TC (OR = 7.8; 95%CI:3.5-17.3), high LDL-C (OR = 9.4; 95%CI:2.8-31.2) and low HDL-C (OR = 10.8; 95% CI = 6.9-17.0) than those adolescents with normal TG and normal WHtR 22 . Other studies also showed that HTW was associated with high TC and low HDL-C in children and adolescents aged 10-18 years [23][24][25] . Adult studies have found that individuals with the HTW phenotype have a greater chance of having low HDL-C and LDL-C compared to individuals with the NTNW phenotype 11,26,27 . We obtained similar results in boys, but in girls, the HTW phenotype was associated with low HDL-C, but not with high TC and high LDL-C. This difference may be related to differences in gender and sex hormone levels, but more definitive underlying mechanisms need to be further investigated.
The clustering of CVRFs among adolescents is known to be associated with accelerated atherosclerosis and an increased risk of many chronic diseases, such as hypertension, hyperglycemia and dyslipidemia in adulthood 4,28 . Therefore, after confirming the association of the HTW phenotype with individual CVRFs, our study further analyzed its association with CVRFC and the results presented that the HTW phenotype was related to an increased risk of CVRFC ≥ 2 (OR = 2.40; 95%CI:1.23-4.58) and CVRFC ≥ 3 (OR = 3.63; 95%CI:1.49-8.86) in adolescents after adjustment for BMI, gender and age. Previous studies in adolescents have shown that the HTW phenotype is a stronger risk factor for CVRFC ≥ 1 (OR = 1.4; 95%CI:1.1-1.7) and CVRFC ≥ 2 (OR = 2.2; 95%CI:1.6-3.0) compared to adolescents with the NTNW phenotype after adjusting for potential confounding variables 19 . Bailey et al. found that in participants aged 10-19 years with HTW phenotype, the odds of having CVRFC ≥ 1 (OR = 4.78; 95%CI:1.32-17.29) and CVRFC ≥ 2 (OR = 7.16; 95%CI:2.38-21.54) were higher than those without the HTW phenotype 24 . This is similar to the results of our study. Another adult study showed that hypertensive adults with the HTW phenotype were significantly more likely to have all CVRFs compared to the NTNW group, and in particular for 8.35 times (95% CI 5. 92-11.79) more likely to have CVRFC ≥ 3 12 .
Regarding the HTW phenotype increased risk of cardiovascular risk may be associated with insulin resistance and endothelial dysfunction. First, the increase in WC, a proxy for abdominal fat, reflects some extent the accumulation of visceral and subcutaneous fat tissue. In the case of central obesity, visceral adipocytes release excess fatty acids and pro-inflammatory adipocytokines such as leptin and tumour necrosis factor-alpha into the portal circulation, leading to increased hepatic adiposity and insulin resistance, which further activates the renin-angiotensin-aldosterone system, increasing sympathetic activity, enhancing procoagulant activity, and inducing endothelial dysfunction, leading to hypertension and other cardiovascular diseases [29][30][31] . And a recent study found that high TG and high WC is a state of insulin resistance in adolescents 32 . Besides, a meta-analysis showed a significant correlation between the HTW phenotype and insulin resistance 20 . When the body has both abdominal obesity and high triglycerides, there may be a superimposed effect on insulin resistance. Insulin resistance has been identified as a major cause of increased cardiovascular risk factors 33 .
After stratification by gender, similar results were found in boys, however, girls with the EW phenotype had a lower risk of CVRFC ≥ 2 and CVRFC ≥ 3 compared with the NTNW phenotype when adjusting for BMI and age. Our further study analyzed that after adjustment for BMI and age, among adolescents with normal TG, the risk of IFG and CVRFC ≥ 2 in girls with enlarged WC were 0.17 times (95%CI:0.05-0.61, P = 0.007) and 0.36 times (95%CI:0.15-0.86, P = 0.022) lower than those girls with normal WC, respectively. Conversely, elevated TG levels increased the risk of IFG and CVRFC in girls with normal WC. It means when adjusted for the effect  www.nature.com/scientificreports/ of BMI, it appeared to suggest that enlarged WC levels were protective against IFG and CVRFC in girls with normal TG. The results in girls were unexpected. But this explains why girls with the HTW phenotype did not have an increased risk of IFG and CVRFC compared with the NTNW phenotype. Data from a 6-year cohort study showed that FPG, TG, TC levels and the prevalence of T2DM decreased with increasing WC in females (P < 0.001) 34 . Another cohort study also showed that females with a WC gain had lower levels of TG and FPG at baseline than those without a gain (P < 0.05) 35 . WC serves as a simple surrogate for abdominal fat, which includes visceral fat tissue (VAT) and subcutaneous fat tissue (SAT). Data from McLaughlin et al. demonstrate that after adjustment for BMI and VAT, SAT is protective for insulin resistance, whereas VAT, after adjustment for SAT and BMI, has the opposite effect 36 . Studies have shown differences in TG lipolysis and turnover in visceral and subcutaneous white adipose tissue, and the proportion of VAT or SAT in abdominal fat determines the main factor of metabolic disorders, and these effects were more pronounced in girls than boys [37][38][39] . In addition, it may be related to testosterone (TT) and sex hormone-binding globulin (SHBG). In experimental studies, TT has been found to be a protective factor against atherosclerosis via suppressing pro-inflammatory cytokine activity and immuno-modulating effects 40 41 . But these are far from explaining this interesting phenomenon, and more large sample studies or studies on the mechanism are still needed.
Previous study results suggest that the HTW phenotype may be as discriminant as the NCEP-ATP III or the IDF criteria and could be used as an initial screening approach to identify individuals with deteriorated cardiometabolic risk markers 42 . Some studies of adolescents suggested HTW phenotype as a simple marker to identify adolescents at risk for metabolic syndrome (MetS) and other metabolic abnormalities 19,23 . Other adult studies have also shown that the HTW phenotype is independently associated with CVRFs and suggest that the HTW phenotype may be a simple and useful tool to screen individuals for future cardiovascular disease risk [9][10][11]43,44 . Liu et al. showed that the HTW phenotype is a reliable tool for identifying MetS, with an AUC of 0.843 (0.824-0.862) in men and 0.839 (0.813-0.865) in women 45 . Another study showed an AUC of 0.81 for TG*WC to predict metabolic syndrome 46 . In the present study, TG combining WC predicted AUC of 0.698 and 0.782 for CVRFC ≥ 2 and CVRFC ≥ 3 in adolescents respectively, which performed better than BMI, WHtR, TG or WC alone. Lee et al. also concluded that the combination of TG and WC has been illustrated as the best indicator of overall MetS in both genders 47 . These findings suggest that the HTW phenotype is not only strongly associated with individual cardiovascular risk factors, but also has important implications for the identification of metabolic syndromes and cardiovascular risk factor clustering.
In this study, we identified the important role of a simple combination of TG and WC in identifying cardiovascular risk factors and cardiovascular risk factors clustering in adolescents, and the protective effect of WC on IFG and cardiovascular risk factors clustering in girls with normal TG. This implies that for adolescents, a simple primary screening with TG and WC indicators can provide some basis for identifying early signs of cardiovascular disease. However, there are several limitations of the study that should be noted. Firstly, this cross-sectional study limits the causal interpretation of the observed associations. Secondly, this study did not assess some confounding factors such as lifestyle and physical activity, which may have influenced our results. Thirdly, regarding the protective effect of EW phenotype in girls on IFG and CVRFC and the confidence interval is wide in Table 5, which may be caused by the small sample size. Finally, the results cannot be generalized to other populations due to the age and ethnicity limitations of the participants. More cohort studies are needed in the future to explore the causal relationship, taking into account some confounding factors (e.g. diet, physical activity and family history of hypertension). This study only explored adolescents aged 12-18 years, and future studies could appropriately widen the age range or even examine whether there are differences between ages. Regarding the particular results in girls, this is a point well worth exploring in depth. In the discussion, we think that there may be an association to subcutaneous fat tissue (SAT), testosterone (TT) and sex hormone-binding globulin (SHBG) in which future studies could consider whether there is a mediating role for these factors.

Conclusions
In conclusion, after adjustment for BMI and age, compared with the NTNW phenotype, boys with the HTW phenotype were at higher risk for cardiovascular risk factors and cardiovascular risk factors clustering, and girls with the EW phenotype had a lower risk of cardiovascular risk factors clustering. TG combining WC performed better than BMI, WHtR, TG or WC alone in predicting cardiovascular risk factor clustering.