Background

Previously, obesity was usually assessed by weight, body mass, and skinfold thickness. Recently, scientists and physicians have focused more on abdominal visceral fat (AVF) accumulation for its evident association with the risk of hypertension, diabetes mellitus (DM) [1,2,3,4], cardiovascular disease [5] and dyslipidemia [6].

Magnetic resonance imaging (MRI) and computed tomography (CT) have been regarded as the best methods for precise measurement of mesenteric, omental and retroperitoneal adipose tissue in the living body. Pouliot and colleagues used the ratio of area of abdominal adipose tissue: femoral adipose tissue obtained by CT to evaluate abdominal visceral obesity [1], but it was expensive, time-consuming and involved exposure to ionizing radiation.

Waist circumference (WC) measurements can be acquired without expenditure. WC has shown an advantage for measurement of total abdominal adiposity, but its association with intraperitoneal fat varies depending on subcutaneous abdominal fat, sex and ethnicity [7]. The Visceral Adiposity Index (VAI) is based on WC, the Body Mass Index (BMI) as well as the levels of triglyceride (TG) and high-density lipoprotein-cholesterol (HDL-c). The VAI has been revealed to be a reliable marker of accumulation of visceral fat associated with cardiometabolic risk [5] and DM [4]. However, VAI calculation is complex, and validation of its power in estimating adipose-tissue distribution has been undertaken only on 26 patients via MRI [5].

The “hyper-triglyceridemic waist (HTW) phenotype” is defined as the simultaneous presence of a large WC and high TG concentration. It was mentioned first by Lemieux et al. in the Quebec City study for identification of a metabolic triad of nontraditional risk factors [increased levels of insulin (INS) and apolipoprotein-B, and small, dense low-density lipoprotein (LDL) particles]. They showed an increased risk for coronary artery disease in men with the HTW phenotype (WC ≥ 90 cm and TG ≥ 2.0 mmol/L) [6].

Subsequently, several studies focused on the HTW phenotype, and it was demonstrated to be associated with atherogenic and cardiovascular risks [2, 8,9,10,11,12,13,14,15], metabolic abnormalities and DM [3, 16,17,18,19,20,21,22]. Although the HTW phenotype was more prevalent in postmenopausal women, the risk of cardiovascular disease in this population did not differ in different phenotypes of the WC and TG groups [12]. Besides, among men with the HTW phenotype (WC > 90 cm, TG > 150 mg/dL), 78.9% of men were demonstrated to have an excess area of visceral adipose tissue > 130 cm2 according to CT [21].

Importantly, the HTW phenotype can be identified readily with little expense, can be used to predict abdominal visceral obesity and the risk of metabolic diseases and is convenient. However, its association with AVF distribution (as tested by MRI or CT) has not been evaluated extensively.

The present study was carried out in Chinese adults to ascertain how the HTW phenotype is associated with abdominal visceral obesity and its relationship with metabolic abnormalities.

Patients and methods

Ethical approval of the study protocol

The study protocol was approved by the Committee of Ethics in Research of the First Hospital of Qinhuangdao (2016C005; Hebei, China). All participants provided written informed consent to take part in our study.

Study population

This cross-sectional study involved 195 individuals referred to the Physical Examination Centre of the First Hospital of Qinhuangdao from 10 October to 10 December 2015. Inclusion criteria were age ≥ 18 years and Chinese ethnicity. Medical history was collected, and values for systolic and diastolic blood pressure were recorded routinely.

Exclusion criteria were ongoing inflammation, hypertension, cardiovascular disease, cerebrovascular disease, peripheral vascular disease, hepatic insufficiency, renal dysfunction, autoimmune disease or metabolic diseases (except for impaired glucose tolerance and use of hypolipidemic drugs in the previous month or anti-DM drugs in the last 24 h).

Participants were divided into four groups according to the TG concentration (in mmol/L) and WC (in cm) for males/females, respectively: A (normal): TG < 1.7 and WC < 85/80; B (“hyper-TG”): ≥1.7 and < 85/80; C (“hyper-WC”): <1.7 and ≥ 85/80; D (HTW phenotype): ≥1.7 and ≥ 85/80.

Anthropometric and body-composition measurements

Height, weight and WC were measured according to a standard procedure reported previously by our research team [23]. The Conicity Index was calculated as WC (m)/[0.109 × square root of weight (kg)/height (m)]. The waist:hip ratio was calculated as WC (m)/hip circumference (m).

Abdominal compartments of body fat were measured using a 1.5-T MRI scanner (Signa; GE Medical Systems, Milwaukee, WI, USA). A single MRI slice was obtained at the navel level of L4/L5 intervertebral spaces [24] using water-suppressed T1-weighted imaging sequences (repetition time = 600 ms, echo time = 13 ms). The AVF area was quantified using the area measurement tool of Advantage Workstation 4.2 (GE Healthcare Technologies, Milwaukee, WI, USA). “Abdominal obesity” was defined as an AVF area ≥ 100 cm2 [25].

Oral glucose tolerance test (OGTT)

After withdrawal of anti-DM drugs for ≥ 24 h, a 75-g OGTT was undertaken in the morning after a 12-h overnight fast. Blood samples were collected through an antecubital vein at 0, 30, 60, 120 and 180 min for determination of plasma glucose (PG, mmol/L) and INS (µIU/mL) concentrations. PG concentrations were tested using an automatic analyzer (7600 series; Hitachi, Tokyo, Japan). INS concentrations were measured by an automatic chemiluminescence immune analyzer (CL-2000i; Mindray, Shenzhen, China).

Indices of PG–INS homeostasis

To characterize the sensitivity and secretion capacity of INS, we used the homeostatic model assessment to estimate insulin resistance (HOMA-IR), which was calculated as fasting insulin (FINS, pmol/L) × fasting plasma glucose (FPG, mmol/L)/22.5. The Insulin Sensitivity Index (ISI) was calculated as 1/(FINS × FPG). The OGTT area under the curve of glucose (AUCG) was calculated as FPG/2 + PG60min + PG120min + PG180min/2. The OGTT area under the curve of insulin (AUCI) was calculated as FINS/2 + INS60min + INS120min + INS180min/2. The Early Insulin Secretion Index was calculated as (INS30/60 min − FINS)/(PG30/60 min − FPG), which were designated as △I30/△G30 and △I60/△G60, respectively.

Tests of lipids, liver function, uric acid (UA) and glycosylated hemoglobin A1c (HbA1c)

After a 12-h overnight fast, blood samples were extracted through an antecubital vein. An automatic analyzer (7600 Series; Hitachi) was used for measurement of the levels of TG, total cholesterol (TC), low-density lipoprotein-cholesterol (LDL-c), HDL-c, alanine aminotransaminase, aspartate amino-transferase, γ-glutamyl transpeptidase (GGT) and UA. The HbA1c level was measured using a hemoglobin testing system (Variant II; Bio-Rad Laboratories, Hercules, CA, USA).

Statistical analyses

Statistical analyses were carried out using SPSS v21.0 (IBM, Armonk, NY, USA). Data are the mean ± standard deviation. Group comparisons were undertaken via one-way ANOVA. Post hoc tests were executed via the Student–Newman–Keuls (S–N–K) test. Frequency distributions were analyzed via Chi square tests. Estimation of the independent contributions of clinical and laboratory variables as well as the HTW phenotype to the variance of AVF accumulation was done via the backward logistic regression method. There were no missing values in our study. P < 0.05 was considered significant.

Results

Clinical and biochemical features of participants

The study cohort comprised 195 eligible individuals [83 men and 112 women; 18–62 (mean, 40.31 ± 10.50) years]. Among the four groups, significant differences were shown with regard to the BMI, blood pressure, AVF area, as well as concentrations of lipids, UA and liver enzymes (P < 0.05 for all) (Table 1).

Table 1 Clinical, laboratory and anthropometric characteristics of 195 subjects and their comparisons in four groups
Full size table

People with the HTW phenotype (group D) were older and had a higher BMI and blood pressure than those with normal WC (groups A and B) but, when compared with group C, no difference was shown (Table 1). Individuals from group C were older than those of group B and also had a higher BMI as compared with those from groups A and B (Table 1).

WC, the Conicity Index, waist:hip ratio and AVF area increased successively in group A, B, C and D. Except for the AVF area of groups A and B, data from each neighboring group were revealed to be significantly different (P < 0.05 for all) (Table 1).

People with the HTW phenotype (group D) had the highest concentrations of TG, UA and GGT and the lowest HDL-c levels among the four groups. The UA levels of group B and C were higher than those of group A, whereas the HDL-c concentrations were lower as compared with those of group A. People with very high TG levels (group B and D) had higher cholesterol levels than those of group A, but no significant difference was found when compared with those of group C. Nevertheless, LDL-c levels showed no significant differences among the four groups (Table 1).

Plasma tolerance, and sensitivity and resistance to INS

PG and INS concentrations were significantly different among the four groups at each time point of a 75-g OGTT (P < 0.05) (Table 2). FPG was the lowest in group A and, simultaneously, FINS was highest in the HTW-phenotype group (group D). After taking 75 g of glucose (p.o.), people with the HTW phenotype (group D) had higher PG and INS concentrations (except for INS30min) than those of group A, increased PG120min and PG180min levels as compared with those of group B, as well as the highest PG peak at 60 min among the four groups.

Table 2 Data of 75-g OGTT test and plasma glucose–insulin homeostasis
Full size table

The HbA1c level, HOMA-IR and ISI showed significant differences among the four groups (P < 0.05 for all) (Table 2). People with the HTW phenotype (group D) had a higher HbA1c level than those of groups A and B, whereas no significant difference was found as compared with group C. They also had the highest HOMA-IR and lowest ISI. Conversely, individuals from group A had the highest ISI and lowest HOMA-IR. There were no differences in △I30/△G30 or △I60/△G60 among the four groups analyzed by one-way ANOVA or the S–N–K test. In addition, participants with the HTW phenotype (group D) had the highest AUCG and higher AUCI value than that of people with normal WC (groups A and B). Group C also had a higher AUCI value as compared with group A.

Binary logistic regression analysis for predicting the AVF area

Binary logistic regression analysis was done to quantify the contributions of clinical data (age, sex (male = 1/female = 2), the BMI, blood pressure), laboratory data (HDL-c, LDL-c, UA, HOMA-IR) and the HTW phenotype to accumulation of abdominal fat. Age, sex, the BMI, HDL-c level, LDL-c level and the HTW phenotype were revealed to be independent risk factors to an increased AVF area, and the odds ratio (OR) of the HTW phenotype was 4.300 (P = 0.034) (Table 3). Omnibus tests of model coefficients were significant with a Chi square value of 133.632 (P < 0.001).

Table 3 Binary logistic regression of AVF area
Full size table

Of 74 individuals with an AVF area ≥ 100 cm2 measured by MRI, 60 participants were predicted to have abdominal obesity using a logistic regression model for estimation, with a sensitivity of 81.08%. In the other 121 participants with an AVF area < 100 cm2, 110 cases were estimated not to have abdominal obesity, with a specificity of 90.91%. The total accuracy was 87.18%.

Discussion

The optimal choice of anthropometric indices for measurement of visceral adipose tissue is important for clinical assessment and scientific research. Previously, simple measurements such as weight, the BMI, WC and waist:hip ratio were used for evaluation of obesity, of which WC seemed more relevant to abdominal obesity. However, separation of visceral fat from subcutaneous fat using this approach was not possible [7], so precise prediction of AVF accumulation could not be achieved.

Scholars have hypothesized that the HTW phenotype could be a simple and effective predictor of abdominal visceral obesity. The prevalence of the HTW phenotype in young adults from southern Brazil was estimated to be 5.9% and 4.5% among men and women, respectively, when defining the HTW phenotype as WC ≥ 90 cm and TG ≥ 177 mg/dL for males, and WC ≥ 85 cm and TG ≥ 133 mg/dL for females [26]. In Iranian adolescents, 7.3% of boys and 5.6% of girls had the HTW phenotype when it was defined as a serum TG concentration ≥ 110 mg/dL and WC ≥ 90th percentile for age and sex in both sexes. With respect to Chinese adults, overall 35.4% of men and 33.6% of women had the HTW phenotype when it was defined as TG ≥ 110 mg/dL and WC ≥ 75 cm for men and women [20]. Apparently, the prevalence of the HTW phenotype can vary extraordinarily among different ages, ethnicities and definitions for the HTW phenotype.

Our study showed that individuals with the HTW phenotype were older and more obese, had a higher Conicity Index, waist:hip ratio and AVF area (155.71 ± 52.50 cm2) and seemed to be associated with higher blood pressure and more metabolic problems (dyslipidemia, impaired liver function and UA metabolism) compared with people with normal WC (groups A and B). Our results are consistent with those of other scholars [6, 16, 18]. The HTW phenotype has been demonstrated to be related to an increased risk of cardiovascular disease [27] and used as a screening tool of the atherogenic metabolic triad [6].

Notably, the HDL-c level was lowest in the HTW-phenotype group, whereas the LDL-c level was not the highest in this group. Hence, an increased risk of cardiovascular disease might not be based on increased levels of LDL-c. Among those without the HTW phenotype, obese people (group C) seemed to be the oldest, whereas those with increased TG levels (group B) were not correlated with aging or hypertension. People with a very high WC (group C) had higher indicators of obesity (the BMI, WC, the Conicity Index, waist:hip ratio and AVF area) than those with very high TG levels (group B). Hence, WC had a greater capacity to influence abdominal obesity than TG levels, so the HTW phenotype could be considered to be a dominant, increased WC supplemented with increased TG levels. A regression model including age, sex, the BMI, LDL-c level, HDL-c level and the HTW phenotype was demonstrated to predict an AVF area > 100 cm2 with satisfactory accuracy (87.18%), sensitivity (84.51%) and specificity (88.71%). The OR of the HTW phenotype was 4.300, which suggests that the HTW phenotype could be a reliable marker for prediction of AVF accumulation.

Individuals with the HTW phenotype had the highest FINS, HbA1c levels, AUCG and HOMA-IR, PG peak at 60 min in the OGTT and the lowest ISI values. Hence, the HTW phenotype could be considered to be related to increased INS resistance, a finding that is in agreement with other studies [4, 6, 28]. Conversely, individuals with normal serum TG levels and WC (group A) had the highest ISI, suggesting that the TG level and abdominal obesity contributed to INS sensitivity, even though the influence appeared mainly after 60 min in the OGTT and the △I30/△G 30 and △I60/△G60 values showed no significant differences among the four groups. Recently, the HTW phenotype was reported to be associated with type-2 DM in China (OR = 6.23), with the strongest relationship found in those without obesity (BMI < 24 kg/m2, OR = 6.54) [22]. We did not undertake analyses of the receiver operating characteristics curve. However, a large-scale study involving 15,540 Chinese adults showed that the cutoff values of the HTW phenotype were TG ≥ 110 mg/dL and WC ≥ 75 cm for prediction of impaired FPG and DM [20].

Considering TG levels and WC separately, obese people (groups C and D) with very high TG levels (group D, HTW phenotype) had a higher AVF area, Conicity Index, waist:hip ratio and HOMA-IR, as well as a lower ISI, when compared with those with normal TG levels (group C). Apparently, the HTW phenotype had a greater capacity for identifying abdominal obesity and INS resistance than that of WC alone.

Interestingly, the factors influencing the HTW phenotype for men and women were demonstrated to be different in a study in Brazil [26]. For males, a sedentary lifestyle, tobacco smoking and obesity were associated with the HTW phenotype; among women, the HTW phenotype was related positively with skin color, family income, obesity and dietary fat intake [26]. However, limited by the sample size, we did not separate men and women in one-way ANOVA or post hoc tests. In binary regression analysis, sex (male = 1, female = 2) contributed to estimation of the AVF area with an OR of 0.196 (P = 0.003), suggesting that females had an advantage with regard to control of abdominal obesity. After adjustment by sex and age, the HTW phenotype remained significant for prediction of an increased AVF area (OR = 4.300).

People of Asian ethnicity tend to have greater abdominal obesity than that of Caucasians with similar BMI [29], so the cutoff value of WC for defining abdominal obesity might differ for different ethnicities, and establishment of an ethnicity-specific HTW phenotype is necessary.

The HTW phenotype appears to be reliable for predicting an increased AVF area and could be closely related to the risk of cardiovascular disease and DM, so its early recognition is important. For people with the HTW phenotype, attempts at controlling WC and reducing lipid uptake are strongly recommended. A strategy could be increasing daily exercise, maintaining a regular lifestyle, eating a low-fat diet and taking hypolipidemic drugs (if necessary). If people with the HTW phenotype have already suffered from metabolic diseases, treatments involving weight loss and lipid reduction will be more aggressive.

Conclusions

The HTW phenotype was shown to be a reliable marker for prediction of AVF accumulation associated with INS resistance. People with the HTW phenotype were demonstrated to be older and more obese, have a higher AVF area and were associated with higher blood pressure, more metabolic problems and INS resistance than people with normal WC.