Association of metabolic syndrome with TyG index and TyG-related parameters in an urban Chinese population: a 15-year prospective study.

BACKGROUND: The metabolic syndrome (Mets) is a multiplex risk factor for atherosclerotic cardiovascular diseases. The aims of the study were to assess the association of the Mets with TyG index and TyG-related parameters in an urban Chinese population.
METHODS: The data were collected in 1992 and then again in 2007 from the same group of 590 individuals (363 males and 227 females) without Mets in 1992. The fasting lipid profile and blood glucose were measured. TyG index and related parameters were calculated, and Mets defined according to the harmonized criteria. The area under the curve (AUC) of receiver operating characteristic curves was used to evaluate TyG index and related parameters for their diagnostic ability to identify people with Mets. Odd ratios (OR) for Mets prediction were calculated using stepwise logistic regression analyses.
RESULTS: The incidence of Mets was 18.64% over the 15-year follow-up period.During 15 years' follow-up, TyG-waist to height ratio (TyG-WHtR) shows the largest AUC for Mets detection (0.686) followed by TyG-waist circumference (TyG-WC) (0.660), TyG-waist-to-hip ratio (TyG-WHpR) (0.564), and TyG index (0.556) in all participants. Gender analysis revealed that TyG-WHtR and TyG-WC have the largest AUC in both genders. TyG-WHtR significantly predicted Mets in all participants, with an unadjusted odds ratio of 5.63 (95% CI 3.23-9.83 P < 0.001). Associations remained significant after adjustment for smoking, drinking, physical exercise and components of Mets.
CONCLUSIONS: TyG-WHtR might be a strong and independent predictor for Mets in all participants in an urban Chinese population. TyG-related markers that combine obesity markers with TyG index are superior to other parameters in identifying Mets in both genders.

Introduction

Metabolic syndrome (Mets) is a cluster of metabolic abnormalities characterized by abdominal obesity, hypertension, dyslipidemia, abnormal glucose metabolism, or previously diagnosed type 2 diabetes [1]. Cardiometabolic abnormalities that are associated with the Mets can increase the risk of cardiovascular disease and type 2 diabetes mellitus [2]. Insulin resistance (IR) is characterized by impaired tissue sensitivity or responsiveness to circulating insulin, which plays an important role in the development of Mets [3]. The triglycerides and glucose (TyG) index combine fasting plasma glucose (FPG) and fasting triglycerides (TG), is a novel tool that has been suggested to help as a surrogate marker for IR [4]. In recent years, the TyG index has been deemed to be more accurate than IR in predicting the risk of insulin resistance related metabolic diseases [5]. Many evidence has shown that there was a strong correlation between the TyG index and type 2 diabetes mellitus, hypertension, cardiovascular events and fatty liver both in China and elsewhere [6-9].

There are several anthropometric measures that can predict Mets, such as waist circumference (WC), waist-to-height ratio (WHtR) and waist-to-hip ratio (WHpR) [10]. In recent years, researchers have focused on TyG-related parameters such as the product of TyG and waist circumference (TyG-WC), TyG and waist-to height ratio (TyG-WHtR), TyG and waist-to-hip ratio (TyG-WHpR) as well as their ability to predict the risk of cardiovascular events [11]. In a cross study, Taiwo H et al. found that TyG-related parameters improved identification and prediction of Mets in Nigerians [12]. However, there is no prospective study to explore the relationship between TyG-related parameters and Mets in an urban Chinese population. Therefore, this study aimed to prospectively determine the predictive value of TyG-related parameters for the Mets in an urban Chinese population.

Methods

Study population

The study population was obtained from a Chinese Multiprovincial Cohort Study (CMCS) in an urban community located in Chengdu, Sichuan province, China. A baseline examination was conducted in 1992 using a risk factor survey developed by the World Health Organization-Multinational Monitoring of Trends and Determinants in Cardiovascular Diseases (WHO-MONICA) [13]. The data were again collected in 2007 from the same group. Detailed information of these participants has already been reported [14-17]. In 1992, each patient’s history of hypertension, diabetes mellitus, hyperlipidemia, heart diseases (coronary artery disease, heart failure or arrhythmia), current smoking and current alcohol consumption, as well as their physical exercise habits, was determined by self-administered questionnaires and confirmed by a physician’s interview. After at least 5-min of rest in a seated position, blood pressure (BP) was measured in the sitting position twice at 2-min interval using an upright standard sphygmomanometer. Waist, height, weight, and body mass index (BMI) were measured. BMI was calculated as body weight (kg) divided by the height squared (m2). Blood was drawn from the antecubital vein in the morning after a 12-h fast for determinations of fasting plasma glucose (FPG), fasting serum TC, LDL-C, HDL-C, and TG. These chemistries were measured at West China Hospital laboratory. The study participants selection diagram are presented in Fig. 1. Since 114 participants were diagnosed with Mets, 7 participants with heart disease in 1992, they were excluded from the analysis. Therefore, only 590 participants with complete data were available and analyzed. This study was approved by the Ministry of Health of China, and the Ethics Committee of West China Hospital of Sichuan University. All participants provided written informed consent. In 2007, we repeated those measurements with the same methods.

Fig. 1

Flow chart of study population

Related definitions

Mets were defined as the new joint interim statement [1], and the presence of any 3 of 5 after mentioned risk factors constituted a diagnosis of MetS: (1) elevated TG was defined as 1.7 mmol/L or greater; (2) BP was defined as systolic BP (SBP) ≥ 130 and/or diastolic BP (DBP) ≥ 85 mmHg and/or those receiving antihypertensive medications; (3) reduced HDL-C was defined as a level less than 1.0 mmol/L for males and a level less than 1.3 mmol/L for females; (4) elevated FPG was defined as 5.6 mmol/L or greater; (5) for Asians, elevated WC was defined as 80 cm or greater for females and 90 cm or greater for males [1, 18]. Smoking: average cigarette consumption ≥ one/day. Alcohol intake: average intake of alcohol ≥ 50 g/day. Physical activity: exercise one or more times per week, at least 20 min each time.

TyG index and related parameters were calculated as follows:

TyG index = Ln[TG (mg/dL)·fasting glucose (mg/dL)/2] [19].

TyG-WC = TyG index*WC

TyG-WHpR = TyG index*WHpR

TyG-WHtR = TyG index*WHtR.

Statistical analysis

Data are presented as the mean ± SD for normally continuous variables and as frequency (%) for categorical variables by gender. Additionally, to explore the relationship between TyG index and TyG-related parameters and risk of Mets, both univariate and multivariate logistic regression analyses were used to estimate the odds ratios (ORs) and 95% CI values. Similarly, the ORs and 95% CIs for the risk of Mets in various parameters across each subgroup were estimated and their interactions were tested. The diagnostic ability of TyG index and TyG index-related parameters to identify people with Mets (as per the harmonized criteria) was determined with the receiver operating characteristic (ROC) curves. Pairwise comparisons between area under the curve (AUC)s for the four parameters were performed. A 2-tailed p < 0.05 was considered significant in all analysis. All analyses were performed using Empower (R) (http://www.empowerstats.com, X&Y solutions, Inc., Boston MA) and R (http://www.R-project.org).

Results

Baseline characteristics of Mets patients and controls

The incidence of Mets was 18.64% over the 15-year follow-up period. Table 1 shows the baseline characteristics of the involved population classified by genders. A total of 590 subjects were included in our study, including 363 (61.68%) males and 227 (38.32%) females. The mean age of males was older than that of females. The males had higher SBP, DBP, High, Weight, Waist circumference, waist hip rate (WHpR) as well as rate of smoking and alcohol intake. Compared with females, males had higher levels of TG. By contrast, the level of HDL-C was lower in the males. Values of TyG index, TyG-WC and TyG-WHpR were higher in the males than in the females.

Table 1

Baseline characteristics of the involved population classified by genders

	Males	Females	P-value
N	363	227
Age	48.87 ± 5.73	46.00 ± 6.13	<0.001
EH	41 (11.29%)	18 (7.93%)	0.185
FBG (mmol/l)	4.26 ± 0.74	4.31 ± 1.12	0.964
Height (cm)	165.16 ± 5.71	154.59 ± 5.45	<0.001
Weight (cm)	62.59 ± 8.01	54.93 ± 6.74	<0.001
Waist (cm)	77.42 ± 7.41	72.19 ± 6.32	<0.001
Hip (cm)	91.18 ± 5.37	91.66 ± 5.27	0.266
WHpR	0.85 ± 0.06	0.79 ± 0.05	<0.001
WHtR	0.47 ± 0.05	0.47 ± 0.04	0.649
BMI	22.94 ± 2.72	22.97 ± 2.40	0.825
SBP (mmHg)	113.40 ± 12.82	109.91 ± 12.97	0.001
DBP (mmHg)	73.15 ± 8.46	71.34 ± 8.30	0.011
TG (mmol/L)	2.04 ± 0.86	1.86 ± 0.73	0.011
TC (mmol/L)	4.42 ± 0.70	4.50 ± 0.80	0.294
HDL-C (mmol/L)	1.24 ± 0.22	1.30 ± 0.24	<0.001
LDL-C (mmol/L)	2.23 ± 0.76	2.32 ± 0.81	0.327
TyG index	8.76 ± 0.39	8.68 ± 0.35	0.020
TyG-WC	555.38 ± 67.19	512.30 ± 55.37	<0.001
TyG-WHtR	4.11 ± 0.48	4.06 ± 0.42	0.257
TyG-WHpR	7.43 ± 0.65	6.84 ± 0.54	<0.001
Smoking	228 (62.81%)	1 (0.44%)	<0.001
Drinking	210 (57.85%)	6 (2.64%)	<0.001
Exercise	78 (21.49%)	46 (20.26%)	0.030

Data are presented as means ± SD or number (percentage)

EH, essential hypertension; FPG, fasting plasma glucose; BMI, body mass index; WHpR, waist-to-hip ratio; WHtR, waist-to-height ratio; SBP, systolic blood pressure; DBP, diastolic blood pressure; TG, triglyceride; TC, total cholesterol; HDL-C, high-density lipoprotein cholesterol; LDL-C, low-density lipoprotein cholesterol; TyG, triglyceride-glucose; TyG-WC, product of TyG and waist circumference; TyG-WHtR, product of TyG; TyG-WHpR, product of TyG and waist-to-hip ratio

Logistic regression analyses for TyG index and TyG-related parameters with Mets risk

In the univariate logistic regression analysis, TyG index and TyG-related parameters were associated with Mets. This association persisted after adjustments for some Mets risk factors (age, gender, smoking, drinking, physical exercise, components of Mets). Before adjustment, TyG-WHtR presented the highest OR in all participants (4.86, 95% CI 2.98–7.95). After adjustment, TyG-WHtR presented the highest OR in all participants (5.63, 95% CI 3.23–9.83). (Table 2).

Table 2

Logistic regression analyses for the relationship between various atherogenic parameters at baseline and incident Mets at follow-up in different models

	Model 1		Model 2		Model 3
	OR (95 % CI)	P-value	OR (95 % CI)	P-value	OR (95 % CI)	P-value
TyG index	2.04 (1.19, 3.49)	0.009	2.41(1.37, 4.26)	0.002	2.43 (1.32, 4.44)	0.004
TyG-WC	1.01 (1.00, 1.01)	<0.001	1.01 (1.01, 1.02)	<0.001	1.01 (1.01, 1.02)	<0.001
TyG-WHtR	4.86 (2.98, 7.95)	<0.001	6.09 (3.57, 10.37)	<0.001	5.63 (3.23, 9.83)	<0.001
TyG-WHpR	1.48 (1.10, 2.01)	0.012	2.44 (1.69, 3.52)	<0.001	2.44 (1.66, 3.61)	<0.001

Model 1: non-adjusted model

Model 2: adjusted for age, gender, smoking, drinking, physical exercise

Model 3: adjusted for age, gender, smoking, drinking, physical exercise and components of Mets (included EH,SBP,DBP and HDL-C).

Mets, metabolic syndrome; EH, essential hypertension; SBP, systolic blood pressure; DBP, diastolic blood pressure; HDL-C, high-density lipoprotein cholesterol; TyG, triglyceride-glucose; TyG-WC, product of TyG and waist circumference; TyG-WHtR, product of TyG; TyG-WHpR, product of TyG and waist-to-hip ratio

To determine the consistency of the relationship between TyG related parameters and risk of Mets, we conducted stratified analyses (Table 3). For the non-adjusted model, TyG-related parameters significantly predicted Mets in both genders. TyG-WHtR was most strongly associated with Mets, the OR for Mets was 9.10 in males (P < 0.001) and 3.46 in females (P = 0.001). In Model 2, after adjusting for age, smoking, drinking and physical exercise, we found TyG-WHtR was the most strongly associated with Mets, the OR for Mets was 9.73 in males (P < 0.001) and 3.57 (P = 0.001) in females. After adjustments for components of Mets included HDL-C, SBP, DBP, and EH, only TyG-WHtR and TyG- WC significantly predicted Mets in both genders. The adjusted OR for TyG-WHtR in males was 9.14 (P < 0.001) compared with 3.18 (P = 0.005) in females.

Table 3

Hazards ratios with 95% confidence intervals for incident Mets increase in various atherogenic parameters in subgroups of gender

	Model 1			Model 2			Model 3
	HR (95%CI)	P value	P value for interaction	HR (95%CI)	P value	P value for interaction	HR (95%CI)	P value	P value for interaction
TyG index
Males	3.18 (1.50, 6.73)	0.003	0.262	3.25 (1.51, 6.99)	0.003	0.251	3.23 (1.44, 7.28)	0.005	0.289
Females	1.68 (0.73, 3.84)	0.220		1.67 (0.72, 3.89)	0.231		1.69 (0.69, 4.15)	0.251
TyG-WC
Males	1.01 (1.01, 1.02)	<0.001	0.338	1.01 (1.01, 1.02)	<0.001	0.295	1.01 (1.01, 1.02)	<0.001	0.194
Females	1.01 (1.01, 1.02)	<0.001		1.01 (1.01, 1.02)	<0.001		1.01 (1.00, 1.02)	<0.001
TyG-WHtR
Males	9.10 (4.35, 19.04)	<0.001	0.069	9.73 (4.55, 20.82)	<0.001	0.064	9.14 (4.16, 20.09)	<0.001	0.062
Females	3.46 (1.65, 7.26)	0.001		3.57 (1.68, 7.59)	0.001		3.18 (1.43, 7.08)	0.005
TyG-WHpR
Males	3.14 (1.92, 5.11)	<0.001	0.118	3.12 (1.90, 5.13)	<0.001	0.127	3.24 (1.92, 5.47)	<0.001	0.096
Females	1.75(1.01, 3.02)	0.044		1.76 (1.01, 3.06)	0.047		1.67 (0.92, 3.01)	0.089

Model 1: non-adjusted model

Model 2: adjusted for age,gender,smoking,drinking,physical exercise

Model 3: adjusted for age,gender,smoking,drinking,physical exercise and components of Mets (included EH,SBP,DBP and HDL-C)

Mets, metabolic syndrome; EH, essential hypertension; SBP, systolic blood pressure; DBP, diastolic blood pressure; HDL-C, high-density lipoprotein cholesterol; TyG, triglyceride-glucose; TyG-WC, product of TyG and waist circumference; TyG-WHtR, product of TyG; TyG-WHpR, product of TyG and waist-to-hip ratio

ROC curve analyses for TyG index and TyG-related parameters with Mets risk

The ROC curve analyses are shown in Fig. 2A–C and the corresponding AUCs (95% confidence interval, CI) in Tables 4, 5 shows the pairwise comparison of the AUCs of TyG index, TyG-WC, TyG-WHpR, and TyG-WHtR for the detection of Mets. In all participants, TyG-WHtR shows the largest AUC for Mets detection (0.686) followed by TyG-WC (0.660), TyG-WHpR (0.564) and TyG-index (0.556) in that order. Analysis revealed that TyG-WHtR has the largest AUC in all participants, suggesting that it has the best discriminating power to identify Mets in comparison with other parameters.

Fig. 2

ROC curves for the parameters for identifying Mets. A ROC curve for each parameter for identifying Mets in all participants. B ROC curve for Mets each parameter for identifying Mets in males. C ROC curve for each parameter for identifying Mets in females. Mets, metabolic syndrome; ROC, receiver operating characteristic; TyG, triglyceride-glucose; TyG-WC, product of TyG and waist circumference; TyG-WHtR, product of TyG and waist-to-height ratio; TyG-WHpR, product of TyG and waist-to-hip ratio

Table 4

The areas under the receiver operating characteristic curve for each parameter for identifying Mets

Variable	AUC	95%CI low	95%CI upp	Specificity	Sensitivity
All participants
TyG index	0.5776	0.5111	0.6345	0.2687	0.8545
TyG-WC	0.6771	0.6184	0.7203	0.2833	0.9364
TyG-WHpR	0.5793	0.5194	0.6323	0.2812	0.8273
TyG-WHtR	0.6967	0.6454	0.7484	0.3396	0.9364
Males
TyG index	0.5981	0.5057	0.6905	0.8439	0.3673
TyG-WC	0.7671	0.6973	0.8370	0.7134	0.7143
TyG-WHpR	0.6960	0.6196	0.7724	0.7229	0.5714
TyG-WHtR	0.7557	0.6850	0.8264	0.6783	0.7143
Females
TyG index	0.5568	0.4772	0.6364	0.3012	0.8852
TyG-WC	0.6956	0.6255	0.7657	0.3554	0.9672
TyG-WHpR	0.5992	0.5201	0.6783	0.5000	0.6885
TyG-WHtR	0.6493	0.5761	0.7225	0.2831	1.0000

Mets, metabolic syndrome; TyG, triglyceride-glucose; TyG-WC, product of TyG and waist circumference; TyG-WHtR, product of TyG; TyG-WHpR, product of TyG and waist-to-hip ratio

Table 5

Pairwise comparison of AUC of the different parameters.

	All	Males	Females
TyG_WHtR~TyG_index
Difference between areas	0.1307	0.1576	0.0925
P	< 0.0001	0.0002	0.0118
TyG_WHtR ~ TyG_WC
Difference between areas	0.0267	0.0114	0.0463
P	0.0380	0.4305	0.0030
TyG_WHtR~TyG_WHpR
Difference between areas	0.1224	0.0654	0.0501
P	<0.0001	0.0090	0.0422

AUC, area under curve; TyG, triglyceride-glucose; TyG-WC, product of TyG and waist circumference; TyG-WHtR, product of TyG; TyG-WHpR, product of TyG and waist-to-hip ratio

Pairwise comparison of the AUCs showed that compared with other parameters, TyG-WHtR was the best in detecting Mets in all the participants (TyG-WHtR vs. TyG index, P < 0.0001; TyG-WHtR vs. TyG-WC, P = 0.0380; TyG-WHtR vs. TyG-WHpR, P < 0.0001). In males TyG-WHtR was as good as TyG-WC (TyG-WHtR vs. TyG index, P = 0.0002; TyG-WHtR vs. TyG-WC, P = 0.4305; TyG-WHtR vs.TyG-WHpR, P = 0.0090) superior to other parameters in identifying Mets. In contrast, TyG-WC was better than TyG-WHtR (TyG-WHtR vs. TyG index, P = 0.0118; TyG-WHtR vs. TyG-WC, P = 0.0030; TyG-WHtR vs.TyG-WHpR, P = 0.0422) in detecting Mets in females.

Discussion

In this 15-year prospective follow-up study, we found that compared with the predictive ability of TyG index, TyG-WC and TyG-WHpR, TyG-WHtR with AUC of 0.683 was superior to other parameters for predicting Mets in all participants. Furthermore, this study demonstrated that TyG-related markers that combine obesity markers with TyG index are superior to other parameters in identifying Mets in both genders. Further, TyG-WHtR showed the highest OR in all participants and both genders before and after adjustment.

IR plays an important role in the pathophysiology of Mets, as it leads to decreased glucose metabolism, impaired insulin action, and alterations in hepatic lipid metabolism [20]. Because testing for insulin sensitivity is expensive, using the product of triglycerides and glucose as a surrogate marker to assess IR might help to minimize costs for clinical practice purpose [21]. TyG index, a product of triglycerides and glucose, was calculated as ln (fasting triglycerides (mg dl-1) × fasting glucose (mg dl-1)/2) [19]. A series of cohort and cross-sectional studies also confirmed that the TyG index can act as a better marker for predicting Mets [22-25].

We also observed that the TyG index is associated with Mets. The AUC of TyG index is 0.556 in our study. The overall AUC for TyG index in our study is lower than Nigerians [12], Pakistan [24] and Korean studies[25]. Although our cohort had a higher TG, lower FPG, similar TyG index, the overall AUC for TyG index in our study was even lower than other Chinese studies [23, 26, 27]. These may imply that there are not just ethnic differences, but regional differences among human subjects with regards to identifying Mets.

Several anthropometric indicators have been linked to Mets [10]. Lim et al. reported that a combination of TyG index and anthropometric indices was more accurate than TyG alone in predicting IR [11]. Taiwo et al. found that TyG-WHtR is better than TyG index and other TyG-related parameters for predicting the risk of Mets in Nigerians [12]. This result was consistent with ours. We found that TyG-WHtR is superior to other parameters for predicting the risk of Mets in an urban Chinese population. The superiority of WHtR on predicting MetS might be attributed to the fact that it takes into account height variability and, therefore, is more accurate at representing central adiposity [28].

Consist with precious study [12], we carried out subgroup analysis by genders and found that TyG-WHtR as well as TyG-WC outperformed other indices in males at 15-year follow-up. Further, TyG-WHtR showed the highest OR before and after adjustment. Therefore, TyG-WHtR appears to be the best of all the parameters among all participants and males. Abdominal obesity includes both subcutaneous and visceral adipose tissue [29]. Visceral (intra-abdominal) fat is found to correlate more with cardiovascular risk, because they produce more fatty acids and secrete inflammatory cytokines and adipokines [30, 31]. Both WC and WHtR are markers of visceral adiposity [11]. Because WHtR corrected for height, it may be better than WC to predict Mets and cardiovascular risk [32, 33]. The present study found that the accumulation of visceral adipose tissue accelerates the epigenetic age mostly mediated by TyG index in males [34]. Our study found that the males’ average age was significantly older than the females. Therefore, we think that the TyG-WHtR was a significant predictor for Mets due to age-related metabolic dysfunctions occurring in adipose tissue in males.

Our study has strength and limitations. In this study, we lacked the information about the drugs used which might influence the levels of serum lipids and the risk for subsequent Mets, and long term usage of these drugs could influence our results. Usually, the individuals take medicine erratically in China, so that might not influence the results in our study. Moreover, Mets is just a complex multifactorial health problem, and it has limited practical utility as a diagnostic or management tool, but it is worthwhile to further elucidate the underlying pathways of the clustering of such a lot of risk factors.

Conclusion

The findings indicate that TyG-WHtR is superior to other parameters for predicting the risk of Mets in an urban Chinese population. The present study also reveals that TyG-related markers that combine obesity markers with TyG index are superior to other parameters in identifying Mets in both genders.