Association of hypertension with the triglyceride–glucose index and its associated indices in the Chinese population: A 6-year prospective cohort study

Abstract

The authors aim to assess the correlation between hypertension and the triglyceride-glucose (TyG) index and its associated indices, and to compare their abilities to identify hypertension. Four thousand eight hundred and sixty-six non-hypertensive participants were enrolled from the China National Health Survey in 2009. The data on new-onset hypertension were gathered in both 2011 and 2015. The TyG index and its associated indices were derived from the fasting triglyceride, blood glucose levels, and anthropometric parameters. Multivariate logistic regression analyses and receiver operating characteristic curve analysis were used. The adjusted odds ratio (OR) and 95% confidence interval (CI) for the new-onset hypertension for the TyG-waist-to-height ratio (TyG-WHtR), TyG-waist circumference (TyG-WC), TyG-waist-to-hip ratio (TyG-WHR), TyG-body mass index (TyG-BMI), and TyG index were 1.379 (1.230–1.546), 1.002 (1.001–1.003), 1.156 (1.069–1.251), 1.007 (1.005–1.009), and 1.187 (1.051–1.341), respectively. In addition, comparing the lowest quartile (Q1) group with the highest quartile (Q4), the adjusted OR and 95% CI for the new-onset hypertension were found to be 1.86 (1.48–2.35), 1.93 (1.53–2.43), 1.71 (1.36–2.16), 2.00 (1.60–2.50), and 1.49 (1.19–1.88) for TyG-WHtR, TyG-WC, TyG-WHR, TyG-BMI, and TyG index, respectively, among all participants. The TyG-WHtR had the largest area under the curve (AUC) for hypertension (AUC, 0.628; 95% CI, 0.614–0.641) in all participants. Stratified analysis also indicated that the TyG-WHtR exhibited the greatest AUC in both males (AUC, 0.608; 95% CI, 0.587–0.629) and females (AUC, 0.648; 95% CI, 0.629–0.666). In conclusions, the TyG index and its associated indices were positively associated with hypertension. Among these indices, TyG-WHtR was the most valuable indicator for predicting hypertension.

INTRODUCTION

Hypertension is the most prevalent modifiable risk factor for cardiovascular disease,¹ which poses a great threat to public health. The incidence of hypertension in China has markedly risen over the course of the last three decades, affecting approximately one-quarter of all Chinese individuals.² Therefore, an efficient and reliable predictor for the occurrence of hypertension is urgently required to prevent this rapidly rising trend of hypertension.

Insulin resistance (IR) refers to a state of decreased sensitivity and responsiveness to the metabolic actions of insulin.³ It is associated with hypertension and has been regarded not only as a pathogenic cause but also as a predictor of hypertension.⁴ The hyperinsulinemic euglycemic clamp is regarded as the preferred method for assessing insulin resistance.⁵ Nevertheless, as a result of the intricate nature and large costs associated with this approach, its utilization is limited in clinical practice. Hence, it is imperative to develop surrogate biomarkers of insulin resistance that are both accessible and reliable. In recent years, the triglyceride-glucose (TyG) index has emerged as a potential surrogate biomarker for insulin resistance.⁶ Accumulating evidence indicates a correlation between the TyG index and cardiovascular disease.⁷ Nevertheless, there is a scarcity of cohort studies that particularly investigate the relationship between the TyG index and high blood pressure.^8–12

Additionally, obesity has been linked to IR and hypertension.^13,14 Obesity indicators, such as body mass index, waist circumference, waist-to-hip ratio, and waist-to-height ratio, have been widely used due to their convenient utilization. Numerous studies have investigated the correlation of TyG-associated indices that combined obesity indicators and the TyG index with IR,¹⁵ diabetes,¹⁶ metabolic syndrome,¹⁷ and coronary artery calcification progression.¹⁸ Nevertheless, there is a lack of prospective cohort studies that investigate the correlation between TyG-associated indices and hypertension, as well as assess the corresponding predictive capacities for hypertension in the Chinese population. Therefore, the objective of this study was to assess the association between the TyG index and its associated indices and the likelihood of developing hypertension, as well as to compare the predictive performance of the TyG index, TyG-BMI, TyG-WC, and TyG-WHtR for the risk of hypertension.

METHODS Study population

The data were obtained from the China Health and Nutrition Survey (CHNS). It was a prospective longitudinal survey launched in 1989 among the Chinese population. Subsequent surveys were administered at intervals of 2–4 years, resulting in a total of 10 follow-up waves conducted in 1989, 1991, 1993, 1997, 2000, 2004, 2006, 2009, 2011, and 2015. Additional information regarding the CHNS can be obtained on the official website: https://www.cpc.unc.edu/projects/china. In this work, we employed three waves of the CHNS dataset spanning from 2009 to 2015 because blood measurements became available for the first time in 2009. First, among the 9516 eligible participants in 2009, those under the age of 18 (n = 836) were excluded. Second, participants with hypertension at baseline in 2009 were excluded (n = 2660) among the 8680 participants. In addition, participants lost to follow-up visits (n = 888) were also omitted. Finally, among the 5132 participants, participants without anthropometric measures and serum biochemical indicators were discarded (n = 266). Consequently, a total of 4866 participants were included in the final study, as depicted in Figure 1. Prior to data collection, informed consent was obtained from all participants involved in the CHNS. Ethical approval for data collection in the CHNS was granted by the institutional review committees of the University of North Carolina at Chapel Hill, the National Institute of Nutrition and Food Safety, and the Chinese Center for Disease Control and Prevention.

FIGURE 1. Flow chart of the participants in the study.

Anthropometrics and blood biomarker measurements

Trained researchers measured body weight, height, waist circumference, and hip circumference according to predefined procedures. BP was measured by experienced physicians using standard mercury sphygmomanometers at each visit. The participants were asked to rest and sit quietly for 5 min. Three measures of seated SBP and DBP were taken on the same arm, and the mean of these three readings was utilized for analysis.¹⁹ Hypertension was defined as systolic and/or diastolic blood pressure over 140/90 mm Hg, antihypertensive medication use, or previous diagnosis by a physician.²⁰ Venous blood samples were collected from individuals in a fasting condition and analyzed at the laboratory of the Ministry of Health at the China-Japan Friendship Hospital under strict quality control. The TyG index and its associated indices were calculated using the following formula: [Image Omitted. See PDF]

TyG was modified by multiplying it by BMI, WC, WHR, and WHtR to create TyG-BMI, TyG-WC, TyG-WHR, and TyG-WHtR, respectively: [Image Omitted. See PDF][Image Omitted. See PDF][Image Omitted. See PDF][Image Omitted. See PDF]

Statistical analysis

Data are presented as the mean ± SD or median (interquartile range) for continuous variables and frequencies (percentages) for categorical variables. Independent two-sample t tests or Mann‒Whitney U tests were used to compare the differences in continuous variables. The differences in categorical variables were compared using the chi-square test. The odds ratio (OR) and 95% confidence interval (95% CI) were calculated using univariate and multivariate logistic regression models. Model 1 was not adjusted. Model 2 adjusted for age, sex, current smoker, consumed alcohol, urban residence, SBP, and DBP. Model 3 adjusted for, in addition to variables included in Model 2, uric acid, hsCRP, eGFR, total cholesterol, hemoglobin concentration, HbA1c, and urea. Additionally, for each quartile of the TyG index and its associated indices, multivariable logistic regression analyses were used to calculate the OR and 95% CI of hypertension, with the lowest quartile (Q1) serving as the reference. Similarly, the OR and 95% CI for the risk of hypertension based on different sex among the TyG index and its associated indices were estimated, and their interactions were tested. To evaluate the predictive value of the TyG index and its associated indices for new-onset hypertension, the area under the curve (AUC) was assessed through receiver operating characteristic (ROC) curve analysis. Pairwise comparisons between AUC for these indices were performed. A two-sided p<.05 was considered statistically significant. All analyses were performed using Empower (R) (http://www.emPowerstats.com, X&Y solutions, Inc., Boston MA) and the Statistical Package for Social Sciences Version 24.0 (SPSS Inc., Chicago, IL, USA).

RESULTS Baseline characteristics of the participants

This study included 4866 participants. A total of 1256 out of the 4866 individuals developed hypertension after a 6-year follow-up period (602 individuals with new-onset hypertension in 2011 and 654 individuals with new-onset hypertension in 2015). Table 1 displays the baseline characteristics of the study population, categorized according to the presence or absence of hypertension. Participants with hypertension were prone to be older, had a higher proportion of males, were current smokers, consumed alcohol and lived in an urban area. Those who developed hypertension had greater BMI, waist circumference, hip circumference, WHR, WHtR, SBP, DBP, uric acid, urea, hsCRP, triglycerides, TC, LDL-C, apolipoprotein B, HbA1c, FBG, and hemoglobin concentration but lower eGFR. TyG index, TyG-BMI, TyG-WC, TyG-WHR, and TyG-WHtR values were significantly different between individuals with and without hypertension.

TABLE 1 Baseline characteristics of the participants according to hypertension.

Characteristics	Overall (n = 4866)	Non-HTN (n = 3610)	HTN (n = 1256)	p value
Age (years)	47.8 ± 13.8	45.8 ± 13.6	53.8 ± 12.6	<.001
Male, n (%)	2173(44.7)	1568(43.4)	605 (48.2)	.004
Current smoker, n (%)	1348(27.7)	968(26.8)	380 (30.3)	.019
Consumed alcohol, n (%)	1565(32.2)	1115(30.9)	450 (35.8)	.001
Urban residence, n (%)	1465(30.1)	1148(31.8)	317 (25.2)	<.001
Height (cm)	160.9 ± 8.4	160.9 ± 8.3	160.7 ± 8.5	.300
Weight (cm)	59.3 ± 10.5	58.5 ± 10.2	61.3 ± 11.1	<.001
BMI (kg/m2)	22.8 ± 3.2	22.5 ± 3.1	23.6 ± 3.3	<.001
Waist circumference (cm)	80.8 ± 9.6	79.8 ± 9.4	83.7 ± 9.7	<.001
Hip circumference (cm)	93.3 ± 7.5	92.8 ± 7.3	94.9 ± 7.9	<.001
WHR	0.86 ± 0.08	0.86 ± 0.80	0.88 ± 0.07	<.001
WHtR	0.50 ± 0.06	0.49 ± 0.05	0.52 ± 0.06	<.001
SBP (mm Hg)	116.2 ± 11.2	114.5 ± 11.0	121.2 ± 10.3	<.001
DBP (mm Hg)	75.8 ± 7.5	74.9 ± 7.5	78.6 ± 6.7	<.001
Uric acid (μmol/L)	283(230–345)	279(225–342)	296(239–359)	<.001
Urea (mmol/L)	5.35 ± 1.47	5.26 ± 1.46	5.58 ± 1.48	<.001
hsCRP (mg/L)	1.00(0.00–2.00)	1.00(0.00–2.00)	1.00 (1.00–2.00)	<.001
eGFR (mL/min/1.73m2)	82.2 ± 15.9	83.6 ± 15.8	78.1 ± 15.5	<.001
Triglycerides (mmol/L)	1.18(0.81–1.81)	1.14(0.78–1.74)	1.33 (0.91–2.02)	<.001
TC (mmol/L)	4.78 ± 0.97	4.73 ± 0.96	4.94 ± 0.99	<.001
LDL-C (mmol/L)	2.92 ± 0.95	2.87 ± 0.92	3.06 ± 1.01	<.001
HDL-C (mmol/L)	1.45 ± 0.43	1.45 ± 0.44	1.43 ± 0.40	.069
Apolipoprotein A-1 (g/L)	1.15 ± 0.36	1.15 ± 0.34	1.16 ± 0.41	.223
Apolipoprotein B (g/L)	0.88 ± 0.25	0.86 ± 0.25	0.93 ± 0.26	<.001
Lipoprotein (a) (mg/L)	78(40–166)	76(39–166)	83 (42–168)	.154
Total protein (g/L)	77.0 ± 5.1	77.0 ± 5.1	77.1 ± 5.3	.420
Albumin (g/L)	47.3 ± 3.4	47.3 ± 3.4	47.4 ± 3.4	.288
HbA1c (%)	5.52 ± 0.79	5.47 ± 0.72	5.68 ± 0.95	<.001
FBG (mmol/L)	5.26 ± 1.31	5.19 ± 1.23	5.48 ± 1.49	<.001
Hemoglobin concentration (g/L)	140.2 ± 20.3	139.4 ± 20.2	142.4 ± 20.6	<.001
White blood cell (10^9/L)	6.21 ± 1.85	6.21 ± 1.91	6.19 ± 1.67	.796
TyG index	8.55 ± 0.69	8.50 ± 0.69	8.68 ± 0.68	<.001
TyG-BMI	195.8 ± 36.1	192.3 ± 35.2	206.1 ± 36.8	<.001
TyG-WC	693.7 ± 116.2	681.2 ± 113.6	729.8 ± 116.1	<.001
TyG-WHR	7.42 ± 1.01	7.33 ± 1.02	7.67 ± 0.94	<.001
TyG-WHtR	4.31 ± 0.71	4.23 ± 0.69	4.54 ± 0.71	<.001

Abbreviations: BMI, body mass index; DBP, diastolic blood pressure; eGFR, estimated glomerular filtration rate; FBG, fasting blood glucose; HDL-C, high density lipoprotein cholesterol; hsCRP, hypersensitive c-reactive protein; HTN, hypertension; LDL-C, low density lipoprotein cholesterol; SBP, systolic blood pressure; TC, total cholesterol; TyG, triglyceride-glucose; WC, waist circumference; WHR, waist-to-hip ratio; WHtR, waist-to-height ratio.

The relationship between hypertension and the continuous TyG index and its associated indices

To elucidate whether the TyG index and its associated indices were associated with hypertension, logistic regression analyses were performed. The TyG-associated indices were individually included in the logistic models. Table 2 shows the logistic regression model for hypertension. The OR and 95% CI of the TyG index, TyG-BMI, TyG-WC, TyG-WHR, and TyG-WHtR for hypertension were 1.445 (1.320–1.581), 1.010 (1.009–1.012), 1.004 (1.003–1.004), 1.399 (1.311–1.493), and 1.818 (1.661–1.990), respectively. After adjusting for age, sex, current smoker, consumed alcohol, urban residence, SBP, DBP, uric acid, hsCRP, eGFR, TC, hemoglobin concentration, urea, and HbA1c in model 3, the TyG index and its associated indices were still associated with a higher risk of hypertension for the entire population.

TABLE 2 Logistic regression analyses for the relationship between TyG-associated indices and hypertension.

	Model 1		Model 2		Model 3
	OR (95%CI)	p value	OR (95%CI)	p value	OR (95%CI)	p value
TyG index	1.445(1.320–1.581)	<.001	1.227(1.113–1.353)	<.001	1.187(1.051–1.341)	.006
TyG-BMI	1.010(1.009–1.012)	<.001	1.007(1.005–1.009)	<.001	1.007(1.005–1.009)	<.001
TyG-WC	1.004(1.003–1.004)	<.001	1.002(1.001–1.003)	<.001	1.002(1.001–1.003)	<.001
TyG-WHR	1.399(1.311–1.493)	<.001	1.182(1.104–1.266)	<.001	1.156(1.069–1.251)	<.001
TyG-WHtR	1.818(1.661–1.990)	<.001	1.392(1.260–1.537)	<.001	1.379(1.230–1.546)	<.001

Model 1: unadjusted.

Model 2: age, sex, current smoker, consumed alcohol, urban residence, SBP, DBP.

Model 3: Model 2+uric acid, hsCRP, eGFR, TC, hemoglobin concentration, urea, HbA1c.

The relationship between hypertension and the categorized TyG index and its associated indices

Each parameter of the TyG-associated indices was used to categorize participants into quartile groups. In unadjusted model 1, the TyG index, TyG-BMI, TyG-WC, TyG-WHR, and TyG-WHtR in the fourth quartile demonstrated a positive association with hypertension (OR = 2.17, 95% CI: 1.80–2.63, p < .001 for TyG index; OR = 2.73, 95% CI: 2.26–3.31, p < .001 for TyG-BMI; OR = 3.36, 95% CI: 2.76–4.09, p < .001 for TyG-WC; OR = 3.04, 95% CI: 2.49–3.70, p < .001 for TyG-WHR; OR = 3.34, 95% CI: 2.74–4.08, p < .001 for TyG-WHtR) when compared to the first quartile (Table 3). In model 2, the association between these variables and hypertension remained significant (Table 3). After adjusting for additional covariates such as uric acid, hsCRP, eGFR, TC, hemoglobin concentration, urea, and HbA1c in model 3, the associations between hypertension and the TyG index, TyG-BMI, TyG-WC, TyG-WHR, and TyG-WHtR were found to be statistically significant in the 4th quartile compared to the 1st quartile (OR = 1.49, 95% CI: 1.19–1.88, p < .001 for TyG index; OR = 2.00, 95% CI: 1.60–2.50, p < .001 for TyG-BMI; OR = 1.93, 95% CI: 1.53–2.43, p < .001 for TyG-WC; OR = 1.71, 95% CI: 1.36–2.16, p < .001 for TyG-WHR; OR = 1.86, 95% CI: 1.48–2.35, p < .001 for TyG-WHtR).

TABLE 3 Logistic regression analyses for hypertension of quartiles of the TyG-associated indices.

	Model 1		Model 2		Model 3
	OR (95%CI)	p value	OR (95%CI)	p value	OR (95%CI)	p value
TyG index
Q1 (<8.06)	1	reference	1	reference	1	reference
Q2 (8.06–8.46)	1.40(1.15–1.71)	.001	1.28 (1.03–1.58)	.022	1.26(1.01–1.56)	.034
Q3(8.46–8.95)	1.81(1.49–2.20)	<.001	1.48 (1.20–1.81)	<.001	1.44(1.17–1.79)	.001
Q4(>8.95)	2.17(1.80–2.63)	<.001	1.59(1.30–1.95)	<.001	1.49(1.19–1.88)	.001
TyG-BMI
Q1(<168.9)	1	reference	1	reference	1	reference
Q2(168.9–191.7)	1.27(1.04–1.56)	.019	1.18(0.95–1.47)	.134	1.17(0.93–1.45)	.163
Q3(191.7–218.7)	2.07(1.70–2.51)	<.001	1.64 (1.33–2.02)	<.001	1.61(1.30–1.99)	<.001
Q4(>218.7)	2.73(2.26–3.31)	<.001	2.06 (1.67–2.54)	<.001	2.00(1.60–2.50)	<.001
TyG-WC
Q1(<605.9)	1	reference	1	reference	1	reference
Q2(605.9–680.6)	1.76(1.43–2.16)	<.001	1.34 (1.08–1.67)	.008	1.33 (1.06–1.66)	.012
Q3(680.6–768.1)	2.36(1.93–2.89)	<.001	1.61 (1.29–1.99)	<.001	1.58 (1.27–1.97)	<.001
Q4(>768.1)	3.36(2.76–4.09)	<.001	2.01 (1.62–2.49)	<.001	1.93 (1.53–2.43)	<.001
TyG-WHR
Q1(<6.72)	1	reference	1	reference	1	reference
Q2(6.72–7.31)	1.79(1.46–2.21)	<.001	1.33 (1.07–1.66)	.011	1.31 (1.05–1.64)	.015
Q3(7.31–7.98)	2.55(2.09–3.12)	<.001	1.63 (1.31–2.02)	<.001	1.59 (1.28–1.98)	<.001
Q4(>7.98)	3.04(2.49–3.70)	<.001	1.80 (1.46–2.23)	<.001	1.71 (1.36–2.16)	<.001
TyG-WHtR
Q1(<3.78)	1	reference	1	reference	1	reference
Q2(3.78–4.25)	1.81(1.47–2.22)	<.001	1.37 (1.10–1.71)	.004	1.36 (1.09–1.70)	.006
Q3(4.25–4.75)	2.40(1.96–2.94)	<.001	1.59 (1.28–1.98)	<.001	1.56 (1.25–1.95)	<.001
Q4(>4.75)	3.34(2.74–4.08)	<.001	1.95 (1.57–2.41)	<.001	1.86 (1.48–2.35)	<.001

Model 1: unadjusted.

Model 2: age, sex, current smoker, consumed alcohol, urban residence, SBP, DBP.

Model 3: Model 2+Uric acid, hsCRP, eGFR, TC, hemoglobin concentration, urea, HbA1c.

Sex-specific TyG index and its associated indices comparison

Table 4 illustrates notable disparities between the male and female populations. The TyG index and its associated indices (TyG-BMI, TyG-WC, and TyG-WHR) exhibited significantly higher values in males than in females, but the TyG-WHtR showed significantly lower values in males than females.

TABLE 4 Sex-specific TyG index and its associated indices comparison.

Variable	Males (n = 2173)	Females (n = 2693)	p value
TyG index	8.62 ± 0.75	8.49 ± 0.64	<.001
TyG-BMI	197.3 ± 37.5	194.6 ± 34.9	.010
TyG-WC	713.9 ± 119.4	677.5 ± 111.0	<.001
TyG-WHR	7.60 ± 0.97	7.27 ± 1.02	<.001
TyG-WHtR	4.28 ± 0.69	4.34 ± 0.72	.001

Sex-specific subgroup analysis between the TyG index and its associated indices to predict hypertension

To assess the reliability of the correlation between TyG-associated indices and the likelihood of developing hypertension, we conducted stratified analyses, as shown in Table 5. It is worth noting that the TyG index exhibited a more pronounced positive association in the male subgroup than in the female subgroup (OR = 1.249, 95% CI: 1.058–1.475 vs. OR = 1.101, 95% CI: 0.919–1.319) (p value for interaction = .312). The association between TyG-associated indices and hypertension was not moderated by sex (p for interaction > .05). In adjusted model 3, the variable TyG-WHtR exhibited the strongest association with new-onset hypertension. The adjusted OR for hypertension was 1.395 in males (p < .001) and 1.366 in females (p < .001).

TABLE 5 Logistic regression analyses for hypertension of the TyG-associated indices by sex.

Variable	Model 1		Model 2		Model 3
	OR (95%CI)	p value	OR (95%CI)	p value	OR (95%CI)	p value
TyG index
Males	1.328 (1.175–1.502)	<.001	1.270 (1.115–1.446)	<.001	1.249 (1.058–1.475)	.008
Females	1.563 (1.366–1.788)	<.001	1.139 (0.981–1.323)	.088	1.101 (0.919–1.319)	.295
p for interaction		.080		.281		.312
TyG-BMI
Males	1.009 (1.006–1.011)	<.001	1.008 (1.005–1.011)	<.001	1.008 (1.005–1.011)	<.001
Females	1.012 (1.009–1.014)	<.001	1.007 (1.004–1.010)	<.001	1.007 (1.004–1.010)	<.001
p for interaction		.122		.542		.539
TyG-WC
Males	1.003 (1.002–1.004)	<.001	1.002 (1.001–1.003)	<.001	1.002 (1.001–1.003)	<.001
Females	1.004 (1.003–1.005)	<.001	1.002 (1.001–1.003)	<.001	1.002 (1.001–1.003)	<.001
p for interaction		.009		.960		.983
TyG-WHR
Males	1.348 (1.225–1.483)	<.001	1.230 (1.112–1.362)	<.001	1.219 (1.085–1.368)	<.001
Females	1.427 (1.304–1.563)	<.001	1.128 (1.026–1.239)	.012	1.112 (1.006–1.229)	.037
p for interaction		.393		.216		.218
TyG-WHtR
Males	1.663 (1.454–1.902)	<.001	1.406 (1.217–1.624)	<.001	1.395 (1.191–1.634)	<.001
Females	1.990 (1.760–2.251)	<.001	1.379 (1.205–1.579)	<.001	1.366 (1.183–1.579)	<.001
p for interaction		.053		.849		.834

Model 1: unadjusted.

Model 2: age, current smoker, consumed alcohol, urban residence, SBP, DBP.

Model 3: Model 2+uric acid, hsCRP, eGFR, TC, hemoglobin concentration, urea, HbA1c.

Assessment of the TyG index and its associated indices for hypertension diagnostic accuracy

The ROC curve analyses are displayed in Figure 2(A–C), and Table 6 contains the matching AUC and 95% CI. TyG-WHtR had the greatest AUC value (0.628) for the identification of hypertension among all participants, followed by TyG-WC (0.624), TyG-BMI (0.614), TyG-WHR (0.612), and TyG-index (0.583). In comparison to other indices, analysis showed that TyG-WHtR had the largest AUC in both females and males, indicating that it had the strongest ability to identify hypertension. According to pairwise comparison of the AUC, TyG-WHtR outperformed other indices in detecting new-onset hypertension in all participants (TyG-WHtR vs. TyG index, p < .001; TyG-WHtR vs. TyG-BMI, p = .011; TyG-WHtR vs. TyG-WHR, p = .002), with the exception of TyG-WC (TyG-WHtR vs. TyG-WC, p = .215). The AUC of TyG-WHtR was significantly larger than that of the TyG index (p < .001), TyG-WC (p = .019) and TyG-WHR (p = .008), except for TyG-BMI (p = .565) in males. The AUC of TyG-WHtR was substantially greater than that of the other four indices in women, as shown in Table 7.

FIGURE 2. ROC curve for these indices for identifying new-onset hypertension.

TABLE 6 The receiver operating characteristic curve for each index for identifying hypertension.

Variable	AUC	95%CI	Specificity	Sensitivity
All participants
TyG index	0.583	0.569–0.597	53.32	59.39
TyG-BMI	0.614	0.600–0.627	58.81	60.35
TyG-WC	0.624	0.610–0.637	47.53	71.26
TyG-WHR	0.612	0.598–0.626	56.48	61.31
TyG-WHtR	0.628	0.614–0.641	63.02	55.81
Males
TyG index	0.570	0.549–0.591	48.28	62.98
TyG-BMI	0.604	0.583–0.625	57.21	61.49
TyG-WC	0.600	0.580–0.621	61.16	53.72
TyG-WHR	0.592	0.571–0.612	50.19	66.45
TyG-WHtR	0.608	0.587–0.629	44.07	72.56
Females
TyG index	0.592	0.573–0.611	42.21	72.04
TyG-BMI	0.621	0.602–0.639	61.02	58.83
TyG-WC	0.640	0.621–0.658	55.24	66.82
TyG-WHR	0.624	0.605–0.642	39.62	80.65
TyG-WHtR	0.648	0.629–0.666	60.38	62.52

TABLE 7 Pairwise comparison of AUC of the different indices.

Variable	All	Males	Females
TyG-WHtR ∼ TyG index
Difference between areas	0.044	0.038	0.055
p value	<0.001	<0.001	<0.001
TyG-WHtR ∼ TyG-BMI
Difference between areas	0.013	0.004	0.026
p value	0.011	0.565	<0.001
TyG-WHtR ∼ TyG-WC
Difference between areas	0.003	0.008	0.008
p value	0.215	0.019	0.019
TyG-WHtR ∼ TyG-WHR
Difference between areas	0.015	0.016	0.023
p value	0.002	0.008	<0.001

DISCUSSION

In this prospective study, the TyG index and its associated indices were evaluated for their capacities to predict the occurrence of hypertension. The results showed that TyG-associated indices were significantly associated with hypertension in both male and female individuals, whereas the TyG index showed a significant correlation with hypertension exclusively in men after controlling for confounding factors. The predictive power of TyG-WHtR for hypertension was the largest in both male and female individuals and was significantly superior to that of other indices.

Recent studies showed that the TyG index was closely linked to a higher risk of hypertension,²¹ which supported our findings. However, unlike previous studies, we found that the TyG index was correlated with the risk of hypertension in male individuals (OR = 1.249, 95% CI 1.058–1.475) but not in female individuals (OR = 1.101, 95% CI 0.919–1.319). A recent study also performed a sex-specific comparison between the TyG index and hypertension and found that the TyG index was significantly correlated with hypertension only in female individuals.²² This conflicted with the results of our study. One potential explanation was that our study enrolled a wider spectrum age of participants, whereas this study mainly enrolled a population of middle-aged and older adults. Another reason could be ethnic differences in that the participants of our study were Chinese, but in that trial, they were Koreans. Different lifestyles in both countries might lead to inconsistent results. Despite the fact that the causes of the different relationships between the TyG index and hypertension in male and female individuals are not fully understood, there is a potential explanation. A prior study conducted in the Chinese population indicated a higher degree of effect of the TyG index on the risk of metabolic syndrome among male individuals,²³ which would increase the susceptibility to hypertension for men.

As an alternative biomarker of IR, the TyG index may exert an influence on susceptibility to hypertension by activating the RAAS and sympathetic nervous system²⁴ or increasing arterial stiffness.²⁵ Although our results showed that the TyG index was correlated with the risk of hypertension in all participants and male individuals, the TyG index for hypertension showed inadequate predictive ability in all participants (AUC, 0.583 [95% CI, 0.569–0.597]) and male individuals (AUC, 0.570 [95% CI, 0.549–0.591]).

The concept of TyG-associated indices was initially proposed by Ko and coworkers,²⁶ who integrated the TyG index with obesity indices of WC, BMI, WHR, and WHtR. Increasing evidence has shown that obesity is also correlated with IR and hypertension.^13,14 Thus, the integration of obesity indicators with the TyG index could provide more information. Our study indicated that the TyG-associated indices were also correlated with the risk of hypertension. In addition, among the TyG-associated indices, the diagnostic performance of TyG-WHtR was shown to be superior to that of TyG-BMI, TyG-WC, and TyG-WHR. The findings were consistent with prior research. A population-based cross-sectional study in Henan, China, demonstrated that the AUC of TyG-WHtR was higher than that of the TyG index, TyG-BMI and TyG-WC in predicting hypertension.²⁷ In addition, a nationwide cohort study from the China Health and Retirement Longitudinal Study also confirmed this result.²⁸ The most likely cause of the occurrence of this phenomenon was that WHtR was considered a more effective measure for identifying abdominal obesity compared to BMI, WC, and WHR due to its comprehensive consideration of both WC and height.^13,29 A study carried out in Thailand performed a comparative analysis of nine obesity indices in relation to hypertension, indicating that WHtR was the most useful indicator of obesity linked with hypertension.³⁰ The biological explanation is that the WHtR can better reflect the distribution of adipose tissue. Adiposity can lead to the pro-inflammatory and pro-oxidative milieu that promotes vascular remodeling.³¹ Besides, the additional autocrine and paracrine activities of adipose tissue also can contribute to inappropriate activation of the RAAS and sympathetic nervous system that promote hypertension.^32,33 Therefore, combining the TyG index and WHtR together can further improve the ability to predict hypertension. Nevertheless, it is worth noting that the best TyG-associated indices for prediction of hypertension remain controversial with incompatible conclusions. A cross-sectional study conducted in 16 834 participants from china showed that TyG-WC exhibited the highest diagnostic efficacy for hypertension, followed by TyG-WHtR, TyG-BMI, TyG-WHR, and TyG index.³⁴ But, due to the nature of the cross-sectional design, causality cannot be determined. Therefore, more cohort studies are needed to determine which of the TyG-associated indicators can better predict new-onset hypertension.

The strengths of this research were the cohort study design, relatively large sample size and sex-specific analysis method, which could establish a causal relationship. However, this research had several drawbacks. First, some unknown factors still existed, even if all relevant confounders were taken into account. Secondly, nearly 20% of participants were lost follow up or lack of baseline data. In addition, the study solely assessed anthropometrics and blood measures (viz, fasting blood glucose and triglycerides) at baseline, without considering their changes that might have occurred over time. Finally, this study was conducted in Chinese adults, and it was unclear whether the findings could be generalized to other ethnicities.

CONCLUSIONS

The TyG index and its associated indices were positively associated with hypertension in the Chinese population. The TyG-associated indices could improve the identification of hypertension. Among these indices, TyG-WHtR was the most valuable indicator for predicting the risk of hypertension.

AUTHOR CONTRIBUTIONS

Changqiang Yang, Yue Song, and Xinquan Wang were listed as the co-first author and contributed equally to the study. Jixin Hou and Peijian Wang were listed as corresponding author. Changqiang Yang and Yue Song analyzed the data and drafted the manuscript. Xinquan Wang and Yi Yang helped with the interpretation of the data and revised the manuscript. Yaqiong Zhou and Dan Wang offered assistance and recommendations for the data analysis. Jixin Hou and Peijian Wang provided direction and revised the manuscript. All authors read and approved the final manuscript.

ACKNOWLEDGMENTS

The authors express their gratitude to the CHNS research team and all volunteers and staff members who participated in this research endeavor.

CONFLICT OF INTEREST STATEMENT

The authors declare that they have no conflict of interest.

PATIENT CONSENT STATEMENT

Informed consent was obtained from all participants involved in the China Health and Nutrition Survey (CHNS).

PERMISSION TO REPRODUCE MATERIAL FROM OTHER SOURCES

No reproduced material from other sources in the study.

CLINICAL TRIAL REGISTRATION

Clinical trial registration was waived for this study due to the use of publicly available data.

DATA AVAILABILITY STATEMENT

The raw data are available (https://www.cpc.unc.edu/projects/china).