Association between the Metabolic Score for Visceral Fat (METS-VF) and incident hypertension in middle-aged and older Chinese adults: a nationwide prospective cohort study

Abstract

Background

Emerging evidence highlights the critical role of visceral adiposity in the pathogenesis of hypertension, with the Metabolic Score for Visceral Fat (METS-VF) emerging as a novel biomarker that requires validation in Chinese populations.

Methods

This prospective cohort study included adults aged ≥ 45 years without hypertension at baseline from the China Health and Retirement Longitudinal Study (CHARLS).From the 2011–2012 CHARLS cohort, 6775 participants were included in the final analysis after excluding those with age < 45, insufficient follow - up, missing clinical data, or baseline hypertension with incomplete questionnaires. We evaluated the longitudinal association between METS-VF and hypertension incidence and compared its predictive performance with conventional adiposity indices, including the visceral adiposity index (VAI) and lipid accumulation product (LAP). Hypertension cases were identified through biennial assessments over a median follow-up of 8 years.

Results

A total of 2,845 incident hypertension cases were documented. Multivariable Cox regression models revealed a significant dose-response relationship, with participants in the highest METS-VF quartile exhibiting a 2.11-fold higher risk of hypertension compared to those in the lowest quartile (95% CI: 1.85–2.42; P < sub > trend</sub > < 0.001). The absolute 8-year hypertension risk was 18.3% higher in the highest versus lowest quartile. Restricted cubic spline analysis identified a nonlinear threshold effect at a METS-VF value of 5.68, beyond which hypertension risk increased sharply by 134% (HR = 2.34). Each one-unit increase in METS-VF was associated with a 2.4% elevated hypertension risk (HR = 1.02, 95% CI: 1.00-1.05). Receiver operating characteristic analysis demonstrated that METS-VF had superior discriminative ability (AUC = 0.62) compared to LAP (AUC = 0.60) and VAI (AUC = 0.57), with statistically significant differences in predictive accuracy (P < 0.05). Subgroup analyses confirmed consistent associations across age, sex, and BMI categories (all P < sub > interaction</sub > > 0.05).

Conclusion

METS-VF serves as an independent predictor of hypertension in middle-aged and older Chinese adults, outperforming conventional indices. The substantial absolute risk difference—an 18.3% higher 8-year hypertension risk in the highest versus lowest METS-VF quartile—underscores its clinical relevance. The identified risk threshold and superior predictive ability support its potential use in clinical risk stratification and preventive strategies.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12889-025-25580-1.

Introduction

Hypertension, a prevalent chronic condition affecting hundreds of millions worldwide, is recognized as a major risk factor for severe health complications, including cardiovascular diseases, stroke, and renal disorders [1, 2]. Studies have demonstrated that hypertension contributes to the development of cardiac pathologies, heart failure, renal insufficiency, and cognitive impairment [1, 2]. Epidemiological data from Poland reveal a hypertension prevalence of 82.5% among women and 78.3% among men aged 65 and older [3]. Globally, the prevalence of hypertension in populations over 60 years continues to escalate annually, with projections indicating that individuals aged 65 and above will constitute 10.9% of the global population by 2030 [4]. In China, epidemiological surveys report a hypertension prevalence rate of 27.65% in middle-aged and older adults, with 19.91% of undiagnosed cases exhibiting elevated blood pressure during baseline measurements [5].The clinical significance of hypertension is particularly pronounced in elderly populations due to its frequent comorbidity with chronic conditions such as diabetes and hyperlipidemia, which synergistically exacerbate cardiovascular risks. Furthermore, inadequately managed hypertension may precipitate acute hypertensive crises requiring urgent medical intervention to prevent life-threatening outcomes [6]. These findings underscore the critical importance of early detection and systematic management of hypertension to mitigate its substantial public health burden.

METS-VF, by integrating multidimensional information including waist circumference, blood lipids, blood glucose, and age, may more accurately reflect visceral fat-driven metabolic risk, thus outperforming simple waist-to-height ratio or insulin resistance indices in identifying high-risk individuals for hypertension. Visceral Adipose Tissue Metabolic Score (METS-VF), an emerging risk assessment tool, integrates visceral adiposity and its metabolic implications to predict hypertension risk [7].The clinical applicability of this metric relative to established anthropometric measures such as body mass index (BMI) and waist circumference has attracted growing scientific interest.Emerging evidence from a provincial longitudinal investigation conducted in Guizhou, China, revealed a statistically significant elevation in hypertension incidence among subjects with increased METS-VF values. Of particular note, the observed correlation demonstrated greater clinical significance within the Han Chinese demographic, highlighting the potential role of ethnic heterogeneity in moderating the METS-VF-hypertension risk relationship [8]. Recent studies have further validated METS-VF as a predictor of cardiovascular outcomes in East Asian populations [9, 10], strengthening its potential utility in this ethnic group.

Nevertheless, a significant research gap remains unaddressed, particularly concerning the paucity of comprehensive, population-level longitudinal investigations within the Chinese context that rigorously assess METS-VF’s prognostic value in hypertension pathogenesis. To bridge this evidence gap, our study utilizes a nationally representative prospective cohort design—aimed at quantifying the longitudinal associations between METS-VF trajectories and hypertension onset, exploring the underlying biological mechanisms mediating this relationship, and ultimately conducting comparative analyses of predictive efficacy between METS-VF and conventional visceral adiposity indices (specifically the lipid accumulation product [LAP] and visceral adiposity index [VAI])—and based on this design, we hypothesize that METS-VF is positively associated with hypertension onset and that its predictive performance is superior to that of the aforementioned traditional indices.

Methods

Study design and population

The China Health and Retirement Longitudinal Study (CHARLS) represents a methodologically rigorous initiative to investigate aging dynamics through nationally representative longitudinal surveillance of households and individuals aged ≥ 45 years. This project was specifically designed to address critical knowledge gaps in population aging research and catalyze interdisciplinary inquiry into gerontological challenges within the Chinese context. The inaugural 2011–2012 national survey employed a multistage probability proportional to size (PPS) sampling framework, structured across 150 county-level administrative divisions and 450 village-level clusters, ultimately enrolling a cohort of 17,000 participants from 10,000 households [9, 10]. Longitudinally maintained through standardized biennial waves (2013–2018), data collection involved face-to-face interviews utilizing validated computer-assisted personal interviewing (CAPI) protocols administered by certified field staff [11]. Ethical governance aligned with international standards, as the study protocol was conducted under full compliance with the Declaration of Helsinki and granted formal approval by the Institutional Review Board of Peking University (IRB00001052-11015). All participants provided written informed consent prior to enrollment. De-identified CHARLS datasets, including comprehensive metadata and technical documentation, are publicly accessible through institutional repositories with controlled access protocols to ensure data integrity and participant confidentiality (http://charls.pku.edu.cn/en).

For this analysis, participants interviewed during the 2011–2012 baseline survey were included, with follow-up assessments in 2013, 2015, 2018, and 2020. A total of 6,775 individuals met the inclusion criteria and were stratified into four subgroups based on METS-VF quartiles (Q1–Q4). Exclusions were applied for the following reasons: missing age data or age < 45 years (n = 538), follow-up duration < 2 years (n = 1,123), incomplete anthropometric data (waist circumference [WC], height, or BMI; n = 1,706), missing lipid or glucose profiles (triglycerides [TG], fasting blood glucose [FBG], or HDL-C; n = 5,145), and baseline hypertension or missing hypertension-related data (n = 2,421) (Fig. 1).

Fig. 1.

A flow chart of the study

Variables

Calculation of METS-VF, VAI, and LAP

The primary exposure variable, METS-VF, was calculated using sex-specific coefficients (1 for males, 0 for females). The following equations were applied to derive METS-VF, visceral adiposity index (VAI), and lipid accumulation product (LAP) [12–14]:

Diagnosis of hypertension

The primary outcome of this study was incident hypertension, defined by meeting any one of the following criteria: (1) systolic blood pressure (SBP) ≥ 140 mmHg; (2) diastolic blood pressure (DBP) ≥ 90 mmHg; (3) self-reported physician-diagnosed hypertension; or (4) self-reported use of antihypertensive medications [15, 16]. Participants meeting any one of these criteria were classified as hypertensive.

Covariates

Potential confounders adjusted in the analysis included demographic, lifestyle, anthropometric, laboratory, and health status variables:

• Demographics: Age, sex, marital status, educational attainment.

• Lifestyle: Smoking status, alcohol consumption.

• Anthropometrics: Height, BMI, WC, grip strength, systolic blood pressure (SBP), diastolic blood pressure (DBP).

• Laboratory parameters: Glycated hemoglobin (HbA1C), FBG, TG, total cholesterol, HDL-C, LDL-C, serum creatinine, C-reactive protein, uric acid.

• Health conditions: Diabetes and depression.

Diabetes was defined as self-reported physician diagnosis, active use of antidiabetic medication, HbA1C ≥ 6.5%, or FBG ≥ 7.0 mmol/L (126 mg/dL) [17, 18]. Depression was assessed using the 10-item Center for Epidemiologic Studies Depression Scale (CES-D-10), a validated instrument for depressive symptom screening [19]. It was not listed as a comorbidity but was considered in the sensitivity analysis.

Statistical analysis

Study participants underwent quartile-based stratification (Q1-Q4) according to METS-VF distribution. Continuous variables demonstrating parametric distributions were presented as arithmetic means ± standard deviations (SD), with interquartile comparisons performed through one-way ANOVA. Non-parametric continuous measures were characterized by median values with interquartile ranges (IQR) and analyzed via Kruskal-Wallis tests. Categorical variables were operationalized as frequency distributions (n, %) and adjudicated using Pearson’s chi-square tests.Prospective associations between METS-VF quartiles and hypertension incidence were evaluated through Cox proportional hazards regression, incorporating both unadjusted and covariate-adjusted models. In strict adherence to STROBE reporting standards [20], three hierarchically structured multivariable models were implemented: Model 1: Crude association without confounder adjustment, Model 2: Minimally adjusted for age and biological sex, Model 3: Comprehensively adjusted for sociodemographic factors (age, sex, marital status, education), behavioral covariates (smoking, alcohol use), anthropometric measures (BMI, grip strength), metabolic biomarkers (LDL-C, TC, FBG, UA, CRP, Scr), and clinical comorbidities (diabetes, depression).

To assess heterogeneity of effects, prespecified subgroup analyses stratified by age (< 60 vs. ≥60 years), sex, smoking status, alcohol consumption, BMI categories (underweight: <18.5 kg/m²; normal: 18.5–23.9 kg/m²; overweight: 24–27.9 kg/m²; obese: ≥28 kg/m²), diabetes status, and depression were conducted, accompanied by formal interaction testing. Nonlinear dose-response relationships were interrogated through restricted cubic spline (RCS) regression, with hazard ratios (HRs) and 95% confidence intervals (CIs) derived across METS-VF percentiles.Comparative predictive capacity analyses between METS-VF, VAI, and LAP were executed using receiver operating characteristic (ROC) curve methodology, with diagnostic performance quantified via area under the curve (AUC) metrics. All statistical procedures were performed in R (v4.2.2) and EmpowerStats (www.empowerstats.com), applying a two-tailed α-level of 0.05 for significance determination.

Results

Baseline characteristics of participants

The analytical cohort comprised 6,775 participants, with baseline demographic and clinical characteristics systematically categorized by METS-VF quartiles (Table 1). Relative to the lowest quartile (Q1), individuals in the highest METS-VF quartile (Q4) demonstrated significantly elevated values across multiple cardiometabolic parameters, including age, BMI, waist circumference, systolic/diastolic blood pressure, glycemic markers (HbA1c, fasting plasma glucose), lipid profiles (total cholesterol, triglycerides, LDL-C), renal function indices (serum creatinine), and inflammatory biomarkers (uric acid, C-reactive protein) (all P < 0.05).

Table 1.

Baseline characteristics of individuals classified by quartiles of the METS-VF

	Q1	Q2	Q3	Q4	P-value
Age, years	56.24 ± 8.67	57.38 ± 8.48	58.21 ± 8.86	62.73 ± 9.54	< 0.001
Gender, %					< 0.001
Male	803 (47.40%)	774 (45.72%)	756 (44.63%)	886 (52.30%)
Female	891 (52.60%)	919 (54.28%)	938 (55.37%)	808 (47.70%)
Educational level, %					0.008
No completion of primary school	781 (46.10%)	786 (46.51%)	791 (46.78%)	879 (51.89%)
Sishu/home school/elementary school	352 (20.78%)	387 (22.90%)	369 (21.82%)	361 (21.31%)
Middle school	370 (21.84%)	343 (20.30%)	358 (21.17%)	307 (18.12%)
High school and above	191 (11.28%)	174 (10.30%)	173 (10.23%)	147 (8.68%)
Marital status, %					< 0.001
Married	1533 (90.50%)	1504 (88.84%)	1518 (89.61%)	1441 (85.06%)
Never married	161 (9.50%)	189 (11.16%)	176 (10.39%)	253 (14.94%)
Smoking status, %					< 0.001
No	1077 (63.77%)	1137 (67.28%)	1202 (71.25%)	1173 (69.57%)
Yes	612 (36.23%)	553 (32.72%)	485 (28.75%)	513 (30.43%)
Drinking status, %					0.849
No	1107 (65.39%)	1098 (64.93%)	1122 (66.27%)	1102 (65.09%)
Yes	586 (34.61%)	593 (35.07%)	571 (33.73%)	591 (34.91%)
Height, m	1.59 ± 0.08	1.58 ± 0.08	1.57 ± 0.08	1.57 ± 0.11	< 0.001
Waist circumference, cm	69.92 ± 12.82	80.32 ± 4.21	86.55 ± 5.18	94.49 ± 7.00	< 0.001
Left hand grip, kg	30.39 ± 9.87	30.51 ± 10.17	30.30 ± 10.37	30.10 ± 11.26	0.561
Right hand grip, kg	31.63 ± 10.40	32.00 ± 10.53	31.71 ± 10.69	31.59 ± 11.68	0.308
Systolic, mmHg	119.47 ± 17.34	121.27 ± 16.42	125.15 ± 17.98	131.33 ± 19.58	< 0.001
Diastolic, mmHg	70.77 ± 10.77	71.86 ± 10.48	73.79 ± 11.08	75.94 ± 11.28	< 0.001
BMI, kg/m2	20.28 ± 2.78	21.94 ± 2.69	23.78 ± 2.86	28.20 ± 62.22	< 0.001
HbA1c, %	5.13 ± 0.66	5.19 ± 0.77	5.25 ± 0.79	5.36 ± 0.82	< 0.001
FPG, mg/dl	103.39 ± 30.26	105.50 ± 29.82	109.50 ± 33.95	114.28 ± 36.31	< 0.001
TC, mg/dl	186.84 ± 37.19	189.37 ± 36.70	195.23 ± 38.89	197.88 ± 40.41	< 0.001
TG, mg/dl	103.15 ± 66.74	115.81 ± 91.09	130.94 ± 100.81	155.39 ± 120.99	< 0.001
LDL, mg/dl	111.10 ± 32.39	113.61 ± 32.82	118.49 ± 35.23	119.32 ± 36.35	< 0.001
HDL, mg/dl	56.83 ± 15.78	54.24 ± 15.65	51.03 ± 14.34	46.87 ± 14.28	< 0.001
Scr, mg/dl	0.76 ± 0.17	0.75 ± 0.17	0.77 ± 0.19	0.80 ± 0.18	< 0.001
UA, mg/dl	4.15 ± 1.15	4.22 ± 1.15	4.39 ± 1.20	4.71 ± 1.27	< 0.001
CRP, mg/dl	2.43 ± 8.25	2.28 ± 7.11	2.63 ± 7.37	2.59 ± 5.73	< 0.001
LAP	46.45 ± 110.04	126.19 ± 108.28	191.36 ± 170.76	289.11 ± 242.18	< 0.001
VAI	3.10 ± 3.69	4.48 ± 9.41	5.37 ± 7.32	7.12 ± 10.78	< 0.001
Hypertension, %					< 0.001
No	1327 (78.335%)	1235 (72.947%)	1119 (66.057%)	928 (54.782%)
Yes	367 (21.665%)	458 (27.053%)	575 (33.943%)	766 (45.218%)
Diabetes, %					< 0.001
No	1653 (97.58%)	1645 (97.16%)	1625 (95.93%)	1585 (93.57%)
Yes	41 (2.42%)	48 (2.84%)	69 (4.07%)	109 (6.43%)
Depression, %					0.016
No	1034 (61.04%)	1097 (64.80%)	1109 (65.47%)	1112 (65.64%)
Yes	660 (38.96%)	596 (35.20%)	585 (34.53%)	582 (34.36%)

Sociodemographic stratification revealed a greater representation of male participants, individuals with limited educational attainment, unmarried status, and non-smokers in the Q4 subgroup. Conversely, this cohort exhibited reduced anthropometric measures (height) and attenuated biochemical profiles for HDL-C, CRP, uric acid, and HbA1c relative to Q1 counterparts (all P < 0.05). Clinically significant disparities emerged in comorbidity prevalence, with the Q4 group showing an elevated burden of diabetes mellitus alongside a reduced prevalence of depressive disorders (both P < 0.05). Most critically, a dose-dependent escalation in hypertension incidence was observed across ascending METS-VF quartiles (P < 0.05).

Association between METS-VF and hypertension incidence

The longitudinal relationship between METS-VF and hypertension risk is systematically quantified in Table 2. A robust positive association emerged between METS-VF increments and hypertension incidence, demonstrating stability across progressively adjusted multivariable models. In the unadjusted Model I, each unit elevation in METS-VF corresponded to a 4.5% increased hypertension risk (HR = 1.045, 95% CI: 1.020–1.071). This association persisted in Model II following age-sex adjustment, albeit attenuated to a 3.8% risk increment per METS-VF unit (HR = 1.038, 95% CI: 1.014–1.063). Full covariate adjustment in Model III maintained statistical significance, revealing a 2.4% risk elevation per unit increase (HR = 1.024, 95% CI: 1.002–1.047). Through “per-SD effect size conversion,” the per-SD HR was calculated via the HR multiplicative effect as approximately 1.056 (95%CI: 1.005–1.063), meaning a 5.6% increase in hypertension risk for every 2.3-unit rise in METS-VF. For unmeasured confounding, E-values indicate an approximate 1.023 E-value based on 1-unit HR and 1.031 based on per-SD HR, suggesting the results are relatively sensitive to unmeasured confounding. However, considering adjusted covariates (e.g., age, gender, metabolic indicators), the actual impact of confounding is limited. In verifying the proportional hazards assumption, the Schoenfeld residual test yielded a global χ² of 3.28 and a P-value of 0.070, with no violation of the assumption, and the temporal stability of the HR in the original model is confirmed. Quartile-based analysis revealed pronounced risk stratification, with the highest METS-VF quartile (Q4) demonstrating a 2.115-fold elevated hypertension risk relative to Q1 (HR = 2.115, 95% CI: 1.850–2.418) in the fully adjusted model. This model incorporated demographic covariates (age, sex, marital status, education), behavioral factors (smoking, alcohol use), metabolic parameters (BMI, LDL-C, TC, FBG, UA, HbA1c), inflammatory markers (CRP), renal function (Scr), physical capacity (bilateral grip strength), and comorbidities (diabetes, depression). A statistically significant dose-response gradient was evident across quartiles (P < sub > trend</sub > < 0.001), confirming METS-VF’s graded relationship with hypertension susceptibility. Nonlinear restricted cubic spline (RCS) analysis delineated a monotonic yet curvilinear association between METS-VF and hypertension risk (P < sub > overall</sub > < 0.001; P < sub > nonlinear</sub > < 0.001; Fig. 2). As shown in Fig. 2, the quartile cutoffs of METS-VF were: Q1 (≤ 3.24), Q2 (3.25–4.89), Q3 (4.90–6.57), and Q4 (> 6.57), with balanced sample sizes across groups. Regarding the Restricted Cubic Spline (RCS) analysis, the inflection point at METS-VF = 5.681 (which may appear unusually right-shifted) was clarified as follows: the distribution of METS-VF was right-skewed, with 75.9% of participants (n = 5128) having METS-VF ≤ 5.681 and 24.1% (n = 1647) having METS-VF > 5.681. After screening for outliers via the Grubbs test (only 0.18% of values were extreme) and re-fitting the model, the inflection point remained stable. Moreover, gender-stratified RCS analyses showed consistent inflection points (males: 5.72; females: 5.65), confirming the reliability of the result. Threshold analysis identified a critical inflection at METS-VF = 5.681 (Table 3), beyond which hypertension risk escalated markedly (HR = 2.344, 95% CI: 2.040–2.694), indicating a nonlinear amplification effect at elevated METS-VF values.

Table 2.

Association between METS-VF and hypertension

HR (95%CI) P-value	Model 1	Model 2	Model 3
Continuous METS-VF	1.045 (1.020, 1.071) < 0.001	1.044 (1.019, 1.070) < 0.001	1.031 (1.009, 1.054) 0.007
METS-VF quartiles
Q1	Reference	Reference	Reference
Q2	1.261(1.100, 1.447) < 0.001	1.254(1.093, 1.439) 0.001	1.213 (1.056, 1.394) 0.006
Q3	1.665(1.460, 1.898) < 0.001	1.662(1.457, 1.894) < 0.001	1.498(1.311, 1.711) < 0.001
Q4	2.481 (2.190, 2.810) < 0.001	2.439(2.153, 2.763) < 0.001	2.044 (1.797, 2.418) < 0.001
P for trend	< 0.001	< 0.001	< 0.001

Model 1: no covariates were adjusted

Model 2: marital status, education level, smoking and drinking were adjusted

Model 3: LDL-C, Scr, CRP, TC, UA, left hand grip, right hand grip, diabetes, dyslipidemia, heart disease, stroke and depression were adjusted

Fig. 2.

The fully adjusted RCS regression model showed a nonlinear positive correlation between eGDR and CVD risk (P-overall < 0.001, P-linear < 0.001)

Table 3.

Threshold effect analysis of METS-VF on hypertension using a two-piecewise linear regression model

Hypertension	Adjust HR (95% CI)	P
METS-VF
Fitting by standard linear model	1.031(1.009,1.054)	0.007
Fitting by two-piecewise linear model
Inflection point	5.129
< 5.129	0.997(0.966,1.000)	0.779
> 5.129	1.198(1.159,1.238)	< 0.001
Log-likelihood ratio	< 0.001

Marital status, education level, smoking, drinking, LDL-C, Scr, CRP, TC, UA, left hand grip, right hand grip, diabetes, dyslipidemia, heart disease, stroke and depression were adjusted

Subgroup analyses

Subgroup and interaction analyses were conducted to evaluate the relationship between METS-VF and hypertension incidence across strata of age, sex, smoking status, alcohol consumption, BMI, diabetes, and depression (Table 4). No statistically significant interaction effects were observed for any of these variables (P for interaction > 0.05), indicating consistent associations between METS-VF and hypertension risk across all subgroups.

Table 4.

Subgroup analysis of the association between METS-VF and hypertension

Characteristics	Number of participants	HR (95%CI)	p	P for interaction
Age (years)				0.683
< 60	3964	1.021 (0.994, 1.048)	0.134
≥ 60	2811	1.031 (0.991, 1.072)	0.129
Gender				0.721
Female	3556	1.028 (0.997, 1.059)	0.075
male	3219	1.019 (0.987, 1.053)	0.239
Smoking				0.493
No	4612	1.030 (1.001, 1.059)	0.041
Yes	2163	1.013 (0.978, 1.050)	0.459
Drinking				0.477
No	4434	1.018 (0.990, 1.045)	0.208
Yes	2341	1.034 (0.997, 1.073)	0.074
BMI				0.612
< 18.5	593	1.005 (0.929, 1.088)	0.901
18.5–23.9	3812	1.006 (0.981, 1.031)	0.642
24–27.9.9	1822	1.003 (0.970, 1.038)	0.844
≥ 28	548	1.045 (0.988, 1.105)	0.121
Diabetes				0.528
No	6508	1.027(1.003, 1.050)	0.024
Yes	267	0.998(0.923, 1.080)	0.967
Depression				0.833
No	4352	1.020 (0.996, 1.045)	0.102
Yes	2423	1.015(0.974, 1.058)	0.486

The model adjusted for age, gender, marital status, education level, smoking, drinking, BMI, LDL-C, Scr, CRP, FBG, TC, UA, HbA1c, left hand grip, right hand grip, diabetes and depression (Excluding variables used for subcomponent stratification)

Predictive performance of METS-VF, LAP, and VAI

Receiver operating characteristic (ROC) curve analysis was employed to compare the diagnostic utility of METS-VF, lipid accumulation product (LAP), and visceral adiposity index (VAI) for hypertension prediction. The area under the curve (AUC) values ranked as follows: METS-VF (AUC = 0.616) > LAP (AUC = 0.603) > VAI (AUC = 0.569) (Table 5; Fig. 3). We compared the predictive value of LAP, VAI, and METS-VF for hypertension in middle-aged and elderly Chinese people, with the results shown in Table 6. While all three indices demonstrated modest predictive capacity for hypertension, METS-VF exhibited statistically superior performance compared to LAP and VAI (P < 0.05).

Table 5.

Sensitivity, specificity, Youden index, cutoff points, and AUC (95% CI) for each index in predicting hypertension risk among adults in China

	AUC	95%CI	Cut off value	Specificity	Sensitivity	P Value
METS-VF	0.616	0.602–0.631	6.150	55.0	62.8	< 0.001
LAP	0.603	0.589–0.618	122.641	57.2	59.0	< 0.001
VAI	0.569	0.554–0.584	3.730	62.2	49.6	< 0.001

Fig. 3.

The ROC curve was used to investigate the potential diagnostic value of METS-VF, LAP and VAI for hypertension

Table 6.

Predictive values of LAP, VAI, and METS-VF for hypertension in middle-aged and elderly Chinese individuals

Model	NRI (95% CI)	P value	IDI (95% CI)	P value
Basic model	Ref	Ref	Ref	Ref
+ LAP	0.078 (0.028–0.116)	< 0.001	0.001(−0.001–0.003)	0.445
+ VAI	0.014 (−0.019–0.053)	0.339	0.002 (0.000–0.005)	0.020
+ METS-VF	0.084(0.011–0.133)	0.020	0.008 (0.003–0.013)	< 0.001

The basic model included marital status, education level, smoking, drinking, LDL-C, Scr, CRP, TC, UA, left hand grip, right hand grip, diabetes, dyslipidemia, heart disease, stroke and depression

Discussion

Recent years have witnessed a growing focus on the association between visceral adiposity and hypertension amid escalating obesity rates. As a metabolically active adipose depot enveloping internal organs, excessive visceral fat accumulation elevates blood pressure through multiple pathways, including insulin resistance, systemic inflammation, endocrine dysregulation (e.g., adipokine-mediated vascular dysfunction), autonomic nervous system activation, and enhanced renal sodium reabsorption [21]. Concurrently, visceral adiposity drives chronic low-grade inflammation and insulin resistance, amplifying risks for metabolic disorders such as diabetes and cardiovascular diseases. Epidemiological evidence consistently links visceral fat deposition to hypertension. For instance, in the U.S., METS-VF has emerged as a robust predictor of hypertension, particularly among males and middle-aged/older adults, with higher scores correlating with increased risks of cardiovascular events and all-cause mortality [22, 23]. South Asian studies validated METS-VF as a reliable tool for assessing visceral adiposity in obese populations [24]. Chinese research further corroborated these findings, showing that elevated METS-VF scores correlate with higher systolic and diastolic blood pressure in obese individuals, while visceral fat accumulation synergistically exacerbates insulin resistance, diabetes, and metabolic syndrome, collectively amplifying hypertension risk [23].This national prospective cohort study demonstrates that METS-VF is significantly associated with the risk of incident hypertension in middle-aged and elderly Chinese individuals, and serves as an important and independent predictor for the development of hypertension. Furthermore, the METS-VF index may hold clinical significance for population-based prevention of hypertension—which is consistent with the predictive value of METS-IR for other metabolic and cardiovascular outcomes [25–27].

Conventional anthropometric indices (e.g., BMI, waist circumference, skinfold thickness) face limitations in distinguishing fat-muscle distribution or requiring specialized techniques. Although the visceral adiposity index (VAI) exhibits a strong correlation with visceral fat volume (r = 0.74), METS-VF demonstrates superior clinical utility in hypertension screening. Its unit increase consistently predicts hypertension risk across diverse populations, enabling early identification of at-risk individuals [7].

Emerging evidence highlights the pivotal role of visceral adiposity in hypertension pathogenesis, driven by its heightened metabolic activity and propensity for ectopic deposition, which underlie cardiometabolic dysregulation [28, 29]. Hypertrophied adipocytes within visceral fat depots release excessive free fatty acids (FFAs) into the liver, inducing insulin resistance and atherogenic dyslipidemia (elevated triglycerides, reduced HDL-C), while impaired preadipocyte differentiation exacerbates lipid overflow into non-adipose organs, perpetuating a vicious cycle of metabolic syndrome, diabetes, and cardiovascular dysfunction [29, 30]. Concurrently, visceral adipose tissue secretes pro-inflammatory cytokines (e.g., TNF-α, IL-6), triggering systemic inflammation that disrupts metabolic homeostasis and amplifies cardiovascular risk [31]. In middle-aged and older adults, age-related visceral fat accumulation synergizes with metabolic decline, characterized by adipocyte hypertrophy, hyperplasia, and diminished preadipocyte differentiation capacity, collectively fostering ectopic lipid deposition and adipose tissue remodeling. This pathogenic milieu promotes endothelial dysfunction via NF-κB-mediated oxidative stress and adipokine imbalance (e.g., hyperleptinemia, hypoadiponectinemia) [32]– [33]. Concurrently, visceral fat-derived angiotensinogen activates the renin-angiotensin-aldosterone system, enhancing vasoconstriction and sodium retention [34].

These mechanisms collectively underscore the centrality of visceral adiposity in hypertension development, particularly in aging populations with progressive metabolic dysfunction.

Compared with previous studies, this study has distinct strengths, including its prospective design, well-characterized cohort, and broader population coverage. While it confirms the critical value of METS-VF as a novel biomarker for hypertension prediction—providing a quantifiable tool to assess adipose distribution and metabolic dysfunction for clinical guidance—our findings also have certain limitations. First, our study population was mainly Han Chinese, which may limit the generalizability of our results to other ethnic groups. Second, although we adjusted for many potential confounders, residual confounding from unmeasured factors (such as dietary quality and physical activity) cannot be excluded. Third, changes in antihypertensive treatment guidelines between 2011 and 2020 may have introduced unmeasured confounding. Fourth, since METS-VF includes metabolic variables that are also components of the hypertension definition, there is a risk of overfitting. Future studies should incorporate multi-ethnic cohorts, longitudinal validation, and mechanistic exploration to improve its applicability and establish standardized thresholds for risk stratification.

Conclusion

The visceral adipose tissue metabolic score (METS-VF) is increasingly recognized as a critical tool for hypertension risk stratification. The robust association between visceral adiposity-derived metabolic dysregulation and hypertension unveils the intricate pathophysiological interplay linking obesity to cardiovascular morbidity. These findings not only deepen our comprehension of visceral fat’s role in cardiometabolic disorders but also provide novel perspectives for targeted prevention and personalized management of hypertension, particularly in aging populations.

Authors’ contributions

Guiyue Wang and Linjie Qiu drafted the initial version of the manuscript. Zhili Wang assisted in revising the manuscript. Jianxun Liu acquired funding and oversaw the overall research implementation. All authors have reviewed and approved the final manuscript.

Funding

This research was supported by the State Key Program of National Natural Science of China (82030124).

Data availability

The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found below: This analysis uses data or information from the Harmonized CHARLS dataset and Codebook, Version D as of June 2021 developed by the Gateway to Global Aging Data. The development of the Harmonized CHARLS was funded by the National Institute on Aging (R01 AG030153, RC2 AG036619, R03 AG043052). For more information, please refer to https://g2aging.org/.

Declarations

Ethics approval and consent to participate

The studies involving humans were approved by public databases-China Health and Retirement Longitudinal Study (CHARLS). Besides, the protocols of CHARLS were approved by the Biomedical Ethics Review Committee of Peking University (IRB00001052-11015). The studies were conducted in accordance with the local legislation and institutional requirements. The participants provided their written informed consent to participate in this study.

Competing interests

The authors declare no competing interests.