Visceral fat metabolism score and cardiovascular-kidney-metabolic syndrome stages in all-cause and cause-specific mortality: insights from UK Biobank

Abstract

Background

The roles of Cardiovascular-Kidney-Metabolic Syndrome (CKM) staging and the emerging obesity marker Visceral Fat Metabolism Score (METS-VF) in cause-specific mortality, as well as their potential interaction, remain to be further investigated.

Methods

396,383 UK Biobank participants were analyzed. Primary outcomes included all-cause and cause-specific deaths, including cardiovascular, respiratory, digestive, neurodegenerative diseases, cancer, and other causes. Independent and joint effects of METS-VF and CKM were evaluated using a COX proportional hazards regression model, with multiple sensitivity analyses conducted to verify robustness. Dose–response relationships were visualized using restricted cubic spline (RCS) curves. Receiver operating characteristic (ROC) analyses were conducted to evaluate the discriminative performance of METS-VF for predicting mortality compared with other obesity-related indices. Multiple sensitivity analyses were further performed to assess the robustness of the findings.

Results

Over an average follow-up of 12.51 years, 34,471 deaths were recorded. After multivariate adjustment, both METS-VF and CKM stage were associated with all-cause mortality. The hazard ratio for the highest METS-VF quartile was 1.70 (95% CI: 1.61–1.80), and for the highest CKM stage was 2.23 (95% CI: 2.12–2.35). Each 0.1 increase in METS-VF raised the risk of death by 8.2% (CVD), 3.5% (respiratory), 8.8% (digestive), 2.7% (cancer), and 6.9% (other causes). RCS analysis showed a non-linear relationship between METS-VF and mortality. ROC analyses further demonstrated that METS-VF exhibited significantly higher predictive performance for both all-cause and cause-specific mortality compared to other obesity-related indices. Joint analysis revealed that participants with stage IV CKM and high METS-VF had the highest risks of all-cause and cause-specific mortality, with synergistic effects noted for CVD, and digestive mortality.

Conclusion

High METS-VF and late-stage CKM were associated with increased risks of mortality from multiple causes. For individuals with late-stage CKM, enhanced assessment of METS-VF is recommended to facilitate timely interventions and reduce mortality risk.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12986-026-01083-7.

Background

Cardiovascular-Kidney-Metabolic (CKM) Syndrome, proposed by the American Heart Association (AHA), aims to highlight the complex pathophysiological interactions among obesity, diabetes, cardiovascular disease (CVD), and chronic kidney disease (CKD), which lead to multi-organ dysfunction and adversely affect patients’ systemic health [1]. The AHA has established a staging criterion ranging from stage 0 to IV based on disease progression, allowing for a dynamic assessment of the development of metabolic disorders and organ damage among patients with CKD [2]. CKM and its components are significant determinants of mortality [2]. According to the 2017 Global Burden of Disease (GBD) report, CKD is responsible for 1.2 million deaths worldwide [3]. The 2019 GBD report found that obesity-related deaths were the highest of all metabolic factors, accounting for approximately 5 million deaths [4]. As the primary causes of death in patients with diabetes and CKD are heart failure and atherosclerotic cardiovascular disease [5, 6], previous studies regarding CKM have predominantly focused on all-cause mortality and CVD mortality. However, non-cardiovascular-specific mortality also warrants further investigation as it constitutes a significant proportion of deaths in high-risk populations with risk factors including obesity, diabetes, hypertension, and CKD [7–9].

CKM syndrome is a continuous and dynamic process that progresses from obesity-related metabolic disturbances to multi-organ damage and ultimately death [2]. In this trajectory, visceral fat accumulation serves not only as a key source of insulin resistance and chronic low-grade inflammation [10], but also exerts direct detrimental effects on multiple organs through lipotoxicity, oxidative stress, and activation of the renin–angiotensin system [11, 12]. In the heart, it promotes myocardial remodeling and dysfunction [13]; in the kidneys, it drives glomerular hyperfiltration, podocyte injury, and interstitial fibrosis [14]; in the liver, it exacerbates steatosis and inflammation, accelerating the progression of metabolic dysfunction associated steatotic liver disease (MASLD) [15]. Furthermore, adipokines and pro-inflammatory cytokines secreted by visceral fat foster a systemic inflammatory milieu that disrupts immune homeostasis and may impair central nervous system function [16]. These multi-organ pathological alterations are interconnected and mutually reinforcing, collectively forming the core biological foundation of CKM progression. However, previous CKM research often pools individuals across stages 0–3 [17–19], overlooking the stage-specific differential impact of visceral fat burden on disease progression and mortality risk. Therefore, integrating CKM staging with quantitative assessment of visceral fat may enable more precise identification of individuals at high risk of death.

The Visceral Fat Metabolism Score (METS-VF) is a novel measurement index to estimate visceral adiposity based on BMI, waist-to-height ratio (WHtR), Metabolism Score for Insulin Resistance (METS-IR), age, and sex [20]. In previous validation studies using dual-energy X-ray absorptiometry (DXA) and magnetic resonance imaging (MRI) as reference standards, METS-VF outperformed other commonly used surrogate indices for visceral fat [20]. Notably, it can assess visceral fat content independently of BMI and waist circumference [21] underscoring its potential utility in clinical and epidemiologiology. Accumulating evidence has linked METS-VF to a range of adverse health outcomes, including type 2 diabetes, non-alcoholic fatty liver disease, impaired lung function, cognitive dysfunction, and chronic kidney disease (CKD) [22–26]. However, research regarding the correlation between the METS-VF and disease-related mortality is limited, especially in the context of CKM. The potential value of applying the METS-VF in the joint assessment of CKM staging has not been fully explored.

Based on the UK Biobank dataset, we aimed to explore the independent and joint effect of different stages of CKM and the METS-VF with all-cause and cause-specific mortality, providing valuable insights into the early identification of populations at high risk of death by analyzing how these factors collectively affect mortality risk.

Methods

Study population

The UK Biobank is a large-scale, prospective, cohort study that recruited over 500,000 individuals aged 37–73 years from 2006 to 2010. Participants underwent comprehensive assessments at one of 22 designated assessment centers nationwide, where detailed information regarding sociodemographic characteristics, lifestyle, health data, and biological samples were collected. All procedures included in this study were approved by the Northwest Multicenter Research Ethics Committee (approval number: 21/NW/0157). All participants provided written informed consent for data collection, analysis, and record linkage.

We initially included 502,166 participants. And then several participants were excluded, including 9,330 participants with missing data related to CKM staging; 70,622 missing the indicators required for calculating the METS-VF; and 25,831 with incomplete covariate data. In total, the data of 396,383 participants were included in the current study (Fig. S1).

Definitions of CKM syndrome stages 0 to IV

CKM is diagnosed in five stages, with detailed definitions provided by the AHA [2]. According to European population standards, stage 0 is characterized by a normal BMI, waist circumference, blood glucose levels, blood pressure, lipid levels, and no evidence of CKD or clinical CVD. Stage I is characterized by one or more of the following criteria: BMI ≥ 25 kg/m², waist circumference ≥ 88 cm for women or ≥ 102 cm for men, or dysglycaemia, defined as fasting blood glucose ≥ 100–124 mg/dL or HbA1c of 5.7%–6.4% in the absence of other metabolic risk factors and CKD. Due to the absence of subclinical CVD data in the UK Biobank dataset, it was not feasible to differentiate between stage II and III CKM. Therefore, in the current study, these stages were combined into a single stage, defined as metabolic risk factors (hypertriglyceridemia (≥ 135 mg/dL), hypertension, metabolic syndrome, diabetes) or CKD at baseline without clinical CVD. Stage IV CKM was defined as the presence of metabolic risk factors or CKD with CVD, including coronary heart disease, heart failure, stroke, peripheral artery disease, and atrial fibrillation. Disease diagnoses were derived from the UK Biobank’s inpatient diagnoses utilizing International Classification of Diseases 9th Revision (ICD-9), ICD-10, Operative procedures 4 (OPCS-4) codes and self-reported medical histories, types of surgeries, medication histories, and blood biochemical tests. When multiple diseases or records of a specific disease were reported, the earliest occurrence was considered the reference point. Detailed disease diagnosis codes are presented in Table S1.

Assessment of the METS-VF

The METS-VF calculation is based on sex, age, BMI, WHtR, and the METS-IR [20, 27]:

Sex was defined as male = 1 and female = 0.

Outcome assessment

In our study, the primary endpoints were all-cause and cause-specific mortality, which included deaths attributed to CVD, respiratory diseases, digestive diseases, neurodegenerative diseases, cancer, and other causes. The dates and causes of death were sourced from death certificates obtained from the National Health Service Information Centre (England and Wales) and National Health Service Central Register (Scotland), with classifications based on the ICD-10 codes. The follow-up period began at baseline and continued until the first occurrence of the study outcome or the conclusion of the follow-up period, whichever occurred first. The specific cut-off dates were 28 February 2018 for Wales; 30 September 2021 for England; and 31 July 2021 for Scotland.

Covariates assessment

The study examined several potential confounding factors including sociodemographic variables, health behaviors, and health status. Sociodemographic variables included age, sex, ethnicity, educational level, employment status, and the Townsend Deprivation Index (TDI), which serves as an indicator of socioeconomic status [28]. The smoking status was categorized into three groups: never smoked, former smoker, and current smoker. Alcohol intake was classified into three categories: never drink, ≤ 16 g per day, and > 16 g per day. Physical activity was deemed sufficient if any of the following criteria were met: ≥ 150 min of moderate activity per week, ≥ 75 min of vigorous activity per week, ≥ 150 min of a combination of moderate and vigorous activities per week, ≥ 5 sessions of moderate-intensity activity per week, or ≥ 1 session of vigorous activity per week. Sleep patterns were defined as normal sleep lasting 7–9 h/day. Diet was assessed according to the latest dietary recommendations for cardiovascular health [29], with adherence to at least five recommendations considered indicative of a healthy diet; specific criteria are detailed in Table S2. BMI was categorized as normal (< 25 kg/m²), overweight (25–29.9 kg/m²), or obese (≥ 30 kg/m²) [30]. Given that low vitamin D levels and C-reactive protein (CRP) are both strongly associated with visceral fat accumulation, insulin resistance, and chronic inflammation [31, 32], and have been independently linked to mortality risk [33, 34], we further adjusted for these two potential confounders. Additionally, to minimize bias in the exposure and outcome association due to pre-existing conditions at baseline, we also adjusted for baseline diagnoses of respiratory, gastrointestinal, neurodegenerative diseases, and cancer.

Statistical analysis

Stacked bar charts and Kaplan-Meier (KM) curves were used to illustrate the difference and trends in all-cause and cause-specific mortality across different stages of CKM and the METS-VF quartiles. Cox proportional hazards regression models were used to evaluate the associations between CKM, the METS-VF, and both all-cause and cause-specific mortality using hazard ratios (HRs) and their 95% confidence intervals (CIs). Trend tests were also conducted. The proportional hazards assumption was tested using Schoenfeld residuals, whereas multicollinearity within the models was examined using variance inflation factors (VIFs). To mitigate the impact of potential confounding factors, three progressively adjusted models were constructed: model 1 was adjusted for sex, age, ethnicity, occupation, education, the TDI, and BMI; model 2 was further adjusted for smoking, alcohol consumption, sedentary behavior, physical activity, diet, and sleep; and model 3 was additionally adjusted for vitamin D levels, C-reactive protein levels, and baseline diseases corresponding to specific causes of death. Restricted cubic spline (RCS) regression was used to explore the dose-response relationship between the METS-VF and the risk of mortality, with four knots selected according to the Akaike Information Criterion. To evaluate the discriminative performance of METS-VF, we constructed receiver operating characteristic (ROC) analysis to compare its ability to predict mortality risk against of other obesity-related indices [35], including waist circumference (WC), waist-to-height ratio (WHtR), lipid accumulation product (LAP), visceral adiposity index (VAI), body roundness index (BRI), waist-weight index (WWI), and trunk fat percentage measured by bioelectrical impedance analysis (TFP_BIA). The statistical significance of differences in area under the curve (AUC) was assessed using the DeLong test.

To investigate the joint effects of the different stages of CKM and the METS-VF on all-cause and specific-cause mortality, the METS-VF were re-classified into high, medium, and low categories and combined with the stages of CKM to create a new categorical variable. The product term of both variables was included in the model and the models were compared with and without the cross-product term using the likelihood ratio test to assess the multiplicative interaction between them. Additionally, the relative excess risk due to interaction and the attributable proportion were used as measures of additive interaction. To further explore the potential modifying effects, subgroup analyses were conducted to assess whether the combined effects of CKM and the METS-VF on various mortality risks varied across health behavioral factors, including smoking, alcohol consumption, sedentary behavior, physical activity, diet, and sleep.

Multiple sensitivity analyses were performed to confirm the robustness of the results. First, to validate the independent risk of the METS-VF for all-cause and cause-specific mortality, participants with BMI ≥ 25 kg/m² were excluded to determine whether the METS-VF had an association with mortality among individuals with a normal BMI. Second, as baseline diseases may influence the corresponding mortality rates, participants with pre-existing conditions at baseline were excluded from the cause-specific mortality analysis. Third, to account for loss to follow-up and competitive relationships formed by different types of death, a competing risk model was used to capture the true risk of mortality more accurately. Fourth, deleting data with missing covariates during the initial stage of the experimental design could introduce potential bias. Therefore, these data were restored using multiple imputations using chained equations, conducting the imputation process five times to minimize the bias associated with missing data. Finally, participants who died within two years of enrolment were excluded to mitigate potential reverse causality effects. All analyses were conducted using R software, version 4.4.3, with P < 0.05 deemed statistically significant.

Results

Baseline characteristics

Participants with a high METS-VF were more likely to be aged > 60 years and male and were more likely to have achieved a lower educational status and diminished socioeconomic status (Table 1). Many participants in the highest METS-VF quartile were retired or unemployed. The BMI increased as the METS-VF increased; however, some participants with normal BMI had a high METS-VF. Participants with a high METS-VF were more likely to smoke and consume alcohol, had longer average daily sedentary durations and insufficient physical activity, and were less inclined to maintain healthy dietary and sleep habits. These individuals had lower vitamin D levels, higher C-reactive protein levels, and were more likely to have stages II-IV CKM, indicating a greater number of metabolic risk factors. A high proportion of participants with a high METS-VF had respiratory diseases, digestive system diseases, neurodegenerative diseases, and cancer at baseline. Most participants had stage II-III CKM (Table S3). As the CKM stage worsened, the trends in baseline characteristics of the participants remained similar to those shown in Table 1. (all P < 0.001).

Table 1.

Baseline characteristics of the 396,383 participants based on METS-VF quartiles

Characteristics	All participants	Individuals classified by quartiles of the METS-VF				P value
		Q1 (< 6.194)	Q2 (6.194–6.748)	Q3 (6.749–7.159)	Q4 (7.160–8.423)
Age, N (%)						< 0.001
38–59	224,670 (56.68%)	70,672 (71.32%)	60,141 (60.69%)	52,583 (53.06%)	41,274 (41.65%)
≥ 60	171,713 (43.32%)	28,424 (28.68%)	38,955 (39.31%)	46,512 (46.94%)	57,822 (58.35%)
Sex, N (%)						< 0.001
male	186,585 (47.07%)	14,195 (14.32%)	40,029 (40.39%)	58,248 (58.78%)	74,113 (74.79%)
female	209,798 (52.93%)	84,901 (85.68%)	59,067 (59.61%)	40,847 (41.22%)	24,983 (25.21%)
Ethnicity, N (%)						< 0.001
white	375,588 (94.74%)	94,926 (95.79%)	93,642 (94.50%)	93,113 (93.96%)	93,847 (94.70%)
non-white	20,855 (5.26%)	4,170 (4.21%)	5,454 (5.50%)	5,982 (6.04%)	5,249 (5.30%)
Education level, N (%)						< 0.001
college or university degree	129,010 (32.55%)	40,508 (40.88%)	34,498 (34.81%)	30,142 (30.42%)	23,862 (24.08%)
middle or high school	149,991 (37.84%)	40,113 (40.48%)	38,824 (39.18%)	37,056 (37.39%)	33,998 (34.31%)
professional qualifications	46,897 (11.83%)	8,384 (8.46%)	10,751 (10.85%)	12,855 (12.97%)	14,907 (15.04%)
others	70,485 (11.83%)	10,091 (10.18%)	15,023 (15.16%)	19,042 (19.22%)	26,329 (26.57%)
Career, N (%)						< 0.001
employment	241,056 (60.81%)	69,094 (69.72%)	63,128 (63.70%)	58,618 (59.15%)	50,216 (50.67%)
unemployment	27,060 (6.83%)	7,045 (7.11%)	5,776 (5.83%)	5,995 (6.05%)	8,244 (8.32%)
retirement	124,722 (31.47%)	22,076 (22.28%)	29,345 (29.61%)	33,576 (33.88%)	39,725 (40.09%)
others	3545 (0.89%)	881 (0.89%)	847 (0.85%)	906 (0.91%)	911 (0.92%)
TDI, Mean ± SD	−1.35 ± 3.06	−1.58 ± 2.93	−1.48 ± 2.99	−1.36 ± 3.06	−0.96 ± 3.22	< 0.001
BMI (kg/m2), Mean ± SD	27.46 ± 4.78	22.98 ± 2.24	25.90 ± 2.39	28.31 ± 2.90	32.67 ± 4.69	< 0.001
BMI (kg/m2), N (%)						< 0.001
< 25	128,778 (32.49%)	81,408 (82.15%)	36,916 (37.25%)	9,947 (10.04%)	507 (0.51%)
25 - <30	170,125 (42.92%)	17,448 (17.61%)	56,787 (57.31%)	64,782 (65.37%)	31,108 (31.39%)
≥ 30	97,480 (24.59%)	240 (0.24%)	5,393 (5.44%)	24,366 (24.59%)	67,481 (68.10%)
Smoking status, N (%)						< 0.001
current	41,446 (10.46%)	9,790 (9.88%)	10,513 (10.61%)	10,622 (10.72%)	10,521 (10.62%)
former	197,001 (49.70%)	45,010 (45.42%)	47,036 (47.47%)	49,679 (50.13%)	55,276 (55.78%)
never	157,936 (39.84%)	44,296 (44.70%)	41,547 (41.93%)	38,794 (39.15%)	33,299 (33.60%)
Alcohol intake (g/day), N (%)						< 0.001
>16	151,409 (38.20%)	31,665 (31.95%)	37,201 (37.54%)	41,093 (41.47%)	41,450 (41.83%)
≤16	123,523 (31.16%)	38,205 (38.55%)	32,735 (33.03%)	28,325 (28.58%)	24,258 (24.48%)
never	121,451 (30.64%)	29,226 (29.49%)	29,160 (29.43%)	29,677 (29.95%)	33,388 (33.69%)
Sedentary (hours), Mean ± SD	4.82 ± 2.46	4.01 ± 2.08	4.57 ± 2.25	5.02 ± 2.42	5.70 ± 2.71	< 0.001
Sleep (hours/day), N (%)						< 0.001
< 7	97,699 (24.65%)	21,380 (21.58%)	23,537 (23.75%)	25,199 (25.43%)	27,583 (27.83%)
7–9	291,610 (73.57%)	76,643 (77.34%)	74,174 (74.85%)	72,112 (72.77%)	68,681 (69.31%)
> 9	7074 (1.78%)	1,073 (1.08%)	1,385 (1.40%)	1,784 (1.80%)	2,832 (2.86%)
Healthy diet, N (%)	192,646 (48.60%)	54,045 (54.54%)	49,281 (49.73%)	45,971 (46.39%)	43,349 (43.74%)	< 0.001
Sufficient activity, N (%)	285,516 (72.03%)	76,811 (77.51%)	74,221 (74.90%)	71,059 (71.71%)	63,425 (64.00%)	< 0.001
Vitamin D (ng/ml), Mean ± SD	19.45 ± 8.42	20.86 ± 8.83	20.03 ± 8.45	19.18 ± 8.20	17.73 ± 7.87	< 0.001
CRP (mg/L), Mean ± SD	2.60 ± 4.36	1.52 ± 3.43	2.20 ± 3.96	2.83 ± 4.41	3.84 ± 5.13	< 0.001
CKM category, N (%)						< 0.001
CKM 0	66,283 (16.72%)	51,106 (51.57%)	13,125 (13.24%)	2,009 (2.03%)	43 (0.04%)
CKM I	60,884 (15.36%)	14,539 (14.67%)	24,882 (25.11%)	16,231 (16.38%)	5,232 (5.28%)
CKM II-III	239,414 (60.40%)	31,570 (31.86%)	56,789 (57.31%)	72,964 (73.63%)	78,091 (78.80%)
CKM IV	29,802 (7.52%)	1,881 (1.90%)	4,300 (4.34%)	7,891 (7.96%)	15,730 (15.87%)
Respiratory disease	61,318 (15.47%)	14,230 (14.36%)	14,928 (15.06%)	15,412 (15.55%)	16,748 (16.90%)	< 0.001
Digestive disease	94,492 (23.84%)	18,244 (18.41%)	21,584 (21.78%)	25,351 (25.58%)	29,313 (29.58%)	< 0.001
Neurodegenerative disease	53,668 (13.54%)	12,629 (12.74%)	12,811 (12.93%)	12,954 (13.07%)	15,274 (15.41%)	< 0.001
Cancer	34,584 (8.72%)	8,698 (8.78%)	8,607 (8.69%)	8,449 (8.53%)	8,830 (8.91%)	< 0.001

P-values for continuous and categorical variables were calculated using the Kruskal-Wallis test and chi-square (χ²) test, respectively

METS-VF, Visceral Fat Metabolism Score; TDI, Townsend deprivation index; BMI, body mass index; CRP, C-reaction protein; CKM, Cardiovascular-Kidney-Metabolic syndrome; SD, standard deviation

Association of CKM and the METS-VF with all-cause and cause-specific mortality

During the follow-up period (median: 12.51 years), 34,471 participants died, including 7,237 due to cardiovascular diseases, 2,466 due to respiratory diseases, 1,318 due to digestive diseases, 2,029 due to neurodegenerative diseases, 16,679 due to cancer-related causes, and 3,447 due to other causes (two causes of death were reported for three participants). Both all-cause and cause-specific mortality increased significantly as the METS-VF increased (Fig. 1a and b). As the CKM stage worsened, the mortality rate increased, especially in participants with stage IV CKM, among whom the incidence rates of all-cause mortality and cause-specific mortality showed the most pronounced increase compared with the stage III CKM (Fig. 1c and d). The cumulative incidence displayed in the KM curves of the METS-VF and stage of CKM and cause-specific mortality support these results (Fig. S2 and Fig. S3).

Fig. 1.

Stacked Bar Charts of All-Cause Mortality by METS-VF Quartiles (a) and CKM Stages (c). Cumulative Incidence Curve of All-Cause Mortality by METS-VF Quartiles (b) and CKM Stages (d). Abbreviations: METS-VF, Visceral Fat Metabolism Score; CKM, Cardiovascular-Kidney-Metabolic syndrome

In the multivariate Cox model, different stages of CKM and METS-VF quartiles were significantly associated with an increased risk of all-cause mortality (Table 2). Compared to stage 0 CKM, the HRs for stage I, stage II-III, and stage IV CKM I were 1.15 (95% CI: 1.09, 1.22), 1.31 (95% CI: 1.25, 1.37), and 2.23 (95% CI: 2.12, 2.35), respectively. Similarly, compared to the lowest METS-VF quartile, the HRs for the second, third, and fourth quartiles were 1.11 (95% CI: 1.07, 1.16), 1.31 (95% CI: 1.25, 1.37), and 1.70 (95% CI: 1.61, 1.80), respectively. For per 0.1 increasement in the METS-VF, the risk of all-cause mortality increased by 4% (Table S4).

Table 2.

Associations of METS-VF and CKM stages with All-Cause mortality

All-cause mortality	N (%)	Model 1 HRs (95% CIs)	Model 2 HRs (95% CIs)	Model 3 HRs (95% CIs)	P for trend
CKM
CKM 0	3,005 (4.53)	ref	ref	ref
CKM I	3,331 (5.47)	1.18 (1.11,1.25)	1.16 (1.10,1.23)	1.15 (1.09,1.22)	< 0.001
CKM II-III	21,204 (8.86)	1.38 (1.33,1.44)	1.35 (1.29,1.41)	1.31 (1.25,1.37)
CKM IV	6,931 (23.26)	2.47 (2.36,2.58)	2.31 (2.20,2.43)	2.23 (2.12,2.35)
METS-VF
Q1	4,746 (4.79)	ref	ref	ref
Q2	6,393 (6.45)	1.23 (1.19,1.28)	1.15 (1.11,1.20)	1.11 (1.07,1.16)	< 0.001
Q3	8,674 (8.75)	1.57 (1.51,1.62)	1.40 (1.34,1.47)	1.31 (1.25,1.37)
Q4	14,658 (14.79)	2.40 (2.32,2.48)	1.89 (1.79,2.00)	1.70 (1.61,1.80)

Q1, Q2, Q3, and Q4 indicated the 1 st, 2nd, 3rd, and 4th quartiles, respectively

Model 1 was adjusted for Age, Sex, Ethnicity, Career, Education and TDI

Model 2 was further adjusted for BMI, Smoking status, Alcohol consumption, Sedentary, Sufficient activity, Healthy diet and Sleep based on Model 1

Model 3 was further adjusted for CRP and Vitamin based on Model 2

P for trend was calculated by treating the ordered categorical variables as continuous variables

METS-VF, Visceral Fat Metabolism Score; CKM, Cardiovascular-Kidney-Metabolic syndrome; HRs, hazard ratios; CIs, confidence intervals

The stage of CKM and the METS-VF were associated with the risk of several cause-specific mortalities (Table 3). After adjusting for all covariates, the HR for stage IV CKM compared to stage 0 CKM was 5.06 (95% CI: 4.46, 5.74) for CVD mortality, 2.63 (95% CI: 2.20, 3.14) for respiratory disease mortality, 2.49 (95% CI: 1.90, 3.26) for digestive disease mortality, 1.25 (95% CI: 1.02, 1.51) for neurodegenerative disease mortality, 1.51 (95% CI: 1.40, 1.62) for cancer mortality, and 2.26 (95% CI: 1.97, 2.60) for mortality due to other causes. Similarly, when comparing the highest METS-VF quartile to the lowest, the HR was 2.36 (95% CI: 2.06, 2.69) for CVD mortality, 1.69 (95% CI: 1.37, 2.07) for respiratory disease mortality, 2.86 (95% CI: 2.11, 3.88) for digestive disease mortality, 1.50 (95% CI: 1.20, 1.88) for neurodegenerative disease mortality, 1.50 (95% CI: 1.39, 1.62) for cancer mortality, and 1.72 (95% CI: 1.48, 2.01) for other causes of mortality. All trends were statistically significant except for CKM and neurodegenerative disease mortality. For each 0.1-unit increase in the METS-VF, the risk of mortality increased by 8.2% for CVD, 3.5% for respiratory disease, 8.8% for digestive disease, 2.7% for cancer, and 6.9% for death due to other causes (Table S4). No statistically significant difference was observed in neurodegenerative disease-related deaths. The VIFs of each variable in the above model were all < 5 (Table S5).

Table 3.

Associations of METS-VF and CKM stages with Cause-Specific mortality

Mortality	CKM 0	CKM I	CKM II-III	CKM IV	P for trend*	Q1	Q2	Q3	Q4	P for trend
CVD					< 0.001					< 0.001
N (%)	364 (0.55)	477 (0.78)	3997 (1.67)	2399 (8.05)		623 (0.63)	1,082 (1.09)	1,777 (1.79)	3,755 (3.79)
Model 1	ref	1.24 (1.08,1.42)	2.03 (1.82,2.27)	6.31 (1.82,2.27)		ref	1.59 (1.44,1.76)	2.44 (2.22,2.67)	4.64 (4.25,5.05)
Model 2	ref	1.22 (1.05,1.42)	1.79 (1.59,2.02)	5.28 (4.65,5.99)		ref	1.33 (1.20,1.48)	1.79 (1.60,2.01)	2.66 (2.33,3.04)
Model 3	ref	1.21 (1.04,1.41)	1.73 (1.53,1.94)	5.06 (4.46,5.74)		ref	1.28 (1.15,1.42)	1.66 (1.48,1.86)	2.36 (2.06,2.69)
Cancer					< 0.001					< 0.001
N (%)	1,698 (2.56)	1,885 (3.10)	10,806 (4.51)	2,290 (7.68)		2,689 (2.71)	3,384 (3.41)	4,318 (4.36)	6,288 (6.35)
Model 1	ref	1.14 (1.07,1.21)	1.33 (1.26,1.40)	1.66 (1.56,1.78)		ref	1.17 (1.11,1.23)	1.42 (1.35,1.49)	1.92 (1.83,2.00)
Model 2	ref	1.13 (1.05,1.22)	1.27 (1.20,1.35)	1.54 (1.43,1.66)		ref	1.11 (1.05,1.18)	1.31 (1.22,1.40)	1.64 (1.52,1.78)
Model 3	ref	1.13 (1.05,1.22)	1.24 (1.17,1.32)	1.51 (1.40,1.62)		ref	1.08 (1.02,1.14)	1.23 (1.15,1.32)	1.50 (1.39,1.62)
Respiratory					< 0.001					< 0.001
N (%)	211 (0.32)	202 (0.33)	1,396 (0.58)	657 (2.20)		357 (0.36)	398 (0.40)	586 (0.59)	1,125 (1.14)
Model 1	ref	1.38 (1.11,1.67)	1.47 (1.26,1.68)	3.01 (2.55,3.57)		ref	1.02 (0.89,1.17)	1.54 (1.27,1.83)	2.39 (1.92,3.01)
Model 2	ref	1.34 (1.09,1.65)	1.39 (1.18,1.63)	2.79 (2.33,3.33)		ref	1.00 (0.86,1.16)	1.43 (1.20,1.69)	2.12 (1.73,2.61)
Model 3	ref	1.30 (1.06,1.61)	1.32 (1.12,1.55)	2.63 (2.20,3.14)		ref	0.93 (0.79,1.07)	1.24 (1.04,1.46)	1.69 (1.37,2.07)
Digestive					< 0.001					< 0.001
N (%)	94 (0.14)	119 (0.20)	819 (0.34)	306 (0.96)		131 (0.13)	211 (0.21)	335 (0.34)	641 (0.65)
Model 1	ref	1.57 (1.09,2.04)	1.66 (1.34,2.06)	2.92 (2.29,3.72)		ref	1.76 (1.39,2.22)	2.86 (2.21,3.32)	4.21 (3.05,5.44)
Model 2	ref	1.51 (1.12,2.04)	1.63 (1.28,2.08)	2.66 (2.03,3.49)		ref	1.63 (1.29,2.06)	2.58 (2.00,3.34)	3.72 (2.74,5.04)
Model 3	ref	1.49 (1.10,2.01)	1.54 (1.21,1.96)	2.49 (1.90,3.26)		ref	1.49 (1.78,1.87)	2.17 (1.68,2.81)	2.86 (2.11,3.88)
Neurodegenerative					0.315					< 0.001
N (%)	257 (0.39)	258 (0.42)	1237 (0.52)	277 (0.93)		351 (0.35)	477 (0.48)	513 (0.52)	688 (0.69)
Model 1	ref	1.17 (0.94,1.40)	1.09 (0.88,1.30)	1.22 (1.02,1.42)		ref	1.21 (1.06,1.39)	1.27(1.09,1.46)	1.63 (1.36,1.93)
Model 2	ref	1.22 (1.01,1.49)	1.09 (0.93,1.27)	1.27 (1.04,1.54)		ref	1.20 (1.03,1.40)	1.24 (1.03,1.49)	1.58 (1.27,1.98)
Model 3	ref	1.23 (1.01,1.50)	1.07 (0.91,1.25)	1.25 (1.02,1.51)		ref	1.18 (1.02,1.38)	1.21 (1.01,1.45)	1.50 (1.20,1.88)
Others					< 0.001					< 0.001
N (%)	382 (0.58)	390 (0.64)	2,951 (1.23)	1,022 (3.43)		596 (0.60)	841 (0.85)	1,146 (1.16)	2,162 (2.18)
Model 1	ref	1.06 (0.97,1.15)	1.45 (1.30,1.62)	2.62 (2.32,2.96)		ref	1.28 (1.15,1.42)	1.61 (1.45,1.78)	2.71 (2.47,2.97)
Model 2	ref	1.03 (0.88,1.20)	1.38 (1.22,1.56)	2.35 (2.05,2.70)		ref	1.16 (1.03,1.30)	1.36 (1.19,1.55)	1.94 (1.67,2.27)
Model 3	ref	1.02 (0.87,1.19)	1.33 (1.18,1.51)	2.26 (1.97,2.60)		ref	1.11 (1.01,1.25)	1.26 (1.10,1.43)	1.72 (1.48,2.01)

Q1, Q2, Q3, and Q4 indicated the 1 st, 2nd, 3rd, and 4th quartiles, respectively

Model 1 was adjusted for Age, Sex, Ethnicity, Career, Education, and TDI

Model 2 was further adjusted for BMI, Smoking status, Alcohol consumption, Sedentary, Sufficient activity, Healthy diet and Sleep based on Model 1

Model 3 was further adjusted for CRP, Vitamin baseline respiratory, digestive, neurodegenerative diseases, and cancer based on Model 2

P for trend was calculated by treating the ordered categorical variables as continuous variables

METS-VF, Visceral Fat Metabolism Score; CKM, Cardiovascular-Kidney-Metabolic syndrome; HRs, hazard ratios; CIs, confidence interval

The RCS analysis revealed a non-linear relationship between the METS-VF and mortality (Fig. 2). The relationships between the METS-VF and respiratory disease mortality and neurodegenerative disease mortality exhibited U-shaped dose-response curves, while those between the METS-VF and other mortalities showed increasing dose-response curves. In ROC analyses, METS-VF demonstrated an AUC of 0.641 for predicting all-cause mortality. Among cause-specific mortality, METS-VF showed the highest discriminative ability for CVD mortality (AUC = 0.704), followed by digestive disease mortality (AUC = 0.681), other causes of mortality (AUC = 0.650), respiratory disease mortality (AUC = 0.643), cancer mortality (AUC = 0.601), and neurodegenerative disease mortality (AUC = 0.568). All these AUC values were significantly higher than those of other obesity-related indices, as confirmed by DeLong tests (Table S6).

Fig. 2.

RCS Analysis of METS-VF with All-Cause and Cause-Specific Mortality. a, CVD Mortality; b, Cancer Mortality; c, Respiratory Mortality; d, Digestive Mortality; e, Neurodegenerative Mortality; f, Other Mortality. RCS was adjusted for Age, Sex, Ethnicity, Career, Education, TDI, BMI, Smoking status, Alcohol consumption, Sedentary, Sufficient activity, Healthy diet, Sleep, CRP, Vitamin, baseline respiratory, digestive, neurodegenerative diseases, and cancer

Joint associations of CKM stage and the METS-VF with all-cause and cause-specific mortality

The risks of all-cause mortality and cause-specific mortality increased in a stepwise manner as the METS-VF increased and the CKM stage worsened (Fig. 3 and Fig. S4). Participants with stage IV CKM with a high METS-VF had the highest risk of all-cause mortality, with an HR of 3.00 (95% CI: 2.79, 3.21) compared to participants with stage 0 CKM and a low METS-VF. Similar risk trends were observed for mortality due to CVD, respiratory diseases, cancer, and other causes, with HRs of 8.20 (95% CI: 6.90, 9.75), 3.82 (95% CI: 2.96, 4.94), 1.89 (95% CI: 1.71, 2.10), and 2.96 (95% CI: 2.30, 3.79), respectively. However, participants with a high METS-VF in stage 0 or stage I CKM had the highest risk of death due to digestive or neurodegenerative diseases, with HRs of 8.28 (95% CI: 3.55, 19.3) and 1.81 (95% CI: 1.30, 2.52), respectively. Nonetheless, stage IV CKM still corresponded to a higher risk of mortality.

Fig. 3.

Risk of Mortality According to CKM Stages Based on Different METS-VF Categories. a, All-Cause Mortality; b, CVD Mortality; c, Digestive Mortality; d, Cancer Mortality. All models were adjusted for Age, Sex, Ethnicity, Career, Education, TDI, BMI, Smoking status, Alcohol consumption, Sedentary, Sufficient activity, Healthy diet, Sleep, CRP, Vitamin, baseline respiratory, digestive, neurodegenerative diseases, and cancer. Abbreviations: METS-VF, Visceral Fat Metabolism Score; CKM, Cardiovascular-Kidney-Metabolic syndrome; HRs, hazard ratios; CIs, confidence intervals

The METS-VF and stage of CKM exhibited multiplicative interactions in terms of all-cause mortality, CVD mortality, and digestive diseases mortality. Further analysis of additive interactions across different groups revealed that the combination of a high METS-VF and stage II-III or stage IV CKM had an additive effect on the mortality risks. No interactions were observed for other types of mortality. Results from the subgroup analysis indicated that the joint effect of a high METS-VF and worse stage CKM on all-cause mortality exhibited effect modifications associated with smoking, alcohol consumption, and sleep, whereas no significant differences were observed in other categories (Tables S7–S12).

Sensitivity analysis

The findings from several sensitivity analyses aligned with those of the primary analysis. When overweight or obese individuals were excluded, a higher METS-VF was associated with a significantly increased risk of all-cause and cause-specific mortality. Therefore, BMI did not attenuate the effect of METS-VF on mortality (Table S13). In another analysis, participants who died within two years of enrolment and those with related diseases at baseline were excluded, a competing risk model for cause-specific mortality was used, and multiple imputations for missing covariates were performed, and the associations of the METS-VF and stages of CKM with mortality risk were consistent with the original results (Tables S14–S17).

Disscussion

Our study indicates that the stage of CKM and the METS-VF are not only associated with all-cause and CVD mortality, but are also strongly linked to cause-specific mortality due to respiratory, digestive, and nervous system diseases, cancer, and other causes. Additionally, the relationship between the METS-VF and mortality risk exhibited non-linear characteristics. Joint analysis revealed that individuals with stage IV CKM and a higher METS-VF had an extremely high risk of all-cause and cause-specific mortality, whereas a high METS-VF demonstrated a synergistic effect in participants with existing metabolic risk factors, further increasing the risk of mortality. In subgroup and sensitivity analyses, this association remained robust. These findings highlight that the METS-VF combined with the stage of CKM has the potential to serve as an indicator for early health risk assessment and proactive interventions to reduce mortality.

Currently, the METS-VF, a novel index for estimating visceral fat, has been confirmed to be associated with the risk of metabolic diseases such as hypertension, diabetes, and CVD. A study based on population from the United States reported that the risks of all-cause mortality, cardiovascular mortality, and cancer mortality increased by 13% (95% CI: 6%−20%), 18% (95% CI: 6%−31%), and 13% (95% CI: 3%−25%), respectively, for every 0.2 increase in the METS-VF [36]. Another longitudinal study regarding the cumulative METS-VF further revealed that participants with a high METS-VF had increased risks of all-cause mortality (HR: 2.78, 95% CI: 2.49–3.17) and cardiovascular mortality (HR: 4.90, 95% CI: 4.36–5.50) [21]. These findings are consistent with the conclusions of the current study and further confirm the effectiveness of the METS-VF as a predictor of mortality risk. Additionally, this study provides new evidence that a higher METS-VF is significantly associated with the risk of mortality from respiratory diseases, digestive diseases, neurodegenerative diseases, and other specific causes of death, and in individuals with normal BMI, higher METS-VF remained significantly associated with an increased risk of mortality.

Although visceral fat measured by DXA is currently regarded as the gold standard for assessing visceral obesity, its reliance on expensive equipment and complex operation limits its applicability in large scale population screenin [37, 38]. In addition, trunk fat percentage derived from BIA can rapidly estimate body fat distribution, but its measurements are susceptible to interference from factors such as device model, individual hydration status, recent food intake, and physical activity [39]. Moreover, the lack of standardized protocols across different manufacturers compromises the comparability and reliability of BIA results [40]. In contrast, METS-VF can be calculated using only six routinely collected clinical parameters, is easy to obtain and low-cost, and is suitable for both clinical practice and large-scale epidemiological studies. Previous studies have shown that, when validated against DXA and MRI as reference standards, METS-VF demonstrates superior performance in assessing visceral fat compared with other commonly used surrogate indices [20]. In our study, METS-VF showed significantly higher predictive performance for both all-cause and cause-specific mortality than other abdominal obesity indices and BIA-derived trunk fat percentage, further confirming its superiority and potential value as a biomarker of visceral adiposity–related risk.

The reasons visceral fat leads to diseases are multifaceted. Research has indicated that visceral fat is closely associated with the risk of respiratory diseases such as asthma and chronic obstructive pulmonary disease [41, 42]. This association may be due to the accumulation of adipose tissue within the airway walls, which leads to increased wall thickness and exacerbates airway inflammation, thereby affecting respiratory function [43]. In the digestive system, deposition of visceral fat and insulin resistance caused by obesity elevate the levels of free fatty acids and glucose in the body, subsequently leading to the accumulation of fat and triglycerides in the liver, which increases the incidence of non-alcoholic fatty liver disease [44]. Furthermore, the pro-inflammatory state mediated by inflammatory factors derived from fat cells may contribute to pancreatitis, inflammatory bowel disease, and various gastrointestinal dysfunctions [45, 46]. Research regarding the nervous system indicates that individuals with higher levels of abdominal fat have a significantly increased risk of cognitive decline and neurodegenerative diseases, such as Alzheimer’s and Parkinson’s diseases [47, 48]. This may be related to the action of pro-inflammatory cytokines within the vagus nerve and across the blood-brain barrier to affect the brain and hypothalamus, as well as the gut microbiota dysbiosis due to obesity [16, 49, 50]. However, evidence associating visceral fat with mortality from these diseases is limited as the progression from such diseases to death is prolonged and more likely to be observed in the elderly population. In the current study, the mean baseline age of the deceased population was 67 years, with a mean age at death of 71.8 years. In contrast, the mean age at baseline enrolment for the non-deceased population was 57 years and the mean age at the end of follow-up was approximately 69.5 years. Therefore, participants who were ill but had not yet died remained at risk of experiencing mortality events in subsequent years. These findings encompass nearly all major causes of death, further underscoring the significant role of the METS-VF in various causes of mortality.

Similar to the METS-VF, the stage of CKM was also found to be significantly associated with the risk of other cause-specific mortality. This indicates that CKM is a complex metabolic disorder which affect the cardiovascular system and impact various organs throughout the body [51–53]. When the CKM staging was combined with METS-VF, a more pronounced joint effect was observed. The combination of an increased METS-VF and a worse stage of CKM further increased the risk of both all-cause and cause-specific mortality. Furthermore, these two indicators exhibited a synergistic effect on all-cause mortality, cardiovascular mortality, and mortality due to digestive system diseases. This synergy may stem from the combined action of chronic inflammation, insulin resistance, and metabolic disorders [54, 55], thereby accelerating functional decline and disease progression across multiple organ systems. Therefore, integrating CKM staging with METS-VF assessment may enhance the identification of individuals at high risk of mortality, particularly among those in later stages of CKM. The subgroup analysis indicated that as the METS-VF increased and the stage of CKM worsened, the impact on mortality could not be significantly alleviated by a healthy lifestyle, suggesting that relying solely on lifestyle improvements may be inadequate to offset the increased mortality risk associated with increased visceral fat and CKM progression. Therefore, early screening, prevention, and management within the population are essential to reduce premature deaths attributable to these factors.

This study had several key findings. First, the use of a large population and extended follow-up period from the UK Biobank database significantly enhanced the robustness of our findings. Second, while previous studies have primarily focused on the associations between CKM stages or METS-VF and all-cause and cardiovascular mortality, a more comprehensive analysis that included respiratory diseases, digestive diseases, neurodegenerative diseases, cancer, and other specific causes of death was conducted in this study, thereby encompassing nearly all possible types of disease-related mortality. Third, the METS-VF, a measure of visceral fat, has demonstrated superior predictive performance compared with other related indicators in prior research and is both low-cost and easily accessible. The METS-VF was analysed in conjunction with CKM in the current study, which aligns with the early screening and prevention recommendations proposed by the AHA. Finally, multiple potential confounding variables were adjusted for in this study, including social, economic, and behavioural factors. In addition, the multicollinearity of the model was evaluated using the variance inflation factor, ensuring the independence of the METS-VF and stage of CKM from various mortality risks.

However, this study has several limitations. First, due to the lack of data regarding subclinical cardiovascular disease in the UK Biobank, CKM stage II and stage III were merged into one group in the current study, which may have underestimated the mortality risk associated with stage III CKM. Second, visceral fat and CKM were measured only at baseline; therefore, changes that may have occurred during follow-up were not recorded. Some participants might have progressed through stages during this period, potentially leading to mortality. Third, the majority of participants in the UK Biobank are of European descent and experience more favorable living conditions and better medical services than other populations, suggesting that the results may not be representative of other populations. Fourth, despite adjustments for potential confounding factors, underlying environmental and genetic factors may have influenced the outcomes. Fifth, although our study demonstrates that METS-VF outperforms other surrogate indices of visceral adiposity in predicting mortality risk, DXA-derived visceral fat measures in the UK Biobank were collected during the third follow-up visit, whereas the parameters required for METS-VF were not assessed concurrently. Consequently, a direct comparison of their predictive performance for mortality was not feasible. Sixth, in our study, cancer mortality was markedly higher than other cause-specific mortality. This is primarily because the UK Biobank cohort had a baseline mean age of 58 years, and the risk of mortality from chronic conditions, typically rises substantially after age 70 [56, 57]; in cohorts predominantly composed of middle-aged individuals, cancer is a more common cause of mortality [58]. Furthermore, with an average follow-up duration of approximately 10–12 years, many participants had not yet reached the high risk age range for other cause-specific mortality. Therefore, future studies should include cohorts enriched with older adults to further validate the predictive performance of METS-VF for various mortality outcomes across different age groups. Finally, although participants who did not survive beyond two years of enrolment were excluded in the sensitivity analysis, the observational nature of the study limits the ability to establish causality, necessitating further interventional studies for validation.

Conclusion

The results of this study indicate a significant association between the METS-VF and stages of CKM with all-cause and cause-specific mortality. Incorporating the METS-VF into the assessment of CKM aligns with the recommendations proposed by the AHA. The clinical use of these indicators would allow for earlier identification of high-risk individuals and the implementation of effective prevention and management measures, thereby reducing premature deaths from various causes.

Abbreviations

AHA

American heart association

AUC

Area under the curve

BMI

Body mass index

BRI

Body roundness index

CI

Confidence interval

CKD

Chronic kidney disease

CKM

Cardiovascular-kidney-metabolic syndrome

CVD

Cardiovascular disease

DXA

Double-energy X-ray absorptiometry

GBD

Global burden of disease

HRs

Hazard ratios

KM

Kaplan-Meier

LAP

Lipid accumulation product

METS-IR

Metabolism score for insulin resistance

METS-VF

Visceral fat metabolism score

RCS

Restricted cubic spline

ROC

Receiver operating characteristic

TDI

Townsend deprivation index

WC

Waist circumference

WHtR

Waist-to-height ratio

WWI

Weight-adjusted waist index

VAI

Visceral adiposity index

VIFs

Variance inflation factors

Author contributions

Jilong Bai, conceptualisation, methodology, data analysis and writing-original manuscript. Panting Wei, validation and visualisation. Yao Zhang, visualisation. Liying Wang, visualisation. Jinyan Liu, visualisation. Wenxu Wang, visualisation. Difei Wang, supervision and writing-review & editing, Wan Yu, supervision and writing-review & editing All authors had full access to all the data in the study and have read and approved the final manuscript.

Funding

This study was supported by the Major Program of the National Natural Science Foundation of China (Grant No. 12474435), the Major Science and Technology Program of Liaoning Province’s “Jie Bang Guai Shuai” initiative (Grant No. 2022JH1/10400001), and the Liaoning Provincial People’s Livelihood Science and Technology Program Joint Project (Grant No. 2021JH2/10300090). The funders had no role in the study design, implementation, data collection, management, analysis and interpretation; preparation, review, or approval of the manuscript; or the decision to submit the manuscript for publication.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

UK Biobank received ethical approval from the North West Multi-center Research Ethics Committee (21/NW/0157), and all participants provided written informed consent at the time of participation. The study adhered to the principles of the Declaration of Helsinki.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.