Association of the skeletal muscle mass to visceral fat area ratio with metabolically healthy obesity and metabolically unhealthy non-obesity: a cross-sectional study based on NHANES 2011–2018

Abstract

Background

The interplay between skeletal muscle loss and excess adiposity has been associated with metabolic dysfunction and may impose a dual burden on physiological homeostasis. This study aimed to investigate the associations of the skeletal muscle mass to visceral fat area ratio (SVR) with metabolically healthy obesity (MHO) and metabolically unhealthy non-obesity (MUNO).

Methods

Data from the 2011–2018 NHANES were analyzed. Obesity phenotypes were defined using body mass index (BMI ≥ 30 kg/m²), in combination with the presence of metabolic syndrome components. Associations of SVR with MHO and MUNO were assessed using multivariable logistic regression and restricted cubic spline (RCS) analyses. Discriminatory performance was evaluated by area under the curve (AUC) values from receiver operating characteristic (ROC) analyses, comparing SVR with other anthropometric and body composition measures. Subgroup analyses were conducted to evaluate potential effect modification, and sensitivity analyses were performed to assess the robustness of the findings.

Results

Among 4,609 adults, 5.06% were classified as MHO and 39.97% as MUNO. After covariate adjustment, higher SVR was associated with greater odds of MHO (OR = 1.48; 95% CI: 1.16–1.89) and lower odds of MUNO (OR = 0.47; 95% CI: 0.37–0.60). RCS analyses suggested linear associations for MHO and MUNO. Subgroup analyses suggested stronger associations for MHO among current or former drinkers, and for MUNO among men and adults aged 45–59 years. ROC analyses showed modest discriminatory performance, with AUC values of 0.65 for MHO and 0.71 for MUNO.

Conclusions

SVR was associated with both MHO and MUNO. These findings suggest that SVR may provide supplementary information for characterizing obesity phenotypes, but its clinical relevance remains uncertain and requires further confirmation in longitudinal and interventional studies.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12944-025-02744-x.

Introduction

Obesity is a persistent, multifactorial disorder characterized by excessive or dysregulated fat accumulation in adipose tissue [1]. The World Health Organization (WHO) has identified obesity as a major global health concern due to its strong associations with metabolic syndrome and increased risk of comorbidities such as cardiovascular diseases (CVD), hypertension, chronic kidney disease (CKD), and type 2 diabetes mellitus (T2DM) [2–4]. Although body mass index (BMI) is commonly used to classify overweight and obesity, its ability to reflect the true metabolic and health consequences of fat accumulation is limited [5]. Substantial differences in metabolic health status and comorbidities may occur among individuals with the same BMI, largely due to BMI’s inability to differentiate fat from lean mass [6, 7]. These limitations are exemplified by the metabolically healthy obesity (MHO) phenotype, wherein some obese individuals maintain normal metabolic function despite excess adiposity [8]. The recognition of MHO highlights the heterogeneity of obesity, indicating that not all individuals with excess adiposity inevitably develop metabolic disorders. By contrast, metabolically unhealthy non-obesity (MUNO) individuals frequently present with insulin resistance, elevated blood pressure, oxidative stress, and atherogenic dyslipidemia [9–11]. These observations underscore the limitations of relying solely on BMI for risk assessment, especially in populations with MHO or MUNO, where it may result in misleading health guidance and management.

Considering the severity and heterogeneity of obesity-related complications, early stratification of individuals with MHO and MUNO is critical for optimizing clinical care and reducing long-term risks. Longitudinal evidence over the past two decades suggests that MHO is often a transient state, with most individuals eventually progressing to metabolically unhealthy phenotypes or subclinical cardiovascular disease [12]. This dynamic progression underscores the limitations of single-time-point phenotyping and emphasizes the need for reliable indicators that can sensitively monitor metabolic health trajectories. Current diagnostic protocols for metabolic syndrome, which depend on repeated blood sampling for multi-component metabolic panels, impose substantial clinical burdens because of frequent blood draws and their invasive nature. Therefore, this gap underscores the urgent need for minimally invasive, rapid-assessment indicators that can capture metabolic health status across obesity phenotypes and provide a practical means to track transitional risks in high-risk populations.

Growing evidence demonstrates a strong association between changes in body composition and metabolic dysregulation [13–15]. Obesity is frequently accompanied by reduced muscle mass, and reductions in appendicular skeletal muscle mass (ASM) have been linked to metabolic dysfunction [16, 17]. Abdominal obesity, especially an increased visceral fat area (VFA), has emerged as a key contributor to metabolic syndrome (MetS) and CVD [18–20]. Unlike subcutaneous fat, visceral fat surrounds internal organs and exhibits substantially higher metabolic activity [21, 22]. Higher VFA is associated with impaired insulin sensitivity, dyslipidemia, and broader metabolic disturbances [23, 24].

Skeletal muscle mass and visceral fat exhibit opposite associations with metabolic health. Skeletal muscle promotes glucose uptake and enhances insulin sensitivity, while visceral fat is closely associated with chronic inflammation and insulin resistance [25, 26]. Individuals with both reduced muscle mass and excess fat, a condition termed sarcopenic obesity, face a higher cardiometabolic risk than those with either abnormality alone [27]. Previous studies have reported that a lower skeletal muscle mass to visceral fat area ratio (SVR) is independently associated with higher 10-year cardiovascular risk beyond BMI, as well as with increased prevalence of metabolic-associated fatty liver disease and liver fibrosis [28, 29]. Emerging evidence also suggests that SVR may be a potential indicator of risk for T2DM and metabolic dysfunction-associated steatotic liver disease (MASLD) [30, 31]. Taken together, these findings suggest that SVR reflects the relative proportion of muscle to visceral fat, two compartments with opposing metabolic effects, and may therefore have potential value in reflecting obesity-related heterogeneity and in differentiating between obesity phenotypes.

However, the value of SVR in characterizing obesity phenotypes has not been fully assessed. Conventional measures such as BMI cannot differentiate lean mass from fat mass and therefore do not adequately capture the heterogeneity of obesity-related metabolic risk [5, 8]. Clarifying the potential utility of SVR may help address this gap, as individuals with the same BMI can present very different metabolic profiles. For instance, individuals with MHO and metabolically unhealthy obesity (MUO), or those with MUNO and metabolically healthy non-obesity (MHNO), may share similar BMI levels but differ substantially in insulin resistance, cardiometabolic complications, and long-term outcomes [32]. If validated as a supplementary indicator, SVR may facilitate earlier identification of high-risk individuals and provide a reference for future exploration of tailored interventions, while also helping to avoid unnecessary treatment in low-risk groups.

Because SVR integrates both skeletal muscle and visceral fat, two components closely related to metabolic health, it may complement BMI in reflecting obesity-related heterogeneity [30, 33]. Evaluating its capacity to differentiate obesity phenotypes could improve understanding of metabolic differences among obese individuals and support more refined risk stratification. Such findings may have implications for public health screening and the prevention of obesity-related complications, although further validation is required before clinical application.

This exploratory cross-sectional study examined the potential association between SVR and obesity phenotypes. Using data from NHANES 2011 to 2018, we estimated the prevalence of MHO and MUNO, performed multivariable-adjusted analyses to explore their associations with SVR, and considered the potential value of SVR in characterizing obesity phenotypes.

Methods

Data source

Administered by the National Center for Health Statistics (NCHS), NHANES collects cross-sectional data to investigate associations between nutritional status, health behaviors, and chronic disease epidemiology. The survey is conducted biennially and includes physical examinations, structured interviews, and assessments of dietary intake, demographic characteristics, clinical measures, and laboratory tests [34]. The NHANES protocol was approved by the NCHS Research Ethics Review Board, and written informed consent was obtained from all participants.

Study population

Baseline clinical data were obtained from four NHANES cycles (2011–2018), selected because the whole-body dual-energy X-ray absorptiometry (DXA) data and complete VFA data were available only during this period. A total of 39,156 participants were initially enrolled across these four cycles. Participants younger than 20 years (n = 16,539) were excluded because the definitions of MHO and MUNO are based on adult metabolic criteria. Consistent with NHANES analytic practice and prior epidemiological studies, an age threshold of ≥ 20 years was applied to minimize the potential influence of late adolescent developmental changes [35–37]. Individuals with incomplete ASM and VFA data (n = 12,019), missing BMI data (n = 33), or key metabolic syndrome data (n = 5,956) were also excluded. After these exclusions, 4,609 participants remained in the final analytical sample (Fig. 1).

Fig. 1.

The selection flowchart of the participants from NHANES 2011–2018

Exposure variable

SVR (kg/cm²) was calculated as ASM divided by VFA. Whole-body composition, including lean mass, adipose tissue, and bone mineral content, was quantified using DXA. ASM was derived by summing the lean mass of both upper and lower limbs, measured according to standardized DXA protocols [38].

Obesity phenotype criteria

Obesity was defined as BMI ≥ 30 kg/m², in accordance with established definitions [39–41]. Metabolic abnormalities were defined as the presence of at least one of the following criteria: (1) systolic blood pressure (SBP) ≥ 130 mmHg, diastolic blood pressure (DBP) ≥ 85 mmHg, or use of antihypertensive medication; (2) fasting plasma glucose (FPG) ≥ 100 mg/dL or use of antidiabetic medication; (3) high-density lipoprotein cholesterol (HDL-C) < 40 mg/dL in men or < 50 mg/dL in women; (4) triglycerides (TG) ≥ 150 mg/dL. Participants were classified into four phenotypes: metabolically healthy obesity (MHO; obese without metabolic abnormalities), metabolically unhealthy obesity (MUO; obese with ≥ 1 metabolic abnormality), metabolically healthy non-obesity (MHNO; non-obese without metabolic abnormalities), and metabolically unhealthy non-obesity (MUNO; non-obese with ≥ 1 metabolic abnormality).

Covariables

Consistent with prior studies [31, 38, 42–46], the final models adjusted for the following covariates: age; gender; race; poverty income ratio (PIR); education; physical activity (PA); drinking status; smoking; CVD; CKD; total cholesterol (TC, mg/dL); low-density lipoprotein cholesterol (LDL-C, mg/dL); estimated glomerular filtration rate (eGFR, mL/min/1.73m²); and use of lipid-lowering medications. TC and LDL-C were dichotomized. Detailed descriptions of covariates and measurement methods are provided in Supplementary Table S1.

Statistical analysis

All analyses were performed in accordance with NHANES analytic guidelines [47]. To account for the complex survey design and ensure nationally representative estimates, sampling strata, primary sampling units, and fasting subsample weights were incorporated [39, 48]. For the 2011–2018 cycles, sampling weights were calculated as WTSAF2YR/4. Weighted analyses were applied to baseline characteristics, multivariable logistic regression models, and subgroup analyses. RCS and ROC analyses were conducted without survey weights, as no standard implementation is currently available for complex, multistage survey designs. Similar analytical approaches have been used in previous NHANES studies [49–51], and the corresponding results should be interpreted as exploratory. To enhance the robustness of the analysis and minimize bias from missing data, covariates with missing values were imputed using the nonparametric missForest algorithm in R, which applies random forest-based iterative imputation to handle mixed data types while preserving multivariate relationships [52, 53]. The extent of missingness for each covariate is summarized in Supplementary Table S2, with a graphical overview shown in Supplementary Figure S1. Continuous variables are presented as means with standard errors (SE), and categorical variables as numbers with weighted percentages. Group differences were evaluated using Kruskal-Wallis tests for continuous variables and chi-squared tests for categorical variables, with Bonferroni corrections applied for multiple comparisons.

To improve interpretability and enable comparisons across covariates, SVR was standardized and analyzed per one standard deviation (SD), consistent with prior NHANES studies [54, 55]. Analyses were conducted separately for obese and non-obese groups. Among obese participants, MHO was compared with MUO; among non-obese participants, MUNO was compared with MHNO. Associations of SVR with MHO and MUNO were assessed using multivariable logistic regression, with odds ratios (ORs) and 95% confidence intervals (CIs) reported. Variance inflation factor (VIF) analysis was used to assess multicollinearity, and covariates with a VIF ≥ 5 were excluded. Three models were constructed: Model 1 was unadjusted. Model 2 adjusted for age, gender, and race, and Model 3 additionally adjusted for PIR, education level, PA, TC, LDL-C, eGFR, smoking status, drinking status, CVD, CKD, and lipid-lowering medication use. Subgroup analyses used multivariable logistic models to evaluate interactions stratified by age, gender, race, PIR, PA, smoking status, drinking status, and lipid-lowering medication use. Non-linear associations of SVR with MHO and MUNO were examined using restricted cubic spline (RCS) regression.

Receiver operating characteristic (ROC) curve analyses were conducted to evaluate the discriminatory ability of SVR compared with other anthropometric and body composition indices, including ASM, VFA, waist circumference (WC), waist-to-height ratio (WHtR), and weight-adjusted waist index (WWI), in differentiating MHO and MUNO. Model performance was further assessed using the Hosmer-Lemeshow goodness-of-fit test for calibration, and reclassification was evaluated by net reclassification improvement (NRI) and integrated discrimination improvement (IDI). Two sensitivity analyses were performed to assess the robustness of the results. First, participants with missing covariate data were excluded to evaluate potential impacts on effect estimates. Second, participants taking lipid-lowering medications were excluded to minimize potential confounding from pharmacological treatment on the associations between obesity phenotypes and other factors.

All analyses were conducted using R software (version 4.3.3; https://www.r-project.org) and EmpowerStats (https://www.empowerstats.com). A two-sided P-value < 0.05 was considered statistically significant.

Results

Baseline characteristics of participants

Baseline characteristics of participants, stratified by obesity phenotype, are presented in Table 1. The mean age of participants was 39.47 years (SE, 0.28), and 51.05% were male and 48.96% were female. The racial distribution was 61.75% Non-Hispanic White, 10.64% Non-Hispanic Black, 10.55% Mexican American, and 17.06% other races or ethnicities. The prevalence of MHO and MUNO was 5.06% and 39.97%, respectively. The mean SVR was 0.31 kg/m² (SE, 0.01), with significant variation across phenotype groups (P < 0.001).

Table 1.

. Baseline characteristics of participants

Characteristic		Obesity Phenotype
	Overall	MHO (n=223)	MUO (n=1378)	MUNO (n=1848)	MHNO (n=1160)	P-valuea	P-adjustb
Age, years	39.47 (0.28)	36.24 (0.91)	41.20 (0.36)	41.71 (0.45)	34.79 (0.44)	<0.001	<0.001
PIR	2.86 (0.06)	2.73 (0.14)	2.64 (0.07)	2.92 (0.08)	3.03 (0.07)	<0.001	0.002
Physical activity, met-min/week	4433.48 (140.85)	4962.83 (579.61)	3994.79 (183.57)	4520.17 (200.54)	4681.13 (312.02)	0.074	>0.999
SVR, kg/cm2	0.31 (0.01)	0.27 (0.01)	0.21 (0.00c)	0.30 (0.01)	0.46 (0.01)	<0.001	<0.001
ASM (kg)	24.09 (0.13)	26.03 (0.50)	27.74 (0.23)	22.96 (0.21)	21.44 (0.21)	<0.001	<0.001
VFA (cm2)	103.01 (1.47)	116.35 (3.79)	154.42 (2.21)	93.18 (1.71)	58.91 (1.33)	<0.001	<0.001
BMI, kg/m2	28.44 (0.16)	33.99 (0.30)	35.90 (0.22)	25.54 (0.10)	23.58 (0.13)	<0.001	<0.001
SBP, mmHg	118.24 (0.36)	115.73 (0.81)	124.27 (0.58)	119.38 (0.51)	110.37 (0.40)	<0.001	<0.001
DBP, mmHg	71.29 (0.29)	69.77 (0.66)	74.98 (0.44)	71.82 (0.42)	66.70 (0.38)	<0.001	<0.001
Fasting blood glucose, mg/dL	104.17 (0.60)	92.82 (0.42)	115.25 (1.17)	106.24 (0.80)	91.03 (0.22)	<0.001	<0.001
Triglyceride, mg/dL	119.08 (2.53)	83.95 (2.49)	153.29 (4.38)	130.56 (3.88)	70.75 (1.19)	<0.001	<0.001
Total cholesterol, mg/dL	190.38 (0.80)	189.44 (2.71)	193.13 (1.35)	194.93 (1.44)	180.58 (1.33)	<0.001	<0.001
HDL-C, mg/dL	53.31 (0.37)	56.16 (0.71)	46.15 (0.47)	51.52 (0.54)	63.37 (0.63)	<0.001	<0.001
LDL-C, mg/dL	114.15 (0.65)	116.45 (2.64)	118.17 (1.01)	118.22 (1.17)	103.08 (1.28)	<0.001	<0.001
eGFR, mL/min/1.73m2	102.55 (0.47)	105.34 (1.48)	102.39 (0.58)	100.60 (0.74)	105.17 (0.82)	<0.001	0.003
Age group, n (%)						<0.001	<0.001
20-44	2905 (61.84)	179 (79.10)	790 (57.18)	1028 (53.62)	908 (76.18)
45-59	1704 (38.16)	44 (20.90)	588 (42.82)	820 (46.38)	252 (23.83)
Gender, n (%)						<0.001	<0.001
Male	2300 (51.05)	82 (40.90)	614 (43.38)	1114 (59.99)	490 (42.27)
Female	2309 (48.96)	141 (59.10)	764 (51.62)	734 (40.01)	670 (57.73)
Race, n (%)						<0.001	<0.001
Non-Hispanic White	1625 (61.75)	74 (58.82)	481 (59.71)	624 (61.55)	446 (64.87)
Non-Hispanic Black	920 (10.64)	57 (14.57)	284 (13.76)	224 (7.55)	177 (10.00)
Mexican American	685 (10.55)	67 (14.65)	359 (14.08)	277 (7.82)	217 (10.39)
Other Race	1379 (17.06)	39 (11.94)	272 (12.34)	683 (20.81)	385 (17.53)
Education Level, n (%)						<0.001	<0.001
Less than high school	860 (13.80)	30 (12.72)	301 (15.88)	393 (15.46)	136 (9.21)
High School	972 (20.97)	42 (20.54)	325 (22.15)	403 (22.80)	202 (16.99)
College or above	2777 (65.23)	151 (66.74)	752 (61.98)	1052 (61.75)	822 (73.81)
PIR, n (%)						0.002	0.051
<=1.3	1409 (22.71)	67 (24.41)	475 (26.08)	553 (21.36)	314 (20.75)
>1.3, <=3.5	1893 (39.15)	89 (41.28)	592 (41.34)	748 (39.16)	464 (36.31)
>3.5	1307 (38.14)	67 (34.31)	311 (32.57)	547 (39.48)	382 (42.95)
Physical activity, n (%)						<0.001	<0.001
Inactive	1432 (28.91)	55 (25.21)	492 (35.26)	591 (29.74)	294 (21.38)
Active	3177 (71.09)	168 (74.79)	886 (64.74)	1257 (70.27)	866 (78.62)
Drinking status, n (%)						0.075	>0.999
Never	535 (8.69)	27 (11.51)	169 (10.15)	208 (8.03)	131 (7.56)
Current or ever	4074 (91.31)	196 (88.49)	1209 (89.85)	1640 (91.97)	1029 (92.44)
Smoking status, n (%)						<0.001	<0.001
Never	2754 (57.59)	155 (71.11)	805 (56.81)	1016 (52.04)	778 (64.32)
Current or ever	1855 (42.41)	68 (28.89)	573 (43.19)	832 (47.97)	382 (35.68)
CVD, n (%)						0.005	0.124
No	4432 (96.48)	221 (98.59)	1032 (95.44)	1763 (95.54)	1146 (98.66)
Yes	177 (3.52)	2 (1.41)	76 (4.56)	85 (4.46)	14 (1.34)
CKD, n (%)						<0.001	<0.001
No	4207 (92.61)	214 (96.01)	1185 (89.03)	1709 (93.88)	1099 (93.96)
Yes	402 (7.39)	9 (3.99)	193 (10.97)	139 (6.12)	61 (6.04)
Lipid-lowering medication use, n (%)						<0.001	<0.001
No	4205 (90.24)	220 (97.62)	1188 (84.98)	1658 (88.72)	1139 (96.93)
Yes	404 (9.76)	3 (2.38)	190 (15.03)	190 (11.28)	21 (3.08)

 Variables are expressed as means (standard errors) for continuous variables to indicate the precision of the estimated mean values and as n (weighted percentage) for categorical variables

Abbreviations: PIR poverty income ratio,met-min/weekmetabolic equivalents-minutes/week,SVRskeletal muscle mass to visceral fat area ratio,ASMappendicular skeletal muscle mass,VFA visceral fat area, BMIbody mass index,SBPsystolic blood pressure,DBPdiastolic blood pressure,HDL-Chigh-density lipoprotein cholesterol,LDL-Clow-density lipoprotein cholesterol,eGFRestimated glomerular filtration rate,CVDcardiovascular disease,CKDchronic kidney disease,MHOmetabolically healthy obesity,MUOmetabolically unhealthy obesity,MUNOmetabolically unhealthy non-obesity,MHNOmetabolically healthy non-obesity

^a Kruskal-Wallis tests for continuous variables; Chi-squared tests for categorical variables

^bBonferroni corrections were applied for comparisons among multiple groups

^cSE values <0.01 are shown as 0.00 due to rounding

Among the 33.84% of participants classified as obese, those with MHO were younger, more often female, and had fewer comorbidities than those with MUO. Conversely, among the 66.16% of participants without obesity, those with MUNO were older, predominantly male, and had more comorbidities than those with MHNO. With respect to lifestyle factors, participants with MHO or MHNO reported higher PA levels and lower smoking rates than those with MUO or MUNO (P < 0.001).

Multivariable logistic regression analysis results

Results from the multivariable logistic regression models are presented in Table 2. In Model 3, after covariate adjustment, SVR was positively associated with MHO when treated as a continuous variable (OR = 1.48; 95% CI: 1.16–1.89; P = 0.002). When SVR was categorized into tertiles, participants in the highest tertile had greater odds of MHO (OR = 2.18; 95% CI: 1.22–3.90; P = 0.012). Conversely, SVR was inversely associated with MUNO when analyzed as a continuous variable (OR = 0.47; 95% CI: 0.37–0.60; P < 0.001). Participants in the highest SVR tertile had lower odds of MUNO (OR = 0.20; 95% CI: 0.15–0.28; P < 0.001). Similar associations were observed when SVR was expressed per 0.1-unit increase (Supplementary Table S3). Additional analyses incorporating SVR and other covariates showed that SVR, age 45–59 years, female sex, and current or former smoking status were significantly associated with MHO (Supplementary Table S4). MUNO was significantly associated with SVR, age 45–59 years, and female sex (Supplementary Table S5). Following adjustment for various potential confounders, RCS analyses suggested linear associations of SVR with MHO (P for nonlinearity = 0.139, Supplementary Figure S2A) and MUNO (P for nonlinearity = 0.078, Supplementary Figure S2B).

Table 2.

Multivariable logistic regression models of SVR with MHO and MUNO

	Model 1		Model 2		Model 3
	OR (95%CI)	P-value	OR (95%CI)	P-value	OR (95%CI)	P-value
MHO
SVR	1.52 (1.26, 1.84)	< 0.001	1.59 (1.25, 2.03)	< 0.001	1.48 (1.16, 1.89)	0.002
SVR Tertiles
Low	1.00 (Reference)		1.00 (Reference)		1.00 (Reference)
Middle	1.36 (0.89, 2.09)	0.166	1.31 (0.81, 2.12)	0.278	1.17 (0.73, 1.87)	0.522
High	2.51 (1.59, 3.97)	< 0.001	2.64 (1.48, 4.72)	0.001	2.18 (1.22, 3.90)	0.012
P for trend		< 0.001		0.001		0.012
MUNO
SVR	0.41 (0.35, 0.49)	< 0.001	0.43 (0.34, 0.54)	< 0.001	0.47 (0.37, 0.60)	< 0.001
SVR Tertiles
Low	1.00 (Reference)		1.00 (Reference)		1.00 (Reference)
Middle	0.36 (0.27, 0.47)	< 0.001	0.34 (0.25, 0.46)	< 0.001	0.39 (0.28, 0.53)	< 0.001
High	0.16 (0.12, 0.21)	< 0.001	0.16 (0.12, 0.22)	< 0.001	0.20 (0.15, 0.28)	< 0.001
P for trend		< 0.001		< 0.001		< 0.001

Model 1 was unadjusted

Model 2 was adjusted for age, gender, and race

Model 3 was adjusted for age, gender, race, PIR, education level, PA, TC, LDL-C, eGFR, smoking status, drinking status, CVD, CKD, and lipid-lowering medication use

ORs and 95% CIs were expressed per 1-SD increase in SVR

For models of MHO, MUO served as the reference group; for models of MUNO, MHNO served as the reference group

Abbreviations: SVR skeletal muscle mass to visceral fat area ratio, MHO metabolically healthy obesity, MUNO metabolically unhealthy non-obesity, OR odds ratio, 95% CI 95% confidence interval, PIR poverty income ratio, PA physical activity, TC total cholesterol, LDL-C low-density lipoprotein cholesterol, eGFR estimated glomerular filtration rate, CVD cardiovascular disease, CKD chronic kidney disease, SD standard deviation

Subgroup analyses

To explore potential differences across demographic groups, stratified analyses were conducted to examine how the association between SVR and MHO or MUNO varied across subpopulations. As shown in Fig. 2A, SVR remained positively associated with MHO across most subgroups. However, drinking status significantly modified this association (P for interaction < 0.05). Among current or ever drinkers, per 1-SD increase in SVR was associated with higher odds of MHO (OR = 1.61; 95% CI: 1.24–2.10), compared with never drinkers.

Fig. 2.

Forest plot showing the subgroup analyses of the association between SVR and MHO (A) and MUNO (B). Note: Age, gender, race, PIR, education level, PA, TC, LDL-C, eGFR, smoking status, drinking status, CVD, CKD, and lipid-lowering medication use were adjusted in the subgroup analyses. ORs and 95% CIs were expressed per 1-SD increase in SVR. PIR, poverty income ratio; SVR, skeletal muscle mass to visceral fat area ratio; MHO, metabolically healthy obesity; MUNO, metabolically unhealthy non-obesity; OR, odds ratio; 95% CI, 95% confidence interval; PIR, poverty income ratio; PA, physical activity; TC, total cholesterol; LDL-C, low-density lipoprotein cholesterol; eGFR, estimated glomerular filtration rate; CVD, cardiovascular disease; CKD, chronic kidney disease

Figure 2B shows significant interactions between SVR and MUNO by age and gender (P for interaction < 0.05). Among age groups, participants aged 45–59 years showed a stronger inverse association between SVR and MUNO (OR = 0.36, 95% CI: 0.26–0.50) compared with those aged 20–44 years (OR = 0.51; 95% CI: 0.40–0.66). In sex-specific analyses, significant associations were observed for both males (OR = 0.37; 95% CI: 0.28–0.48) and females (OR = 0.55; 95% CI: 0.42–0.72), with a slightly stronger effect in males. These subgroup associations were consistent when SVR was expressed per 0.1-unit increase (Supplementary Figure S3).

ROC curve analyses

ROC curve analyses were conducted to evaluate the discriminatory ability of SVR in comparison with other anthropometric and body composition indicators (Fig. 3; Table 3). For MHO, SVR showed a modest discriminatory capacity with an AUC of 0.65 (95% CI: 0.61–0.69), slightly higher than VFA (0.63), WHtR (0.63), WC (0.62), and WWI (0.61), while outperforming ASM (0.56). For MUNO, SVR achieved the highest AUC of 0.71 (95% CI: 0.69–0.73), exceeding VFA (0.68), WHtR (0.66), WC (0.65), WWI (0.62), and ASM (0.57), and meeting the threshold for acceptable discrimination [56]. Overall, SVR demonstrated relatively better discriminatory performance than several commonly used anthropometric and body composition indices, particularly for MUNO, although its diagnostic value was modest and requires further validation.

Fig. 3.

ROC curves of six anthropometric and body composition indicators for differentiating MHO (A) and MUNO (B). SVR, skeletal muscle mass to visceral fat area ratio; BMI, body mass index; MHO, metabolically healthy obesity; MUNO, metabolically unhealthy non-obesity; ASM, appendicular skeletal muscle mass; VFA, visceral fat area; WHtR, waist-to-height ratio; WWI, weight-adjusted waist index; WC, waist circumference; 95% CI, 95% confidence interval; AUC, area under the curve

Table 3.

The AUC of anthropometric and body composition indicators for differentiating between MHO and MUNO

Indicator	Obesity Phenotype		AUC	95%CI	Cutoff value	Specificity	Sensitivity	PPV	NPV
SVR (kg/cm 2 )		MHO	0.65	(0.61–0.69)	0.20	0.59	0.66	0.21	0.92
		MUNO	0.71	(0.69–0.73)	0.29	0.71	0.59	0.76	0.52
VFA (cm 2 ) a		MHO	0.63	(0.60–0.67)	110.79	0.83	0.36	0.25	0.89
		MUNO	0.68	(0.66–0.70)	77.21	0.64	0.64	0.74	0.53
ASM (kg) a		MHO	0.56	(0.52–0.60)	27.31	0.47	0.65	0.17	0.89
		MUNO	0.57	(0.55–0.59)	21.17	0.58	0.59	0.69	0.47
WHtR a		MHO	0.63	(0.59–0.67)	0.65	0.56	0.65	0.19	0.91
		MUNO	0.66	(0.64–0.68)	0.52	0.61	0.64	0.72	0.51
WWI a		MHO	0.61	(0.57–0.65)	10.77	0.37	0.81	0.18	0.92
		MUNO	0.62	(0.60–0.64)	10.48	0.65	0.56	0.72	0.48
WC (cm) a		MHO	0.62	(0.58–0.66)	109.25	0.62	0.57	0.20	0.90
		MUNO	0.65	(0.63–0.67)	86.65	0.59	0.64	0.71	0.50

Diagnostic performance of anthropometric and body composition indicators for differentiating between MHO and MUNO. Classification efficacy was quantified by AUC

^a The AUC of this indicator was significantly different from that of SVR (P < 0.05)

Abbreviations: SVR skeletal muscle mass to visceral fat area ratio, BMI body mass index, MHO metabolically healthy obesity, MUNO metabolically unhealthy non-obesity, ASM appendicular skeletal muscle mass, VFA visceral fat area, WHtR waist-to-height ratio, WWI weight-adjusted waist index, WC waist circumference, 95% CI 95% confidence interval, AUC area under the curve, PPV positive predictive value, NPV negative predictive value

To further evaluate model performance, reclassification and calibration analyses were conducted. Incorporating SVR into the basic model improved risk discrimination for both MHO and MUNO, as indicated by positive NRI and IDI values (Supplementary Table S6). Calibration plots and Hosmer-Lemeshow tests suggested no evidence of lack of fit, with P-values > 0.05 for both models (Supplementary Figure S4).

Sensitivity analyses

To assess the robustness of the results, two sensitivity analyses were conducted. The first excluded participants with missing covariate data, and the second excluded those taking lipid-lowering medications. In both analyses, SVR remained positively associated with MHO and inversely associated with MUNO across all alternative datasets (Supplementary Tables S7 and S8). Details on lipid-lowering medication use among participants are provided in Supplementary Table S9.

Discussion

Using data from the 2011–2018 NHANES cycles, we examined associations of SVR with obesity phenotypes in 4,609 adults aged 20–59 years. Within the study population, 5.06% met criteria for MHO and 39.97% for MUNO. Among participants with obesity, higher SVR was associated with a greater likelihood of MHO, whereas among the non-obese participants, SVR was inversely associated with MUNO. RCS analyses suggested approximately linear associations of SVR with MHO and MUNO. Subgroup analyses indicated that the positive association of SVR with MHO was more evident among current or former drinkers, while the inverse association with MUNO appeared stronger in males and in participants aged 45–59 years.

The epidemiological characteristics of MHO and MUNO have been widely examined, yet reported prevalence rates vary considerably across studies. For instance, a multi-regional cross-sectional study in Russia involving 21,121 participants reported MHO and MUNO prevalences of 41.5% and 34.4%, respectively [57]. In contrast, a nationwide cohort study in South Korea, including 514,866 individuals, reported baseline prevalence of 10% for MHO and 32.2% for MUNO [58]. A smaller cross-sectional study in Hoveyzeh, Khuzestan Province, Iran (n = 7,464) reported MHO prevalence of 6.54% and MUNO prevalence of 33.06% [59]. These findings differ from our results, which indicated a lower prevalence of MHO and a higher prevalence of MUNO. Such discrepancies may be attributable to differences in sample size, inclusion criteria, definitions of metabolic health, and methods of obesity classification across studies. Notably, some previous studies may have overestimated the prevalence of MHO and underestimated that of MUNO. To address these issues, this study applied rigorous criteria to define metabolic status and classify obesity [39] and adjusted for multiple potential confounders. Findings from this analysis suggest associations between SVR and both MHO and MUNO. These results may offer insights into the interplay between body composition and metabolic function; they may also have implications for refining risk assessment in obesity-related conditions.

Although widely used to measure excess body weight, BMI provides limited insight into metabolic status. Previous studies have reported that some individuals classified as non-obese by BMI may still present with cardiometabolic disease, while others classified as obese may remain metabolically healthy [60]. These observations highlight the limitations of BMI in capturing obesity-related risks. Furthermore, obesity phenotypes are dynamic, and long-term monitoring combined with personalized management may help reduce the risk of developing metabolic diseases [32]. Therefore, further clinical investigations are warranted to clarify how SVR is associated with MHO and MUNO.

Most studies examining obesity and metabolic dysfunction have assessed skeletal muscle and visceral adipose tissue separately. A reduction in skeletal muscle mass combined with elevated visceral adiposity has been associated with obesity and related metabolic disturbances [16–20]. In addition, greater muscle mass has been associated with improved insulin sensitivity and metabolic function [61, 62]. A prior observational study reported that lower skeletal muscle mass was associated with increased risk of MetS [63]. A prospective study among Japanese Americans found that greater visceral fat was associated with higher risk of T2DM, with risk increasing in proportion to visceral fat volume [64]. These findings highlight the interplay between muscle mass and adipose distribution in metabolic health.

In recent years, growing attention has been given to the pathophysiological interactions between muscle and adiposity. SVR, which incorporates both ASM and VFA, has been suggested as a potential indicator of metabolic risk. Evidence from several studies suggests that SVR is associated with components of MetS, including WC, SBP, and LDL-C [33]. Lower SVR has also been associated with T2DM, MetS, MASLD, cardiometabolic disease, dyslipidemia, and insulin resistance [27, 30, 31, 33, 65]. These findings provide insights into associations between SVR and obesity phenotypes and suggest potential relevance for metabolic risk assessment.

However, the biological mechanisms linking SVR with MHO and MUNO remain complex and not fully clarified. Several hypotheses have been suggested. One hypothesis is that higher SVR may facilitate glucose uptake and limit fat accumulation, thereby improving insulin sensitivity. Because skeletal muscle is a major site of glucose uptake, greater muscle mass may increase energy expenditure and improve glucose metabolism [66, 67]. Another hypothesis is that lower SVR may be associated with altered secretion of muscle-derived cytokines (myokines) and greater visceral adiposity. Myokines play key roles in oxidative processes and immune regulation, and impaired function of these molecules has been linked to reduced insulin sensitivity and diminished fatty acid metabolism [68]. Excess visceral fat may release large amounts of circulating lipids and inflammatory mediators, which can interfere with glucose regulation and contribute to chronic low-grade inflammation [69, 70]. Visceral fat may also promote lipid accumulation in myocytes, potentially leading to mitochondrial dysfunction and impaired insulin signaling [71]. Finally, SVR may reflect the dynamic interplay between muscle mass and adipose tissue. Exercise-induced myokines may inhibit adipogenesis and regulate adipokine secretion [72], thereby supporting metabolic health. Conversely, visceral adipose tissue may release inflammatory mediators, such as interleukin-6 (IL-6) and tumor necrosis factor alpha (TNF-α), which can accelerate muscle protein breakdown, reduce lean mass, and contribute to metabolic dysfunction [73].

The observed associations between SVR and obesity phenotypes such as MHO and MUNO may partly reflect influences of the gut microbiota. Certain microbial communities have been suggested to support muscle anabolism, regulate lipid metabolism, and limit visceral fat deposition, potentially favoring higher SVR and better metabolic health. In contrast, dysbiosis has been associated with visceral adiposity, insulin resistance, and metabolically unhealthy states [74–77]. Recent studies suggest that gut microbiota may exert additional regulatory effects through interactions with muscular and adipose systems [78–80]. As a potential regulator of host metabolism and systemic energy balance, the gut microbiota has been associated with obesity progression, insulin resistance, and chronic low-grade inflammation [81–83]. Moreover, some evidence suggests that specific gut microbial profiles may indirectly influence body composition through effects on lipid metabolism and adipose distribution, which may in turn alter metabolic phenotypes [84–86]. These observations suggest that future studies integrating gut microbial profiles with body composition measures may help clarify the biological mechanisms related to SVR and obesity phenotypes.

Subgroup analyses indicated that individuals with current or past alcohol consumption showed a stronger association of SVR with MHO. This observation is consistent with earlier reports suggesting that moderate alcohol use may improve insulin sensitivity and favorably influence lipid metabolism, potentially reducing visceral adiposity and increasing SVR [87, 88]. Specifically, bioactive polyphenols in wine, such as resveratrol, have been suggested to exert antioxidant effects that may enhance insulin sensitivity and reduce adipose-tissue inflammation, thereby potentially mitigating visceral adiposity [89, 90]. Similarly, moderate beer consumption has not been consistently associated with increased risk of major chronic diseases and may be linked to some cardiovascular benefits [91]. Furthermore, alcohol’s influence on lipid metabolism, particularly its reported ability to increase plasma HDL-C, may contribute to metabolic homeostasis, given that low HDL-C is strongly associated with cardiometabolic risk [92]. However, the dose-response relationship and long-term effects of alcohol intake on SVR and overall metabolic health remain unclear and require further investigation.

Significant interactions between SVR and MUNO were observed when stratified by age and sex. When analyzed by age group, the inverse association between SVR and MUNO appeared stronger in individuals aged 45–59 years compared with those aged 20–44 years. This pattern may partly reflect the combined effects of age-related muscle atrophy and progressive visceral fat accumulation in middle age [93, 94]. Advancing age is often accompanied by reductions in skeletal muscle mass and strength. This decline has been attributed in part to reduced levels of anabolic hormones, such as androgens and estrogens, which may impair the efficiency of skeletal muscle protein synthesis. Concurrently, lifestyle factors such as sedentary behavior and suboptimal nutrition, along with a decline in metabolic rate, may further contribute to visceral adiposity. These factors may contribute to the stronger association observed between SVR and MUNO in individuals aged 45–59 years [95–97]. In contrast, among individuals aged 20–44 years, who generally have less muscle loss and slower fat accumulation, the association between SVR and MUNO appeared weaker.

Although SVR and MUNO were inversely associated in both sexes, the association appeared stronger in men than in women. This disparity may partly reflect sex-specific patterns of fat distribution and hormonal regulation. Men tend to accumulate more visceral adiposity than women at equivalent BMI levels [98]. Visceral fat, characterized by higher adipocyte activity and pro-inflammatory properties, has been associated with increased cytokine release, which may exacerbate insulin resistance and elevate metabolic risk. In older men, reduced skeletal muscle mass has also been associated with a higher prevalence of MetS [63]. In contrast, premenopausal women may experience estrogen-related protection of insulin sensitivity [99, 100]. Elevated estrogen levels in this population have been reported to promote subcutaneous rather than visceral fat storage and to suppress pro-inflammatory cytokine release from adipocytes, thereby supporting metabolic health. Collectively, these mechanisms may help attenuate the adverse metabolic consequences associated with a low SVR. Therefore, SVR may have potential utility in identifying MUNO among middle-aged adults, particularly men, although further research is required to clarify its relevance.

Building on these findings, further research is needed to clarify the potential clinical relevance of SVR in the context of metabolic health. As a non-invasive indicator incorporating both skeletal muscle mass and visceral fat, SVR may provide supplementary information beyond BMI and other anthropometric indices in characterizing obesity phenotypes. Although its discriminatory performance was modest, SVR could still contribute to a more refined understanding of metabolic heterogeneity. Future longitudinal studies should examine whether changes in SVR over time are associated with changes in metabolic risk, which may help clarify its potential value in risk monitoring. However, additional validation is required before any clinical application can be considered.

Strengths and limitations

This study has several strengths. First, data on the associations between SVR, MHO, and MUNO in U.S. populations are scarce, and this study offers preliminary evidence that may help address this gap. Second, use of the NHANES database, which includes a demographically diverse U.S. population, enabled evaluation of associations in a well-characterized and heterogeneous sample. Finally, systematic adjustment for a wide range of demographic, clinical, and lifestyle factors, together with serum biomarkers, provides additional confidence in the observed associations.

Several limitations should be acknowledged. First, previous studies have described the prevalence of sarcopenia and metabolic abnormalities in older adults [101–104], and recent epidemiological data suggest a rising incidence of MetS and sarcopenic obesity among pediatric and adolescent populations [105–107]. However, the present study was restricted to adults aged 20–59 years; therefore, the results cannot be directly extrapolated to younger or older age groups. In addition, the exclusion of participants with missing body composition or metabolic data reduced the analytic sample, which may further limit the generalizability of the findings. Future studies should incorporate broader age ranges, larger sample sizes, and more recent data to validate these associations. Second, although comprehensive adjustments were made for demographic, clinical, and lifestyle factors, residual confounding from unmeasured or unidentified variables cannot be ruled out. Third, the cross-sectional design precludes causal inference, making it difficult to evaluate temporal dynamics or bidirectional associations. Finally, while SVR showed modest discriminatory ability (AUC = 0.65 for MHO and 0.71 for MUNO), these values suggest that SVR should be regarded as a supplementary indicator rather than a definitive diagnostic tool. This limitation highlights the need for cautious interpretation and further validation before any clinical application is considered.

Conclusion

This cross-sectional study found that SVR was positively associated with MHO and negatively associated with MUNO. These findings suggest that SVR may provide supplementary information for characterizing obesity phenotypes. However, given its modest discriminatory performance, further prospective and interventional studies are required to clarify the clinical relevance of SVR and to assess whether improvements in SVR are associated with reduced metabolic risk.

Abbreviations

ASM

Appendicular Skeletal Muscle Mass

AUC

Area Under the Curve

BMI

Body Mass Index

CI

Confidence Interval

CKD

Chronic Kidney Disease

CVD

Cardiovascular Disease

DBP

Diastolic Blood Pressure

DXA

Dual-energy X-ray Absorptiometry

eGFR

Estimated Glomerular Filtration Rate

FPG

Fasting Plasma Glucose

HDL-C

High-density Lipoprotein Cholesterol

IDI

Integrated Discrimination Index

IL-6

Interleukin-6

LDL-C

Low-density Lipoprotein Cholesterol

MASLD

Metabolic dysfunction-associated Steatotic Liver Disease

MET

Metabolic Equivalents

MetS

Metabolic Syndrome

MHNO

Metabolically Healthy Non-obesity

MHO

Metabolically Healthy Obesity

MUNO

Metabolically Unhealthy Non-obesity

MUO

Metabolically Unhealthy Obesity

NCHS

National Center for Health Statistics

NHANES

National Health and Nutrition Examination Survey

NPV

Negative Predictive Value

NRI

Net Reclassification Index

OR

Odds Ratio

PA

Physical Activity

PIR

Poverty Income Ratio

PPV

Positive Predictive Value

RCS

Restricted Cubic Spline

ROC

Receiver Operating Characteristic

SBP

Systolic Blood Pressure

SD

Standard Deviation

SE

Standard Errors

SVR

Skeletal muscle mass to visceral fat area ratio

T2DM

Type 2 diabetes mellitus

TC

Total Cholesterol

TG

Triglycerides

TNF-α

Tumor Necrosis Factor Alpha

VFA

Visceral Fat Area

VIF

Variance Inflation Factor

WC

Waist Circumference

WHO

World Health Organization

WHtR

Waist-to-Height Ratio

WWI

Weight-adjusted Waist Index

Authors’ contributions

R.W.: Performed statistical analysis, interpretation, and manuscript drafting.L.C., P.Z., and H.C.: Participated in data collection. C.W., Q.Z., and Z.S.: Provided methodological validation and analytical guidance. Z.H.: Participated in study conception and design. C.W. and Z.H.: Provided critical content revision to improve quality assurance. The final manuscript has been thoroughly reviewed and approved by all authors.

Funding

There was no funding support for our study.

Data availability

This study conducted analyses of publicly accessible datasets, with all data obtained from NHANES, https://wwwn.cdc.gov/nchs/nhanes/.

Declarations

Ethics approval and consent to participate

The NCHS institutional review board provided ethical approval. Participants were fully informed and signed consent forms, and research procedures adhered strictly to the Declaration of Helsinki standards.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.