Determinants of Metabolic Syndrome Among Rural Older Adults: A Cross-Sectional Analysis of the 2023 Korea National Health and Nutrition Examination Survey

Abstract

Metabolic syndrome (MetS) is common in later life and shaped by modifiable lifestyle and clinical factors, yet data specific to rural older adults are limited. This cross-sectional study analyzed rural Koreans aged ≥65 years (unweighted n = 467) from the 2023 Korea National Health and Nutrition Examination Survey, incorporating the complex survey design (strata, clusters, and weights). MetS was defined using National Cholesterol Education Program Adult Treatment Panel III criteria with Asian-specific waist cutoffs (≥3 of 5 components). Sociodemographic, behavioral, and clinical characteristics were compared by MetS status using design-based tests, and complex-sample logistic regression estimated adjusted odds ratios (aORs) with 95% confidence intervals (CIs). The survey-weighted prevalence of MetS was 42.8%. Compared with those without MetS, participants with MetS had higher body mass index (BMI) and waist circumference, more hypertension and diabetes, higher triglycerides, and lower high-density lipoprotein cholesterol; low-density lipoprotein cholesterol did not differ meaningfully. In multivariable models, BMI ≥25 kg/m² (aOR 9.08; 95% CI 6.01–13.71, p ≤ 0.001), hemoglobin A1c ≥ 7.0% (aOR 4.42; 95% CI 1.75–11.16, p = 0.003), and vitamin D deficiency <20 ng/mL (aOR 2.32; 95% CI 1.23–4.35, p = 0.012) were independently associated with higher odds of MetS, whereas meeting the World Health Organization physical activity guideline was inversely associated (aOR 0.50; 95% CI 0.26–0.96, p = 0.039). These findings highlight adiposity, suboptimal glycemic control, and vitamin D deficiency as key, potentially modifiable correlates of MetS in rural older adults and support promotion of guideline-level physical activity as part of integrated cardiometabolic risk management in rural settings.

1. Introduction

Metabolic syndrome (MetS) is a cluster of cardiometabolic abnormalities—central adiposity, atherogenic dyslipidemia, elevated blood pressure, and dysglycemia—linked to cardiovascular disease, type 2 diabetes, and premature mortality [1]. Standard definitions derive from the National Cholesterol Education Program Adult Treatment Panel III (NCEP ATP III) [2], with Asian-specific waist cutoffs to better reflect risk distribution in Koreans [3]. In later life, age-related changes in body composition and insulin signaling heighten vulnerability to MetS and its complications [4].

Globally, MetS is common in older adults and contributes to geriatric morbidity, functional decline, and health-care use [1]. In Korea, population-based data indicate a substantial burden among older adults with implications for primary care and community services [5]. Rural residence may further shape risk via socioeconomic context, behaviors, built environment, and access to prevention and continuity of care [6,7]. Yet determinants of MetS in rural older adults remain less well characterized; prior studies vary in measurement and adjustment, limiting comparability [8]. In particular, MetS prevalence estimates can differ substantially depending on the diagnostic definition and related analytic choices, which complicates direct comparisons across studies and populations [8].

Evidence points to modifiable factors with plausible biology. Adiposity, especially central fat, promotes insulin resistance and low-grade inflammation; in Asian populations, risk rises at lower BMI, supporting Asia–Pacific thresholds (e.g., ≥25 kg/m²) [9,10]. Glycemic control indexed by hemoglobin A1c (HbA1c) relates to cardiometabolic risk, and clinical standards emphasize its role for risk management in diabetes care [11]. Vitamin D (25-hydroxyvitamin D [25(OH)D]) may influence insulin secretion/sensitivity and inflammation; lower 25(OH)D is frequently linked to MetS components, though causality is uncertain [12]. Physical activity favorably modifies adiposity, glucose regulation, and lipid profiles; the World Health Organization recommends ≥150 min/week of moderate-intensity activity for older adults [13]. Rural contexts can complicate these relationships: transportation barriers, fewer facilities, limited specialty care, and seasonal/occupational patterns affect activity, diet, and chronic-disease management [7].

Methodologically, analyses of national surveys should incorporate complex design features (stratification, clustering, and sampling weights) to ensure valid inference, and KNHANES provides standardized national data suitable for design-based analysis [14]. Although several studies have examined metabolic syndrome in Korean adults and older populations, rural-focused evidence remains limited and has been difficult to compare across studies due to heterogeneity in MetS definitions, inconsistent operationalization of key modifiable exposures, and incomplete application of design-based survey methods in rural subgroup analyses [8,14]. Therefore, our study extends prior work by providing a rural, subgroup-specific, survey-weighted analysis using the most recent KNHANES 2023 data restricted to adults aged ≥65 years and applying harmonized definitions and prespecified covariate adjustment to identify actionable correlates of MetS in rural settings.

The study aim was to examine sociodemographic, behavioral, and clinical correlates of MetS among rural adults aged ≥65 years using the 2023 Korea National Health and Nutrition Examination Survey within a complex-sample analytical framework; we hypothesized that (1) adiposity (BMI ≥ 25 kg/m²), (2) suboptimal glycemic control (HbA1c ≥ 7.0%), and (3) vitamin D deficiency (<20 ng/mL) would be positively associated with MetS, whereas (4) meeting the WHO physical activity guideline would be inversely associated after adjustment for confounders [9,11,12,13].

2. Materials and Methods

2.1. Study Design and Data Source

This cross-sectional study was reported in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guideline [15]. This study used data from the 2023 Korea National Health and Nutrition Examination Survey (KNHANES), released in December 2024. KNHANES is an annual, nationally representative survey of the non-institutionalized Korean population conducted by the Korea Disease Control and Prevention Agency (KDCA) [14]. It uses a complex, multistage, stratified cluster sampling design with survey weights, and data are collected through standardized health interviews, physical examinations, and nutrition assessments administered by trained personnel under rigorous quality-control procedures [14].

2.2. Study Population

In 2023, 6929 individuals participated in KNHANES. Of these, 1836 were aged ≥65 years. Rural residence was identified using Korean administrative classifications (eup or myeon). Among older adults, 527 resided in rural areas. Rural older adults aged ≥65 years with complete data on all variables required to define metabolic syndrome and prespecified covariates were included, resulting in a final analytic sample of 467 participants (unweighted). Thus, 60 participants (11.4%) were excluded due to missing values under complete-case analysis.

2.3. Variables and Definitions

MetS was defined according to the National Cholesterol Education Program Adult Treatment Panel III (NCEP ATP III), adapted for Asian populations [2]. Participants were classified as having MetS if they met ≥3 of the following: (1) abdominal obesity (waist circumference ≥90 cm in men or ≥85 cm in women) [3], (2) hypertriglyceridemia (triglycerides ≥150 mg/dL), (3) reduced HDL cholesterol (<40 mg/dL in men or <50 mg/dL in women), (4) elevated blood pressure (systolic ≥130 mmHg or diastolic ≥85 mmHg, or antihypertensive use), and (5) elevated fasting glucose (≥100 mg/dL or antidiabetic medication) [2].

Serum vitamin D status was assessed using serum 25-hydroxyvitamin D [25(OH)D]; concentrations ≥20 ng/mL were considered sufficient and <20 ng/mL classified as deficient [16].

Physical activity was determined by self-reported adherence to the World Health Organization guideline of at least 150 min per week of moderate-intensity aerobic activity; participants meeting this criterion were considered physically active per WHO standards [13].

Other covariates included sex (male/female), age group (<72 or ≥72 years), education (middle school graduate or higher vs. less than middle school), income (upper 50% vs. lower 50%), current smoking (yes/no), alcohol consumption in the past year (yes/no), sleep duration (≥7 vs. <7 h/day), and body mass index (BMI), categorized as <25 or ≥25 kg/m² in accordance with Asia–Pacific criteria [9].

2.4. Statistical Analysis

All analyses were conducted using IBM SPSS Statistics (version 30.0). To account for the complex survey design of KNHANES, sampling weights, strata, and primary sampling units were incorporated in all descriptive and inferential analyses. Continuous variables were summarized using weighted means and standard errors, and their distributions were assessed using descriptive diagnostics (e.g., weighted histograms and skewness/kurtosis) to identify potential extreme deviations. Group differences were examined using the design-based Wald F test for continuous variables and the Rao–Scott chi-square test for categorical variables. Factors associated with metabolic syndrome were evaluated using complex-sample multivariable logistic regression, and multicollinearity was assessed using variance inflation factors. Sex-stratified analyses showed similar patterns in men and women, and no statistically significant sex-by-exposure interactions were observed.

2.5. Ethical Considerations

This study conducted a secondary analysis of publicly available, de-identified data from the 2023 KNHANES. The KNHANES protocol is reviewed and approved annually by an institutional review board, and all participants provided written informed consent at the time of the original survey. For this secondary analysis, the study protocol was reviewed by the Daejin University Institutional Review Board and was determined to be exempt (Approval No. 1040656-202512-HR-02-15; approved on 2 December 2025). Because only de-identified public-use data were analyzed, no additional informed consent was required.

3. Results

3.1. Sociodemographic and Lifestyle Characteristics by Metabolic Syndrome Status

Under the complex-sample design, sociodemographic and lifestyle profiles were broadly comparable between participants with and without MetS. Mean age did not differ by MetS status (72.62 vs. 72.52 years; p = 0.866), and the proportion of men was nearly identical (49.7% vs. 49.0%; p = 0.893). Educational attainment (≥middle school: 38.8% vs. 42.8%; p = 0.257) and household income (≥lower-middle: 45.5% vs. 46.8%; p = 0.795) were similarly distributed. Health behaviors were also comparable, including current smoking (11.3% vs. 11.8%; p = 0.871) and alcohol use in the past year (both 50.5%; p = 0.995). By contrast, adherence to the World Health Organization (WHO) physical activity guideline was lower among those with MetS (10.9% vs. 17.8%; p = 0.032), whereas sleeping ≥7 h did not differ (48.6% vs. 43.6%; p = 0.370). All estimates are survey-weighted and tested with design-based procedures (

Table 1. Sociodemographic and lifestyle characteristics by metabolic syndrome status (unweighted n = 467).

Variable	No MetS M or % (SE)	MetS M or % (SE)	p
Age, years, M	72.52 (0.34)	72.62 (0.45)	0.866
Male (yes), %	49.0 (2.8)	49.7 (2.7)	0.893
Education ≥ middle school (yes), %	42.8 (2.1)	38.8 (3.5)	0.257
Household income ≥ lower-middle (yes), %	46.8 (3.5)	45.5 (4.2)	0.795
Current smoker (yes), %	11.8 (1.9)	11.3 (2.5)	0.871
Alcohol use in past year (yes), %	50.5 (3.6)	50.5 (4.5)	0.995
Meets WHO physical activity guideline (yes), %	17.8 (2.6)	10.9 (2.3)	0.032
Sleep duration ≥7 h/night(yes), %	43.6 (3.1)	48.6 (3.9)	0.370

Values are survey-weighted. Continuous variables are presented as mean (SE) with p-values from complex-sample GLM (Wald F); categorical variables are presented as % (SE) with p-values from Rao–Scott adjusted χ² (reported as modified F). The complex-sample design (strata, PSUs, and weights) was applied. Percentages may not sum to 100% due to rounding. Abbreviations: MetS, metabolic syndrome; M, mean; SE, standard error; WHO, World Health Organization; PA, physical activity; PSU, primary sampling unit.

).

3.2. Clinical and Anthropometric Measures by Metabolic Syndrome Status

Participants with MetS exhibited more adverse anthropometric and clinical profiles than those without MetS. Mean BMI was higher in the MetS group (25.77 vs. 22.78 kg/m²; p < 0.001), and general obesity (BMI ≥25 kg/m²) was more prevalent (61.7% vs. 18.2%; p < 0.001). Central adiposity was pronounced, with larger mean waist circumference (91.45 vs. 82.60 cm; p < 0.001) and a higher prevalence of sex-specific abdominal obesity (80.9% vs. 24.1%; p < 0.001). Systolic blood pressure was modestly higher in the MetS group (130.80 vs. 126.16 mmHg; p = 0.001), while diastolic pressure did not differ (74.16 vs. 73.56 mmHg; p = 0.487). Hypertension prevalence was substantially higher among participants with MetS (90.9% vs. 59.2%; p < 0.001). Diabetes was also more common (76.0% vs. 38.9%; p < 0.001) (

Table 2. Clinical and anthropometric measures by metabolic syndrome status (unweighted n = 467).

Variable	No MetS M or % (SE)	MetS M or % (SE)	p
BMI, kg/m2, M	22.784 (0.138)	25.771 (0.205)	<0.001
BMI ≥ 25 kg/m2 (yes), %	18.2 (1.7)	61.7 (3.2)	<0.001
Waist circumference, cm, M	82.603 (0.459)	91.445 (0.555)	<0.001
Abdominal obesity (sex-specific cut-offs; yes), %	24.1 (3.1)	80.9 (2.8)	<0.001
SBP, mmHg, M	126.16 (1.00)	130.80 (1.11)	0.001
DBP, mmHg, M	73.56 (0.53)	74.16 (0.85)	0.487
Hypertension (yes), %	59.2 (4.0)	90.9 (1.9)	<0.001
Diabetes (yes), %	38.9 (3.0)	76.0 (3.4)	<0.001

Survey-weighted estimates. Continuous variables are presented as mean (SE) with p-values from complex-sample GLM (Wald F); categorical variables are presented as % (SE) with p-values from Rao–Scott adjusted χ² (reported as modified F). The complex-sample design (strata, PSUs, and weights) was applied. Percentages may not sum to 100% due to rounding. Abbreviations: MetS, metabolic syndrome; M, mean; SE, standard error; BMI, body mass index; SBP/DBP, systolic/diastolic blood pressure; PSU, primary sampling unit.

).

3.3. Laboratory Biomarkers by Metabolic Syndrome Status

Laboratory profiles were consistently more adverse among participants with MetS. Mean HbA1c and fasting plasma glucose were higher (HbA1c: 6.136% vs. 5.725%, p < 0.001; FPG: 113.38 vs. 98.81 mg/dL, p < 0.001), with greater proportions above clinical thresholds (HbA1c ≥6.5%: 24.3% vs. 8.3%, p < 0.001; HbA1c ≥7.0%: 12.5% vs. 4.1%, p = 0.005; FPG ≥100 mg/dL: 71.0% vs. 35.9%, p < 0.001). Triglycerides were higher (158.26 vs. 98.51 mg/dL; p < 0.001) with more hypertriglyceridemia (≥150 mg/dL: 47.0% vs. 6.3%; p < 0.001), whereas HDL-C was lower (45.57 vs. 56.75 mg/dL; p < 0.001) with more low HDL-C (sex-specific; 57.7% vs. 9.0%; p < 0.001). LDL-C was slightly lower in MetS and not different statistically (98.04 vs. 105.24 mg/dL; p = 0.053). Vitamin D status was less favorable: mean 25(OH)D was lower (25.20 vs. 28.33 ng/mL; p = 0.001), and deficiency (<20 ng/mL) was more prevalent (33.6% vs. 19.8%; p = 0.017) (

Table 3. Laboratory biomarkers by metabolic syndrome status (unweighted n = 467).

Variable	No MetS M or % (SE)	MetS M or % (SE)	p
HbA1c, %, M (SE)	5.725 (0.048)	6.136 (0.067)	<0.001
HbA1c ≥ 6.5% (yes), %	8.3 (1.6)	24.3 (2.8)	<0.001
HbA1c ≥ 7.0% (yes), %	4.1 (1.2)	12.5 (2.2)	0.005
Fasting plasma glucose, mg/dL, M (SE)	98.81 (1.12)	113.38 (1.78)	<0.001
FPG ≥ 100 mg/dL (yes), %	35.9 (3.3)	71.0 (4.1)	<0.001
Triglycerides, mg/dL, M (SE)	98.51 (3.10)	158.26 (5.09)	<0.001
TG ≥ 150 mg/dL (yes), %	6.3 (1.5)	47.0 (3.3)	<0.001
HDL-C, mg/dL, M (SE)	56.75 (0.49)	45.57 (1.10)	<0.001
Low HDL-C (sex-specific; yes), %	9.0 (1.7)	57.7 (3.9)	<0.001
LDL-C, mg/dL, M (SE)	105.24 (2.35)	98.04 (2.19)	0.053
25(OH)D, ng/mL, M (SE)	28.33 (0.71)	25.20 (0.73)	0.001
Vitamin D deficiency (<20 ng/mL; yes), %	19.8 (2.5)	33.6 (5.0)	0.017

Survey-weighted estimates. Continuous variables are presented as mean (SE) with p-values from complex-sample GLM (Wald F); categorical variables are presented as % (SE) with p-values from Rao–Scott adjusted χ² (reported as modified F). The complex-sample design (strata, PSUs, and weights) was applied. Percentages may not sum to 100% due to rounding. Abbreviations: MetS, metabolic syndrome; M, mean; SE, standard error; FPG, fasting plasma glucose; TG, triglycerides; HDL-C/LDL-C, high-/low-density lipoprotein cholesterol; 25(OH)D, 25-hydroxyvitamin D; PSU, primary sampling unit.

).

3.4. Multivariable Predictors of Metabolic Syndrome (Complex-Sample Logistic Regression)

In the fully adjusted complex-sample model, four predictors were independently associated with MetS: BMI ≥25 kg/m² (aOR = 9.08; 95% CI 6.01–13.71; p < 0.001), HbA1c ≥7.0% (aOR = 4.42; 95% CI 1.75–11.16; p = 0.003), vitamin D deficiency <20 ng/mL (aOR = 2.32; 95% CI 1.23–4.35; p = 0.012), and meeting the WHO physical activity guideline (inverse association: aOR = 0.50; 95% CI 0.26–0.96; p = 0.039). Other covariates—sex, age (≥72 years), education, household income, current smoking, alcohol use in the past year, sleep ≥7 h, and LDL-C ≥130 mg/dL—were not significant after adjustment (all p > 0.05). Overall model fit under the survey design was acceptable (modified F(12, 10) = 9.954; p < 0.001; design df = 21), with pseudo-R² values of 0.244 (Cox–Snell), 0.327 (Nagelkerke), and 0.205 (McFadden). CSLOGISTIC provides design-based Wald F tests but not Hosmer–Lemeshow or classification indices. Sex-stratified analyses showed similar patterns in men and women, and no statistically significant sex-by-exposure interactions were observed (

Table 4. Predictors of metabolic syndrome: complex-sample logistic regression (unweighted n = 467).

Predictor (Comparison)	aOR	95% CI	p
Female (vs. Male)	1.408	0.956–2.075	0.080
Age ≥ 72 y (vs. <72 y)	0.898	0.427–1.887	0.766
Education ≥ middle school (vs. Elementary)	1.374	0.780–2.421	0.256
Household income ≥ lower-middle (vs. low)	0.876	0.486–1.580	0.646
Current smoker (vs. non-smoker)	0.930	0.412–2.101	0.855
Alcohol use in past year (vs. none)	1.241	0.721–2.132	0.418
Meets WHO PA guideline (vs. not meets)	0.498	0.258–0.962	0.039
Sleep ≥ 7 h (vs. <7)	1.036	0.562–1.658	0.894
HbA1c ≥ 7.0% (vs. <7.0)	4.422	1.751–11.163	0.003
BMI ≥ 25 kg/m2 (vs. <25)	9.081	6.014–13.712	<0.001
25(OH)D < 20 ng/mL (vs. ≥20)	2.315	1.229–4.348	0.012
LDL-C ≥ 130 mg/dL (vs. <130)	0.883	0.513–1.517	0.636

Values are survey-weighted and estimated using complex-sample multivariable logistic regression. Abbreviations: MetS, metabolic syndrome; aOR, adjusted odds ratio; CI, confidence interval; HbA1c, glycated hemoglobin; BMI, body mass index; LDL-C, low-density lipoprotein cholesterol; WHO, World Health Organization; PA, physical activity; PSU, primary sampling unit.

).

4. Discussion

4.1. Principal Findings in Context

In this design-based national analysis, rural older adults had a high MetS burden (42.8%) with clustering of central adiposity, hypertriglyceridemia, low HDL-C, and impaired glycemia. After multivariable adjustment, BMI ≥ 25 kg/m², HbA1c ≥ 7.0%, and 25(OH)D <20 ng/mL were associated with higher odds of MetS, whereas meeting the World Health Organization (WHO) physical activity guideline was inversely associated. LDL-C did not differ significantly—a plausible result given its exclusion from MetS criteria and typically weaker cross-sectional gradients than triglycerides or HDL-C. Collectively, these findings prioritize abdominal adiposity, dysglycemia, and atherogenic dyslipidemia as prevention targets in rural elders, while LDL-centric strategies are less informative for MetS risk stratification.

4.2. Modifiable Pathways and Biologic Plausibility

Greater adiposity—particularly central adiposity—remained the dominant correlate of MetS, aligning with evidence that excess visceral fat contributes to insulin resistance, chronic low-grade inflammation, and an atherogenic lipid profile [10]. In Asian populations, cardiometabolic risk can accrue at lower BMI than in Western cohorts, supporting Asia–Pacific thresholds (e.g., ≥25 kg/m²) for obesity-related risk stratification and justifying our modeling approach [9].

Suboptimal glycemic control, indexed by elevated HbA1c, reflects both fasting and postprandial dysglycemia. Age-related changes in body composition and metabolic regulation in later life may further impair glucose disposal and insulin secretion, reinforcing the observed association between higher HbA1c and MetS [4,11]. Importantly, while HbA1c ≥7.0% is clinically meaningful as an indicator of poorer glycemic control, it is not itself a diagnostic component of MetS; thus, it should be interpreted as a marker of metabolic dysregulation that co-occurs with MetS-related physiology rather than as a criterion for case definition.

Vitamin D deficiency was also independently related to MetS, consistent with meta-analytic evidence linking low 25-hydroxyvitamin D [25(OH)D] to MetS and its components and with mechanistic reviews suggesting roles in insulin sensitivity and inflammatory signaling [12,16]. While causality cannot be inferred from cross-sectional data, identifying vitamin D deficiency in high-risk older adults and correcting deficiency in accordance with guideline-based thresholds and safety considerations is clinically reasonable [17].

Finally, meeting the WHO physical activity guideline was inversely associated with MetS in our fully adjusted model. This is consistent with evidence from systematic reviews and meta-analyses of clinical trials showing that exercise improves multiple MetS indicators, including lipids and glycemic measures, even when weight loss is modest [18]. Because these pathways are interconnected (e.g., activity reduces central adiposity and improves insulin sensitivity, which may lower HbA1c), multicomponent strategies that address activity, weight management, and metabolic monitoring concurrently are likely to yield the greatest benefit.

4.3. Rural Context and Translation to Practice/Policy

Rural living conditions can shape cardiometabolic risk through structural barriers, including transportation constraints, fewer opportunities for preventive counseling, limited access to exercise facilities, and seasonal or occupational patterns that influence habitual activity and outdoor exposure. In addition, local food environments and market accessibility may affect diet quality and the feasibility of sustained weight management [19].

Translation to practice may therefore benefit from community-anchored, low-cost, geriatric-friendly approaches: waist- and BMI-informed risk screening using Asia–Pacific thresholds; brief but repeated lifestyle counseling; individualized glycemic management emphasizing safe targets and avoidance of hypoglycemia; assessment and correction of vitamin D deficiency when clinically indicated; and WHO-aligned activity prescriptions (e.g., group walking, chair-based aerobic activity, and resistance training) delivered through senior centers, local community venues, or primary-care extenders, supported by simple self-monitoring tools (e.g., step goals or pedometers) [20].

At the policy level, investments that increase access to safe, age-friendly walking routes and indoor exercise options during inclement seasons, strengthen transportation support, and improve food access (e.g., mobile markets or local cooperative partnerships) may shift risk distributions at scale. Programs embedded within existing rural public-health infrastructure and delivered through nurses or community health workers using brief, repeated contacts may enhance uptake and equity, particularly among the oldest-old and individuals with mobility limitations.

4.4. Strengths, Limitations, and Future Directions

Strengths include standardized national measurements, harmonized definitions, and survey-weighted modeling that supports valid inference for rural older adults. Several limitations warrant emphasis. First, the cross-sectional design precludes causal inference and reverse causality cannot be ruled out; for example, individuals with MetS may be less physically active or have reduced outdoor exposure due to comorbidity burden or functional limitations, which could also influence vitamin D status. Second, our complete-case analytic approach may introduce selection bias if excluded participants differed systematically in socioeconomic vulnerability or health status; reporting the proportion excluded due to missing data can help contextualize this possibility. Third, some predictors are conceptually related to MetS pathways (e.g., BMI and HbA1c). Although BMI is not a diagnostic criterion for MetS and HbA1c was included as an indicator of glycemic control rather than as a component of the MetS definition, partial overlap with adiposity- and glycemia-related mechanisms may contribute to effect sizes. BMI was retained in the model due to its clinical and public health relevance as a pragmatic indicator of overall adiposity that complements waist circumference–based central adiposity for risk stratification. Future work should test robustness using sensitivity analyses (e.g., alternative HbA1c cut-points such as ≥6.5%, modeling BMI continuously, or evaluating models that vary adiposity-related proxies) and prospective designs to strengthen causal inference. Future studies should assess the robustness of our findings using alternative HbA1c thresholds (e.g., ≥6.5%) in sensitivity analyses.

Looking forward, repeated KNHANES waves and linked administrative data could clarify temporal ordering among vitamin D status, adiposity trajectories, activity patterns, and incident MetS. Pragmatic, community-deliverable trials combining activity promotion, weight-management counseling, and targeted vitamin D strategies—implemented through primary care and community venues—are warranted to evaluate scalable models suitable for rural contexts. Incorporating social determinants (food access, income volatility, caregiving burden) and environmental exposures (seasonality, built environment) may further illuminate upstream drivers of heterogeneity and sharpen precision targeting.

5. Conclusions

In this design-based national analysis of rural older adults, MetS was common. Adiposity (BMI ≥25 kg/m²), suboptimal glycemic control (HbA1c ≥7.0%), and vitamin D deficiency (<20 ng/mL) were independently associated with higher odds of MetS, whereas meeting the World Health Organization physical activity guideline was protective; LDL-C did not differ meaningfully. Our harmonized definitions and complex-sample methods strengthen internal validity and comparability for this rural subgroup. These findings prioritize central weight management, individualized glycemic optimization, vitamin D assessment and correction when indicated, and promotion of WHO-level activity through community-deliverable programs. At the systems level, improving walkability, access to age-friendly exercise spaces, transportation, and healthy food options may reduce structural barriers and enhance equity in rural settings. Given the cross-sectional design, longitudinal studies and pragmatic trials are needed to test scalable, integrated strategies that prevent or attenuate MetS in rural older adults.