Associations Between Metabolic Syndrome Index, Systemic Inflammatory Index, and Arthritis Subtype in the U.S. and Chinese Adults: A Mediation Analysis

1. Introduction

Arthritis constitutes a major source of chronic pain and disability globally, placing an escalating burden on healthcare systems worldwide [1, 2]. By 2030, an estimated 67 million adults in the United States alone are projected to receive a physician diagnosis of arthritis, representing nearly one‐quarter of the adult population [3]. Although traditionally classified into distinct etiopathogenic categories—such as mechanical degeneration, autoimmune dysregulation, and crystal deposition—growing evidence highlights metabolic dysregulation as a common pathogenic thread across various forms of arthritis [4, 5]. Notably, insulin resistance (IR), a core component of metabolic syndrome, has been implicated in promoting chronic low‐grade inflammation and disrupting articular homeostasis, suggesting its potential as a shared risk factor among different arthritis subtypes [6, 7]. Nevertheless, whether metabolic dysfunction contributes uniformly or differentially across arthritis types remains to be systematically elucidated. Furthermore, research has also explored natural therapeutic options for arthritis management. Plant‐derived compounds (Filipendula ulmaria) [8], combinatorial nutritional formulation [9], and flavonoids such as luteolin [10] have each shown anti‐inflammatory effects relevant to arthritic conditions. These findings complement the mechanistic perspective by highlighting inflammation as a shared therapeutic target.

The triglyceride‐glucose (TyG) index, calculated from fasting triglycerides and glucose, is widely recognized as a reliable and practical surrogate marker of IR [11]. More recently, the TyG–body mass index (TyG–BMI)—a product of the TyG index and BMI—has been proposed as an integrated metric that combines lipid‐glucose metabolism with overall adiposity, potentially offering superior predictive capacity for cardiometabolic diseases [12]. Given the close interplay between adipose tissue dysfunction, systemic inflammation, and IR, these indices may hold particular relevance in the pathogenesis of arthritis. Concurrently, chronic low‐grade inflammation serves as a hallmark of both metabolic dysregulation and many arthritic conditions [5, 13, 14]. The systemic immune‐inflammation index (SII), which incorporates neutrophil, lymphocyte, and platelet counts, has emerged as a comprehensive biomarker reflecting systemic inflammatory and immune status [15]. However, the extent to which SII explains the link between metabolic dysfunction and arthritis risk has not been formally evaluated.

To address these gaps, this study leverages two independent large‐scale cohorts: a nationally representative sample from the U.S. National Health and Nutrition Examination Survey (NHANES) and a hospital‐based cohort from China. We hypothesized that both the TyG index and TyG‑BMI are dose‑dependently associated with the risk of multiple arthritis subtypes, that systemic inflammation (SII) partially mediates these associations, and that the predictive performance of these indices may differ between the U.S. and Chinese populations. We aimed to elucidate the metabolic‑inflammatory pathways linking these indices to arthritis and to evaluate their potential as practical screening tools for risk stratification in diverse clinical settings through comprehensive analyses.

2. Methods

3. Result

4. Discussion

This study comprehensively analyzed arthritis subtypes across two ethnically and clinically distinct populations. By integrating novel metabolic indices with an established inflammatory marker, our findings provide multi‐dimensional evidence that metabolic dysfunction, particularly IR and adiposity, is robustly associated with arthritis risk and that this relationship is partially mediated by systemic inflammation.

This study found that higher TyG and TyG–BMI indices were significantly associated with greater odds of various arthritis types, thereby extending and strengthening the accumulating evidence connecting metabolic dysregulation with musculoskeletal disorders [20, 21]. The TyG index, a well‐established surrogate marker of IR, has been linked to the risk of OA, potentially mediated by mechanisms such as chronic low‐grade inflammation, oxidative stress, and dysregulated chondrocyte metabolism [20, 22, 23]. The even stronger associations observed for TyG–BMI, particularly in the U.S. cohort, suggested that the confluence of IR and increased adiposity exerts a synergistic detrimental effect. Adipose tissue is an active endocrine organ secreting pro‐inflammatory adipokines (e.g., leptin and resistin) and cytokines that could perpetuate joint inflammation and degradation, providing a plausible biological pathway for our observations [24]. The consistency of these associations across two ethnically and geographically distinct cohorts strengthens the argument for a fundamental metabolic component in arthritis etiology beyond traditional risk factors. Notably, the ORs for TyG–BMI in the Chinese cohort, particularly for RA (OR = 4.77) and gout (OR = 5.46), were substantially higher than those for the TyG index. This numerical disparity is partly attributable to the multiplicative structure of the TyG–BMI, which yields a broader numerical scale than TyG alone. To exclude the possibility that these high ORs arose from multicollinearity, we assessed variance inflation factors (VIFs) in the regression models; all VIFs were below 2, confirming that collinearity did not unduly inflate the estimates. Therefore, the observed stronger associations for TyG–BMI more plausibly reflect a synergistic effect of combined IR and adiposity rather than a statistical artifact.

Our mediation analysis offers mechanistic insight by identifying SII—a composite marker reflecting systemic immune‐inflammatory balance—as a significant partial mediator. This was especially consistent for the TyG–BMI index in both cohorts. IR and visceral adiposity are known to drive a state of meta‐inflammation, characterized by activation of innate immune pathways (e.g., NLRP3 inflammasome) and elevated circulating levels of interleukin‐1β, interleukin‐6, and tumor necrosis factor‐alpha [25, 26]. These inflammatory mediators are pivotal in the synovitis and cartilage breakdown seen in RA and OA [27, 28]. Therefore, our results posited that metabolic abnormalities may contribute to arthritis risk, at least in part, by fueling a pro‐inflammatory systemic environment.

An intriguing and potentially clinically relevant finding is the divergent predictive performance of the metabolic indices between the two cohorts. In the U.S. NHANES cohort, the TyG–BMI demonstrated superior predictive ability for arthritis subtypes, whereas in the Chinese cohort, the TyG index alone showed the strongest predictive value. This discrepancy may be attributed to several inter‐population differences. First, the mean BMI in the U.S. cohort was notably higher (28.96 vs. 24.75 kg/m²), placing it in the overweight range. In a population with prevalent adiposity, a combined index integrating IR and BMI may capture a more comprehensive metabolic risk profile [29]. Conversely, in the Chinese cohort with a lower average BMI, IR, possibly driven by factors like dietary patterns or genetic predisposition, might be the more dominant driver of arthritis risk, diminishing the added value of BMI in the index [30]. This aligns with the notion that the TyG index can identify IR even in nonobese individuals [31]. Second, ethnic variations in body composition, fat distribution (e.g., higher visceral adiposity at lower BMIs in Asian populations), and genetic susceptibility to metabolic‐inflammatory crosstalk could contribute to these differences [32, 33]. This highlights the importance of context and population characteristics when selecting metabolic screening tools for risk stratification.

Several limitations should be acknowledged. First, the cross‑sectional design precludes causal inference, and reverse causation is plausible. Second, residual confounding from unmeasured factors (e.g., medication use, physical activity, dietary quality, and genetic susceptibility) may persist. Third, the substantial reduction in the NHANES sample due to missing data may introduce selection bias; the hospital‑based Chinese cohort may overrepresent severe cases with enriched comorbidities, and diagnostic reliance on self‑report or clinical records without standardized imaging may introduce heterogeneity in outcome definitions—both of which limit direct comparability with NHANES. Fourth, cross‑population comparisons are further complicated by temporal mismatches (NHANES 2005–2018 vs. Chinese 2022–2025), which may introduce unadjusted period effects. Fifth, mediation analysis was performed on contemporaneously measured variables; thus, indirect effects reflect statistical associations rather than causal pathways. Finally, the predictive performance of the indices was modest (AUCs 0.599–0.719), supporting their role as screening tools rather than diagnostic markers. These findings warrant validation in prospective population‑based cohorts across diverse ethnic groups.

5. Conclusion

This study provided cross‐population evidence that metabolic indices—particularly TyG and TyG–BMI—are significantly associated with the risk of multiple arthritis subtypes and that systemic inflammation partially mediates this relationship. Our findings highlighted the potential utility of simple, low‐cost screening tools for identifying high‐risk individuals in diverse clinical and public health settings.

Author Contributions

Xiumin Jiang: conceptualization, methodology, formal analysis, software, visualization, writing – original draft, writing – review and editing. Danting Wang: data curation, methodology. Zhuwei Zheng: data curation, software. Guangpu Zhu: data curation, software. Yuchao Zhang: data curation, visualization. Xianglong Feng: data curation, visualization. Huizhong Fu: data curation, supervision. Jiehui Fu: visualization, project administration. Shan Gao: visualization, project administration. Xiaowen Lian: project administration, supervision, writing – review and editing. Weiquan Zeng: conceptualization, funding acquisition, project administration, supervision, writing – review and editing.

Funding

This research was funded by the National Science Foundation of China (Grant 82575107), Fujian Province Science and Technology Innovation Joint Fund Project (Grant 2024Y9523), Traditional Chinese Orthopedics Open subject of FJTCM (Grant XGS2023002), High‐Level Key Discipline Construction Project in TCM Rehabilitation under the National Administration of Traditional Chinese Medicine (Grant zyyzdxk‐2023102), NATCM’s Project of High‐level Construction of Key TCM Disciplines (Traditional Chinese Orthopedics) (Grant zyyzdxk‐2023106), and Fujian Rehabilitation Industry Research Institute Technology Innovation Platform Construction Project (Grant 2015Y2001).

Ethics Statement

This study was conducted in accordance with the Declaration of Helsinki. For the U.S. cohort, publicly available, de‐identified data from the National Health and Nutrition Examination Survey (NHANES) were used, with prior approval by the NCHS Research Ethics Review Board and documented participant consent. For the Chinese hospital‐based cohort, the study protocol was approved by the Institutional Review Board of Fujian University of Traditional Chinese Medicine Affiliated Rehabilitation Hospital. All data were anonymized throughout the study to ensure participant confidentiality.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting Information

Additional supporting information can be found online in the Supporting Information section.

Supporting information

Jiang, Xiumin , Wang, Danting , Zheng, Zhuwei , Zhu, Guangpu , Zhang, Yuchao , Feng, Xianglong , Fu, Huizhong , Fu, Jiehui , Gao, Shan , Lian, Xiaowen , Zeng, Weiquan , Associations Between Metabolic Syndrome Index, Systemic Inflammatory Index, and Arthritis Subtype in the U.S. and Chinese Adults: A Mediation Analysis, Mediators of Inflammation, 2026, 9114075, 12 pages, 2026. 10.1155/mi/9114075

Data Availability Statement

Data will be made available upon request.