1 Introduction

Osteoarthritis (OA) is defined by the Osteoarthritis Research Society International (OARSI) as a heterogeneous joint disease characterized by progressive loss of articular cartilage, sclerosis of subchondral bone, and osteophyte formation. The pathological process of OA is accompanied by synovial inflammation and meniscus degeneration.¹ As the musculoskeletal disease with the highest disability rate in the world, the epidemiological burden of OA continues to climb. The 2023 Global Burden of Disease Study (GBD) showed that there are 654 million people with OA worldwide, with an age-standardized prevalence of 8.4 %. Notably, the prevalence rate exceeds 33 % in individuals over 65 years old.² Gender differences are significant: the risk of knee OA is 1.72 times (95 % CI: 1.58–1.88) higher in women than in men, and the risk of hip OA is 1.24 times (95 % CI: 1.12–1.38) higher in women, which may be related to fluctuations in estrogen levels and anatomical differences in joints.³ Economic models predict that OA will cause an annual economic loss of USD 78 billion in the United States by 2040, with direct medical costs accounting for 0.25 %–0.50 % of GDP.⁴ This public health crisis urgently necessitates the exploration of new prevention and treatment strategies from the perspective of modifiable risk factors.

The etiological framework of OA exhibits complex interactions, including the synergistic effects of genetic susceptibility, mechanical load, and metabolic disorders. Genetic studies have revealed that loci such as GDF5 rs143383 (OR = 1.30) and COL11A1 rs2615977 (OR = 1.18) increase OA risk by modulating chondrogenesis signaling pathways.^5,6 Obesity, as a core mechanical factor, accelerates cartilage degradation by increasing joint load (with every 1 kg increase in body weight, the risk of knee OA rises by 15 %) and promoting the secretion of pro-inflammatory adipokines (e.g., leptin and adiponectin).^7,8 Additionally, metabolic syndrome, including insulin resistance and hyperlipidemia, can induce mitochondrial dysfunction in cartilage cells, resulting in a 2.5-fold increase in reactive oxygen species (ROS) production.⁹ Notably, the “inflammaging” theory provides a new perspective on OA pathogenesis, indicating that senescent-associated secretory phenotype (SASP) factors (such as IL-6 and MMP-13) disrupt cartilage homeostasis through paracrine effects, and the elimination of senescent chondrocytes can slow OA progression by 40 %–60 %.^10,11 However, a lack of reliable biomarkers that integrate inflammation and metabolic imbalances hinders the optimization of OA risk stratification and early intervention strategies.

In recent years, the red blood cell distribution width-to-serum albumin ratio (RAR) has emerged as a novel biomarker that integrates inflammatory responses and nutrient metabolism, demonstrating clinical relevance across various disease contexts. Large-scale cohort studies indicate that elevated RAR is significantly associated with an increased risk of all-cause mortality in the general population (HR = 1.83–2.08).¹² In patients with chronic kidney disease, each 1-unit increase in RAR is associated with a 192 % increase in the risk of end-stage renal disease (HR = 2.92).¹³ The predictive value of RAR for metabolic diseases is particularly pronounced; for instance, in diabetic patients, a RAR >3.22 mL/g correlates with a 2.59-fold increase in mortality risk compared to the normal group.¹⁴ Moreover, a dose-response relationship has been established between RAR levels and carotid plaque formation in patients with coronary artery disease (OR = 1.24).^15,16 RAR not only predicts short-term mortality in acute myocardial infarction (AUC = 0.738)¹⁷ and acute respiratory failure (HR = 1.22),¹⁸ but also correlates independently with 28-day mortality in acute pancreatitis patients (Hazard Ratio, HR = 2.72).¹⁹ Additionally, the prognostic value of RAR has been confirmed in heart failure patients (HR = 3.66).²⁰In terms of its pathophysiological mechanisms, an increased red cell distribution width (RDW) reflects abnormal erythropoiesis, oxidative stress, and chronic inflammation, whereas hypoalbuminemia is closely related to malnutrition, endothelial dysfunction, and systemic inflammatory responses. By integrating these two indicators, RAR provides a more accurate characterization of the systemic imbalances within the “inflammation-metabolism” axis. However, no research has yet explored the association between RAR and the development of OA, and the absence of this crucial investigation limits the optimization of OA risk stratification models and hinders the development of precision intervention strategies based on metabolic regulation of inflammation.

Based on the above background, this study aims to systematically evaluate the association between RAR and OA risk for the first time, addressing the following research gaps: (1) to assess whether RAR is associated with increased risk of OA; (2) to explore the dose-response relationship and potential nonlinear thresholds of RAR; (3) to reveal the differences in sensitivity to RAR across different populations (such as gender, race, and metabolic status). By analyzing the mechanisms underlying the association between RAR and OA, this research provides a theoretical basis for developing OA prevention and treatment strategies focused on regulating inflammation and metabolism, and offers new insights for optimizing OA risk stratification using routine blood indicators in clinical practice.

2 Materials and methods

2.1

2.1 Study population and exclusion criteria

This study conducted a retrospective analysis utilizing the National Health and Nutrition Examination Survey (NHANES) database. Since the data used are derived from publicly available datasets and the study does not involve direct intervention with participants, ethical review board approval was not required.This study included individuals without OA as well as those diagnosed with OA.The dataset encompasses three components: questionnaires, laboratory results, and medical examinations, including information on Alcohol Use, Standard Biochemistry Profile, Body Measures, Blood Pressure & Cholesterol, Complete Blood Count with 5-Part Differential, Demographic Variables and Sample Weights, Medical Conditions, and Smoking Status. A total of 101,316 participants were included from ten NHANES cycles (1999–2000 through 2017–2018). After initial screening, we excluded 37,451 participants lacking data on albumin and red blood cell distribution width (RDW), 23,401 participants without OA-related data, 427 individuals missing demographic information (age, sex, race, education, marital status, and poverty-to-income ratio (PIR), and 4,135 participants due to missing information on other confounders (including smoking, alcohol use, body mass index (BMI), globulin, albumin-to-globulin ratio (AGR), hypertension, diabetes, asthma, congestive heart failure (CHF), coronary heart disease (CHD), angina, liver disease, and cancer).Ultimately, 35,902 participants were enrolled in the study, as detailed in Fig. 1.

Close

Fig. 1 Research flowchart.

Export to PPT

3 Results

3.3

3.3 Non-linear dose-response relationship between RAR and OA

Analysis using the restricted cubic spline (RCS) model indicated a significant non-linear association between RAR and OA risk (overall test P < 0.001, non-linear test P < 0.001). The risk of OA began to increase significantly when the RAR level reached 3.03 (OR = 1), with risk increasing further with higher RAR levels (Fig. 2).

Close

Fig. 2 Dose-response relationship between RAR and OA risk, evaluated using restricted cubic spline (RCS) analysis. The analysis adjusted for potential confounding factors, including age, sex, race, education, marital status, poverty-to-income ratio, smoking status, alcohol consumption, BMI, globulin levels, AGR, hypertension, diabetes, asthma, CHF, CHD, angina, liver disease, and cancer.

Export to PPT

3.4

3.4 Subgroup heterogeneity analysis

Interaction tests revealed significant heterogeneity in the association between RAR and OA based on gender, race, and metabolic status (interaction P < 0.05)(Fig. 3). The effect size was 3.6 times larger in men (OR = 1.07, P = 3.6 × 10^−28) compared to women (interaction P = 3.7 × 10^−19). The association was strongest among non-Hispanic whites (OR = 0.91, P = 1.4 × 10^−30). Higher education levels and poverty income ratios increased the effect size by 37 % and 43 %, respectively (P < 1 × 10^−20). Metabolic disorders significantly attenuated the RAR effect (36 % reduction in the obesity group, P = 7.4 × 10^−14), while elevated AGR (OR = 1.63, P = 9.8 × 10^−26) and smoking (interaction P = 0.041) increased risk.

Close

Fig. 3 Analysis of the dose-response relationship between RAR and OA risk across various subgroups.

Export to PPT

It is noteworthy that the association between RAR and OA risk did not demonstrate significant differences in patients with hypertension, diabetes, asthma, angina, liver diseases, and cancer, as well as in individuals without these conditions (interaction P values were all greater than 0.05). However, compared to patients with heart failure and coronary heart disease, an increase in RAR was associated with a more significant risk of OA in patients without heart failure and coronary heart disease.

4 Discussion

This study utilized data from the NHANES, comprising 35,902 participants, of whom 4,285 were diagnosed with OA. The findings revealed significant demographic differences between OA patients and those without OA (p < 0.05). The OA cohort primarily consisted of female participants (63.2 %) and showed an elderly mean age of 64.1 ± 13.3 years, with a high percentage of non-Hispanic whites (65.7 %). Socioeconomically, OA patients had a higher poverty income ratio (2.75 ± 1.60) and a greater widowhood rate (17.8 %). Additionally, metabolic characteristics highlighted a markedly higher obesity rate (46.1 % vs. 33.3 %) among OA patients, alongside lower albumin levels (41.66 ± 3.30) and elevated red blood cell distribution width (RDW) (13.40 ± 1.32). Notably, RAR was significantly increased (3.24 ± 0.49 vs. 3.11 ± 0.49), potentially indicating the roles of chronic inflammation and nutritional imbalance in OA pathogenesis.

RAR serves as a prominent biomarker, widely utilized in clinical research on conditions such as myocardial infarction, CHD, heart failure, and diabetes mellitus. Studies have consistently shown that elevated RAR correlates with worsened inflammatory states and patient prognoses. Specifically, RAR is associated with in-hospital all-cause mortality and adverse events in populations with myocardial infarction and heart failure.^21,22 Similarly, in diabetes patients, an increasing RAR value is correlated with higher risk and poorer outcomes, especially among high-risk groups. Importantly, RAR has been associated with deteriorating cardiac function and increased complication risks in myocardial and coronary diseases, underscoring its value as both an inflammatory marker and a critical tool for risk assessment across various pathologies.²³

Despite the established role of RAR in multiple diseases, its clinical significance diverges across conditions. In patients with heart failure, although RAR is used to gauge mortality risk, its complexity arises from responses to various clinical parameters.²⁴ In contrast, RAR elevation in OA often reflects chronic inflammation and nutritional states rather than mere physiological changes, highlighting the need for individualized analyses tailored to disease-specific contexts and mechanisms when assessing RAR.

Previous studies have classified blood biomarkers associated with the risk of osteoarthritis (OA) into three primary categories. Firstly, certain biomarkers have been identified as having a weak correlation with OA, demonstrating insignificant associations; notable examples include the neutrophil-to-albumin ratio (NPAR) and C-reactive protein (CRP).^25,26Secondly, a subset of biomarkers reveals a significant negative correlation with OA risk, including the C-reactive protein-albumin-lymphocyte index (CALLYI), serum translycopene concentration, and the thyroid hormone sensitivity index.^27–29Finally, various hematological indicators are positively correlated with OA risk. Specifically, serum iron and ferritin levels exhibit odds ratios (OR) of 1.231 and 1.280, respectively, in the highest quartiles.³⁰ Additionally, high-sensitivity C-reactive protein (HSCRP) levels also demonstrate a significant positive correlation within the highest quartile (OR = 1.59), underscoring the critical role of inflammation in the pathogenesis of OA.³¹

In this context, the findings of our study demonstrated that RAR exhibited a stronger association than previously reported biomarkers. The OR for the fourth quartile of RAR showed values of 3.36 (unadjusted, Model I), 2.12 (adjusted, Model II), and 1.91 (fully adjusted, Model III), remaining significantly high despite controlling for covariates (95 % CI: 1.58–2.31). This suggests that RAR's independent effect on OA risk is not only pronounced but may also provide greater clinical relevance as a biomarker than previously documented.

Nonetheless, while our study underscores RAR's potential as a biomarker for OA, several limitations must be acknowledged. The cross-sectional design precludes definitive conclusions regarding causal relationships, which may impact interpretation of our results. Additionally, the reliance on self-reported data could introduce bias, and not all possible confounders—including lifestyle and genetic factors—were controlled, limiting the generalizability of our findings. Future prospective studies should integrate dynamic biomarker changes with clinical OA progression to elucidate the exact role of RAR in OA pathogenesis. Despite these limitations, RAR's potential in clinical application remains significant. As a simple, cost-effective, and widely available biomarker, RAR is poised to play an essential role in screening and monitoring high-risk OA populations.

5 Conclusions

This study confirms a significant positive dose-response relationship between the red blood cell distribution width-to-albumin ratio (RAR) and the risk of OA, with an odds ratio of 1.91 for participants in the highest quartile (Q4) after adjusting for confounding factors. While these findings suggest a strong association, it is important to note that causality cannot be established at this time. Therefore, prospective or longitudinal studies are essential to further investigate the clinical relevance of RAR as a biomarker for OA risk.

CRediT authorship contribution statement

Qifan Chen: Conceptualization, Methodology, Formal analysis, Writing – original draft. Jiaxing Zeng: Conceptualization, Methodology. Hanhua Wu: Software, Validation. Bufan Li: Investigation. Yu-Nan Man: Data curation. LuYang Zhong: Visualization. Mao-Lin He: Writing – review & editing. Shuizhong Cen: Writing – review & editing.

Guardian/Patient's consent

The data for this study were sourced from the NHANES database, where the NHANES official has obtained approval from the appropriate institutional committees and informed consent from all participating subjects. Therefore, this study does not need to provide proof of informed consent from the subjects.

Ethical statement

The data for this study were sourced from the NHANES database, where the NHANES official has ensured that all procedures for collecting information from human subjects comply with relevant laws and institutional guidelines. Therefore, this study does not require additional ethical or moral statements.

Funding

This study was supported by the Guangdong Basic and Applied Basic Research Foundation (2022A1515012118), the Science and Technology Program of Guangzhou (No. 2024A04J4794), and the National Science Fund for Distinguished Young Scholars (82202657).