Red blood cell distribution width-to-albumin ratio and osteoarthritis: A cross-sectional and cohort study

Abstract

Background

The red cell distribution width-to-albumin ratio is a novel composite biomarker reflecting systemic inflammation and nutritional status. This study aimed to evaluate the association between the red cell distribution width-to-albumin ratio and the prevalence of osteoarthritis as well as long-term mortality risk among US adults, thereby exploring the potential value of the red cell distribution width-to-albumin ratio as a marker for osteoarthritis risk assessment and prognostic evaluation.

Methods

Data were obtained from the National Health and Nutrition Examination Survey in the United States, covering the years 2003–2016 and comprising 31,994 adult participants. The association between the red cell distribution width-to-albumin ratio and the prevalence of osteoarthritis was examined using logistic regression models. Additionally, prospective Cox regression models were used to evaluate the relationship between the red cell distribution width-to-albumin ratio and all-cause and cardiovascular mortality among patients with osteoarthritis. The models were adjusted for multiple important covariates, including age, sex, body mass index, race, lifestyle factors (such as smoking and alcohol consumption), and chronic comorbidities (such as diabetes, hypertension, and cardiovascular disease history). Restricted cubic spline analysis was performed to examine the nonlinear association between red cell distribution width-to-albumin ratio and mortality risk. Kaplan–Meier survival analysis, subgroup analyses, and interaction assessments were conducted to validate the robustness of the results.

Results

Among 31,994 participants, 3150 were identified with osteoarthritis; of these, 3147 had follow-up records and were included in the survival analysis. After adjustment for key covariates, red cell distribution width-to-albumin ratio remained positively associated with the prevalence of osteoarthritis (adjusted odds ratio = 1.13, 95% confidence interval: 1.01–1.25). Compared with the lowest quartile (Q1), the highest quartile (Q4) of red cell distribution width-to-albumin ratio was significantly associated with an increased prevalence of osteoarthritis (odds ratio = 1.27, 95% confidence interval: 1.09–1.48). During a median follow-up of 92 months, higher red cell distribution width-to-albumin ratio significantly predicted all-cause mortality (hazard ratio = 2.11, 95% confidence interval: 1.74–2.56) and cardiovascular mortality (hazard ratio = 2.08, 95% confidence interval: 1.68–2.58) among patients with osteoarthritis. Further analysis revealed a significant nonlinear J-shaped association between red cell distribution width-to-albumin ratio and all-cause mortality risk among patients with osteoarthritis, with a threshold value of 2.84, above which mortality risk increased markedly (p < 0.05).

Conclusions

This study suggests that the red cell distribution width-to-albumin ratio is associated with osteoarthritis prevalence and mortality risk among adults with osteoarthritis. Red cell distribution width-to-albumin ratio may represent a potential risk marker for osteoarthritis-related risk assessment and prognostic evaluation.

Keywords

Osteoarthritis mortality prevalence red cell distribution width-to-albumin ratio inflammation and nutrition

Introduction

Osteoarthritis (OA) is the most common chronic joint disease, affecting over 250 million people worldwide and representing a significant global public health burden.¹ With population aging and rising obesity rates, the prevalence and impact of OA continue to increase, leading to substantial disability, reduced quality of life, and increased mortality among affected individuals.^2,3 Epidemiological studies have shown that compared with the general population, patients with OA have higher all-cause and cardiovascular mortality.^4,5

Inflammation and malnutrition have been increasingly recognized as crucial contributors to OA pathogenesis and progression.^6,7 Chronic low-grade inflammation promotes cartilage degradation and joint destruction, while malnutrition—reflected by impaired protein and micronutrient status—further exacerbates disease severity and impedes recovery.^8,9 Traditional markers such as C-reactive protein (CRP), albumin, and the neutrophil-to-lymphocyte ratio have been widely used to reflect systemic inflammation or nutritional status, but their predictive value for OA outcomes remains limited.^10,11

Recently, the red cell distribution width-to-albumin ratio (RAR) has emerged as a novel composite indicator reflecting both systemic inflammation (via red cell distribution width (RDW)) and nutritional status (via serum albumin). RAR has been shown to be associated with adverse outcomes in various chronic diseases, including cardiovascular disease, diabetes, and chronic obstructive pulmonary disease.^12–14 However, its role in osteoarthritis and OA-related mortality has not been comprehensively evaluated.

Recently, researchers have developed composite indices that combine various inflammatory parameters to better predict OA prognosis. For example, Xiong et al.¹⁵ proposed the Inflammatory Burden Index, which integrates multiple inflammatory blood cell counts, and found it to be significantly associated with mortality risk among patients with OA. However, there remains a lack of simple and practical indicators that reflect both inflammation and nutrition. As a convenient ratio derived from routine blood tests, RAR may fill this gap.

Therefore, this study aimed to systematically evaluate the association between RAR and the prevalence of OA among US adults and to further explore, using prospective follow-up data, the predictive value of RAR for all-cause and cardiovascular mortality among patients with OA. By addressing this gap, we aim to provide evidence supporting the utility of RAR as a practical tool for risk stratification and prognostic assessment in osteoarthritis.

Methods

Study population

This study utilized seven consecutive cycles of publicly accessible National Health and Nutrition Examination Survey (NHANES) data collected in the United States from 2003 to 2016, encompassing a wide range of regions and demographic groups. To achieve a nationally representative sample, NHANES applies a stratified, multistage probability sampling methodology. All participants underwent physical examinations, completed health- and nutrition-related questionnaires, and provided laboratory specimens. The NHANES program was approved by the Ethics Review Board of the National Center for Health Statistics, and all participants provided written informed consent. The data contain no personally identifiable information.

Inclusion criteria for this study were adults aged ≥20 years with complete data on RAR, covariates, and OA outcomes. Exclusion criteria included age <20 years (n = 31,837), missing RAR data (n = 4019), and missing key covariate data (n = 3208). The final analytic sample comprised 31,994 individuals; the detailed participant selection flowchart is shown in Figure 1. For survival analyses among OA patients, only those with self-reported OA diagnosis and available follow-up mortality outcome data were included (n = 3147).

Figure 1.

Flowchart of the study participants.

Definition of OA

OA was identified based on self-reported questionnaire data from NHANES. The variable codes differed across survey years (MCQ190, MCQ191, and MCQ195). Participants were asked whether a physician had ever informed them that they had arthritis and, if so, to specify the type.

Assessment of RAR

RDW (%) was measured in NHANES laboratories by trained technicians using a Coulter counter. Serum albumin concentration (g/dL) was determined using the bromocresol purple (BCP) method. All measurements were performed in accordance with standardized NHANES laboratory protocols. Laboratory indicators were obtained from baseline NHANES measurements, and the selected indicators were not restricted to fasting subsamples. RAR was calculated as follows¹⁶:

RAR = RDW (%) / Albumin (g / dL)

Mortality data

Mortality outcomes were ascertained through linkage between NHANES data and the US National Death Index, with follow-up through 31 December 2019. The primary outcomes were all-cause mortality and cardiovascular mortality (ICD-10 codes I00–I99).

Assessment of covariates

Covariates included in the analysis were as follows. Sociodemographic variables included sex, age, ethnicity, educational level, marital status, and poverty income ratio (PIR). Lifestyle factors included smoking status, alcohol consumption, and recreational physical activity. Smoking status was defined according to self-reported lifetime smoking of at least 100 cigarettes. Alcohol consumption was defined according to questionnaire-reported lifetime consumption of at least 12 alcoholic drinks in any 1 year. Recreational physical activity was defined according to participation in moderate- or vigorous-intensity leisure-time physical activity during a typical week; participants reporting such activity were classified as physically active. Chronic disease conditions included cardiovascular disease (defined as a self-reported physician diagnosis of congestive heart failure, coronary heart disease, angina, myocardial infarction, or stroke), diabetes (self-reported diagnosis, current use of glucose-lowering medication, HbA1c ≥ 6.5%, or fasting blood glucose ≥126 mg/dL), and hypertension (self-reported diagnosis or current use of antihypertensive medication). All covariate information was obtained from NHANES questionnaires or laboratory measurements.

Statistical analysis

Population demographics and clinical characteristics were described according to OA status. To ensure national representativeness, analyses incorporated NHANES-recommended complex survey design elements, including sampling weights (WTMEC2YR), strata (SDMVSTRA), and primary sampling units (SDMVPSU). Categorical variables were presented as frequencies and percentages, whereas continuous variables were reported as mean ± standard deviation (SD). The relationship between RAR and OA prevalence was analyzed using multivariable logistic regression, adjusting for potential confounding factors. Results were presented as odds ratios (ORs) with corresponding 95% confidence intervals (CIs). Age-stratified analyses of OA prevalence were performed using three age groups: <45 years, 45–60 years, and >60 years. For analyses of mortality risk among patients with OA, Kaplan–Meier (KM) survival curves were constructed to compare mortality across RAR quartiles, and Cox proportional hazards models were used to estimate hazard ratios (HRs) and 95% CIs. Dose–response relationships were evaluated using restricted cubic spline (RCS) models to assess potential nonlinear associations between RAR and mortality risk, including analyses of inflection points and threshold effects. Subgroup and interaction analyses were performed for key stratification variables. All statistical analyses were performed using R software (version 4.3.3). All tests were two-sided, with p < 0.05 considered statistically significant. To further address potential residual confounding, we constructed an extended sensitivity model (Model 4) by additionally adjusting for alanine aminotransferase, aspartate aminotransferase, white blood cell count, uric acid, serum creatinine, and recreational physical activity. As an additional sensitivity analysis for mortality outcomes, participants who died within the first 24 months of follow-up were excluded to reduce the potential influence of reverse causation.

Results

Study population and baseline characteristics

A total of 31,994 eligible adults were included in the final analysis, which, after applying sampling weights, represented approximately 192,082,846 US community-dwelling residents (Table 1). Among them, 3150 participants were identified as having osteoarthritis (OA), accounting for 9.85% of the total sample. As shown in Table 1, significant differences were observed in multiple baseline characteristics between the OA and non-OA groups. Participants with OA were older, had a higher body mass index (BMI), and had higher proportions of non-Hispanic White individuals, higher educational attainment, and higher family income. Participants with OA also had a higher prevalence of smoking, hypertension, cardiovascular disease, and diabetes (all p < 0.001). RAR levels were significantly higher among participants with OA than among those without OA (p < 0.001). No statistically significant differences were observed between the groups with respect to marital status or alcohol consumption.

Table 1.

Baseline characteristics of participants with or without osteoarthritis.

Characteristic	Overall n = 31,994N = 192,082,846	Osteoarthritis		p-value
Characteristic	Overall n = 31,994N = 192,082,846	No (n =28,844)	Yes (n =3150)	p-value
Sex, n (%)				<0.001
Male	48.3%	49.8%	35.0%
Female	51.7%	50.2%	65.0%
Age, (year)	46.94 ± 16.73	45.19 ± 16.25	61.47 ± 13.28	<0.001
BMI	28.81 ± 6.73	28.60 ± 6.62	30.53 ± 7.40	<0.001
Race ethnicity, n (%)				<0.001
Mexican American	8.1%	8.8%	2.7%
Other Hispanic	4.9%	5.2%	2.0%
Non-Hispanic White	69.6%	67.7%	85.0%
Non-Hispanic Black	10.6%	11.2%	5.8%
Other Race	6.8%	7.1%	4.5%
Educational level, n (%)				0.007
College graduate or above	60.5%	60.2%	63.6%
High school or equivalent	22.9%	23.0%	21.8%
Less than high school	16.5%	16.8%	14.6%
Marriage status, n (%)				>0.9
Married or living with partner	64.4%	64.4%	64.3%
Widowed, divorced, separated, or never married	35.6%	35.6%	35.7%
PIR, n (%)				<0.001
<1.3	21.3%	21.8%	16.6%
1.3–3.5	42.8%	42.4%	45.9%
≥3.5	35.9%	35.8%	37.4%
Smoking status, n (%)				<0.001
Smoker	46.0%	45.1%	53.0%
Nonsmoker	54.0%	54.9%	47.0%
Drinking status, n (%)				0.7
Drinker	71.2%	71.3%	70.9%
Nondrinker	28.8%	28.7%	29.1%
Diabetes, n (%)				<0.001
Yes	11.9%	11.0%	19.6%
No	88.1%	89.0%	80.4%
Hypertension, n (%)				<0.001
Yes	31.1%	28.2%	55.0%
No	68.9%	71.8%	45.0%
CVD, n (%)				<0.001
Yes	8.4%	7.2%	18.5%
No	91.6%	92.8%	81.5%
RAR	3.07 ± 0.44	3.06 ± 0.44	3.17 ± 0.43	<0.001

Categorical variables are presented as percentages, with p-values calculated using survey-weighted chi-square tests. Continuous variables are expressed as mean ± standard deviation (SD), with p-values derived from survey-weighted linear regression analysis.

BMI: body mass index; CVD: cardiovascular disease; PIR: poverty income ratio; RAR: red cell distribution width-to-albumin ratio.

Association between RAR and OA prevalence

In the unadjusted model, RAR was positively associated with OA prevalence (OR = 1.61, 95% CI: 1.50–1.73, p < 0.001). After adjustment for confounders, this association remained significant (OR = 1.13, 95% CI: 1.01–1.25, p = 0.027). Stratified analyses showed that, compared with the lowest quartile of RAR (Q1, RAR ≤ 2.83), the prevalence of OA was 15%–27% higher across the Q2–Q4 groups (p for trend = 0.003, Table 2).

Table 2.

Odds ratios (95% CIs) for the prevalence of osteoarthritis according to quartiles of RAR in NHANES from 2003 to 2016.

Characteristic	Model 1 (OR (95% CI))	Model 2 (OR (95% CI))	Model 3 (OR (95% CI))
RAR	1.61 (1.50, 1.73), p < 0.001	1.16 (1.05, 1.28), p =0.005	1.13(1.01, 1.25), p =0.027
RAR (quartile)
Q1	Reference	Reference	Reference
Q2	1.66 (1.40, 1.97), p < 0.001	1.15(0.97, 1.35), p =0.110	1.15 (0.97, 1.35), p =0.100
Q3	2.10 (1.80,2.46), p < 0.001	1.18 (1.01, 1.39), p =0.035	1.16 (0.99, 1.36), p = 0.068
Q4	2.51 (2.17, 2.92), p < 0.001	1.31 (1.12, 1.52), p < 0.001	1.27 (1.09, 1.48), p =0.002
p trend	<0.001	<0.001	0.003

Model 1: Unadjusted for covariates.

Model 2: Adjusted for age, sex, and race/ethnicity.

Model 3: Further adjusted for sex, age, BMI, race/ethnicity, educational level, marital status, poverty income ratio (PIR), smoking status, alcohol consumption, as well as hypertension, cardiovascular disease, and diabetes.

95% CI: 95% confidence interval; OR: odds ratio; RAR: red cell distribution width-to-albumin ratio.

Age-stratified analysis of OA prevalence

Age-stratified analyses showed that the positive association between RAR and OA prevalence was more evident among participants aged 45–60 years and >60 years, whereas estimates among participants aged <45 years were less stable (Supplementary Tables 1 to 3).

Mortality outcomes and survival analysis in OA patients

During a median follow-up of 92 months, there were 778 all-cause deaths and 223 cardiovascular deaths among 3147 patients with OA. KM survival curves showed significant differences in mortality rates across RAR quartiles (Figure 2), with higher RAR quartiles demonstrating lower survival probabilities.

Figure 2.

(a) Kaplan–Meier survival analysis plot for all-cause mortality with quartile groups of RAR. (b) Kaplan–Meier survival analysis plot for cardiovascular mortality with quartile groups of RAR.

The findings of the Cox proportional hazards analyses are summarized in Table 3. Across the unadjusted, partially adjusted, and fully adjusted models, the HRs for all-cause mortality associated with RAR were 2.46 (95% CI: 2.15–2.81, p < 0.001), 2.25 (95% CI: 1.80–2.81, p < 0.001), and 2.11 (95% CI: 1.74–2.56, p < 0.001), respectively. For cardiovascular mortality, the corresponding HRs were 2.56 (95% CI: 2.21–2.96, p < 0.001), 2.27 (95% CI: 1.77–2.92, p < 0.001), and 2.08 (95% CI: 1.68–2.58, p < 0.001). When stratified by quartiles, individuals in the highest RAR quartile (Q4, RAR > 3.43) exhibited a 2.73-fold higher risk of all-cause mortality (HR = 2.73, 95% CI: 2.15–3.47, p < 0.001) and a 3.03-fold higher risk of cardiovascular mortality (HR = 3.03, 95% CI: 1.97–4.66, p < 0.001) compared with those in the lowest quartile (Q1).

Table 3.

Hazard ratios (95% CIs) for all-cause and cardiovascular mortality in osteoarthritis according to quartiles of RAR in NHANES from 2003 to 2016.

Characteristic	Model 1 (HR (95% CI))	Model 2 (HR (95% CI))	Model 3 (HR (95% CI))
All-cause mortality
RAR	2.46 (2.15, 2.81), p < 0.001	2.25 (1.80, 2.81), p < 0.001	2.11 (1.74, 2.56), p < 0.001
RAR (quartile)
Q1	Reference	Reference	Reference
Q2	1.55 (1.16, 2.07), p = 0.003	1.21(0.92, 1.60), p =0.168	1.15 (0.87, 1.51), p =0.330
Q3	2.20 (1.71,2.82), p < 0.001	1.65 (1.28, 2.13), p < 0.001	1.61 (1.25, 2.08), p < 0.001
Q4	4.61 (3.66, 5.80), p < 0.001	3.22 (2.53, 4.11), p < 0.001	2.73 (2.15, 3.47), p < 0.001
p trend	<0.001	<0.001	<0.001

Cardiovascular mortality
RAR	2.56 (2.21, 2.96), p < 0.001	2.27 (1.77, 2.92), p < 0.001	2.08 (1.68, 2.58), p < 0.001
RAR (quartile)
Q1	Reference	Reference	Reference
Q2	1.30 (0.78, 2.16), p = 0.323	1.01(0.61, 1.66), p =0.972	0.93 (0.57, 1.52), p =0.786
Q3	2.34 (1.35,4.07), p = 0.002	1.76 (1.00, 3.08), p = 0.049	1.76 (0.99, 3.11), p = 0.052
Q4	5.18 (3.41, 7.87), p < 0.001	3.61 (2.36, 5.52), p < 0.001	3.03 (1.97, 4.66), p < 0.001
p trend	<0.001	<0.001	<0.001

Model 1: Unadjusted for covariates.

Model 2: Adjusted for age, sex, and race/ethnicity.

95% CI: 95% confidence interval; HR: hazard ratio; RAR: red cell distribution width-to-albumin ratio.

Dose–response relationship and nonlinear analysis

RCS analyses demonstrated a nonlinear association between RAR and all-cause mortality risk (p for nonlinearity < 0.05). As shown in Figure 3(a), a threshold value of RAR = 2.84 was identified. All-cause mortality risk increased significantly when RAR exceeded 2.84 (HR = 2.04, 95% CI: 1.65–2.50, p < 0.05), whereas no significant association was observed when RAR ≤ 2.84 (p > 0.05). No nonlinear association was found between RAR and cardiovascular mortality (p for nonlinearity > 0.05, Figure 3(b)).

Figure 3.

(a) RCS plot of the relationship between RAR and all-cause mortality in osteoarthritis. The x-axis represents ln-transformed RAR, and the threshold reported in the text corresponds to the back-transformed original RAR value (RAR = 2.84). (b) RCS plot of the relationship between RAR and cardiovascular mortality in osteoarthritis.

Subgroup and interaction analyses

Analyses stratified by sex, age, and chronic disease status indicated that the relationship between RAR and mortality remained stable across subgroups. No significant interactions were identified, as all p values for interaction exceeded 0.05 (Figure 4).

Figure 4.

(a) Relationship between RAR and risk of all-cause mortality in each subgroup: point estimates represent HR, and line segments represent 95% CI. There was no significant interaction among all subgroups (all p for interaction >0.05). (b) Relationship between RAR and risk of cardiovascular mortality in each subgroup: point estimates represent HR, and line segments represent 95% CI. There was no significant interaction among all subgroups (all p for interaction > 0.05).

Sensitivity analysis

In Model 4, after further adjustment for selected biochemical, hematologic, and physical activity variables, the overall pattern of associations remained consistent with the main analyses (Supplementary Table 4). After excluding participants who died within the first 24 months of follow-up, the associations between RAR and both all-cause and cardiovascular mortality remained generally consistent with the primary findings (Supplementary Table 5).

Discussion

This study utilized nationally representative NHANES data and included a total of 31,994 adult participants. We found that RAR was significantly positively associated with OA prevalence and mortality risk. Among patients with OA, the relationship between RAR and all-cause mortality was nonlinear. Using smooth curve fitting and threshold effect analysis, we identified an inflection point at RAR = 2.84. When RAR exceeded this threshold, the risk of death increased significantly. This finding suggests that, as low-grade inflammation and malnutrition worsen, the prognosis of patients with OA deteriorates markedly. In addition, our subgroup analyses did not identify any significant interaction effects among the stratified variables, indicating that the associations between RAR and OA risk and prognosis were consistent across different populations. These findings suggest that RAR may be a potential marker associated with OA risk and long-term prognosis. In particular, the observed threshold of approximately 2.84 may help identify patients with OA who are at higher mortality risk, although this cutoff should be validated in future studies before broader clinical application.

Currently, studies examining the association of RAR with OA prevalence and mortality are limited. A recent cross-sectional study based on NHANES 2005–2018 data that included 2698 adults with diabetes first reported a significant positive association between RAR and OA prevalence, with a nonlinear relationship.¹⁷ The threshold effect analysis in that study showed that the inflection point for RAR was approximately 3.69. On the left side of the inflection point, each 1-unit increase in RAR was associated with an OR of approximately 2.61 (95% CI: 1.77–3.83) for OA prevalence. However, because of its cross-sectional design, that study was unable to assess the impact of RAR on long-term mortality outcomes among patients with OA. Therefore, our study makes an important contribution to this field. We not only expanded the sample size but also adopted a design combining cross-sectional and prospective cohort approaches, systematically validating, for the first time, the predictive value of RAR for long-term mortality risk among patients with OA, including all-cause and cardiovascular mortality. Our results confirmed that, after adjustment for multiple confounders, higher RAR levels independently predicted higher mortality risk, and this association remained consistent across subgroups. These findings are consistent with those of the aforementioned cross-sectional study and extend the potential clinical application of RAR from risk assessment to prognostic evaluation. In summary, our study expands the evidence regarding the relationship between RAR and OA and provides a basis for the future use of RAR in clinical risk stratification and individualized intervention strategies for patients with OA.

In recent years, RAR, as an integrated marker reflecting inflammation and nutritional status, has demonstrated important prognostic significance in both the general population and various chronic diseases.¹² For example, a cross-sectional analysis of 31,417 US adults revealed a significant association between RAR levels and kidney stone prevalence.¹⁸ After stratification by RAR quartiles, the risk of kidney stones in the highest quartile was approximately 62% higher than that in the lowest quartile (multivariable OR = 1.62), and a nonlinear threshold relationship was observed (RAR inflection point = 3.23). Similarly, Hao et al.,¹² through a prospective analysis of approximately 506,000 participants from NHANES and the UK Biobank, confirmed that elevated RAR was independently associated with a higher risk of all-cause mortality in the general population. This association was evident across different causes of death and exhibited a nonlinear dose–response relationship. Likewise, in patients with chronic obstructive pulmonary disease (COPD), a cohort study by Cao et al.¹⁴ (n = 1549) compared the prognostic value of various inflammatory and nutritional indicators and found that RAR had the strongest predictive value for mortality. Specifically, among patients with COPD, the risk of all-cause mortality in the highest RAR quartile was 2.45 times that in the lowest quartile (HR = 2.45, 95% CI: 1.90–3.17). Random survival forest analysis further confirmed that RAR had the greatest predictive value among all indicators examined. In addition, RAR has demonstrated similar prognostic significance in other metabolic and psychiatric disorders. Liu et al.¹³, in an NHANES-based study, identified elevated RAR as an independent predictor of mortality among patients with diabetes. For each quartile increase in RAR, the multivariable HR was 1.80 (95% CI: 1.57–2.05), and participants in the highest quartile had an approximately 159% higher risk of death than those in the lowest quartile (HR = 2.59, 95% CI: 2.18–3.09). That study also found, using RCS analysis, that the RAR risk threshold in the diabetic population was approximately 3.22, above which survival declined significantly. Similarly, in a prospective analysis of 2519 patients with depression, RAR was also confirmed to be closely associated with adverse prognosis.¹⁹ Each 1-unit increase in RAR nearly doubled the risk of all-cause mortality (HR = 1.98, 95% CI: 1.62–2.42), while the risk of cardiovascular mortality also increased significantly (HR = 1.73, 95% CI: 1.19–2.51). Overall, current evidence indicates that the high-inflammation/low-nutrition state reflected by RAR is associated with adverse outcomes, including disease onset, progression, and mortality, across various populations. The similar findings observed among patients with OA in our study further support the potential role of RAR as a broadly applicable prognostic indicator across multiple chronic diseases.

RAR consists of two components: RDW and serum albumin, which can be regarded as indicators of erythrocyte kinetic variability and nutritional status, respectively. In a chronic inflammatory environment, multiple cytokines act synergistically to increase RAR. Inflammatory factors such as IL-6 and TNF-α can inhibit erythrocyte maturation and bone marrow hematopoiesis, resulting in increased RDW.^20–22 Previous studies have shown that RDW levels are positively correlated with pro-inflammatory cytokines. Under conditions of heightened inflammation, RDW tends to increase with IL-6 and TNF-α levels.²⁰ Therefore, chronic low-grade inflammation can lead to greater erythrocyte size heterogeneity, reflected by elevated RDW. This may indicate ineffective hematopoiesis or disordered erythropoiesis. On the other hand, serum albumin, as a negative acute-phase protein, decreases significantly in states of inflammation and malnutrition. Albumin is not only a useful indicator of nutritional status but also has multiple biological functions, including antioxidant and anti-inflammatory effects and maintenance of capillary endothelial stability.^23,24 Under normal circumstances, albumin molecules can clear free radicals, stabilize the endothelial glycocalyx structure, and modulate immune responses by binding inflammatory mediators.²⁵ When chronic inflammation persists, albumin synthesis is inhibited and degradation increases, resulting in decreased serum albumin and a concomitant reduction in antioxidant and anti-inflammatory capacities.²⁵ Therefore, elevated RAR essentially reflects the composite state of “increased RDW + decreased albumin,” i.e. high inflammation and low nutrition. Among patients with OA, persistent systemic inflammation and malnutrition are closely associated with disease onset and progression.²⁶ For example, some studies have reported that the mean RDW in patients with severe knee OA is significantly higher than that in patients with mild or moderate disease, suggesting that changes in erythrocyte kinetics occur alongside the progression of inflammation and worsening symptoms in OA.²⁷ Similarly, low albumin status is closely associated with frailty, reduced quality of life, and increased mortality among older adults. In large cohorts of hospitalized patients, long-term mortality was significantly higher among those with low albumin levels than among those with normal albumin levels.²⁸ After several years of follow-up, the mortality rate among patients with normal admission albumin was approximately 29%, whereas the rates among those with mild and marked hypoalbuminemia reached 67% and 83%, respectively. Decreased albumin is often accompanied by muscle wasting, impaired immune function, and other manifestations, serving as a warning signal of poor prognosis in older adults with chronic diseases.^29,30 In summary, patients with OA exhibiting both “increased RDW + decreased albumin” often indicate a persistent state of chronic inflammation and malnutrition, which may accelerate cartilage destruction and deterioration of joint function and increase the risk of complications and mortality.

This study has several important strengths. First, the data source was the nationally representative NHANES database, which provides a large sample size, strong representativeness, and detailed long-term follow-up and outcome data, thereby enhancing the robustness and generalizability of the findings. Second, to our knowledge, this study is the first to systematically evaluate the association between RAR and mortality risk among patients with OA, filling an important gap in the literature and providing new perspectives for OA risk assessment and individualized intervention strategies.

However, several limitations should be considered when interpreting our findings. First, because the association between RAR and OA prevalence was evaluated using a cross-sectional design, the temporal sequence and causal direction cannot be determined. Although significant associations were observed, it remains unclear whether elevated RAR contributes to OA development or whether OA-related inflammatory and nutritional changes influence RAR levels. Second, residual confounding cannot be fully excluded. Although we adjusted for multiple demographic, lifestyle, and comorbidity-related factors, some potentially relevant variables, including detailed physical activity intensity and duration, dietary intake, sleep patterns, circadian-related factors, frailty-related measures, medication use, and other inflammatory or nutritional indicators, were not uniformly captured in the current analytic framework. These unmeasured factors may have influenced both OA status and mortality outcomes. Third, OA diagnosis was based on self-reported physician or healthcare professional diagnosis from NHANES questionnaire data, which may have introduced misclassification. This questionnaire-based definition did not include radiographic confirmation, orthopedic evaluation, anatomical site, or disease severity information. Similar concerns may also apply to other self-reported disease variables, including cardiovascular disease and hypertension, for which detailed clinical verification and severity information were unavailable. Fourth, RAR is a nonspecific indicator that may be influenced by multiple biological processes and clinical conditions, including anemia, chronic infections, liver disease, systemic inflammation, and nutritional status. Therefore, RAR should be interpreted cautiously and in conjunction with other clinical information. Finally, the study population consisted of US adults with specific racial, ethnic, and healthcare backgrounds; therefore, the generalizability of our findings to other countries and populations requires confirmation in large independent cohorts.

Conclusion

In summary, higher RAR was associated with OA prevalence and mortality risk among a nationally representative sample of US adults. Given the observational design and questionnaire-based definition of OA, RAR should be interpreted as a potential risk-related marker. Further prospective studies involving clinically confirmed OA are needed to validate its prognostic value and clinical applicability.

Supplemental Material

sj-docx-1-imr-10.1177_03000605261465293 - Supplemental material for Red blood cell distribution width-to-albumin ratio and osteoarthritis: A cross-sectional and cohort study

Supplemental material, sj-docx-1-imr-10.1177_03000605261465293 for Red blood cell distribution width-to-albumin ratio and osteoarthritis: A cross-sectional and cohort study by Jiaming Tang, Jing Li and Rulai Yang in Journal of International Medical Research

Footnotes

Acknowledgments

We would like to acknowledge the participants and investigators of the National Health and Nutrition Examination Survey (NHANES).

ORCID iD

Jiaming Tang

Ethics approval and consent to participate

This study was conducted in accordance with the ethical guidelines of the Declaration of Helsinki. The detailed methods and protocols of NHANES were approved by the CDC/NCHS Research Ethics Review Board and are publicly available through the CDC website. These protocols include informed consent procedures for all participants. All methods in this study were performed in accordance with relevant guidelines and regulations. This study was exempt from human subjects ethical review because the data are publicly available and de-identified.

Author contributions

The study was designed by Jiaming Tang and Jing Li. Data analysis and initial drafting of the manuscript were carried out by Jiaming Tang and Jing Li. The manuscript was revised by Rulai Yang. All authors read and approved the final manuscript.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was supported by the National Program on Key Basic Research Project (2022YFC2703401).

Declaration of conflicting interests

The authors declare that they have no competing interests.

Data availability

The datasets generated and analysis during the current study are available in the NHANES, .

Supplemental material

Supplemental material for this article is available online.

References

GBD 2016 Disease and Injury Incidence and Prevalence Collaborators. Global, regional, and national incidence, prevalence, and years lived with disability for 328 diseases and injuries for 195 countries, 1990–2016: a systematic analysis for the Global Burden of Disease Study 2016. Lancet 2017; 390: 1211–1259.

Cross

Smith E, Hoy D, et al. The global burden of hip and knee osteoarthritis: estimates from the global burden of disease 2010 study. Ann Rheum Dis 2014; 73: 1323–1330.

Glyn-Jones

, Palmer AJR, Agricola R, et al. Osteoarthritis. Lancet 2015; 386: 376–387.

, Jia J, Zhao B, et al. Objectively measured daily steps and health outcomes: an umbrella review of the systematic review and meta-analysis of observational studies. BMJ Open 2024; 14: e088524.

Nüesch

, Dieppe P, Reichenbach S, et al. All cause and disease specific mortality in patients with knee or hip osteoarthritis: population based cohort study. BMJ 2011; 342: d1165.

Aigner

Söder

Gebhard

, et al. Mechanisms of disease: role of chondrocytes in the pathogenesis of osteoarthritis – structure, chaos and senescence. Nat Clin Pract Rheumatol 2007; 3: 391–399.

Giorgino

, Albano D, Fusco S, et al. Knee osteoarthritis: epidemiology, pathogenesis, and mesenchymal stem cells: what else is new? An update. Int J Mol Sci 2023; 24: 6405.

Michael

JW-P

Schlüter-Brust

Eysel

. The epidemiology, etiology, diagnosis, and treatment of osteoarthritis of the knee. Dtsch Arztebl Int 2010; 107: 152–162.

Kuang

Chiou

, et al. Magnesium in joint health and osteoarthritis. Nutr Res 2021; 90: 24–35.

10.

Geng

Zhang

. CRP-Albumin-Lymphocyte index (CALLYI) as a risk-predicting biomarker in association with osteoarthritis. Arthritis Res Ther 2025; 27: 57.

11.

Liu

, Xu C, Driban JB, et al. Whole grain consumption and risk of radiographic knee osteoarthritis: a prospective study from the Osteoarthritis Initiative. Rheumatology (Oxford) 2023; 62: 1834–1840.

12.

Hao

, Jiang S, Tang J, et al. Ratio of red blood cell distribution width to albumin level and risk of mortality. JAMA Netw Open 2024; 7: e2413213.

13.

Liu

J, Wang X, Gao TY

, et al. Red blood cell distribution width to albumin ratio associates with prevalence and long-term diabetes mellitus prognosis: an overview of NHANES 1999–2020 data. Front Endocrinol (Lausanne) 2024; 15: 1362077.

14.

Cao

Miao

Chen

. Association of inflammation and nutrition-based indicators with chronic obstructive pulmonary disease and mortality. J Health Popul Nutr 2024; 43: 209.

15.

Xiong

Z, Xu W, Wang Y

, et al. Inflammatory burden index: associations between osteoarthritis and all-cause mortality among individuals with osteoarthritis. BMC Public Health 2024; 24: 2203.

16.

B, Li M, Yu Z

, et al. Red blood cell distribution width to albumin ratio (RAR) is associated with low cognitive performance in American older adults: NHANES 2011–2014. BMC Geriatr 2025; 25: 157.

17.

Lei

W, Mixue G, Huqiang D

, et al. Red blood cell distribution width to albumin ratio is associated with osteoarthritis prevalence among US adults with diabetes using data from NHANES 2005 to 2018. Sci Rep 2025; 15: 21529.

18.

L, Zhang Y, Chen D

, et al. Association of red blood cell distribution width to albumin ratio with the prevalence of kidney stones among the general adult population. Immun Inflamm Dis 2024; 12: e70070.

19.

Zhu

X-F

Y-Q

Long

Z-W

, et al. Association between RAR and the prevalence and prognosis of depression: a population-based study. J Affect Disord 2025; 380: 1–9.

20.

Moreno-Torres

, Sánchez-Chica E, Castejón R, et al. Red blood cell distribution width as a marker of hyperinflammation and mortality in COVID-19. Ann Palliat Med 2022; 11: 2609–2621.

21.

Joosten

LAB,

Netea MG, Kim SH, et al. IL-32, a proinflammatory cytokine in rheumatoid arthritis. Proc Natl Acad Sci USA 2006; 103: 3298–3303.

22.

Horta-Baas

Romero-Figueroa

MDS

. Clinical utility of red blood cell distribution width in inflammatory and non-inflammatory joint diseases. Int J Rheum Dis 2019; 22: 47–54.

23.

Niedziela

, Hudzik B, Szygula-Jurkiewicz B, et al. Albumin-to-globulin ratio as an independent predictor of mortality in chronic heart failure. Biomark Med 2018; 12: 749–757.

24.

Mirsaeidi

Omar

Sweiss

. Hypoalbuminemia is related to inflammation rather than malnutrition in sarcoidosis. Eur J Intern Med 2018; 53: e14–e16.

25.

, Liu T, Tian M, et al. Albumin, an interesting and functionally diverse protein, varies from ‘native’ to ‘effective’ (Review). Mol Med Rep 2024; 29: 24.

26.

Vieira

, Godziuk K, Barazzoni R, et al. Hidden malnutrition in obesity and knee osteoarthritis: assessment, overlap with sarcopenic obesity and health outcomes. Clin Nutr 2025; 48: 111–120.

27.

Taşoğlu

Bölük

Şahin Onat

, et al.

Is blood neutrophil-lymphocyte ratio an independent predictor of knee osteoarthritis severity?

Clin Rheumatol 2016; 35: 1579–1583.

28.

Akirov

Masri-Iraqi

Atamna

, et al. Low albumin levels are associated with mortality risk in hospitalized patients. Am J Med 2017; 130: 1465.e11–1465.e19.

29.

Eckart

, Struja T, Kutz A, et al. Relationship of nutritional status, inflammation, and serum albumin levels during acute illness: a prospective study. Am J Med 2020; 133: 713–722.e7.

30.

Picca

Coelho-Junior

Calvani

, et al. Biomarkers shared by frailty and sarcopenia in older adults: a systematic review and meta-analysis. Ageing Res Rev 2022; 73: 101530.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.07 MB