There is a potential genetic and epigenetic link between diabetes and arthritis. The current study delves into the intricate connection between diabetes and arthritis through utilizing both empirical data from the NHANES and causal inference methods through Mendelian Randomization. This research endeavor seeks to unravel the complex interplay between these conditions, with a particular focus on Rheumatoid Arthritis (RA), which has been identified as a significant association in our findings. However, after adjusting for HT, the relationship between diabetes and arthritis loses its statistical significance. The use of Mendelian Randomization further strengthens our understanding of the genetic predisposition linking Type 1 DM and RA, while it remains less evident in the context of residual arthritis and diabetes. This study sheds light on the potential genetic factors that contribute to the development of arthritis in individuals with diabetes, shedding new light on the pathogenesis of this complex disorder.

The prevalence of arthritis was found to be almost twice as high in diabetic patients compared to non-diabetic patients, indicating a potential relationship. This association, particularly with RA, supports prior studies linking diabetes and arthritis [22,23,24,25]. Bolla et al. identified a relationship between metabolic syndrome and rheumatoid arthritis, mirroring our observations [26]. Conversely, Okais et al. found diabetes prevalence in RA patients comparable to controls [27], suggesting potential variations in study design or demographics. Mahroug et al. highlighted a link between rheumatoid cachexia and RA disease parameters [28], hinting at shared metabolic disturbances. Sewell et al. reported increased odds of obesity, diabetes, HT, and dyslipidemia in arthritis patients [29], consistent with our findings and emphasizing the intricate relationship between these conditions.

During the regression analysis, total energy intake and physical activity were incorporated into the adjusted model. These variables are recognized as being correlated with both diabetes and arthritis, acting as potential confounders that may distort the genuine relationship between these conditions if not appropriately considered [12, 30]. Total energy intake serves as a reflection of overall dietary patterns, which can exert an influence on metabolic health, subsequently affecting the susceptibility to both diabetes and arthritis. Similarly, physical activity is pivotal in maintaining optimal weight, insulin sensitivity, and joint health, all of which are directly implicated in the onset and progression of both conditions.

Although the logistic regression model showed a significant gross association between arthritis and diabetes, upon adjusting for confounding factors such as age, gender, and race, the strength of this association was attenuated, yielding a marginally significant OR of 1.14. Interestingly, when we further accounted for HT, the association became statistically insignificant with an OR of 1.08. Notably, among the various arthritis subtypes, only Rheumatoid Arthritis (RA) maintained a significant association with diabetes, evidenced by an OR of 1.12. These findings add nuance to this relationship between diabetes and arthritis by showing that it may be confounded by HT, a common comorbidity in both conditions.

The observed interaction between DM and HT in relation to ‘Total’ arthritis and ‘Other’ types of arthritis is noteworthy. When examining DM and HT individually, both conditions have been found to elevate the risk of specific subtypes of arthritis. Nevertheless, upon combining DM and HT, a surprising negative correlation with arthritis incidence was observed. Initially, we postulated the existence of a shared mediator that independently promotes arthritis incidence through both DM and HT. However, upon further investigation, it became evident that the combined influence of DM and HT on arthritis does not exhibit a synergistic effect, where the combined risk exceeds the sum of their risks (i.e., 1 + 1 > 2). Instead, an antagonistic relationship was identified, indicating that the combined effect of DM and HT on arthritis risk is less than the sum of their individual effects (i.e., 1 + 1 < 2). Interestingly, this interaction was not observed for Rheumatoid Arthritis (RA) and Osteoarthritis (OA), indicating that the relationship between DM, hypertension, and these specific arthritis types might be independent or modulated by other factors. This finding is particularly relevant for clinicians as it underscores the importance of comprehensive management of both DM and hypertension in patients with arthritis, and is consistent with guidelines that recommend focusing on diabetes control to manage comorbidities [14, 15]. Further research is needed to elucidate the underlying common influencing factors and clinical implications of these interactions.

Some previous literature seems to suggest a risk factor of diabetes for osteoarthritis [31, 32]. However, our conclusion indicates that there is no definite relationship between the two. The main reason is that this study controlled for confounding factors such as age, gender, BMI, especially hypertension. Additionally, our data was obtained from the NHANES database, which minimized the potential data bias caused by small sample sizes in clinical trials. Undoubtedly, basic experiments demonstrated that a hyperglycemic microenvironment can alter chondrocyte phenotypes and induce oxidative stress, indicating a potential link between diabetes and arthritis [31]. Considering that there exists a significant disparity between ideal experimental conditions and the complex environments of real life, this conclusion seems invalid. Thus, further support from additional clinical trials and theoretical data is required to establish a correlation between diabetes and osteoarthritis.

Notably, our findings suggest there was a strong association between diabetes and arthritis among young adult subjects under 45 years of age, but the association weakened with age, especially after age 60. This contradicts previous research findings suggesting a high comorbid burden of arthritis and prediabetes among people ≥ 65 years of age, females, and non-Hispanic Whites [33]. Age-dependent insulin resistance causes a gradual increase in the prevalence of diabetes in the population. Similarly, the prevalence of arthritis increases owing to age. This could potentially be attributed to a range of factors, including but not limited to, biological predispositions, lifestyle choices, or differential access to healthcare. These findings underscore the importance of targeted interventions and tailored healthcare strategies for these specific demographic groups.

We conjecture that the discordance in the degree of association between the two diseases in young adults and older adults may be attributable to distinct pathological factors that contribute to arthritis. In young adults, mechanical trauma and oxidative stress cause chondrocytes to acquire a deterioration phenotype and produce a series of cytokines and inflammatory mediator-induced arthritis [34]. In elderly patients, the viability of cartilage and subchondral bone cells decreases with age [35]. Programmed death such as autophagy and apoptosis occur, resulting in diabetic-independent progression of OA [36]. Moreover, elderly patients have poorer physical conditions and more confounding factors [30], resulting in a weak correlation between the prevalence of arthritis and diabetes at the age of 60 years. Simultaneously, certain data were omitted during data collection, which may influence the correlation between the two variables.

Considering different arthritis patients, OA has a clear age-related relationship, with OA prevalence increasing with age. However, RA and other types of arthritis have a specific pathogenesis and have little to do with age. Patients with OA are significantly associated with diabetes under the age of 60, whereas patients with RA and other types of arthritis have a correlation between arthritis and diabetes only under the age of 45, indicating that there are different pathogenesis types of arthritis affecting the prevalence of diabetes.

The female patients in this study were found more frequently comorbid with arthritis, which is consistent with previous literature [11,12,13, 23, 24], and there is a clear association between diabetes and arthritis in females across all subtypes of arthritis. The proportion of males with diabetes is higher than that of females. However, neither total arthritis nor individual subtypes of arthritis were associated with diabetes in males. On the one hand, there are gender differences in the absolute levels of sex hormones in diabetes because of gender differences in biological processes linking diabetes and sex hormone production [37]. Clear sex differences were observed in how sex steroid hormones may modulate the risk of DM [38, 39]. Diabetes, Type 2 DM, obesity, and sex differences majorly affect intracellular glucose handling [40], which plays an important role in cartilage matrix formation and chondrocyte metabolism. On the other hand, the male participants in this study were more commonly associated with unhealthy eating habits and chronic stressful environments [41]. These factors may result in more interference between arthritis and diabetes in male patients. The interpretation of results requires more caution.

About half of the adults with prediabetes and arthritis are either physically inactive or obese, increasing their risk of developing Type 2 Diabetes [33]. Excessive physical activity is a risk factor for arthritis. However, we found that non-arthritis patients in this study were more physically active. This could result from arthritis symptoms or an underlying factor in the comorbidity of arthritis and diabetes. Arthritis can hinder the ability of adults with prediabetes to engage in physical activity to induce Type 2 Diabetes. A combination of arthritis and other chronic diseases such as obesity, has been linked to higher levels of physical activity [42]. Moderate physical activity may improve physical function and mobility and lowers blood sugar levels and body weight, thereby decreasing the risk of Type 2 Diabetes and arthritis symptoms.

Furthermore, the MR analyses provide a novel insight into the causal pathways between diabetes and arthritis. Our MR analyses suggested the significant association between Type 1 Diabetes and RA that common genetic factors or pathways may underlie both conditions. This is in line with the known role of chronic inflammation in both diseases [9]. Besides, diabetes does not appear to be a genetic risk factor for other arthritis.

One of the strengths of our study is the use of MR, which minimizes the risk of confounding and reverse causality, thus providing more robust evidence for causality [16,17,18]. Additionally, the large sample size and the use of nationally representative NHANES data enhance the generalizability of our findings. This study has certain limitations. Firstly, potential biases related to self-reported data. Our study relies on self-reported data from participants, which may be subject to recall bias and social desirability bias. These biases can affect the accuracy of the data and the study results [43]. Secondly, unmeasured confounding. While we included multiple known confounders in our analysis, there may still be unmeasured confounding factors that could influence the relationship between the independent and dependent variables [44]. Thirdly, lack of temporality in cross-sectional design. Our study uses a cross-sectional design, which limits our ability to establish causality. Since data were collected at a single point in time, we cannot determine the temporal sequence of the relationships between variables [45]. Fourthly, we lacked conversion of variables during utilizing sensitivity analyses to assess robustness. This approach involves converting continuous variables into categorical variables or vice versa to observe how these transformations impact the model outcomes. Fifthly, the NHANES database in the US is a horizontal questionnaire survey database, and it does not conduct individual follow-up research. This was a retrospective study without evidence for a specific cause-and-effect relationship between arthritis and diabetes. Sixthly, the effects of drugs were not considered in this study. Further research is needed on hypoglycemic drugs used by diabetic patients, NSAID drugs used by arthritis patients, and the relationships between different drugs and between drugs and diseases. Finally, there was missing data concerning the confounding variables. Further prospective cohort studies or animal models can better exemplify the question.