Rheumatoid arthritis (RA) is a common chronic rheumatoid immune disease characterized by synovitis, cartilage, and symmetrical joint damage (1). RA can affect individuals of any age and its prevalence rate is estimated at 0.5–1% worldwide (2). The pathogenic factors of RA are unclear, which may be related to infection, hormone secretion disorder, genetics, etc. (3, 4). Coffee is one of the most popular beverages in the world, with about 85% of Americans reportedly consuming at least one caffeinated beverage per day (5). Because of the widespread popularity and easy availability of coffee, the public and the scientific community are extraordinarily interested in the relationship and interaction between coffee and essential health. Coffee contains hundreds of bioactive compounds, including caffeine, caffeic acid, chlorogenic acid, etc. (6).

At present, a large number of studies have indicated that coffee can affect the risk of some chronic diseases such as depression, type 2 diabetes, Parkinson's disease (7–9), and RA. Among them, we are very interested in the relationship between coffee intake and RA risk. Previous prospective studies have shown that coffee intake was not associated with RA risk (10–12). However, Mikuls et al. found that decaffeinated coffee intake was independently positively correlated with RA risk, while the caffeinated coffee intake was not associated with RA rise (13) and similar results were obtained by a meta-analysis by Asoudeh et al. (14). Overall, their results were not consistent.

Mendelian randomization (MR) is considered a method comparable to the randomized controlled study (15), which uses genetic variation as an instrumental variable (IV) to deduce the causality between the outcome and exposure. Genetic variation is randomly assigned at conception to effectively avoid the confounding bias and reverse causal bias of traditional epidemiological studies (16). Moreover, as the effect of genetic variation is much longer than clinical intervention, the actual causal effect estimated by MR is more significant and more accurate than the clinically observed causal effect, including the meta-analysis of clinical research and prospective studies (17).

Therefore, this study intended to use the extensive database of the National Health and Nutrition Examination Survey (NHANES) to conduct a cross-sectional study on the correlation between coffee intake and RA risk, followed by the MR study to evaluate the causal relationship between coffee intake and RA from the level of genetic variation.

This present study used epidemiological analysis and two-sample MR analysis to explore the causality between coffee intake and RA risk. Our results did not support a causal association between coffee intake and RA risk.

RA is a chronic autoimmune inflammatory disease. Based on susceptible genes, RA initiates T cell activation and autoimmune reaction by infection, smoking, etc. A large number of inflammatory cytokines, autoantibodies, and oxygen free radicals increase, resulting in the inflammatory injury of joint tissues, synovial hyperplasia, and structural destruction of bone and cartilage (31). Firstly, its occurrence and development are closely related to inflammation. A prospective cohort study of RA in older women showed that caffeine-free coffee intake was independently positive correlated with RA incidence (13). Findings from observational studies further showed that coffee intake was associated with an increased risk of RA. Possible mechanisms include promoting the production of rheumatoid factor (32), antagonizing adenosine receptors, and inhibiting cytokine anti-inflammation (33, 34). Furthermore, Oxidative stress can lead to inflammatory infiltration of neutrophils, and free radicals can be used as oxidants and inflammatory mediators to participate in RA pathological process. Yashin et al. found that coffee contains unique antioxidants such as polyphenols, caffeine alcohol, and melanoid, which play an indispensable role in scavenging free radicals, inducing DNA repair, and activating detoxification enzymes (35–37). Secondly, immune regulation is also a vital influencing factor of RA. Sharif et al. suggested that coffee intake may have a protective effect on autoimmune diseases (38). Oxidative stress can decrease the number of activated peripheral blood mononuclear cells (PBMCs) and down-regulate immune function. The antioxidant activity of caffeic acid and its derivatives can maintain an adequate amount of PBMCs and contribute to maintaining immune function (39). Park et al. verified this theory through experiments on the effects of caffeic acid phenethyl ester on the immune system of mice (40). On the contrary, many animal experiments have proved that caffeine intake will inhibit the proliferation of B cells and T cells (41), affect antibody production and destroy the body's immune function (42).

In addition, a large cross-sectional study in South Korea reported that coffee consumption was not related to the prevalence of RA among Koreans (20). This result was consistent with our NHANES cross-sectional study. We speculate that coffee is a complex mixture, and the biological mechanisms by which it affects disease are multifactorial. On the one hand, caffeine in coffee can antagonize adenosine receptors and inhibit the production of cytokines, thus inhibiting anti-inflammatory effects (33, 34). However, antioxidants such as polyphenols and cafestol in coffee can inhibit the oxidation of the body and produce anti-inflammatory effects (35). On the other hand, the antioxidant activity of caffeic acid and its derivatives can maintain the effective amount of PBMCs and help maintain the body's immune function (39). Conversely, caffeine intake will inhibit the proliferation of T cells and B cells and destroy the body's immune function (41). In conclusion, due to the diversity of the pathogenesis of RA affected by coffee intake, the effects of multiple mediators (positive and negative) may be offset, resulting in coffee intake not being associated with RA. At present, the relationship between coffee intake and RA risk remains controversial. The main reason for this dispute may be that prospective and cross-sectional studies are easily affected by confounding factors such as environment and selection bias, so deterministic causality can not be obtained.

MR has been a popular and effective method for causal inference in recent years. It uses genetic variation as an IV to deduce the causality between outcome and exposure, which can effectively avoid the confounding bias of traditional epidemiological studies (43). Consulting the literature, we found only one MR research report on coffee consumption and RA risk—the result showed no causal relationship between coffee consumption and RA (44). However, only four SNPs in the study explained a small part of the difference in coffee consumption. A weak IV may bias the causal estimation of two-sample MR toward zero. This study selected the GWAS data set with the most prominent coffee intake and RA samples to solve the bias problem and finally screened 27 SNPs. The IVW, MR-Egger, weighted median, weighted mode, and simple mode were used to analyze the causality between the two samples. The results still indicated no causal relationship between coffee intake and RA risk. We speculate that people are an organic whole with a steady-state solid regulatory system, and the effect of coffee intake on inflammation and immunomodulatory function is not enough to cause damage to human tissues and organs. Therefore, coffee intake is unlikely to have a causality with the risk of RA.

The main advantages of this study include a large cross-sectional study based on NHANES and a two-sample MR analysis study. Cross-sectional studies can explore the relationship between coffee consumption and RA risk from the real-world epidemiological level based on population self-question-answering. MR overcomes the inherent effects of residual confusion, reverse causal deviation, and measurement errors in traditional epidemiological studies. The second advantage is that compared with previous MR studies on coffee and RA, we screened more SNPs and reduced the causal estimation bias caused by weak SNPs. However, this study has some inevitable limitations. Firstly, the cross-sectional study may have self-reported recall bias on RA diagnosis and coffee intake. Second, the results did not apply to other populations due to deviations from the data limited to American and European populations. Finally, valid information on coffee intake, including brewing method, type of coffee, and quantity, is very significant in exploring the causality between coffee intake and exposure genes, so it is necessary to consider these factors in further analysis of coffee and RA.

Our study did not support a causal association between coffee intake and RA risk. However, it is necessary to consider valid information on coffee intake, including brewing method, type of coffee, and quantity, in further analysis of coffee intake and RA.

All National Health and Nutrition Examination Survey files are available from the American Centers for Disease Control and Prevention database (URL https://wwwn.cdc.gov/nchs/nhanes/Default.aspx, accessed on 6 March 2022). All the Mendelian randomization study files are available from GWAS (URL https://gwas.mrcieu.ac.uk, accessed on 17 March 2022). For all the original data, please see the Supplementary material Data Sheet 2.ZIP.

The studies involving human participants were reviewed and approved by the National Center for Health Statistics Research Ethics Review Board. The patients/participants provided their written informed consent to participate in this study. Written informed consent was obtained from the individual(s) for the publication of any potentially identifiable images or data included in this article.

BP and PG designed the study, wrote, reviewed and edited the manuscript. CZ and LM analyzed data. ZZ and XZ reviewed and edited the manuscript. ZZ is the guarantor of this work. All authors approved the final version of the manuscript to be published.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.