Testing relationship between tea intake and the risk of rheumatoid arthritis and systemic lupus erythematosus: a Mendelian randomization study

Abstract

Background

We performed Mendelian randomization (MR) to assess the causal effect of tea intake on rheumatoid arthritis (RA) and systemic lupus erythematosus (SLE).

Methods

Genetic instruments for tea intake were obtained from a large genome-wide association study (GWAS) dataset of the UK Biobank. Genetic association estimates for RA (6236 cases and 147,221 controls) and SLE (538 cases and 213,145 controls) were obtained from the FinnGen study through the IEU GWAS database.

Results

MR analyses using the inverse-variance weighted method showed that tea intake was not associated with risk of RA [odds ratio (OR) per standard deviation increment in genetically predicted tea intake = 0.997, 95% confidence interval (CI) 0.658–1.511] and SLE (OR per standard deviation increment in genetically predicted tea intake = 0.961, 95% CI 0.299–3.092). Weighted median, weighted mode, MR-Egger, leave-one-out and multivariable MR controlling for several confounding factors including current tobacco smoking, coffee intake, and alcoholic drinks per week yielded completely consistent results. No evidence of heterogeneity and pleiotropy was found.

Conclusion

Our MR study did not suggest a causal effect of genetically predicted tea intake on RA and SLE.

Keywords
Mendelian randomization; Rheumatoid arthritis; Systemic lupus erythematosus; Tea intake

Introduction

Rheumatoid arthritis (RA) and systemic lupus erythematosus (SLE) are two common immune mediated inflammatory disorders. RA affects approximately 0.5–1% of the global population, characterized by symmetrical polyarthritis, chronic synovial inflammation, gradual joint deterioration, and chronic disability [¹]. SLE can affect both sexes and all ethnicities [²]. The hallmark of SLE includes the production of autoantibodies, aberrant formation of immune complexes, increased levels of pro-inflammatory cytokines, and chronic inflammation of multiple organ systems [³]. The pathogenesis of RA and SLE is complex. Genetic, environmental, and lifestyle factors can affect the risk of RA and SLE. Although lifestyle factors such as alcohol drinking and cigarette smoking have been extensively studied, the effect of tea intake on the risk of RA and SLE remains largely unknown.

Tea is one of the most commonly consumed beverages all over the world. It consists of a number of bioactive compounds including polyphenols and caffeine. These compounds possess antioxidative, anti-inflammatory, immunosuppressive, and cancer-preventative properties [⁴–⁷]. Some experimental studies showed that tea and its active ingredients such as epigallocatechin-3-gallate (EGCG) could ameliorate animal models of RA by regulating Th17/Treg balance, inducing B cell apoptosis, decreasing pro-inflammatory cytokine production, and inhibiting proliferation of activated synovial fibroblasts [⁸–¹¹]. Evidence from experimental studies also demonstrated that tea extract alleviated experimental lupus nephritis through promotion of the nuclear factor E2-related factor 2 signaling pathway, inhibition of renal nucleotide-binding domain-like receptor protein 3 activation, and regulation of systemic regulatory T-cell activity [¹²]. However, observational clinical studies evaluating the association of tea intake with RA and SLE have yielded inconsistent results [¹³–¹⁹].

Mendelian randomization (MR) is an analytical method applying genetic variants as instrumental variables to examine causal associations from observational data. Genetic variants are randomly distributed at conception based on Mendel's law. MR analyses are thus less susceptible to possible confounding factors and reverse causation. In the epidemiology field, MR is increasingly used for assessing the causal effect of a risk or protective factor on an outcome. Prior MR studies have evaluated the causal relationships of several lifestyle factors such as coffee intake, alcohol drinking, and smoking with the risk of RA and SLE [²⁰–²⁴], whereas the role of tea intake remains unclear. In the present study, we applied a two- sample MR approach to assess the causal effect of genetically predicted tea intake on the risk of RA and SLE.

Methods

Two-sample MR was performed using summary-level data from published genome-wide association studies (GWASs). Ethical approval was not necessary for this MR study; we did not use individual-level data.

Genetic data for exposures

The GWAS summary statistics for tea intake in individuals of European ancestry (447,485 participants) based on the UK Biobank were obtained from the IEU GWAS database at https://gwas.mrcieu.ac.uk/datasets/ukb-b-6066/. Briefly, the UK Biobank is a large prospective cohort of more than 500,000 participants (aged 40 to 69 years); they were recruited across England, Wales, and Scotland between 2006 and 2010 [²⁵]. As part of the UK Biobank diet Survey, self-reported tea intake was ascertained using a touchscreen question: "How many cups of tea do you drink each day? (include black and green tea)" Participants who answered > 99 or < 0 were excluded, and those answering > 20 were asked to confirm. Genotyping of the participants was undertaken using Affymetrix UK Biobank Axiom array and imputed against the UK10K, 1000 Genomes Phase 3 and Haplotype Reference Consortium panels [²⁶]. The GWAS was adjusted for age, sex, genotyping array, and the principal components by the original GWAS researchers.

Single nucleotide polymorphisms (SNPs) that strongly associated with tea intake at genome-wide significance (P < 5 × 10⁻⁸) were selected as instrumental variables. Before performing MR analyses, SNP instruments were clumped at r² < 0.001 using the R package "TwoSampleMR" version 0.5.6 (https://github.com/MRCIEU/TwoSampleMR). In addition, palindromic SNPs with intermediate allele frequencies were removed. ^{Additional file 1: Table S1} shows the information on genetic variants included as instruments for tea intake.

Current tobacco smoking summary-level data were derived from the Medical Research Council-Integrative Epidemiology Unit (MRC-IEU) consortium (462,434 participants), which is available at https://gwas.mrcieu.ac.uk/datasets/ukb-b-223/. Summary statistics for alcoholic drinks per week were obtained from the GWAS and Sequencing Consortium of Alcohol and Nicotine use (335,394 participants) at https://gwas.mrcieu.ac.uk/datasets/ieu-b-73/[²⁷]. Genetic instruments for coffee intake were derived from the MRC-IEU consortium (428,860 participants) at https://gwas.mrcieu.ac.uk/datasets/ukb-b-5237/. All data used were from individuals of European ancestry. ^{Additional file 2: Table S2}, ^{Additional file 3: Table S3} and ^{Additional file 4: Table S4} show the information on genetic variants included as instruments for current tobacco smoking, alcoholic drinks per week, and coffee intake.

Genetic data for outcomes

The association of genetic instruments with RA was retrieved from the FinnGen study (6236 RA cases and 147,221 controls) at https://gwas.mrcieu.ac.uk/datasets/finn-b-M13_RHEUMA/. The association of genetic instruments with SLE was also obtained from the FinnGen study (538 SLE cases and 213,145 controls), which is available at https://gwas.mrcieu.ac.uk/datasets/finn-b-M13_SLE/. Launched in 2017, the FinnGen study is a nationwide cohort study which aims to collect and analyze genome and health data from 500,000 Finnish biobank participants. RA and SLE were identified according to International classification of diseases (ICD) codes retrieved from nationwide registries in Finland. Genetic associations in FinnGen were evaluated with adjustment for age, sex, genotype batch, and 10 principal components by the FinnGen researchers. Linkage disequilibrium proxies (r² > 0.8) were applied if the SNP instruments were missing from the outcome dataset. We harmonised the exposure and outcome datasets to ensure that the genetic associations reflect the same effect allele.

Statistical analysis

Our MR study is based on three key assumptions (Additional file 5: Figure S1). To assess if the SNP instruments obtained were robustly associated with tea intake, we calculated the F statistic as previously described by Noyce and colleagues [²⁸]. We considered an F-statistic of more than 10 as sufficient to conduct MR analyses, which is conventionally accepted in the field [²⁹]. In the main MR analyses, we combined the causal estimates for all instrumental variables with the use of the inverse-variance weighted method. It provides reliable causal estimates when directional pleiotropy is absent [³⁰]. To assess the robustness of our main MR results to pleiotropy, the weighted median, weighted mode, MR-Egger, and MR pleiotropy residual sum and outlier (MR-PRESSO) methods were used in sensitivity analyses. The weighted median method can provide a robust evaluation as long as at least 50% of the weight stems from valid instruments satisfying the MR assumptions [³¹]. The weighted mode method can estimate the causal relationship allowing for even the majority of instrumental variables to be pleiotropic [³²]. The MR-Egger approach may be used even though all genetic variants are invalid instrumental variables [³⁰]. We estimated the average directional pleiotropy across instrumental variables with the intercept P value of the MR-Egger regression. The MR-PRESSO method (https://github.com/rondolab/MR-PRESSO) can detect horizontal pleiotropic outliers [³³]. Besides these robust methods, to minimize any possible confounding introduced by individual SNPs, a leave-one-out analysis was conducted in our sensitivity checks. To look for signs of heterogeneity, Cochran's Q statistic and I² were calculated [³⁴].

To further control for potential pleiotropic effects, a multivariable MR was carried out [³⁵]. We took into account several covariates including current tobacco smoking, alcoholic drinks per week, and coffee intake. The inverse-variance weighted method was used for multivariable MR. All statistical tests were two sided. All MR analyses were undertaken using R software version 4.0.4 (R Foundation for Statistical Computing, Vienna, Austria).

Results

As shown in ^{Additional file 1: Table S1}, among genetic variants included as instruments for tea intake, rs2472297 has the strongest association with tea intake (P = 2.30 × 10⁻¹⁰⁹). The 40 SNPs for tea intake corresponded to a F-statistic of 62.99, explaining 0.56% of the variance in tea intake. In the primary MR analyses, tea intake was not associated with risk of RA (OR per standard deviation increment in genetically predicted tea intake = 0.997, 95% CI 0.658–1.511, P = 0.988) and SLE (OR per standard deviation increment in genetically predicted tea intake = 0.961, 95% CI 0.299–3.092, P = 0.947) using the inverse-variance weighted method (Table 1 and Fig. 1). The MR-Egger, weighted median and weighted mode MR estimates yielded similar results to the inversevariance weighted analyses (Table 1 and Fig. 1), suggesting that genetically predicted tea intake was not causally associated with RA and SLE. In the main MR analyses, we found absence of heterogeneity across the individual MR estimates derived from the 40 SNPs (tea intake and RA: Q-statistic = 49.267, P = 0.126; tea intake and SLE: Q-statistic = 39.853, P = 0.432). The MR-Egger intercept test indicated no evidence of directional pleiotropy (tea intake and RA: intercept of 0.003, P = 0.724; tea intake and SLE: intercept of 0.007, P = 0.775) (Table 1). In addition, using the MR-PRESSO method, we did not find the presence of pleiotropy (tea intake and RA: P-value global test 0.122; tea intake and SLE: P-value global test 0.44). The leave-one-out analyses did not identify any outliers for all estimates (^{Additional file 6: Table S5}), indicating that the MR results were not driven by any single SNP. ^{Additional file 7: Table S6} shows the inverse-variance weighted MR results of the causal effect of genetically predicted current tobacco smoking, alcoholic drinks per week, and coffee intake on RA and SLE.

Table 1
Univariable MR estimates of genetically predicted tea intake on the risk of rheumatoid arthritis and systemic lupus erythematosus

Exposure	Outcome	Number of instruments	MR method	OR	Association 95% CI	P value	Pleiotropy Egger intercept	P value	Heterogeneity Q-statistic	P value
Tea intake	Rheumatoid arthritis	40	Inverse-variance weighted	0.997	0.658–1.511	0.988			49.267	0.126
			MR-Egger	0.859	0.341–2.162	0.748	0.003	0.724	49.103	0.107
			Weighted median	0.770	0.422–1.403	0.393
			Weighted mode	0.880	0.429–1.805	0.728
Tea intake	Systemic lupus erythematosus	40	Inverse-variance weighted	0.961	0.299–3.092	0.947			39.853	0.432
			MR-Egger	0.685	0.051–9.154	0.776	0.007	0.775	39.766	0.391
			Weighted median	0.407	0.069–2.396	0.320
			Weighted mode	0.475	0.072–3.145	0.445

• CI confidence interval, OR odds ratio, MR Mendelian randomization

Fig. 1
Univariable Mendelian randomization analyses for the association of genetically instrumented tea intake with RA and SLE. CI, confidence interval; IVW, inverse-variance weighted; OR, odds ratio; RA, rheumatoid arthritis; SLE, systemic lupus erythematosus

We further carried out a multivariable MR to evaluate the causal effect of tea intake on RA and SLE by adjusting for current tobacco smoking, alcoholic drinks per week, and coffee intake. The results did not suggest any causal association of tea intake with RA after adjusting for current tobacco smoking (OR = 0.992, 95% CI 0.676–1.459, P = 0.968), alcoholic drinks per week (OR = 1.046, 95% CI 0.680–1.613, P = 0.837), and coffee intake (OR = 0.980, 95% CI 0.660–1.191, P = 0.902) (Fig. 2). Similarly, there was no causal association between tea intake and SLE after adjusting for current tobacco smoking (OR = 1.097, 95% CI 0.343–3.521, P = 0.876), alcoholic drinks per week (OR = 1.173, 95% CI 0.391–3.525, P = 0.774), and coffee intake (OR = 1.246, 95% CI 0.347–4.436, P = 0.738) (Fig. 2).

Fig. 2
Multivariable Mendelian randomization analyses for the association of genetically instrumented tea intake with RA and SLE. CI, confidence interval; OR, odds ratio; RA, rheumatoid arthritis; SLE, systemic lupus erythematosus

In summary, the main MR analyses using the inversevariance weighted method and sensitivity analyses using several other reliable MR methods, leave-one-out analyses, and multivariable MR provided consistent results, suggesting that genetically predicted tea intake was not causally associated with RA and SLE.

Discussion

Tea intake is among the most widely consumed beverages in the world. Our study estimated the causal association of tea intake with RA and SLE in individuals of European ancestry. Across the primary and all sensitivity MR estimates, our study provided no support for a causal effect of tea intake on the risk of RA and SLE.

Tea and its extract such as caffeine and polyphenols have been showed to display antiinflammatory, immunosuppressive, and antioxidative properties. Although tea or its extract was reported to ameliorate animal models of RA and SLE in some experimental studies, observational clinical studies failed to replicate animal findings in humans [¹⁴, ¹⁶, ³⁶]. Several studies even observed that tea intake was associated with increased risk of RA and SLE [¹⁵, ¹⁸, ³⁷]. Observational studies are susceptible to biases such as residual confounding and selection bias [³⁸]. It is possible that the prior observational studies may not reliably control for these biases. In some case–control studies, control subjects were not selected for matching RA or SLE patients on confounding factors. For example, the study by Kiyohara et al. [¹⁵] recruited university students and staffs from nursing homes and clinics as control subjects; they had different characteristics compared with SLE patients. In addition, some observational studies collected information on tea intake and confounding factors only at baseline, which may lead to misclassification. Furthermore, tea is frequently consumed in combination with cigarettes and alcohol. It is thought that conventional observational studies have difficulty in separating the effect of tea from cigarettes and alcohol when evaluating the effect of tea intake on RA and SLE. Besides these shortness, conventional observational studies are not able to determine the causal association of tea intake with the risk of RA and SLE. Therefore, reliable conclusions may not be drew from conventional observational studies.

MR is a promising approach in genetic epidemiology. It can minimize potential biases in conventional observational studies and perform causal inference. To clarify the causal association, we performed a two-sample MR using summary-level statistics from large GWAS datasets. We selected SNPs that were strongly associated with tea intake as instruments; the potential for weak instrument bias was reduced since all instruments had sufficiently large F-statistics (> 10). The main MR analyses using the inverse-variance weighted method showed null causal associations of tea intake with RA and SLE and no evidence of heterogeneity. The sensitivity analyses using the MR-Egger, weighted median, weighted mode, and multivariable MR analyses yielded completely consistent results with those of the main MR analyses. The MREgger intercept test, the MR-PRESSO global test, and the leave-one-out analyses indicated low likelihood of pleiotropy. Thus, our MR findings were robust against violations of the MR assumptions.

These findings are in agreement with two prior metaanalyses evaluating the association between tea and risk of RA [³⁹, ⁴⁰]. Lee et al. [³⁹] summarized available evidence from three studies (cohort or case–control design) including 638 patients with RA and 113,998 controls. They found no association between tea intake and the incidence of RA in the overall meta-analysis (relative risk (RR) = 0.88, 95% CI = 0.62–1.24, P = 0.463) and subgroup analysis according to study design (case–control design: RR = 1.00, 95% CI = 0.58–1.72, P = 1.000; cohort design: RR = 0.68, 95% CI = 0.27–1.71, P = 0.410). Similarly, the meta-analysis by Asoudeh et al. [⁴⁰] using prospective cohort studies (823 RA cases and 191,313 controls) did not observe significant relation between tea consumption and the risk of RA (RR = 1.05, 95% CI 0.73–1.53). There were no meta-analyses estimating the relationship between tea and SLE risk. Our findings were also in line with a recently published MR study by Chen and colleagues who showed no effect of tea consumption on RA (sample size: 58,284, OR = 1.24, 95% CI 0.81–1.91) [⁴¹]. Compared with that MR study, we increased the statistical power by using a much larger study sample (447,485 participants). Additionally, we controlled for several confounding lifestyle factors including smoking, alcohol drinking, and coffee consumption using multivariable MR, which was important but neglected by Chen and coworkers [⁴¹].

This MR study is subject to some limitations. Firstly, we assessed the effects of black tea and green tea together but did not separate their effects, due to the questionnaire design in the original GWAS study. Black tea and green tea differ in terms of chemical composition [⁴¹]. The major polyphenolic compounds of black tea include thearubigins and theaflavins, whereas green tea is rich in catechins such as epicatechin-3-gallate and epicatechin and EGCG [⁴²]. Limited evidence showed that consumption of black tea and green tea exhibited a different association with SLE risk [¹⁵]. We hope that future MR studies can evaluate the causal effect of black tea and green tea intake on RA and SLE separately when more data becomes available. Secondly, we were not able to assess non-linear associations since two-sample MR can only assess linear associations. Future MR analyses are needed to evaluate the non-linear relationships of tea intake with RA and SLE. Thirdly, it is known that one of the shortcomings of MR is limited statistical power. We cannot rule out completely the possibility that the null association of tea intake with RA and SLE was owing to a small causal role which was undetectable in this MR analysis. However, it is certain that our MR results did not suffer from weak instrument bias since the instruments used were strong (F-statistic > 10). Fourthly, since different ethnic populations such as Europeans and Asians have different tea intake habits, more research is required to investigate the generalizability of our results to races other than Europeans.

Conclusion

In sum, we did not find any effect of genetically predicted tea intake on the risk of RA and SLE based on summary-level data from large GWAS in individuals of European ancestry.

• Funding
  There is no funding for this study.
• Declarations
  Ethics approval and consent to participate
  Not applicable.
• Consent for publication
  Not applicable.
• Publisher's Note
  Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Availability of data and materials

This Mendelian randomization study is based on summary-level data that have been made publicly available in the IEU GWAS database (https://gwas.mrcieu.ac.uk/).

Acknowledgements

Not applicable.

References

• ¹
  Cush JJ. Rheumatoid arthritis: early diagnosis and treatment. Rheum Dis Clin North Am. 2022;48(2):537–47.
• ²
  Barber MRW, Drenkard C, Falasinnu T, Hoi A, Mak A, Kow NY, et al. Global epidemiology of systemic lupus erythematosus. Nat Rev Rheumatol. 2021;17(9):515–32.
• ³
  Mathias LM, Stohl W. Systemic lupus erythematosus (SLE): emerging therapeutic targets. Expert Opin Ther Targets. 2020;24(12):1283–302.
• ⁴
  Neyestani TR, Gharavi A, Kalayi A. Selective effects of tea extract and its phenolic compounds on human peripheral blood mononuclear cell cytokine secretions. Int J Food Sci Nutr. 2009;60(Suppl 1):79–88.
• ⁵
  Peairs A, Dai R, Gan L, Shimp S, Rylander MN, Li L, et al. Epigallocatechin-3-gallate (EGCG) attenuates inflammation in MRL/lpr mouse mesangial cells. Cell Mol Immunol. 2010;7(2):123–32.
• ⁶
  Bañuls-Mirete M, Ogdie A, Guma M. Micronutrients: essential treatment for inflammatory arthritis? Curr Rheumatol Rep. 2020;22(12):87.
• ⁷
  Song C, Xu S, Chang L, Zhao X, Mei X, Ren X, et al. Preparation of EGCG decorated, injectable extracellular vesicles for cartilage repair in rat arthritis. Regen Biomater. 2021;8(6):67.
• ⁸
  Marotte H, Ruth JH, Campbell PL, Koch AE, Ahmed S. Green tea extract inhibits chemokine production, but up-regulates chemokine receptor expression, in rheumatoid arthritis synovial fibroblasts and rat adjuvant-induced arthritis. Rheumatology0. 2010;49(3):467–79.
• ⁹
  Liu D, Li P, Song S, Liu Y, Wang Q, Chang Y, et al. Pro-apoptotic effect of epigallo-catechin-3-gallate on B lymphocytes through regulating BAFF/PI3K/Akt/mTOR signaling in rats with collagen-induced arthritis. Eur J Pharmacol. 2012;690(1–3):214–25.
• ¹⁰
  Wu D, Wang J, Pae M, Meydani SN. Green tea EGCG, T cells, and T cell-mediated autoimmune diseases. Mol Aspects Med. 2012;33(1):107–18.
• ¹¹
  Riegsecker S, Wiczynski D, Kaplan MJ, Ahmed S. Potential benefits of green tea polyphenol EGCG in the prevention and treatment of vascular inflammation in rheumatoid arthritis. Life Sci. 2013;93(8):307–12.
• ¹²
  Tsai PY, Ka SM, Chang JM, Chen HC, Shui HA, Li CY, et al. Epigallocatechin-3-gallate prevents lupus nephritis development in mice via enhancing the Nrf2 antioxidant pathway and inhibiting NLRP3 inflammasome activation. Free Radic Biol Med. 2011;51(3):744–54.
• ¹³
  Mikuls TR, Cerhan JR, Criswell LA, Merlino L, Mudano AS, Burma M, et al. Coffee, tea, and caffeine consumption and risk of rheumatoid arthritis: results from the Iowa Women's Health Study. Arthritis Rheum. 2002;46(1):83–91.
• ¹⁴
  Karlson EW, Mandl LA, Aweh GN, Grodstein F. Coffee consumption and risk of rheumatoid arthritis. Arthritis Rheum. 2003;48(11):3055–60.
• ¹⁵
  Kiyohara C, Washio M, Horiuchi T, Asami T, Ide S, Atsumi T, et al. Modifying effect of N-acetyltransferase 2 genotype on the association between systemic lupus erythematosus and consumption of alcohol and caffeine-rich beverages. Arthritis Care Res. 2014;66(7):1048–56.
• ¹⁶
  Washio M, Fujii T, Kuwana M, Kawaguchi Y, Mimori A, Horiuchi T, et al. Lifestyle and other related factors for the development of mixed connective tissue disease among Japanese females in comparison with systemic lupus erythematosus. Mod Rheumatol. 2014;24(5):788–92.
• ¹⁷
  Rambod M, Nazarinia M, Raieskarimian F. The impact of dietary habits on the pathogenesis of rheumatoid arthritis: a case-control study. Clin Rheumatol. 2018;37(10):2643–8.
• ¹⁸
  Lamichhane D, Collins C, Constantinescu F, Walitt B, Pettinger M, Parks C, et al. Coffee and tea consumption in relation to risk of rheumatoid arthritis in the women's health initiative observational cohort. J Clin Rheumatol. 2019;25(3):127–32.
• ¹⁹
  Westerlind H, Palmqvist I, Saevarsdottir S, Alfredsson L, Klareskog L, Di Giuseppe D. Is tea consumption associated with reduction of risk of rheumatoid arthritis? A Swedish case-control study. Arthritis Res Ther. 2021;23(1):209.
• ²⁰
  Bae SC, Lee YH. Coffee consumption and the risk of rheumatoid arthritis and systemic lupus erythematosus: a Mendelian randomization study. Clin Rheumatol. 2018;37(10):2875–9.
• ²¹
  Bae SC, Lee YH. Alcohol intake and risk of rheumatoid arthritis: a Mendelian randomization study. Z Rheumatol. 2019;78(8):791–6.
• ²²
  Jiang X, Alfredsson L. Modifiable environmental exposure and risk of rheumatoid arthritis-current evidence from genetic studies. Arthritis Res Ther. 2020;22(1):154.
• ²³
  Jiang X, Zhu Z, Manouchehrinia A, Olsson T, Alfredsson L, Kockum I. Alcohol consumption and risk ofcommon autoimmune inflammatory diseases-evidence from a large-scale genetic analysis totaling 1 million individuals. Front Genet. 2021;12:687745.
• ²⁴
  Wang P, Dan YL, Wu Q, Tao SS, Yang XK, Wang DG, et al. Non-causal effects of smoking and alcohol use on the risk of systemic lupus erythematosus. Autoimmun Rev. 2021;20(9):102890.
• ²⁵
  Bycroft C, Freeman C, Petkova D, Band G, Elliott LT, Sharp K, et al. The UK Biobank resource with deep phenotyping and genomic data. Nature. 2018;562(7726):203–9.
• ²⁶
  Cole JB, Florez JC, Hirschhorn JN. Comprehensive genomic analysis of dietary habits in UK Biobank identifies hundreds of genetic associations. Nat Commun. 2020;11(1):1467.
• ²⁷
  Liu M, Jiang Y, Wedow R, Li Y, Brazel DM, Chen F, et al. Association studies of up to 1.2 million individuals yield new insights into the genetic etiology of tobacco and alcohol use. Nat Genet. 2019;51(2):237–44.
• ²⁸
  Noyce AJ, Kia DA, Hemani G, Nicolas A, Price TR, De Pablo-Fernandez E, et al. Estimating the causal influence of body mass index on risk of Parkinson disease: a Mendelian randomisation study. PLoS Med. 2017;14(6):e1002314.
• ²⁹
  Burgess S, Thompson SG. CRP CHD genetics collaboration avoiding bias from weak instruments in Mendelian randomization studies. Int J Epidemiol. 2011;40(3):755–64.
• ³⁰
  Bowden J, Davey Smith G, Haycock PC, Burgess S. Consistent estimation in mendelian randomization with some invalid instruments using a weighted median estimator. Genet Epidemiol. 2016;40(4):304–14.
• ³¹
  Burgess S, Bowden J, Fall T, Ingelsson E, Thompson SG. Sensitivity analyses for robust causal inference from mendelian randomization analyses with multiple genetic variants. Epidemiology. 2017;28(1):30–42.
• ³²
  Hartwig FP, Davey Smith G, Bowden J. Robust inference in summary data Mendelian randomization via the zero modal pleiotropy assumption. Int J Epidemiol. 2017;46(6):1985–98.
• ³³
  Verbanck M, Chen CY, Neale B, Do R. Detection of widespread horizontal pleiotropy in causal relationships inferred from Mendelian randomization between complex traits and diseases. Nat Genet. 2018;50(5):693–8.
• ³⁴
  Bowden J, Del Greco MF, Minelli C, Davey Smith G, Sheehan N, Thompson J. A framework for the investigation of pleiotropy in two-sample summary data Mendelian randomization. Stat Med. 2017;36(11):1783–802.
• ³⁵
  Burgess S, Thompson SG. Multivariable Mendelian randomization: the use of pleiotropic genetic variants to estimate causal effects. Am J Epidemiol. 2015;181(4):251–60.
• ³⁶
  Pattison DJ, Symmons DP, Lunt M, Welch A, Luben R, Bingham SA, et al. Dietary risk factors for the development of inflammatory polyarthritis: evidence for a role of high level of red meat consumption. Arthritis Rheum. 2004;50(12):3804–12.
• ³⁷
  Mazzucca CB, Scotti L, Cappellano G, Barone-Adesi F, Chiocchetti A. Nutrition and rheumatoid arthritis onset: a prospective analysis using the UK biobank. Nutrients. 2022;14(8):1554.
• ³⁸
  Khasawneh LQ, Al-Mahayri ZN, Ali BR. Mendelian randomization in pharmacogenomics: the unforeseen potentials. Biomed Pharmacother. 2022;150:112952.
• ³⁹
  Lee YH, Bae SC, Song GG. Coffee or tea consumption and the risk of rheumatoid arthritis: a meta-analysis. Clin Rheumatol. 2014;33(11):1575–83.
• ⁴⁰
  Asoudeh F, Dashti F, Jayedi A, Hemmati A, Fadel A, Mohammadi H. Caffeine, coffee, tea and risk of rheumatoid arthritis: systematic review and dose-response meta-analysis of prospective cohort studies. Front Nutr. 2022;9:822557.
• ⁴¹
  Chen S, Chen T, Chen Y, Huang D, Pan Y, Chen S. Causal association between tea consumption and bone health: a mendelian randomization study. Front Nutr. 2022;9:872451.
• ⁴²
  Khan N, Mukhtar H. Tea polyphenols in promotion of human health. Nutrients. 2018;11(1):39.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s42358-023-00290-7.

Additional file 1: Table S1.

Summary results of the SNPs associated with tea intake.

Additional file 2: Table S2.

Summary results of the SNPs associated with current tobacco smoking.

Additional file 3: Table S3.

Summary results of the SNPs associated with alcoholic drinks per week.

Additional file 4: Table S4.

Summary results of the SNPs associated with coffee intake.

Additional file 5: Figure S1.

Schematic representation of the Mendelian randomisation design. Our Mendelian randomization study is based on three key assumptions: (1) instruments are associated with tea intake; (2) instruments have no associations with confounding factors; and (3) instruments directly affect rheumatoid arthritis and systemic lupus erythematosus through tea intake.

Additional file 6: Table S5.

Leave-one-out analyses using the IVW method.

Additional file 7: Table S6.

The IVW MR estimates on the causal effect of genetically predicted current tobacco smoking, alcoholic drinks per week, and coffee intake on rheumatoid arthritis and systemic lupus erythematosus.

Publication Dates

• Publication in this collection
  01 Sept 2023
• Date of issue
  2023

History

• Received
  05 July 2022
• Accepted
  24 Feb 2023
• Published
  10 Mar 2023