Dose–response relationships of tea and coffee consumption with gout: a prospective cohort study in the UK Biobank

Abstract

Objectives

The association of tea or coffee consumption with gout is inconsistently reported. Few prospective studies have explored their dose–response relationship. Therefore, we aimed to quantitatively investigate the association between tea, coffee and the risk of developing gout.

Methods

The study included 447 658 participants in the UK Biobank who were initially free of gout. Tea and coffee consumption were assessed at baseline. We used Cox proportional hazards models to estimate the associations between tea/coffee consumption and incident gout, with restricted cubic spline added to the Cox models to evaluate the dose–response relationships.

Results

During a median follow-up period of 13.42 years, we recorded 3,053 gout cases. The associations between tea, coffee and gout were nonlinear, with a significant reduction in the risk by ∼ six cups/day of tea and three cups/day of coffee. Compared with those who were not tea and coffee drinkers, those who consumed >6 cups/day of tea or coffee were associated with 23% (HR 0.77, 95% CI, 0.66, 0.91) and 40% (HR 0.60, 95% CI, 0.47, 0.77) lower risks of gout, respectively, and both caffeinated and decaffeinated coffee consumption were associated with a decreased risk. Moreover, hyperuricaemia may modify the association between coffee and gout. Compared with non-coffee consumers with hyperuricaemia, those with ≥4 cups/day coffee intake without hyperuricaemia had the lowest risk (HR 0.34, 95% CI, 0.28, 0.41).

Conclusion

Consumption of tea or coffee had a strong nonlinear association in gout risk reduction. Hyperuricaemia status had a potential effect modification on the association of coffee intake with gout.

Introduction

Gout is an inflammatory arthritis caused by the deposition of monosodium urate crystals in synovial joints due to elevated serum urate (SU) levels [1], and the overall disease burden remains substantial and may be growing [2]. There are several risk factors for gout, including obesity [3], hypertension [4] and dietary behaviours [5], such as alcohol consumption and purine-rich foods.

As a disease where disruption of purine metabolism is the main pathway, reducing urate synthesis and promoting urate excretion are effective approaches to prevent and treat gout [1]. Thus, several available drugs have been used for the treatment of gout [6]. However there is usually an inherent risk of drug toxicity, interactions and polypharmacy in patients with multiple comorbidities [7], which manifests itself in a number of side effects such as liver damage and nephrotoxicity [8, 9]. Therefore, it is necessary to find some complementary strategies to treat gout. To some extent, dietary interventions have been shown to alleviate the symptoms of hyperuricaemia [7]. Plant-based functional ingredients, such as polyphenols and alkaloids, have also been introduced into the diet through vegetables, fruits, coffee, tea, etc. [10, 11], which may be beneficial for patients with hyperuricaemia.

Tea and coffee are broadly consumed beverages worldwide [12]. Given this widespread use, the public health effects they bring are worthy of attention. Tea and coffee intake are closely related, with ∼70% of participants consuming both [13]. On the other hand, they are distinct beverages with overlapping components, such as caffeine, as well as different biologically active components, which may have diverse roles in the development of gout [14]. Indeed, a few previous studies have found inverse associations between coffee consumption and serum urate levels as well as gout risk [15–17]. However, despite animal studies suggesting that tea consumption lowers serum urate [7, 18], epidemiological evidence between tea consumption, serum urate levels and gout risk remain inconsistent [16, 19, 20].

Hence, using multidimensional cohort data from the UK Biobank, we prospectively assessed the dose–response associations of tea and coffee intake with gout risk. Specifically, we also explored whether these associations were modified by hyperuricaemia at baseline, as well as investigating their joint associations with gout.

Methods

Study design and population

This prospective study used data from the UK Biobank [21]. The UK Biobank was approved by the Northwest Multicentre Research Ethics Committee, and that approval covers this research. All participants in the UK Biobank provided written informed consent. For the present analyses, 54 764 participants were excluded, including 11 466 with prevalent gout, 38 702 with prevalent cancer, and 4,596 with missing data on tea and coffee intake, which left 447 658 participants for the current study.

Exposure assessment

Tea and coffee consumption were assessed at baseline using a touchscreen questionnaire. Participants were asked ‘How many cups of tea do you drink each day (including black and green tea)?’ or ‘How many cups of coffee do you drink each day (including decaffeinated coffee)?’ They selected either the number of cups, ‘<1’, ‘Do not know’, ‘Prefer not to answer’, or a specific number per day. Participants were asked to confirm their responses if they reported drinking >10 cups daily.

Incident gout

Gout was determined based on information (ICD-10 codes: M10.0, M10.2, M10.3, M10.4, M10.9) from hospital admission records containing admission and diagnosis data obtained from hospital event statistics in England, morbidity record data in Scotland, and patient event databases in Wales.

Covariates

We used the baseline questionnaire to assess the following potential confounders: age, sex, ethnicity, qualifications, Townsend deprivation index (TDI), income (less than £18 000, £18 000–£30 999, £31 000–£51 999, £52 000–£100 000, and >£100 000), BMI (kg/m²), smoking status (never, former and current), drinking status (never, former and current), physical activity (MET-h/week), sleep duration (h/day), intake of processed meat, fruits, vegetables and fish, tea and coffee consumption (cups/day), prevalent hypertension, prevalent diabetes, high-density lipoprotein (HDL), low-density lipoprotein (LDL), serum urate level at baseline, diuretic use, use of anti-hypertensive drugs, use of anti-hyperlipidemic drugs and use of anti-diabetic drugs. The TDI is an area-based proxy measure for socioeconomic status provided in the UK biobank study, and negative TDI values indicated relative affluence. A healthy diet score was calculated based on the following diet factors: vegetable intake ≥3 servings/day; fruit intake ≥3 servings/day; fish intake ≥2 servings/week; and unprocessed red meat intake ≤2 servings/week. Each point was given for each favourable diet factor, and the healthy diet score ranged from 0 to 4. Diabetes was defined according to medical records (ICD-10 codes: E10-14), glycosylated haemoglobin ≥6.5%, fasting blood glucose ≥7.0 mmol/l, self-reported or use of anti-diabetic drugs. Hypertension was defined on the basis of medical records (ICD-10 codes: I10-13), systolic blood pressure (SBP) ≥140 mmHg or diastolic blood pressure (DBP) ≥90 mmHg, self-reported or on anti-hypertensive medications.

Statistical analyses

Baseline characteristics of 447 658 participants were described across tea and coffee intake as percentages for categorical variables and means (s.d.) for continuous variables. The duration of follow-up was calculated from the baseline date to the diagnosis at the date of the outcome, death, loss of follow-up, or on 30 June 2022, whichever occurred first. To analyse the association between tea and coffee (overall, decaffeinated and caffeinated) consumption categories and new onset gout, we defined tea and coffee consumption into the following categories: 0, 0.5–1, 2–3, 4–6 and >6 cups/day. We used Cox proportional hazard models to estimate the associations of tea and coffee consumption categories with the incidence of gout. The proportional hazards assumptions of the Cox model were tested using the Schoenfeld residual method and were satisfied. Restricted cubic spline models were used to evaluate the dose–response relationships between tea, coffee and incident gout, with four knots at the 25th, 50th, 75th and 95th centiles. The covariates adjusted by the model were as described above. Additionally, we examined the multiplicative interaction between tea or coffee consumption and other factors by incorporating the two variables and their cross-product terms in the same model. The joint associations between coffee consumption and the status of hyperuricaemia on gout risk were assessed by creating dummy variables based on the combined exposures to both factors.

We performed five sensitivity analyses to assess the robustness of our study results: censoring participants when the event of interest was observed within the first 2 years of follow-up; baseline diabetic patients were excluded; the models were adjusted using the healthy diet scores described above instead of the independent diet covariates; limiting the analysis to participants with non-hyperuricaemia at baseline; and performing the analyses with additional detailed adjustments for alcohol consumption (daily or almost daily, three or four times a week, once or twice a week, 1–3 times a month, special occasions only and never drinkers). Last, the E-value was estimated to examine the magnitude of an unmeasured confounding factor that could affect the association between tea or coffee consumption and gout by random chance [22].

Statistical tests were two-tailed, and α levels of 0.05 were considered statistically significant. Analyses were performed using R, version 4.2.1.

Results

The sample included 447 658 participants with a median follow-up time of 13.42 years. Of these, a total of 3,053 incident gout events occurred during the follow-up period. The baseline characteristics of the participants are provided in Table 1. The mean age was 56.20 ± 8.13 years and 244 655 (54.65%) participants were females. In total, 65 975 (14.74%) participants were non-tea drinkers, and 99 465 (22.22%) were non-coffee drinkers.

Tea and coffee with gout risk

We classified tea and coffee intake into the following categories: 0, 0.5–1, 2–3, 4–6 and >6 cups/day, and examined the association of each category with gout risk (Fig. 1). In adjusted multivariate Cox models, we found that higher tea and coffee consumption was associated with a lower risk of gout. Drinking >6 cups of tea reduced the risk of gout by 23% (95% CI, 9%, 34%) compared with non-tea drinkers, while drinking >6 cups of coffee was associated with a 40% (95% CI, 23%, 53%) reduction compared with non-coffee drinkers (Table 2). Our analyses in further classifying coffee according to the absence or presence of caffeine rendered similar results (Supplementary Table S1, available at Rheumatology online). Drinking >6 cups of decaffeinated coffee or caffeinated coffee per day was associated with 52% (95% CI, 4%, 76%) and 38% (95% CI, 20%, 53%) lower risk of gout, respectively, compared with non-consumers.

Nonlinear association

Restricted cubic spline models were used to evaluate the relationship between tea and coffee with gout. In the multi-adjusted models, nonlinear associations of tea (P for nonlinear = 0.004) and coffee (P for nonlinear <0.0001) with gout were observed (Fig. 2). Tea consumption of six and coffee intake of three cups/day were associated with the lowest hazard ratio (HR) for incident gout, with a relatively rapid decrease in risk before the nadir but a slower reduction if drinking more. Similar nonlinear associations of decaffeinated (P for nonlinear = 0.0004) or caffeinated (P for nonlinear <0.0001) coffee intake with a reduced risk of gout were also documented (Supplementary Figs S1 and S2, available at Rheumatology online).

Association of tea and coffee with serum urate

We hypothesized that the association between tea and coffee intake and reduced risk of gout was in part through serum urate concentrations, so that we performed further cross-sectional dose–response analyses. As a result, nonlinear associations of tea (P for nonlinear <0.0001) and coffee (P for nonlinear <0.0001) with serum urate concentrations were observed in multi-adjusted models (Supplementary Fig. S3, available at Rheumatology online). We observed that increased tea or coffee consumption was significantly associated with lower serum urate concentrations. Similar results were also seen in the analyses of decaffeinated and caffeinated coffee (Supplementary Fig. S4, available at Rheumatology online).

Subgroup analyses

We evaluated the association between categorized tea or coffee consumption and gout in subgroups (Supplementary Table S2, available at Rheumatology online). We found that the strength of the association between tea or coffee and gout was comparable across age, sex, BMI, smoking and drinking status, sleep duration and physical activity level groups (P for interaction>0.05). Interestingly, the test revealed that the association between coffee and gout was modified by the presence of hyperuricaemia (P for interaction <0.05), rather than tea and gout (P for interaction = 0.705). In parallel, we analysed the effect modification between tea and coffee intake of interest to the study without finding it to be statistically significant (P for interaction = 0.169). Nevertheless, in a further joint association analysis, we found that participants who consumed ≥4 cups/day of both tea and coffee had a reduced risk of gout by ∼50% (HR, 0.52; 95% CI, 0.42, 0.64) compared with non-coffee consumers who drank <4 cups/day of tea (Supplementary Fig. S5, available at Rheumatology online).

Coffee consumption, hyperuricaemia and incident gout

Based on the findings above, we further finely divided coffee intake into 0, 0.5–1, 2–3 and ≥4 categories to explore the interaction between coffee and hyperuricaemia status (Table 3). As a result, the protective role of coffee appears to be stronger in non-hyperuricemia individuals (Supplementary Fig. S6, available at Rheumatology online), resembling in the analyses of coffee with and without caffeine (Supplementary Table S3, available at Rheumatology online). Additionally, we tested the joint association between coffee, hyperuricaemia status and gout (Fig. 3), consequently finding an HR of 0.34 (95% CI, 0.28, 0.40) for those with ≥4 cups/day of coffee intake and free of hyperuricaemia compared with non-coffee drinkers with hyperuricaemia.

Additional and sensitivity analyses

Censoring participants when the event of interest occurred within the first 2 years of follow-up, furtherly excluding baseline diabetic patients, adjusted the models using the healthy diet scores instead of the independent diet covariates and other sensitivity analyses mirrored the results of the primary analysis (Supplementary Table S4, available at Rheumatology online). Compared with non-consumers, the E values for the association of >6 cups of tea and coffee intake per day with new-onset gout ranged from 1.90–2.36, and 2.01–3.57, respectively, in the primary and sensitivity analyses (Supplementary Table S5, available at Rheumatology online).

Discussion

In this large prospective cohort study, we assessed the dose–response association of tea or coffee (overall, decaffeinated and caffeinated) intake with gout and explored whether these associations were modified by hyperuricaemia at baseline. It was found that there was a nonlinear association between tea or coffee consumption and gout, with significant gains in gout risk reduction from as few as six cups of tea or three cups of coffee per day. Drinking ≥6 cups/day of tea or coffee was associated with a much lower gout risk of 23% or 40%, respectively. The association of coffee with gout was modified by baseline hyperuricaemia status, as demonstrated by the decreased risk being stronger in non-hyperuricemic individuals.

Our findings agreed with the previously reported inverse association between coffee consumption and the risk of incident gout among women [15] and men [17], although their adjusted results were not adjusted for tea consumption. Unlike coffee, the epidemiological evidence for association between tea and gout appears inconsistent, with some studies indicating that tea consumption is not associated with the risk of gout or hyperuricaemia [15–17], and others suggesting the hypouricemic effect of tea but not a reduction in the risk of either [23]. As such, randomized controlled trials and well-designed prospective studies to describe this association are needed [19]. Based on these findings, we used a prospective cohort design with a relatively large sample size, controlling for potential confounders to explore the dose–response association between tea drinking and gout, which may be useful in clarifying the relationship.

Although the detailed mechanisms are inconclusive, several potential pathways have been proposed to explain the associations between tea or coffee intake and the reduced risk of gout. Coffee is a major source of caffeine, a methylxanthine that has been shown to competitively inhibit xanthine oxidase in rats [24], thereby reducing serum urate concentrations. Coffee intake may also reduce gout risk through non-caffeine components in other mechanisms, such as reducing inflammation [15], oxidative stress [25], glucose and insulin levels [26], and insulin resistance [27], which may partially explain the hypouricemic effects of tea and decaffeinated coffee. Indeed, in combination with caffeine, coffee is also a major source of phenolic chlorogenic acid, a strong antioxidant that also improves insulin sensitivity and lowers blood glucose concentrations [26]. In addition, decaffeinated coffee appears to delay intestinal absorption of glucose and increase glucagon-like peptide 1 concentrations in a human intervention study [28], while glucagon-like peptide 1 is known for its beneficial effects on glucose-induced insulin secretion and insulin action [29]. The health benefits of tea are mainly attributed to bioactive compounds such as polyphenols, alkaloids, pigments and free amino acids [19], which account for 18–36%, 3–5%, 0.3–2% and 2–4% of the dry weight of tea leaves, respectively [30, 31]. Animal studies have suggested that these tea extracts can inhibit xanthine oxidase [7], are anti-inflammatory [32], and have good hypoglycaemic effects [33], which play key roles in the development of gout [19].

An interesting finding in our study was that the inverse association between coffee and gout was stronger in non-hyperuricemic patients than in patients with hyperuricaemia, either in the full type, decaffeinated, or caffeinated one. The causal association between serum urate and gout has been extensively studied [34], and the state of hyperuricaemia triggers inflammatory responses [35], oxidative stress, and endocrine dysfunction [36], which may diminish the gout risk-reducing function of beverages that may target serum urate as their primary target, such as coffee. Nevertheless, further exploration is needed to obtain a clearer picture of the underlying mechanisms.

The strengths of the current study include a large sample size and prospective study design, which may make the association results more convincing. Notably, we investigated the dose–response relationship between drinking tea or coffee and reduced risk of gout, such that common health behaviour interventions may benefit a larger population from a public health perspective. There are still several limitations that should be noted. First, this is an observational study so the association cannot be interpreted as causal. Second, tea and coffee intakes are self-reported at baseline, which may not reflect long-term consumption patterns, although intake of both may remain somewhat stable for a long period of time in European adults [37]. Future studies with repeated measurements may be helpful in addressing this issue. Third, if the associations we found are likely to be true, we must also acknowledge that the gout risk-reducing effect of tea may be weaker than that of coffee, as demonstrated by the consumption frequency of >4 cups of tea vs <3 cups of coffee per day. Fourth, similar to most observational studies, the bias that may be caused by unknown and unmeasured confounding factors remains. Finally, most of the UK Biobank participants were white British, which might preclude the generalizability of the findings to the general population but does not affect the internal validity of the study.

Supplementary material

Supplementary material is available at Rheumatology online.

Data availability

This research was conducted using the UK Biobank resource under the project ID 66137. Raw data used for this study are available from the UK Biobank resource.

Contribution statement

H.G. conceived and designed the study and performed statistical analyses. J.H. and T.W. managed and curated the UK Biobank dataset. All authors interpreted the results. H.G. drafted the manuscript. All authors edited the manuscript for intellectual content. All authors take responsibility for the integrity of the study.

Funding

This research was supported by Peking University Outstanding Discipline Construction Project on Epidemiology and Biostatistics (L.L., T.W.), Fujian provincial health technology project (Grant No. 2020CXB009, T.W.), the Natural Science Foundation of Fujian Province, China (Grant No. 2021J01352), National Key Research and Development Program of China (2020YFC2002900, J.H.), and the China Postdoctoral Science Foundation (Grant No. BX2021021; Grant No. 2022M710249, M.W.).

Disclosure statement: The authors have declared no competing interests.

Acknowledgements

We appreciate the participation of the hundreds of thousands of individuals in the UK Biobank study who made this study possible.

References

Author notes