Dose-Response Associations Between Diet and Risk of Rheumatoid Arthritis: A Meta-Analysis of Prospective Cohort Studies

Abstract

To provide a systematic and quantitative summary of dietary factors and rheumatoid arthritis (RA) risk. A systematic review and meta-analysis included prospective cohort studies from 2000 to 2024 reporting relative risks (RRs) with 95% confidence intervals (CIs) for RA incidence relating to 32 different dietary exposures. Linear and non-linear dose-response analyses were conducted. Thirty studies were included, involving 2,986,747 participants with 9,677 RA cases. Linear dose-response analysis suggested that each 2-unit per week increase in total alcohol intake was linked to 4% risk reduction (RR (95%-CI), heterogeneity (I²), NutriGrade score: 0.96 (0.94, 0.98), 58%, moderate certainty), and beer consumption was associated with a 10% reduction per 2 units/week increase (0.90 (0.84, 0.97), 0%, very low certainty). Each 2-unit/week increase in total alcohol intake was associated with a 3% decrease in seropositive RA risk (0.97 (0.96, 0.99), 28%, moderate certainty). Increased intakes of fruit (per 80 g/day) and cereals (per 30 g/day) were associated with 5% (0.95 (0.92, 0.99), 57%, moderate certainty) and 3% (0.97 (0.96, 0.99), 20%, moderate certainty) reduced risk, respectively. Conversely, tea consumption showed a 4% increased risk per additional cup/day (1.04 (1.02, 1.05), 0%, moderate certainty). Non-linear associations were observed for total coffee, vegetables, oily fish, and vitamin D supplementation. Data on dietary patterns and specific micronutrients were limited. The findings suggest that moderate alcohol consumption and a higher intake of fruits, oily fish, and cereals are associated with a reduced risk of RA, while tea and coffee may be linked to an increased risk. Optimising dietary intake of certain food components may reduce RA risk, despite moderate-quality evidence.

Keywords: diet; dose–response; meta-analysis; nutrient intake; rheumatoid arthritis.

Figure 1

PRISMA flow chart summarising article retrieval and screening process.

Figure 2

The summary of pooled HRs (95% CIs) for different dietary exposures and incident RA comparing the highest with the lowest categories. (A) Sankey diagram illustrating the 21 food groups, beverages, and 11 nutrients across 28 included cohort studies. (B) OncoPrint chart displaying the pooled hazard ratios (HRs) for incident RA, incident seropositive RA, and incident RF-positive RA, according to 37 food components and nutrients. Green colour represents a beneficial effect on RA incidence; red colour represents a detrimental effect on RA incidence. This figure does not indicate statistically significant results.

Figure 3

Meta-analysis of alcohol consumption and risk of RA comparing the highest with the lowest categories. Diamonds represent pooled estimates from random effects analysis.

Figure 4

Non-linear dose–response association of food component consumption with risk of RA using restricted cubic splines. Relations for food components of (A) total alcohol, (B) total coffee, (C) sugar-sweetened soda, (D) fruits, (E) vegetables, and (F) oily fish. Solid lines represent the fitted non-linear trend and blue shade represents the pointwise 95% CIs.