Abstract
OBJECTIVE: To determine whether tea consumption is associated with risk of type 2 diabetes (T2D), to elucidate potential effect modification by genetic susceptibility, and to examine metabolic biomarkers as potential mediators for the association.
METHODS: We conducted a cohort analysis of 382,946 participants in the UK Biobank. Multivariable-adjusted age-stratified Cox proportional hazards regression models were used to calculate hazard ratios (HRs) and 95% confidence intervals (CIs) for the association between tea consumption and T2D risk.
RESULTS: During a median follow-up of 14.0 years, 16,100 incident cases of T2D were documented. Greater tea consumption was associated with a modestly reduced risk of T2D. Compared with no tea consumption, the HR (95% CI) of T2D was 0.93 (0.89, 0.98), 0.86 (95% CI: 0.81, 0.90) and 0.87 (95% CI: 0.82, 0.92) for tea consumption of 2 to 3, 4 to 5, or 6+ cups per day, respectively (P for trend < 0.0001). The observed inverse association between tea consumption and T2D risk was generally consistent across cohort subgroups, including groups with different genetic susceptibility to T2D (P for interaction = 0.64). Mediation analyses estimated that 49.7% (95% CI: 37.1%, 62.4%), 11.2% (95% CI: 4.2%, 26.3%), 3.6% (95% CI: 1.5%, 8.5%), 5.3% (95% CI: 2.6%, 10.4%), and 2.8% (95% CI: 0.8%, 9.6%) of the inverse tea association may be effected through BMI, waist-hip ratio, non-high-density lipoprotein cholesterol, blood pressure, and C-reactive protein, respectively. On the other hand, consumption of artificially sweetened tea was positively related to risk of T2D.
CONCLUSIONS: Greater tea consumption was associated with decreased risk of T2D, independent of other risk factors and genetic predisposition. In contrast, consumption of artificially sweetened tea was positively associated with risk of T2D. Our findings provide evidence relevant to the primary prevention of T2D, supporting the potential of tea consumption as a component of a healthy diet.