BACKGROUND: There is growing evidence that the Mediterranean (Medi) diet may lower the risk of type 2 diabetes mellitus (T2DM). Whether this association is due to the Medi diet by itself or is mediated by a diet-associated lower rate of overweight is uncertain. Our aim was to disentangle these relationships among UK adults. METHODS: Based on 21 585 participants from the UK Biobank cohort, the adherence to the Medi diet (high fruits, vegetables, legumes, cereals, fish, olive oil; low meat, dairy products; and intermediate alcohol intakes) was assessed (range 0-18). Data on diabetes were self-reported, and overweight was defined as a body mass index (BMI) ≥ 25 kg/m². A mediation analysis was implemented to disentangle the role of overweight in the Medi diet-T2DM relationship. RESULTS: The average baseline Medi diet score was 8.8 [standard deviation (SD) 2.6]. During a mean follow-up of 6.1 years, 473 individuals developed T2DM. A higher adherence to a Medi diet (+1 point) was associated with 14% decreased risk of T2DM [hazard ratio (HR): 0.86, 95% confidence interval (CI): 0.82-0.90]. This association split into an indirect effect of 10%, mediated by lower odds of overweight (HR: 0.90, 95% CI: 0.87-0.92), and a direct effect of the Medi diet of 4% (HR: 0.96, 95% CI: 0.93-0.99), regardless of the effect mediated by overweight. CONCLUSIONS: Considered as a single mediator, reduced overweight mainly contributes to the association between greater Medi diet adherence and lower risk of T2DM on this British subsample. However, the direct effect of the diet on the risk of T2DM, even weaker, should not be overlooked.< Réduire