Epigenetic and genetic factors modifying the association between diet quality and incident type 2 diabetes: the EPIC-Potsdam cohort

Abstract Background Benefits of healthy diets for type 2 diabetes (T2D) prevention are established. It remains unclear whether such recommendations should be tailored to specific subgroups according to their susceptibility to the disease. This study aimed to assess whether the association of diet quality with T2D risk differs across subgroups with different genetic or epigenetic susceptibility. Methods We used data from a case-cohort within the European Prospective Investigation into Cancer and Nutrition (EPIC)–Potsdam cohort (random subsample with genetic data of 2204 participants, 750 verified incident T2D cases; random subsample with epigenetic data of 1065 participants, 676 verified incident T2D cases). The Alternative Healthy Eating Index (AHEI-2010), Dietary Approaches to Stop Hypertension (DASH), Dietary Inflammatory Index (DII) and the Mediterranean Diet Pyramid score (MedPyr) were used to assess diet quality. A blood DNA methylation risk score (MRS) was used to reflect epigenetic risk. Genetic risk was characterized using global polygenic risk scores (PRS) and pathway-specific polygenic risk scores (pPRS). Cox proportional hazards models were used to assess the modification of the diet quality-diabetes risk association by the risk scores. Results Higher adherence to the AHEI-2010 was associated with a lower risk of T2D, while higher MRS, PRS and pPRS predictably indicated higher risk in this population. We detected significant modification of the diet quality-T2D association by the MRS for AHEI-2010 ( p = 0.008) and MedPyr ( p < 0.0001) and DII ( p = 0.010) but only the AHEI-2010 × MRS interaction was robust across all sensitivity analyses. However, we did not find evidence of interaction between diet quality and genetic risk. Conclusions The associations of healthy diets with T2D could depend on the diabetes risk captured by blood DNA methylation profiles, but they do not seem to depend on the genetic predisposition for T2D. Confirmation of the effect modification by the MRS in independent populations is further required before considering clinical application.