Objective: To systematically review prospective cohort studies about the association between dietary patterns and type 2 diabetes mellitus (T2DM) incidence, and to quantify the effects using a meta-analysis.
Design: Databases such as PubMed, ISI Web of Science, SCOPUS and Google Scholar were searched up to 15 January 2015. Cohort studies which tried to examine the association between empirically derived dietary patterns and incident T2DM were selected. The relative risks (RR) and their 95 % confidence intervals for diabetes among participants with highest v. lowest adherence to derived dietary patterns were incorporated into meta-analysis using random-effects models.
Results: Ten studies (n 404 528) were enrolled in the systematic review and meta-analysis; our analysis revealed that adherence to the 'healthy' dietary patterns significantly reduced the risk of T2DM (RR=0·86; 95 % CI 0·82, 0·90), while the 'unhealthy' dietary patterns adversely affected diabetes risk (RR=1·30; 95 % CI 1·18, 1·43). Subgroup analysis showed that unhealthy dietary patterns in which foods with high phytochemical content were also loaded did not significantly increase T2DM risk (RR=1·06; 95 % CI 0·87, 1·30).
Conclusions: 'Healthy' dietary patterns containing vegetables, fruits and whole grains can lower diabetes risk by 14 %. Consuming higher amounts of red and processed meats, high-fat dairy and refined grains in the context of 'unhealthy' dietary patterns will increase diabetes risk by 30 %; while including foods with high phytochemical content in these patterns can modify this effect.
Keywords: Dietary patterns; Meta-analysis; Systematic review; Type 2 diabetes mellitus.