Adherence to predefined dietary patterns and risk of developing type 2 diabetes in the Canadian adult population

OBJECTIVE: Diet quality indices are increasingly being used in epidemiologic research. However, no studies have addressed whether adherence to Canadian dietary guidelines is longitudinally associated with decreased risk of type 2 diabetes in a population-based sample. The objective of this study is to examine the association between the Healthy Eating Index (HEI) and incident type 2 diabetes in the Canadian population. METHODS: We used data from Ontario respondents to the 2004 Canadian Community Health Survey linked to health administrative data (n=4,755). Adherence to the HEI was analyzed with a 24-hour dietary recall. Type 2 diabetes was ascertained through the Ontario Diabetes Database, and tracked up to 12.1 years from baseline. Cox proportional hazards models were used to estimate type 2 diabetes risk as a function of HEI score. Given obesity's potential role as a mediator, we explored the effects of removing body mass index from the final model. RESULTS: High HEI adherence was not associated with a reduction in diabetes risk overall (hazard ratio [HR], 0.97; 95% confidence interval, 0.62 to 1.50), nor in separate strata of men (HR, 0.94) or women (HR, 1.03). Additional adjustment for body mass index attenuated the multivariable adjusted hazard ratios toward the null. CONCLUSIONS: This is the first study to prospectively explore the relationship between adherence to the dietary recommendations of the HEI and diabetes risk in a representative, population-based sample. Our analyses challenge previous findings and highlight the utility of linked data to evaluate the role of healthy dietary patterns in relation to population-level morbidity