Identification of change-points in the relationship between food groups in the mediterranean diet and overall mortality: an 'a posteriori' approach

Background: Adherence to Mediterranean diet has been shown to be associated with a better health and greater survival. The aim of the present study was to identify change-points in the relationship between food groups composing Mediterranean diet and overall mortality. Methods: The population of the Greek EPIC prospective cohort study (23,349 adult men and women in the Greek EPIC sample who had not previously been diagnosed as having cancer, coronary heart disease or diabetes mellitus at enrolment) was analysed. Segmented logistic regression analysis was conducted to examine the association between each of the food groups contributing to the Mediterranean diet score and overall mortality. Results: This analysis allowed the determination of the following change-points: among men: 1 change-point for vegetables, legumes, cereals, fish and seafood and dairy products and 2 change-points for fruit and nuts, meat and meat products and ethanol; among women: 1 change-point for legumes and fish and seafood and 2 change-points for the remaining food groups. These cut-off points were used to construct an 'a posteriori' score that may be better in capturing the health-promoting potential of the traditional Mediterranean diet. Conclusion: Identification of change-points in the relationship between components of the Mediterranean diet and mortality can be used to increase the discriminatory ability of a widely used Mediterranean diet score in relation to mortality.