Cumulative average dietary pattern scores in young adulthood and risk of incident type 2 diabetes: the CARDIA study

Aims/hypothesis: The evidence for the role of contemporary dietary patterns, trends and predominant aspects of energy intake in a typical American diet and in type 2 diabetes risk is limited. Therefore, we examined the association between dietary pattern scores created to reflect the 2015 Dietary Guidelines for Americans (DGA) Scientific Report, a Palaeolithic (Palaeo) diet, a diet high in ‘empty calories’, and the A Priori Diet Quality Score (APDQS) (cohort reference) and type 2 diabetes risk over time. Methods: We carried out a prospective analysis of 4719 young adult black and white men and women from the Coronary Artery Risk Development in Young Adults (CARDIA) study with repeated dietary histories collected at study years 0, 7 and 20. Using multivariable Cox proportional hazards regression models, we examined the association between time-dependent cumulative average dietary pattern scores and incident type 2 diabetes. Results: During the 30 year follow-up period, 680 (14.4%) incident cases of type 2 diabetes occurred. There was no association between the 2015 DGA, Palaeo or empty calorie scores and type 2 diabetes risk in the overall population. Participants in the fourth quartile of the APDQS, reflecting a more healthful dietary pattern, had a 45% lower risk of type 2 diabetes compared with those in the lowest quartile (HR 0.55 [95% CI 0.41, 0.74]). In stratified analyses there was an inverse association for the 2015 DGA in non-smokers per SD (HR 0.86 [95% CI 0.74, 0.99]) and an inverse association for the empty calorie score in white women (HR 0.76 [95% CI 0.60, 0.96]) as well as in a subgroup analysis of the Palaeo index of participants who maintained a high score over 20 years (per SD, HR 0.59 [95% CI 0.39, 0.88]). Conclusions/interpretation: Higher levels of the APDQS, which largely aligns with the 2015 DGA, were strongly inversely associated with 30 year type 2 diabetes risk in the CARDIA cohort; the results from the other patterns were nuanced and need to be considered in the context of the study and potential biases.