Background: A healthy diet may be protective against cognitive decline by mechanisms that involve improved vascular risk factors such as hypertension and dysglycemia, and reduced systemic inflammation. In this population-based study, we hypothesized that midlife diet pattern would be associated with cognitive decline over 21-years.
Methods: This study included 13,603 participants in the ARIC population-based cohort recruited from four U.S. sites who were aged 45 to 64 at baseline (1987-89) when diet was measured. Participants recorded diet using a 66-item food frequency questionnaire. Two dietary patterns, called “Meat and Fried” and the “Balanced Diet”, were named after the most representative foods that emerged from constructs derived from a principal component analysis of 30 food groups. A higher diet pattern score represented greater adherence. Cognitive testing, including the digit symbol substitution, the word fluency and delayed word recall tests, were combined to a z-score at each visit (visits 2, 1990-92; 4, 1996-98 and 5, 2011-2013). Test scores for participants not attending subsequent visits were imputed using Multiple Imputation by Chained Equations to account for cohort attrition. Cognitive performance at visit 2 was compared by tertile (T) of each diet pattern. Using mixed effects models with a random slope and intercept , we determined the 21-year change in cognitive function by diet pattern tertile, adjusting for demographics and medical history.
Results: At visit 2, adherence to the Meat and Fried pattern was associated with lower cognitive test scores (z-score T3: -0.172, SD 0.985; T1: 0.149, SD 0.981, p-trend <0.001). Adherence to the Balanced Diet was not associated with differences in cognitive performance (z-score T3: 0.013, SD 0.988; T1 -0.036, SD 1.001, p-trend 0.10). 21-year change in cognitive function did not differ by adherence to diet pattern with adjustments (difference of the change in z-score for Meat and Fried, T3 vs. T1: 0.02, [CI -0.05 to 0.08]; Balanced Diet T3 vs. T1: -0.03, [CI -0.09 to 0.02]).
Conclusion: Although participants with a diet pattern high in meat and fried foods had lower cognition at time of first assessment, diet patterns at midlife did not carry independent associations with cognitive decline.