Résumé
Few studies have investigated the long-term impact of overall dietary patterns (DP) on cognition. We evaluated the association between empirically derived DP in midlife and cognitive performance 13 y later. Dietary data were based on 24-h dietary records obtained from a subsample of the Supplementation en Vitamines et Mineraux Antioxydant Study. Cognitive performance was assessed via a battery of neuropsychological tests that included verbal fluency, the RI-48 cued recall test, the trail-making test, and forward and backward digit span. Three composite variables, for global cognitive function, verbal memory, and executive functioning, were built. The multivariate analyses were adjusted for baseline characteristics (age, gender, intervention group, education, alcohol and energy intake, number of dietary records, physical activity, BMI, tobacco use, self-reported memory troubles, diabetes, hypertension, and, for women, menopausal status and hormone therapy use), follow-up time, history of cardiovascular disease, and depressive symptoms. Adjusted means +/- SEM of composite variables across quartiles (04 vs. 01) of DP were estimated using ANCOVA. A healthy and a traditional DP were identified. In the multivariate model, the healthy pattern was associated with better global cognitive function (50.1 +/- 0.7 vs. 48.9 +/- 0.7; P-trend = 0.001) and verbal memory (49.7 +/- 0.4 vs. 48.7 +/- 0.4; P-trend = 0.01). These relationships were stronger in participants scoring below the gender-specific median values for energy intake (<2490 kcal for men and <1810 for women) than in those scoring at or above those values. Adherence to a healthy DP in middle life may help preserve global cognitive function, especially verbal memory, when total energy intake is regulated. J. Nutr. 142: 909-915, 2012.