Abstract
Objective: Existing research suggests walnut intake may be associated with better cognitive function in older adults, yet few studies utilise longitudinal data from observational studies of ageing populations. Our objective was to estimate the association between whole walnut intake and cognitive change in a representative sample of older Americans. Design: Secondary analysis of the Health and Retirement Study and Health Care and Nutrition Study. Walnut consumption was defined as a categorical measure (none, low intake (0·01–0·08 1 oz. servings per day) and moderate intake (>0·08 1 oz. servings per day)) and cognitive function was measured using the Telephone Interview for Cognitive Status. Latent growth modelling estimated the association between walnut consumption and trajectories of cognitive status over a 4-year observational period. Sensitivity analyses assessing non-random dropout and Monte Carlo power analyses were conducted to contextualise results. Setting: The USA. Participants: A sample of 3632 US adults aged 65 years and older. Results: Those reporting any walnut consumption had greater cognitive scores at baseline than those not consuming walnuts (low walnut consumption, b = 1·53, se = 0·21, P < 0·001; moderate walnut consumption, b = 2·22, se = 0·27, P < 0·001), but walnut consumption was not associated with cognitive change. Walnut consumption was positively associated with socioeconomic status and health behaviours as well as intake of nutrients identified to have neuroprotective benefits. Conclusions: We identified an association between walnut consumption and cognitive function in older adults, although we did not find that walnut consumption was protective against age-related cognitive decline.