Total sugar intake and cognition in community‐dwelling older adults

Background Impaired glucose metabolism has been implicated in the progression of Alzheimer’s disease. Data from the Rush Memory and Aging Project was used to examine the relationship between sugar intake and cognition. Method Data were included from 837 respondents in the Rush Memory and Aging Project with complete food frequency questionnaire data consisting of 144 items. Cognition was assessed using a battery of 19 cognitive tests. Missing data were imputed using multiple imputation with chained equations. Linear random effects regression models were used to examine the relationship between total calorie‐adjusted quintile of sugar intake and global cognition, working memory, semantic memory, episodic memory, perceptual orientation and perceptual speed. Linear random effect models included race/ethnicity, age, sex, and time interactions for education (years), total energy (kilocalories/day), alcohol intake (grams/week), high‐density lipoprotein cholesterol (HDL‐C) and a continuous base term for time. Cox proportional hazards models were used to examine the relationship between total sugar intake and hazard of Alzheimer’s dementia, and to explore potential effect measure modification and confounding by race/ethnicity, age, sex, HDL‐C, total energy, alcohol intake, education and history of diabetes and hypertension. Variables were considered confounders if their inclusion resulted in >10% change in the crude hazard ratio. Effect measure modification was examined by including a time interaction term in Cox models and a performing a Wald test. Result In adjusted linear random effect models (Table 1) continuous quintile of sugar intake was inversely related to global cognition (β=‐0.0073; SE=0.0019; p=<0.001), episodic memory (β=‐0.0078; SE=0.0026; p=0.0026), semantic memory (β=‐0.0096; SE=0.0023; p=<0.001), working memory (β=‐0.0056; SE=0.0023; p=0.015) and perceptual speed (β=‐0.0090; SE=0.0025; p<0.001). There were 200 cases of Alzheimer’s dementia over 5,098 person years of follow‐up (IR: 39.2 cases per 1,000 person‐years). In crude Cox proportional hazards models (Table 2), those in the fifth quintile of total sugar intake were 1.84 (95% CI: 1.18; 2.87) as likely to develop Alzheimer’s dementia as those in the first quintile. A significant linear trend relationship was observed (p=0.006). Conclusion There is some evidence to suggest that total sugar intake may adversely affect cognitive function over time, and it may increase risk of Alzheimer’s dementia.