We randomly assigned 23 older adults with mild cognitive impairment to either a high carbohydrate or very low carbohydrate diet. Following the 6-week intervention period, we observed improved verbal memory performance for the low carbohydrate subjects (p = 0.01) as well as reductions in weight (p < 0.0001), waist circumference (p < 0.0001), fasting glucose (p = 0.009), and fasting insulin (p = 0.005). Level of depressive symptoms was not affected. Change in calorie intake, insulin level, and weight were not correlated with memory performance for the entire sample, although a trend toward a moderate relationship between insulin and memory was observed within the low carbohydrate group. Ketone levels were positively correlated with memory performance (p = 0.04). These findings indicate that very low carbohydrate consumption, even in the short term, can improve memory function in older adults with increased risk for Alzheimer's disease. While this effect may be attributable in part to correction of hyperinsulinemia, other mechanisms associated with ketosis such as reduced inflammation and enhanced energy metabolism also may have contributed to improved neurocognitive function. Further investigation of this intervention is warranted to evaluate its preventive potential and mechanisms of action in the context of early neurodegeneration.
Copyright © 2012 Elsevier Inc. All rights reserved.