BACKGROUND: Alzheimer's disease (AD) is a progressive neurodegenerative disorder that primarily strikes the elderly. Studies in both humans and animal models have linked the consumption of cholesterol and saturated fats with amyloid-beta (Abeta) deposition and development of AD. Yet, these studies did not examine high fat diets in combination with reduced carbohydrate intake. Here we tested the effect of a high saturated fat/low carbohydrate diet on a transgenic mouse model of AD. RESULTS: Starting at three months of age, two groups of female transgenic mice carrying the "London" APP mutation (APP/V717I) were fed either, a standard diet (SD) composed of high carbohydrate/low fat chow, or a ketogenic diet (KD) composed of very low carbohydrate/high saturated fat chow for 43 days. Animals fed the KD exhibited greatly elevated serum ketone body levels, as measured by beta-hydroxybutyrate (3.85 +/- 2.6 mM), compared to SD fed animals (0.29 +/- 0.06 mM). In addition, animals fed the KD lost body weight (SD 22.2 +/- 0.6 g vs. KD 17.5 +/- 1.4 g, p = 0.0067). In contrast to earlier studies, the brief KD feeding regime significantly reduced total brain Abeta levels by approximately 25%. Despite changes in ketone levels, body weight, and Abeta levels, the KD diet did not alter behavioral measures. CONCLUSION: Previous studies have suggested that diets rich in cholesterol and saturated fats increased the deposition of Abeta and the risk of developing AD. Here we demonstrate that a diet rich in saturated fats and low in carbohydrates can actually reduce levels of Abeta. Therefore, dietary strategies aimed at reducing Abeta levels should take into account interactions of dietary components and the metabolic outcomes, in particular, levels of carbohydrates, total calories, and presence of ketone bodies should be considered.