Objectives: To investigate the effect of a medical food (Souvenaid) on body mass index (BMI) and functional abilities in patients with mild Alzheimer's disease (AD). Design/setting/participants/intervention /measurements: These analyses were performed on data from a 12-week, double-blind, randomized, controlled, multicenter, proof-of-concept study with a similarly designed and exploratory 12-week extension period. Patients with mild AD (Mini-Mental State Examination score of 20-26) were randomized to receive either the active product or an iso-caloric control product While primary outcomes included measures of cognition, the 23-item Alzheimer's Disease Cooperative Study-Activities of Daily Living (ADCS-ADL) scale was included as a secondary outcome. Both ADCS-ADL and BMI were assessed at baseline and Weeks 6, 12 and 24. Data were analyzed using a repeated-measures mixed model. Results: Overall, data suggested an increased BMI in the active versus the control group at Week 24 (ITT: p = 0.07; PP: p = 0.03), but no treatment effect on ADCS-ADL was observed. However, baseline BMI was found to be a significant treatment effect modifier (ITT: p = 0.04; PP: p = 0.05), and an increase in ADCS-ADL was observed at Week 12 in patients with a 'low' baseline BMI (ITT: p = 0.02; PP: p = 0.04). Conclusions: These data indicate that baseline BMI significantly impacts the effect of Souvenaid on functional abilities. In addition, there was a suggestion that Souvenaid increased BMI. © 2011 Serdi and Springer Verlag France.