Objective: To investigate the extent that baseline cognitive impairment and intake adherence affected the 13-item Alzheimer's Disease Assessment Scale - cognitive subscale (ADAS-cog) intervention response of a medical food in Alzheimer's Disease (AD) patients. Desigrt/setting/participants /intervention/measurements: This analysis was performed on data from a proof-of-concept study, consisting of a 12-week, double-blind, randomized, controlled, multicenter trial, followed by a similarly designed 12-week extension study. Patients with mild AD (Mini-Mental State Examination [MMSE] score of 20-26) were randomized to receive active or control product as a 125 ml daily drink. One of the co-primary outcome measures was the 13-item ADAS-cog. In this analysis, the study population was divided into two subgroups: patients with 'low' baseline ADAS-cog scores (<25.0) and patients with 'high baseline ADAS-cog scores (≥25.0). Repeated Measures Models (RMM) were used to determine the relationship between ADAS-cog score and intervention. Results: A significant treatment effect (F[1,319]=4.0, p=0.046) was shown in patients with 'high baseline ADAS-cog, but not in patients with 'low' baseline ADAS-cog (F[1,250]= 1.25, p=0.265). Overall, intake adherence was significantly correlated with ADAS-cog improvement in the active product group (correlation coefficient=-0.260; p=0.019), but not the control group. Conclusion: These data indicate that baseline ADAS-cog significantly influenced the effect of Souvenaid intervention on ADAS-cog outcome. A higher intake of active study product was also associated with greater cognitive benefit. These findings highlight the potential benefits of Souvenaid in AD patients and warrant confirmation in larger, controlled studies. © 2011 Serdi and Springer Verlag France.