Objectives: To assess the longitudinal effects of acetyl-L-carnitine (ALC) on patients diagnosed with Alzheimer's disease. Design: Longitudinal, double-blind, parallel-group, placebocontrolled. Setting: Twenty-four outpatient sites across the United States. Participants: A total of 334 subjects diagnosed with probable Alzheimer's disease by NINCDS-ADRDA criteria. These data were originally reported by Thal and colleagues (1996). Measurements: Cognitive subscale of the Alzheimer Disease Assessment Scale (ADAS) given every 3 months for 1 year. Results: The average rate of change was estimated using the trilinear approach, which allows for periods of both change and stability. Both the ALC group and the placebo group exhibited the same mean rate of change on the ADAS (0.68 points/month). However, a multiple regression analysis revealed a statistically significant Age × Drug interaction characterized by younger subjects benefiting more from ALC treatment than older subjects. Further analyses suggested that the optimal, though not statistically significant, cutpoint for ALC benefit was 61 years of age. Conclusions: ALC slows the progression of Alzheimer's disease in younger subjects, and the use of the trilinear approach to estimate the average rate of change may prove valuable in pharmacological trials.