Jae Yong Chung (1), AnHye Kim (2), Howard Lee (2), In-Jin Jang (2)
1) Seoul National University Bundang Hospital, Seongnam, Korea, (2) Seoul National University Hospital, Seoul, Korea
Objectives: Alzheimer’s Disease progressively decreases cognitive function and Alzheimer Disease Assessment Scale (ADAS-Cog) is the standard score used to assess cognition in clinical trials. Better understanding of the time course of placebo response and natural disease progression and their covariates would be essential for designing better clinical trials. In addition, ethnic difference may exist in the placebo response. The aim of this work is to investigate the ADAS-Cog placebo response in Alzheimer’s disease and evaluate the effect of Asian ethnicity and other covariates by population modeling approach.
Methods: Data were obtained from an open database of the CAMD (Coalition Against Major Disease) Alzheimer trial database run by Critical Path Institute. In the placebo arms of 5 clinical trials, 63 Asian and 69 African patients were eligible and study-matching Caucasian patients (n=150) were randomly selected. ADAS-cog were modeled to investigate the disease progression and placebo response as a function of time. Nonlinear Mixed Effects Modelling Approach was applied to these data using NONMEM V7.2. The following models were explored:
ADAS = ADAS0 + S*t – A* (1 – exp(-k*t)) + ε, where ADAS0=ADAS-Cog at baseline, S=disease progression slope, A=magnitude of placebo contribution, k=onset rate of placebo response, ε=error. For each parameter, between-subject variability was tested and covariate analysis was investigated.
Results: The covariate analysis showed that the effects of ethnicity, baseline age, and participation duration on disease progression rate were significant. Goodness-of-fit plots and the visual predictive check showed a good adquacy between observed data and predicted (simulated) data. Maximum placebo effect was estimated to be 8.55. The disease progressoin slope of Caucasian (0.018/day) was 50% greater than that of Asian (0.012/day). The confidence intervals produced by non-parametric bootstrap analyses confirmed the significance of the difference. The slope was multiplied by (1 – 0.0373*(Age – 75)) and (1 – 0.00089*(participation duration – 546)). The covariate effects on placebo response could not be identified.
Conclusions: The model adequately describes ADAS-cog progression in patient with placebo treatment. Our results suggest that the natural disease progression can be slower in Asian than Caucasian and inversely correlated with age at baseline, and slower in patients participating longer period. These findings could help optimize the design of clinical trials for Alzheimer’s disease.
Reference: PAGE 23 (2014) Abstr 3060 [www.page-meeting.org/?abstract=3060]
Poster: Drug/Disease modeling - CNS