American Conference on Pharmacometrics . San Diego, CA. April 3-6, 2011

1. An Improved Model for Disease Progression in Subjects from Alzheimer’s Disease Neuroimaging InitiativeMahesh N. Samtani, Michael Farnum, Victor Lobanov, Eric Yang, Nandini Raghavan, Allitia DiBernardo, Vaibhav Narayan Johnson & Johnson Pharmaceutical R&D, Raritan, NJ, USA. Objectives Newly Identified Covariates [A] Hippocampal Volume; [B] Ventricular Volume; [C] Trail B Test • The objective of the current analysis is to develop a non-linear mixed effects model for disease progression in Alzheimer’s Disease (AD). • Allow estimation of the typical disease progression parameters in the target population along with their inter- and intra-individual variability and incorporate the effects of influential covariates. • Expand the knowledge gained from previously published models and assesses previously untested assumptions about linear disease progression in AD. • Offer a differing perspective by evaluating an expanded set of previously untested covariates on disease progression parameters. Argument Against a Linear Progression Model Rate of Progression Increases from Year 1 to 2 by 1.74 point/yr in ADNI AD Subjects Final Covariate Model Apoe & Chol are 0/1 exponents depending on APOE ε4 non-carrier/carrier and normal/high cholesterol status. HVOL, VVOL, TRAB, & CHOL refer to hippocampal volume, ventricular volume, trail B test, and serum cholesterol [A] VPC & [B] Progression Rate as a Function of Covariates A B • 6 subjects with progression rate >15 points/yr Summary of Structural Models Tested Description of 6 Subjects with Progression Rate >15 points/yr Logistic Structural Model Selected Conclusions • The model describes AD disease progression as a function of previously unidentified influential covariates. • Covariates affecting baseline disease status were years since disease onset, hippocampal volume and ventricular volume. • Disease progression rate was influenced by age, total cholesterol, APOE ε4 genotype, trail making test score (part B), as well as current levels of impairment as measured by ADAS-cog. • Rate of progression was slower for mild and severe Alzheimer’s patients compared with moderate Alzheimer’s patients who exhibited faster rates of deterioration. • Modeling the longitudinal changes in ADAS-cog scores offered several potential benefits: • It allowed for a more precise understanding of the natural history of cognitive decline in the disease. • It may allow optimization of future trial designs to detect disease modifying effects of drugs. • It allowed for more precise correlation of cognitive states with structural and chemical biomarkers. • It facilitated identification of risk factors, demographics, and other covariates that affected disease progression, which may serve as stratification variables in future clinical trials. • The model could estimate the between subject variability and residual variability associated with disease progression, which is a critical factor in determining adequate sample size for future clinical trials. • Availability of long term data allowed assessment of the linear disease progression assumption. Results suggested that the rate of decline is nonlinear and incorporation of nonlinear effects provided an improved model of AD progression. • The logistic model described here could also find applications in other chronic illnesses where disease scores are constrained to lie within a certain theoretical range. • The logistic model has the ability to capture clinician-rated instruments for assessing disease status that may act as a saturable system, where the intrinsic properties of the functional assessment follow a non-linear path. • The 6 subjects with the fastest progression rate (>15 points/year) have combinations of the risk factors predicted by the model to predispose them to faster progression. Covariates Introduced in the Model Based on Prior Knowledge [A] Years Since Disease Onset; [B] Age; [C] APOE4; and [D] Serum Cholesterol Acknowledgements: Data used in this analysis were obtained from the Alzheimer’s Disease Neuroimaging Initiative (ADNI) database (www.loni.ucla.edu/ADNI) on November 7th, 2009. As such, the investigators within the ADNI contributed to the design & implementation of ADNI and/or provided data but did not did not contribute to this analysis or poster preparation. A complete listing of ADNI investigators can be found at: www.loni.ucla.edu\ADNI\Collaboration\ADNI_Authorship_list.pdf American Conference on Pharmacometrics. San Diego, CA. April 3-6, 2011