Mixed Model of Cross-Sectional and Longitudinal Data with Application to Modeling Cognitive Aging in a Cohort of APOE e4 Carriers and Noncarriers

• Activity Number: 170
• Type: Contributed
• Date/Time: Monday, August 2, 2010 : 10:30 AM to 12:20 PM
• Sponsor: Biometrics Section
• Author(s): Amylou C. Dueck*+ and Joseph G. Hentz and Brie N. Noble and Eric M. Reiman and Richard J. Caselli
• Companies: Mayo Clinic and Mayo Clinic and Mayo Clinic and Banner Alzheimer's Institute and Mayo Clinic
• Keywords: Longitudinal data ; Cross-sectional data ; Mixed model ; SAS PROC MIXED ; APOE e4 ; Alzheimer's disease

Abstract

To compare the longitudinal effect of aging on 4 cognitive measures between carriers and noncarriers (NC) of an allele (APOE e4) associated with the risk of late-onset Alzheimer's disease in a cross-sectional and longitudinal sample, we applied a statistical method which simultaneously models the cross-sectional and longitudinal effect of age (Ware et al, 1990). The longitudinal acceleration of cognitive decline for each measure was compared between carriers and NC using a quadratic cross-sectional/longitudinal model with random intercept term. Tables and graphics (including a longitudinal "porcupine" plot) to evaluate model fit and communicate results to a clinical audience will be presented. Consistent with the clinical hypothesis, the model found differences between carriers and NC in the acceleration of memory decline (P=0.03) but not the 3 non-memory measures (Caselli et al, 2009).