Associations of Amyloid, White Matter Hyperintensities, and Hippocampal Volume with Cognitive Trajectories in the Oldest‐Old: The 90+ Study

Abstract

AbstractBackgroundAmyloid pathology, vascular disease pathology, and hippocampal atrophy are associated with cognitive trajectories in older adults. However, there is almost no prior evidence on how these pathologies influence cognition in the oldest‐old.MethodWe included 216 individuals from The 90+ Study, a longitudinal study of aging and dementia in people aged 90 years and older, who had 18F‐florbetapir PET and MRI imaging. We examined the association of amyloid, white matter hyperintensities (WMH) volume, and hippocampal volume (HV) with baseline cognition and longitudinal cognitive decline. Amyloid burden was measured using the standardized uptake value ratio (SUVR) in the precuneus/posterior cingulate with white matter mask as reference. WMH volume and HV were normalized by intracranial volume (IV) and normalized WMH were log‐transformed. SUVR and HV for each individual were Z‐scored using the sample mean and standard deviation. Global cognitive performance was measured by Mental State Examination (MMSE) and modified MMSE (3MS) tests, repeated every six months. We defined baseline (time=0) as the visit closest to the PET imaging and included all visits starting from one year before baseline (time=‐1). We used linear mixed‐effects models with a random intercept to estimate the effect of pathologies on cognitive trajectories. All models included baseline age, sex, education, APOE‐ε4, time, time squared, amyloid, WMH, HV, and interactions between linear time and pathology variables.ResultAt baseline, participants were 93.2 years old on average, 65.3% were females, 10.6% were APOE‐ε4 carriers, and 66.2% had normal cognition (Table 1). Higher HV was associated with higher MMSE and 3MS scores at baseline. Both lower amyloid burden and higher HV were associated with slower rate of cognitive decline longitudinally. WMH was not associated with baseline cognition or cognitive trajectory. Parameter estimates are shown in Table 2 and illustrative cognitive trajectories predicted for a 93.2‐year‐old female with APOE‐ε4 alleles and a college degree predicted at varying degrees of amyloid, HV, and WMH are shown in Figure 1.ConclusionAmyloid burden and HV are associated with longitudinal cognitive trajectories, highlighting the utility of amyloid and HV in predicting future cognitive decline for the oldest‐old.