Longitudinal tau accumulation in cognitively normal older adults is associated with baseline age, pathology, and activation

Abstract

AbstractBackgroundPredictors of longitudinal tau accumulation have not yet been fully characterized, but are critical for understanding progression from normal aging to Alzheimer’s disease (AD). We investigated whether baseline AD pathology, demographic risk factors, and activation during a memory task predicted longitudinal tau deposition in cognitively normal older adults (OA).MethodTau pathology was measured with [18F] Flortaucipir (FTP) at ≥2 time points (44 2TPs; 24 3TPs; 2 4TPs; 2.5±1.1 years follow‐up) in 70 OA (mean age 77 years, 59% female). FTP scans were processed with a longitudinal pipeline using an eroded white matter reference region. Linear mixed models were used to calculate FTP slope in entorhinal cortex (EC), inferior temporal gyrus (IT), and a MetaROI composed of AD‐vulnerable temporal regions. Aβ was measured with global PiB DVR. A subset of 37 OA received baseline fMRI while performing a memory task, and task activation (all images vs. perceptual baseline contrast) was calculated for EC and hippocampus.ResultThere was significant group‐level tau accumulation in EC, IT, and MetaROI across the full sample, and within Aβ+ and Aβ‐ groups (one‐sample t‐tests, all ps<0.001; Figure 1). Older age at baseline was specifically associated with increased FTP slope in EC (Figure 2A). Higher baseline PiB DVR was associated with increased FTP slope in EC, IT, and MetaROI (Figure 2B). ApoE ε4+ was associated with greater FTP slope in IT and MetaROI, but not EC (Figure 2C). Multiple regression models including age, sex, ApoE, and baseline EC FTP and PiB DVR showed a significant interaction between baseline EC FTP and PiB DVR, and an independent effect of ApoE ε4+, in predicting FTP slope in IT and MetaROI (predictor ps<0.01). In the subset with fMRI data, greater task activation in EC and hippocampus was associated with increased EC FTP slope (Figure 3).ConclusionAge was a specific risk factor for tau accumulation in EC, while baseline Aβ and ApoE ε4+ were related to tau accumulation in neocortical regions. Positive associations between activation and tau change within EC support theories of activity‐related tau accumulation. Our findings help elucidate potential underlying causes of tau accumulation in aging.