Longitudinal hippocampal atrophy is associated with cognitive decline independently of amyloid and tau in neocortex

Abstract

AbstractBackgroundHippocampal volume (HV) atrophy is a well‐known biomarker of cognitive impairment. However, it is not specific for Alzheimer’s disease (AD). To disentangle the contributions of AD and other neuropathologies to cognitive decline, we investigated the longitudinal associations between HV and cognition, adjusting for Aβ and tau in participants from the Harvard Aging Brain Study who were clinically normal (CN) at inclusion.MethodsSerial MRI (HV [1.3‐7.0y], PiB‐PET (Aβ, [1.9‐8.5y]), Flortaucipir‐PET (tau, [0.8‐6.0y]), and the Preclinical Alzheimer Cognitive Composite (PACC, [3.0‐8.8y]) were observed over an average five‐year follow‐up in 128 CN participants, including ten who progressed to symptomatic AD during the study. Longitudinal HVs were processed using Freesurfer v.6 and adjusted for intracranial volume. PiB was measured in a neocortical aggregate, Flortaucipir in inferior temporal (IT) and entorhinal cortex (EC). PET data were expressed as PVC‐SUVr scaled to subcortical white matter. We predicted cognition and imaging data over time with random intercept and slope in linear mixed‐models and extracted PACC, PiB, FTP, and HV slopes for each subject. Baseline imaging and slope data were entered in age‐adjusted linear regressions to evaluate their associations with PACC slope.ResultsWithout covariate, faster HV atrophy was correlated with faster PACC decline (Fig.1, R2=0.28, p<0.0001). Adjusting for age, the correlation was reduced, but significant (R2=0.20, p<0.0001, Table 2). Entering all predictors into sequential regression models, we observed that HV atrophy was associated with PACC decline independently of PiB and Tau (Table 3, model‐D). HV atrophy was explained by EC Tau, but not by PiB or IT Tau (model‐C, Fig.2). When comparing all biomarkers, HV slope uniquely accounted for 10% of the variance in PACC decline. IT Tau explained most of the association between EC Tau, PiB, and PACC. Altogether, 45% of the variance in PACC decline was explained by longitudinal changes in imaging biomarkers.ConclusionsIn older adults, longitudinal hippocampal atrophy is associated with cognitive decline, independently of amyloid and tau, suggesting that non‐AD neuropathologies (TDP43, vascular) contribute to hippocampal‐mediated cognitive decline. Serial HV measures, in addition to AD‐specific biomarkers, may help evaluate the contribution of non‐AD pathologies that cannot be measured in‐vivo but are important for cognition.