Associations between tau‐PET and neurodegeneration with cognitive and behavioural changes over time in amyloid‐positive individuals

Abstract

AbstractBackgroundTau pathology is strongly associated with clinical symptoms, possibly (partly) through the manifestation of neurodegeneration. However, independent pathways between tau pathology and clinical symptoms have been suggested. Longitudinal studies investigating the role of tau vs neurodegeneration on cognitive and behavioral changes across the AD clinical spectrum are relatively lacking.MethodWe included n = 482 individuals from the Swedish‐BioFINDER‐2‐study who had undergone longitudinal neuropsychological assessments and [18F]RO948‐tau‐PET and 3T‐MRI, of whom n = 157 were amyloid‐beta positive (A+) cognitively unimpaired (CU), and n = 326 A+ cognitively impaired (CI) individuals, and calculated SUVRs and cortical thickness in multiple regions(i.e. medial‐temporal, lateral‐temporal, medial‐parietal, lateral‐parietal, frontal, occipital, whole‐brain). Mean follow‐up time was 1.63(1.29) years. Linear‐Mixed‐Effects‐models with random‐intercepts and fixed‐slopes were performed, adjusting for age, sex, and education, applying FDR‐correction, with both baseline and rate‐of‐change in tau‐PET or MRI (the annualized difference between baseline and 2 year follow‐up) as predictors and MMSE, mPACC5 or the Mild Behavioral Inventory (MBI) as outcomes. Post‐hoc analyses included corrections for the other modality and serial mediation analyses (n = 1000 bootstrap iterations).ResultLinear‐Mixed‐Effects‐models with baseline tau‐PET or MRI as predictors showed stronger associations between tau‐PET and cognitive and behavioral changes over time than cortical thickness, which were more pronounced in CI than CU individuals (Figure 1). Analyses with rate‐of‐change in tau‐PET or MRI as predictors showed stronger associations for tau‐PET in CU individuals, but superiority of MRI in CI individuals, which was confirmed when correcting for the other modality, i.e. assessing independent effects (Figure 2). No significant associations were found between rate‐of‐change in tau‐PET or MRI and longitudinal MBI (all p>0.05; data not shown). Serial mediation analyses revealed no mediating effect of rate‐of‐change in tau‐PET or MRI between baseline tau‐PET and cognition over time in CU, while in CI, rate‐of‐change in MRI did show a significant mediating effect (p = 0.002; Figure 3).ConclusionTau‐PET showed stronger associations with cognitive and behavioral changes over time than cortical thickness at baseline across the AD clinical spectrum. However, assessing rate‐of‐change, MRI showed superiority over tau‐PET in CI individuals, suggesting differential contributions of tau pathology vs neurodegeneration in different disease stages.