191 Longitudinal tau PET and atrophy old (without correction for multiple testing). Taken together, in Aβ+ individuals we found robust associations between baseline tau pathology and longitudinal cortical thickness, while no association between change in tau pathology and longitudinal cortical thickness was found. This might be explained by the difference in sample size in the statistical models for baseline tau PET BPND (n=147) vs annual change in tau PET BPND (n=61) as determinant. It may also be that our follow-up sample was somewhat biased, as the most advanced individuals might have dropped out more frequently. Another reason may be that there is a temporal delay for the neurotoxic effects of tau to manifest, making baseline tau pathology more important for the occurrence of neurodegeneration when compared to change in tau. Lastly, the magnitude of annual change in tau PET BPND is generally modest, leading to a difference in variability which may affect the ability of finding statistically significant effects. Our results demonstrated an increase in tau pathology over time, irrespective of baseline amyloid status. Although current hypothetical models propose that amyloidosis is an upstream driver of tau accumulation [4, 43, 44], and tau pathology is generally only found to be accumulating in Aβ+ individuals, there are some studies showing significant cortical tau accumulation in Aβ- individuals [40, 45]. Accumulation of tau pathology in Aβ- individuals may be driven by processes related to ageing, since positive associations between rates of tau accumulation and age were found among cognitively unimpaired Aβ- individuals [44]. It could also be that Aβ- individuals with accumulating tau pathology do actually have amyloid pathology, but at subthreshold or below detection levels, as it has previously been shown that in individuals who were nominally Aβ-, both the rate of Aβ accumulation and the baseline Aβ load predicted tau deposition in cortical Braak regions associated with AD [46, 47]. Future studies into longitudinal tau accumulation in the context of amyloid pathology may therefore consider looking at continuous amyloid levels rather than binary amyloid status. Relative cerebral blood flow did not change over time during the two-year follow-up period. Although longitudinal changes in rCBF have not been studied previously using [18F]flortaucipir R 1 in other cohorts, we might compare our findings with studies investigating CBF using SPECT, 15O-H2O PET and MRI. Previous findings were indicative of both increases and decreases in rCBF over time in individuals without dementia who had high-amyloid load [24] and decreases in (fast progressing) AD patients [48]. These findings are in contrast with the lack of change over time in our study and also in contrast with our hypothesis, where we assumed that CBF alterations occur in between tau accumulation and atrophy in the pathophysiological development of AD, thus expecting changes in rCBF to occur, especially when changes in cortical thickness are observed. One explanation for this discrepancy might be that changes in rCBF occur in different directions (in- or decreases) on 7
RkJQdWJsaXNoZXIy MjY0ODMw