156 Chapter 6 DISCUSSION Our main findings are that both baseline tau PET and higher rate of tau accumulation strongly predicted rate of cognitive decline across multiple domains in both CI and CU individuals and that these effects were more apparent after correction for rCBF. Associations between rCBF and cognitive decline were only subtle, as lower rCBF at baseline was predictive of steeper memory decline in CU individuals, and steeper decrease in rCBF predicted a steeper decline on global cognitive functioning (MMSE) in CI individuals only. We found that annual change in rCBF predicted global cognitive decline in CI individuals. Studies relating changes in CBF to changes in cognition in AD are sparse. One study found that decreases in frontal CBF over time (as measured using single photon emission CT [SPECT]) correlated with change in overall cognitive function and the CAMCOG sub tests of language and praxis [39]. Other studies investigating CBF (using SPECT or arterial spin labeling MRI [ASL]) in the context of cognitive decline found a relationship between rCBF reductions in widespread cortical regions, including parietotemporal and frontal lobes, and faster decline in (modified) MMSE scores [18, 40]. These findings are largely in line with results from our study, showing an association between decreases in global CBF and steeper decline on MMSE. Lastly, a recent study using ASL found no association between changes in CBF with longitudinal neurocognitive changes in a cohort including similar participants in terms of cognitive status (CU and CI along the AD spectrum) [17]. These results partially match our findings, since we only found an effect for MMSE in CI individuals, but not for memory or executive functioning, and no effects in CU individuals. Lower baseline rCBF was not clearly predictive for cognitive test scores over time, except for a steeper decline on RAVLT immediate recall in CU individuals. These results partially match findings from a previous study investigating CBF in relation to cognition over time using 18F-florbetapir R 1 in a cohort of CU individuals [15]. This previous study found trend significant (Puncorrected<0.05) associations between lower baseline rCBF with worse trajectory for tests for memory (RAVLT immediate and delayed), attention (TMT-A), and global cognition (MMSE). This may support the notion that R1 may serve as a tracer-independent measure of rCBF (in relation to cognition). For tau pathology we found that baseline values were predictive for longitudinal cognition in CI and CU individuals. In CU, tau pathology particularly did well in predicting memory performance over time, while in CI there was no specificity for cognitive domain. These findings are well in line with previous literature and established development of AD pathology and related clinical aspects, with earliest changes in (medial) temporal regions and the memory domain and later pathological changes in neocortical areas and a wide range of cognitive domains affected [2,
RkJQdWJsaXNoZXIy MjY0ODMw