185 Summary and general discussion biomarkers and cognition. Some other studies report a worse baseline cognitive performance associated with biomarker positivity (9, 10), but associations are subtle. In contrast to the lack of cross-sectional associations between biomarkers and cognition, we consistently found biomarkers at baseline to be predictive of cognitive decline over time, which highlights the added value of longitudinal data. Biomarkers in the ATN classification, especially a positive amyloid status, were associated with cognitive decline and clinical progression to MCI or dementia in chapter 2. Not only clear-cut amyloid positivity, but also grey zone amyloid burden was associated with worse memory performance over time in chapter 3. Finally, different N biomarkers added predictive value for clinical progression to MCI or dementia beyond amyloid and p-tau in chapter 4. The chapters in this thesis included cognitive data covering 3-4 years follow-up on average, which could be viewed as relatively short since the development of dementia is a gradual process which occurs over decades (11). Studies with a follow-up duration of more than a decade illustrate that results can be different for the short-term (for example shorter than five years) and long-term, highlighting the benefits of longer follow-up duration (12, 13). However, the follow-up duration in this thesis is similar to or longer than most other studies (14-18), and we were able to robustly show the predictive value of biomarkers. Because the overall progression rate to MCI or dementia was low, the results show the predictive value of biomarkers on a group level. A longer follow-up period could lead to a higher number of individuals progressing to MCI or dementia, which would enable us to more accurately estimate risks associated with biomarkers on an individual level. We additionally incorporated longitudinal imaging data in chapter 6 and 7, to study brain changes over time. In chapter 6, we found amyloid accumulation and rCBF were longitudinally associated and were both independently related to cognitive decline. In chapter 7, we found biomarker status changed over time in a substantial percentage of individuals. The order in which they changed was highly variable and did not necessarily follow the generally accepted hypothetical sequence of A → T → N. It is essential to study longitudinal trajectories of biomarkers in larger groups of participants in order to gain more insight in the pathophysiological processes in the development of AD. This could also help the design of clinical trials. Failures of anti-amyloid trials in symptomatic individuals lead to the targeting of individuals even earlier in the disease process, and more knowledge about the relevance of amyloid accumulation in the negative range could help participant selection (19, 20). Even in the relatively short follow-up period of 2-3 years, we were able to find relevant associations between longitudinal trajectories of imaging markers and we provided insight in the order of biomarker changes. However, a longer follow-up duration with a higher number of time points would enable us to more accurately estimate specific biomarker trajectories. 8
RkJQdWJsaXNoZXIy MjY0ODMw