Brain-age mediates the association between modifiable risk factors and cognitive decline early in the AD continuum
Background: Neuroimaging-derived brain-age is a useful biomarker to study the brain’s biological aging process. Brain-age has shown cross-sectional associations with cognitive function and modifiable risk factors for dementia. We aimed to study, in cognitively unimpaired (CU) individuals, the mediat...
| Autores: | , , , , , , , , , , , , , , |
|---|---|
| Tipo de recurso: | artículo |
| Fecha de publicación: | 2023 |
| País: | España |
| Institución: | Universitat Politècnica de Catalunya (UPC) |
| Repositorio: | UPCommons. Portal del coneixement obert de la UPC |
| Idioma: | inglés |
| OAI Identifier: | oai:upcommons.upc.edu:2117/406371 |
| Acceso en línea: | https://hdl.handle.net/2117/406371 https://dx.doi.org/10.1002/alz.078202 |
| Access Level: | acceso abierto |
| Palabra clave: | Brain -- Aging Cervell -- Envelliment Àrees temàtiques de la UPC::Informàtica::Aplicacions de la informàtica::Bioinformàtica |
| Sumario: | Background: Neuroimaging-derived brain-age is a useful biomarker to study the brain’s biological aging process. Brain-age has shown cross-sectional associations with cognitive function and modifiable risk factors for dementia. We aimed to study, in cognitively unimpaired (CU) individuals, the mediating role of brain-age in the association between modifiable risk factors and cognitive changes, and the impact of AD pathology on this role. Method: We included 416 CU individuals from the ALFA+ study with available structural MRI, measurements of the global cognitive Preclinical Alzheimer’s Cognitive Composite (PACC) (370 individuals had a follow-up PACC assessment 3.28±0.27 years later), and lifestyle and cardiovascular risk factors assessments. We computed brain-age delta as the difference between chronological and predicted brain-age using a previously pre-trained machine learning algorithm on structural MRI data. Partial Least Squares Path Modeling (PLS-PM) was employed to investigate the mediation effect of brain-age delta between a computed latent variable from modifiable risk factors (cardiovascular, mental health and mood, metabolic/endocrine disease history, and alcohol consumption factors; Table 1) and a latent variable from longitudinal PACC. Statistical bias adjustment was performed to control for the confounding effects of age and sex by using multiple linear regression. The analysis was performed on the whole sample (ALL) and after stratification by amyloid-ß (Aß) status. Participants were classified as amyloid-ß positive (Aß+) if CSF Aß42/40<0.071 (Milà-Alomà et al, 2020). Significance of mediation was examined by 95% confidence interval bootstrap. Result: The effect of risk factors (latent risk) on longitudinal cognition (latent cognition) was partially mediated by brain-age delta only in Aß+ individuals (57.286%) (Figure 1C). Higher latent risk was associated with older brain-age delta (p = 0.003) and older brain-age delta was associated with worse latent cognition (p = 0.007). Conversely, the indirect effect was not significant in ALL nor in Aß- individuals, where brain-age delta was not significantly associated with latent cognition (Figure 1A-B). Still, higher latent risk was associated with older brain-age delta (ALL: p<0.001; Aß-: p = 0.004) and associated with worse latent cognition (ALL: p = 0.012, Aß-: p = 0.001). Conclusion: In CU individuals in the AD continuum, biological brain aging mediates the association between modifiable risk factors and cognitive decline. |
|---|