Distinct patterns of atrophy and their determinants among the elderly population: insights from the Gothenburg H70–1944 Birth cohort

Abstract

Introduction: Distinct patterns of brain atrophy exists in common neurodegenerative disorders. However, whether these patterns are present also in non-clinical populations remains elusive, yet crucial to detect people at-risk of dementia. This cross-sectional study aimed to identify specific patterns of grey matter atrophy and their determinants among cognitively intact older individuals. Methods: We included 746 septuagenarians from the Gothenburg H70-1944 Birth cohort who did not have dementia or neuropsychiatric disorders and with available MRI. Unsupervised classification using Random Forest was applied to identify distinct patterns of brain atrophy based on grey matter volume and thickness from 34 cortical and 7 subcortical regions. Linear and multinomial logistic regression models were used to compare the clusters pairwise (cluster 1 vs. remaining) and estimate their associations with sociodemographic and vascular risk factors, medical conditions, lifestyle, and genetic factors. Furthermore, the clusters were compared in relation to cognition and biomarkers (MRI, DTI, CSF, or blood) of neurodegeneration, cerebrovascular pathology, inflammation, and lipidic abnormalities. Results: We identified 5 clusters that significantly differed in atrophy severity and typicality (location). Cluster 1 was the most prevalent (n=278, 37.3%). As it showed no atrophy differences from the overall sample, Cluster 1 was set as the reference group. Cluster 2 (n=142, 19%) exhibited diffused atrophy, primarily concentrated in frontal regions. Cluster 3 (n=121, 16.2%) and Cluster 4 (n=157, 21%) were characterized by the least atrophy, particularly in frontotemporal and temporal regions. Cluster 5 (n=48, 6.4%) showed more atrophy in parietal-occipital regions.Moreover, regression analyses indicated that small vessel disease (white matter hyperintensities and lacunes), vascular risk factors (smoking, overweight/obesity, alcohol risk consumption), cardiometabolic disorders, lipid alterations, depression, and education play a key role in separating the clusters. Discussion: Our study revealed the presence of distinct patterns of atrophy in elderly individuals otherwise cognitively healthy. Small vessel disease, vascular risk factors, and cardiometabolic conditions may contribute to these patterns. These findings highlight the importance of vascular health to preserve cognitive function and brain structure in old age.