Late-life depression as a vulnerability factor in the relation between brain abnormalities and cognitive functioning: The smart-medea study

Background: Late-life depression (LLD) is associated with brain abnormalities and risk of dementia, including Alzheimer's disease, but the nature of this relationship is still not well understood. Proposed explanations focus on LLD having a direct relationship with brain abnormalities, whether cause or consequence. We propose an alternative model hypothesizing that LLD is a vulnerability factor in the relationship between brain abnormalities and risk of cognitive decline and dementia. Our aim was to investigate whether LLD modified the relationship between brain abnormalities and cognitive performance. Methods: Within the SMART-Medea study, 691 participants with a history of arterial disease (6269 years) had depression assessment, cognitive testing, and 1.5tesla brain MRI. LLD was defined as a 12-month diagnosis of a major depressive disorder using the CIDI interview or a score of ≥6 on the Patient Health Questionnaire-9. Automated segmentation was used to quantify brain tissue volumes and cerebrospinal fluid. Hippocampal volume was manually segmented. We defined 4 groups based on LLD and brain abnormalities absence/presence (poorest quartile of distribution). ANCOVA was used to estimate the mean Z-scores for memory performance, information processing speed, and executive functioning, adjusting for age, sex, education and premorbid intellectual functioning. Results: LLD was present in 21% of the sample. The mean MMSE score was 28.6 (SD=1.7). Compared to participants without LLD and without global brain atrophy (indicated by brain parenchymal fraction (BPF)), participants with LLD but without atrophy had significantly poorer memory performance (mean difference (95% CI) in z-score -0.24 (-0.43 to -0.05)), while no significant difference was observed for participants with atrophy but without LLD. The largest difference was observed in participants with both LLD and BPF atrophy (mean difference in z-score -0.44 (-0.80 to -0.08)). Similar patterns were observed for the other brain parameters and cognitive domains (Figure). Conclusions: With the caveat of the cross-sectional design and absence of statistically significant interaction, this study provides first evidence that late-life depression may be a vulnerability factor in the relation between brain abnormalities and cognitive decline.