Background: Late-life depression (LLD) is related to an increased risk of developing dementia; however, the biological mechanisms explaining this relationship remain unclear.
Objective: To determine whether the relationship between LLD and dementia can be best explained by the glucocorticoid cascade or vascular hypothesis.
Methods: Data are from 4,354 persons (mean age 76±5 years) without dementia at baseline from the AGES-Reykjavik Study. LLD was assessed with the MINI diagnostic interview (current and remitted major depressive disorder [MDD]) and the Geriatric Depression Scale-15. Morning and evening salivary cortisol were collected (glucocorticoid cascade hypothesis). White matter hyperintensities (WMH; vascular hypothesis) volume was assessed using 1.5T brain MRI. Using Cox proportional hazard models, we estimated the associations of LLD, cortisol levels, and WMH volume with incident all-cause dementia, AD, and non-AD dementia.
Results: During 8.8±3.2 years of follow-up, 843 persons developed dementia, including 397 with AD. Current MDD was associated with an increased risk of developing all-cause dementia (HR = 2.17; 95% CI 1.66-2.67), with risks similar for AD and non-AD, while remitted MDD was not (HR = 1.02; 95% CI 0.55-1.49). Depressive symptoms were also associated with increased risk of dementia, in particular non-AD dementias. Higher levels of evening cortisol increased risk of dementia, but this was independent of MDD. WMH partially explained the relation between current MDD and dementia risk but remained increased (HR = 1.71; 95% CI 1.34-2.08).
Conclusion: The current study highlights the importance of LLD in developing dementia. However, neither the glucocorticoid cascade nor the vascular hypotheses fully explained the relation between depression and dementia.
Keywords: Cerebrovascular disorders; cohort studies; dementia; depression.