Depression prodromal to dementia
: mechanisms, biomarkers and potential drug targets

Student thesis: Doctoral ThesisDoctor of Philosophy


The association between depression occurring in late life (in individuals aged 60-65 and above) and subsequent onset of dementia has been demonstrated consistently in a number of cohort studies and meta-analyses. However, much remains unknown at the moment about the determinants of this association. In particular, it is debated whether depression in late life represents a prodrome or risk factor for dementia, or a reaction to emerging cognitive deficits, and whether depression with an onset in later life (LOD) is a stronger predictor of future dementia compared to early- and midlife-onset depression (EOD) with recurring episodes in later life. Furthermore, little is known about the specific factors of a late-life depressive episode – especially biological markers - that might predict progression to dementia.

The present thesis investigates the relationship between depression in later life and dementia or cognitive decline by means of two systematic reviews and the analysis of three longitudinal cohorts. Using a retrospective cohort derived from the Clinical Record Interactive Search (CRIS) dataset and linked datasets, the clinical phenotypes and specific symptoms of late-life depression predicting future progression to dementia were explored. Second, the longitudinal associations between the trajectories of depressive symptoms, anxiety symptoms, and four cognitive domains were explored in a longitudinal prospective internet-based population sample (the PROTECT study). Finally, the role of an array of plasma inflammatory markers in late-life depression and subsequent progression to dementia was examined using a clinical inpatient sample of elderly depressed patients (PRODE) and a control group of non-depressed cognitively intact elderly individuals (COGNORM).

The two cohorts with diagnosis of dementia as an outcome measure demonstrated a rate of progression to dementia of around 26%. In neither cohort substantial evidence was found to support the hypothesis that late-onset depression is associated with higher risk of conversion to dementia than earlier onset. Clinical symptoms of depression most strongly associated with subsequent onset of dementia were hallucinations and irritability. A number of plasma inflammatory markers were higher in participants with late-life depression compared to controls; however, none of these markers were predictive of progression from late-life depression to dementia during 3 years of follow-up.

The study exploring longitudinal associations between the trajectories of PHQ-9 scores, GAD-7 scores, and four cognitive domains found that depressive symptoms were predictive of worse performance on verbal working memory and episodic memory tests, while anxiety was negatively associated with performance on spatial working memory (SWM) tests. Stratification into participants with and without a history of depression demonstrated substantial differences, namely depressive symptoms only predicted poorer scores on verbal memory tests in never-depressed participants, similarly for the relationship between anxiety and SWM; however, the associations between both depression and anxiety and measures of general intelligence were much stronger in the group of participants reporting a history of depression.

This thesis confirms a substantial increase in dementia risk in patients with late-life depression. The likelihood of progression to dementia was not significantly higher in late-onset depression compared to early-onset depression, although mechanisms may be different. Future research should strive to differentiate between the phenotypes of late-life depression and early dementia with more precision. Although no evidence of the role of plasma inflammatory factors as predictors of progression from depression to dementia was found, patients with late-life depression had higher levels of several inflammatory markers compared to control subjects. Future research should continue to examine whether both inflammatory and non-inflammatory, biomarkers can predict progression from late-life depression to dementia.
Date of Award1 Jul 2022
Original languageEnglish
Awarding Institution
  • King's College London
SupervisorDag Aarsland (Supervisor) & Allan Young (Supervisor)

Cite this