Is depression a risk factor for dementia? a translational research

Frederico Simões do Couto de Oliveira Fernandes

Research output: Types of ThesisDoctoral Thesis


Introduction Depression and dementia are very common and disabling conditions. Most dementia conditions are irreversible, whereby the identification and correction of the risk factors seems to be of paramount importance. Among the several risk factors identified so far, depression emerges as an important target. Several casecontrol and cohort studies yielded heterogeneous results, but the meta-analyzes performed found approximately a two fold increase in the risk for dementia in depressed patients. However, the quality of the studies varies widely, and the accuracy of the diagnosis of depression is frequently not the ideal. Furthermore, the nature of the risk is not clear and the following issues have been repeatedly raised: (a) depressive symptoms are quite common in dementia and could be a symptom of this disorder, and not a true, and early, risk factor, (b) depression and dementia are heterogenous disorders, with distinct biologies, and eventually the risk is different, depending on the specific disorder involved, and (c) what is the role of antidepressants? That is, does the risk diminish if the depression is cured? Antidepressants have neuroprotective properties at the molecular level, but the evidence is much less consistent in vivo. To answer these questions, two studies were performed: a clinical longitudinal controlled study and a preclinical behavioral study. Methods A cohort of 322 depressed patients (exposed cohort), recruited for a taxonomic study of depression between 1977-84, was built. Subjects without depression, admitted for surgery at the same time as the exposed subjects, were the group not exposed to depression. Subjects were contacted again between 2009 and 2013, to assess their dementia status. The risk for dementia in the depressed cohort was compared to the risk in the surgical cohort using binary regression, and the odds ratio were computed (OR). The same analysis was performed in subjects younger than 45 years old (considered to have early onset depression). To quantify the association between different depression subtypes (namely melancholic, anxious, and psychotic) and dementia, crude and adjusted hazard ratios (HR) were obtained with 95% confidence intervals (95% CI) using Cox proportional hazards regression. The preclinical behavioral study assessed the cognitive effect, using the Morris Water Maze test (MWM), of escitalopram in rats submitted to a maternal separation protocol (MS; MS is a protocol that induces depressive-like behaviors). A two-way ANOVA was carried out for analyzing probe trial time (using MS and and escitalopram as treatment factors), computing main effects and interactions. Two-way ANOVA repeated measures was used for the learning curve of MWM. Results In 133 (41.3%) depressed subjects, followed-up for a mean (standard deviation) of 25.7 (7.2) years, the diagnosis of dementia could be established or excluded. Among these, 44 (33.1%) developed dementia versus 20 (15.0%) among the subjects with no depression at baseline, and this result is significant [OR 2.50 (1.14-5.49; 95% CI); p=0.022]. Subjects with early onset depression had an increased risk for dementia when compared to the surgical cohort patients [OR 6.85 (95% C.I. 1.38-34.00); p=0,019]. Patients suffering from depression with melancholic features had an increased risk of developing dementia compared to those depressed without melancholic features [HR 3.64 (1.78-11.26; 95% CI); p=0.025]. In the preclinical study, all groups of animals showed a significant learning effect in the MWM over time, but no differences have been found upon treatment. However, escitalopram treatment significantly increased the time spent on the platform quadrant in the probe trial in the MS group [F(1.23)=10.764; p=0.004], thus seeming to have improved the memory. Discussion The main results of current study are that depression is a risk factor for dementia, with a risk magnitude in line with the longest longitudinal studies with an accurate diagnosis of depression. Two limitations were considered: the lack of formal cognitive assessment at the baseline, and the number of subjects lost to follow. However, the depressed cohort had a low age at baseline, what altogether with the long follow up, makes unlikely the depressed patients were demented at baseline. The erosion in this study is high, but in line with previous studies with a similar design and follow up time, and the differences between those with and those without a known outcome were minimal and were taken into account. When exploring the nature of this risk, (a) these results support the hypothesis that depression is an early risk for dementia, again in line with the studies with a stronger design, (b) depression with melancholic features was found as the only depression feature or subtype that was associated with an increased risk for dementia. Melancholia can have a permanent deleterious effect on cognition, but this was the first study showing a higher risk for dementia. Melancholia is associated with hypercortisolism, and it is known that high cortisol damages the hippocampus, providing a biological rational for these findings. However, no biological assessment of HPA activity were made, and this is a limitation of this study. The inclusion of biological markers would support a biological explanation, but would not interfere with the conclusions of the study, and (3) chronic treatment with escitalopram improved hippocampal dependent memory, in a model that induces depressive-like behaviors (MS). Our results are line with the neuroprotective action of antidepressants, but take a step further by showing that escitalopram also improves cognition in vivo. Transposition of results from animal studies to humans has limitations, but animal studies allow the use of models that are not easily amenable or ethically allowed to humans, and permitting a reliable evaluation of a number of internal and external factors, such as pharmacological interventions. Conclusions The results presented seem to support a role of depression as a risk factor for dementia, and add novel information regarding the nature of this risk. According to these results, depression is not merely a prodrome of dementia, but an early risk factor, and melancholia is the only subtype associated with an increased risk. Also, they point to a neuroprotective action of escitalopram in depression.
Original languageEnglish
QualificationDoctor of Philosophy
Awarding Institution
  • University of Lisbon
  • Figueira, Maria Luisa, Supervisor, External person
  • Ribeiro , Joaquim Alexandre, Co-supervisor, External person
Award date22 Jul 2016
Publication statusPublished - 23 Jul 2016
Externally publishedYes


  • Depression
  • Dementia
  • Risk factor
  • Melancholia
  • Antidepressants


Dive into the research topics of 'Is depression a risk factor for dementia? a translational research'. Together they form a unique fingerprint.

Cite this