Scientists from the Stevens Institute for Neuroimaging and Informatics at the University of Southern California studied how depression and antidepressant use are related to the volume of individual parts of the hippocampus - a brain region critical for memory and one of the first to be affected in Alzheimer's disease. The results were published in the journal Translational Psychiatry.
The study involved 2,009 people aged 50 to 90 (mean age about 65) without cognitive impairment. All underwent high-resolution MRI of the hippocampus. Depression was diagnosed in 630 participants: either by the Geriatric Depression Scale (10 points or higher) or by self-report of a doctor's diagnosis. The remaining 1,379 people were considered not to have depression.
The researchers measured the volume of three hippocampal subfields: CA1, the subiculum and a composite CA23DG region comprising CA2, CA3 and the dentate gyrus. Some participants also underwent PET scanning to assess accumulation of amyloid (684 people) and tau protein (386 people) - proteins associated with Alzheimer's disease.
It turned out that people with depression on average had a smaller CA23DG volume. No association was found with CA1 or the subiculum. This association was weak and persisted after adjustment for body mass index, physical activity, tau level and the presence of the APOE4 gene variant, which raises Alzheimer's risk. When amyloid was added to the analysis, the association ceased to be statistically significant, although its magnitude remained the same - the authors explain this by the smaller number of participants with amyloid PET scans.
It is important to emphasize: this is a correlation - a statistical association that by itself does not prove cause-and-effect relationships. The study design does not allow the claim that depression causes a reduction in hippocampal volume or vice versa.
Among participants with depression, those taking antidepressants had smaller CA1 and CA23DG volumes compared with those not taking them. After accounting for body mass index and physical activity, these associations weakened to a borderline level. When medication use and depression were analyzed together, the association of depression itself with smaller CA23DG disappeared, while antidepressant use remained associated with smaller CA23DG volume.
The authors suggest that these results may reflect so-called "confounding by indication": people with more severe depression are more often prescribed medications. Participants taking antidepressants reported fewer current depressive symptoms than those not taking them, which supports the effectiveness of treatment. However, the researchers did not have data on depression severity before treatment began or on the duration of medication use.
The severity of current depressive symptoms and the age of depression onset were not associated with the volume of any of the subfields. In an additional analysis, the association of depression with smaller volume was stronger in men than in women, but the authors treat this cautiously: the formal test for sex differences was not statistically significant.
The scientists did not compare different types and doses of antidepressants. They note that the possibility of the drugs themselves affecting hippocampal structure should be tested in randomized controlled trials. In addition, depression diagnoses and medication use were self-reported by participants rather than taken from medical records, and the amyloid and tau analyses relied on smaller subgroups.
"It is important to study Alzheimer's disease in the context of mental health, since the prevalence of depression increases with age. In addition, it is necessary to investigate how depression and antidepressant use affect the hippocampus - a brain region that is severely affected in Alzheimer's disease. Understanding the link between depression and the development of Alzheimer's disease is important for finding preventive measures against brain atrophy and dementia," the authors conclude.