Depression and dementia: interrogating the causality of the relationship

BACKGROUND: Depression is often cited as a major modifiable risk factor for dementia, though the relative contributions of a true causal relationship, reverse causality and confounding factors remain unclear. This study applied a subset of the Bradford Hill criteria for causation to depression and dementia including strength of effect, specificity, temporality, biological gradient and coherence. METHODS: A total of 491 557 participants in UK Biobank aged between 40 and 69 at enrolment and followed up for a mean duration of 12.4 years were studied. Diagnoses of depression and dementia were ascertained from linked health records, self-reports and death certificate registration. Depressive symptoms were measured at enrolment using a combination of questions based on the Patient Health Questionnaire-9 depression screening questionnaire. Regional grey matter volumes were measured using T1-weighted MRI in 41 929 participants. RESULTS: Depression was a strong risk factor for incident dementia with an OR of 1.76 (95% CI 1.63 to 1.90), a relationship which was found to be specific to depression rather than commonly proposed confounders. Depressive symptoms increased rapidly in the 10 years prior to dementia diagnosis. The severity of depressive symptoms showed a dose-response relationship with dementia risk. Depression at older ages correlated with reduced grey matter volume in an Alzheimer's pattern whereas younger onset depression was associated with reduced grey matter volume in the frontal lobes and cerebellum. CONCLUSIONS: This study provides evidence that the link between depression and dementia is due to reverse causation with a smaller component of causation with clear evidence of both mechanisms driving the association.