A substantial body of evidence linking late-life depression and dementia is now available. However, precise estimates of the relative risk attributable to late-life depression assessed with specific screening instruments at specified thresholds have not been previously produced.

Summarise dementia risks associated with depression.

Conduct a systematic review of the literature to produce precise and specific risk estimates for all cause dementia, Alzheimer's disease (AD), and vascular dementia (VaD).

The PubMed, PsycInfo, and Cochrane databases were systematically searched. Studies assessing incident dementia using validated measures of clinical depression or depressive symptomatology from prospective population studies were selected. The most specific analyses were conducted using both continuous symptomatology ratings and categorical measures of clinical depression based on single instruments with defined cut-offs.

The literature search yielded 121,301 articles, of which 36 were eligible. Included studies provided a combined sample size of 66,532 individuals including 6593 dementia, 2797 AD, and 585 VaD cases. Random-effects summary estimates showed that the risk associated with depression did not differ by type of dementia. The most widely used instrument was the CES-D. A clinical threshold of 20 produced similar estimates for all-cause dementia (HR 1.83, 95% CI 0.95–3.52) and for AD (HR 1.97, 95% CI 0.96–4.04). Estimates based on other thresholds and continuous measures produced consistent results.

Reliable dementia risk estimates associated with late-life depression can be produced and do not differ between dementia types. Such estimates should be used in evidence-based medicine practice to assess individual risk and to inform policy on interventions to decrease risk in the population.

The authors have not supplied their declaration of competing interest.