King's College London

Research portal

Associations of neuropsychiatric symptoms and antidepressant prescription with survival in Alzheimer's disease

Research output: Contribution to journalArticle

Christoph Mueller, Jonathan Huntley, Brendon Stubbs, Andrew Sommerlad, André F. Carvalho, Gayan Perera, Robert Stewart, Nicola Veronese

Original languageEnglish
JournalJournal Of The American Medical Directors Association
Early online date31 Aug 2017
StateE-pub ahead of print - 31 Aug 2017

King's Authors


Objective: Depression is associated with increased mortality in community samples. The use of antidepressant medication may also increase mortality, however, it is still unclear whether taking antidepressants before or after a diagnosis of dementia influences survival.

Design: Retrospective.

Setting: A cohort with a diagnosis of Alzheimer disease (AD) from a large mental health and dementia care database in South London, linked to hospitalization and mortality data.

Participants: Mild dementia (Mini-Mental State Examination ≥18/30) at the point of diagnosis.

Measurements: We ascertained antidepressant prescription, either in the 6 months before or after dementia diagnosis, and used the HoNOS65+, a standard clinician-rated measure of patient well-being, to determine depression severity and other neuropsychiatric, physical health, and functional difficulties. We conducted a survival analysis, adjusted for potential confounders and addressed possible confounding by indication through adjusting for a propensity score.

Results: Of 5473 patients with AD, 22.8% were prescribed an antidepressant in a 1-year window around dementia diagnosis. Of these, 2415 (44.1%) died in the follow-up period [mean (standard deviation) 3.5 (2.4) years]. Prescription of an antidepressant, both before and after dementia diagnosis, was significantly associated with higher mortality after adjusting for a broad range of potential confounders including symptom severity, functional status, and physical illness (hazard ratio 1.22; 95% confidence interval 1.08–1.37 for prescription prior to dementia diagnosis; 95% confidence interval 1.04–1.45 for prescription post dementia diagnosis). In stratified analyses, risks remained significant in those without neuropsychiatric symptoms.

Conclusions: The prescription of antidepressants around the time of dementia diagnosis may be a risk factor for mortality.

View graph of relations

© 2018 King's College London | Strand | London WC2R 2LS | England | United Kingdom | Tel +44 (0)20 7836 5454