Objectives Antidepressants are commonly used in dementia. Depression is a frequent and important co‐morbidity in dementia, and antidepressants are often used to treat depression and more widely. However, there are questions about their utility in depression in dementia and other behavioural and psychological symptoms of dementia. The aim of this narrative review is to summarize the evidence on whether there is therapeutic value in prescribing antidepressants to people with dementia. Methods A PubMed search was performed to identify randomized controlled trials that prescribed antidepressants to people with dementia, either in the treatment of behavioural and psychological symptoms of dementia (depression, anxiety, agitation/aggression, psychosis and apathy) or for secondary outcomes (quality of life, carer burden, activities of daily living, cognition, clinical severity and adverse events). Results Thirty‐six randomized controlled trials were identified (participant n = 3386). A consistent finding in well‐designed blinded placebo controlled trials in dementia is the lack of positive effect of antidepressants on outcomes of interest, including depression. One large well‐designed study has reported a significant reduction in agitation in people with dementia, but at the expense of clinically significant adverse events. Otherwise, change observed in open trials is also seen in the placebo group, suggesting that any effect is not attributable to the prescription of antidepressants. Conclusions It is striking how few data there are on indications other than depression. We should question the use of antidepressants in dementia. Definitive trials of clinical effectiveness of specific indications such as anxiety and agitation in dementia and discontinuation of antidepressants in dementia are needed. Copyright © 2016 John Wiley & Sons, Ltd.