Antipsychotics for agitation in dementia: are they safe?

Do they help?

Some trials show a modest effect. In a 13-week trial of 344 people with dementia, low-dose risperidone reduced behaviour symptoms, especially aggression, more than placebo, although the share of people who improved by the trial's own definition was not significantly higher. [1]

In a 2023 trial of 345 people with Alzheimer's dementia, agitation scores fell by 22.6 points over 12 weeks on brexpiprazole and by 17.3 points on placebo: a difference of 5.32 points. Most of the improvement seen on the drug was also seen on placebo. [2]

The risks

A 2005 review of placebo-controlled trials of the newer (“atypical”) antipsychotics found they may be associated with a small increased risk of death compared with placebo. [3]

In a study of 90,786 veterans with dementia, the extra risk of death over 180 days was largest for haloperidol and risperidone: about one extra death for every 26 and every 27 people treated. For olanzapine it was about one in 40, and for quetiapine about one in 50. [4]

A 2024 study of 173,910 people with dementia in England linked antipsychotic use to higher risks of stroke, blood clots, heart attack, heart failure, fractures, pneumonia and kidney injury. The risks were highest soon after starting: in the first 90 days, 4.48% of new users developed pneumonia, against 1.49% of similar people not taking them. [5]

People with dementia with Lewy bodies can react badly. In a small 1992 study, 13 of 16 patients with Lewy body dementia who were given these drugs had adverse reactions, seven of them severe. [6]

What else has been tried

Good-practice guidelines say a non-drug approach should usually come first. In the CALM-AD study, 318 people with Alzheimer's and agitation had 4 weeks of brief psychosocial therapy (social contact, personalised music, or removing whatever set the agitation off) and their agitation scores improved by 5.6 points. There was no comparison group at that stage, so the study cannot say how much of the change the therapy caused. [7]

For agitation at bath time, a randomized trial in nursing homes in Oregon and North Carolina found that person-centred showering and a towel bath in bed cut aggressive incidents by 53% and 60%. [8]

Among other medicines, the antidepressant citalopram reduced agitation and caregiver distress in a 2014 trial of 186 people, but the authors warned that its effects on thinking and on the heart may limit its use. Its close relative escitalopram was not effective in a 2025 trial. [9, 10]

Stopping

Guidelines recommend keeping antipsychotic treatment to under 12 weeks, then tapering and stopping. In UK health records, people who tapered off after 12 weeks had lower estimated risks of delirium and fractures over 24 months than people who kept taking them. [11]

Among US nursing home residents with dementia who had taken antipsychotics for more than 90 days, those who stopped had fewer hospital admissions, including for pneumonia and urinary infections, than matched residents who carried on. Deaths were also lower, but not by a statistically clear margin. [12]

Both are observational studies: people who stopped may have differed from those who did not. Talk to the prescriber before changing or stopping a dose. [11, 12]