Drugs for agitation in dementia: the good, the bad and the useless

This is a summary of a longer piece by our co-founder Dr Andres Fonseca originally published in Mental Elf: Which (if any) drugs should we use in agitated or aggressive behaviour in dementia?

This is a dilemma I frequently face when I am called out to see someone with dementia on the ward or living in the community. On the one hand I am thinking that anything I use can potentially have serious side-effects and will probably lead to increased health risks and increased mortality. On the other hand I have a very distressed person and a very distressed group of caregivers. I know that very soon the situation will be unsustainable if I do nothing and I just rely on the carers’ coping skills.

Screen for pain and delirium—I hear you say. You are right, of course. Relationships with staff or other residents and the actual environment can play a role and should be explored. OK, so now I have done all that and I still have a very distressed person and carers at the end of their tether. Now is the time to break out those non-pharmacological strategies, right? Well, maybe. Not many staff members will know about them, they may not work all that well, and patients may not cooperate. Plus, whenever I have tried to introduce this idea I am met by smoldering looks from the staff who are feeling demoralised and exhausted from long hours of looking after a very distressed person. I am now asking them to redouble their efforts. Their faces usually say: 'Oh, yes, easy for you to say, but try doing a shift and then tell me how you feel about not using any drugs'. It always requires all my powers of persuasion to get them to do this.

Eight in 10 people with dementia will have neuropsychiatric symptoms, which means that after all has been tried I will be contemplating using medication in quite a few cases.

What is the most effective least harmful alternative is what you want to know in that situation.  Dr Dallas Seitz and his colleagues in Ontario, Canada Seitz looked at 7310 studies that could potentially help them answer this question. They ended up selecting 29 that were of good enough quality to use.

Of those 29 studies most (15) were on antipsychotics, including risperidone (6), olanzapine (4), quetiapine (3) and aripiprazole (3). Table 1 summarises the results (the result column lists the treatment with the most benefits in the trial. The withdrawals column lists the treatment with most withdrawals due to any cause including adverse events and mortality)


There were 3 studies on cholinesterase inhibitors: one on donepezil and 2 on rivastigmine. Unfortunately all the rivastigmine trials had antipsychotics as their controls. Table 2 summarises the results.

There were 4 studies on anticonvulsants: two on divalproex, one on carbamazepine and 1 on oxcarbamazepine. The results are on table 3.

Seven studies probing a variety of other drugs also made it into the review. The summary is on table 4 below.

Overall risperidone, olanzapine, and aripiprazole may be of benefit although the effect was small and not consistent across all trials. There were additional single small positive studies with carbamazepine, estrogen, cyproterone acetate, propranolol, and prazosin. The amount of benefit tended to be very small. Only a few studies said anything about results I would actually care about, such as symptoms going completely away. There were more people stopping treatment with those treatments that showed a good effect, so there usually is a trade off between benefits and side-effects.

From other evidence we know that antipsychotics increase the risk of death and stroke. Antipsychotics also increase sleepiness, falls, and injuries due to falls including hip fractures. There is also more and more evidence that they may make people with dementia more forgetful and less independent.

What should I do after reading this to help my distressed patient and overwhelmed carers? There is some evidence for the use of a few agents, but most of them can cause adverse effects to my patient. The paper might help me to talk to carers about potential pitfalls of treatment and point out how they are not always so good; that might help me persuade them to try things other than drugs. Unfortunately there is not much evidence for treatments other than drugs yet. I can try medication with clear treatment objectives and a way to measure change in a time-limited trial with a firm commitment to stop treatment if there is no improvement. If there is improvement there is evidence that I may be able to stop it without symptoms coming back, so I should try to do that as soon as possible. From reading this review it's clear that we don't know much about how our well our drugs work and we also need to directly compare them to treatments not using any drugs.

If you have a friend or a relative that suffers from dementia tell us if any of these drugs has been used in their case and how it worked for them. Also, if there are things other than drugs that have helped let us know as well.