Key message: This review suggests that many older people with dementia and neuropsychiatric symptoms who have been taking antispychotic medicines for three months or more can be taken off them without negative effects on their behaviour
Older people with dementia are surely amongst the most vulnerable in our society. For those looking after such individuals, either as health professionals or as spouses and children perhaps, how best to care for them can be the subject of uncertainty and debate. The use of antipsychotic drugs to help control difficult behaviour (which generally tends to come and go) is one of these areas of uncertainty. A large study published in the BMJ last year added to a growing body of evidence on the risks of using these drugs, which include having a stroke or dying, in elderly people in nursing homes (see the links below for The Mental Elf’s blog on this). In the UK there is an increasing emphasis on trying non-drug alternatives and the good practice guidelines from NICE, issued back in 2006, suggested that non-drug treatment should be the first choice for people with dementia displaying neuropsychiatric symptoms (NPS). This term includes symptoms such as agitation, aggression, hallucinations, wandering, shouting, apathy and many more.
A new review from the Cochrane Dementia and Cognitive Improvement Group has just been published which looked for evidence that might help us understand whether older people with dementia can be successfully taken off antipsychotic medicines and how this might affect their NPS. They were able to include nine randomized controlled trials (RCTs) with 606 people aged 65 years or older, randomized to continue with an antipsychotic medicine which they had been taking for at least three months or to take a placebo (sugar pill). Seven RCTs took place in nursing homes, one in an outpatient setting and one in both. Different antipsychotic drugs and doses were used and both abrupt and gradual withdrawal schedules were used.
What did they find?
- 8 of 9 RTCs reported no overall difference between groups in withdrawal success (remaining in the trial and off antipsychotic drugs) or NPS
- One small pilot study of people who had responded to haloperidol found that the time to relapse was significantly shorter in the group taken off the drug and the subsequent trial reported increased risk of worsening NPS in people with psychosis or agitation who had responded well to risperidone when it was withdrawn
- Data from 2 RCTs on the full Neuropsychiatric Inventory score could be combined; no difference found between groups
How good was the evidence?
The quality of studies as assessed by the risk of bias was generally good but
- there were few studies available
- all studies had problems recruiting enough frail older people so the statistical power of the studies was low; they may not have detected clinically important differences between groups
- studies differed in the participants, the methods used and the outcomes assessed so mostly it was not possible to combine data
What does this change?
At our recent Anniversary Symposium, Dr Ben Goldacre pointed out that this is a key question users of evidence should be asking and that producers of evidence should anticipate. Well, the limitations of the evidence probably mean that it doesn’t change very much, and the review authors themselves urge caution in interpreting their conclusions, but it’s another bit of evidence that says we’re probably on the right path in looking for alternatives to using these drugs to manage NPS in older people with dementia. Uncertainties remain about the benefits of withdrawing antipsychotics in terms of NPS but the risk of side effects and in particular the marked increase in deaths amongst those taking them call for continued efforts to try alternative approaches and particularly for those with less severe symptoms.
Mind the gap
Ben Goldacre also reminded us that those undertaking and disseminating Cochrane reviews are in a great position to note where there are gaps in the evidence (these might be gaps where trial data has been withheld or where high quality research has yet to be undertaken) and to shout about it. Gaps noted by the authors of this review which could be addressed in future trials are how the effects of withdrawing antipsychotic drugs may vary depending on the drug used and also the influence of other drugs being taken with them, for many people will be taking other medicines too. They suggest that side effects should be measured systematically in studies as this has not happened in most of those which are currently available.
On the subject of side effects, I was very interested to read a blog written last week for The Mental Elf by Clive Adams, the Co-ordinating Editor of the Cochrane Schizophrenia Group, on a study which used a novel approach to investigate anticholinergic side-effects (such as dry mouth and blurred vision) of antipsychotic drugs in people with schizophrenia, using data from trials in Cochrane reviews. You can find it from the link below.
Links:
Declercq T, Petrovic M, Azermai M, Vander Stichele R, De Sutter AIM, van Driel ML, Christiaens T. Withdrawal versus continuation of chronic antipsychotic drugs for behavioural and psychological symptoms in older people with dementia. Cochrane Database of Systematic Reviews 2013, Issue 3. Art. No.: CD007726. DOI: 10.1002/14651858.CD007726.pub2.
Cochrane summary and podcast of this review http://summaries.cochrane.org/CD007726/withdrawal-of-chronic-antipsychotic-drugs-for-behavioural-and-psychological-symptoms-in-older-people-with-dementia
National Collaborating Centre for Mental Health; National Institute for Health and Clinical Excellence (commissioner). Dementia: a NICE-SCIE guideline on supporting people with dementia and their carers in health and social care. London: British Psychological Society and the Royal College of Psychatrists; 2006. (NICE CG42). [Issued November 2006]. Available from URL: http://guidance.nice.org.uk
The Mental Elf blogs: Comparative risks of antipsychotics amongst nursing home residents. Andre Tomlin, March 9th 2012; Old side effects and old drugs; old side effects and new drugs. Clive Adams, April 4th 2013.
Evidently Cochrane 
There has been good old fashioned Cochrane review evidence for some time that antipsychotics have limited effectiveness in many cases, and can have dangerous side effects, (Ballard 2006 http://summaries.cochrane.org/CD003476/atypical-antipsychotics-benefit-people-with-dementia-but-the-risks-of-adverse-events-may-outweigh-the-benefits-particularly-with-long-term-treatment) And the guidelines on their use reflect this.
The evidence that non-drug treatments work may be growing, but until there’s as much weight of evidence as there is for drug treatments, and until training and support for people to administer effective non-drug treatments is in place, all the recommendations in the world won’t change the culture of over-prescription and chronic use of these drugs – at least not very fast.
This 2009 Cochrane review about aromatherapy for dementia illustrates this point beautifully (Holt 2009 http://summaries.cochrane.org/CD003150/the-one-small-trial-published-is-insufficient-evidence-for-the-efficacy-of-aroma-therapy-for-dementia)
A Cochrane review can’t review the results of trials that haven’t yet been done – it can only review the best `available’ evidence. But Cochrane could advocate more vigorously for more trials to be done in areas where the evidence is promising, interest amongst patients and health professionals high, but not enough research has been done to translate knowledge into changes in clinical practice. Cochrane could also help researchers with the design of high-quality trials which are needed to fill in the gaps in our knowledge – and this doesn’t just mean randomized trials.
If Cochrane reviews contained a “Summary of Research Recommendations” table in conjunction with the existing mandatory “Summary of Findings” table, reviews could really spell out exactly where the research has not yet been done, which core patient-relevant outcomes should be measured routinely in future trials, and give constructive guidance on the best study design and methods to deal with the remaining uncertainties revealed by the review.
Cochrane could also pinpoint exactly where research has been done but the results have not been made available. A bit more naming and shaming (all very Cochraney and polite, of course) would not go amiss…
Perhaps Cochrane could also more routinely create short plain language editorial summaries of reviews covering different interventions for the same condition – for example a round-up of the evidence for all treatments (both drug and non-drug) for agitation in dementia. It would then be much easier for people to see the big picture and make well-informed decisions without having to read each review separately. Just a thought!