Antipsychotics should be used with caution in dementia
A randomised, placebo-controlled discontinuation trial from the UK has found an increased long-term risk of mortality in those patients with dementia who remained on antipsychotic medication.
A total of 165 patients were recruited, of whom 37 withdrew prior to the onset of the 12-month trial period. Of the remaining 128 patients, 64 continued to receive antipsychotic medication and 64 were given a placebo.
After 12 months, a modified intention to treat analysis (restricted to those who started their allocated treatment) found that 70% of those who continued treatment were still alive compared with 77% of the placebo group (45 and 49 survivors respectively).
A telephone assessment was used to extend the period of follow-up by a minimum of 12 months following the end of the intervention. The long-term differences in mortality were found to be much greater: at 24 months only 46% of the treatment group had survived compared with 71% in the placebo group. Over the extended follow-up period there was a significant reduction in survival among those patients who continued their treatment (HR 0.58).
Some caution is necessary in interpreting these findings. For the primary outcome of 12-month mortality, the difference between the two groups was very small and insignificant. Much larger differences were found when the comparison period was extended (to a maximum of 54 months), but the variable length of follow-up meant that increasingly small numbers were being compared.
A meta-analysis has reported a small but significant increase in the mortality risk during 10-12 week RCTs of atypical antipsychotics (OR 1.54, NNH 100).1 Similarly, a large retrospective cohort study found that initiation of atypical antipsychotic treatment significantly increased the risk of death at 30 days both in community (HR 1.31) and residential (HR 1.55) patients.2 The present study indicates a continuing risk where treatment is continued for 12 months or longer.
Most of the patients were prescribed risperidone (67%) or haloperidol (26%) before randomisation. The CSM has advised that the risk of stroke is three times higher with risperidone and olanzapine and that they should not be used to treat behavioural symptoms in dementia.3 The present study, however, did not find evidence of an increase in cerebrovascular deaths. There are several other potential causes of death in patients taking antipsychotics, including arrhythmias, aspiration pneumonia, thromboembolism and falls. All atypical antipsychotics are associated with an increased risk, and the risk is greater for conventional antipsychotics.2
This trial provides a further reminder that we should regularly review our patients with dementia and seek to withdraw antipsychotic medication at the earliest opportunity.
Ballard C, Hanney ML, Theodoulou M et al. The dementia antipsychotic withdrawal trial (DART-AD): long-term follow-up of a randomised placebo-controlled trial. Lancet 2009; DOI: 10.1016/S1474-4422(08)70295-3Reviewer
Dr Phillip Bland