Skip navigation

Atypical antipsychotics and stroke in dementia

 

Clinical bottom line

Evidence from randomised trials shows a small increased risk of stroke in patients with dementia and agitation treated with atypical antipsychotics risperidone and olanzapine, but based on a small number of events. A large observational study failed to find any increased risk, and probably limits the size of the effect to no more than two extra strokes per 1000 person years of treatment.


Reference

N Herrman et al. Atypical antipsychotics and tisk of cerebrovascular accidents. American Journal of Psychiatry 2004 161: 1113-1115.


Background

Concerns have been raised about the risk of stroke when using atypical antipsychotic drugs in older people with dementia suffering from agitation and behavioural problems. In Canada in 2002, the USA in 2003, and the UK in 2004, authorities or companies have written to doctors with guidance or advice about the risk, or have advised stopping their use. In the past these drugs were used off-license because they were considered to have fewer common but serious adverse events compared with older drugs, but now placebo-controlled trials have been conducted, especially with risperidone and olanzapine.

These placebo-controlled trials have demonstrated a small increased risk of stroke in patients treated with atypical antipsychotics. To date we have no published systematic review to quantify the data.

For risperidone, a letter to Canadian physicians about risperidone put the risk of stroke or transient ischaemic attack at 4%, compared with 2% with placebo. For olanzapine, information from trials indicates a similar higher risk (15/1178; 1.3%) over placebo (2/478; 0.4%).

From these data, it looks as if the excess incidence of stoke runs at about 1%, but we are lacking a published analysis (in October 2004). Observational information can provide additional information, and there is one study.

Study

The administrative healthcare database in Ontario contains over 1.4 million people over age 65 years. Databases for prescriptions, hospital care and physician services are linked, and individual patient data can be obtained on an anonymous basis. It can therefore link prescriptions with outcomes.

Users of risperidone, olanzapine and typical antipsychotics were identified, and those who were 66 years or older, who had at least two successive prescriptions for at least 30 days were included. Exposure was determined as the period of exclusive use of any study drug. Failure to refill prescriptions, or hospital admission for stroke, or exposure to another medication, or death all represented termination events. Hospital admission for stroke were identified.

Results

During the period there were 11,400 antipsychotic-na´ve subjects who began typical antipsychotics, risperidone, or olanzapine, and there were 13,300 years of follow up. There were 92 admissions for stroke, with no difference in the exposure-adjusted rate between risperidone, olanzapine, or typical antipsychotics (Table 1).

Table 1: Atypical antipsychotics in dementia

 
Typical antipsychotic
Risperidone
Olanzapine
Number
1,015
6,964
3,421
Strokes
10
58
24
Rate per 1000 person years
5.7
7.8
5.7
Risk ratio adjusted for exposure
1.0
1.4 (0.7 to 2.8)
1.1 (0.5 to 2.3)
Patient characteristics
Age (years)
81
83
81
Female (%)
66
69
69
Long-term care residence (%)
33
43
43
Hospital admission prior year
47
46
46
Five year stroke admission (%)
12
13
11
Five year heart failure admission (%)
36
39
34
Five year MI or ischaemia admission (%)
34
34
31

 

Table 1 shows that the patients were old, mainly women, and were frail, with about 40% in long-term care. They had an average of 10 prescription drugs in the last year, almost half had had a hospital admission in the last year, and there were many admissions over the previous five years for cardiovascular disease or stroke.

Comment

It is likely (though not known for certain) that the vast majority of the prescriptions were for behaviours associated with dementia rather than for schizophrenia. Compliance was not known, nor was the incidence of any mild stroke or transient ischaemic attack that did not result in a hospital admission.

Is it possible that the risk of stroke with risperidone or olanzapine is raised, but that this study failed to find it? It found a risk of stoke of about 6 per 1,000 patient years in what was a high-risk population. The assumed increased risk of stoke with atypical antipsychotics from the trials was about 1%, or even more. If they were one year in length, that would be an increased risk of 10 per 1,000 patient years, so this study should have found it.

The trouble is that the small number of events in the trials is inadequate to give good information about the magnitude of an increased risk. Perhaps the best we can say is that on the basis of this good observational study, the risk of increased strokes from atypical antipsychotics in dementia is less than the trials indicate.