Cerebrovascular accidents in elderly people treated with antipsychotic drugs: a systematic review

After 2002, an association between stroke and antipsychotic use was reported in clinical trials and large database studies. This review considers previous quantitative reviews, newly published clinical trials, and recent observational cohort and case-control studies, and focuses on the clinical significance of the risk for stroke, the difference between typical and atypical antipsychotics, the possible at-risk patient profile and the timing of stroke after exposure. A search of MEDLINE covering the period from 1966 to June 2009 was carried out using selected keywords. Inclusion criteria were (i) quantitative reviews on stroke and antipsychotics; (ii) double-blind, placebo-controlled clinical trials involving patients with dementia treated with antipsychotics; and (iii) observational database cohort studies and observational case-control studies investigating the association between stroke and antipsychotics. Clinical trials were excluded if they were single-blind or if patients were affected by dementia and/or other neurological illnesses. Four reviews with aggregate data, 2 meta-analyses, 13 randomized, double-blind, controlled trials, 7 observational cohort studies and 4 observational case-control studies were selected and analysed. The incidence of cerebrovascular accidents (CVAs) was found to be very low in aggregate reviews and meta-analyses (2-4%). When the number collected was sufficiently high, or different drug treatments were grouped together, the higher rate in subjects exposed to antipsychotics was statistically significant. Inspection of other randomized controlled clinical trials, not included in aggregate reviews and meta-analyses, reported similar rates of CVAs. The majority of observational cohort studies compared typical and atypical antipsychotics and no significant class differences were found. A comparison with non-users was carried out in some cohort studies. In case-control studies, the probability of CVAs in users compared with non-users was in the range of 1.3- to 2-fold greater. Preliminary data also indicate that the highest risk of stroke is related to the first weeks of treatment, and a risk profile for stroke is emerging, such as older age, cognitive impairment and vascular illness. Different pathophysiological pathways may be involved, ranging from the facilitation of thrombosis, pre-existing cardiovascular factors, sedation and a common diathesis for stroke of dementia, schizophrenia and affective illness. Before prescribing an antipsychotic, clinicians should weigh all the risk factors for a given patient and consider not only the indications as provided by the regulatory agencies, but also the overall effectiveness of typical and atypical antipsychotics.