Use of antipsychotics to treat behavioural symptoms of dementia has been associated with increased risks of mortality and stroke. Little is known about individual patient characteristics that might be associated with bad or good outcomes.Aims
We examined the risperidone clinical trial data to look for individual patient characteristics associated with these adverse outcomes.Method
Data from all double-blind randomised controlled trials of risperidone in dementia patients (risperidone n = 1009, placebo n = 712) were included. Associations between characteristics and outcome were analysed based on crude incidences and exposure-adjusted incidence rates, and by time-to-event analyses using Cox proportional hazards regression. Interactions between treatment (risperidone or placebo) and characteristic were analysed with a Cox proportional hazards regression model with main effects for treatment and characteristic in addition to the interaction term.Results
Baseline complications of depression (treatment by risk factor interaction on cerebrovascular adverse event (CVAE) hazard ratio (HR): P = 0.025) and delusions (P = 0.043) were associated with a lower relative risk of CVAE in risperidone-treated patients (HR = 1.47 and 0.54, respectively) compared to not having the complication (HR = 5.88 and 4.16). For mortality, the only significant baseline predictor in patients treated with risperidone was depression, which was associated with a lower relative risk (P<0.001). The relative risk of mortality was increased in risperidone patients treated with anti-inflammatory medications (P = 0.021).Conclusions
Only anti-inflammatory medications increased mortality risk with risperidone. The reduced risks of CVAE in patients with comorbid depression and delusions, and of mortality with depression, may have clinical implications when weighing the benefits and risks of treatment with risperidone in patients with dementia.