Fluoxetine to improve functional outcomes in patients after acute stroke: the FOCUS RCT

Background: Our Cochrane review of selective serotonin inhibitors for stroke recovery indicated that fluoxetine may improve functional recovery, but the trials were small and most were at high risk of bias. Objectives: The Fluoxetine Or Control Under Supervision (FOCUS) trial tested the hypothesis that fluoxetine improves recovery after stroke. Design: The FOCUS trial was a pragmatic, multicentre, parallel-group, individually randomised, placebo-controlled trial. Setting: This trial took place in 103 UK hospitals. Participants: Patients were eligible if they were aged ≥ 18 years, had a clinical stroke diagnosis, with focal neurological deficits, between 2 and 15 days after onset. Interventions: Patients were randomly allocated 20 mg of fluoxetine once per day or the matching placebo for 6 months via a web-based system using a minimisation algorithm. Main outcome measures: The primary outcome was the modified Rankin Scale at 6 months. Patients, carers, health-care staff and the trial team were masked to treatment allocation. Outcome was assessed at 6 and 12 months after randomisation. Patients were analysed by their treatment allocation as specified in a published statistical analysis plan. Results: Between 10 September 2012 and 31 March 2017, we recruited 3127 patients, 1564 of whom were allocated fluoxetine and 1563 of whom were allocated placebo. The modified Rankin Scale score at 6 months was available for 1553 out of 1564 (99.3%) of those allocated fluoxetine and 1553 out of 1563 (99.4%) of those allocated placebo. The distribution across modified Rankin Scale categories at 6 months was similar in the two groups (common odds ratio adjusted for minimisation variables 0.951, 95% confidence interval 0.839 to 1.079; p = 0.439). Compared with placebo, patients who were allocated fluoxetine were less likely to develop a new episode of depression by 6 months [210 (13.0%) vs. 269 (16.9%), difference –3.78%, 95% confidence interval –1.26% to –6.30%; p = 0.003], but had more bone fractures [45 (2.9%) vs. 23 (1.5%), difference 1.41%, 95% confidence interval 0.38% to 2.43%; p = 0.007]. There were no statistically significant differences in any other recorded events at 6 or 12 months. Health economic analyses showed no differences between groups in health-related quality of life, hospital bed usage or health-care costs.