A Multicentre randomised trial to establish the effect (s) of routine administration of Fluoxetine in patients with a recent stroke

Design: Multicentre randomised, parallel group, double-blind, placebo-controlled clinical trial. Setting: Patients with an acute stroke admitted to hospital or seen as outpatients, recruited by NHS hospitals in the UK. Target population: UK patients who had a stroke within the previous 2 to 15 days. Inclusion criteria: Age at least 18 years; clinical and brain imaging compatible with ischaemic stroke or intracerebral haemorrhage within 2-15 days; persisting focal deficit severe enough to warrant treatment (from the patient’s or carer’s perspective) with a once daily trial capsule for 6 months. Exclusions: Subarachnoid haemorrhage; unlikely to be available for 12 month follow-up; patient unable to understand spoken or written English AND no close relative to help with follow-up forms; other life threatening illnesses; pregnant; breast-feeding; child bearing age not taking contraception; previous epileptic seizures; previous self-harm; allergy or contraindication to fluoxetine; current or recent depression (within 1 month) requiring treatment with SSRI; currently taking an SSRI or other medication with serious interaction with fluoxetine. Health technology: Fluoxetine 20mg daily or matching placebo for 6 months, orally or via a feeding tube if the patient is unable to swallow. All patients will receive usual stroke care. We chose fluoxetine 20mg as this was the most commonly used drug and dose in our Cochrane review of Selective serotonin reuptake inhibitors (SSRI) for stroke recovery. It has a longer half life than other SSRIs so withdrawal syndrome is rare. Six months treatment is in the range of previous trials (4 weeks to 1 year) and is the period during which most neurological recovery occurs. Measurement of costs and outcomes: The primary outcome is dependency (modified Rankin scale (mRS)) at 6 months after randomisation. An NHS perspective will be used for measuring and valuing health service use. Several other patient-centred outcomes (including EuroQol (EQ5D-5L)) will be measured at 6 months and 12 months to determine whether any benefits persist to 12 months, and cost-effectiveness. Survival times will be adjusted using EQ5D-5L to calculate quality adjusted life years (QALYs). We will estimate one year cumulative costs of in-patient episodes and community care. Justification of outcomes: The modified Rankin Scale (mRS) at 6 months (our primary outcome) is a simple, time efficient measure to categorise functional outcome, is the most commonly used primary outcome in multicentre stroke trials and can be completed by postal questionnaire, telephone and by proxies. Its predictive validity and reliability have been well-studied. Our secondary outcomes are patient centred, have been used extensively in other stroke trials, and have acceptable validity and reliability in stroke. They are: Stroke Impact Scale (also tested for proxies), Mental Health Inventory 5 (anxiety and depression), vitality subscale of the Short-Form 36 (fatigue), EQ5D-5L (health related quality of life), new diagnosis of depression (by clinician responsible for patient’s care), adverse events, health and social care resource use. These outcomes are collected by post or telephone by the trial co-ordinating centre, thus avoiding expensive, time-consuming face-to-face follow-up. Our postal/telephone follow-up at 12 months (mRS and all secondary outcomes) will determine whether any benefits persist. Long-term follow-up for hospital admission and mortality will be by flagging patient records with GRO in Scotland and Medical Research Information in the rest of the UK. Our mechanism for reporting adverse events have been approved by MHRA. We have searched the COMET database using the term ‘stroke’ and found no relevant core outcomes to inform our decision. Sample size: Based on the distribution of mRS at 6 months in our start-up phase, and using an ordinal sample size method, we have 90% power (alpha=0.05) to detect a Commons Od