Fluoxetine Disappoints in Two New Stroke Trials

May 14, 2020

Two new large, randomized, placebo-controlled trials of the use of the antidepressant fluoxetine for stroke patients have both shown that the drug has no effect on functional recovery in comparison with placebo.

The EFFECTS and AFFINITY trials were both presented at a European Stroke Organization Conference (ESOC) virtual webinar on May 13. The webinar is being conducted in place of the ESOC meeting, which has been postponed until November because of the COVID-19 crisis.

The EFFECTS and AFFINITY trials are part of a three-trial program to investigate the effects of fluoxetine in stroke patients. The first of these trials ― FOCUS ― has already been reported; that trial also showed a neutral result.

The results of the three trials are remarkably similar, with no suggestion of improvements in stroke recovery but a small increase in risk for bone fractures with fluoxetine.

"We plan to pool our results from these three new trials in an individual patient data meta-analysis, enabling us to perform preplanned analyses, which are required to confirm or refute a more modest benefit or harm," the lead investigator of the EFFECTS trial, Erik Lundström, MD, associate professor of neurology, Uppsala University, Sweden, told Medscape Medical News. "But until these results are published, we do not recommend further fluoxetine trials for stroke recovery."

Commenting for Medscape Medical News, the co-chair of the ESOC webinar, Jesse Dawson, MD, professor of stroke medicine, University of Glasgow, United Kingdom, said: "We now have three large clinical trials showing no effect of use of selective serotonin reuptake inhibitors [SSRIs] alongside usual rehabilitation after stroke. Some of the trials included many people with minor stroke who were most likely destined for a good outcome anyway, and people with more severe disability were less well represented," he said.

"Pooled analysis of the three trials will help us decide if there is any potential benefit in people with more severe stroke," he said. "However, given the increased risk of fracture, any effect would have to be fairly large to offset this, and this appears unlikely.

"The reduction in depression seen in these studies is interesting, and whether we can better address this with SSRIs in selected people after stroke would be interesting to explore," Dawson added.

Lundström noted that fluoxetine had shown evidence of neuroprotection and neurogenesis in animal studies, and a preliminary small, randomized clinical study (FLAME) involving 118 patients that was published in 2011 suggested an improvement in functional outcomes for patients treated with the drug.

A Cochrane review in 2012 of all the available data suggested that SSRIs might improve recovery after stroke, but when only high-quality trials were considered, the effect was less convincing. The review concluded that more data were needed. The current program of three trials that together included almost 6000 patients was designed to answer the question more definitively.

The first of these ― the UK FOCUS trial ― included 3000 patients and was published in the Lancet in 2018. That trial showed no effect of fluoxetine in comparison with placebo on functional recovery, as reflected in change in modified Rankin Scale (mRS) score distribution. The common odds ratio was 0.95 (95% confidence interval [CI], 0.84 – 1.08).


The EFFECTS trial included 1500 patients from 35 stroke centers in Sweden who had experienced an ischemic or hemorrhagic stroke within the past 2 to 15 days and who had a persisting focal neurologic deficit severe enough to warrant treatment from the physician's and the patient's perspective. The median National Institutes of Health Stroke Scale (NIHSS) score was 3. The patients were randomly assigned to receive either fluoxetine 20 mg once daily or placebo for 6 months.

For the primary outcome ― functional recovery, as measured on the mRS at 6 months ― there was no significant difference between the group that received fluoxetine and the patients who received placebo. The common odds ratio was 0.94 (95% CI, 0.78 – 1.13).

Safety data showed that there were fewer episodes of new depression in the fluoxetine group (7.2% vs 10.8%), but more hyponatremia (1.47% vs 0.13%) and bone fractures (3.7% vs 1.5%).

Lundström concluded that the results do not support the routine use of fluoxetine after acute stroke.

Asked about the increase in fracture risk, he said, "We were a bit surprised about this, but FOCUS also found an increased rate of fractures with fluoxetine, and the drug is believed to affect osteoclasts and osteoblasts, so we think it is something biological that is causing this."

Asked whether the patients' deficits may not have been severe enough at baseline for there to have been evidence of benefit from treatment, Lundström acknowledged that this was a possibility. "Our patients had mainly mild strokes ― median NIHSS score of 3. Whereas in the FLAME trial, which did suggest a benefit of fluoxetine, the patients were much more affected, with a mean NIHSS score of 13 at baseline.

"We aimed to include mild strokes in this trial because we wanted a universal therapy. But we will go back and look at this in a combined analysis of all three studies together."


The AFFINITY trial was presented by Graeme Hankey, MD, professor of neurology, the University of Western Australia, Perth. In that trial, which was conducted in 43 centers in Vietnam, Australia, and New Zealand, 1280 stroke patients who had received a clinical diagnosis of recent stroke (within 2–15 days) were randomly assigned to receive fluoxetine 20 mg daily or placebo for 6 months. The mean baseline NIHSS score was 6.

Results were similar to the FOCUS and EFFECTS trials ― there was no significant difference in functional outcomes between the patients who received fluoxetine and those who received placebo at 6 months, as measured by distribution of mRS scores. The common odds ratio was 0.94 (95% CI, 0.76 – 1.15), with a value below 1 favoring placebo.

The only secondary outcome that was significantly different between the two groups was a small improvement in mood in the patients who were treated with fluoxetine.

In terms of adverse events, fluoxetine was associated with increases in epileptic seizures (1.56% vs 0.31%), falls with injury (3.12% vs 1.10%), and new bone fractures (2.96% vs 0.94%).

The EFFECTS and AFFINITY trials were funded from noncommercial sources. Lundström and Hankey have disclosed no relevant financial relationships.

European Stroke Organization Conference Webinar. Presented May 13, 2020.

For more Medscape Neurology news, join us on Facebook and Twitter.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.