SSRI No Help for Depression After Stroke

Daily treatment with a selective serotonin reuptake inhibitor (SSRI) had no effect on stroke survivors' risk of depression, which decreased on its own by 6 months, a secondary analysis of the AFFINITY trial found.

Among study participants who had experienced a recent stroke, fluoxetine (Prozac) use did not reduce clinically significant symptoms of depression, which at baseline affected 18.9% of the fluoxetine group and 18.5% of the placebo group and dropped at 26 weeks to 7.0% and 8.2% (P=0.41), respectively, according to the study published in JAMA Neurology.

Cumulative prevalence of clinically significant symptoms during the study was 20.2% of the fluoxetine group and 21.1% of the placebo group (P=0.70), reported Osvaldo Almeida, PhD, of University of Western Australia in Perth, and colleagues.

Likewise, rates of newly developed, self-assessed depression during the study were similar in both fluoxetine- and placebo-treated groups (13.0% vs 14.8%, P=0.43).

"The results of the AFFINITY trial showed that the prevalence of clinically significant symptoms of depression decreases steadily during the first 6 months after a stroke associated with minimal to moderate neurologic deficits, and such a decrease is not affected by the routine daily use of 20 mg of fluoxetine hydrochloride," the group concluded.

Thus, their analysis failed to replicate the positive findings from other fluoxetine trials in the setting of stroke.

Interest in potential benefits of fluoxetine's role post-stroke depression -- affecting one in every three patients within a year of the stroke -- had been fueled a decade ago when the FLAME trial reported that 20 mg daily of fluoxetine enhanced motor recovery and decreased the proportion of stroke survivors with depression at 3 months. While functional improvement was not replicated in subsequent trials, benefits in depression were observed again in the FOCUS and EFFECTS trials.

The main finding of AFFINITY had been fluoxetine's lack of a improvement on functional outcome after stroke -- but an increase in falls, bone fractures, and epileptic seizures.

For now, "use of fluoxetine to prevent post-stroke depression should be considered only if the evidence of benefit outweighs the potential risk of harm," the investigators said.

In an accompanying editorial, Michael Hill, MD, MSc, and Sean Dukelow, MD, PhD, both of the University of Calgary in Alberta, noted that in clinical practice, "pharmacologic treatments for depression are started a bit (or much) later than the 15 days after stroke used in the AFFINITY trial and after clinician-administered assessments of depression are conducted."

For their secondary analysis, Almeida and colleagues assessed the prevalence of depression in 1,221 patients enrolled in AFFINITY within 2-15 days of experiencing a stroke. Participants had been recruited from centers in Australia, New Zealand, and Vietnam between January 2013 and June 2019, with 614 randomly assigned to 20 mg of fluoxetine hydrochloride daily, and 607 to placebo.

The groups were balanced in baseline demographic and clinical measures. Participants in each study arm averaged 64 years of age, and approximately 63% were men; median stroke severity was moderate, with an NIH Stroke Scale score of 6.

Notably, in contrast to earlier studies using a non-structured clinician diagnosis of depression, AFFINITY investigators relied on the self-administered Patient Health Questionnaire (PHQ)-9 to assess for signs and symptoms of depression.

"That our ... repeated assessments with a valid depression instrument failed to replicate [previous fluoxetine trial] results showed that agreement between the PHQ-9 definition of depression and participant-reported diagnosis of depression is poor," Almeida and coauthors noted.

Indeed, of the AFFINITY participants with PHQ-9 scores below 9 at baseline, there was a trend toward less clinician-diagnosed depression in the fluoxetine group (4.3% vs 7.0%, P=0.05).

"Diagnosing major depression in the first 15 days after stroke likely has lower specificity than waiting longer and allowing time for [a more comprehensive assessment]," study authors added.

Fluoxetine also did not reduce the secondary outcome of proportion of participants who were prescribed a non-trial antidepressant or nonpharmacologic treatment.

Interaction with time was not significant, nor were results affected by patient factors including type of stroke, sex, history of depression, and country.

Almeida and colleagues cautioned that these results apply primarily to patients with stroke who had mild to moderate neurologic deficits, thus were able to complete the PHQ-9. They acknowledged that the loss of one-third of participants by week 26 may have rendered the study underpowered to show a modest reduction in depression risk.

"It may also be worthy to consider that, like any major illness, stroke will be commonly associated with an adjustment disorder as patients go through a process of grieving, before adapting and compensating," according to Hill and Dukelow.

"The implication is that the effect size sought may have been too large -- the authors calculated that the study had adequate power to detect a 7% risk difference. Clinically smaller changes, measured using this instrument or perhaps with a combination of existing instruments, could still be important to stroke recovery," the editorialists said.

Comment