Frequently Asked Questions

Vortioxetine had the strongest and most consistent evidence for procognitive effects in major depressive disorder. In the largest trial reviewed, 8 weeks of vortioxetine 10 mg/day and 20 mg/day was significantly superior to placebo for improving global cognition in adults with moderate major depressive disorder (N = 598), with the largest effects on the Digit Symbol Substitution Test (DSST) at d = 0.51 for 10 mg and d = 0.52 for 20 mg.

The review also notes that about half to two-thirds of vortioxetine's cognitive benefit appeared to be a direct effect on cognition rather than simply a consequence of improvement in depressive symptoms. Across the serotonergic agents reviewed, vortioxetine was the agent most consistently associated with improvement in objective cognitive performance.

Overall, the review found that tricyclic antidepressants do not appear to be procognitive in major depressive disorder. Imipramine showed some memory improvement in small studies, but these gains were linked to depression severity, suggesting the effect may not have been a direct cognitive benefit. Amitriptyline did not improve psychomotor performance after 4 weeks in one randomized trial and was associated with significantly worse verbal learning than fluoxetine in 2 studies.

Based on the studies reviewed, the authors concluded that further research on tricyclic antidepressants for cognitive impairment in major depressive disorder is not warranted.

Vortioxetine showed mixed results for functional outcomes, although some studies suggested a relationship between better cognition and better functioning. In one 8-week randomized trial, vortioxetine significantly improved functional capacity versus placebo in 508 patients with major depressive disorder (P < .001), but 2 other studies did not find a significant direct effect on functional capacity.

Even with these mixed results, cognitive improvement with vortioxetine was positively correlated with functional change in 2 studies: one reported r = 0.21 (P = .02) between cognitive performance and performance-based functional capacity, and another found a partial correlation of r = 0.31 (P = .006) between DSST performance and functioning after 8 weeks of vortioxetine or escitalopram treatment.

Bupropion showed positive effects on memory and some measures of processing speed in the studies included in this review. In one open-label study, 8 weeks of bupropion in patients with severe major depressive disorder (N = 20; mean HDRS = 24.8) significantly improved memory and some measures of mental processing speed, with greater improvement in responders than nonresponders.

In a separate study comparing bupropion with escitalopram, 8 weeks of bupropion in patients with moderate major depressive disorder (N = 36; mean HDRS = 23.4) was associated with significant improvements in verbal and nonverbal memory, but not working memory or composite memory. Functional improvement also occurred, although it was not significantly greater than with escitalopram, and change in immediate verbal memory directly influenced psychosocial functioning.

Among SNRIs and related agents, the review found the most supportive evidence for levomilnacipran, desvenlafaxine, and some studies of duloxetine. In a placebo-controlled study, 8 weeks of levomilnacipran versus placebo in severely depressed patients (N = 429; mean MADRS = 35.2) significantly improved attention and reaction time, with larger benefits in those who had greater baseline cognitive impairment.

Desvenlafaxine was associated with significant improvement in cognitive flexibility, processing speed, and global cognition after 8 weeks of open-label treatment (N = 36), including a global cognition effect size of d = 0.43, P = .003. Duloxetine improved several cognitive domains in open-label or active-comparator studies, but in an 8-week placebo-controlled trial (N = 508) it did not separate from placebo on executive functioning, processing speed, or attention.

The review suggests that older SSRIs such as fluoxetine and fluvoxamine were generally not procognitive. Although 2 small active-comparator studies found that fluoxetine improved verbal learning more than amitriptyline after 6 weeks, a larger 24-month randomized placebo-controlled trial of fluoxetine (N = 202) found no significant improvement in cognitive performance over time.

Fluvoxamine improved attention and reaction time after 4 weeks in one trial of severely depressed patients (N = 116), but that improvement was mediated by change in depression severity (r = 0.44, P < .05). The review therefore interpreted older SSRIs as not having clear direct procognitive effects.

Several non-antidepressant agents showed possible cognitive benefits, but the evidence was limited and often based on small or exploratory studies. Lisdexamfetamine augmentation for 9 weeks in remitted major depressive disorder (N = 59) significantly improved executive functioning versus placebo, although it did not improve composite cognition.

Other signals included modafinil augmentation, which improved one test of executive functioning after 4 weeks in an open-label study (N = 31); erythropoietin, which improved recall and recognition memory more than placebo over 8 weeks in treatment-resistant depression with effects maintained across 14 weeks (N = 39); l-theanine, which improved executive functioning and verbal memory after 8 weeks in an open-label study (N = 20); and low-dose caffeine, where 60 mg but not 120 mg was associated with improved cognition in a male-only sample (N = 90). By contrast, 12 weeks of intranasal insulin in treatment-resistant major depressive disorder (N = 35) showed no cognitive benefit.

The review focused on objective cognitive measures because self-reported cognitive symptoms may not accurately reflect actual cognitive performance. The authors state that self-reported cognitive improvement has lower validity because the relationship between reported and actual cognitive impairment is weak, while cognitive complaints are strongly related to depressive symptom severity.

For that reason, studies were only included if they measured objective cognitive functioning. The authors identified this strict use of objective measures as a key strength because it provides a more realistic assessment of procognitive effects in major depressive disorder.

This was a systematic review of clinical trials evaluating antidepressants and other pharmacologic agents for cognitive impairment in adults with major depressive disorder. The authors searched MEDLINE, PsycINFO, and Embase through Ovid on May 7, 2019, screened 2,045 papers, assessed 53 full-text articles, and included 26 studies.

Eligible studies included randomized controlled trials and open-label trials in adults aged 18 to 65 years with DSM-III, DSM-IV, DSM-5, or ICD-10/11 major depressive disorder, and they had to use objective cognitive outcome measures. The final sample included 13 randomized controlled trials, 7 open-label trials, and 6 active-comparator studies, with study durations ranging from 1 week to 24 months and sample sizes ranging from 17 to 598 participants.

The main limitations were small samples, short follow-up, open-label designs, and wide variation in the cognitive tests used across studies. Seven of the 26 included studies were open-label, many studies were short-term, and sample sizes ranged widely from 17 to 598 participants.

The authors also noted that most studies did not use objective cognition as a primary outcome, so many relied on brief cognitive batteries that may not capture the full range of impairments seen in major depressive disorder. Additional concerns included possible bias favoring newer drugs, recruitment bias related to symptom profiles, and the inclusion of participants without objective cognitive impairment, which could weaken observed treatment effects.