Neurocognitive aspects of ketamine and esketamine on subjects with treatment-resistant depression: A comparative, randomized and double-blind study

Major Depressive Disorder (MDD) is highly prevalent and often accompanied by cognitive impairments in domains such as executive functioning, attention, memory, verbal fluency and visual learning. A substantial proportion of patients—commonly termed treatment-resistant depression (TRD)—do not respond to standard monoaminergic antidepressants, and both ketamine (racemic mixture) and its S-enantiomer esketamine have emerged as rapid-acting treatments that can produce response or remission within 24–72 hours. Despite clinical efficacy at subanesthetic doses, concerns remain about their cognitive effects because higher, anaesthetic or recreational doses of ketamine produce cognitive deficits; the literature on cognitive outcomes at antidepressant doses is mixed, with studies reporting no impairment and some reporting improvements in specific domains such as visual memory, processing speed and verbal learning. Araújo-De-Freitas and colleagues set out to fill a gap in the literature by directly comparing the cognitive effects of intravenous ketamine and esketamine in patients with TRD. The study reports neuropsychological outcomes collected before and after a single infusion of either drug, testing whether cognitive trajectories differ between the two treatments and whether, when pooled, these drugs affect cognition across a one-week follow-up period. The cognitive evaluation was a secondary outcome of a larger randomised, double-blind, active-controlled, bicentric non-inferiority trial of antidepressant efficacy.

This investigation used data from a bicentric, randomised, double-blind, active-controlled trial with two parallel groups comparing intravenous ketamine (racemic ketamine hydrochloride) to S(+)-ketamine (esketamine). Randomisation was 1:1. Neurocognitive assessment was specified as a secondary outcome. The trial followed CONSORT guidance and was registered (UMIN000032355). Assessments were performed at baseline (within seven days before infusion), 24 hours (day 1) and seven days (day 7) after a single infusion. Participants were adults (≥18 years) with DSM-determined MDD using the MINI-Plus and met criteria for TRD defined as failure of at least one adequate antidepressant treatment for at least 12 weeks. Exclusion criteria included recent or concurrent electroconvulsive therapy, psychotic disorders, intellectual deficit or dementia, unstable cardiac disease, current illicit drug use, history of ketamine abuse, inability to consent and IQ <70. Ongoing antidepressant medications were maintained but could not be changed within 15 days before randomisation or during the seven-day follow-up. Interventions consisted of a single intravenous infusion over 40 minutes of either ketamine 0.5 mg/kg or esketamine 0.25 mg/kg, both diluted in 100 mL saline. The esketamine dose was chosen as half the racemate dose on the basis that S(+)-ketamine is approximately twice as potent as racemic ketamine and prior work suggested doses around 0.25 mg/kg provide antidepressant efficacy. Cognitive outcomes were measured with a neuropsychological battery validated for the Brazilian population. Instruments included the Wechsler Abbreviated Scale of Intelligence (WASI) for estimated IQ, the WAIS-III Digit Span (direct and reverse) for verbal short-term and working memory, the Corsi Block-Tapping Task for visuospatial short-term and working memory, the Five Digit Test (FDT) for processing speed, inhibitory control and cognitive flexibility, and the Rey Auditory Verbal Learning Test (RAVLT) for auditory-verbal episodic memory and related indices (immediate recovery, immediate recall, delayed recall, learning over trials, interference measures, forgetting speed and recognition). Neuropsychological raw scores were converted into Z-scores. To compare groups and timepoints, the investigators fitted linear mixed-effects models (lme4 package in R) with fixed effects for assessment time, drug group, educational level and change in depressive symptoms quantified as the area under the curve of MADRS scores with respect to ground (MADRS AuCg). Participant identity and repeated timepoints were modelled as random effects. Demographic comparisons used t-tests and χ2 tests for categorical variables.

Seventy-nine subjects were screened: 16 failed inclusion criteria, four declined participation and eight were excluded for other reasons (five for low WASI scores or incomplete neuropsychological data; three for a therapeutic failure), yielding 51 participants who completed the neuropsychological assessments at baseline, day 1 and day 7. The analysed sample was 60.8% female with a mean age of 46.94 ± 14.59 years. When comparing ketamine and esketamine groups directly using linear mixed-effects models adjusted for educational level and MADRS AuCg, the only statistically significant group difference was observed for the Corsi Block-tapping Task Backward: Number Correct. No other cognitive measure showed a between-group difference, indicating largely comparable cognitive effects of the two drugs at the tested doses. Because of the overall similarity, the investigators analysed both treatments combined as a single group to evaluate temporal changes. Several cognitive measures improved across time (baseline, 24 h, day 7). Visuospatial short-term memory showed improvement on the Corsi Block-tapping Task Forward: Number Correct from baseline to day 1 and to day 7. Processing speed and executive-related indices on the FDT improved: Switching time improved from baseline to day 1 and to day 7 (P < 0.01 for both comparisons), Flexibility improved from baseline to day 7 (P < 0.01), Reading time improved from baseline to day 7 (P < 0.01), and Counting time improved from baseline to day 1 and to day 7 (P < 0.01 for both). No significant changes were seen for inhibitory control and working memory tasks. Auditory-verbal episodic memory measures on the RAVLT also showed positive changes: Immediate recovery increased (baseline vs day 1 and baseline vs day 7; and day 1 vs day 7 all P < 0.01), Immediate recall increased (baseline vs day 1 and baseline vs day 7; P < 0.01), Delayed recall increased (baseline vs day 1 and baseline vs day 7; P < 0.01), and Recognition improved (baseline vs day 7; P < 0.01). Conversely, Learning Over Trials (LOT) and Proactive Interference decreased (baseline vs day 1 and baseline vs day 7; P < 0.01 for both). Forgetting speed on the RAVLT decreased from baseline to day 1 (P < 0.01) and then increased from day 1 to day 7 (P < 0.01). Change in depressive symptom burden (MADRS AuCg) was associated with poorer performance on several memory and visuospatial measures: Corsi forward and backward Number Correct and RAVLT Immediate recovery and Immediate recall. By contrast, higher MADRS AuCg was associated with better performance on the FDT Counting time. Educational level was associated with FDT Choosing time and RAVLT Recognition. These associations were reported within the mixed-effects modelling framework used for the main analyses.

Araújo-De-Freitas and colleagues interpret their findings as indicating minimal cognitive difference between ketamine and esketamine when used at antidepressant subanesthetic doses in TRD, the only group difference being on one visuospatial working memory measure (Corsi Backward Number Correct). They note this single finding contrasts with prior work in which esketamine outperformed ketamine at anaesthetic doses, and they propose, without asserting causation, that the R-enantiomer of ketamine might contribute to differences in visuospatial working memory—citing preclinical data that R-ketamine (arketamine) can reverse cognitive deficits in animal models. Across the pooled sample, the investigators found cognitive improvements after infusion in visuospatial short-term memory, processing speed, cognitive flexibility and several auditory-verbal episodic memory measures. The authors place these improvements in the context of earlier studies reporting similar procognitive effects after ketamine or esketamine treatment. They acknowledge apparent contradictions in some RAVLT subscores (LOT and proactive interference decreased) and explain these as calculation artefacts driven by an increased A1 trial score; specifically, higher A1 performance changes the derived LOT and interference indices. When depressive symptom change over time was controlled for using MADRS AuCg, higher overall depression severity related to poorer performance on several memory and visuospatial tasks while associating with better performance on one processing-speed measure; the authors interpret this as suggesting that sustained depressive burden, even when decreasing, can adversely influence some neuropsychological functions. Limitations acknowledged by the study team include cognition being a secondary outcome of the parent trial, potential practice or learning effects from repeated neuropsychological testing and the absence of an untreated control group to quantify those practice effects. They note that most battery components were selected to minimise learning effects, though RAVLT is more susceptible. The authors also state that the study did not assess the isolated effects of R-ketamine and recommend future trials designed to compare enantiomers and to evaluate R-ketamine neurocognitive outcomes. Concluding, they report that at the tested subanesthetic antidepressant doses both ketamine and esketamine were cognitively safe in TRD patients and, in this sample, associated with improvements in several cognitive domains, a finding they regard as clinically and scientifically relevant because of the potential to restore functioning rapidly in TRD patients.