Acute effects of ketamine and esketamine on cognition in healthy subjects: A meta-analysis

Ketamine is a dissociative anaesthetic with emerging therapeutic applications in pain and depressive disorders, but it is also associated with psychiatric and cognitive adverse effects. Earlier experimental and non-placebo-controlled studies reported ketamine-related impairments in memory, learning and attention, and more recent placebo-controlled work has documented deficits in executive and sensorimotor functions. Ketamine acts primarily as an NMDA receptor antagonist; its two enantiomers, (S)-ketamine and (R)-ketamine, differ in NMDA binding affinity. Prior literature suggested that cognitive effects vary with dose and resultant plasma concentrations, and that sex and age might also modulate pharmacokinetics and cognitive vulnerability, but a quantitative synthesis across domains and moderating variables had not been conducted. Zhornitsky and colleagues set out to fill this gap by performing a multivariate meta-analysis of placebo-controlled experimental studies in healthy volunteers to quantify acute ketamine-induced cognitive changes across eleven predefined cognitive domains. The study also aimed to examine whether enantiomer type, route of administration, plasma level, infusion dose, sex composition and mean age of samples moderated cognitive effects, thereby informing the immediate cognitive risks associated with ketamine in controlled settings.

The investigators searched PubMed, Embase and Web of Science using keywords around ketamine and cognitive constructs, and supplemented the search with cross-referencing. Screening and data extraction were performed independently by two authors with a final consensus reached among three investigators. Studies were eligible if they: administered ketamine experimentally to healthy volunteers; used a randomised placebo comparison; measured acute cognitive performance with validated neuropsychological tests; and were published before 1 September 2021. There were no language or age restrictions. Reporting followed PRISMA guidance. Neuropsychological outcomes were grouped into 11 domains: attention, executive functions, response inhibition, social cognition, speed of processing, verbal fluency/language, verbal learning, verbal memory, visual learning/memory, visuospatial abilities and working memory. Assignment of individual tests to domains was based on prior classifications and the authors' previous meta-analyses. Quantitative synthesis used R with the metafor package. A multivariate random-effects model with a random intercept nested within studies accounted for within-study correlations arising from multiple outcomes per study. Effect sizes were computed as Cohen's d (positive values indicating worse performance under ketamine relative to placebo); conventionally 0.2, 0.5 and 0.8 indicate small, medium and large effects. Heterogeneity was assessed with Q and I2 statistics. Publication bias was evaluated with funnel plots and Egger's regression. Meta-regressions tested continuous moderators (mean age, % males, plasma ketamine concentration in ng/ml and infusion doses in mg/kg), and subgroup analyses examined route of administration (IV, IM, IN) and ketamine type (racemic versus S-ketamine). The authors also checked whether IV studies that measured psychotic-like symptoms differed in dose or plasma levels from those that did not.

From an initial pool of 123 potentially eligible studies, 56 placebo-controlled experimental studies met inclusion criteria and were retained for analysis, comprising 1175 participants (mean age 27.3 years; 63.3% male). Most studies administered ketamine intravenously (52 studies), with three intramuscular and one intranasal study. Ketamine formulation was reported as racemic in 11 studies, S-ketamine in 12, and unspecified in 33. Approximately half of IV studies measured psychotic-like symptoms; those that did not differ from those that did in mean plasma levels (135.4 ± 52.2 versus 135.8 ± 58.2 ng/ml; p = 0.983) and infusion doses (0.008 ± 0.004 versus 0.007 ± 0.003 mg/kg; p = 0.416). The primary pooled analysis included 244 cognitive measures and found that acute ketamine produced an overall moderate impairment in cognition, Cohen's d = 0.465, SE = 0.054, 95% CI 0.360 to 0.571, t = 8.65, p < 0.001. Heterogeneity was large (Q = 1019.4, p < 0.001; I2 = 74.3%). Funnel-plot inspection and Egger's test indicated potential publication bias (Egger's t = 14.95, p < 0.001), with smaller studies showing larger standardised effects. Domain-specific analyses showed significant ketamine-related deficits in most cognitive domains except response inhibition and visuospatial abilities. The largest impairments were observed for verbal learning (d = 0.714), social cognition (d = 0.670), verbal memory (d = 0.580) and visual learning/memory (d = 0.538). Moderate effects were reported for executive functions (d = 0.484), working memory (d = 0.490) and attention (d = 0.426). Smaller-to-moderate deficits were found for speed of processing (d = 0.360) and verbal fluency/language (approximate d = 0.312). Response inhibition and visuospatial domains did not show significant impairment. Meta-regression analyses identified a positive association between plasma ketamine concentration and the magnitude of overall and domain-specific cognitive deficits. Infusion dose was likewise positively associated with cognitive impairment. In contrast, mean age and percentage male in study samples were not associated with effect size. Subgroup comparisons found no statistically significant differences between racemic ketamine and S-ketamine (racemic d = 0.324, 95% CI -0.031 to 0.678; S-ketamine d = 0.631, 95% CI 0.253 to 1.008; Q = 1.08, p = 0.30) nor among routes of administration (IM, IV, IN; Q = 0.287, p = 0.87), though the authors caution that non-IV studies were few.

Zhornitsky and colleagues interpret the findings as evidence that acute ketamine administration in healthy volunteers produces broad cognitive impairment across multiple domains, with the most pronounced effects on verbal learning and memory and social cognition, and little or no effect on response inhibition and visuospatial ability. They note that the pattern of deficits parallels observations in chronic ketamine users and overlaps with cognitive impairments seen in schizophrenia, supporting the use of ketamine as an experimental model of psychosis. Dose dependence emerged as a key moderator: higher infusion doses and greater plasma concentrations were associated with larger cognitive deficits, consistent with the pharmacology of an agent whose effects range from mild psychoactivity to anaesthesia. The authors discuss implications for clinical practice, particularly in depression treatment where a common IV dose (0.5 mg/kg) produces plasma concentrations similar to many included studies; they emphasise that acute impairments observed in healthy volunteers may not directly translate to patients, since some clinical trials in treatment-resistant depression report mixed or transient cognitive effects and some uncontrolled studies even suggest later cognitive improvement. The authors acknowledge several limitations: evidence of publication bias skewed by small studies reporting large effects; a predominance of IV studies with only four non-IV studies, limiting route comparisons; small numbers of studies for some domains such as response inhibition, social cognition and visuospatial abilities, which could bias estimates; absence of a standardised study quality assessment; and substantial heterogeneity, implying that unmeasured factors contributed to variability. They therefore recommend cautious interpretation and call for further research, especially on social cognition and on cognitive outcomes in patient populations and different dosing regimens. The discussion also contrasts ketamine's transient cognitive effects with those reported for electroconvulsive therapy, noting that ECT-related deficits can be more prolonged. Finally, the authors highlight the clinical necessity of monitoring cognitive domains found to be most affected and of paying attention to dose and plasma levels in future studies.