Relationship of ketamine’s antidepressant and psychotomimetic effects in unipolar depression

Sos and colleagues situate their work within growing evidence that ketamine, an NMDA (N-methyl-D-aspartate) receptor antagonist, produces rapid antidepressant effects at subanesthetic doses while also causing transient psychotomimetic and dissociative phenomena. Earlier research has established ketamine's rapid mood benefits but left the mechanistic link between those antidepressant effects and the concurrent psychotomimetic effects unclear. The authors note that industry efforts have sought NMDA antagonists that retain antidepressant activity without psychotomimetic effects, but the association between these two domains had not been prospectively explored. This study aimed to evaluate ketamine’s antidepressant efficacy versus placebo in inpatients with major depressive disorder and, a priori, to test whether greater psychotomimetic effects during infusion predict larger subsequent antidepressant responses. The investigators also measured ketamine and nor-ketamine plasma levels to explore whether blood concentrations related to either psychotomimetic or antidepressant effects. The work used a double-blind, placebo-controlled, crossover design to address these questions prospectively in a clinical sample of hospitalised patients with moderate-to-severe depression.

Participants were right-handed, ketamine‑naive inpatients aged 18–65 with DSM‑IV major depressive disorder confirmed by the M.I.N.I., and with baseline Montgomery‑Åsberg Depression Rating Scale (MADRS) scores of at least 20. All were on stable antidepressant medication for at least three weeks prior to admission and remained on the same regimens during the study. Exclusion criteria included suicidal risk, current Axis I or II comorbidity, serious unstable medical or neurological illness, lifetime personal or first/second‑degree family history of psychotic disorder, and recent electroconvulsive therapy. The study was approved by the institutional review board and registered (EudraCT number provided). The trial used a two‑week, double‑blind, placebo‑controlled, crossover format with sessions one week apart. Each participant received one ketamine and one placebo (0.9% saline) infusion in randomised order. Ketamine (racemic hydrochloride) was given intravenously as a 30‑minute regimen comprising a 0.27 mg/kg loading dose over 10 minutes followed by 0.27 mg/kg over 20 minutes (total 0.54 mg/kg). Infusions took place 08:00–10:00 and venous blood for ketamine and nor‑ketamine was drawn 5 minutes before and at 10 and 30 minutes after infusion start. Psychotomimetic symptoms were assessed with the Brief Psychiatric Rating Scale (BPRS) before and 30 minutes after infusion; MADRS was rated at baseline and at days 1, 4 and 7 after each session. Two experienced raters performed MADRS assessments after training to an intra‑class correlation >0.80. Ketamine and nor‑ketamine serum concentrations were measured by a validated gas chromatography–mass spectrometry (GC‑MS) method with internal standardisation (ketamine‑D4); lower limits of quantification were 50 ng/ml for ketamine and 8 ng/ml for nor‑ketamine. Statistical analyses used a modified intention‑to‑treat (ITT) set including patients with baseline and at least one post‑baseline visit after crossover. A general linear model for a two‑period crossover design tested treatment, period, sequence and time effects, with BPRS change during ketamine infusion included as a covariate; Bonferroni post‑hoc tests were applied. Prescott’s test evaluated binary responder status (≥50% MADRS reduction) across treatments. Pearson correlations were used to examine associations between BPRS change during infusion and MADRS change at days 1, 4 and 7, and between plasma levels and psychometric changes.

Thirty‑eight patients were screened; 30 met inclusion criteria and were randomised (flip of a coin) and 27 received intended treatment and were included in ITT analyses (9 in the ketamine→placebo sequence, 18 in the placebo→ketamine sequence). Three participants discontinued before completing both periods (two owing to worsening after placebo; one maintained placebo response and did not receive ketamine). The two sequence groups differed on baseline MADRS (t=2.23, p=0.03) but were otherwise comparable on reported demographics and clinical variables. Safety observations indicated that ketamine was generally well tolerated. Expected subanesthetic effects—dissociation/perceptual disturbances, transient confusion, mild blood pressure increases, emotional blunting and euphoria—occurred but resolved quickly in most participants, typically within 30 minutes and never beyond 60 minutes. No serious adverse events were reported. Efficacy analyses using the crossover general linear model showed a significant treatment effect (F(1,24)=5.87, p=0.03), a time effect (F(3,72)=5.58, p=0.002) and a treatment×time interaction (F(3,72)=4.11, p=0.01), with no significant sequence or period carry‑over effects. Post‑hoc comparisons found ketamine superior to placebo at all post‑infusion visits: day 1 (p<0.001), day 4 (p=0.002) and day 7 (p=0.02). Mean between‑treatment MADRS score differences (ketamine minus placebo) were 5.7 points (95% CI 3.4–7.9) at day 1, 4.7 points (95% CI 2.5–7.0) at day 4 and 4.0 points (95% CI 1.8–6.2) at day 7. Cohen’s d effect sizes were reported as 0.62, 0.57 and 0.44 at days 1, 4 and 7, respectively. Responder analyses (≥50% MADRS reduction) showed significantly more responders after ketamine than after placebo: day 1 ketamine 10/27 (37.0%) versus placebo 1/27 (3.7%), Prescott’s test p=0.008; day 4 ketamine 11/27 (40.7%) versus placebo 1/27 (3.7%), p=0.003; day 7 ketamine 10/27 (37.0%) versus placebo 3/27 (11.1%), p=0.02. Ten patients were responders to ketamine on at least two visits and five maintained response from day 1 through day 7; no placebo recipients showed that persistence. Regarding the relationship between psychotomimetic and antidepressant effects, BPRS change during ketamine infusion correlated with MADRS improvements at day 7 (Pearson r=−0.40, p=0.04) and showed trends at day 1 (r=−0.37, p=0.06) and day 4 (r=−0.36, p≈0.07). Analyses of BPRS subscales did not yield significant correlations with MADRS changes. Plasma assays showed undetectable ketamine and nor‑ketamine at baseline for both sessions and measurable increases after infusion: ketamine 10 min 306±136 ng/ml and 30 min 237±95 ng/ml; nor‑ketamine 10 min 11±7 ng/ml and 30 min 50±21 ng/ml. No differences in ketamine or nor‑ketamine concentrations were found between clinical responders and non‑responders, and there were no correlations between total BPRS change and plasma levels. One significant correlation emerged between the BPRS Anergia Factor (items reflecting emotional withdrawal, motor retardation, blunted affect and disorientation) and nor‑ketamine level at 10 minutes (r=−0.47, p<0.05).

Sos and colleagues interpret their findings as evidence of a substantive association between the acute psychotomimetic effects of ketamine and its subsequent antidepressant benefit. They highlight that the intensity of transient altered mental state during infusion, indexed by BPRS change, predicted a greater reduction in core depressive symptoms over the following week, with the strongest correlation observed at day 7. The magnitude and character of psychotomimetic effects in this sample were similar to prior ketamine studies in unipolar and bipolar depression. The authors situate their results within mechanistic hypotheses linking NMDA antagonism to downstream signalling changes thought to mediate antidepressant action, notably increased glutamate throughput to AMPA receptors; they also mention that changes in extracellular glutamate or serotonin in prefrontal cortex could underlie psychotropic effects. The absence of correlations between ketamine or nor‑ketamine blood levels and either psychotomimetic or antidepressant responses is taken to support the view that early receptor interactions trigger downstream processes not directly indexed by plasma concentrations. The single finding of an inverse correlation between nor‑ketamine at 10 minutes and the BPRS Anergia Factor is noted and discussed in the context of NMDA blockade producing negative‑symptom‑like effects. Several limitations acknowledged by the investigators temper interpretation. The sample size, while comparable with other randomised ketamine trials, remains modest. The use of an inactive saline placebo could have compromised blinding because it lacks psychotomimetic properties; the authors note that an active control would have been preferable. Randomisation by coin flip rather than block randomisation produced unequal group sizes and a baseline MADRS imbalance between sequences. Finally, the authors call for further research to test whether individual sensitivity of NMDA receptors predicts acute subjective effects and longer‑term antidepressant outcomes.

The study concludes that there is a substantial relationship between ketamine’s transient psychotomimetic effects and its antidepressant efficacy following a single subanesthetic infusion. The authors propose that both clinical effects may arise from early actions at NMDA receptors and that these downstream antidepressant processes are not directly reflected in ketamine or nor‑ketamine plasma levels. They recommend further studies to determine whether NMDA receptor sensitivity can predict acute subjective responses and antidepressant outcomes.