Do the dissociative side effects of ketamine mediate its antidepressant effects?

A single subanesthetic infusion of ketamine produces rapid antidepressant effects in major depressive disorder (MDD) and bipolar depression that can last about one to two weeks. However, ketamine commonly causes transient dissociation, psychotomimetic symptoms and sympathomimetic hemodynamic changes (for example, raised blood pressure). Other NMDA receptor antagonists that lack these subjective and physiological side effects have shown antidepressant effects too, but generally with smaller magnitude than ketamine. This pattern leaves uncertain whether ketamine’s dissociative, psychotomimetic or sympathomimetic effects are necessary for, or simply accompany, its antidepressant efficacy. Luckenbaugh and colleagues set out to test whether acute sympathomimetic (vital sign) changes and hypoglutamatergic effects manifested as psychotomimetic or dissociative symptoms are associated with ketamine’s antidepressant response. The investigators hypothesised that greater acute sympathomimetic and hypoglutamatergic effects would correlate with larger reductions in depressive symptoms after a single subanesthetic ketamine infusion.

The analysis pooled data from 108 treatment-resistant inpatients in a current major depressive episode (74 with MDD; 34 with bipolar disorder), all without psychotic features and diagnosed using the Structured Clinical Interview for DSM‑IV. Participants were required to have at least moderate symptom severity at screening (≥18 on the 21-item Hamilton Depression Rating Scale [HDRS] or ≥20 on the MADRS) and a history of at least one failed antidepressant trial. Exclusion criteria included lifetime psychotic-spectrum disorder, recent substance dependence/abuse and unstable serious medical illness. Subjects were psychotropic-free for at least two weeks prior to infusion (five weeks for fluoxetine), except that bipolar patients remained on therapeutic lithium or valproate levels. Each subject received a single intravenous infusion of ketamine 0.5 mg/kg over 40 minutes. Psychiatric ratings were obtained with the HDRS (depression), Young Mania Rating Scale (YMRS), Brief Psychiatric Rating Scale (BPRS) and the Clinician‑Administered Dissociative States Scale (CADSS). Vital signs (systolic and diastolic blood pressure, pulse) were recorded every 5 minutes during the first 40 minutes using a standard monitor with the subject reclining. Psychiatric assessments were performed at baseline and at 40, 80, 110 and 230 minutes after infusion start, with additional post-infusion time points including Days 1 and 7. For statistical analyses, linear mixed models with restricted maximum likelihood estimation and a compound symmetry covariance structure examined changes in vital signs over the first 40 minutes and clinical ratings over the first 230 minutes, adjusting for phase-specific baseline and including drug-by-time interactions. Bonferroni-adjusted post hoc tests examined time-point differences. Using all ketamine-treated subjects, Pearson correlations were computed between absolute changes from baseline to 40 minutes in CADSS, BPRS (total and positive symptoms), YMRS and vital signs, and percent change in HDRS at 230 minutes, Day 1 and Day 7. Significance was set at two-tailed p<0.05.

All 108 participants received a single ketamine infusion; patients randomised to adjunctive riluzole were excluded from analyses at Days 1 and 7. The sample had moderate-to-severe current depression with a mean episode duration of 55.7 months (SD=97.2), was 85% Caucasian (n=92) and approximately half female (n=54). A linear mixed model found a significant drug-by-time interaction showing HDRS scores were lower following ketamine than placebo from 40 to 230 minutes (F=3.14, df=3,333, p=0.026). Acute subjective and physiological effects peaked during or shortly after infusion. CADSS (dissociation), YMRS and BPRS positive symptom scores were significantly increased on ketamine at 40 minutes only. Significant increases in diastolic blood pressure (F=264.95, df=1,640, p<0.001), systolic blood pressure (F=429.80, df=1,652, p<0.001) and pulse (F=175.22, df=1,609, p<0.001) were observed on ketamine versus placebo from 5 to 40 minutes. Correlation analyses examined whether these acute changes predicted antidepressant outcome. Greater increase in CADSS at 40 minutes was significantly associated with greater percent improvement in HDRS at 230 minutes (r= -0.35, p=0.007) and at Day 7 (r= -0.41, p=0.01), but not at Day 1 (r= -0.21, p=0.18). (As reported, the negative correlation indicates that larger intra-infusion increases in CADSS were linked to larger reductions in HDRS scores.) Changes in YMRS, BPRS total or BPRS positive symptoms at 40 minutes were not significantly correlated with HDRS percent change at any examined time point. Likewise, changes in systolic and diastolic blood pressure and pulse were not associated with depressive symptom change. Additional correlations showed that CADSS change at 40 minutes correlated positively with BPRS positive symptoms (r=0.31, p=0.02) and BPRS total (r=0.34, p=0.008), but not with YMRS (r=0.22, p=0.11). The strongest observed association among acute measures was between BPRS positive symptoms and YMRS (r=0.63, p<0.001). No significant relationships emerged between vital sign changes and CADSS, YMRS or BPRS measures. Finally, ketamine and norketamine blood levels did not correlate significantly with changes in depression or dissociation.

Luckenbaugh and colleagues interpret the results as showing that, among several acute effects of ketamine, only dissociative side effects were associated with a more robust and sustained antidepressant response. Ketamine produced expected transient increases in pulse, blood pressure, psychotomimetic and dissociative symptoms; however, only acute dissociation correlated with HDRS improvement at 230 minutes and Day 7. The investigators propose that acute dissociation might serve as a clinical biomarker predicting ketamine efficacy, while psychotomimetic and sympathomimetic effects do not. The authors discuss possible mechanistic differences underlying these effects. They note that psychotomimetic and hemodynamic effects have been linked in some work to dopamine release, whereas dissociation may reflect ketamine-induced enhancement of glutamate release resulting from disinhibition of cortical GABAergic interneurons. Under the dominant mechanistic hypothesis for ketamine, this glutamate surge increases AMPA-to-NMDA throughput and synaptic potentiation; individuals who manifest greater dissociation might have larger presynaptic glutamate responses and thus greater antidepressant benefit. The paper also notes evidence that dissociative and psychotomimetic symptoms may involve different brain regions, suggesting avenues for mechanistic studies. The authors acknowledge several limitations. This was a secondary analysis combining data from both unipolar and bipolar participants and from double-blind and one open-label study, which may affect generalisability. Adequacy of blinding is a major concern because ketamine’s conspicuous subjective and physiological effects could have unblinded participants and raters; the investigators suggest future studies might use active controls with comparable dissociative or hypertensive effects but no antidepressant activity. The extracted text also notes that prior studies have reported discrepant correlations, potentially due to small samples, mixed diagnoses, mandatory treatment resistance, or differing ketamine doses. Importantly, although the CADSS–HDRS association was statistically significant, CADSS change accounted for only a fraction of the variance in antidepressant response, so these data do not establish that intra-infusion dissociation is necessary or causally responsible for ketamine’s antidepressant effect. The authors conclude that prospective mechanistic investigations are needed to disentangle these relationships.