Antidepressant effects of ketamine in depressed patients

A growing body of preclinical research implicates NMDA-class glutamate receptors in the pathophysiology of major depression and in the mechanism of action of antidepressant treatments. Animal studies have shown that NMDA receptor antagonists can produce antidepressant-like effects in many, but not all, paradigms, while chronic administration of conventional antidepressants produces changes in NMDA receptor function and subunit mRNA expression. Early clinical work with agents that modulate NMDA-related sites (for example, amantadine and D-cycloserine) provided some support for this hypothesis but was limited by design and drug-specific issues; overall clinical data on selective NMDA antagonists remain sparse. Berman and colleagues conducted the present study to test whether ketamine hydrochloride, a potent NMDA receptor antagonist, produces antidepressant effects in patients with major depressive episodes. The study aimed to compare a single low-dose intravenous ketamine infusion with saline in a randomised, double-blind, crossover design and to assess mood, depressive symptoms, perceptual effects and safety over the subsequent 72 hours.

The investigators recruited nine patients (4 men, 5 women; mean age 37 ± 10 years, range 23–56; 2 Hispanic, 7 Caucasian) who met DSM-IV criteria for a major depressive episode (eight with recurrent unipolar major depression, one with bipolar disorder, depressed). Two participants had remote histories of alcohol abuse in remission; one had comorbid panic disorder. All patients were medically healthy, drug-free and unmedicated for at least two weeks before the first treatment, as confirmed by medical assessment and toxicology. Written informed consent was obtained and the study received institutional review board approval. Participants underwent two treatment days separated by at least one week in a randomised, double-blind, crossover fashion. On each test day they received an intravenous infusion over 40 minutes of either ketamine hydrochloride at a total dose of 0.5 mg/kg or saline. Four participants were randomly assigned to receive ketamine on the first test day. Two participants terminated before completing the second treatment day (one before placebo, one before ketamine), leaving seven subjects who completed both conditions; the extracted text also describes some analyses that refer to eight subjects, an inconsistency in the reported denominators. Mood and symptom assessments were made at multiple time points. The primary clinician-rated measure was the 25-item Hamilton Depression Rating Scale (HDRS) at baseline and at 80 minutes, 230 minutes, 24 hours, 48 hours and 72 hours after infusion start. Self-report using the Beck Depression Inventory (BDI) was obtained at the same time points plus 10, 40 and 110 minutes. Perceptual/intoxication effects were measured with a visual analogue scale for feeling "high" (VAS-high) at baseline and at 10, 40, 80 and 110 minutes. The Brief Psychiatric Rating Scale (BPRS) was administered at baseline and 10, 40, 80, 110 and 230 minutes. Data were analysed with repeated-measures analyses of variance, using a Huynh–Feldt correction, testing condition (ketamine versus saline), time and condition-by-time interaction effects on dependent variables.

Analyses focused on the seven participants who completed both active and sham treatments; two patients withdrew prior to completing the second condition. Intravenous ketamine produced significantly greater reductions in HDRS scores than saline, with a significant condition-by-time interaction (F = 3.97, df 5,30, p = .02). Mean baseline HDRS scores reported in the extracted text were 33.0 ± 6.7 for the ketamine condition and 26.9 ± 5.8 for the placebo condition; this baseline difference was not statistically significant (paired t test, p = .10). The extracted text also notes one additional patient who completed only the active treatment and experienced marked improvement (baseline HDRS 33, final HDRS 16) but was excluded from the primary analyses. Self-reported depressive symptoms on the BDI showed robust decreases after ketamine (mean baseline 29.5 ± 8.2; final 16.8 ± 10.5) but not after saline (baseline 23.0 ± 8.2; final 25.2 ± 6.0). For the BDI there was a significant drug-by-time effect (F = 5.7, df 8,48, p = .0001). Reported response rates during the three-day follow-up period indicated that four of eight patients demonstrated ≥50% decreases in HDRS scores after active treatment versus one of eight after sham; however, the Fisher Exact test reported in the extracted text did not reach statistical significance (p > .05). Ketamine-associated mood improvement generally returned toward baseline by one to two weeks after infusion, although one subject who showed prolonged improvement began antidepressant medication before fully returning to baseline. Ketamine produced marked acute perceptual and euphoric effects. VAS-high scores were significantly greater after ketamine with significant condition-by-time, time and condition effects (condition-by-time F = 16.9, df 4,24, p = .0001). These subjective "high" ratings returned to baseline by 110 minutes post-infusion; all subjects showed maximum VAS-high scores >50 mm during ketamine testing and none did during sham. Ketamine also increased BPRS scores, particularly positive symptom items, but the extracted tests for BPRS-positive effects did not reach conventional significance (condition-by-time F = 4.14, df 5,30, p = .068; time F = 3.97, df 5,30, p = .06; condition F = 3.71, df 1,30, p = .10). Changes in BPRS or VAS-high did not correlate with percent decreases in HDRS (R2 < .05, p > .65), suggesting acute perceptual effects were temporally dissociated from antidepressant response. Item-level HDRS comparisons (uncorrected for multiple comparisons) indicated significant decreases during active treatment for depressed mood (p = .0025), suicidality (p = .02), helplessness (p = .008) and worthlessness (p = .015).

Berman and colleagues interpret their findings as evidence that a single low-dose intravenous ketamine infusion is associated with rapid and robust reductions in depressive symptoms, with improvement emerging over the three days following administration. They note that while ketamine is a high-affinity NMDA receptor antagonist, it also has lesser affinities for opioid receptors and weak dopamine-transporter antagonist activity; consequently, antidepressant effects might arise from direct NMDA antagonism or secondary effects on monoamine and opioid systems. The investigators emphasise that ketamine also produced profound but transient cognitive disturbances and euphoria, consistent with prior observations in other populations. Importantly, the mood improvements persisted beyond the short-lived perceptual "high," returning to baseline after one to two weeks in most patients, which the authors take as evidence that the antidepressant effect is temporally dissociated from acute intoxication. They caution, however, that the perceptual effects made it easy for participants to distinguish active from placebo treatment, compromising blinding and potentially biasing self-reports and placebo responses. The authors situate their results within animal-model data supporting NMDA antagonists' antidepressant-like effects and note that rapid onset might in part reflect the intravenous route of administration. They acknowledge limitations to clinical applicability, including ketamine's psychotomimetic effects and abuse potential, and suggest that NMDA antagonists without prominent psychotomimetic properties (for example, memantine or eliprodil) merit further clinical testing for antidepressant activity.