Ketamine safety and tolerability in clinical trials for treatment-resistant depression

Murrough and colleagues pooled patient-level data from three clinical trials conducted between 2006 and 2012 at two academic centres to evaluate the safety, tolerability, and acceptability of low-dose intravenous ketamine in treatment-resistant depression. The pooled dataset comprised 205 infusions of racemic ketamine (0.5 mg/kg administered over 40 minutes) given to 97 enrolled participants meeting DSM-IV criteria for major depressive disorder; because some individuals took part in more than one trial, 84 unique subjects are presented in certain analyses. Eligibility required failure to respond to at least two or three FDA-approved antidepressants per the Antidepressant Treatment History Form. Exclusion criteria included recent substance abuse or dependence, lifetime psychosis or any psychotic disorder, bipolar disorder, developmental disorder, recreational ketamine or phencyclidine use, uncontrolled hypertension, arrhythmia or other significant cardiac disease, and unstable medical illness. Subjects were washed off psychotropic medications 2–4 weeks prior to randomisation; physical exam, vital signs, ECG, blood tests, urinalysis, and urine toxicology were used to confirm medical stability and absence of illicit substances. Each infusion was administered under the supervision of an anaesthesiologist with standard telemetry (pulse oximetry, heart rate, blood pressure, end-tidal CO2). Behavioural and safety measures were collected before infusion, at 40 minutes (immediately post-infusion), and at 240 minutes; a subset had 24-hour assessments. Antidepressant response was defined as a ≥ 50% improvement on the Montgomery–Åsberg Depression Rating Scale (MADRS) at 24 hours. Psychotomimetic symptoms were measured with the 4-item positive symptom subscale of the Brief Psychiatric Rating Scale (BPRS+), dissociation with the Clinician-Administered Dissociative States Scale (CADSS), and general side effects with the SAFTEE-SI or PRISE instruments. A protocol allowed the anaesthesiologist to treat sustained blood pressure elevations (> 180/100 mm Hg) or heart rate > 110 bpm and to discontinue the infusion if three consecutive measurements remained above limits despite intervention. Longer-term follow-up was attempted by telephone at varying intervals (range 8 months to 6 years) to assess persistent adverse effects and acceptability; 46 of 84 unique subjects completed this interview. Statistical analyses were performed in SPSS v20 using repeated-measures ANOVA for within-subject changes, paired t tests to localise differences, Mann–Whitney U or independent t tests for between-group comparisons, χ2 tests for categorical associations, and Pearson correlations for relationships among continuous change scores.

Across the pooled sample, the 24-hour antidepressant response rate was 67% (65 of 97 participants), with a mean MADRS reduction of 19.0 ± 11.7 points from baseline. Responders were older on average than nonresponders (mean age 50.4 ± 11.7 years versus 43.2 ± 12.4 years; P = .01); other demographic and clinical characteristics did not differ significantly between groups in univariate analyses. Retention was high: overall attrition was 3.1% (3 of 97 subjects). Of 205 infusions, four (1.95% of infusions; affecting 4.1% of subjects) were discontinued because of adverse events. Two discontinuations were for antihypertensive-resistant blood pressure elevations (maxima 180/115 mm Hg and 187/91 mm Hg) that normalised after stopping the infusion. A third subject requested discontinuation for increased anxiety; a fourth experienced transient hypotension and bradycardia during blood draw and was hospitalised for overnight cardiac monitoring. That single serious adverse event (SAE) represented 0.49% of infusions. Adverse effects peaked within 120 minutes after infusion and largely resolved by 240 minutes and 24 hours. Mean peak systolic blood pressure during infusion was 141.9 ± 21.2 mm Hg, a mean increase of 19.6 ± 12.8 mm Hg (P < .001); mean peak diastolic pressure was 86.4 ± 12.7 mm Hg, a mean increase of 13.4 ± 9.8 mm Hg (P < .001). Transient tachycardia was also observed; there was no change in blood oxygen saturation (P = .3). Protocol-defined increases in blood pressure or pulse occurred in 29.8% of participants (25/84), and 14.3% (12/84) received medication intervention, typically a short-acting antihypertensive, for a mean duration of 14.4 ± 12.6 minutes. Hemodynamic changes were associated with body mass index ≥ 30 (P = .03) and with a higher absolute ketamine dose (mean 45.8 ± 9.3 mg versus 41.5 ± 11.1 mg; P = .03). Hemodynamic change, comorbid medical conditions, and mean ketamine dose did not differ between responders and nonresponders. Psychotomimetic symptoms measured by the BPRS+ increased slightly from 4.2 ± 1.3 at baseline to 4.5 ± 1.3 at 40 minutes (P = .001), a change the authors characterise as very mild on the scale range of 4–28. Dissociative symptoms measured by the CADSS increased from 0.7 ± 2.9 at baseline to 9.8 ± 13.9 at 40 minutes (P < .001); the mean peak CADSS sits in a range the authors describe as mild. For a subset of 72 infusions with 24-hour data, delayed effects did not differ significantly from baseline. Changes in CADSS and BPRS+ did not differ between responders and nonresponders (P = .2 for both), and Pearson correlations showed no association between change in MADRS score and change in CADSS or BPRS scores. Longer-term follow-up interviews were completed by 46 of 84 unique subjects at a mean interval of 2.9 ± 1.9 years (range 8 months to 6 years) from first exposure. Most participants did not report persistent physical, emotional, or psychological symptoms; one individual reported increased anxiety and dysphoria lasting just over 1 month. No participants reported increased craving for ketamine or increased illicit substance use. On a 5-point acceptability scale, mean acceptability was 3.7 ± 1.5 (median 4), with primary reasons for rating ketamine as "not at all acceptable" including lack of antidepressant benefit, increased anxiety, or dissociative effects.

Murrough and colleagues interpret these pooled data as indicating that subanaesthetic, low-dose intravenous ketamine (0.5 mg/kg over 40 minutes) is safe and generally well tolerated in patients with treatment-resistant unipolar depression when administered in a controlled medical research setting. The authors highlight a 67% 24-hour response rate and note that most general side effects and acute behavioural changes resolved within four hours of administration. A single SAE requiring overnight cardiac monitoring was reported; otherwise adverse effects tended to be transient. The investigators emphasise the relatively high prevalence of transient hemodynamic changes: nearly 30% of participants experienced protocol-defined increases in blood pressure or pulse and 14.3% required pharmacological intervention. These elevations were associated with obesity (BMI ≥ 30) and with larger absolute ketamine dose, prompting the suggestion that doses below 0.5 mg/kg may be appropriate for some patients and that optimal dosing requires further study. Small but statistically significant increases in psychotomimetic and dissociative symptoms were observed during infusion; the magnitude of psychotomimetic change was characterised as very mild while dissociative responses were mild on average but pronounced in a minority. Longer-term follow-up showed no evidence of persistent dissociative or psychotomimetic effects, nor of increased substance use among the participants who could be re-contacted. The authors acknowledge several important limitations. Combining patient-level data from studies with different designs, including both controlled and uncontrolled trials, limits confidence in pooled summary estimates. Adverse event rates were derived from two different assessment instruments, which may bias reporting toward shared items. The dataset reflects short-term administration and excludes patients with current substance use, lifetime psychotic or manic symptoms, unstable medical illness, or concomitant psychotropic treatment, limiting generalisability to broader clinical populations. The investigators also note the paucity of data on longer-term or repeated ketamine exposure, and on interactions with concurrent antidepressant medications. They call for additional research to characterise safety and efficacy over longer durations, to examine repeated-infusion effects on hemodynamics, and to determine optimal dosing strategies for individual patients. In conclusion, the study team presents these data as the first systematic analysis of acute and some longer-term neuropsychiatric and physical adverse effects of ketamine in treatment-resistant depression, concluding that, within a controlled research environment, subanaesthetic ketamine carries a low and acceptable short-term risk profile but requires adequate monitoring and further study to define long-term safety.