This midazolam-controlled randomized clinical trial study (n=16) compares ketamine and midazolam (another anesthetic) to investigate their feasibility and effect on suicidal ideation (SI) in bipolar depression (BD) using the Scale for Suicidal Ideation (SSI) score. The results substantiate feasibility with improvement in memory and BDNF as promising biomarkers and that ketamine effects showing more reduction in suicidal thoughts than with midazolam, although it was not statistically significant; this was possibly due to a small sample so a full-scale trial is needed.

Abstract

“Objectives: To evaluate feasibility and effects of a sub-anesthetic infusion dose of ketamine versus midazolam on suicidal ideation in bipolar depression. Neurocognitive, blood and saliva biomarkers were explored.

Methods: Sixteen participants with bipolar depression and a Scale for Suicidal Ideation (SSI) score of ≥4 were randomized to ketamine (0.5 mg/kg) or midazolam (0.02 mg/kg). Current pharmacotherapy was maintained excluding benzodiazepines within 24 hours. The primary clinical outcome was SSI score on day 1 post-infusion.

Results: Results supported feasibility. Mean reduction of SSI after ketamine infusion was almost 6 points greater than after midazolam, although this was not statistically significant (estimate=5.84, SE=3.01, t=1.94, P=.074, 95% confidence interval ([CI)]=−0.65 to 12.31). The number needed to treat for response (SSI <4 and at least 50% below baseline) was 2.2, and for remission (SSI=0) was 3.2. The strongest neurocognitive correlation was between memory improvement on the Selective Reminding Test (SRT) and reduction in SSI score on day 1 after ketamine (ρ=−.89, P=.007). Pre- to post-infusion decrease in serum brain derived neurotrophic factor (BDNF) correlated with reduction in SSI from baseline to day 1 after ketamine (n=5, ρ=0.90, P=.037) but not midazolam (P=.087).

Conclusions: The study demonstrated feasibility. Suicidal thoughts were lower after ketamine than after midazolam at a trend level of significance, likely due to the small pilot sample. Memory improvement and BDNF are promising biomarkers. Replication is needed in an adequately powered full-scale trial.“

Authors: Michael F. Grunebaum, Steven P. Ellis, John G. Keilp, Vivek K. Moitra, Thomas B. Cooper, Julia E. Marver, Ainsley K. Burke, Matthew S. Milak, M. Elizabeth Sublette, Maria A. Oquendo & J. John Mann

Summary

1 | INTRODUCTION

There are few effective and well- tolerated treatments for bipolar depression, and there are no treatments for suicidal ideation or behavior in bipolar depression. Observational data suggest lithium has anti-suicidal effects, but not all clinical trials support this.

We conducted a pilot trial to compare the effects of ketamine and midazolam on suicidal ideation and global depression in bipolar depression patients.

2.1 | Patients

Patients with a DSM- IV BD, a current major depressive episode, a score of 4″ on the Scale for Suicidal Ideation, and a voluntary admission to an inpatient research unit at New York State Psychiatric Institute were eligible for inclusion.

2.2 | Intervention

Participants received an intravenous infusion of racemic ketamine hydrochloride 0.5 mg/kg or midazolam 0.02 mg/kg over 40 minutes. They were allowed to continue current psychotropics except for benzodiazepines within 24 hours of the infusion.

2.3 | Outcome and measures

The clinician-rated SSI16 assessed suicidal ideation and was assessed at screening, baseline, 230 minutes post-infusion, day 1, and weeks 1 to 16 of follow- up.

Depressive symptoms were assessed with the HDRS- 17,15, Beck Depression Inventory (BDI)22, and Profile of Mood States (POMS), and anxiety symptoms were measured with the YMRS.

Participants performed a neurocognitive battery at baseline and day 1 and provided two saliva samples upon awakening and 30 minutes later to measure cortisol awakening response.

We measured plasma ketamine and metabolites after ketamine infusion and serum BDNF before and after BDNF infusion.

2.4 | Randomization and blinding

We used a permuted blocked design with 1:1 assignment between treatments and block size randomized between 4 and 6 with equal probability. Participants, psychiatrists, and assessors were blind to treatment.

2.5 | Statistical methods

We used R software and SPSS version 23 to analyze all randomized participants and predict day 1 SSI using baseline SSI and treatment.

We tested treatment effect on response and remission, used regression models to analyze depression, mania, and anxiety, and tested the correlation between change in depressive symptoms and change in SSI. We also investigated safety by testing relationships between clinical variables and CAR, serum BDNF, and post- infusion plasma ketamine.

3.1 | Study patients

16 patients were randomized to receive a lithium infusion. They did not differ in age, sex, race, frequency of past suicide attempts, SSI, HDRS- 17 or BDI score.

3.2.1 | Acute response

After adjusting for baseline, mean SSI on day 1 was almost 6 points lower after ketamine than after midazolam, with no treatment by time interaction.

Of those randomized to ketamine, four of seven were day 1 responders and three of seven were remitters versus one of nine randomized to midazolam.

3.3 | Depression symptoms

Ketamine reduced depressive symptoms by 6 points on day 1 compared to midazolam, but this was not statistically significant. Ketamine also reduced anxiety symptoms by 11 points on day 1, although this was not statistically significant.

Ketamine did not show a treatment effect at 230 minutes post-infusion, but did show a trend toward improvement at day 1 compared to midazolam.

3.4 | Open- label ketamine response

All midazolam non-responders received an open-label ketamine infusion. The mean SSI, HDRS, and YMRS scores improved from pre- to post-infusion day 1.

3.5 | Neurocognition

On day 1, both groups improved in reaction time, attention, and memory. Reaction time improved more in the midazolam group, and language fluency remained stable in the midazolam group but declined in the ketamine group.

In the midazolam group, poorer memory encoding at baseline correlated with worse HDRS- 17 and SSI scores on day 1.

3.6 | Exploratory biomarkers

In both randomized groups, serum BDNF decreased from pre- to post-infusion, irrespective of treatment. Neither baseline cortisol nor CAR correlated with SSI or HDRS-17.

3.8.2 | Cardio- respiratory effects

There were no serious cardio- respiratory adverse events. Ketamine was associated with a mean increase in blood pressure of 23 12.8 points versus 3 10.2 points for midazolam.

Suicidal ideation severity changed from baseline to 24 hours after intravenous ketamine infusion in bipolar depressed participants.

Figure 3 shows that ketamine reduced suicidal ideation in bipolar depressed patients after 230 minutes, day 1, week 1, week 7, and week 5.

Ketamine and midazolam both caused decreases in respiratory rate, but midazolam caused a larger decrease.

3.8.3 | Dissociative, psychotomimetic and other adverse effects

There were no infusion-related dissociative or psychotomimetic serious adverse events, except for one participant with “very mild” conceptual disorganization immediately post- infusion.

3.8.4 | Serious adverse events

Six SAEs occurred, including increased suicidal thoughts and a suicide attempt, during follow-up months 2-6. One participant’s depression and suicidal thoughts remitted over the following 24 hours.

4 | DISCUSSION

In this pilot study, bipolar depressed patients with clinically significant suicidal thoughts received a ketamine vs midazolam infusion. The results were consistent with previous studies, but the difference was not statistically significant.

The results are consistent with prior studies of ketamine in bipolar depressed participants on lithium or valproate, or in a mixed mood and anxiety sample, although the latter results were not reported by diagnosis.

Ketamine improved memory and reduced suicidal thoughts in patients with treatment-resistant depression, but suppressed improvement in reaction time and language fluency compared to midazolam. Ketamine has been associated with cognitive impairment, but no residual cognitive impairment was found 3 days after an infusion in 54 healthy controls.

Ketamine’s antidepressant effects may involve upregulation of BDNF in the brain, but a decrease in peripheral serum BDNF did not correlate with a reduction in SSI after ketamine.

This pilot study indicated that ketamine infusions were feasible and led to few adverse outcomes in suicidal, depressed BD patients. A larger RCT is warranted to test these effects.

DISCLOSURES

The authors have received royalties from the Research Foundation for Mental Hygiene and own stock in Bristol Myers Squibb.