Identification of an optimal dose of intravenous ketamine for late-life treatment-resistant depression: a Bayesian adaptive randomization trial

This double-blind, randomized study (n=33) sought to identify the optimal dose of intravenous ketamine for late-life (mean age=62) treatment-resistant depression (TRD). Varying doses of ketamine (0.1 mg/kg-0.5 mg/kg) were compared to an active placebo (midazolam 0.03 mg/kg). It was found that 0.5 mg/kg is an effective initial IV ketamine dose in TRD.

Abstract

“Evidence supporting specific therapies for late-life treatment-resistant depression (LL-TRD) is necessary. This study used Bayesian adaptive randomization to determine the optimal dose for the probability of treatment response (≥50% improvement from baseline on the Montgomery-Åsberg Depression Rating Scale) 7 days after a 40 min intravenous (IV) infusion of ketamine 0.1 mg/kg (KET 0.1), 0.25 mg/kg (KET 0.25), or 0.5 mg/kg (KET 0.5), compared to midazolam 0.03 mg/kg (MID) as an active placebo. The goal of this study was to identify the best dose to carry forward into a larger clinical trial. Response durability at day 28, safety and tolerability, and effects on cortical excitation/inhibition (E/I) ratio using resting electroencephalography gamma and alpha power, were also determined. Thirty-three medication-free US military veterans (mean age 62; range: 55-72; 10 female) with LL-TRD were randomized double-blind. The trial was terminated when dose superiority was established. All interventions were safe and well-tolerated. Pre-specified decision rules terminated KET 0.1 (N = 4) and KET 0.25 (N = 5) for inferiority. Posterior probability was 0.89 that day-seven treatment response was superior for KET 0.5 (N = 11; response rate = 70%) compared to MID (N = 13; response rate = 46%). Persistent treatment response at day 28 was superior for KET 0.5 (response rate = 82%) compared to MID (response rate = 37%). KET 0.5 had high posterior probability of increased frontal gamma power (posterior probability = 0.99) and decreased posterior alpha power (0.89) during infusion, suggesting an acute increase in E/I ratio. These results suggest that 0.5 mg/kg is an effective initial IV ketamine dose in LL-TRD, although further studies in individuals older than 75 are required.”

Authors: Marijn Lijffijt, Nicholas Murphy, Sidra Iqbal, Charles E. Green, Tabish Iqbal, Lee C. Chang, Colin N. Haile, Nithya Ramakrishnan, Dylan A. Fall, Alan C. Swann, Rayan K. Al Jurdi & Sanjay J. Mathew

Summary

INTRODUCTION

Monoaminergic drugs are considered first-line treatment for major depressive disorder, but can take eight weeks or longer to achieve a response. Late-life treatment-resistant depression is an unmet need.

A single 40 min IV infusion of 0.5 mg/kg of KET can achieve antidepressant effects within 24 h in adult TRD. This effect is mediated by an initial temporary block of NMDARs located on fast firing, inhibitory GABAergic parvalbumin (PV+) interneurons with subsequent disinhibition of glutamate release.

Four studies have examined the clinical effects of KET in individuals with LL-TRD, with mixed results. Little is known about optimal, safe doses for LL-TRD, and NMDAR receptor engagement of KET in older patients.

This study used Bayesian adaptive randomization to determine the optimal dose of KET for LL-TRD, and used change in Montgomery-sberg Depression Rating Scale (MADRS) score at seven days post-infusion as the primary endpoint. The study was stopped for superiority if salient pre-specified criteria were met.

In a double-blind study, participants with LL-TRD received a single IV infusion of KET at 0.5, 0.25, or 0.1 mg/kg against a psychoactive control. KET 0.5 mg/kg outperformed the alternative conditions at the primary endpoint.

Primary endpoint

Figure 1 shows MADRS total scores as a function of condition and time, proportions of participants with a treatment response, and posterior distributions of the probability of a day-seven treatment response.

Sixteenof33randomizedparticipants achieved a day-seven treatment response, and all day-seven responders but one (MID) achieved day-seven remission. KET 0.5 showed a higher probability of achieving a day-seven treatment response than MID, and KET 0.25 showed minimal evidence of differing from MID.

Durability

KET 0.5 had superior response durability, with 7 of 8 patients continuing to meet response criteria at day 28, compared to 1 of 2 patients for KET 0.25 and 2 of 6 patients for MID.

Safety and tolerability

Pre-existing medical conditions were moderate to low across interventions. One participant made a suicide attempt four days after MID infusion, which was determined to be unrelated to the intervention.

Resting state EEG

The complete analytical report is summarized in Supplementary Table 3. Figure 2 presents frontal gamma power, posterior alpha power, and posterior alpha peak frequency as functions of condition and time.

Day 1 model showed strong evidence for a time-by-condition interaction for gamma power. KET 0.5 showed an initial increase followed by a decrease continuing until the 240-min measurement.

24-h model: frontal gamma power appeared comparable for KET0.5andMID, and durability model: moderate certainty of an increase in gamma power.

24-h model: KET 0.5 had high probability of greater alpha power than MID, but evidence for time-by-condition interaction was weak.

Day 1 model showed moderate evidence for a time-by-condition interaction for posterior alpha peak frequency. Evidence for an initial decrease followed by an increase in alpha peak frequency for MID was moderately strong.

The evidence for an interaction between time and condition on day one was weak. There was strong evidence that the geometric mean of plasma BDNF was higher for KET 0.25 and KET 0.1 than MID on day-seven.

Exploratory correlational analysis

A recent study showed that baseline gamma power was related to clinical response to KET in patients with TRD, and that greater gamma power reactivity was associated with a stronger reduction in MADRS at 7 days post infusion for KET.

DISCUSSION

In this study, we compared three doses of IV KET to midazolam (0.03 mg/ kg) in a sample of 62 adults with LL-TRD. KET 0.5 proved superior to the other conditions, and responders to KET 0.5 at day-seven remained responders at day 28 follow-up.

Improvement is often inadequate for functional recovery, but remission to KET 0.5 mg/kg is a positive indicator of possible sustained remission at later measurements.

The gamma frequency band of the EEG showed a sharp increase over the course of KET infusion, which then gradually reset over a period of four hours, indicating enhanced cortical excitability relative to inhibition after blockade of NMDAR. In earlier work, baseline gamma power moderated clinical response to KET at 230 minutes. In this study, increased reactivity at 30 minutes from baseline was associated with decreased depression severity at day-seven post-infusion for KET 0.5 but not MID.

Resting state alpha oscillations reflect the ability of an upstream system to exert control over the gain measured in downstream local neuronal clusters. KET 0.5 and MID suppressed posterior resting state alpha power at the end of infusion consistent with prior studies.

MID and KET are both positive allosteric modulators of GABAA receptors located on glutamate receptors, but MID had a low probability of a sustained antidepressant effect at day 28. This may be a placebo effect, but MID may also have antidepressant properties.

The current study used Bayesian methods to determine superiority across doses, and used psychotropic medication-free TRD patients to ensure uncontaminated interpretation of treatment response and neurophysiological measures.

Limitations of this study include that patients were U.S. military veterans and predominantly male, that there was a lack of enrollment of participants from the oldest age cohort, that EEG and BDNF outcomes should be interpreted with caution, and that aspects of illness course may have influenced response across treatment conditions. This highlights a design limitation in clinical trials that prevents simultaneous collection of data in adequate quantity for detailed understanding of biomarkers.

A single infusion of KET 0.5 mg/kg was superior to lower doses and midazolam in achieving a treatment response at day-seven and in maintaining a treatment response up to day 28.