A randomized, double-blind, active placebo-controlled study of efficacy, safety, and durability of repeated vs single subanesthetic ketamine for treatment-resistant depression

Ketamine, a glutamate receptor antagonist approved for anesthesia, has been investigated for rapid antidepressant effects in treatment-resistant depression (TRD). A single subanesthetic ketamine infusion (0.5 mg/kg over 40 min) can improve mood within hours, with peak benefit at approximately 24 hours and waning effects over 5–8 days. Open-label and saline-controlled studies suggested that multiple ketamine infusions may enhance response rates and prolong durability relative to a single infusion; however, consensus guidelines have noted insufficient evidence to support frequent ketamine administration. This study aimed to rigorously compare the efficacy, safety, and durability of six repeated ketamine infusions versus a single ketamine infusion (after five active placebo infusions with midazolam) in adults with TRD.

Prior studies of ketamine in TRD demonstrated rapid antidepressant effects after a single subanesthetic infusion, peaking at 24 hours and diminishing within one week. Open-label and saline-controlled trials suggested that repeated ketamine infusions could yield higher response/remission rates and longer-lasting benefits than a single infusion, but methodological limitations (open-label designs, inadequate blinding with saline) tempered conclusions. Consensus guidelines accordingly cited a lack of robust evidence to endorse frequent ketamine dosing, underscoring the need for randomized, controlled trials with active placebo to assess efficacy, safety, and durability of repeated ketamine.

Design: Randomized, double-blind, active placebo-controlled trial conducted at the Minneapolis VA Medical Center (April 2015–March 2019). Participants were randomized (permuted blocks) to: (a) six ketamine infusions, or (b) five midazolam infusions followed by a single ketamine infusion, over 12 days (Monday–Wednesday–Friday schedule). The final infusion was ketamine for both arms and remained blinded to subjects and raters at the 24-hour post-last-infusion assessment. Participants: Outpatients aged 18–75 with DSM-IV MDD (SCID), current major depressive episode, and nonresponse to at least two adequate antidepressant trials (ATHF). Antidepressant regimens (including augmentation and psychotherapy) were stable for ≥6 weeks pre-consent and throughout the study. Exclusions included lifetime PTSD, mild–moderate TBI, psychotic disorders, bipolar disorder, other primary Axis I disorders, recent (≤6 months) substance use disorder, imminent suicide/homicide risk with intent, MMSE ≤27, positive urine toxicology or pregnancy, and unstable medical conditions. Interventions: Ketamine 0.25 mg/kg IV over 40 minutes; midazolam 0.045 mg/kg IV over 40 minutes. Vital signs and safety were closely monitored during and after infusions. Discharge criteria required resolution of clinically significant side effects per modified Aldrete score ≥9. Standard rescue medications and safety procedures were available. Assessments: Efficacy and safety assessments on days 0, 1, 3, 5, 8, 10, and 12 during the infusion phase; post-treatment follow-up weekly for 4 weeks, every 2 weeks for 8 weeks, then every 4 weeks for 12 weeks (up to 6 months). Primary outcome: change in MADRS score at 24 h after the last infusion (T+24). Secondary outcomes: MADRS change over time; response (≥50% reduction from baseline) and remission (MADRS <10) rates and their trajectory; durability of response up to 6 months; CGI-Severity/Improvement; pain NRS; BAI; CEQ. Side effects: PRISE; psychotic symptoms (BPRS positive subscale); dissociation (CADSS); mania (YMRS); suicide risk (C-SSRS). Rater training for MADRS achieved interrater reliability of 0.92. Sample size and statistics: Power analysis (ANCOVA with treatment, time, and interaction; 5% alpha) informed by prior open-label six-infusion data indicated 21 patients per group to detect a 10-point between-group difference in MADRS change. Repeated measures ANOVA tested differences from baseline to T+24 after last infusion and after the fifth infusion. Multilevel models compared groups across all MADRS measures. Two-sided tests with p < 0.05. Ethics approval by Minneapolis VA HCS IRB (protocol 4533-B).

Participants: 178 screened; 62 consented; 58 randomized; 54 initiated treatment and completed primary endpoint (29 in five midazolam + single ketamine; 25 in six ketamine). Baseline demographics and clinical characteristics were similar between groups. Primary outcome: No significant difference between groups in MADRS change at 24 h after the last infusion (i.e., after six total infusions; both arms received ketamine as the final infusion). Trajectory analyses: Across all MADRS assessments, there was a significant effect of group over time. The six-ketamine group showed significantly lower MADRS scores than the midazolam group prior to infusion 5, at 24 h after infusion 5, and prior to infusion 6. Response/remission: Response and remission rates did not differ significantly between groups at 24 h after the last infusion. However, prior to the final infusion, remission (after infusion 4) and response (after infusion 5) favored the six-ketamine group compared with the midazolam group before the latter received its single ketamine infusion. Durability: Among responders, median time-to-relapse was 2 weeks in the single-ketamine (after midazolam) arm and 6 weeks in the six-ketamine arm; this difference was nominal and not statistically significant. Safety/tolerability: Repeated infusions were relatively well-tolerated. Common adverse events during the infusion phase included, in the ketamine group: general malaise (28%), decreased energy (28%), increased blood pressure (24%), headache (24%), fatigue (24%), nausea/vomiting (20%), anxiety (20%), and poor concentration (20%); and in the midazolam + single ketamine group: anxiety (24.1%), decreased energy (17.2%), increased blood pressure (17.2%), fatigue (13.8%), and headache (13.8%).

The trial did not demonstrate superiority of six repeated ketamine infusions over a single ketamine infusion (administered after five midazolam infusions) at the primary endpoint of 24 hours after the final infusion, likely because both groups had received ketamine immediately prior to that assessment. Nevertheless, longitudinal analyses showed that repeated ketamine produced greater antidepressant effects than midazolam across the first five infusions, with higher remission and response rates at interim time points before the midazolam arm received ketamine. Durability among responders tended to be longer after repeated ketamine (median 6 weeks) compared with a single infusion (median 2 weeks), but this difference was not statistically significant. The safety profile of repeated ketamine was acceptable, with expected transient increases in blood pressure and dissociative/sedative-type adverse events that resolved with monitoring. Overall, the findings suggest that while repeated ketamine confers advantages over active placebo during the multi-infusion phase, the advantage diminishes when the comparison group also receives ketamine, underscoring the need to optimize dosing schedules and maintenance strategies to achieve sustained benefit with minimal exposure.

Six repeated ketamine infusions conferred stronger antidepressant effects than active placebo during the infusion series but were not superior at 24 hours after the last infusion when the comparison group also received a single ketamine dose. Durability among responders nominally favored repeated ketamine without statistical significance. Repeated ketamine was generally well-tolerated. Future research should define optimal dosing frequency, total number of infusions, maintenance/booster strategies, and patient selection to maximize sustained antidepressant efficacy while minimizing exposure and adverse effects.

- Primary endpoint timing occurred after both groups had received ketamine, reducing the ability to detect differences between six versus single infusion conditions at that time point. - Single-center VA sample with a predominance of male participants may limit generalizability. - Modest sample size (N=54 completing primary endpoint) may have limited power to detect between-group differences in remission/response and durability. - Active placebo design with subsequent ketamine exposure in the control arm complicates interpretation of longer-term comparative efficacy. - Cognitive and neuroimaging outcomes were collected only in a subsample and are reported separately.