Ketamine for bipolar depression: a systematic review

Bipolar depression is a major contributor to global disability and treatment-resistant bipolar depression (TRBD) is common yet undertreated, with few well-established therapeutic options beyond electroconvulsive therapy and repetitive transcranial magnetic stimulation. Earlier research has shown rapid antidepressant effects of single sub‑anaesthetic intravenous racemic ketamine in unipolar major depression, including short‑term reductions in suicidal ideation, but the evidence base for ketamine in bipolar depression has been smaller and less clear. Attempts to prolong ketamine's acute effects with other glutamatergic agents have produced inconsistent results, and questions remain about optimal dosing, formulation, and long‑term outcomes in bipolar disorder. Bahji and colleagues set out to update prior reviews by systematically synthesising experimental studies that examined the efficacy and tolerability of ketamine for bipolar depression. The authors focused on controlled and nonrandomised experimental studies in adults and aimed to summarise outcomes related to depressive symptom change, response rates, and adverse events, while assessing study quality and gaps that warrant further research.

This work was conducted as a systematic review registered on the Open Science Framework and followed PRISMA reporting guidance. The investigators searched MEDLINE, Embase, PsycINFO, CENTRAL and the Cochrane Database via Ovid from database inception to 13 December 2019, and also searched clinical trial registries (ClinicalTrials.gov, EU Clinical Trials Register, Australian and New Zealand Clinical Trials Registry) and reference lists to identify published, ongoing or unpublished trials. Eligibility criteria included randomised controlled trials and nonrandomised experimental studies of ketamine in adults (≥18 years) with bipolar depression; observational designs, reviews, secondary analyses, commentaries and studies that did not separate bipolar from unipolar depression were excluded. Any ketamine formulation and combination with psychotropic medications or psychotherapies were eligible in principle. Two authors independently screened records using Covidence, retrieved full texts, and resolved disagreements by consensus. Data extraction used a piloted Excel form completed by two co‑authors, prioritising conservative analyses (preference for intention‑to‑treat where available); the reviewers collected information on population characteristics, intervention details (dose, route, number of doses), comparators, outcomes, withdrawals and adverse events, and contacted authors when data were missing. Risk of bias for randomised trials was assessed with the Cochrane Risk of Bias tool examining selection, performance, detection, attrition and reporting biases, with assessments conducted independently by two authors. Although a meta‑analysis had been planned, only six eligible studies were identified (three randomised controlled trials and three open‑label single‑arm studies), so the authors presented a narrative synthesis and tabulated study characteristics and outcomes. The search initially yielded 2,494 records; after deduplication and screening, 361 full texts were assessed and six studies met inclusion criteria.

Six experimental studies met inclusion criteria, comprising 135 participants (53% female; mean age 44.7 years, SD 11.7). Three studies were double‑blind randomised controlled trials and three were open‑label, single‑arm studies. By country, half of the trials were conducted in the United States (k = 3) and two were from Poland. Five studies used DSM‑IV or DSM‑5 criteria for bipolar disorder; one study did not report diagnostic criteria. All included trials administered adjunctive intravenous racemic ketamine at 0.5 mg/kg while participants continued a primary mood stabiliser. Dosing regimens varied: three studies delivered a single dose, two studies used a two‑test‑dose design (one ketamine and one placebo two weeks apart), and one study administered six doses across two weeks. Most trials recruited participants with TRBD; common exclusions were serious medical comorbidity, pregnancy or breastfeeding, psychosis and substance use disorder. Across the six studies, 61% of participants who received ketamine at some point achieved a response, defined as ≥50% reduction in baseline depression severity (77/126 as reported). Study‑level response rates ranged from 52% to 80%. In the three controlled trials, the pooled response among control participants was 5% (2/42). Depression rating scores improved over time in all studies, although in a trial that used midazolam as an active control in a non‑TRBD sample the between‑group difference was not statistically significant. Single‑dose ketamine effects generally did not extend beyond two weeks, whereas the study that delivered six doses over two weeks suggested more durable benefit. Tolerability was generally acceptable across studies. Two participants developed manic symptoms (one participant receiving ketamine and one receiving placebo). Significant dissociative symptoms were reported in two trials, typically peaking around 40 minutes after infusion; the other four trials did not report substantial dissociation or mania. Regarding study quality, the three randomised, double‑blind trials with concealed allocation were judged at very low risk of bias, while the three open‑label single‑arm studies lacked control groups and were more susceptible to participation and other biases.

Bahji and colleagues interpret the assembled evidence as preliminary but supportive of a role for adjunctive intravenous racemic ketamine (0.5 mg/kg) in producing rapid antidepressant effects in adults with bipolar depression, including TRBD. The pattern of rapid onset observed in these bipolar samples mirrors findings from unipolar depression studies, but the evidence for bipolar disorder is based only on racemic intravenous ketamine and derives from a small number of trials with short follow‑up. The authors highlight that there are no published studies of intranasal esketamine specifically in bipolar depression, creating uncertainty about whether findings for racemic ketamine generalise to other formulations or routes. They recommend head‑to‑head comparisons between racemic ketamine and esketamine and suggest measuring blood levels of ketamine and norketamine to explore pharmacokinetic differences. Concerning safety, dissociation and affective switching to mania or hypomania are identified as the principal concerns; however, most trials did not observe these adverse events and the limited data do not establish a clear risk of manic switching after single sub‑anaesthetic doses. The authors note that trials to date are typically short and small, which limits the ability to detect less common adverse outcomes such as treatment‑emergent mania. Key limitations acknowledged by the reviewers include the inability to perform a quantitative meta‑analysis due to the small number of trials, potential publication bias, very limited data beyond two weeks of follow‑up, restrictive trial eligibility that limits generalisability to real‑world bipolar populations, and heterogeneity across studies in participant treatment resistance status, treatment setting, and dosing schedules. The authors therefore call for larger, longer randomised controlled trials that examine different ketamine formulations, compare routes of administration, assess longer‑term outcomes and monitor safety signals comprehensively.

Ketamine is described as an innovative, rapidly acting experimental treatment for bipolar depression with preliminary evidence supporting short‑term efficacy for intravenous racemic ketamine as an adjunct to mood stabilisers. No studies to date have tested other formulations such as intranasal esketamine in bipolar depression; the intravenous route also presents practical limitations. Long‑term benefits remain insufficiently explored and the overall level of proof is judged to be low. The authors conclude that further randomised controlled trials are needed to clarify efficacy, compare formulations and routes, and establish safety and longer‑term outcomes in bipolar disorder.