Efficacy and Safety of Ketamine vs Electroconvulsive Therapy Among Patients With Major Depressive Episode: A Systematic Review and Meta-analysis

Major depressive episodes (MDEs) are common, disabling, and frequently resistant to standard antidepressant treatments: more than 30% of patients do not achieve remission after several medication trials. Electroconvulsive therapy (ECT) is endorsed by professional guidelines as a highly effective intervention for treatment-resistant depression, but its use is limited by logistical barriers, stigma, and concerns about cognitive adverse effects despite procedural refinements. Since 2000, multiple small-to-medium clinical trials have reported rapid antidepressant effects of low-dose intravenous racemic ketamine, a glutamatergic N-methyl-D-aspartate (NMDA) receptor antagonist, raising the possibility that ketamine could be an alternative to ECT for some patients. Taeho and colleagues set out to synthesise and quantify head-to-head clinical trial evidence comparing ketamine and ECT for patients with MDE. The study aimed to estimate pooled effect sizes for depressive symptom improvement (primary outcome), and to compare cognition/memory outcomes and safety profiles (secondary outcomes), using a systematic review and meta-analytic approach to inform comparative efficacy and tolerability in the acute treatment phase.

The investigators registered their protocol on PROSPERO (CRD42022338045) and conducted a systematic search of PubMed, MEDLINE, the Cochrane Library, and Embase from database inception to April 19, 2022, supplemented by manual searches of US and European clinical trial registries and Google Scholar. No language restrictions were applied. The review followed PRISMA reporting guidance. Pre-specified inclusion criteria required human clinical trials enrolling patients with depression diagnosed by standard criteria (eg, DSM or ICD), direct comparisons between ECT and ketamine, and use of standardised measures of depressive symptom severity (eg, MADRS, HDRS). Secondary efficacy outcomes included suicidal ideation and cognition/memory measures; safety outcomes included serious adverse events and specified adverse event types. Excluded studies were non-human or lacked standardised outcome measures. Two reviewers independently screened titles/abstracts and full texts, extracted data using a pilot-tested form, and resolved discrepancies by discussion. Extracted items included study and participant characteristics, intervention details, and quantitative outcome data. The investigators checked carefully for overlapping samples. Risk of bias was assessed with the Cochrane Risk of Bias tool version 2 across five domains (randomisation process; deviations from intended interventions; missing outcome data; measurement of outcomes; selection of reported results). Publication bias was evaluated with funnel plots and the Egger and Begg tests. For continuous depression severity outcomes the authors calculated Hedges' g standardised mean differences (SMDs) to combine different rating scales; relative risks (RRs) were used for binary safety outcomes. Study weights used inverse-variance methods. Heterogeneity was quantified with Cochran's Q and I2; a random-effects model was used when I2 exceeded 50%, and a fixed-effects model otherwise (with both reported when I2 equalled 50%). Because some studies reported multiple severity measures, a two-stage approach was used: a three-level meta-analysis to combine multiple measures within studies in stage one, and pooling one effect size per study in stage two to avoid sample duplication. Meta-regression and subgroup analyses examined moderators including sample size, age, sex percentage, geographic region, randomisation status, presence of psychotic features, and ECT treatment type. Analyses were performed in R (version 4.2.1) using the meta, rmeta, and metafor packages, with a 2-sided P < .05 considered statistically significant.

The search identified 1,248 articles; after screening and full-text review, 6 clinical trials met inclusion criteria, yielding 340 patients (162 assigned to ECT and 178 to ketamine). Five trials were single-site and one was multicentre. Five studies were randomised clinical trials and one was an open-label naturalistic trial. Sample sizes ranged from 18 to 186 and mean ages ranged from about 37.5 to 52.5 years. All studies recruited patients who were ECT candidates; five enrolled unipolar MDE only, and one included unipolar or bipolar depression. Treatment series for both modalities were completed within about a month in the included trials. Primary efficacy: Across 11 effect sizes from the six studies, pooled results favoured ECT over ketamine for acute reduction in depressive symptoms. The overall pooled SMD was -0.69 (95% CI, -0.89 to -0.48), with Cochran Q P = .15 and I2 = 39%, indicating moderate heterogeneity. When stratified by instrument, ECT was superior on MADRS (SMD -0.59; 95% CI, -0.85 to -0.33), HDRS (SMD -0.83; 95% CI, -1.22 to -0.44), and Beck Depression Inventory (SMD -0.86; 95% CI, -1.50 to -0.22). Suicidal ideation: Only one trial formally assessed suicidal ideation (Beck Scale for Suicidal Ideation). Baseline scores were similar across groups and each treatment group showed within-group reductions, but there were no significant between-group differences at 1-week and 1-month follow-up (P = .99 and P = .69 respectively). Cognition and memory: Two studies reported cognitive outcomes. One study found better neurocognitive performance in the ketamine group versus ECT (Cohen's d = 0.40; P = .04), with advantages in attention, verbal memory, and executive function but no differences in immediate or visual memory. The other study, using the Wechsler Memory Scale, reported no between-group differences in memory performance. Safety and adverse events: Four of six studies explicitly reported adverse events. Only one trial reported serious adverse events (including suicide attempts and deaths): numbers were not statistically different between groups (23 of 90 in ECT vs 14 of 91 in ketamine; P = .09), and counts of suicide attempts (6 vs 4) and suicide deaths (1 vs 0) did not differ. For specific adverse events pooled across studies, ketamine had lower risks than ECT for headache (RR 0.37; 95% CI, 0.18-0.76) and muscle pain (RR 0.23; 95% CI, 0.13-0.38). Transient dissociative or depersonalisation symptoms were reported more frequently with ketamine (reported RR approximately 5 in the extracted text), although the confidence interval for that estimate was not clearly reported in the extraction. The reviewers distinguished intramuscular and oral ketamine in one study to avoid duplication. Moderator and bias assessments: Meta-regression and subgroup analyses did not identify significant moderators among the tested variables (sample size, age, male sex percentage, region, randomisation status, ECT type, presence of psychotic features). Methodological quality across studies was rated low to moderate: randomisation was used in all but one study, allocation concealment was adequate in four, and four studies had concerns about deviations from intended interventions, largely due to blinding difficulties. One study reported differential dropout favouring ECT. Tests for publication bias (Egger P = .32; Begg and Mazumdar P = .57) did not suggest small-study effects.

Taeho and colleagues interpret the pooled evidence as indicating that ECT is more efficacious than ketamine for acute reduction of depression severity in patients with MDE, based on a pooled SMD of -0.69 favouring ECT. Nevertheless, the authors stress that treatment choice should remain individualised because ketamine may provide more rapid early antidepressant effects in some patients; three included studies qualitatively reported faster initial improvement with ketamine, whereas one reported faster recovery with ECT. Longer-term outcomes were sparsely reported. Two trials included follow-up of three months or longer: one found no difference in depression severity at three months, and another found similar remission rates at 12 months, suggesting that acute advantages may diminish over time. The investigators note that relapse prevention strategies differ between modalities and that well-powered research on sustained maintenance for ketamine is lacking; they cite evidence from esketamine withdrawal trials as an example of a maintenance approach that could inform future ketamine strategies. Regarding cognition and memory, results were inconsistent and underpowered: one study showed a small-to-moderate cognitive advantage favouring ketamine, while another found no difference. The authors conclude that current data are insufficient for firm conclusions about cognitive outcomes and call for larger studies. Safety reporting was limited and mostly acute. The available data indicate distinct adverse-event profiles: ketamine had lower rates of headache and muscle pain while ECT had lower rates of blurred vision, vertigo, ocular motor disturbances, and transient dissociative symptoms. Only one trial reported serious adverse events in detail, and the meta-analysis could not adequately assess long-term adverse effects or patient tolerability/acceptability of differing side-effect profiles. The authors therefore recommend future research to evaluate long-term safety, tolerability, and patient-centred acceptability. Key limitations acknowledged by the study team include small sample sizes, short follow-up durations, heterogeneity in inclusion criteria (eg, inclusion of bipolar or psychotic depression in some trials), and variability in ketamine and ECT protocols (dosing, route, frequency, ECT stimulus parameters). They also note the absence of direct comparisons with esketamine. Two larger ongoing randomised trials (CAN-BIND and ELEKT-D) were identified as likely to provide more definitive evidence when completed. The authors conclude that while ECT may be superior for acute symptom reduction, methodological limitations temper certainty and further large, well-controlled studies are needed to inform clinical decision-making.

This meta-analysis of six clinical trials with 340 patients suggests that ECT may be superior to ketamine for improving acute depression severity. However, the authors caution that methodological limitations — including small and heterogeneous studies, limited follow-up, and incomplete adverse-event reporting — constrain definitive conclusions. Larger ongoing comparative trials are expected to provide more robust evidence to guide treatment choices for patients with major depressive episodes.