Is ketamine an appropriate alternative to ECT for patients with treatment resistant depression? A systematic review

Major depressive disorder is common and a substantial contributor to global disability. A sizeable minority of patients—more than 30% according to the introduction—do not achieve remission after multiple antidepressant trials and are classified as having treatment resistant depression (TRD). Electroconvulsive therapy (ECT) is regarded in guidelines as the gold-standard intervention for TRD because of its high efficacy, but it has disadvantages: repeated general anaesthesia, variable availability, public stigma, and a risk of cognitive adverse effects that many patients find distressing. Relapse after an initial ECT course is also common, so alternatives that match ECT’s efficacy but have more acceptable side-effect profiles and improved relapse prevention are clinically desirable. Veraart and colleagues frame ketamine, an N-methyl-D-aspartate (NMDA) receptor antagonist, as a candidate alternative because of consistent evidence for rapid antidepressant effects in major depression, including in severe TRD, and generally transient acute side effects. The review aims to identify and summarise studies that directly compare ketamine treatment with ECT in patients with depression, to address whether ketamine could be an effective and acceptable alternative to ECT and inform where ketamine might sit within stepped treatment algorithms.

The review followed PRISMA guidance. Searches of PubMed/MEDLINE and the Cochrane Trials Library were carried out on 21 September 2020 using the terms ketamine AND (ECT OR electroconvulsive therapy), with no other restrictions. Two investigators (JV and SS) screened titles independently and searched bibliographies of relevant studies; clinicaltrials.gov was also checked for ongoing trials. Eligibility followed a PICOS approach. Participants were patients with major depressive disorder. Interventions included ketamine in any formulation, dose, frequency or route versus ECT. Primary outcomes were standardised ratings of depressive symptom severity at treatment end-points (for example HDRS or MADRS); secondary outcomes included symptom trajectories at other time points, response and remission rates, cognitive assessments, tolerability and side effects. Study designs eligible were randomized controlled trials, controlled clinical trials and open-label trials. Data extraction was performed independently by two investigators, who collected study design, population, intervention details and outcome measures. Methodological quality was assessed using the RoB 2 tool for randomized trials and the ROBINS-I tool for non-randomized studies. Corresponding authors were contacted for additional information where relevant.

The search identified 137 manuscripts; after screening and full-text assessment, six studies met inclusion criteria. Three were single-blind randomized trials with blind raters; three were naturalistic open-label studies. The overall methodological quality varied and several studies carried appreciable risk of bias. One single-blind randomized study randomised 18 patients to three IV ketamine infusions (0.5 mg/kg) or three brief-pulse bilateral ECT sessions (n = 9 per arm). Ketamine produced significantly lower Beck Depression Inventory and HDRS-25 scores than ECT after the first and second treatments and at 72 hours post-treatment, but there were no significant differences after the third treatment or at 1 week. Both interventions were well tolerated. Limitations included small sample size, short follow-up and the fact that only three ECT sessions were administered, whereas typical ECT courses involve more sessions. A second single-blind randomized trial enrolled 32 patients (16 assigned to 0.5 mg/kg IV ketamine, 16 to bilateral ECT twice weekly to remission); only 10 ketamine and 12 ECT patients completed the study and analyses were restricted to completers. After five sessions HDRS scores decreased by a mean of 8.2 points in the ketamine group and 11.4 points in the ECT group. Numerically, ECT had greater benefit after six sessions (mean HDRS decrease 7.8 points ECT vs 4.5 points ketamine). HDRS scores in the ketamine group tended to return toward baseline within three months whereas scores in the ECT group remained comparable to early post-treatment values. Wechsler Memory Scale scores increased in the ketamine group and declined in the ECT group, though differences were not statistically significant. Reported acute ketamine adverse events included dizziness, headache, blurred vision, numbness and depersonalisation; ECT-related complaints included muscle and joint pain and headache. This trial was limited by high dropout, small groups, unclear blinding (patients could likely infer treatment), and unclear reporting of total ECT sessions. The third randomized trial compared right unilateral or bifrontal ECT with six alternate-day 0.5 mg/kg IV ketamine infusions in 26 patients with unipolar or bipolar depression. Baseline HDRS scores were similar. Patients receiving ECT showed a greater HDRS reduction (group × time interaction F = 4.79; p < 0.001) and achieved response and remission faster (log-rank statistics both p < 0.01). Cognitive performance measured with a battery for ECT-related deficits improved after ketamine compared to baseline (p = 0.017) but not after ECT. The study had a small sample, three ketamine dropouts, and heterogeneity in ECT electrode placement that may have affected outcomes. Among the three open-label studies, one enrolled 35 unipolar TRD patients (17 ECT, up to 12 sessions, median = 8; 18 ketamine, up to three 0.5 mg/kg infusions weekly). Ketamine produced significant HDRS reductions at 2 hours and 24 hours after each infusion and at 1 week after the last infusion (all p < 0.001). At 2 hours after the first infusion 76.5% of ketamine patients met response criteria, and 60% were responders at 1 week post-treatment; 50% of the ECT group showed response after the final ECT. Seven ketamine patients discontinued for reasons including elevated blood pressure, unpleasant experience or lack of efficacy; no ECT discontinuations occurred. The study was not prospectively designed to compare the treatments and was vulnerable to selection bias. A naturalistic matched study treated 31 hospitalised TRD patients with a 2-week ketamine infusion series and compared them to 31 matched ECT controls; after exclusions the analysed sample was 50. Baseline MADRS scores were lower in the ketamine group (M = 26.40, SD = 4.94) than the ECT group (M = 31.17, SD = 7.28; p = 0.010), and episode duration was longer in the ketamine group. At mid-treatment (three ketamine infusions or six ECT sessions) percent MADRS reduction was greater in the ketamine arm (mean -47.45%, SD = 23.43) than ECT (mean -34.86%, SD = 21.29; p < 0.05). At end of treatment reductions were similar (ECT mean -55.70%, SD = 23.63; ketamine mean -49.88%, SD = 27.30). Ketamine patients showed significantly better performance on attention, verbal memory and executive function; the composite cognitive effect modestly favoured ketamine (d = 0.40). Limitations included the non-randomised design, baseline differences and lack of follow-up. A further open-label study of 44 patients (17 ECT, 27 ketamine) reported significant HDRS reductions in both groups (ECT from 21.41 to 15.35, T = 3.07, p < 0.01; ketamine from 20.15 to 8.93, T = 10.73, p < 0.01). Ketamine was given as four infusions 2–3 days apart; ECT continued to maximal response with an average of 14 sessions. No safety data were reported and the study focused on neural correlates rather than a direct efficacy comparison. Across studies, acute ketamine adverse events commonly reported were dizziness, headache, blurred vision, numbness, depersonalisation, vertigo, diplopia and nausea; dissociation and transient blood pressure elevations led to some dropouts. Cognitive outcomes generally favoured ketamine in the short term. The authors also identified three ongoing trials via clinicaltrials.gov.

Veraart and colleagues conclude that current direct-comparison evidence is limited and heterogeneous, so no definitive statement can be made about ketamine as an alternative to ECT for TRD. The six included studies do not yield a consistent picture: one of three randomized trials found superior outcomes with ECT, while two found no significant differences but carried notable risk of bias. Several reports suggest a more rapid antidepressant effect with ketamine, which could be clinically important, but the best-assessed randomized trial found faster recovery with ECT. The only study with longer-term follow-up indicated less durable antidepressant effects after ketamine versus ECT, although concurrent or additional ECT around follow-up time points was not always clear. The investigators emphasise that ketamine appears to have a more favourable short-term neurocognitive side-effect profile than ECT, but they caution that chronic or high-dose ketamine exposure carries risks including memory impairment, urinary tract toxicity and potential for misuse. Acute ketamine adverse effects and dissociation, and the need for repeated or maintenance administrations to sustain benefit, are highlighted as pragmatic concerns that could influence acceptability and long-term safety. By contrast, ECT’s side effects include cognitive impairment and anaesthesia-related risks. The balance of benefits and harms therefore remains an open question. Methodological limitations across the literature are emphasised: small sample sizes, variable and sometimes low ketamine dosing frequencies (for example once-weekly regimens that may underpower ketamine’s efficacy), heterogeneity in ECT parameters (frequency, electrode placement, pulse width, dose relative to seizure threshold and anaesthetic agents), inconsistent reporting of prior treatment history and levels of treatment resistance, inclusion/exclusion differences (for instance bipolar or psychotic depression), and short follow-up intervals. Blinding is inherently challenging in comparisons of ketamine and ECT, and allocation bias is likely in non-randomised studies. These issues reduce confidence in available findings. Given these uncertainties, the authors call for well-powered randomized trials in severely treatment-resistant populations typical of those referred for ECT, with careful characterisation of prior treatment history, standardised intervention protocols, adequate follow-up to assess durability and relapse prevention strategies, systematic assessment of tolerability and patient experience, and attention to both acute and long-term adverse events. They note three ongoing trials (including CAN-BIND NCT03674671, a Swedish trial NCT02659085, and the ELEKT-D trial NCT03113968) that aim to address many of these gaps. In summary, current evidence does not convincingly justify considering ketamine as an equally effective replacement for ECT in TRD, although the potentially more favourable short-term cognitive profile of ketamine warrants further investigation.