Ketamine versus ECT for Nonpsychotic Treatment-Resistant Major Depression

Major depressive disorder is a leading cause of disability and a substantial proportion of patients do not respond adequately to antidepressant medications; treatment-resistant major depression is commonly defined as failure to respond to two or more adequate antidepressant trials. Electroconvulsive therapy (ECT) has a long-established record of effectiveness for treatment-resistant depression but is underused because of limited availability, stigma and concerns about cognitive adverse effects. Intravenous subanesthetic ketamine (commonly 0.5 mg/kg) has emerged over the past two decades as a rapidly acting antidepressant option that does not require general anaesthesia, but its comparative effectiveness versus ECT and its cognitive safety profile remain uncertain, especially in patients without psychotic features and in outpatient settings. Anand and colleagues therefore designed a pragmatic, multisite randomized noninferiority trial (the ELEKT-D trial) to determine whether twice-weekly intravenous ketamine over 3 weeks was noninferior to thrice-weekly ECT over 3 weeks for adults with nonpsychotic treatment-resistant major depression. The principal aim was to compare response rates on a patient-rated depression scale at the end of the initial treatment phase, with secondary assessments of clinician-rated depression scales, remission rates, cognitive outcomes and patient-reported quality of life, and a 6-month follow-up of responders receiving clinically prescribed maintenance treatments.

This was a prospective, open-label, randomised, noninferiority trial conducted at five US sites (an urban community hospital, a Veterans Administration hospital, and three university hospital centres). Eligible participants were outpatients or inpatients aged 21–75 years who met DSM-5 criteria for major depressive disorder without psychotic features, had an unsatisfactory response to at least two adequate antidepressant trials, and had a baseline MADRS score greater than 20. Verification of treatment resistance used referring-provider information, standard guidelines and clinical interview; full inclusion/exclusion criteria were reported in the trial protocol. After screening, participants were randomised 1:1, stratified by site, to ketamine or ECT. Masking was not performed because of the treatments' differing natures: patients, treating clinicians and research staff were aware of assignments. The ketamine regimen was intravenous ketamine 0.5 mg/kg infused over 40 minutes twice weekly for 3 weeks, with allowance for dose modification if clinically indicated. The ECT regimen was administered three times per week, with an initial recommended right unilateral ultrabrief pulse at six times the seizure threshold (titrated at the first visit) and allowance to change electrode placement or settings per clinical judgment; treatment could end early or be discontinued by patients. The primary outcome was response on the Quick Inventory of Depressive Symptomatology–Self-Report (QIDS-SR-16), defined as a ≥50% reduction from baseline to the end-of-treatment visit (within 3 days after the last session of the initial 3-week phase). The trial used a noninferiority margin of -10 percentage points (ketamine’s response proportion could be no more than 10 percentage points lower than ECT’s). Key secondary outcomes included response and remission on the clinician-rated MADRS, remission thresholds on both scales, Clinical Global Impression and Patient Global Impression scores, several cognitive measures (GSE-My, SMCQ, Hopkins Verbal Learning Test–Revised [HVLT-R]) and quality-of-life scores; dissociation was measured with the Clinician-Administered Dissociative States Scale (CADSS). The study was powered to enrol about 400 patients (planned 200 per group) to provide 80% power under assumed response rates and attrition. The principal analysis was performed in a modified intention-to-treat population, defined as patients who received at least one treatment session and had at least one QIDS-SR-16 measurement during the initial phase. Noninferiority was tested with the Farrington–Manning score statistic (one-sided alpha 0.025); sensitivity analyses included intention-to-treat analyses with multiple imputation for missing primary outcomes. Longitudinal QIDS-SR-16 trajectories were modelled using linear mixed-effects models including treatment, assessment, treatment-by-assessment interaction and site as fixed effects and random intercepts for patients. Adverse events were summarised by group proportions. The trial was funded by PCORI and overseen by an executive committee and data and safety monitoring board.

Between March 2017 and September 2022, 403 patients were randomised (200 to ketamine, 203 to ECT). Thirty-eight participants withdrew before starting assigned treatment; the modified intention-to-treat population comprised 195 patients who received ketamine and 170 who received ECT. The overall sample had a mean age of 46 years, 51.1% were women and most (89.1%) were outpatients; baseline depression severity was moderate to severe. Primary outcome: A QIDS-SR-16 response (≥50% reduction) occurred in 108 of 195 patients (55.4%) in the ketamine group and in 70 of 170 patients (41.2%) in the ECT group, a between-group difference of 14.2 percentage points (95% CI, 3.9 to 24.2). The Farrington–Manning statistic indicated P<0.001 for the noninferiority of ketamine to ECT. Sensitivity analyses in the intention-to-treat population using multiple imputation produced consistent results (55.4% vs 41.6%, difference 13.8 percentage points; 95% CI, 3.9 to 23.8), and other imputation scenarios similarly supported noninferiority. Depression scales and remission: On the clinician-rated MADRS, response rates were 50.8% (99/195) with ketamine and 41.4% (70/169) with ECT (difference 9.3 percentage points; 95% CI, -0.9 to 19.4). Least-squares mean change in QIDS-SR-16 from baseline to end-of-treatment was -9.0±0.4 for ketamine versus -7.2±0.4 for ECT (difference -1.8 points; 95% CI, -2.8 to -0.8). Corresponding MADRS changes were -15.3±0.7 and -13.1±0.7 (difference -2.2 points; 95% CI, -4.1 to -0.3). Remission rates by QIDS-SR-16 were 32.3% for ketamine and 20.0% for ECT; by MADRS they were 37.9% and 21.8%, respectively. Cognitive outcomes and quality of life: Patient-reported memory (GSE-My) at end-of-treatment was lower (worse) in the ECT group than in the ketamine group (3.2±0.1 vs 4.2±0.1; difference 1.1 points; 95% CI, 0.9 to 1.2). The SMCQ also favoured ketamine with a mean between-group difference of 9.0 points (95% CI, 5.1 to 13.0). On the HVLT-R delayed-recall T-score, decline from baseline was greater after ECT than after ketamine (-9.7±1.2 vs -0.9±1.1; difference 8.8 points; 95% CI, 5.7 to 11.9). By the 1-month follow-up these HVLT-R differences had largely returned to similar levels between groups. Quality-of-life improvements from baseline to end-of-treatment were similar (mean increase 12.3 points with ketamine vs 12.9 points with ECT; difference -0.6 points; 95% CI, -3.4 to 2.1). Treatment exposure and changes: Completion rates during the initial phase were high: 92.3% (180/195) of ketamine patients completed all six planned infusions, and 92.9% (158/170) of ECT patients received six to nine sessions. ECT was converted from right unilateral to bilateral placement in 39% of patients; ketamine dosing remained unchanged for nearly all participants. Adverse events and safety: No deaths occurred. During the initial treatment phase, at least one moderate or severe adverse event occurred in 49 of 195 patients (25.1%) in the ketamine group and in 55 of 170 patients (32.4%) in the ECT group. Dissociative symptoms as measured by CADSS were consistently higher in the ketamine group across treatment days. Musculoskeletal adverse events were more frequent with ECT (5.3% vs 0.5%). New suicidal ideation was reported by six patients during the initial phase (4 ketamine, 2 ECT), and one suicide attempt occurred during follow-up in a patient in the ketamine group. Follow-up: Responders entered a 6-month follow-up during which clinicians prescribed maintenance treatments as clinically indicated. Depression scores tended to increase during follow-up. Relapse (QIDS-SR-16 >11) was observed more frequently after initial ECT response than after ketamine: at month 1 relapse occurred in 19.0% of ketamine responders vs 35.4% of ECT responders; at month 3, 25.0% vs 50.9%; and at month 6, 34.5% vs 56.3%. Cognitive deficits noted after ECT largely recovered by the 1-month visit and remained similar thereafter. Quality-of-life gains were maintained over follow-up.

Anand and colleagues interpret their findings as showing that twice-weekly intravenous subanesthetic ketamine administered over 3 weeks was noninferior to thrice-weekly ECT for adults with nonpsychotic treatment-resistant major depression, using a QIDS-SR-16 response (≥50% reduction) as the primary outcome. They emphasise that ketamine produced at least comparable short-term antidepressant effects and was associated with less short-term decline in measured memory performance than ECT; the latter deficits after ECT largely resolved by the 1-month follow-up. Both treatments yielded similar improvements in patient-reported quality of life after the initial treatment phase. The authors place their results in the context of earlier trials and meta-analyses that suggested inferior remission with ketamine versus ECT, noting important differences: this trial excluded psychotic depression, had a larger sample size and was conducted mainly in outpatients (89%), whereas prior trials included mixed populations and more inpatient settings. They propose that lower-than-expected response and remission rates with ECT in this trial may reflect the outpatient, nonpsychotic population and the initial use of unilateral ultrabrief ECT, and they note that ECT tends to be more effective in older adults, psychotic depression and inpatient settings. Limitations acknowledged by the investigators include the open-label design (which could influence outcomes), starting ECT with right unilateral placement and allowing later conversion to bilateral (which may have reduced ECT efficacy relative to starting with bilateral ECT), the possibility that more ECT sessions would increase response but also adverse events, the absence of placebo control, flexibility in treatment methods, and a follow-up period during which maintenance therapies were not standardized but were determined by treating clinicians. They also note higher withdrawal before treatment initiation in the ECT group, reflecting patient preference and logistical factors. The trial’s pragmatic design is presented as a strength because it improves generalisability to routine clinical practice. In their concluding statement the authors report that subanesthetic intravenous ketamine was noninferior to ECT for treatment-resistant major depression without psychosis, while reiterating the caveats and the need for future studies to assess comparative effectiveness in older patients, bipolar depression and inpatient or emergency settings.