Antidepressant and neurocognitive effects of serial ketamine administration versus ECT in depressed patients

Depressive disorder produces a high disease burden worldwide and many patients do not achieve an adequate or sufficiently rapid response with conventional antidepressants. In addition to mood symptoms, cognitive deficits—particularly in attention and executive function—are common, often persist after mood remission, and are linked to poorer functional recovery. Electroconvulsive therapy (ECT) is the established acute treatment for treatment-resistant depression and is relatively rapid in onset, but it is associated with transient cognitive side effects, particularly in episodic memory. Ketamine, an NMDA receptor antagonist, has emerged as a fast-acting antidepressant, yet its short-term effects on cognition in a clinical treatment context remain incompletely characterised. Earlier clinical studies are limited by small samples, single-infusion designs, and few direct comparisons with ECT, and real-world data are sparse. Basso and colleagues set out to compare clinical efficacy and acute neurocognitive effects of serial ketamine infusions versus ECT in a naturalistic sample of depressed inpatients. A secondary, exploratory aim was to examine whether baseline clinical or cognitive characteristics predict treatment outcome for either intervention. The study therefore addresses both speed and magnitude of antidepressant effect and parallel changes in multiple cognitive domains under conditions closer to routine clinical practice than tightly controlled randomised trials.

This was a naturalistic, non-randomised comparative study of depressed inpatients at the Department of Psychiatry, Charité — Universitätsmedizin Berlin. From April 2014 onwards 56 patients received ketamine in the larger programme; 31 met the study inclusion criteria and were matched by age and gender to 31 patients who had received standard ECT between August 2013 and June 2016. After excluding pairs for very few ketamine infusions (≤3), unusually high numbers of ECT (>16), or missing neurocognitive data, the final analytic sample comprised 50 patients: 25 treated with ketamine and 25 treated with ECT. All participants provided informed consent. The extracted text does not report randomisation or a concurrent control group. Inclusion criteria for ketamine (as reported) excluded lifetime ketamine exposure or recreational ketamine use, recent cardiovascular disease, significant untreated medical conditions (e.g. anaemia, thyroid disorder), increased intracranial pressure or glaucoma, pregnancy, and major psychiatric or neurological comorbidity such as dementia, epilepsy, schizophrenia, psychosis, or PTSD. All patients had a primary diagnosis of major depressive disorder and at least two prior adequate antidepressant trials in the current episode. ECT was administered three times weekly, typically over four weeks, using right unilateral ultra-brief pulse stimulation (0.3 ms) on a MECTA 5000Q device under general anaesthesia (propofol or etomidate) with succinylcholine for muscle relaxation; seizure threshold was titrated at first treatment. Mean number of ECT sessions was 12.36 (SD = 1.75, range 9–16). Ketamine was delivered intravenously three times weekly on the same schedule over two weeks, as R-ketamine 0.5 mg/kg infused over 40 minutes, with monitoring until acute psychotomimetic effects subsided; mean infusions were 6.76 (SD = 1.23, range 6–9). Clinical assessments used the German Montgomery-Åsberg Depression Rating Scale (MADRS). Symptom ratings were performed at baseline (T0), mid-treatment (T1: after six ECT sessions or three ketamine infusions) and at treatment end (T2: after about 12 ECT sessions or six ketamine infusions). Response was defined as ≥50% reduction in MADRS and remission as MADRS ≤10. Neurocognitive testing evaluated attention, immediate verbal short-term memory, verbal memory, visual memory and executive functions; tests were administered at baseline and at treatment end for both groups, and additionally at mid-treatment for the ECT group. Raw cognitive scores were transformed into percentile ranks and a composite cognitive score was constructed by averaging domain scores. Assessors were trained clinical personnel and inter-test intervals were usually about 14 days. Statistical analyses used repeated-measures ANOVA/ANCOVA frameworks in SPSS v24. For symptom change a 2 × 2 (Intervention × Time) repeated-measures ANOVA compared percentage change in MADRS from T0 to T1 and T0 to T2, with baseline MADRS included as a covariate. Cognitive change (percentage change T0 to T2) was compared by ANCOVAs with baseline MADRS as covariate when warranted by correlations; otherwise univariate ANOVAs were used. Effect sizes (Cohen's d) were reported for within- and between-group contrasts. Assumptions (normality, homogeneity of variance, linearity) were tested and parametric tests were used when appropriate or robust.

Participant flow and baseline characteristics: After exclusions the study analysed 25 patients in the ketamine arm and 25 in the ECT arm. The groups were similar in age, gender, education and most demographic variables, but differed on clinical variables: ketamine-treated patients had lower baseline MADRS scores (less severe depression) and longer current episode durations (on average twice as long) than ECT patients. Baseline cognitive performance did not differ between groups, and baseline MADRS was not associated with baseline cognition except for verbal memory (moderate negative correlation, r = -0.315, p = .029, N = 48). Antidepressant efficacy: Both interventions produced clinically meaningful symptom reductions. At mid-treatment (T1) mean MADRS was 13.38 (SD = 5.27, N = 24) in the ketamine group versus 19.52 (SD = 7.07, N = 25) in the ECT group. At treatment end (T2) mean MADRS was 13.40 (SD = 6.89, N = 25) for ketamine and 13.75 (SD = 7.69, N = 24) for ECT. A significant Treatment Group × Time interaction was observed, F(1,43) = 6.93, p = .012, partial η2 = 0.139, indicating a faster antidepressant response with ketamine; between-group effect size for this interaction was d = 0.80. Specifically, percentage MADRS reduction from T0 to T1 was greater for ketamine (mean = -47.45%, SD = 23.43, n = 24) than for ECT (mean = -34.86%, SD = 21.29, n = 22). By T2, symptom reduction did not differ significantly between groups, indicating comparable end-of-treatment efficacy. Neurocognitive outcomes: Cognitive domain results varied by treatment. For ECT, a significant decline was observed in verbal memory (large effect sizes reported), whereas visual memory unexpectedly improved (small to moderate effect), possibly reflecting practice effects on that specific test. Executive functions remained relatively stable in the ECT group. In the ketamine group, performance improved significantly in attention (small effect), visual memory (small to moderate effect) and executive functions (small to moderate effect); however, one measure of verbal memory (delayed recall) showed a small but significant decline. Across domains, differences with p ≤ .05 had effect sizes generally greater than d = 0.5; the overall composite cognitive score difference reflected a smaller effect (d ≈ 0.38–0.40). When baseline MADRS was included as a covariate, the composite score comparison favoured ketamine (ECT: mean change -1.84%, SD = 20.34, n = 24; ketamine: mean change 4.07%, SD = 8.35, n = 24), but excluding MADRS from the model rendered the composite difference non-significant (F(1,46) = 1.73, p = .194, n = 48). Relationship between cognition and clinical response: Baseline cognitive performance did not predict symptom change to T2 in either treatment group. Change in cognitive performance and change in MADRS scores were not correlated, suggesting that cognitive improvements were not simply secondary to mood improvement. No systematic associations were found between baseline MADRS and baseline cognition except for the verbal memory correlation noted above. Other reported details: Mean number of ECT sessions was 12.36 (SD = 1.75, range 9–16); mean number of ketamine infusions was 6.76 (SD = 1.23, range 6–9). The extracted text does not provide a detailed adverse-event table or long-term follow-up outcomes.

Basso and colleagues interpret their findings as showing that serial ketamine administration and ECT are both effective acute treatments for depression in a naturalistic inpatient sample, but ketamine produces a faster initial antidepressant effect. Symptom reduction after one week of ketamine (three infusions) was greater than after one week of ECT (six sessions), whereas by the end of the respective courses (two weeks of ketamine versus four weeks of ECT) both groups had similar levels of symptom improvement. The investigators highlight the clinical relevance of a rapid onset of action for situations requiring urgent symptom reduction, such as high suicide risk, and note that additional ketamine infusions after the first week produced little further symptom reduction in their sample. With regard to cognition, the authors report divergent patterns: ECT was associated with a decline in verbal memory but relatively preserved executive function and even improvement in visual memory (which they suggest may reflect practice effects). Ketamine was associated with improvements in attention, visual memory and executive function but a small decline in delayed verbal recall. Overall, ketamine-treated patients showed better short-term global cognitive outcomes than those who received ECT (moderate effect size on the composite), leading the authors to characterise serial ketamine administration as having a small procognitive effect in the short term and a preferable cognitive profile compared with ECT. The study authors caution several limitations that temper interpretation. The design was naturalistic and non-randomised: patients chose treatment, no control group was included, and participants received concurrent and partially changing medications; these factors introduce potential selection and expectation biases. The ketamine group had less severe baseline depression but a longer episode duration; baseline MADRS was included as a covariate and did not materially alter key domain findings, but residual confounding cannot be excluded. The sample was heterogeneous with psychiatric and medical comorbidities, and the requirement to complete neurocognitive testing excluded patients with severe functional impairment, limiting generalisability. Assessments took place 1–3 days after treatment end, so longer-term cognitive and clinical trajectories were not assessed. The authors also note that practice effects may explain some cognitive improvements and that prior literature on ketamine's cognitive effects is mixed; one prior study found transient verbal memory impairment shortly after a single infusion, which may be consistent with their small verbal memory decline. In terms of implications, the investigators suggest that for middle-aged, chronic non-psychotic depressed patients—particularly those with treatment-related cognitive impairment—serial ketamine could be a favourable acute option because it combines rapid mood improvement with better short-term cognitive outcomes than ECT. They recommend randomised controlled trials with larger samples and longer-term follow-up (e.g. 6–12 months) to confirm these findings and to address maintenance strategies given known relapse rates after both ECT and ketamine.