A Randomized Double-Blind Midazolam-Controlled Trial of Low-Dose Ketamine Infusion in Patients With Treatment-Resistant Depression and Prominent Suicidal Ideation

Su and colleagues situate their study in the context of rising global suicide rates and a growing literature on rapid antisuicidal effects of low-dose ketamine. They note prior randomised trials and meta-analyses showing rapid reductions in suicidal ideation, but also highlight heterogeneity in those samples and in treatment refractoriness; some prior trials included many participants who were not in major depressive episodes or had low levels of treatment resistance, which may have influenced durability of response. The authors further point to inconsistent findings about how long ketamine's antisuicidal effect endures in people with treatment-resistant depression (TRD). This study therefore set out to test whether a single low-dose ketamine infusion (0.5 mg/kg) produces rapid and sustained antidepressant and antisuicidal effects in outpatients with TRD and prominent suicidal ideation, compared with an active control (0.045 mg/kg midazolam). Su and colleagues also aimed to examine whether clinical features of refractoriness — level of treatment resistance, duration of the current depressive episode, and number of prior antidepressant failures — moderated ketamine's effects. The hypothesis was that ketamine would show rapid benefits that would be greater and more durable among patients without severe refractoriness or chronic episodes.

This was a randomised, double-blind, midazolam-controlled trial conducted in adult outpatients aged 20–64 years diagnosed with major depressive disorder (DSM-5) who had not responded adequately to at least two antidepressants and who had prominent suicidal ideation, defined as a score of ≥4 on MADRS Item 10. Ninety patients were screened; six were excluded (three for hypertension, one for type 2 diabetes mellitus, two who declined consent), and 84 were randomised to a single intravenous infusion of either 0.5 mg/kg ketamine or 0.045 mg/kg midazolam. The extracted text does not clearly report additional exclusion criteria beyond the brief notes given. Depressive symptoms were assessed with the Montgomery–Åsberg Depression Rating Scale (MADRS) at baseline, 40, 80 and 240 minutes postinfusion, and on Days 2, 3, 5, 7 and 14. Treatment response was defined as ≥50% reduction in MADRS at Day 2 or Day 3. Suicidal ideation was measured with the Columbia–Suicide Severity Rating Scale Ideation Severity Subscale (CSSRS-ISS) at baseline, 240 minutes, and Days 2, 3, 5, 7 and 14; and with the self-reported Positive and Negative Suicide Ideation Inventory (PANSI) at baseline, 240 minutes, and Days 2, 3, 7 and 14. Level of treatment refractoriness was categorised using the Maudsley Staging Method (MSM) as low (≤7), moderate (8–10) or high (≥11). Statistical analysis used generalized estimating equations (GEE) with treatment (ketamine vs midazolam) as a between-patient factor and time as a within-patient factor, including interactions. The investigators also examined the moderating roles of treatment refractoriness (low/moderate vs high), episode duration (<24 vs ≥24 months) and number of failed antidepressants (≤4 vs >4) on symptom trajectories. Given prior evidence that antisuicidal effects may be short-lived, additional analyses restricted to follow-up <7 days were performed. Two-tailed P < .05 was considered significant. Analyses were performed using SPSS version 17.

Of 90 screened patients, 84 were randomised after exclusions; baseline characteristics were similar across groups for age, sex, BMI, age at illness onset, history of suicide attempt and treatment refractoriness, although the midazolam group had slightly higher mean total MADRS scores at baseline (38.26 ± 3.83) than the ketamine group (35.83 ± 4.53, p = .010). Baseline scores for MADRS Item 10, CSSRS-ISS, and PANSI subscales did not differ by group. On the primary outcome of depressive symptoms (MADRS total), GEE models identified a significant antidepressant effect favouring ketamine up to Day 14 (p = .035). Treatment response rates (≥50% reduction on MADRS at Day 2 or Day 3) were greater in the ketamine group (35.7%) than in the midazolam group (11.9%, p = .020). Stratified analyses showed that the antidepressant effect was present in patients with low or moderate treatment refractoriness (p = .004) but not in those with high refractoriness (p = .371). Similarly, benefit was observed particularly among patients whose current depressive episode had lasted <24 months (p = .015) and among those with ≤4 failed antidepressant treatments (p = .023); no significant antidepressant effect was seen in patients with episode duration ≥24 months (p = .329) or >4 failed antidepressants (p = .167). Regarding suicidal outcomes, more patients receiving ketamine achieved full remission of suicidal ideation (CSSRS-ISS = 0) from Day 2 (14 vs 4, p = .015) through Day 5 (11 vs 3, p = .038) compared with midazolam. GEE models indicated significant antisuicidal effects of ketamine measured by CSSRS-ISS (p = .040) and MADRS Item 10 (p = .023) lasting at least to Day 5; however, these clinician-rated effects diminished thereafter. Subgroup analyses mirrored antidepressant findings: antisuicidal benefits were present in low/moderate refractoriness (CSSRS-ISS p = .001; MADRS Item 10 p < .001) but not in high refractoriness. Only patients with episode duration <24 months and ≤4 failed antidepressants derived antisuicidal benefit (CSSRS-ISS p = .042 and p = .027; MADRS Item 10 p = .005 and p < .001, respectively). On self-reported measures, the PANSI negative-suicide-ideation subscale (PANSI-NSI) showed a significant group difference that persisted to Day 7 (p = .039), suggesting reduction in negative suicidal ideation; the PANSI protective-suicide-ideation subscale (PANSI-PSI) did not differ between groups (p = .077). Adverse effects that occurred during infusion and resolved by 80 minutes included derealisation (29 vs 7, p < .001), dizziness (24 vs 5, p < .001) and crying (6 vs 0, p = .026) in ketamine versus midazolam. There were three serious events of attempted suicide requiring hospitalisation: one in the ketamine group (Day 10) and two in the midazolam group (Days 8 and 12).

Su and colleagues interpret their findings as showing that a single low-dose ketamine infusion produced a rapid antidepressant effect that persisted up to 14 days in outpatients with TRD and prominent suicidal ideation, while antisuicidal effects on clinician-rated measures were apparent but attenuated after about 5–7 days. They emphasise that treatment timing and refractoriness moderated outcomes: only patients with low or moderate treatment refractoriness, current episode duration <24 months, or ≤4 prior antidepressant failures benefitted reliably from ketamine. The authors place these results alongside prior studies showing variable durability of ketamine's antisuicidal effect and with previous work identifying number of prior failed treatments and episode duration as predictors of response to other interventions. They note consistency with earlier findings that highly treatment-refractory patients may show limited benefit from ketamine. Su and colleagues also argue that the PANSI results indicate ketamine primarily reduces negative suicidal ideation rather than increasing protective or positive suicidal-coping cognitions. Regarding clinical implications, the investigators propose two recommendations drawn from their data: optimise timing of ketamine treatment in TRD according to disease and treatment course, and consider biweekly infusion for suicide prevention because the antisuicidal effect appears to last only about 5 days. They acknowledge key limitations of the study: ketamine was used as an add-on treatment rather than as a monotherapy, so concomitant medications were not discontinued; only a single infusion was administered, so durability beyond two weeks and effects of repeated dosing were not assessed. The authors therefore call for further studies of repeated low-dose ketamine to evaluate sustained antidepressant and antisuicidal effects in severe depression with suicidal ideation.