Antisuicidal Response Following Ketamine Infusion Is Associated With Decreased Nighttime Wakefulness in Major Depressive Disorder and Bipolar Disorder

Vande Voort and colleagues frame the study within the substantial public-health burden of suicide and the limited pharmacologic options known to reduce suicide risk. They note that disrupted sleep, particularly difficulty maintaining sleep, is consistently associated with increased suicidal thoughts and death by suicide, and that most prior work has relied on self-report rather than objective measures such as nighttime electroencephalography (EEG). The introduction also observes that although ketamine, an N-methyl-D-aspartate (NMDA) receptor antagonist, has been shown to produce rapid reductions in suicidal ideation, the biological mechanisms underlying this antisuicidal effect are unknown. This study therefore set out to test whether objective changes in nocturnal wakefulness (EEG-verified minutes awake during the night) are associated with an antisuicidal response to a single subanesthetic ketamine infusion in patients with treatment-resistant major depressive disorder or bipolar depression who had baseline suicidal ideation. The investigators focused on hourly wakefulness between midnight and 04:59 AM because earlier work from their group and epidemiologic data suggested this time window is particularly relevant to suicide risk.

This work is a secondary analysis of data from clinical trials conducted between 2006 and 2013 (ClinicalTrials.gov identifier: NCT00088699). The analytic sample comprised 34 inpatients (20 men, 14 women; mean age 44.97 years, SD = 13.28) with baseline suicidal ideation and treatment-resistant depression: 23 with major depressive disorder (MDD) and 11 with bipolar disorder. Diagnoses were established by face-to-face interview and confirmed with the Structured Clinical Interview for DSM-IV (SCID). Inclusion required current major depressive episode and a Montgomery–Åsberg Depression Rating Scale (MADRS) score indicating at least moderate severity; patients with imminent, acute suicide risk were excluded from one of the trials. All MDD participants were medication-free for at least 2 weeks (5 weeks for fluoxetine); bipolar patients were maintained on lithium or valproate at therapeutic levels. A healthy control group (n = 22; 11 men, 11 women; mean age 37.14 years, SD = 10.21) with no current or past Axis I diagnoses was also assessed. All patients received a single intravenous ketamine infusion (0.5 mg/kg over 40 minutes). Polysomnography (PSG) was recorded the night before and the night after infusion using standard EEG leads (C3/A2 and C4/A1), electrooculograms and electromyograms. A blinded reviewer scored EEGs in 30-second epochs as awake or asleep; hourly minutes awake from 00:00 to 04:59 were calculated to capture interruptions in sleep continuity. A subset of patients (n = 11) received riluzole the night after ketamine; planned analyses examined both the full sample (n = 34) and the ketamine-only subsample (n = 23). Suicidal ideation was assessed using the Hamilton Depression Rating Scale (HDRS) suicide item administered 60 minutes before infusion and the morning after infusion; confirmatory analyses used the MADRS suicide item and the first five items of the Scale for Suicide Ideation (SSI5). An antisuicidal response was defined a priori as going from any suicidal ideation at baseline (HDRS suicide item > 0) to no suicidal ideation (HDRS suicide item = 0) one day after ketamine. Given the distribution of minutes awake, the primary analysis used a negative binomial generalized linear mixed model (GLMM) with a log link to compare nocturnal wakefulness between responders and nonresponders. Time was modelled in 1-hour intervals from midnight to 04:59, baseline nocturnal wakefulness (the night before ketamine) was included as a time-dependent covariate, and a first-order autoregressive covariance structure was selected. Sensitivity analyses compared responders to healthy controls, adjusted for diagnosis (MDD vs bipolar), and adjusted for antidepressant response (defined as ≥50% reduction in HDRS score with the suicide item removed). Due to small sample sizes, a Satterthwaite approximation was used for degrees of freedom and robust estimation checked model assumptions.

At baseline, responders and nonresponders did not differ on demographic, clinical, or EEG-verified nocturnal wakefulness measures, except that melancholic depression differed between groups. The primary GLMM, controlling for baseline sleep, showed a significant difference in minutes awake the night after ketamine between antisuicidal responders and nonresponders (F 1,22 = 5.04, P = .04), with responders demonstrating fewer minutes awake; the corresponding effect size was large (Cohen d = 0.96). The time-by-responder interaction was a trend (P = .08). Analyses restricted to the ketamine-only subsample (n = 23) yielded comparable results: the main effect of response was at trend level (P = .059, d = 1.15) while the time-by-responder interaction reached statistical significance (F 4,37 = 3.93, P = .009). Post hoc testing in this subsample identified a significant difference between responders and nonresponders at the 03:00–03:59 hour (p = .02; d = 0.88). Comparing antisuicidal responders to healthy controls produced a trend toward a difference (F 1,40 = 3.15, P = .08, d = 0.56), with no significant group-by-time interaction. Confirmatory analyses using other suicide measures were consistent in direction. Using the MADRS suicide item (n = 36) the main effect of response was significant (P = .04, d = 0.88). Using the SSI5 (n = 24) the main effect was not statistically significant (P = .20) despite a similar effect size (d = 0.88), suggesting limited power in the smaller sample. When diagnosis was added to the model (HDRS suicide item), the main effect of antisuicidal response remained significant (P = .02, d = 0.94) while diagnosis and interaction terms were nonsignificant. Adjusting for antidepressant response attenuated the antisuicidal–wakefulness association to a trend (P = .080, d = 0.49); neither antidepressant response nor its interaction with time was significant.

Vande Voort and colleagues interpret their findings to mean that depressed patients who experience a rapid antisuicidal response to a single subanesthetic ketamine infusion show reduced nocturnal wakefulness the following night, a sleep pattern that approximated that of healthy volunteers. The authors suggest this reduction in sleep fragmentation could represent one neurobiological pathway by which ketamine reduces suicidal ideation, and they place the result in the context of prior work showing rapid decreases in suicidal thoughts after ketamine and associations between nocturnal wakefulness (especially between midnight and 04:59 AM) and next-day suicidal ideation. The discussion outlines plausible mechanistic links without claiming proof: potential pathways include glutamatergic effects, increases in brain-derived neurotrophic factor (BDNF) with downstream effects on synaptogenesis and sleep architecture, and relationships between low-grade inflammation, glutamate signalling, sleep disturbance, and suicidal behaviour. Neuroimaging studies associating suicidality with infralimbic cortex metabolism are cited to indicate that circuits related to suicidal ideation may differ from those underlying overall mood symptoms. The authors acknowledge that controlling for antidepressant response reduced the strength of the association, so the extent to which the antisuicidal effect is independent of broader antidepressant effects remains inconclusive in this sample. Strengths reported by the investigators include objective EEG measurement of sleep, use of multiple suicidal ideation metrics for corroboration, and inclusion of healthy control comparisons. They also enumerate important limitations: exclusion of patients with imminent acute suicide risk, lack of generalisability beyond treatment-resistant depression, focus solely on minutes awake rather than full sleep architecture (eg, REM, slow-wave sleep), limited sample size (which affects power in subgroup and confirmatory analyses), lack of matching between depressed and control samples, and inability to draw causal inferences. The authors call for replication and further work to examine other sleep variables, the role of depressive subtypes such as melancholic depression, relationships with BDNF and inflammatory markers, and whether improving disrupted sleep is a specific mechanism of ketamine's antisuicidal effects or a common pathway shared with other treatments.