A Pilot Study of Ketamine Infusion after Suicide Attempt: New Frontiers in Treating Acute Suicidality in a Real-World Medical Setting

Suicide remains a major public health problem with limited options that act rapidly to reduce intense suicidal thoughts. The Introduction frames suicide as multifactorial and emphasises that prior attempts are the strongest single risk factor for later suicide. Conventional treatments for mood disorders generally take weeks to act, inpatient admission carries its own risks and there is a recognised need for rapid-acting interventions that can be deployed in real-world medical settings to stabilise patients after a suicide attempt. Carlos and colleagues set out to evaluate whether a single, subanesthetic intravenous ketamine infusion (0.5 mg/kg) administered during medical stabilisation after a suicide attempt could rapidly reduce suicidal ideation and depression and whether those effects would be sustained over a six-month follow-up. The study was framed as an open-label, deployment-focused pilot in a heterogeneous, transdiagnostic sample of medically hospitalised suicide attempters, with exploratory comparisons to a matched case-control cohort drawn from the same clinical setting.

This was a single-centre, open-label pilot study approved by the University of Pittsburgh IRB and pre-registered (clinicaltrials.gov NCT04154150). Sixteen adults aged 18–65 who had recently attempted suicide and were medically hospitalised were enrolled between December 2019 and January 2021 after identification via Consultation-Liaison (CL) psychiatry consults. Inclusion required capacity to consent and medical suitability for intravenous ketamine. Key exclusions included acute psychosis, delirium, recent MAOI use, pregnancy or breastfeeding, very low reading ability, ketamine hypersensitivity, certain contraindicated medications, recent ECT, or risk of severe withdrawal. Primary diagnoses (via the MINI) were mixed across unipolar and bipolar mood disorders and 19% had one or more substance-use disorders. A matched case-control group (n = 16) meeting the same eligibility criteria was identified retrospectively from CL consults; these individuals were not enrolled and contributed only EMR-derived outcome data. The intervention was a single intravenous infusion of racemic ketamine 0.5 mg/kg in saline, administered over 40 minutes by experienced inpatient nurses with continuous cardiovascular monitoring. Pre-specified vital-sign thresholds and availability of an experienced prescribing physician and CL psychiatrists provided safety oversight. After infusion, participants were transferred to standard psychiatric inpatient care within 1–4 days where usual pharmacological and behavioural treatments were delivered. For four days post-infusion, participants were randomised within the study to an automated cognitive training or sham programme (n = 8 each); these arms were pooled for analyses due to low power. Clinical assessments were collected at baseline (pre-infusion) and at 1, 5 and 12 days, and 1, 3 and 6 months post-infusion. Rater-administered measures included the MADRS for depression (and its suicidal ideation item), the Columbia Suicide Severity Rating Scale (CSSRS) and, after the first four patients, the Scale for Suicide Ideation (SSI). Self-report measures included the Adult Suicide Ideation Questionnaire (ASIQ) and the QIDS-SR. Retention varied across timepoints (100% day 1; 94% day 5; 63% day 12; 56% month 1; 69% month 3; 75% month 6). Analyses compared each post-infusion timepoint to baseline using paired t-tests; effect sizes (Cohen's d) were reported. An exploratory structured chart review of the broader UPMC electronic medical record captured outcomes for both enrolled patients and the matched case-controls, including length of psychiatric stay, therapy group attendance, outpatient follow-up attendance, rehospitalisations, substance-related events and suicide re-attempts.

Safety: Adverse events within 24 hours were generally mild and consistent with known subanesthetic ketamine effects. One patient developed moderate anxiety, distress and nausea during infusion and requested discontinuation; ondansetron was given and the participant recovered rapidly. That patient's data were retained in analyses under an intention-to-treat principle. No infusions required emergency medical interventions based on the pre-specified monitoring thresholds. No deaths occurred during follow-up. Rapid clinical outcomes in the ketamine sample: By 1 day post-infusion, total depression and suicidal ideation scores fell substantially from baseline across all measures, with authors reporting large-to-very-large effect sizes (see figures/tables). At day 1, 56.3% of patients met a treatment responder definition (≥50% reduction in MADRS total score), and 75% had full remission of passive or active suicidal ideation on the CSSRS (worst SI score = 0). These improvements were maintained at day 5 and day 12. Over the longer term, mean reductions from baseline on most measures were sustained at 1, 3 and 6 months among participants who completed follow-up, with the exception of the QIDS-SR at months 3 and 6 where sustained change was not observed. Using the CSSRS during follow-up interviews, two patients (12.5%) reported at least one repeated suicide attempt captured in those assessments. Exploratory EMR comparisons to matched case-controls: In chart-based outcomes over six months, no statistically significant differences emerged between the ketamine-treated group and matched case-controls on several indices. Length of initial psychiatric inpatient stay averaged 11.63 days (SD 15.76) for ketamine patients versus 7.33 days (SD 5.42) for controls (t29 = -1.0, p = 0.326). Number of therapy groups attended during the inpatient stay averaged 32.63 (SD 59.72) versus 18.86 (SD 12.31) (t28 = -0.85, p = 0.405). Attendance at the first scheduled outpatient visit was 33% in the ketamine group and 50% in controls (Fisher's exact p = 0.46). Five patients (15.6%) had a documented suicide re-attempt in the EMR during follow-up (three ketamine, two controls; Fisher's exact p = 1.00). Psychiatric rehospitalisations occurred in nine patients (six ketamine, three controls; Fisher's exact p = 0.433) and medical rehospitalisations in nine patients (three ketamine, six controls; Fisher's exact p = 0.44). Substance-related inpatient stays, accidental overdoses and positive urine screens occurred in small numbers distributed across both groups; no significant between-group differences in outpatient visit frequency or types were observed (p's ≥ 0.394).

Carlos and colleagues interpret the pilot data as demonstrating that a single subanesthetic IV ketamine infusion administered during medical stabilisation after a suicide attempt is feasible, safe in this setting and associated with rapid, large reductions in depressive symptoms and suicidal ideation in a high-risk, transdiagnostic cohort. The rapid improvements evident at 24 hours were observed prior to transfer to specialised psychiatric care, which the authors note as consistent with ketamine's known fast onset of action. Longer-term reductions in research-assessed symptoms were reported to be maintained across a six-month window for most measures, although missing data and attrition limited complete longitudinal assessment. In positioning the findings relative to earlier work, the authors note that prior ketamine trials often excluded patients with acute medical needs, substance-use disorders or other comorbidities; by contrast, this deployment-focused study intentionally included such patients to improve generalisability to real-world settings. They point to prior randomized and open-label ketamine research showing rapid antisuicidal effects and argue that the present pilot extends those observations to medically hospitalised recent attempters. The paper also situates ketamine among other emerging rapid-acting agents (for example buprenorphine, brexanolone/zuranolone and oral ketamine trials) and highlights differences in inclusion criteria, duration and generalisability across studies. Key limitations acknowledged by the authors include the open-label design, small sample size and inability to disaggregate ketamine-specific effects from concurrent psychotropic medications and subsequent inpatient treatment-as-usual. Attrition and incomplete follow-up for some participants could bias results; the extracted text does not provide complete data on missingness mechanisms. The exploratory EMR comparisons did not show significant differences between treated patients and matched controls on several utilisation and safety outcomes, and the authors interpret those null findings as initial reassurance about safety and absence of clear iatrogenic harms in this small pilot sample. Finally, the authors emphasise that their findings are preliminary and that a larger, randomised trial is underway to address these limitations and better determine efficacy and longer-term effects.

The authors conclude that a single subanesthetic IV ketamine infusion given during routine medical stabilisation after a suicide attempt may rapidly reduce the intensity of depression and suicidal ideation, with many patients showing large improvements within 24 hours and sustained reductions on most measures through six months. Recognising the pilot study's limitations, Carlos and colleagues report that a larger randomized controlled trial (R01MH124983; clinicaltrials.gov NCT04578938) is in progress to randomise 200 medically admitted suicide attempters to ketamine infusion versus treatment-as-usual, with structured assessments over 12 months to more rigorously evaluate safety, efficacy and longer-term outcomes.