This double-blind, randomised, placebo-controlled pilot trial (n=20) tested a single IV dose of ketamine for suicidal thoughts in adolescents attending the emergency department. It found the study was feasible and well tolerated, but ketamine did not clearly reduce suicidal ideation at 40 minutes, although fewer participants were admitted to hospital at the initial visit.
Papers cited by this study that are also in Blossom
Background
Suicidal ideation (SI) is a common reason for emergency department (ED) visits by adolescents. While intravenous (IV) ketamine rapidly reduces SI in adults, its efficacy in adolescents remains unstudied. We assessed the feasibility of a trial of a single dose IV ketamine to reduce adolescent SI in the ED.
Methods
This double-blind, randomized, placebo-controlled pilot trial was conducted from Jan-May 2024. Medically stable adolescents aged 12 to 17 years with moderate-to-severe SI were eligible. Participants were randomized to IV ketamine (0.5 mg/kg; max 50 mg) or IV normal saline (0.5 ml/kg; max 50 mL), infused over 40 min. They were monitored for 120 min and then received usual ED mental health care. The primary outcome was trial feasibility (enrolment and follow-up success). The primary clinical outcome was SI severity at 40 min post-infusion (T-40), measured using the Beck Scale for SI (SSI5), Montgomery-Asberg Depression Rating Scale item 10 (MADRS10) and Beck Depression Inventory item 9 (BDI9). Additional outcomes included hospital admission, adverse events, 30-day ED revisits and death.
Results
Twenty participants were eligible and were enrolled. All participants completed the infusion and day-1 follow-up; 90% completed day-7 follow-up. No serious adverse events occurred. While SI severity did not differ significantly between groups at T-40 (SSI5: p = 0.06; MADRS10: p = 0.19; BDI9: p = 0.18), fewer participants randomized to ketamine were hospitalized at the initial visit (risk difference 40%, 95%CI: 7, 69%).
Conclusions
Recruiting adolescents to an ED-based IV ketamine study for SI is feasible. A larger trial is needed to clarify potential clinical benefits.
Suicidal ideation is a frequent and serious reason for emergency department visits among adolescents, and current emergency care mainly focuses on safety assessment and discharge planning rather than rapid symptom relief. The paper notes that intravenous ketamine can rapidly reduce suicidal ideation in adults, but this approach had not yet been studied in adolescents in the emergency department. Although emerging studies had suggested possible benefits of repeated ketamine protocols for adolescents with depression and suicidality, there were no published evaluations of a rapid-acting intervention delivered in the emergency department for this age group. Schlegelmilch and colleagues therefore set out to assess whether a trial of single-dose intravenous ketamine for suicidal ideation in adolescents presenting to a paediatric emergency department would be feasible, and to gather preliminary data on clinical effects. The study was designed as a pilot to inform a future definitive trial, with attention to both feasibility outcomes and short-term changes in suicidal ideation.
Schlegelmilch and colleagues conducted a single-centre, double-blind, randomised, placebo-controlled pilot trial in the Children’s Hospital of Eastern Ontario emergency department between January and May 2024. The site was a tertiary academic centre with an annual emergency department census of about 75,000 visits. The trial used 1:1 allocation and was registered on ClinicalTrials.gov. Adolescents aged 12 to 17.99 years were eligible if they presented with a mental health complaint, endorsed item 5 of the Ask Suicide-Screening Questionnaire indicating current suicidal thoughts, had moderate-to-severe suicidal ideation defined by a Beck Scale for Suicide Ideation first five items score of at least 3, and were medically stable. Exclusions included acute intoxication, pregnancy or breastfeeding, need for restraint, intellectual disability, a non-psychiatric neurological disorder, active psychosis, contraindications to ketamine, involuntary psychiatric hold, or previous enrolment. Participants received either intravenous ketamine 0.5 mg/kg, up to 50 mg, or intravenous normal saline 0.5 ml/kg, up to 50 mL, infused over 40 minutes. The pharmacy prepared identical-looking study packets according to a computer-generated randomisation list, and participants, caregivers, clinical staff, research staff, outcome assessors, and investigators were all blinded. The primary feasibility outcome was enrolment and completion of infusion and follow-up. The primary clinical outcome was suicidal ideation severity at 40 minutes after infusion start, measured with three tools: the first five items of the Beck Scale for Suicide Ideation (SSI5), item 10 of the Montgomery-Åsberg Depression Rating Scale (MADRS10), and item 9 of the Beck Depression Inventory (BDI9). Secondary measures included suicidal ideation at later timepoints, hospital admission and length of stay, emergency department revisits within 30 days, later hospitalisations, death within 30 days, depressive symptoms using the Children’s Depression Rating Scale-Revised, and dissociation using the simplified Clinician-Administered Dissociative States Scale. The trial was not powered to detect efficacy. The researchers enrolled 20 adolescents to obtain feasibility information and preliminary estimates of outcome distributions. Analyses followed the intention-to-treat principle. Baseline characteristics and safety data were summarised descriptively. Between-group differences in suicidal ideation at 40 minutes were analysed using analysis of covariance adjusting for baseline suicidal ideation, sex, age, and depression severity. Later timepoints were analysed similarly in exploratory analyses without correction for multiple comparisons. Cohen’s d was calculated from adjusted mean differences and pooled standard deviations.
P R E S S emergency care. Our objective was to evaluate the feasibility of conducting a trial utilizing IV ketamine for SI in youth presenting to a pediatric ED.
We conducted a single-centre, double-blind, randomized, placebo-controlled pilot trial in a 1:1 allocation ratio from January to May 2024 in the Children's Hospital of Eastern Ontario (CHEO) ED, a tertiary care academic centre with an ED census of ~75,000 visits per year. The trial was registered at clinicaltrials.gov (ref: NCT06366334) on April 10, 2024.
Adolescents aged 12 to 17.99 years were eligible if they presented to the ED with a mental health complaint, answered 'yes' to item 5 of the Ask Suicide-Screening Questionnaire (ASQ: 'Are you having thoughts of killing yourself right now?'), had moderate-to-severe SI (Beck Scale for Suicide Ideation-first 5 items [SSI5], score ≥3), and were medically stable. Adolescents were excluded if acutely intoxicated, pregnant or breastfeeding; required chemical of physical restraint; had an intellectual disability, a non-psychiatric neurologic disorder, active psychosis, or contraindications to ketamine; were under an involuntary psychiatric hold; or were previously enrolled.
Participants were randomized to receive IV ketamine (0.5 mg/kg; max 50 mg) or IV normal saline (0.5 ml/kg; max 50 mL), infused over 40 minutes.
The pharmacy prepared sequentially numbered, identical appearing packets according to a computer-generated randomization list provided by the study statistician. Ketamine and normal saline were prepared as clear, colorless solutions in identical volumes. Research nurses allocated participants by selecting the "next-in-line" study packet. Caregivers, participants, health care providers, research staff (including outcome assessors), and investigators remained blind to group allocation.
Create a free account to open full-text PDFs.
Wilkinson, S. T., Ballard, E. D., Bloch, M. H. et al. · American Journal of Psychiatry (2017)
Zhou, Y., Lan, X-F., Wang, C. et al. · Child and Adolescent Psychiatry (2024)
Ballard, E. D., Luckenbaugh, D. A., Walls, T. S. et al. · Journal of Psychiatric Research (2015)
Of 129 potentially eligible adolescents seen during research-nurse hours, 73 were screened, 20 were eligible, and all 20 consented and were enrolled. All randomised participants were included in the analysis. Every participant completed the infusion and day-1 follow-up, and 18 of 20 (90%) completed day-7 follow-up; the two losses to follow-up were in the placebo group. The sample included adolescents with prior mental health contact and many were already taking antidepressants. Baseline depressive symptoms were reported as higher in the ketamine group than in the placebo group on the Children’s Depression Rating Scale-Revised. For the primary clinical outcome at 40 minutes, suicidal ideation did not differ significantly between groups on any of the three measures. On the SSI5, ketamine showed a greater reduction than placebo after adjustment, but the difference did not reach conventional significance at 40 minutes (adjusted mean difference -2.5, 95% CI -5.2 to 0.1; p=0.06). The same pattern was seen on the SSI5 at 120 minutes, where the difference became statistically significant (-2.9, 95% CI -5.8 to 0.00; p=0.049), with large effect sizes favouring ketamine (Cohen’s d of -0.9 and -1.2 at 40 and 120 minutes). However, MADRS10 and BDI9 did not show statistically significant between-group differences at 40 or 120 minutes. There were also no significant differences at 80 minutes or day 1 for any suicidal ideation measure, and at day 7 the MADRS10 favoured placebo (adjusted mean difference 2.2, 95% CI 0.1 to 4.3; p=0.04). The authors note that point estimates through day 1 generally favoured ketamine. Regarding clinical disposition, 4 of 10 participants in the placebo group were admitted at the index visit versus 0 of 10 in the ketamine group, corresponding to a risk difference of 40% (95% CI 7 to 69%). Median length of stay among those admitted was 2.5 days. During the 30-day follow-up, emergency department revisits for mental health were reported in 7 of 10 ketamine participants and 1 of 10 placebo participants (risk difference 60%, 95% CI 18 to 84%); two ketamine participants were hospitalised at a repeat visit. No deaths occurred. Safety findings were reassuring in that no serious adverse events occurred. One ketamine participant vomited in the emergency department. Dissociation was reported by some participants before infusion and more often during infusion, especially in the ketamine group (9 of 10 versus 4 of 10 in placebo during infusion), although dissociation also occurred in placebo recipients.
The authors interpret the study primarily as evidence that a randomised emergency department trial of intravenous ketamine for adolescent suicidal ideation is feasible. They emphasise that the pilot was not powered to determine efficacy, so the findings on symptom change and service use should be viewed as preliminary and hypothesis-generating rather than definitive. They state that the suicidal ideation results provide some early signals that ketamine may reduce suicidal thoughts quickly, especially on the SSI5, where effect sizes were moderate to large and point estimates up to day 1 favoured ketamine. At the same time, the lack of clear separation on the other measures, and the fact that some differences did not reach statistical significance, led the authors to caution that the measures may not behave consistently in this setting. They suggest that the different tools capture different facets of suicidality and that the validity and responsiveness of individual items, particularly in acutely distressed adolescents, are not well established. The authors compare their findings with earlier work in adults and with emerging adolescent ketamine research outside the emergency department, noting that there had been no prior published emergency department studies of rapid-acting interventions in youth. They suggest that some adolescents in crisis may have transient situational distress rather than more biologically rooted, treatment-resistant illness, which could help explain heterogeneity in response. They also speculate that ketamine may be more helpful in those with specific clinical or biological risk profiles, and recommend stratifying future studies accordingly. The discussion identifies several limitations and uncertainties. The sample was very small, effect estimates may be unstable and potentially inflated, and the study was conducted at a single centre. The researchers also note baseline imbalances in symptom severity across some measures, the uncertainty around the best instrument for measuring rapid change in adolescent suicidal ideation, and the possibility that dissociation and other acute effects could complicate assessment. They also note that although no ketamine participants were admitted at the index visit, they returned to the emergency department more often within 30 days, which could reflect baseline severity or altered help-seeking behaviour. For future work, the authors estimate that a two-arm trial with moderate effect size assumptions would require about 64 participants per group for 80% power, and that allowing for attrition and site variability, a multi-centre study of roughly 150 to 180 participants would be reasonable. They also say that future trials should monitor downstream outcomes such as emergency revisits and later admissions.
Suicidal ideation (SI) severity was assessed using three instruments: the first five items of the Beck Scale for Suicide Ideation (SSI5; range 0-10),the suicide-specific item (item 10) from the Montgomery-Åsberg Depression Rating Scale (MADRS10; range 0-6),and suicide-specific item (item 9) from the Beck Depression Inventory (BDI9; range 0-3).These single items were selected because they directly assess suicidal thoughts within their respective validated depression scales and have been used in prior ketamine studies to capture rapid changes in suicidality. There is currently no gold standard measure for detecting rapid SI changes in adolescents in acute care settings. While all three measures have demonstrated responsiveness to rapid change in adult ketamine trials, their performance in adolescents receiving ketamine in the emergency department is unknown. Each tool has distinct strengths and limitations: the SSI5 offers multi-item depth but may pose comprehension challenges for youth in acute crisis; the MADRS10 and BDI9 are single-item measures that are pragmatic and efficient but may not capture the multidimensional nature of suicidal ideation. Given this uncertainty regarding responsiveness, feasibility, and appropriateness in our population, we administered all three measures in this pilot study to inform selection of the most suitable primary outcome for a future definitive trial. Depressive symptoms were assessed using the Children's Depression Rating Scale-Revised (CDRS-R; range 17-113, with scores >60 indicating severe depression).Dissociation was measured using the simplified Clinician-Administered Dissociative States Scale (CADSS-6; total score >6 or ≥4 on a single item).
After the 120-minute monitoring period, adolescents received routine ED mental health care. Final disposition decisions were made by the hospital's mental health team and were not influenced by study personnel or procedures. The primary outcome was study feasibility, measured by enrolment success and completion of the infusion/follow-up. The primary clinical outcome was SI severity at 40 minutes after the infusion started, assessed using each of the three SI instruments (SSI5, MADRS10, BDI9). Additional outcomes were SI severity at all subsequent times, hospital admission and length of stay, ED revisits within 30 days, subsequent hospitalizations, and death within 30 days. Adverse events were monitored. The study protocol has been published.
To provide sufficient information on the feasibility outcomes and estimates of the distribution of SI severity measures, we enrolled 20 adolescents. The trial was not powered to detect clinical efficacy of ketamine therapy. Baseline participant characteristics, primary feasibility outcomes, and safety data were reported using descriptive statistics. Measures of central tendency and variance for each SI assessment tool were calculated. Analysis of covariance was used to report between-group differences in SI severity at 40 minutes, adjusting for baseline SI severity, sex, age, and CDRS-R scores. Between-group differences in SI severity at 80-minutes, 120-minutes, 1-day, and 7-days post-intervention was analyzed using analysis of covariance without adjustment for multiple comparisons in this exploratory analysis. Cohen's d was calculated using the adjusted mean difference divided by the pooled standard deviation. Additional analyses examined differences in the proportions hospitalized and their length of stay. The proportions with a return ED visit for mental health, subsequent hospitalization, or death within 30 days were also compared between groups. All statistical hypothesis tests were two-sided; p-values less than 0.05 were considered statistically significant. Efficacy analyses were based on the intention-to-treat principle. All analyses were conducted using R statistical software.
Of 129 potentially eligible adolescents visiting the ED during research nurse hours, 73 were screened by a research nurse; 20 were eligible and all consented to participate. All randomized patients were included in the analysis. All participants completed the infusion and day-1 follow-up. Eighteen (90%) completed day-7 follow-up. The two lost to follow-up were in the placebo group (Figure). A minority of participants had an ED mental health visit in the previous month or a prior mental health hospitalization, while half were on antidepressants. Baseline mean (SD) depression scores measured by CDRS-R were 85.8 (8.3) and 69.7 (9.9) for the ketamine and placebo groups, respectively. Baseline SI severity measured by SSI5 was similar between groups, but different when measured by MADRS-10 and BDI9. (Table) Adjusting for baseline SI of the corresponding measure, depression, age and sex, ketamine was associated with a greater reduction in SSI5 scores, compared with placebo at 40 minutes (adjusted mean difference -2.5, 95%CI: -5.2, 0.1; p=0.06). At 120 minutes, the difference was similar but statistically significant (-2.9, 95%CI: -5.8, 0.00; p=0.049) (Table). These SSI5 differences corresponded to large standardized effect sizes (Cohen's d = -0.9 and -1.2 respectively, favouring ketamine). At these timepoints (40and 120-minutes), differences for MADRS10 and BDI9 were not statistically significant. At day-7, the adjusted mean difference in MADRS10 scores favoured placebo (2.2, 95%CI: 0.1, 4.3; p=0.04). No statistically significant between-group differences were observed at 80-minutes or day-1 for any SI measure, or at day-7 for SSI5 or BDI9. Standardized effect sizes for all comparisons are presented in Table. The unadjusted mean SI severity scores for each measure and time period are presented in Table, along with standardized effect sizes. Four (40%) participants in the placebo group were hospitalized at the enrollment visit versus none in the ketamine group (risk difference [RD] 40%, 95%CI: 7, 69%). Their median length of stay was 2.5 days (IQR 1.5, 4.5). The number returning to the ED within 30 days was 7/10 (70%) for the ketamine group and 1/10 (10%) for placebo (RD 60%, 95%CI: 18, 84%); of
these, two ketamine group participants were hospitalized at a repeat ED visit. No deaths occurred in the study period. One participant in the ketamine group vomited in the ED. Some participants reported dissociation before the infusion began (3/10 ketamine, 1/10 placebo; RD 20%, 95%CI: -17, 53%). During the infusion, dissociation was reported by 9/10 (90%) and 4/10 (40%) in the ketamine and placebo groups, respectively (RD 50%, 95%CI: 9, 77%). Although the trial was not powered to detect clinical efficacy, our findings provide preliminary signals that ketamine may influence important outcomes. No statistically significant difference in SI was observed at 40 minutes, but all point estimates up to and including day 1 favoured ketamine. Moderate to large effect sizes were observed for SSI5 at 40 and 120 minutes, suggesting a possible early reduction in suicidal ideation following ketamine administration. However, effect size estimates in small pilot trials are inherently unstable and may overestimate true effects. These findings should therefore be interpreted cautiously and viewed as hypothesis-generating.
For planning a future definitive trial, assuming a moderate standardized effect size (d=0.5), a two-arm randomized trial with 80% power and α=0.05 would require approximately 64 participants per group. Given potential attrition and site variability, a multi-centre trial enrolling 150-180 participants would be reasonable. No participants in the ketamine group required admission at the index visit, compared with 40% in the placebo group, suggesting a possible role for ketamine in reducing the immediate need for inpatient care. These findings are exploratory and should only be interpreted as hypothesis-generating.
The tools we used to assess SI capture different facets of suicidality, which may explain discrepancies in baseline scores and changes over time across measures. The psychometric validity of individual items, particularly in acute care settings, is less well established.Adolescents in crisis may struggle to interpret or respond to the questions in these tools or hesitate to disclose suicidal thoughts, limiting the sensitivity of these tools to rapid symptom change.Although all participants met criteria for moderate-to-severe SI, for some, SI may have reflected transient situational distress rather than chronic, treatment-resistant depression. This distinction is relevant, as ketamine's proposed mechanisms (e.g., enhanced glutamatergic signaling and synaptic plasticity) are theorized to target biologically rooted mood disorders. Future studies should consider stratifying participants based on clinical and biological risk factors to better identify adolescents most likely to benefit from ketamine therapy. While no participants in the ketamine group were hospitalized at the index visit, they returned to the ED more frequently within 30-days compared to those in the placebo group. This may reflect higher baseline depressive symptom severity, as indicated by their CDRS-R scores, or could indicate that ketamine therapy influenced subsequent help-seeking behaviours. Any future trial would need to monitor outcomes such as ED revisits and subsequent admissions following the enrolment visit.