Intravenous

A recent meta-analysis conducted by Wilkinson and colleagues¹³ described ketamine trials for patients with major depression, bipolar depression, or post-traumatic stress disorder for suicidal ideation (SI; Table). The authors identified 10 ketamine trials for psychiatric disorders that met the study inclusion criteria of single-dose trials, had a comparator group (placebo or midazolam), and utilized the IV route. The study authors assessed the effect of ketamine to resolve SI as compared with a control using a general linear mixed model with covariates including age, sex, race, treatment setting, diagnosis, and concomitant psychotropic medications.

TABLE Summary of published reports for ketamine and self-injurious behavior

View Fullscreen

In total, data were obtained for 298 patients from the selected trials. Of those, 167 were included due to baseline SI. Patients included in the analysis had more severe depressive symptoms at baseline as measured by the MADRS (33.4 vs 25.9, P < .001), HAM-D (20.5 vs 16.3, P < .001), Quick Inventory of Depressive Symptomatology – Self-Report¹⁵ (17.7 vs 14. 2, P < .001), and Beck Depression Inventory¹⁶ scales (29.2 vs 21.1, P < .001). Of the included patients, those who received ketamine did not differ significantly from control groups with regard to age, sex, diagnoses, inpatient status, or proportion receiving concomitant psychotropic medications. The two groups did not differ in their baseline MADRS scores either (33.8 vs 32.9, P = .388).

Patients who received ketamine had lower scores on clinician-rated scores of suicidality on days 1 and 7 using HAM-D or MADRS (group-by-time interaction, chi-square = 50.6, P < .001). The number needed to treat for ketamine for being free of SI ranged from 3.1 to 4.0 for time points between days 1 and 7 after ketamine infusion. Changes associated with the individual SI item on these scales was significantly correlated with overall depressive symptoms (r² = .411, t = 10.73, P < .001). When controlling for change in depressive symptoms, the effect on SI remained significant (time-by-treatment interaction, chi-square = 10.84, P = .028).

This meta-analysis should be considered in the setting in which there were relatively few patients in the final analysis. In addition, the measurement of suicidal ideation in this patient population often involved a single item on either the MADRS or HAM-D. These measures are not as sensitive specific suicidal ideation assessments and were not originally validated to be used for this purpose. Other considerations include the relatively short follow-up of 7 days and the patients not being deemed as “imminent suicide risk.”

Another meta-analysis was conducted in 2017 by Bartoli and colleagues.¹⁷ This meta-analysis assessed 5 single-arm trials of IV ketamine at reducing SI acutely (within 4 hours of infusion). This study utilized a different set of studies than the Wilkinson group¹³ as they utilized single-arm studies. Bartoli et al¹⁷ included 99 patients, 63 of which received IV bolus of ketamine 0.2 mg/kg (over 1 minute), and the remaining 36 received a ketamine 0.5 mg/kg infusion (over 40-45 minutes). A decrease in SI was noted by a standard mean difference of –0.92 (95% confidence interval [CI] –1.4 to –0.44, P < .001). The study found a nonsignificant reduction in SI in the ketamine bolus group as compared with the ketamine infusion group (standard mean difference: -2.11 vs -0.86, P = .27).

The results of this study should be considered with several limitations. There was no comparator group in any of the 5 included ketamine trials, making it difficult to discern the absolute effect of ketamine on SI. In addition, the authors primarily considered the end point of 4 hours postinfusion. Although the acute management of SI is important, the degree of durability of this effect is also important to consider. Although the inclusion criteria for the meta-analysis were trials with specific and valid measurements of suicidality, outcomes for specific scales were not reported, nor were the differences in scales used among the trials. Finally, since publication, 1 of the trials¹⁸ in the included analysis was retracted from publication due to inaccuracy in the reported data.

A randomized, midazolam-controlled trial¹⁹ to assess SI in depressed patients was published recently. This study assessed suicidality using the Scale for Suicidal Ideation (SSI) 24 hours after infusion of ketamine or midazolam. They included patients with a score of 4 or greater on the SSI, which is considered a significant cutoff for SI. The scale consists of 19 items rated 0 to 2 for a maximum score of 38. Depression was also assessed utilizing HAM-D and Beck Depression Inventory.

The study¹⁹ included 82 participants who were randomized to 1 of the treatment arms although 1 participant in each group dropped out prior to day 1 assessment; thus, 40 patients were analyzed in each group. The groups were similar at baseline with the exception of percentage of patients with borderline personality disorder (8% in midazolam group vs 28% in ketamine group, P = .03). The baseline scores for the SSI in ketamine and midazolam groups were 14.3 and 15.7, respectively. The primary outcome was SSI score, which was reduced by 4.96 points in the ketamine group compared with the midazolam group (95% CI 2.33-7.59, P < .001). Including borderline personality disorder as a covariate did not significantly alter the results (estimate = 4.76, 95% CI 1.89-7.63, P = .002). Responders were defined as those with a 50% or greater reduction in the SSI score. Fifty-five percent of ketamine patients responded at day 1 postinfusion as compared with 30% of the midazolam group (odds ratio: 2.85, 95% CI 1.14-7.15, P = .024). Although ketamine demonstrated a significantly greater response rate, it should be noted that the midazolam group, serving as an active placebo, demonstrated a stronger response than would be expected.

This trial was limited by the patient population, who were not necessarily treatment-resistant as in other ketamine trials, which may have contributed to the high response to midazolam treatment. In addition, there was a higher proportion of patients with borderline personality disorder in the ketamine group; however, subsequent analysis did not reveal a significant interaction with the results.

Several other randomized controlled trials of ketamine's antisuicidal properties have been conducted^20-23 and have demonstrated reduction in SI. One of these studies,²⁰ interestingly, occurred in patients with a recent cancer diagnosis but no past psychiatric history. The study demonstrated a significant reduction in SI as measured by both the Beck Scale for Suicidal Ideation as well as the MADRS-SI question as compared with a midazolam-controlled group. Kudoh and colleagues²¹ assessed patients with MDD who were undergoing orthopedic surgery. This study utilized ketamine as part of an anesthetic regimen and, thus, utilized higher initial doses of ketamine, 1 mg/kg. The study compared patients with MDD who did not receive ketamine as part of their general anesthesia to those who did receive ketamine. The study found that depressed patients who received ketamine had a significantly improved HAM-D score overall and a significantly reduced score on the individual SI item of the HAM-D than patients who did not receive ketamine in their general anesthesia. In a cohort of treatment-resistant MDD patients,²² 53% scored 0 on 3 explicit suicide measures after treatment with IV ketamine at a dose of 0.5 mg/kg over 40 minutes. A proof-of-concept study²³ also demonstrated a greater reduction in SSI score in patients with bipolar disorder treated with 0.5 mg/kg of ketamine over 40 minutes as compared with those treated with a midazolam control.

Intravenous ketamine has also been studied and demonstrated benefit in patients with treatment-resistant depression and SI in 3 open-label, single-arm studies.^24-26 Two of these studies^24,26 demonstrated that depressed patients scored significantly lower on the MADRS-SI item (averaging a 1- to 2-point reduction) after receiving a parenteral infusion of ketamine. Thakurta and colleagues²⁵ showed a reduction in SI utilizing a validated scale for SI; however, the demonstrated benefit seemed shorter than in other studies with benefit lasting approximately 1 day.

Finally, 8 publications^27-34 of case reports/series report both subjective and objective improvements in SI following IV infusion of ketamine (typically at 0.5 mg/kg over 40 minutes) although 1 study²⁷ utilized esketamine at a dose of 0.6 to 0.8 mg/kg. Even though the generalizability of these cases is low and warrants caution with application of the results, it is important to note that some of these cases represent patients who have presented to the hospital immediately following a highly lethal suicide attempt.^27,29,31 In another case series, Niciu et al³² describes an initial improvement in depressive symptoms in 2 patients with comorbid obsessive-compulsive disorder several hours after ketamine infusion. However, after approximately 1 day, these patients had worsening of depressive symptoms, including subjective reports of SI without plan or intent despite negative suicidality upon hospitalization.

Intramuscular

One case report³⁵ of IM ketamine has been reported to have subjective benefit in a patient with bipolar disorder and SI in the emergency department setting (Table). This patient received a ketamine 0.5 mg/kg IM injection with subjective improvement in SI. Of note, this patient also had a co-occurring pain disorder that the authors hypothesized would also improve with IM ketamine. The patient was discharged from the hospital in 3 days and had no SI at the follow-up appointment 30 days later.

Intranasal

A recent multicenter, double-blind, randomized, placebo-controlled trial³⁶ of IN esketamine found a reduction in SI as measured by the MADRS-SI item, 4 hours after the dose (Table). This effect was not significant at 24 hours postdose. Besides the formulation, this trial differs from other trials in that it utilized a fixed dose of esketamine for each patient. In addition, the patients included had a baseline suicidality that warranted inpatient admission. Although the MADRS-SI item demonstrated benefit at 4 hours, the clinician global judgment of suicide risk rating never differentiated from placebo at any time point.

Another study³⁷ that assessed the utility of IN ketamine reported results of a retrospective review of survey data. This study documented the experience from an outpatient provider in treating primarily pediatric and adolescent patients (mean age = 15 years) with bipolar disorder with the fear of harm phenotype. The fear of harm phenotype is characterized by treatment-resistant bipolar disorder that often presents with physical aggression and separation anxiety. Patients in this study were treated with IN ketamine for an average of 1.71 ± 1.36 years. During dose titration, patients were assessed clinically twice weekly. A principal component analysis of the survey data showed that SI improvement was among the most impactful measures assessed in patients who were treated with IN ketamine with a mean dosage of 165 mg every 2 to 5 days.

Finally, a case report³⁸ of a patient who was previously enrolled in an IV ketamine study but continued treatment with IN ketamine was published in 2016. This case report documents a patient with a significant past history of substance abuse and multiple suicide attempts. The prescribed dose of IN ketamine in this case is much higher than other reported IN ketamine doses. The patient did not significantly improve on ketamine or other treatment modalities and eventually died due to presumed suicide involving a motor vehicle accident.

Oral/Sublingual

Several case reports^39,40 have been published documenting the use of oral ketamine in patients with bipolar depression or depressive symptoms (Table). These reports use weight-based dosing of ketamine ranging from 1.5 to 3 mg/kg as well as fixed doses ranging from 16 to 128 mg/d. Although these reports have documented improvement in the MADRS-SI item and subjective improvement in SI, these cases should be carefully considered in the setting of great variability of dosing.