This randomised, double-blind, placebo-controlled trial (n=50) tested low-dose sublingual buprenorphine after ketamine in adults with major depressive disorder and suicidal thoughts. Buprenorphine led to a greater and longer-lasting reduction in suicidal ideation than placebo, while depression scores were similar and no serious treatment-related side effects occurred.
Objective
Ketamine rapidly reduces suicidal ideation in major depressive disorder (MDD), but its effects are transient. Preclinical and clinical studies suggest that ketamine’s antidepressant and antisuicidal effects may be partly mediated by mu-opioid receptor (MOR) modulation. The authors investigated the efficacy and safety of low-dose sublingual buprenorphine, a partial MOR agonist, as a follow-on treatment to prolong the effects of intravenous ketamine.
Methods
This was a randomized, double-blind, placebo-controlled trial conducted at a single outpatient center in the United States. Adults with MDD and a total score ≥6 on the Scale for Suicide Ideation (SSI) were randomly assigned in a 1:1 ratio to receive either sublingual buprenorphine (0.2 to 0.8 mg/day) or a matched placebo for 4 weeks, beginning 48 hours after a single open-label intravenous ketamine infusion (0.5 mg/kg over 40 minutes). The primary outcome was the change in SSI total score, assessed weekly from day 1 through day 31.
Results
From November 2020 to March 2025, 50 participants (68% female) received ketamine, of whom 45 completed at least 1 week of follow-on treatment. Both groups showed significant reductions in SSI total scores, with greater improvement in the buprenorphine group (mean change, −11.6, SD=5.8; N=23) than the placebo group (mean change, −6.3, SD=7; N=22) (Glass delta=0.76, 95% CI=0.11, 1.39). Mixed-effects modeling showed a significant time-by-treatment interaction (p<0.001). Depression scores did not differ significantly between groups. No serious treatment-related adverse events occurred.
Conclusions
This randomized controlled trial provides the first evidence that a pharmacological intervention, buprenorphine, significantly sustains and enhances the antisuicidal effects of ketamine in MDD. These findings offer a potentially scalable and safe therapeutic option for a population at risk of suicide.
Papers cited by this study that are also in Blossom
Berman, R. M., Cappiello, A., Anand, A. et al. · Biological Psychiatry (2000)
The authors begin by noting that suicide rates and suicide-related emergency visits have continued to rise, while pharmacological options specifically aimed at reducing suicidal ideation in major depressive disorder remain limited. They highlight ketamine as a rapid-acting treatment that can reduce suicidal ideation within hours, but emphasise that its benefits are usually short-lived and often require repeated infusions, which brings concerns about cost, safety data, and misuse. They also describe earlier work suggesting that ketamine’s antidepressant and antisuicidal effects may involve mu-opioid receptor pathways, including low-dose buprenorphine studies that showed some reduction in suicidal ideation but did not fully resolve symptoms. Against that background, Tucciarone and colleagues set out to test whether low-dose sublingual buprenorphine could be used after a single intravenous ketamine infusion to prolong ketamine’s antisuicidal effect. The main question was whether participants with major depressive disorder and clinically significant suicidal ideation would show greater reductions in suicidal ideation with ketamine followed by buprenorphine than with ketamine followed by placebo. The study was designed as a randomised, double-blind, placebo-controlled trial to evaluate both efficacy and safety in an outpatient setting.
The researchers conducted a single-centre, randomised, double-blind, placebo-controlled trial at Stanford University School of Medicine. Adults aged 18-70 years with major depressive disorder or bipolar II disorder, currently in a major depressive episode of at least 8 weeks, were eligible if they had suicidal ideation at a clinically significant level, confirmed by the Scale for Suicide Ideation (SSI) and the Columbia-Suicide Severity Rating Scale. Participants also had to have had inadequate response to at least one antidepressant given at an adequate dose and duration, or intolerance to two or more antidepressants. Some stable antidepressants, specifically selective serotonin reuptake inhibitors or serotonin-norepinephrine reuptake inhibitors, were allowed if doses had been unchanged for at least 4 weeks. Exclusion criteria included pregnancy, breastfeeding, opioid use disorder, recent substance use disorder other than nicotine, current opioid use, a positive urine drug test, and severe psychiatric comorbidities such as borderline personality disorder, eating disorders within the past year, or psychotic symptoms. Participants were recruited from a Stanford clinic database, public announcements, and referrals. Randomisation was 1:1 to ketamine followed by buprenorphine or ketamine followed by placebo, with masking of participants, treating investigators, and outcome assessors. All participants first received a single open-label intravenous ketamine infusion at 0.5 mg/kg over 40 minutes. Forty-eight hours later, they began 4 weeks of sublingual troches containing either buprenorphine or placebo. Dosing started at 0.2 mg/day and could be increased weekly by 0.2 mg/day to a target range of 0.6-0.8 mg/day, depending on tolerability and clinical response. Participants self-administered the medication at home, and adherence was monitored by dose counts and returned troches. After the 31-day treatment period, participants were followed for 2 additional weeks off medication to observe discontinuation effects. The primary endpoint was change in SSI total score from day 1 to day 31, measured weekly. Secondary outcomes were changes in depression severity on the Montgomery-Åsberg Depression Rating Scale (MADRS) and the 17-item Hamilton Depression Rating Scale (HAM-D). Safety assessments included weekly adverse event review, and withdrawal symptoms were assessed during follow-up using the Clinical Opiate Withdrawal Scale (COWS) and the Subjective Opiate Withdrawal Scale (SOWS). The main efficacy analysis used a modified intention-to-treat sample and linear mixed-effects models with treatment, time, and treatment-by-time interaction terms.
Between November 2020 and March 2025, 79 people were assessed and 50 were ultimately randomised after ketamine infusion; 45 participants completed at least 1 week of follow-on treatment and were included in the primary analysis. Of those who received ketamine, 68% were female, the mean age was 37.6 years, and the sample had substantial illness burden, including a mean baseline MADRS score of 34.1 and mean baseline SSI score of 15.0. Most participants were White, and 62% met criteria for treatment-resistant depression. Average adherence to study medication was 87%, and the mean final buprenorphine dose in the active arm was 0.77 mg/day. During the ketamine phase alone, suicidal ideation fell rapidly: SSI scores decreased by a mean of 9.1 points from day 1 to day 3, and 54% of participants met the response criterion of at least a 50% reduction in SSI. Depression scores also improved, with MADRS scores decreasing by a mean of 13.4 points and 38% meeting the response criterion. Across the 4-week extension phase, SSI scores improved in both groups, but more strongly with buprenorphine. From day 1 to day 31, the mean SSI change was -11.6 in the ketamine-plus-buprenorphine group versus -6.3 in the ketamine-plus-placebo group, with a reported Glass delta of 0.76 (95% CI 0.11 to 1.39). The linear mixed-effects model showed a significant time-by-treatment interaction (p<0.001). Group differences emerged by day 10 and were largest around days 24 and 31. At day 31, 78% of participants in the buprenorphine group met SSI response criteria compared with 48% in the placebo group, corresponding to a number needed to treat of 3. Depression improved in both groups, but the between-group difference was not statistically significant. MADRS scores fell by 15.8 points in the buprenorphine group and 8.5 points in the placebo group, but the time-by-treatment interaction was not significant (p=0.104). HAM-D results were similar, with no significant interaction (p=0.18). Thus, the main signal was specific to suicidal ideation rather than depression severity. No serious treatment-related adverse events occurred. Common side effects included dizziness, oral burning, headache, and nausea, with some symptoms reported more often in the buprenorphine arm. In the 2-week post-treatment follow-up, suicidal ideation and depression scores rose after stopping study medication in the buprenorphine group, with significant increases from day 31 to day 38, but there was no significant worsening during the second week off treatment. Clinician-rated opioid withdrawal remained in the minimal-to-none range in both groups, and self-rated withdrawal was also minimal and did not differ significantly. At day 45, half of the participants who answered a blinding question guessed their treatment correctly, suggesting masking was only partly preserved.
The authors interpret the findings as evidence that low-dose buprenorphine can extend and strengthen ketamine’s antisuicidal effect in major depressive disorder. They describe this as the first randomised controlled trial showing that a pharmacological follow-on intervention can maintain ketamine’s benefit for suicidal ideation. In their view, the sequential ketamine-plus-buprenorphine approach performed better than buprenorphine alone in previous work and may be more effective than follow-on treatment without ketamine pre-treatment. They place the results in the context of earlier attempts to prolong ketamine’s effects, noting that several pharmacological strategies had not shown clear advantage over placebo, whereas one behavioural intervention had preserved ketamine’s antidepressant effects but had not been tested in a suicidal population. They also compare their findings with the earlier low-dose buprenorphine trial by Yovell and colleagues, arguing that their participants reached lower suicidal ideation scores and had a larger reduction, possibly because of the higher final buprenorphine dose, different population, outpatient setting, and exclusion of borderline personality disorder. They further suggest that the separation between suicidal ideation and depression outcomes supports the idea that ketamine’s antisuicidal effects may partly operate independently of its antidepressant effects. The authors acknowledge several limitations. The trial was not powered to determine whether response to ketamine predicted response to buprenorphine, and expectancy was not measured. The sample was relatively small, suicidal ideation at baseline was somewhat lower than in some earlier studies, and the findings may not generalise to people with comorbid substance use disorders because such patients were excluded. They also note uncertainty about the optimal duration of buprenorphine treatment, the possibility of tolerance or dependence with longer use, and whether tapering might reduce the transient symptom worsening seen after stopping medication. Further randomised trials are needed to confirm efficacy, test antidepressant effects in larger samples, and compare this approach with repeated ketamine infusions or other strategies. The authors also emphasise that the study was outpatient, but their findings may still have relevance across clinical settings.
This randomized, double-blind, placebo-controlled trial was conducted at Stanford University School of Medicine. The Institutional Review Board at Stanford University School of Medicine approved the protocol and consent form. Eligible participants were males and females 18-70 years of age with a diagnosis of MDD or bipolar II disorder, currently experiencing a major depressive episode of at least 8 weeks in duration, as defined by DSM-5. All diagnoses were confirmed during baseline assessments by board-certified psychiatrists and the Mini International Neuropsychiatric Interview. To ensure that participants had received prior treatment with a standard antidepressant during the current episode, they were required to have a documented inadequate response to at least one antidepressant administered at an adequate dosage and duration (≥8 weeks) or a confirmed intolerance to two or more antidepressant medications. The adequacy of antidepressant dosage and duration was assessed according to the Massachusetts General Hospital Antidepressant Treatment Response Questionnaire criteria. We deemed it ethically inappropriate to enroll participants presenting with SI who had not been adequately treated during the current episode. Concurrent use of a selective serotonin reuptake inhibitor or serotonin-norepinephrine reuptake inhibitor was permitted, at a stable dosage for a minimum of 4 weeks prior to enrollment and continued throughout the study. Participants were not allowed to initiate or change their current medications during the study. Participants were required to meet a minimum threshold score of 6 on the Scale for Suicide Ideation (SSI)and to be in good general health. The original inclusion criterion was an SSI total score ≥11, but this was lowered to 6 to facilitate recruitment efforts during the COVID-19 pandemic. An SSI score of 6 or higher is commonly used as the cutoff for clinically significant suicidal ideation, as previous studies have shown that this threshold provides the highest accuracy in predicting future suicidal behavior. Of the 50 participants enrolled, six had baseline SSI total scores below 12. Suicidal ideation was confirmed using the Columbia-Suicide Severity Rating Scale (C-SSRS). Exclusion criteria included pregnancy, breastfeeding, history of opioid use disorder, any substance use disorder (excluding nicotine) within 6 months of screening, current or chronic use of opioids, positive urine drug test, severe psychiatric comorbidities (including borderline personality disorder, eating disorders within the past year, or a history of psychotic symptoms), and recent participation in other clinical trials. Complete inclusion and exclusion criteria are outlined in the trial protocol. Participants were recruited through the Stanford Depression Research Clinic database in REDCap, as well as through public announcements and external referrals within the Stanford Department of Psychiatry and Behavioral Sciences. Gender data were self-reported at baseline, with participants selecting from the options female, male, or other. Outpatients with MDD or bipolar depression who had SI were enrolled after providing written consent.
Participants were randomly assigned in a 1:1 ratio to receive either ketamine followed by buprenorphine or ketamine followed by placebo, using a computer random number generator. No stratification or minimization factors were applied. The assignments were kept by an unblinded staff member and the pharmacy, who did not have access to participants after enrollment. Participants, treating investigators, and outcome assessors were masked to treatment allocation. The buprenorphine and placebo troches had identical appearances.
All participants received a single open-label intravenous ketamine infusion (0.5 mg/kg over 40 minutes) administered in a clinical research unit (see Figurein the online supplement). Dissociative symptoms were assessed using the Clinician-Administered Dissociative States Scale (CADSS)at baseline, immediately after the infusion at 40 minutes, and at 80 and 120 minutes. Vital signs were monitored similarly. Forty-eight hours later (day 3), participants were started on low doses of buprenorphine or placebo for 4 weeks under double-blind conditions. Sublingual troches containing 0.2 mg of buprenorphine or placebo were produced by Mariner Pharmacy (San Mateo, California) and delivered to Stanford University's Depression Research Clinic on a per-patient, per-week basis. The troches were administered using a flexible dosing regimen. Dosages were increased by 0.2 mg/day each week, aiming to reach a minimum of 0.6 mg/day and a maximum of 0.8 mg/day, as tolerated and based on clinical response, following an adapted protocol from Yovell et al.. Participants received their study drug supply during weekly clinical follow-up visits and self-administered it at home, with instructions to allow the drug to fully dissolve sublingually over approximately 20 minutes before swallowing. We closely monitored the doses delivered to participants each week and counted the returned spare troches to assess adherence. At the end of 31 days, participants were monitored for an additional 2 weeks without receiving study drugs to evaluate discontinuation.
The primary outcome was the change in SSI total score, assessed weekly from day 1 to day 31 (end of week 4). The SSI is a 21-item clinician-rated scale, with each item offering three response options scored from 0 (not present) to 2 (maximum severity of SI). Items 1 through 5 serve as screening items. If a participant scores zero on both items 4 and 5-indicating no intent to make an active suicide attempt and a willingness to preserve their life in a life-threatening situation-the clinician is instructed to skip items 6 through 19 and proceed directly to items 20 and 21. These final two items assess the lifetime occurrence of suicide attempts and the intensity of the desire to die during the most recent attempt. They are collected for informational purposes only and are not included in the SSI total score. The total score is derived by summing items 1 through 19, yielding a possible range of 0-38, with higher scores indicating greater suicidal ideation. Secondary outcomes included changes on the Montgomery-Åsberg Depression Rating Scale (MADRS)and the 17-item Hamilton Depression Rating Scale (HAM-D), assessed weekly from day 1 to day 31. Safety outcomes included the weekly assessment of adverse events during the treatment phase. In addition, for days 38-45, clinical ratings were repeated, as was the evaluation of opioid withdrawal symptoms after the end of the 4-week treatment period, using the Clinical Opiate Withdrawal Scale (COWS)and the Subjective Opiate Withdrawal Scale (SOWS).
An a priori power analysis indicated that 60 subjects were needed to achieve 80% power to detect effect sizes of ≥0.35 on the primary outcome, defined as changes in SSI total scores from day 1 to 31. Forty-one months after study onset, the Data and Safety Monitoring Board (DSMB) conducted its review. Due to slow enrollment during the COVID-19 pandemic, only 36 participants had completed the study by that time. The DSMB reported no safety concerns and recommended continuation until 50 participants were enrolled. Analyses included a modified intention-to-treat population, defined as all randomized participants who received 1 week of treatment and completed the week 1 assessment. A linear mixed-effects model tested the primary hypothesis of efficacy, measured by change in SSI total scores. The model included random effects for patient and fixed effects for treatment group, time as a continuous variable (days of study, and the time-by-treatment group interaction, along with an unstructured covariance matrix. For the secondary outcome of depression, a similar linear model evaluated changes in MADRS depression scores. An alpha of 0.05 (two-tailed) was used to determine statistical significance. All statistical analyses were conducted using SPSS, version 30. Additional details are provided in the online supplement.
Between November 2020 and March 2025, 79 individuals were assessed for eligibility. Of these, 21 were excluded. Fifty-eight participants were enrolled, of whom eight did not proceed to ketamine infusion. Fifty participants were randomized prior to treatment to receive intravenous ketamine plus buprenorphine or intravenous ketamine plus placebo. Five participants (10%) discontinued the study after the ketamine infusion, including the single participant enrolled who had a diagnosis of bipolar II disorder. A total of 45 participants (23 in the buprenorphine group, 22 in the placebo group) completed at least 1 week of follow-on treatment and assessment and were included in the analysis of the primary outcome (Figure). Baseline demographic and clinical characteristics are summarized in Table. Of the 50 participants who received the ketamine infusion, 34 (68%) were female, and the mean age was 37.6 years (SD=11.2). Thirty-six participants (72%) identified as White, and 10 (20%) identified as Hispanic or Latino. Thirty-one participants (62%) met criteria for treatmentresistant depression, defined as two or more antidepressant treatment failures at an adequate dosage and duration. The study population exhibited a high burden of psychiatric comorbidities. The mean baseline MADRS total score was 34.1 (SD=4.6), the mean baseline SSI total score was 15.0 (SD=4.5), and the mean number of prior suicide attempts was 1.3 (SD=3.6). The sample also had substantial prior exposure to interventional psychiatric treatments: five participants (10%) had previously failed to benefit adequately from electroconvulsive therapy, 21 (42%) had failed to benefit adequately from repetitive transcranial magnetic stimulation, and eight (16%) had received prior ketamine treatment. At day 31, the mean final dosage of buprenorphine for the active group was 0.77 mg/day (SD=0.07). Participants demonstrated an overall average weekly medication adherence rate of 87%.
Fifty participants completed the ketamine phase of treatment. SSI total scores decreased significantly from day 1 to day 3 after ketamine infusion (mean=-9.1, SD=6.6; t=9.8, df=49, p<0.001). At day 3, 54% of participants met the criterion for response (i.e., ≥50% reduction). MADRS total scores also significantly decreased from day 1 to day 3 (mean=-13.4, SD=11.0; t=8.5, df=48, p<0.001), with 38% of participants meeting the criterion for response.
Suicidal ideation. From day 1 to day 31, SSI total scores significantly decreased in both groups (ketamine+buprenorphine: mean=-11.6, SD=5.8, p<0.001; ketamine+ placebo: mean=-6.3, SD=7.0, p<0.001), with a larger reduction observed in the ketamine+buprenorphine group (Glass delta=0.76, 95% CI=0.11, 1.39). The linear mixed-effects model indicated a statistically significant time-by-group interaction (p<0.001) due to greater reduction in SSI total scores among participants randomized to ketamine+buprenorphine (Figure). Group differences on mean SSI total scores were observed as early as day 10, and the largest group differences were observed at days 24 and 31. Inspection of the data from days 3 to 31 indicated that the ketamine+placebo group showed a worsening of SI, whereas the ketamine+buprenorphine group exhibited continued improvement (day 10: ketamine+buprenorphine, mean=6.2, SD=5.7; ketamine+placebo: mean=9.3, SD=8.1; Glass del-ta=0.38, 95% CI=-0.22, 0.98; day 17: ketamine+buprenorphine, mean=4.3, SD=4.9; ketamine+placebo, mean=8.7, SD=7.4; Glass delta=0.58, 95% CI=0.03, 1.2; day 24: ketamine+buprenorphine, mean=3.5, SD=4.7; ket-amine+placebo, mean=9.1, SD=7.4; Glass delta=0.75, 95% CI=0.12, 1.4); day 31: ketamine+buprenorphine, mean=3.6, SD=4.8; ketamine+placebo, mean=8.7, SD=7.8; Glass delta=0.65, 95% CI=0.02, 1.3). At day 31, 78% of participants receiving ketamine+buprenorphine met the response criterion (i.e., ≥50% reduction in SSI score), and 48% of participants receiving ketamine+placebo met the response criterion (χ 2 =4.5, df=1, p=0.035); the number needed to treat was 3. The results of the efficacy analyses did not change when including or excluding the five participants who discontinued the study after the ketamine infusion. Their clinical characteristics did not differ significantly from those of the larger sample, as shown in Table. For the intentionto-treat sample, mixed modeling showed a statistically significant change over time for both groups (both p values <0.001) and a statistically significant group-by-time interaction (p<0.05). Depression. From day 1 to day 31, statistically significant decreases in MADRS total scores were observed in both groups (ketamine+buprenorphine: mean=-15.8, SD=12.8, p<0.001; ketamine+placebo: mean=-8.5, SD=10.4, p<0.001). The difference between groups was not statistically significant (timeby-treatment interaction, p=0.104). MADRS total scores were not significantly lower in the ketamine+buprenorphine group. The largest group differences were observed at days 24 and 31 (Figure) (day 24: ketamine+buprenorphine, mean=19.2, SD=11.9; ketamine+placebo, mean=24.3, SD=9.9; Glass delta=0.51, 95% CI=-0.10, 1.1; day 31: ketamine+buprenorphine, mean=18.8, SD=11.4; ketamine+placebo, mean=24.6, SD=11.3; Glass delta=0.51, 95% CI=-0.11, 1.1). At day 31, 57% of participants receiving ketamine+buprenorphine met the response criterion (i.e., ≥50% reduction in MADRS), and 30% of participants receiving ketamine+placebo met the response criterion. Results from depression ratings on the HAM-D were similar to those observed on the MADRS; the group-by-time interaction was not significant (p=0.18) (Figure).
There were no unanticipated or serious adverse events with the ketamine infusion. CADSS mean scores at 40 minutes are shown in Figurein the online supplement. Side effects were observed in both the buprenorphine and placebo groups. The most frequently reported side effects by individual participants were dizziness ( buprenorphine group, 21.7%; placebo group, 0%), oral burning (buprenorphine group, 17.4%; placebo group, 0%), headache (buprenorphine group, 17.4%; placebo group, 22.7%), and nausea (buprenorphine group, 13.0%; placebo group, 0%) (Table).
Postintervention suicidal ideation and depression scores. After the main study phase (days 1-31), participants were followed for an additional 2 weeks (days 38 and 45) off medication or Åsberg Depression Rating Scale; MDD, major depressive disorder; MDE, major depressive episode; SNRI, serotonin-norepinephrine reuptake inhibitor; SSI, Scale for Suicide Ideation; SSRI, selective serotonin reuptake inhibitor; TMS, transcranial magnetic stimulation. b Participants were instructed not to take any benzodiazepines within 24 hours before the ketamine infusion and on the day of the infusion. c One participant in the placebo group reported 25 suicide attempts. Excluding this outlier, the mean number of suicide attempts was 1.0 (SD=1.3). placebo. Mean SSI and MADRS total scores increased at 1 and 2 weeks postintervention in the ketamine+buprenorphine group, but not in the ketamine+placebo group (Table). In the ket-amine+buprenorphine group, post hoc analyses indicated that the observed increase in SSI total scores from day 31 to day 38 was statistically significant (mean=2.5, SD=4.8; t=2.4, p=0.03). The increase in MADRS total scores from days 31 to 38 was also statistically significant (mean=6.1, SD=9.3; t=2.9, p=0.009). In contrast, in the ketamine+placebo group, there was a small but not statistically significant reduction, rather than an increase, in MADRS and SSI scores. There was no significant worsening in SSI and MADRS total scores during the second week of discontinuation (days 38-45). Overall, SSI and MADRS total scores were statistically significantly lower in both groups at days 38 (p<0.01) and 45 (p<0.01) than at day 1. Withdrawal. Clinician-rated withdrawal scores on the COWS obtained in the follow-up phase were in the "minimal to none" range in both the ketamine+buprenorphine group (mean=1.47, SD=1.9) and the ketamine+placebo group (mean=0.50, SD=0.61) (Table), and the difference was statistically significant (Welch-corrected p=0.044). Patient-rated withdrawal scores on the SOWS were higher than clinician ratings in both the ketamine+buprenorphine group (mean=8.0, SD=5.3) and the ketamine+placebo group (mean=7.2, SD=5.4). SOWS ratings were in the minimal range and were not statistically different between groups (Table). Evaluation of treatment blind. Two weeks postintervention (day 45), participants were asked to endorse one of the following: "I think I was taking active medication"; "I think I was taking placebo"; "I am unsure if I was taking an active medica-tion or a placebo." Of 40 participants (21 in the ketamine+buprenorphine group, 19 in the ketamine+placebo group), 20 (50%) guessed their treatment condition correctly. Fourteen participants (66%) in the ketamine+buprenorphine group correctly guessed that they received buprenorphine, and six (32%) correctly guessed that they had received placebo. Fourteen participants (five in the ketamine+buprenorphine group, nine in the ketamine+placebo group) endorsed being unsure of their treatment assignment, and six guessed that they had been randomized to the other treatment condition.
In our trial, buprenorphine added to treatment as usual was significantly more effective than placebo in further reducing SI following ketamine infusion. To our knowledge, this is the first RCT to demonstrate that a pharmacological intervention can maintain and augment ketamine's antisuicidal effects in MDD. Previous efforts to extend ketamine's therapeutic benefits using pharmacological agents such as lamotrigine, lithium, and riluzolefailed to show a statistically significant advantage over placebo. An ad hoc analysis evaluating d-cycloserine in individuals with treatment-resistant depression showed no benefit on overall HAM-D scores, but revealed a preliminary signal suggesting an advantage of d-cycloserine over placebo on a single suicidal ideation item of the HAM-D. An Automated Self-Association Training (ASAT) protocol was the first behavioral intervention shown to maintain the antidepressant effects of ketamine for up to 30 days in individuals with treatment-resistant depression. However, participants in that study did not present with SI. The durability seen in the present study suggests a novel therapeutic strategy to mitigate the devastating consequences of depression with SI. Our findings are consistent with those of Yovell et al., who reported a reduction in SI with low-dose buprenorphine compared to placebo. However, our study showed a greater overall decrease in SI, with participants reaching mean SI scores below the entry threshold at day 31. Notably, our participants received a higher final mean dosage of buprenorphine compared to those in the Yovell et al. study (0.77 mg/day vs. 0.44 mg/day), which may partially explain the larger reduction in SI total scores observed in our study (-11.6 versus -7.1). We also observed a substantially lower dropout rate of 12% (N=6) at week 4, compared with the 29% reported by Yovell et al. Differences in study populations may have contributed to these discrepancies, as the Yovell et al. sample included a higher proportion of individuals with borderline personality disorder. This diagnostic group was excluded from our trial, which might have impacted both treatment response and retention. Reduction in SI in our study was more rapid and greater overall in participants who received ketamine before buprenorphine compared to those in the Yovell et al. study who received only buprenorphine. At week 4, we observed a 76% reduction in SI in the ketamine+buprenorphine group, compared to a 43% reduction in the ketamine+placebo group. In contrast, Yovell et al. reported an approximate 47% reduction in the buprenorphine group and an 11% reduction in the placebo group over the same time period. The difference between our ketamine+placebo group and the placebo group in the Yovell et al. study suggests that ketamine alone may account for a sustained 32% reduction in SI 4 weeks after infusion. The sustained response to ketamine+placebo we observed is consistent with the prolonged antisuicidal response to ketamine added to treatment as usual, as reported by Grunebaum et al.. As shown in our results, ketamine produced a 60% reduction in SI within 48 hours. Together, these findings suggest that the sequential combination of ketamine followed by buprenorphine may be more effective than adjunctive buprenorphine therapy alone. In both the Yovell et al. study and ours, depression was a secondary endpoint, and the observed reduction was not significantly different between groups. Other RCTs of low-dose buprenorphine have failed to demonstrate significant antidepressant effects in treatment-resistant depression. Prior data have suggested that ketamine's antisuicidal effects may be partially independent of its antidepressant effects, and our findings are consistent with these observations. One possible explanation is that ketamine exerts rapid effects on specific symptoms that relate particularly to SI, such as anhedonia, hopelessness, rumination, and impaired decision making. Such symptoms may reflect alterations in glutamatergic neurotransmission, dopamine turnover, opioid activity, and associated neurocircuitry in relevant brain regions. There is continuing debate over whether ketamine's antidepressant properties are primarily through glutamatergic synaptic plasticity or engagement of opioid receptor pathways. A recent randomized crossover study using functional magnetic resonance spectroscopy demonstrated that pretreatment with the opioid antagonist naltrexone attenuated ketamine-induced glutamatergic activation and its subsequent antidepressant effects, providing direct evidence that opioid system activation may be mechanistically necessary for ketamine's therapeutic action. This trial has several strengths that enhance its scientific and clinical relevance. The individualized titration approach-adjusting buprenorphine dosage based on tolerability and efficacy-supports its practical application. The low dosages required and the favorable side effect profile of buprenorphine further support its potential as a treatment option, even when used alongside other antidepressant medications. Inclusion of 2-week posttreatment safety monitoring allowed us to assess treatment beyond 4 weeks and potential withdrawal symptoms, important features not evaluated previously. The mild and transient worsening of depression and SI after discontinuation point to the need to follow patients after stopping but are consistent with the drug conferring benefit. After 4 weeks of buprenorphine, participants did not demonstrate typical physical opioid withdrawal. The design and findings from our trial have significant public health implications. All participants received one open-label intravenous ketamine infusion, ensuring that individuals at acute suicide risk were provided access to a potentially life-saving treatment. We allowed for continued antidepressant treatment throughout, which likely contributed to high participant retention and adherence. Last, the agents tested are approved and available for use in the United States and other countries. Careful screening for a history of substance use disorder is essential prior to the initiation of treatment. In this study, individuals with a history of substance use disorder were excluded, and all participants underwent urine drug screening before enrollment. As a result, our findings cannot be extended to individuals with comorbid substance use disorders. We aimed to replicate and extend Yovell and colleagues' 4-week intervention study, which was conducted in a more acutely ill population, and similarly limited our trial duration to 4 weeks. However, unlike Yovell et al., we formally followed participants for an additional 2 weeks after the buprenorphine phase to monitor for withdrawal symptoms and potential worsening of SI. Further research is needed to determine the maximum duration of buprenorphine treatment that can effectively reduce SI while minimizing the risks of tolerance or dependence. All participants were required to have a primary diagnosis of MDD with documented resistance to at least one antidepressant treatment or intolerance to two or more antidepressants. Several common psychiatric disorders were permitted, such as generalized anxiety disorder and posttraumatic stress disorder, but not borderline personality disorder. Yovell et al.primarily included individuals with borderline personality disorder, although they observed efficacy in participants both with and without the diagnosis. While their trial was conducted in an inpatient setting, ours was exclusively outpatient. Together, these findings suggest that buprenorphine may have generalizable efficacy in reducing SI across diverse clinical populations and treatment settings. There were limitations to this study. We did not power the trial to evaluate whether response to ketamine predicts response to buprenorphine. In addition, we did not assess treatment expectancy, an important psychological factor known to influence outcomes in clinical trials. Future studies should incorporate expectancy measures to clarify their role in modulating response. We queried participants regarding their drug assignment after discontinuation. A worsening of symptoms after stopping may have led them to infer they had been on active drug. The intensity of SI in our sample was somewhat lower than that reported by Yovell et al., but comparable to that observed in other studies. Our relatively small sample size warrants caution in interpreting the findings. Additional RCTs are needed to confirm the robust effects of buprenorphine follow-on treatment for extending the antisuicidal effects of ketamine, and larger-scale studies would help determine whether buprenorphine can produce significant antidepressant effects. Long-term studies could also compare the ketamine plus buprenorphine combination with repeated ketamine infusions as well as studying participants who are not on antidepressant treatment as usual. A future study could include a placebo-controlled ketamine infusion arm, to better isolate the ketamine effect. Last, further studies are needed to determine whether longer treatment durations and/or tapering are required in individuals to avoid transient worsening upon discontinuation. In summary, this RCT provides the first evidence that lowdose buprenorphine can effectively extend the antisuicidal effects of a single intravenous ketamine infusion over 4 weeks in individuals with MDD. These findings offer a potentially scalable and safe therapeutic option for a population at risk of suicide.
Create a free account to open full-text PDFs.
Newport, D. J., Carpenter, L. L., Mcdonald, W. M. et al. · American Journal of Psychiatry (2015)
Williams, N. R., Heifets, B. D., Bentzley, B. S. et al. · Molecular Psychiatry (2019)
Price, R. B., Spotts, C., Panny, B. et al. · American Journal of Psychiatry (2022)
Wilkinson, S. T., Ballard, E. D., Bloch, M. H. et al. · American Journal of Psychiatry (2017)