Rapid and sustained decreases in suicidality following a single dose of ayahuasca among individuals with recurrent major depressive disorder: results from an open-label trial

Suicide is a leading cause of death worldwide and often occurs in the context of major depressive disorder (MDD). Earlier research indicates that standard first-line treatments for MDD, such as selective serotonin reuptake inhibitors (SSRIs) and cognitive behavioural therapy (CBT), produce limited or delayed reductions in suicidality, and some interventions may even worsen suicidal thoughts in specific subgroups. Rapid-acting pharmacological options such as ketamine have shown promise for anti‑suicidal effects but raise questions about long-term efficacy and safety, leaving a need for additional rapid-acting interventions that target both depressive symptoms and suicidality. Ayahuasca, a traditional psychoactive brew containing DMT and β-carbolines (e.g. harmine), has been reported in preliminary studies to reduce depressive symptoms and may have potential to reduce suicidality when administered in a supportive context. The present study therefore examined whether a single dose of ayahuasca produces rapid (acute) and sustained (post-acute) reductions in suicidality among individuals with recurrent MDD. The authors conducted a secondary analysis of an open‑label clinical trial to test two hypotheses: that suicidality would decrease acutely after ayahuasca administration, and that decreases would be sustained up to 21 days post‑treatment. A single-dose, inpatient setting with repeated suicidality assessments allowed investigation of both immediate and short‑term outcomes.

Zeifman and colleagues recruited participants via local advertisements and referrals from private psychiatric clinics. The trial was conducted on an inpatient psychiatric unit in Brazil. All participants underwent a 2‑week medication washout in which they refrained from psychiatric medications and recreational substances prior to ayahuasca administration. The extracted text does not clearly report detailed inclusion or exclusion criteria beyond recruitment sources and the washout requirement. Drug administration occurred individually in a quiet, dimly lit room with participants seated in a reclining chair. Each participant received a single oral dose of ayahuasca at 2.2 ml/kg. The brew, analysed by GC/MS, contained 0.8 mg/ml DMT and 0.21 mg/ml harmine, with no harmaline detected; tetrahydroharmine content was not analysed. Sessions lasted about 4 hours; no formal psychological preparation or in‑session psychotherapeutic intervention was provided. Participants were observed and discharged if no complications were apparent within 24 hours. Suicidality was measured using item 10 of the clinician‑administered Montgomery–Åsberg Depression Rating Scale (MADRS‑SI), which rates current suicidality from 0 to 6. The sum of the remaining nine MADRS items (MADRS‑total nonSI) was used to index non‑suicide depressive symptoms. Assessments were conducted 10 minutes before administration (baseline), at 40, 80, 140 and 180 minutes during the acute session, and at 1, 7, 14 and 21 days post‑intervention. The primary outcome was change in MADRS‑SI. Individuals with MADRS‑SI = 0 at baseline were excluded from the main analyses and followed for any emergence of suicidality. The statistical approach used two fixed‑effects linear mixed models with unstructured covariance matrices: one model tested acute changes using baseline and the four intra‑session time points, the other tested post‑acute changes using baseline and the four follow‑up days. Missing data were handled in the mixed models using restricted maximum‑likelihood estimation; for effect sizes and correlation analyses the authors used last observation carried forward (LOCF). Within‑subject paired t tests compared baseline with each subsequent time point. Hedges' g (within‑subject) was computed for effect sizes and Pearson correlations examined associations between changes in MADRS‑SI and changes in MADRS‑total nonSI. Analyses were conducted in IBM SPSS Statistics Version 25 with alpha set at 0.05, two‑tailed.

The full trial sample comprised 17 individuals. Two participants had no suicidality at baseline (MADRS‑SI = 0) and were therefore excluded from the primary analyses; of these two, one showed a transient increase to MADRS‑SI = 1 at day 1 and then returned to 0 at subsequent time points. The analytic sample for the suicidality analyses therefore included 15 participants, whose baseline mean MADRS‑SI was 2.40 (SD = 1.35) as reported in the extracted text. Acute effects: The linear mixed model for the intra‑session (acute) phase showed a significant main effect of time (F(4,13.85) = 7.17, p = 0.002). Within‑subject t tests indicated significant decreases in MADRS‑SI between baseline and each acute time point (40, 80, 140 and 180 minutes post‑administration). Effect sizes increased over the session, ranging from medium at 40 minutes to large at 80, 140 and 180 minutes, with the largest acute effect size at 180 minutes (Hedges' g = 0.89). At 40 minutes post‑administration there was a significant correlation between change in suicidality and change in non‑suicide depressive symptoms (MADRS‑SI vs MADRS‑total nonSI), r = 0.69, p = 0.005. Post‑acute effects: The post‑acute linear mixed model likewise showed a significant effect of time (F(4,10.69) = 15.65, p < 0.001). Paired comparisons demonstrated significant reductions in MADRS‑SI at 1, 7, 14 and 21 days compared with baseline. Effect sizes for these post‑acute decreases were large at all assessed days, with the largest observed at day 21 (Hedges' g = 1.75). A significant association between change in suicidality and change in non‑suicide depressive symptoms was also observed at day 21, r = 0.52, p = 0.049. The extracted text does not provide a detailed adverse‑event table; it reports that participants were discharged if no complications were observed within 24 hours but does not catalogue specific adverse events in this extract.

The investigators report that a single dose of ayahuasca was associated with rapid acute reductions in clinician‑rated suicidality during the session and sustained decreases persisting through 21 days post‑administration. Effect sizes ranged from medium at the earliest time point to large thereafter, with the largest effect at 21 days (g = 1.75). The authors highlight that these sustained reductions are notable because some other rapid‑acting interventions (for example, ketamine) may require repeated dosing to maintain effects, which raises cost and safety considerations. The study team also observed significant correlations between reductions in suicidality and reductions in non‑suicide depressive symptoms both acutely and at 21 days, suggesting related changes across symptom domains. Regarding mechanisms, the authors discuss biological hypotheses tied to ayahuasca's constituents: harmine has shown antidepressant and neurotrophic effects in preclinical work (including increases in BDNF and neurogenesis), and DMT's action at cortical 5‑HT2A receptors may increase glutamatergic tone and neuroplasticity. Human neuroimaging studies are cited as indicating modulation of networks implicated in introspection and emotion, such as the default mode network. The authors also note anti‑inflammatory effects of psychedelics as a putative pathway worthy of investigation given links between inflammation and suicidality. Several limitations are acknowledged. The sample was small and the design open‑label, so placebo effects or natural fluctuations in suicidality cannot be excluded. Follow‑up was limited to 21 days, leaving longer‑term durability unknown. Suicidality was measured using a single MADRS item rather than a comprehensive suicidality instrument, and the original trial did not specifically recruit individuals with high suicidality. The authors recommend randomized, controlled trials with larger samples that recruit participants with suicidality as a primary focus and use more extensive suicidality assessments. They further suggest that providing greater psychological support or integration alongside pharmacological administration may be important for safety and maximising therapeutic benefits.