The Impact of Ayahuasca on Suicidality: Results From a Randomized Controlled Trial.

Suicide is described as a major and growing public health problem, particularly among people with major depressive disorder (MDD) and those with comorbid MDD and borderline personality disorder (BPD). The authors note limitations of existing interventions for suicidality — including delayed onset of effect, limited availability, adverse effects, the need for ongoing administration, and substantial non-response among some patients — and highlight the unmet need for fast-acting, durable, and safe novel treatments. Previous experimental work with ketamine shows rapid reductions in suicidality but concerns remain regarding durability and abuse potential; observational and some open-label studies of psychedelics (including psilocybin and ayahuasca) suggest associations with lower suicidality and improvements in related constructs such as depression and hopelessness, but experimental placebo-controlled data on suicidality are lacking. Jha and colleagues therefore report secondary analyses of a double-blind, randomised, placebo-controlled trial in individuals with treatment-resistant unipolar MDD, testing whether a single dose of ayahuasca reduces clinician-assessed suicidality. The primary hypothesis was that ayahuasca would produce reductions in suicidality that were sustained from 1 to 7 days post-administration. The investigators also planned exploratory analyses of the relationship between changes in suicidality and changes in non-suicide-related depressive symptoms.

This report presents secondary analyses of a double-blind, parallel-arm, randomised placebo-controlled trial conducted at a university hospital in Brazil. Adults aged 18–60 with unipolar MDD and treatment resistance (defined as inadequate response to two or more antidepressants from different classes) were eligible. Key exclusions included prior psychedelic experience, current medical disease, pregnancy, imminent suicide risk, substance abuse, neurological disorder, and personal or family history of schizophrenia or bipolar disorder. Participants were recruited by referral and advertisement, screened by a psychiatrist using the Portuguese SCID-IV, and randomised 1:1 in blocks of 10. Investigators and participants were blind to allocation; blinding was augmented by excluding participants with prior psychedelic use, using an active placebo that mimicked taste and mild gastrointestinal effects, and reassigning assessors after the intervention. Antidepressants were discontinued prior to the intervention (on average about 2 weeks, depending on half-life) and for 7 days afterwards; daily benzodiazepine use was permitted except during the acute phase. The intervention took place in a quiet, dimly lit room with a standardised music playlist. Participants received a single 1 ml/kg dose of ayahuasca (analysed concentrations per ml: N,N-DMT 0.36 ± 0.01 mg, harmine 1.86 ± 0.11 mg, harmaline 0.24 ± 0.03 mg, tetrahydroharmine 1.20 ± 0.05 mg) or placebo (per ml: 0.1 g yeast, 0.02 g zinc sulfate, 0.02 g citric acid). Two investigators remained nearby to provide support; sessions lasted around 8 hours and most participants returned home the same day, with four electing inpatient stay for the 7-day follow-up period. Suicidality was assessed by a clinician using the suicidality item of the Montgomery–Åsberg Depression Rating Scale (MADRS-SI, item 10; range 0–6) at baseline, and at 1, 2, and 7 days after dosing. The remaining nine MADRS items were summed to create a measure of non-suicide-related depressive symptoms (MADRS-total nonSI). Analyses used a modified intention-to-treat set including all participants who received the intervention. A fixed-effects linear mixed model examined MADRS-SI at days 1, 2, and 7 with baseline MADRS-SI entered as a covariate; an unstructured covariance was specified and restricted maximum-likelihood estimation was used to account for two missed follow-up visits. The model tested main effects of time and intervention and their interaction. Between-group and within-group Cohen’s d effect sizes were calculated for MADRS-SI and MADRS-total nonSI, and Pearson correlations were used to explore the relationship between change scores at day 7. Significance was set at p < 0.05 (two-tailed).

The analytic sample was small and clinically severe: participant mean age was 42.04 (SD = 11.66), 72% were female, 59% Caucasian, 52% unemployed, 55% reported a lifetime suicide attempt, and 76% met criteria for a comorbid DSM-IV Axis II disorder; nine individuals (31%) had BPD. Baseline suicidality on the MADRS-SI averaged 2.35 (SD = 1.91). Group sizes reported in outcome analyses indicate 14 participants in the ayahuasca group and 15 in the placebo group. In the linear mixed model, there was a significant main effect of time on MADRS-SI (F(2,25.72) = 3.38; p < .05), indicating suicidality decreased across the sample over the week following the intervention. The main effect of intervention (ayahuasca versus placebo) trended toward significance (p = .088), but did not reach the conventional threshold. The time × intervention interaction was not significant (p = .678). Missing follow-up data affected two assessments and were handled via restricted maximum-likelihood estimation. Effect-size estimates provided further detail. Between-group effect sizes (ayahuasca versus placebo) for reductions in MADRS-SI were medium at each time point: day 1 Cohen’s d = 0.58 (95% CI -1.32 to 0.17), day 2 d = 0.56 (95% CI -1.30 to 0.18), and day 7 d = 0.67 (95% CI -1.42 to 0.08). Within the ayahuasca group there were large within-group effect sizes for MADRS-SI decreases: day 1 d = 1.33 (95% CI 1.25 to 3.18; n = 14), day 2 d = 1.42 (95% CI 1.50 to 3.74; n = 13), and day 7 d = 1.19 (95% CI 1.21 to 3.50; n = 14). The placebo group showed small within-group effects on days 1 and 7 (day 1 d = 0.00, 95% CI -1.09 to 1.63; n = 13; day 7 d = 0.04, 95% CI -0.90 to 1.04; n = 15) and a medium effect at day 2 (d = 0.64, 95% CI 0.06 to 1.23; n = 14). At day 7, the correlation between change in suicidality (MADRS-SI) and change in non‑suicide-related depressive symptoms (MADRS-total nonSI) within the ayahuasca group approached significance (r = .53, p = .053), while the same association in the placebo group was non-significant (r = .16, p = .579). In the small subgroup of participants with BPD who received ayahuasca (n = 5), all five had clinically significant suicidality at baseline (MADRS-SI ≥ 4); none met that threshold at days 1 or 2, and one (20%) did at day 7. The investigators report no serious adverse events among participants with BPD. Overall, the investigators interpret these results as mixed: a temporal decline in suicidality was observed, and effect-size estimates support potential benefit for ayahuasca, but the primary test of treatment effect did not reach statistical significance, likely reflecting limited power.

Jha and colleagues interpret their findings cautiously. They emphasise that this is the first placebo-controlled trial to report clinician-assessed suicidality outcomes after ayahuasca administration and that suicidality decreased across the sample over the week following dosing. Although the intervention effect only trended toward significance, the medium between-group and large within-group effect sizes for decreases in MADRS-SI are taken as suggestive that ayahuasca may have a fast-acting antisuicidal effect that can be observed as early as 1 day post-administration and that may persist to at least 7 days. The authors situate their results relative to prior research: rapid antisuicidal effects are well documented for ketamine, and observational and open-label studies of psychedelics have linked lifetime or administered use to reduced suicidality and improvements in depression and emotion dysregulation. Neurobiological and psychological mechanisms proposed to account for ayahuasca’s effects include reductions in emotion dysregulation, increases in mindfulness-related capacities (e.g., acceptance and decentering), enhanced blood flow in emotion-regulation brain regions, and promotion of neuroplasticity in prefrontal circuits. The day‑7 association (r = .53, p = .053) between reductions in suicidality and reductions in non‑suicide-related depressive symptoms within the ayahuasca group is presented as preliminary evidence that effects on suicidality may be at least partly mediated by improvements in depressive symptoms or overlapping mechanisms. Key limitations are acknowledged. The trial had a small sample size, reducing statistical power and precluding inclusion of potentially important covariates (for example, benzodiazepine use). Baseline suicidality was low in the placebo group, raising the possibility that regression to the mean or placebo effects influenced results despite statistical control for baseline scores. For safety reasons people with imminent suicide risk were excluded, so generalisability to highly suicidal populations is unknown. The psychological intensity of psychedelic experiences complicates blinding, and while an active placebo and other measures were used, successful masking cannot be guaranteed. The investigators also note the absence of qualitative data or formal mediator analyses in this secondary report. Finally, the authors argue that their findings justify further research: larger, better‑powered randomised controlled trials with longer follow-up, inclusion of participants with higher baseline suicidality, evaluation of mechanisms and mediators (including qualitative work), refinement of placebo controls to improve blinding, and exploration of safety and efficacy in people with BPD are all recommended. They highlight ayahuasca’s low abuse potential as a relative advantage compared with agents such as ketamine, but stress that longer-term data are required to determine durability and safety in more severely suicidal populations.