A placebo-controlled study of the effects of ayahuasca, set and setting on mental health of participants in ayahuasca group retreats

Ayahuasca is a traditional Amazonian brew prepared from Psychotria viridis (containing the serotonergic 5-HT2A agonist N,N-dimethyltryptamine, DMT) and Banisteriopsis caapi (containing β-carboline MAO inhibitors such as harmine and harmaline). Its use has spread into Western contexts where people attend ayahuasca retreats seeking spiritual, self-developmental, or therapeutic outcomes. Observational field studies have reported acute cognitive and affective benefits and improvements in well-being after ayahuasca use, and clinical trials have reported rapid antidepressant effects in treatment-resistant depression, but many naturalistic studies lack placebo control and therefore cannot rule out expectation or setting effects. Uthaug and colleagues set out to disentangle pharmacological effects of ayahuasca from non-pharmacological influences of set and setting in a naturalistic, placebo-controlled observational study conducted within group retreats. The primary objective was to compare psychological and empathic outcomes before and after ceremonies in participants who received either freeze-dried ayahuasca or a placebo, testing the hypothesis that set and setting would affect both groups while pharmacological effects would be evident only in the ayahuasca group.

The investigators conducted a naturalistic, single-blind, placebo-controlled study at six ayahuasca retreats run by a single organisation across the Netherlands, Spain, and Germany. Invitations to participate were limited to ‘‘students of an ayahuasca school’’ associated with the host organisation; these individuals were typically experienced ayahuasca users who train as facilitators. Inclusion criteria recorded in the extract were fluency in English, age over 18, and written informed consent. The research team observed the ceremonies but did not manage their organisation, preparation, or drug administration. Study participants were randomly assigned by the host organisation to receive either freeze-dried ayahuasca capsules (N = 14) or placebo capsules (N = 16); at one site a liquid drink (ayahuasca or placebo drink) was used because capsules were unavailable. Participants received seven capsules with the option of up to three additional ‘‘booster’’ capsules about 2 hours after the first dose. Placebo capsules contained inert ingredients (cocoa powder, vitamins, turmeric, quinoa, traces of coffee, potato flour) intended to mimic appearance and taste; no detailed production or storage information was available in the extract. Alkaloid concentrations of three ayahuasca capsules were measured post hoc using HPLC-TOF MS calibrated with DMT, harmine, and harmaline reference standards, but the extract does not reproduce the capsule concentration table. Blinding procedures entailed that both study participants and the facilitators administering the substances were blind to assignment; the extract describes debriefing of investigators, participants and facilitators the next day. Baseline assessments were performed shortly before the ceremony and follow-up assessments the morning after; the research battery took around 30 minutes at each time point. The test battery included the Multifaceted Empathy Test (MET), the Depression, Anxiety and Stress Scale-21 (DASS-21), the Brief Symptom Inventory-18 (BSI-18), and the Five Facets Mindfulness Questionnaire (FFMQ-15). The Ego Dissolution Inventory (EDI) and the 5-Dimensional Altered States of Consciousness Rating Scale (5D-ASC) were collected only post-session to retrospectively characterise the psychedelic experience. Statistical analysis used mixed-model ANOVA with within-subject factor time (baseline, post-session), between-subject factor treatment (ayahuasca, placebo), and their interaction. EDI and 5D-ASC ratings were analysed by ANOVA with ayahuasca experience as a covariate; Pearson correlations examined associations between experience ratings and change scores. The significance threshold was set at p = 0.05.

Thirty participants completed the study (12 males, 18 females; mean age 40.18 years, SD 10.10). Most were European (N = 28) and highly experienced with ayahuasca (mean prior uses 23.7, SD 15.58); many reported prior use of other psychoactive substances and most had some contemplative practice (e.g., meditation). Educational levels varied from high school to PhD. Common motivations for participation included self-understanding and problem resolution. The extract does not report attrition or adverse-event data. Blinding checks indicated that correct treatment guesses by study participants were 57.1% in the ayahuasca group and 68.7% in the placebo group; facilitators guessed correctly in 38.5% and 87.5% of ayahuasca and placebo cases, respectively. Chi-square tests showed no significant difference in correct versus incorrect guesses for facilitators (χ2(1) = 3.57, p = 0.06) or participants (χ2(1) = 2.13, p = 0.14). There was no significant difference between facilitator and participant correct-guess rates (t27 = 0.81, p = 0.42), but facilitator and participant guesses were correlated (r = 0.54, p = 0.003). Mixed-model ANOVA identified significant main effects of time (baseline versus post-session) on stress (DASS-21; F1,26 = 8.27, p = 0.008, partial η2 = 0.24), depression (DASS-21; F1,26 = 6.53, p = 0.017, partial η2 = 0.20), and anxiety (BSI-18; F1,26 = 5.12, p = 0.032, partial η2 = 0.24). Mean change scores (95% CI) reported were −5.6 (−9.8 to −0.15) for stress, −4.9 (−8.9 to −0.73) for depression, and −2.1 (−2.4 to −0.22) for anxiety, indicating reductions after the ceremony across both treatment groups. An interaction between treatment and time for DASS-21 depression scores approached significance (F1,26 = 4.11, p = 0.053, partial η2 = 0.14), suggesting a trend toward larger depression reductions in the placebo group. A significant treatment × time interaction occurred for implicit arousal to negative stimuli on the MET (F1,16 = 5.11, p = 0.038, partial η2 = 0.20), indicating that ayahuasca increased emotional arousal to negative emotional stimuli relative to placebo. No FFMQ mindfulness facets changed with time or treatment. Mean EDI scores did not differ between groups: ayahuasca mean 32.39% (SD 23.50), placebo mean 30.66% (SD 27.54). Mean 5D-ASC dimension and subscale scores were generally low in both groups (roughly 6–27% across dimensions in placebo and 10–27% in ayahuasca); group differences on these retrospective measures were not significant when ayahuasca experience was included as a covariate. When experience was removed as a covariate, some 5D-ASC subscales approached significance and insightfulness reached significance (F1,25 = 5.86, p = 0.023) favouring higher ratings in the ayahuasca group. There was no relationship between number of prior ayahuasca uses and ratings of the psychedelic experience. Correlational analyses across both groups initially found no significant associations between change scores (depression, stress, anxiety, emotional empathy) and EDI or 5D-ASC ratings. After removing a single outlier, significant negative associations emerged between change in depression and ratings of anxious ego dissolution (r = −0.59, p = 0.001) and between change in stress and anxious ego dissolution (r = −0.42; the extract does not provide a p-value for this latter correlation). These results suggest that larger acute anxious-dissolution-type experiences were linked to greater symptom reductions in the sample examined.

Uthaug and colleagues interpret their primary finding as evidence that mental health improvements observed after participation in naturalistic ayahuasca ceremonies—reductions in self-reported stress, depression, and anxiety—occurred irrespective of whether participants received ayahuasca or placebo, implicating strong non-pharmacological influences such as set and setting. The authors note that group dynamics, ritual context, explicit intentions and expectations, and facilitator-led suggestions can all potentiate placebo responses and may have shaped outcomes in both conditions. They highlight that participants were experienced ayahuasca users with longstanding expectations about benefits, which the investigators argue likely amplified expectancy-driven improvements. At the same time, the study detected a pharmacological signal: ayahuasca increased implicit emotional arousal to negative stimuli on the MET, a putatively less suggestion-sensitive behavioural measure. The authors propose that this empathic change may be clinically relevant because low empathy is implicated in stress-related psychopathology, and increasing empathy could be therapeutic in mood and personality disorders. They also discuss an observed association—across groups—between the intensity of altered-state ratings on the 5D-ASC and magnitude of symptom reduction, suggesting that the strength of the subjective experience predicts benefit regardless of whether that experience was pharmacologically or psychologically induced. Several limitations acknowledged by the investigators temper conclusions. The sample comprised highly experienced ayahuasca users, which may have biased expectancy and limited sensitivity to pharmacological effects; doses were not adjusted for body weight and, based on later comparison with published dosing regimens, the DMT content per capsule likely produced lower psychedelic intensity than in some clinical trials. The research team had no control over ceremony organisation, dose preparation, administration, or storage, and full capsule composition and stability data were not available in the extract. Blinding was imperfect but not statistically distinguishable between correct and incorrect guesses. The authors note that the current design cannot separate whether ayahuasca modulates sensitivity to set and setting or whether set and setting alone produce the observed benefits; they suggest a factorial 2 × 2 design (drug/placebo × set-and-setting/no-set-and-setting) would be required to do so. Finally, the relatively low 5D-ASC and EDI scores indicate that the mean psychedelic intensity was modest, so differential pharmacological effects might have been more apparent at higher doses. In concluding remarks within the Discussion, the investigators state that both non-pharmacological (placebo-like) and pharmacological factors contributed to changes in participants’ mental state following the retreats. They recommend further controlled studies that systematically probe expectancy, set and setting elements, dose, and user experience (including novice samples) to clarify mechanisms underlying therapeutic effects attributed to ayahuasca and other psychedelics.