Adverse effects of ayahuasca: Results from the Global Ayahuasca Survey

Ayahuasca is a traditional Amazonian plant decoction containing harmine, harmaline, tetrahydroharmine (monoamine oxidase inhibitors) typically combined with DMT-containing plants such as Psychotria viridis. Its ritual use has spread internationally in recent decades, often retaining features of traditional practice such as facilitator‑administered group ceremonies. Biomedical interest has increased because observational studies and clinical trials report potential benefits for affective disorders, substance dependence, and other psychological outcomes. However, existing literature has tended to emphasise positive outcomes and comprises relatively small, selective samples, leaving the rates, severity, persistence and contextual determinants of adverse effects underrepresented and uncertain. Bouso and colleagues designed the present study to address these gaps by using data from a large, multi‑country online survey to quantify the frequency of acute physical and short‑to‑medium term mental health adverse effects associated with ayahuasca, and to examine how those adverse effects relate to history of ayahuasca use, clinical and sociodemographic variables, context of consumption, and the intensity of the acute spiritual experience. The aim was to provide a broader empirical picture of ayahuasca’s risk profile across diverse settings and user groups.

The study used data from the Global Ayahuasca Project, an online multilingual survey (English, Portuguese, Spanish, German, Italian, Czech) collected from 2017 to 2019. Eligible participants were adults (≥18 years) from more than 50 countries who had used ayahuasca at least once and provided informed consent. Recruitment was non‑random and relied on outreach through relevant organisations, retreat centres, churches, online groups, social media and conference flyers; no financial incentives were offered. Duplicate responses were removed and partially completed surveys retained. Ethical approval was obtained from the University of Melbourne Human Research Ethics Committee. For the present analyses, only respondents without missing data on the adverse effects items were included, yielding analytic samples of 8,216 for acute physical adverse effects and 7,839 for adverse mental health effects. The survey collected demographics, lifetime mental health diagnoses, detailed ayahuasca use history (frequency, lifetime uses, context), and used a modified nine‑item Short Index of Mystical Orientation (SIMO) to quantify the intensity of the acute spiritual experience. Acute physical adverse effects were captured by a checkbox list of 10 specified symptoms plus an “other” option, with a follow‑up query about whether medical attention was required. Short‑to‑medium term mental health, emotional or perceptual changes in the weeks or months after ceremonies were measured using a modified PHQ‑4 four‑point scale expanded with six items derived from the ayahuasca literature; respondents reporting increases were asked whether they sought professional support and whether they considered the experience part of a positive growth/integration process. Analytically, the investigators first used principal component analysis (PCA) to reduce heterogeneity in the adverse effect items: a varimax rotation was applied to physical effects and a promax rotation to mental health effects. Identified factor scores and binary indicators of any adverse physical or mental health effect were then modelled using logistic regression to assess associations with history of ayahuasca use (age of initiation, doses/year, lifetime and last‑year use), clinical variables (self‑reported anxiety, depression, substance and alcohol use disorders, number of physical health conditions), SIMO score, and context of consumption (religious, traditional shamanic, non‑traditional supervised, non‑supervised). Sociodemographic covariates were included and only significant controls are reported in result tables. Highly skewed variables (doses/year, lifetime use, last year use) were log‑transformed (natural log) prior to modelling. Statistical significance was set at p < .05.

Sample characteristics and preliminary factor structure. The analytic samples comprised 8,216 respondents for physical adverse effect analyses and 7,839 for mental health adverse effect analyses; just over 46% of respondents were female, mean age was >40 years, the majority held a university degree, were commonly married, and many respondents were from Brazil. PCA of physical adverse items yielded three factors (explained variance 21.4%, 10.5%, 10.5%): a general symptom factor (abdominal pain, vomiting/nausea, breathing difficulties, chest pains, headache), an arthromyalgical factor (stiff/swollen joints, aching muscles, coughing/wheezing), and a neurological factor (fits/seizures, fainting). PCA of mental health items produced two factors (explained variance 44.5% and 14.3%): an emotional‑cognitive factor (anxiety, worry, anhedonia, disconnection, reality‑testing difficulties, nightmares/disturbed thoughts) and a psychotomimetic factor (auditory/visual experiences, visual distortions, feeling attacked by spirit world). Frequency and characteristics of adverse effects. Acute physical effects on at least one occasion were reported by 5,742 participants (69.9%). The general symptom cluster was the most frequent (68.2%), with vomiting/nausea noted as the predominant physical effect; other individual physical symptoms (for example headache) occurred at lower rates (17.8% or less). A minority (2.3%) reported requiring medical attention for physical adverse effects. Neurological‑type physical events (fainting 4.1%, fits/seizures 1.3%) were infrequent but present in roughly 5% overall. Adverse mental health effects on at least one occasion were reported by 4,341 participants (55.4%). The emotional‑cognitive factor was reported by 3,293 participants (42.0%) and altered perception/psychotomimetic experiences by 3,004 participants (38.3%). The single most commonly endorsed mental item was “hearing or seeing things that others do not hear or see” (2,236; 28.5%), followed by “feeling disconnected or alone” (1,650; 21.0%). Severe responses (the top category on the 4‑point scale) were uncommon; the highest severe frequency was 4.4% for visual distortions (n = 342). For a majority of reported adverse mental health effects the duration was less than one week. The extracted text reports 11.9% of participants indicated needing professional support for adverse mental health effects, and among those reporting adverse mental health effects 87.6% judged them to be completely (76.3%) or somewhat (11.3%) part of a positive process of growth or integration. Associations from logistic regression. Acute physical adverse effects were positively associated with greater lifetime ayahuasca use, substance use disorder, number of physical health conditions, and non‑supervised context of use (all p ≤ .05, compared with religious context). Age of first use and last year use also increased risk to a lesser extent; doses/year was negatively associated (p = .05). Female sex and higher education were positively associated with reporting physical adverse effects, while older age at survey was negatively associated. Examining physical effect factors, general symptoms were linked to older age at first use, greater lifetime and last‑year use, physical health conditions and non‑supervised context, with fewer doses/year increasing risk. Arthromyalgical effects were more likely among respondents with physical health conditions, an anxiety disorder, higher lifetime use and older age at initiation; alcohol use disorder was negatively associated. Acute spiritual experience modestly increased risk for arthromyalgical effects and all non‑religious contexts raised arthromyalgical risk. Neurological effects were associated with number of lifetime uses and physical health conditions and showed a small association with the acute spiritual experience; they were unrelated to context of use. For adverse mental health effects, having an anxiety disorder, a physical health condition, and a stronger acute spiritual experience increased risk, as did all non‑religious contexts (traditional shamanic, non‑traditional supervised, and non‑supervised). Emotional‑cognitive effects were significantly associated with anxiety disorder, acute spiritual experience, and non‑religious consumption; psychotomimetic effects were associated with older age at initiation, higher lifetime use, physical health conditions, acute spiritual experience and non‑religious contexts. Sociodemographically, female sex increased the likelihood of adverse mental health effects, while younger age at initial use and being married were associated with lower risk. Supplementary analyses showed religious context users tended to report higher doses/year and lifetime use but lower frequencies of anxiety disorder, depressive disorder and physical health conditions; most religious context respondents were Brazilian, while other contexts had a higher proportion of European respondents.

Bouso and colleagues interpret their findings to mean that adverse effects associated with ayahuasca are common but are generally mild, transient and often viewed by users as part of a growth or integration process. Vomiting and nausea were the most frequent physical responses and are frequently framed within traditional and some non‑traditional ceremonies as expected purging with perceived spiritual benefit rather than an untoward event. A small proportion of respondents required medical attention for physical effects and around one in nine reported needing professional mental health support; severe mental adverse effects were uncommon and most mental effects resolved within a week. The investigators note that adverse physical and mental health outcomes showed distinct patterns of association. Physical adverse effects were linked to somatic vulnerability (more physical health conditions), aspects of recent and lifetime use, and to consumption in non‑supervised contexts, whereas adverse mental health effects were related to prior anxiety disorder, the intensity of the acute spiritual experience, and non‑religious contexts. Neurological‑type events were infrequent but flagged as important because of their potential severity and the theoretical risk of convulsions in neurologically vulnerable individuals; the authors emphasise caution for people with epilepsy or brain disease. The study’s limitations acknowledged by the authors include its retrospective, self‑report online design and consequent potential inaccuracies, self‑selection bias, inability to determine the exact plant composition of brews consumed, and geographic over‑representation of Latin American, particularly Brazilian, participants. These limitations restrict causal inference and generalisability. The authors recommend future research collect prospective data on severity and duration of adverse effects, examine mediation and indirect pathways among the study variables, and evaluate how adverse experiences relate to subsequent improvement or deterioration in psychiatric conditions. They also suggest that findings have practical implications for screening, safer provision and integration support, and for public health and regulatory approaches that must bridge traditional practices and biomedical standards. A single participant quote included by the authors summarises one common perspective: “I have had numerous experiences where ayahuasca has brought difficult patterns into my awareness in my daily life, which is never comfortable but always results in growth in the end.”