Psychedelic-induced hypomania and mania: a systematic review and meta-analysis

The introduction sets out that serotonergic psychedelics, including psilocybin, LSD, DMT/ayahuasca, mescaline, and related compounds such as MDMA, have re-emerged as candidates for treating mood, anxiety, and substance use disorders. At the same time, concerns remain that these agents could trigger psychotic worsening or manic/hypomanic switches, particularly in people with bipolar-spectrum vulnerability. The authors note that although classic psychedelics act mainly through 5-HT2A receptor agonism, their effects on serotonergic, dopaminergic, and glutamatergic systems make manic activation biologically plausible. However, earlier evidence is limited, heterogeneous, and insufficient to quantify risk, especially because most clinical trials exclude people with bipolar disorder or a family history of mania. The paper therefore aims to systematically review retrospective and prospective human studies of psychedelic- or MDMA-associated dysphoria/euphoria, hypomania, mania, and later transition to bipolar disorder. The authors aim to distinguish short-term mood changes from persistent syndromes, identify moderating factors such as psychiatric comorbidity and family history, and assess the strength and quality of the evidence across different study designs. MDMA was included as a related compound because it is often discussed alongside psychedelics in the treatment literature.

The study was conducted as a PRISMA 2020 systematic review, registered under CRD420251160656. The researchers searched PubMed, Embase, PsycINFO, and PsycARTICLES from inception to 26 January 2026 without date or language restrictions. They also screened ClinicalTrials.gov and manually checked reference lists of included papers and relevant reviews. Search terms combined bipolar- and mood-related terms with psychedelics including psilocybin, LSD, ayahuasca, DMT, mescaline, and MDMA. Eligible studies involved adults aged 18-65 years exposed to classic psychedelics or MDMA in clinical, naturalistic, or recreational settings. The review included randomised and non-randomised trials as well as observational designs such as cohort, case-control, cross-sectional, registry-based, and longitudinal studies. Case reports, small case series, conference abstracts, editorials, commentaries, reviews without original data, and non-human studies were excluded. Studies were required to report manic or hypomanic symptoms, manic episodes, or diagnostic transition to bipolar disorder using clinical scales, structured interviews, registry diagnoses, or clearly defined self-report. Mixed samples were only included if bipolar-specific outcomes were reported separately or if at least 75% of the sample had bipolar disorder. Screening and data extraction were done independently in duplicate by two reviewers, with disagreements resolved by discussion or third-reviewer adjudication. The extracted variables included study design, sample size, participant characteristics, diagnostic status, drug exposure, setting, dosing, follow-up, timing of outcome assessment, adverse events, and in trials, information on preparation, integration, monitoring, and discontinuation due to adverse events. The primary outcomes were manic or hypomanic symptoms, manic episodes, and transition to bipolar disorder; secondary outcomes included depressive symptoms, persistence or resolution of manic symptoms, psychiatric hospitalisation, and treatment discontinuation. Risk of bias was assessed independently using NIH tools for cross-sectional studies, ROBINS-I for observational and non-randomised studies, and RoB 2 for randomised trials. The authors also graded overall evidence using the Oxford Centre for Evidence-Based Medicine framework. For quantitative synthesis, they performed a random-effects meta-analysis of proportions in R using logit-transformed proportions and restricted maximum likelihood estimation. Heterogeneity was assessed with I² and τ², and publication bias was explored visually with funnel plots, although formal tests were not conducted because fewer than 10 studies were available.

Twenty-three studies met inclusion criteria: three cross-sectional surveys, seven observational studies, three papers reporting two uncontrolled clinical trials, and two randomised controlled trials. Four additional articles on transition to bipolar disorder after hallucinogen-related presentations contributed to the synthesis. Most studies were naturalistic or self-reported, and most were judged to have serious risk of bias. The review also identified ongoing feasibility trials in bipolar II disorder and one phase II DMT study, although these were not part of the main evidence synthesis. In the general population and in genetically or familially vulnerable samples, the studies did not show a uniform association between psychedelic use and manic symptoms. Instead, higher symptom burden tended to appear in people with greater genetic liability, family history of bipolar or psychotic disorders, or other vulnerability markers. The extracted text reports that symptom increases were generally subsyndromal rather than sustained manic episodes. For example, one genetically informed twin study found no overall association between lifetime psychedelic use and manic symptoms, but a stronger association among those with higher polygenic risk for schizophrenia or bipolar I disorder. In individuals with diagnosed or self-reported bipolar disorder, findings were mixed. In a large web-based survey of 541 people with bipolar disorder who had used psilocybin, 32.2% reported new or worsened symptoms in the week after use, including manic symptoms, sleep disturbance, or anxiety; emergency medical service use was rare at 3.3%. A pooled analysis of three prospective naturalistic cohorts found that 16.4% met criteria for clinically meaningful deterioration after psychedelic use, and 12.5% of participants with self-reported bipolar disorder were classified as negative responders. By contrast, a retrospective within-subject analysis in 105 people with bipolar disorder found reductions in depressive symptoms and no statistically significant increase in manic, psychotic, or anxiety symptoms over three months. Another longitudinal study found that illegal or unsupervised use was associated with increases in manic and psychotic symptoms over two months, especially in bipolar I disorder, with frequency of use and challenging experiences also acting as moderators. The controlled clinical studies in bipolar II depression were comparatively reassuring. Trials by Aaronson and colleagues, Rosenblat and colleagues, and Meshkat and colleagues reported antidepressant improvements with no persistent manic episodes and stable or unchanged Young Mania Rating Scale scores. However, these studies were small, tightly controlled, and involved strict exclusion criteria. In the randomised trials in major depressive disorder and alcohol use disorder, the authors report transient euphoric or dysphoric states in some participants during follow-up. In one trial, 4 participants (5%) experienced such transient mood states between 2 days and 3 weeks after dosing, while another reported a single dysphoric mood case over 32 weeks. The pooled proportion of euphoric mood states in these randomised trials was approximately 5.8%. Four registry-based cohort studies examined transition to bipolar disorder after hallucinogen-related psychiatric presentations. Reported transition rates ranged from about 3% to 12.4% depending on the cohort, the definition of exposure, and follow-up duration. In the meta-analysis of transition after hallucinogen-induced mood or psychotic disorder, the pooled prevalence of transition to bipolar disorder was 4% (95% CI 2-8%; N = 7478; I² = 32.1%). When one large registry study was excluded in a subgroup analysis, the pooled prevalence was 6% (95% CI 3-10%; N = 193; I² = 0%). The authors note that the pooled estimate was heavily influenced by the largest registry cohort. Overall, the results suggested a small but statistically significant prevalence of later bipolar diagnosis after severe hallucinogen-related presentations, but the evidence was heterogeneous and largely hypothesis-generating.

The authors interpret the evidence as showing a pattern in which psychedelic exposure does not appear to increase manic symptoms uniformly across the population, but may pose a situational risk in people with pre-existing bipolar or psychosis vulnerability. They distinguish between short-term affective destabilisation and persistent bipolar illness, arguing that most manic or hypomanic reactions reported in the literature are acute, self-limited, and concentrated in higher-risk subgroups, especially those with bipolar I disorder, familial or genetic liability, polysubstance use, sleep deprivation, or unsupervised and illegal use. In contrast, carefully screened and monitored clinical settings appear to carry a much lower observed risk. The discussion places these findings alongside earlier research on antidepressant-associated mood switching and the exclusion of bipolar-spectrum patients from most psychedelic trials. The authors suggest that the literature is broadly consistent with a diathesis-stress model rather than a psychedelic-specific causal mechanism for bipolar disorder. They emphasise that registry-based transitions to bipolar disorder usually emerge over years and are not clearly specific to hallucinogens. The review also discusses a theoretical kindling or sensitisation model, in which repeated exposure might lower the threshold for later episodes, but states that no included study directly tested this and that current longitudinal evidence does not show progressive worsening with repeated psychedelic use. Key limitations are substantial. Most studies were observational, uncontrolled, cross-sectional, or based on self-report, so causal inference is weak. The authors highlight likely self-selection, recall bias, exposure misclassification, residual confounding, and short follow-up as major problems. They also note that severe adverse outcomes may be underrepresented, that many trials exclude the very patients of greatest interest, and that only a small number of studies were suitable for meta-analysis. Because search strategies did not include full-text records, some studies may have been missed. Publication bias is also considered likely. The authors conclude that better prospective, well-controlled studies with systematic adverse-event reporting are needed, especially to clarify long-term outcomes, repeated exposure, and dose-response relationships.

The authors conclude that psychedelics are associated with a low but clinically meaningful risk of transient hypomanic or manic symptoms in people with bipolar vulnerability, particularly in naturalistic or unsupervised settings. They state that the available literature does not support chronic mania or de novo bipolar disorder being induced by psychedelic exposure. They also argue that any observed mania is more consistent with unmasking latent bipolar vulnerability or short-term affective destabilisation than with a true substance-induced bipolar disorder. Finally, they recommend more careful screening for affective instability and longitudinal monitoring, alongside rigorous prospective studies, before broader clinical implementation.