The association between naturalistic use of psychedelics and co-occurring substance use disorders

Alcohol and drug addiction impose a large and growing public‑health burden, and existing treatments for substance use disorders (SUD) have limited effectiveness for many patients. Earlier clinical and epidemiological research, including mid‑20th century LSD trials and more recent studies of psychedelic‑assisted psychotherapy, has suggested that classic psychedelics (principally LSD, psilocybin and mescaline) can produce enduring improvements in mood, anxiety and addictive behaviours. Proposed mechanisms include serotonin 5‑HT2A receptor agonism that increases glutamatergic transmission and neuroplasticity, psychological effects of mystical or peak experiences, and modulation of impulsivity; contemporary frameworks also emphasise the importance of context or "set and setting" in shaping outcomes, and point to longstanding ceremonial use of substances such as peyote in communal healing rituals. Amit and colleagues framed this work around a gap in population‑level evidence: while some large naturalistic studies report positive mental‑health associations for psychedelic use, there are few large, representative analyses that compare different classic psychedelics individually against a broad SUD outcome. The study therefore aimed to test whether lifetime naturalistic exposure to classic psychedelics (examined both as a class and separately as LSD, psilocybin and peyote/mescaline) was associated with lower prevalence of past‑year substance dependence or abuse after adjustment for sociodemographic, psychiatric and other drug‑use confounders, with nicotine dependence included as a secondary outcome.

This was a cross‑sectional analysis of the 2017 National Survey on Drug Use and Health (NSDUH), a nationally representative survey of the US civilian non‑institutionalised population aged 12 years and older. The analytic sample reported in the paper comprised 56,276 respondents. Primary exposure variables were lifetime use of classic psychedelics. The investigators created a combined variable coded positive if a respondent reported ever using LSD, psilocybin, peyote or mescaline, and they also analysed each substance individually; peyote and mescaline were combined into a single variable because mescaline is the active compound in peyote. The 2017 NSDUH uniquely included recency categories for LSD, so the authors used an additional predictor capturing how recently respondents had last used LSD (past 30 days, past 12 months but >30 days, more than 12 months prior). The primary outcome was past‑year illicit drug or alcohol dependence or abuse, reflecting the DSM‑IV constructs that approximate substance use disorder criteria; nicotine dependence (measured by the Nicotine Dependence Syndrome Scale, NDSS) was included as a secondary outcome because it captures craving. Control variables included past‑year mental‑illness severity and the demographic and behavioural covariates used in a prior NSDUH analysis (age, sex, race/ethnicity, education, household income, marital status, a self‑reported tendency to test oneself with risky behaviour, and lifetime non‑medical use of ten other drug categories such as cannabis, opiates, cocaine, stimulants, MDMA and inhalants). The survey lacked measures for some potentially important confounders (for example, set and setting, frequency and dose of psychedelic use, and lifetime exposure to an extremely stressful event), which were therefore not included. Statistical analyses used binary logistic regression to estimate unadjusted and adjusted odds ratios (OR, aOR). Models were run for the combined psychedelic exposure and separately for each compound, and analogous models were fitted with nicotine dependence as the dependent variable. A separate series of regressions used LSD recency as the predictor. Analyses were performed in SPSS v27. The authors stated that with n = 56,276 and 21 variables, power should be high; significance was set at p < 0.05 with a Bonferroni correction (p < 0.0167) applied when comparing individual psychedelics. All tests were two‑sided.

The sample included 56,276 respondents (27,037 males, 48%; 29,239 females, 52%). Lifetime exposure to any classic psychedelic was reported by 6,362 individuals (11.3%); lifetime prevalence was 8.3% for LSD, 8.2% for psilocybin and 2.3% for peyote/mescaline. Unadjusted analyses showed strong positive associations between lifetime psychedelic exposure and past‑year illicit drug or alcohol dependence or abuse: any classic psychedelic, OR = 6.03 (p < 0.001); LSD, OR = 5.73 (p < 0.001); psilocybin, OR = 6.02 (p < 0.001); peyote/mescaline, OR = 2.97 (p < 0.001). After adjusting for demographic, psychiatric and other drug‑use covariates, these associations were attenuated. For lifetime exposure to any classic psychedelic the adjusted odds ratio was aOR = 1.21 (p < 0.001). When analysed by compound, adjusted estimates differed markedly: LSD aOR = 1.41 (p = 0.024) and psilocybin aOR = 1.135 (p = 0.031), neither meeting the Bonferroni‑corrected threshold of p < 0.0167; by contrast, lifetime peyote/mescaline was associated with lower odds of past‑year illicit drug or alcohol dependence or abuse (aOR = 0.68, p < 0.001). For nicotine dependence the unadjusted associations were also positive (any psychedelic OR = 4.76, p < 0.001; LSD OR = 4.97; psilocybin OR = 4.42; peyote/mescaline OR = 3.67, all p < 0.001). These associations did not remain statistically significant after adjustment, although two adjusted estimates showed suggestive trends that did not survive the Bonferroni correction: LSD aOR = 1.17 (p = 0.0168) and peyote/mescaline aOR = 0.80 (p = 0.018). Analyses using LSD recency revealed that more recent LSD use was associated with markedly higher unadjusted odds of past‑year SUD: past 30 days OR = 18.34 (p < 0.001), past 12 months (but >30 days) OR = 15.65 (p < 0.001), and last use >12 months prior OR = 4.47 (p < 0.001). After adjustment, recency within the past year remained associated with greater odds of past‑year illicit drug or alcohol dependence or abuse (past 30 days aOR = 1.97, p < 0.001; past 12 months aOR = 2.17, p < 0.001), whereas the adjusted aOR for last LSD use >12 months prior was 0.94 and not significant. For nicotine dependence the unadjusted recency associations were elevated (for example, 4.35 times greater odds for past‑30‑day users), but adjusted results were mostly non‑significant; one small adjusted effect was reported for users who last used LSD more than a year prior (aOR = 1.14, p < 0.05). The authors emphasised that inclusion of control variables substantially attenuated the crude associations, and they reported that several covariates behaved as expected: older age and female sex were generally associated with lower odds of SUD, past‑year mental illness and penchant for risk‑taking were strong positive confounders, most other lifetime drug uses predicted higher odds of SUD (with a few exceptions), and socioeconomic variables (income and education) did not show clear directional trends in this analysis.

Amit and colleagues interpreted their findings as showing that crude population associations between lifetime psychedelic use and past‑year substance dependence or abuse are strongly confounded by sociodemographic, psychiatric and other drug‑use factors. Unadjusted models indicated substantially higher prevalence of SUD among psychedelic users, consistent with the common co‑occurrence of psychedelic and other substance use, but adjustment for covariates attenuated most associations. A key and novel observation was that associations differed by compound: lifetime peyote/mescaline use was associated with a lower adjusted odds of past‑year illicit drug or alcohol dependence or abuse (aOR = 0.68, p < 0.001), whereas adjusted associations for LSD and psilocybin were small and not statistically robust after multiplicity correction. The authors propose two non‑exclusive explanations. One is pharmacological: mescaline is a phenylalkylamine and may have different receptor selectivity (for example more selective 5‑HT2 receptor agonism) compared with indoleamines such as LSD and psilocybin, and differences at receptors such as 5‑HT2C might plausibly affect addictive processes. The other explanation emphasises set and setting: peyote/mescaline is often consumed in sacramental or communal contexts (for example within the Native American Church) that could provide therapeutic structure and social support absent from much naturalistic LSD or psilocybin use; the NSDUH lacks measures of context, frequency and dose, so this possibility could not be tested. The authors acknowledged several limitations. The primary SUD outcome derived from DSM‑IV dependence or abuse does not capture DSM‑5 criteria on craving, though the NDSS nicotine outcome did include craving; important potential confounders (set and setting, dose, frequency and lifetime exposure to severe stressors) were not available in the survey; and the cross‑sectional design precludes establishing temporality or causation. They noted that using lifetime psychedelic exposure and past‑year outcomes makes it likely that many exposures preceded outcomes but cannot substitute for prospective incidence data. The authors suggested future research directions: cohort studies to assess relative risk of developing SUD after psychedelic exposure, and controlled comparisons of LSD, psilocybin and mescaline within a unified set‑and‑setting protocol to help disentangle pharmacological from contextual effects. Overall, they emphasised that different classic psychedelics may have different associations with substance‑use outcomes and that further, more granular research is required to understand why.

The study tested the hypothesis that naturalistic lifetime use of classic psychedelics would be associated with lower likelihood of substance use disorder after adjustment for confounders. This pattern was not observed for LSD or psilocybin but was observed for peyote/mescaline, which was associated with reduced adjusted odds of past‑year illicit drug or alcohol dependence or abuse. Amit and colleagues conclude that naturalistic use of different psychedelic compounds may have different associations with SUD and recommend further comparative research that accounts for pharmacological differences and contextual factors.