Exploring protective associations between the use of classic psychedelics and cocaine use disorder: a population-based survey study

Cocaine use disorder (CUD) is described as a substantial public‑health problem in the United States, associated with elevated morbidity and mortality and with limited effective pharmacological treatments. Earlier research has suggested that naturalistic use of classic psychedelics is associated with lower odds of some deleterious outcomes and that psychedelics may have promise in treating addictive disorders, but population‑level evidence specifically linking individual classic psychedelics to CUD has been limited and inconsistent. Jones and colleagues set out to test whether lifetime use of four commonly reported classic psychedelics — peyote, mescaline, psilocybin and LSD — is associated with reduced odds of past‑year CUD. Using a large, nationally representative survey dataset, the investigators aimed both to examine associations with overall past‑year CUD and to probe which of the 11 DSM‑IV criteria for CUD might show protective associations, should any substance emerge as significant in the main analysis.

This was a cross‑sectional, survey‑based analysis of pooled data from the National Survey on Drug Use and Health (NSDUH) spanning 2015–2019. The analytic sample comprised adults aged 18 and older; the extracted text reports a total sample size of 214,505. The main exposures were lifetime (yes/no) self‑reported use of four classic psychedelics: peyote, mescaline, psilocybin and LSD. The investigators selected these substances because they are the most commonly queried classic psychedelics in the NSDUH. Models adjusted for an extensive set of covariates commonly used in population surveys of substance use. Demographic covariates included sex, age group, race/ethnicity, detailed educational attainment, marital status and annual household income. Additional covariates included self‑reported engagement in risky behaviour, lifetime use of other substances (MDMA/ecstasy, heroin, PCP, inhalants, pain relievers, tranquilizers, stimulants, sedatives, and marijuana) and a comorbid diagnosis of hallucinogen use disorder (based on DSM‑IV criteria). The primary outcomes were past‑year CUD (abuse or dependence) and each of the 11 DSM‑IV criteria for CUD. The investigators used survey‑weighted multivariable logistic regression to account for the NSDUH complex sampling design, implemented with the Survey package in R (version 4.1.2). The prespecified analytic plan tested whether lifetime use of each psychedelic was associated with lowered odds of past‑year CUD; if a substance showed a protective association, sensitivity analyses were run to test associations with individual DSM‑IV criteria while retaining the same covariates. The team also conducted post‑hoc chi‑square analyses comparing demographic characteristics of lifetime peyote users who had versus had not used cocaine.

Preliminary descriptive analyses divided the sample by past‑year CUD status; the extract reports 1,017 individuals met criteria for past‑year CUD. Those with CUD were more likely to be single and to have lower formal educational attainment, among other demographic differences noted by the investigators. In the main multivariable, survey‑weighted logistic model predicting past‑year CUD, peyote was the sole substance associated with lower odds of CUD. Lifetime peyote use was associated with a greater than 50% reduction in the odds of past‑year CUD (adjusted odds ratio, aOR: 0.47). All other substances in the model, including the other classic psychedelics (mescaline, psilocybin, LSD), either showed no association with CUD or were associated with increased odds of CUD. Because peyote was the only substance linked to reduced odds of CUD, the investigators performed sensitivity analyses examining peyote use against each of the 11 DSM‑IV CUD criteria, controlling for the full set of covariates. Lifetime peyote use was associated with lowered odds for seven of the 11 individual criteria; the reported strength of these criterion‑level associations corresponded to large reductions in odds, with aORs in the range reported by the authors (approximately 0.26–0.47). Finally, post‑hoc chi‑square tests comparing peyote users who had versus had not used cocaine identified significant demographic differences across marital status, education level, age, sex and race, while household income did not differ between those groups. The extracted text indicates tables summarised frequencies and model results but does not provide the full tabulated values for all comparisons.

Jones and colleagues interpret their principal finding as a preliminary, population‑level association between lifetime peyote use and lowered odds of past‑year CUD and of several CUD diagnostic criteria. They note that this association was specific to peyote in their analyses; other classic psychedelics did not show the same protective relationship and LSD was observed to associate with increased odds of CUD in this dataset. The investigators emphasise several important limitations. Chief among these is that the analyses are cross‑sectional and correlational, so causal inferences cannot be drawn. Temporal precedence is unclear because peyote use was measured over a lifetime while CUD was measured over the past year; the lifetime variable also precludes assessment of recency and frequency of peyote use. The NSDUH sampling frame omits key populations at elevated risk for substance disorders — people experiencing homelessness, currently incarcerated individuals and active‑duty military — which may limit generalisability and leave unmeasured demographic confounders. The extracted text also notes that full adjustment for third‑variable factors (for example, pre‑existing personality traits, spiritual orientation or other unmeasured social factors) is limited by the available survey items. The authors discuss plausible pathways that could underlie both protective and harmful associations. They outline potential harms linked to classic psychedelic use in population‑level data, including acute adverse experiences (“bad trips”), hallucinogen use disorder, possible increases in psychosis risk and rare persistent perceptual disturbances (HPPD), and suggest that such harms may help explain why some substances (notably LSD) were associated with increased odds of CUD. Regarding protective mechanisms, the investigators propose several hypotheses: pharmacological differences between whole‑plant peyote (a complex mix of alkaloids) and isolated mescaline, pre‑existing personality and spiritual characteristics among peyote users that correlate with lower risk, and contextual factors such as the typical ‘‘set and setting’’ for peyote use. The post‑hoc finding of demographic differences between peyote users who have and have not used cocaine leads the authors to emphasise that third‑variable demographic factors likely contribute to the observed associations. Taken together, the authors present the results as hypothesis‑generating: peyote use is associated with lower odds of CUD in this representative sample, but further longitudinal, pharmacological and contextual research is required to determine causality, identify mediators and to clarify differential risks across classic psychedelics.

The investigators conclude that naturalistic lifetime peyote use is associated with lower odds of past‑year CUD and with reduced odds for a majority of DSM‑IV CUD criteria in this US population‑based sample. They call for longitudinal studies to clarify temporality, as well as research into genetic, pharmacological, personality and contextual factors that could explain the association. The authors also recommend future work to examine potential harms, particularly the finding that LSD was linked to increased odds of CUD, and characterise how different classic psychedelics may differ in risk and protective profiles.