Is psychedelic use associated with cancer?: Interrogating a half-century-old claim using contemporary population-level data

Barnett and colleagues open by revisiting concerns from the late 1960s that LSD and other psychedelics might cause chromosomal damage and thereby elevate cancer risk. Early laboratory and small human studies reported chromosomal breaks and rearrangements after LSD exposure, and some case reports described leukemia or Philadelphia-like chromosomes in users. Subsequent critiques identified methodological problems (notably unrealistic in vitro exposures, polydrug use and impurity confounding in non-medical users), and later reviews and population studies largely failed to corroborate a carcinogenic effect. Nonetheless, contemporary epidemiological investigations have been limited, and lingering public and clinical concern remains as psychedelic research and use expand. This study therefore set out to examine whether lifetime psychedelic use is associated with lifetime diagnosis of any cancer or of hematologic cancers specifically. Using recent, nationally representative data from the US National Survey on Drug Use and Health (NSDUH) pooled across 2015–2019, the investigators assessed lifetime use of a range of psychedelics (including LSD, psilocybin, DMT, mescaline, peyote, MDMA, 2C-B, 5-MeO-DIPT and AMT) and tested associations with self-reported lifetime cancer diagnoses, while controlling for multiple sociodemographic, health and substance-use covariates. The analysis also considered psychedelic subclasses by chemical structure (tryptamines, lysergamides, phenethylamines).

The study used pooled data from adult (age 18+) respondents to the 2015–2019 NSDUH, a stratified, multistage, nationally representative survey of the civilian non-institutionalised US population. Barnett and colleagues limited the sample to respondents aged 18 and older and combined five years of data to increase sample size for rare outcomes; the final unweighted sample comprised 210,021 unique respondents. The NSDUH employs audio computer-assisted self-interviewing (ACASI) for sensitive questions and uses sample weights to produce nationally representative estimates. Dependent variables were participant-reported lifetime diagnosis of any cancer and lifetime diagnosis of hematologic cancer (NSDUH groups leukemia, blood cancer and lymphoma together). Primary independent variables were lifetime use (ever, even once) of individual psychedelics and derived variables for lifetime use of chemical classes: tryptamines (5-MeO-DIPT, AMT, DMT, psilocybin), lysergamides (LSD), and phenethylamines (2C-B, mescaline, MDMA, peyote, San Pedro). The investigators selected a broad set of covariates known to relate to cancer risk or to substance use, including age, sex, race/ethnicity, BMI, education, income, marital status, rural/urban residence, health insurance, histories of cirrhosis, hepatitis B/C, HIV/AIDS, multiple measures of lifetime tobacco use, mean age of first alcohol use (used as a proxy for lifetime alcohol exposure), and lifetime use of a range of other substances (marijuana, cocaine, stimulants, sedatives, tranquilizers, heroin, prescription opioids, inhalants, PCP, ketamine). Analytically, descriptive weighted and unweighted statistics characterised differences between lifetime psychedelic users and non-users (Pearson's χ2 for categorical variables, Student's t-test for continuous). Four multivariable logistic regression models examined associations between psychedelic exposure and outcomes: (1) any classic psychedelic use versus any cancer, (2) the three structural classes versus any cancer, (3) any classic psychedelic use versus hematologic cancer, and (4) the three structural classes versus hematologic cancer. Models adjusted for the covariates listed above. Calculations were performed in RStudio using the survey package; statistical significance was set at p < 0.05 and no multiple-comparisons correction was applied (p values reported to four decimal places to allow reader-applied corrections).

Of 210,021 respondents, 37,222 (17.72%) reported lifetime psychedelic use. On univariate comparisons, lifetime psychedelic users differed from non-users across many demographic and health variables: they were more often aged 26–64, male, non-Hispanic White or certain other non-Hispanic racial groups, more highly educated, higher income, more frequently uninsured, lower BMI, urban residents, more often never married or divorced/separated, more likely to report risky behaviour and higher lifetime prevalence of HIV/AIDS, cirrhosis, and hepatitis B/C, and more likely to report lifetime use of other psychoactive substances; they also had a lower mean age of first alcohol use (all reported p < 0.0001 for many comparisons). Weighted prevalence of lifetime any-cancer diagnosis was 5.3% among lifetime psychedelic users versus 6.5% among non-users (p < 0.0001). By structural class, weighted lifetime cancer prevalence was 4.9% in tryptamine users versus 6.5% in non-tryptamine users (p < 0.0001), 5.8% in lysergamide users versus 6.4% in non-lysergamide users (p = 0.0269), and 5.0% in phenethylamine users versus 6.5% in non-phenethylamine users (p < 0.0001). For hematologic cancers, weighted prevalence was 0.4% in lifetime psychedelic users versus 0.5% in non-users (p = 0.3438). Prevalences by class were: tryptamine 0.5% versus 0.5% (χ2 p = 0.7698), lysergamide 0.4% versus 0.5% (p = 0.5148), and phenethylamine 0.4% versus 0.5% (χ2 p = 0.3147). In multivariable logistic regression models adjusting for age, sex, race/ethnicity, education, income, marital status, risky behaviour and an extensive set of substance-use and health covariates, no statistically significant associations were found. Specifically, model 1 (any classic psychedelic use and any cancer) yielded p = 0.2926. In model 2 (structural classes and any cancer), tryptamine use p = 0.1760, lysergamide use p = 0.1172, and phenethylamine use p = 0.5434. Model 3 (any psychedelic use and hematologic cancer) produced p = 0.7496. For model 4 (structural classes and hematologic cancer), tryptamine p = 0.0932 and lysergamide p = 0.1727; the extracted text does not clearly report the phenethylamine p value for this model. The authors report full regression outputs in supplemental appendices.

Barnett and colleagues interpret their findings as indicating no epidemiological evidence that lifetime use of psychedelics—either overall or when divided into tryptamine, lysergamide or phenethylamine classes—is associated with lifetime diagnosis of any cancer or of hematologic cancers. They note that although lifetime psychedelic users displayed many characteristics that increase cancer risk (for example, higher rates of tobacco and other substance use, risky behaviour, certain medical comorbidities) they also tended to be younger, have lower BMI and higher socioeconomic status on average; these competing patterns likely contributed to the lower crude cancer prevalence seen among psychedelic users on univariate testing. After multivariable adjustment, apparent protective associations disappeared, and the investigators highlight younger average age among users as an important confounder given the strong association of older age with cancer diagnosis in the adjusted models. The authors situate their results alongside prior work: their findings align with a recent NSDUH-based study that found no association between lifetime classic psychedelic use and past-year cancer, and they contrast with an earlier 1997 case-control study that reported an association between LSD and non-Hodgkin's lymphoma in men; Barnett and colleagues suggest that limited covariate control in that older study may explain the discrepant result. They also note the inability of NSDUH data to isolate non-Hodgkin's lymphoma specifically but reason that, given lymphoma proportion within hematologic cancers, their null finding would likely hold for non-Hodgkin's lymphoma. Limitations are emphasised: reliance on retrospective self-report (with potential underreporting), exclusion of institutionalised populations, lack of data on number of lifetime psychedelic exposures, dose, purity, and timing relative to cancer diagnosis (so temporal sequencing cannot be established), and possible under-ascertainment of cancer among uninsured respondents. The authors acknowledge that survey participants may have consumed impure substances or misidentified compounds in non-medical settings. Strengths cited include the large, nationally representative pooled sample and the ability to adjust for a wide range of potential confounders. Finally, they recommend that future epidemiological research collect finer-grained measures of psychedelic use (age at first use, frequency and dose, time since last use, and data on additional compounds such as ibogaine and ayahuasca) to further clarify any potential long-term carcinogenic risks.

Using data from over 200,000 adult respondents to the NSDUH (2015–2019), Barnett and colleagues found no evidence that lifetime use of lysergamide, phenethylamine, or tryptamine psychedelics is associated with lifetime diagnosis of any cancer or of hematologic cancer. While acknowledging important limitations—most notably lack of information on dosage, frequency, purity and timing of psychedelic exposures—the authors conclude that these population-level data do not support a carcinogenic effect of psychedelics and call for more detailed epidemiological data collection in future studies.