This observational survey study of university students (n=13,837) explored the association between ADHD symptoms and lifetime psychoactive substance use. It finds a significant correlation between high levels of ADHD symptoms and the use of various substances, including ketamine and psilocybin mushrooms.
In a cross sectional study of 13,837 university students, we aimed to explore the association between attention deficit hyperactivity disorder (ADHD) symptoms and lifetime psychoactive substance use (LPSU) on a wide range of illicit substances. Logistic and Hurdel multivariable regressions were used. ADHD symptoms were significantly associated with the lifetime use of ketamine, magic mushrooms, poppers, and nine other psychoactive substances. There was an association between ADHD symptoms and both LPSU and truncated count of lifetime psychoactive substance use. High levels of ADHD symptoms are associated with the use of a large variety and multiple LPSU.
University students have high prevalence of psychoactive substance use and face attendant risks such as mental health problems, academic difficulties, and risky behaviours. Earlier research has established links between attention deficit/hyperactivity disorder (ADHD) and substance use, particularly for alcohol, tobacco, cannabis and cocaine, but evidence is sparse for a wider range of illicit substances. The authors note gaps in the literature for substances such as ketamine, crack, nitrous oxide and poppers, and an absence of studies treating ADHD symptoms as a predictor of the count of lifetime psychoactive substances used. Jean and colleagues set out to examine, in a large sample of French university students, whether ADHD symptoms are associated cross-sectionally with 1) lifetime use of a broad panel of psychoactive substances (including several illicit drugs not widely studied before), and 2) the count of lifetime psychoactive substances used. The working hypothesis was that higher ADHD symptom levels would be linked both to a wider variety of substances used and to polysubstance use.
The analysis used baseline data from the i-Share cohort, an online health study of French university students. Eligible participants were students registered in French higher-education institutions; recruitment employed campus stands, emails, lectures, flyers, social media and student canvassing. After online registration, participants completed a self-administered baseline questionnaire covering sociodemographic, academic, family and health information. The analytic sample comprised those aged 18–30 years who registered between February 2013 and June 2022 and provided informed consent. Lifetime use was assessed for 15 substances: alcohol, tobacco, cannabis, cocaine, ecstasy, amphetamines, crack, heroin, LSD, ketamine, GHB, magic mushrooms, nitrous oxide, poppers and other drugs. Responses were coded yes/no (with refusals treated as missing). A binary variable "any psychoactive substance use" indicated any lifetime use, and a count variable summed the 15 individual substance-use indicators to capture number of substances ever used. ADHD symptoms were measured with the Adult ADHD Self-Report Scale (ASRS) v1.1 (French version). The six-item ASRS yields a global score from 0 to 24 based on frequency of symptoms over the past six months; the instrument is rooted in DSM-IV criteria and has established validity and reliability. Covariates included demographic factors (age, sex), student-related variables (academic level, pre-college education type, social support, employment), family variables (number of siblings, parents' education, parental separation, parental support in childhood) and psychiatric comorbidities (history of suicide attempt, depression, OCD, anxiety, eating disorder, reading disorder, disability). Only participants with no missing data on each psychoactive substance variable were considered for the primary analyses. Missingness in some covariates was handled via multiple imputation using the MICE algorithm; analyses were performed on imputed datasets. The main statistical approach comprised 15 separate logistic regression models (one per substance) with ASRS score as the exposure and substance-specific lifetime use as the outcome. To model the count of lifetime substances used, the investigators employed a Hurdel regression model composed of a logistic part (association with any use) and a truncated Poisson part (association with multiplicity among users), the authors arguing this fits the data better than a standard Poisson or zero-inflated model because it separates the decision to use from the count among users. Variable selection for adjustment used significance testing followed by identification of interactions and a final absolute shrinkage and selection operator (LASSO) to limit collinearity. Numeric variables, including ASRS, were standardised. Results are reported as unadjusted and adjusted odds ratios (uOR, aOR) and as exp(beta) for the count component with 95% confidence intervals. The extracted text states that further details on imputation, model specification and software appear in supplements.
Of 18,113 registered participants, 13,837 met inclusion criteria and had non-missing ASRS data; 76% were women (n = 10,485). Mean age was 20.28 years (sd = 2.22) and mean ASRS score was 10.84 (sd = 3.99). Overall, 93% (n = 12,930) reported any lifetime psychoactive substance use. The most commonly reported lifetime substances were alcohol (92%, n = 12,687), tobacco (55%, n = 7,665), cannabis (53%, n = 7,298), poppers (22%, n = 3,093) and nitrous oxide (22%, n = 3,024). In substance-specific adjusted logistic models, the ASRS score was positively and significantly associated with lifetime use for the majority of substances. The extracted text states that all substances except alcohol and heroin showed significant positive associations with ASRS score. Illustrative adjusted odds ratios reported include ketamine: aOR = 1.66 (95% CI 1.38–2.00), magic mushrooms: aOR = 1.35 (95% CI 1.23–1.49), and poppers: aOR = 1.15 (95% CI 1.10–1.20). Alcohol (aOR = 1.12, 95% CI 0.98–1.28) and heroin (aOR = 1.43, 95% CI 0.96–2.13) had point estimates above 1 but did not reach statistical significance as reported. Using the Hurdel model, higher ASRS scores were associated both with the odds of any psychoactive substance use (adjusted aOR = 1.14, 95% CI 1.07–1.23) and with a higher truncated count of lifetime substances among users (adjusted exp(beta) = 1.10, 95% CI 1.09–1.12), indicating that greater ADHD symptoms related to using a larger number of different substances. Sensitivity analyses produced similar patterns: ketamine aOR = 1.59 (95% CI 1.26–2.00); magic mushrooms aOR = 1.31 (95% CI 1.17–1.47); poppers aOR = 1.12 (95% CI 1.07–1.18); alcohol aOR = 1.02 (95% CI 0.86–1.20). Note on inconsistency in the extracted text: a later "Main findings" paragraph lists substances associated with ADHD symptoms and then states "Only alcohol and crack did not show a significant association," which contradicts earlier statements. The extraction does not clarify whether crack was significantly associated; the primary association paragraph reported alcohol and heroin as the non-significant exceptions. This inconsistency is present in the provided extraction and could not be resolved from the available text.
The investigators interpret their results as confirming and extending prior literature linking ADHD to psychoactive substance use. They note that earlier studies showed associations with tobacco, cannabis and cocaine and that alcohol use per se has not consistently been associated with ADHD (although alcohol use disorder is). By demonstrating associations across a broad panel of substances — including several rarely investigated drugs such as ketamine, GHB, nitrous oxide, magic mushrooms and poppers — the authors argue their findings broaden understanding of ADHD-related risk for illicit drug use. A proposed mechanism discussed is the role of impulsivity commonly observed in individuals with ADHD, which may increase propensity for risky behaviours including illegal substance use. The authors suggest that no illicit substance examined appears to be an exception to the general relationship between ADHD symptoms and lifetime use, based on their analyses. Strengths highlighted include the large sample size, the wide range of substances examined and adjustment for many potential confounders including psychiatric history and family and socio-demographic variables. Limitations acknowledged are reliance on self-reported substance use (risking information bias and possible underreporting), limited statistical power for substances with few users (notably heroin), the predominance of women in the sample which may limit generalisability to the broader student population, and the cross-sectional design which precludes determination of temporality or causal inference. In terms of implications, the authors recommend that the presence of any, and particularly multiple, psychoactive substance use in students should prompt consideration of assessing for ADHD and offering treatment when appropriate. They call for prospective studies to examine whether ADHD predicts initiation, escalation and persistence of use across a wide range of substances, and for research into the role of ADHD in progression to substance use disorders and in treatment outcomes for addiction services.
Higher levels of ADHD symptoms are associated with lifetime use of a wide variety of psychoactive substances and with a greater number of different substances used among French university students. The authors conclude that screening for ADHD in students who use multiple substances could help detect previously unrecognised ADHD and provide opportunities for more effective care.