Self-reported PTSD is associated with increased use of MDMA in adolescents with substance use disorders

Lukas and colleagues situate their study in the context of frequent co-occurrence between post-traumatic stress disorder (PTSD) and substance use disorders (SUDs) in adolescents, noting that 20–54% of adolescent SUD patients may meet PTSD criteria while about 30% of adolescent PTSD patients present with SUD. They outline three explanatory frameworks for this co-occurrence: shared genetic or environmental risk, increased exposure to traumatic events (TEs) via high‑risk behaviour, and the self‑medication hypothesis in which PTSD symptoms precede and motivate substance use. The authors note that prior work has established greater SUD severity in adolescents with co-occurring PTSD, but that little is known about the use of specific substances in relation to specific PTSD symptom clusters (intrusion, avoidance, hyperarousal) in this age group. This cross‑sectional exploratory study therefore set out to compare past‑month substance use frequencies across three subgroups of treatment‑seeking adolescent SUD patients — those without trauma exposure (NoTE), those with trauma exposure but not PTSD (TE), and those with probable PTSD — and to examine whether use of particular substances relates to the three PTSD symptom clusters. A secondary aim was to compare ages of first substance use across groups and to test whether onset of PTSD symptoms preceded first use of particular substances, as predicted by the self‑medication hypothesis.

The study used a cross‑sectional design embedded in routine diagnostic procedures at a German outpatient clinic for adolescent substance abuse. Between November 2017 and November 2020, 234 treatment‑seeking adolescents consented; the analytic sample comprised n = 121 participants (42% female) who completed the required questionnaires. Participants were assigned to three groups based on self‑report on the UCLA PTSD Reaction Index for DSM‑IV (German version): NoTE (no reported traumatic event), TE (reported TE but did not meet PTSD criteria), and PTSD (probable PTSD when Criterion A plus intrusion, avoidance and hyperarousal criteria were all met). Internal consistency of the UCLA subscales was reported as good to acceptable (α = .82 and .81 for Criteria A and C; α = .77 and .76 for B and D). Substance use was assessed by clinical psychologists using a study‑specific interview that recorded the number of days each substance was used in the past month and the age of first use. The substances with data analysed were tobacco, alcohol, cannabis, MDMA, and amphetamine ('speed'); methamphetamine, cocaine, opioids, benzodiazepines and solvents were excluded from frequency analyses because of very low or zero reported past‑month use. Diagnostic status for SUD was evaluated with the MINI‑KID (DSM‑5 criteria). Caregivers provided sociodemographic information including age, gender, education (ISCED), and household income categories. Statistical analyses compared prevalence and frequency of substance use across the three groups. Chi‑square tests assessed prevalence differences, Kruskal‑Wallis tests assessed group differences on continuous substance frequency variables, and Mann‑Whitney U tests were used for pairwise follow‑ups when omnibus tests were significant. Within the TE and PTSD groups, Mann‑Whitney U tests examined associations between presence of each PTSD symptom cluster (intrusion, avoidance, hyperarousal) and substance use frequency. Six paired‑sample t‑tests compared age of PTSD symptom onset with age of first substance use. The significance threshold was α < 0.05, and the authors applied the Bonferroni‑Holm procedure to correct for multiple testing. Effect sizes were reported where applicable. The study procedures received institutional ethics approval and informed consent was obtained from participants and guardians.

The final sample included n = 121 adolescents. Group sizes are reported elsewhere in the extraction (PTSD n = 38, TE n = 48, NoTE n = 35 according to the abstract), and the three groups did not differ in distribution of SUD diagnoses, education level, or parental income. A significant difference across groups was observed in the proportion of female participants (X2(2) = 11.2, p = .004). No gender differences were detected for age of first substance use or past‑month use of the substances analysed. The most commonly reported trauma types were violence (26%) and sexual abuse (22%). Analyses of substance use identified MDMA as the substance that differed between groups. The proportion reporting past‑month MDMA use varied across groups (X2(2) = 10.60, p = .005, d = .62), with the probable PTSD group showing the highest prevalence. No group differences were found for past‑month tobacco, alcohol, cannabis or amphetamine use. Days of past‑month MDMA use also differed across groups (H(2) = 9.9, p = .007, η2 = .07); pairwise comparisons showed the PTSD group used MDMA more frequently than the NoTE group (U = 510.5, p = .016, η2 = .04) and the TE group (U = 710.0, p = .010, η2 = .04). The TE and NoTE groups did not differ in MDMA days (U = 839.5, p = .992). Within the TE and PTSD subgroups, MDMA use frequency was significantly higher among participants meeting the avoidance symptom‑cluster criterion than among those who did not (U = 7.68, p = .008, η2 = .73). No associations were detected between MDMA frequency and the intrusion or hyperarousal clusters. Regarding ages of onset, groups differed in age of first substance use (H(2) = 11.3, p = .003, η2 = .08); the PTSD group reported an earlier age of first use than the NoTE group (U = 372.0, p = .001, η2 = .14) and the TE group (U = 653.5, p = .022, η2 = .06). The authors also report that, on average, reported PTSD symptom onset preceded first MDMA use by about 2.2 years (t(26) = -2.89, p = .008), although they note that only a subset of participants could be included in these onset comparisons. Methamphetamine and several other substances were excluded from frequency analyses because of very low or absent past‑month reporting.

Lukas and colleagues interpret their findings as consistent with the self‑medication hypothesis in this sample of treatment‑seeking adolescents with SUD: adolescents with probable PTSD began substance use at an earlier age, were more likely to have used MDMA in the past month, and used it more frequently than peers without PTSD. The authors emphasise that MDMA use was particularly associated with the avoidance symptom cluster, and that reported PTSD symptom onset preceded initiation of MDMA use by an average of about two years in the subset analysed, which they view as further support for the notion that PTSD symptoms may motivate subsequent MDMA use. They discuss plausible mechanisms linking avoidance symptoms and MDMA: avoidance includes affective flattening and detachment, while MDMA can produce heightened empathy and prosocial feelings and is often used in social contexts, which might subjectively counter avoidance‑related distress. The authors also consider why MDMA was not associated with hyperarousal or intrusion: acute MDMA effects (sympathetic activation) could exacerbate hyperarousal, and MDMA‑related transient memory impairment might reduce intrusive memories. An alternative explanation raised is that regular MDMA use could worsen psychopathology — via neuropsychological effects, sleep disruption, depressive or psychotic‑like symptoms, withdrawal phenomena, or exposure to adulterated substances — and thereby contribute to development or worsening of PTSD symptoms. The authors compare their adolescent findings with adult literature that more often links alcohol, benzodiazepines or other substances with avoidance, and suggest developmental and contextual differences in adolescent substance use (different substance preferences and settings) may account for the discrepancy. They propose clinical implications: elevated MDMA use among adolescent SUD patients might signal untreated PTSD and clinicians should routinely assess whether substances are being used as self‑medication. The authors note several limitations they acknowledge: the cross‑sectional, retrospective design prevents causal inference; past‑month use reflects only a snapshot; reliance on self‑report may produce recall or desirability biases and may overestimate probable PTSD; small numbers limited some onset analyses; the convenience, treatment‑seeking sample limits generalisability beyond adolescent SUD patients; inability to control for sociodemographic confounders in primary non‑parametric analyses; and multiple testing risks despite Bonferroni‑Holm correction. They therefore recommend longitudinal and more comprehensive measurement in future research.

The authors conclude that, among adolescent treatment seekers with SUD, probable PTSD is associated with greater likelihood of past‑month MDMA use and higher frequency of MDMA use compared with peers without PTSD. This pattern may reflect self‑medication targeted at avoidance symptoms or, conversely, a contribution of MDMA use to worsening avoidance; regardless of directionality, clinicians should carefully assess for PTSD and for substance use that may function as self‑medication in this vulnerable population.