Residual neurocognitive features of long-term ecstasy users with minimal exposure to other drugs

Halpern and colleagues frame the study against a background of extensive animal evidence that MDMA ("ecstasy") can be neurotoxic, particularly to serotonergic systems, and a substantial body of human field studies that have reported residual cognitive impairments in illicit ecstasy users—most consistently lowered verbal memory. The authors note that many prior naturalistic investigations are vulnerable to confounding factors that could inflate apparent drug-related deficits, including differences in participation in the "rave" subculture (with associated sleep and fluid deprivation), failure to screen for recent substance use, heavy lifetime exposure to other drugs among ecstasy users, and premorbid differences such as lower baseline intelligence or greater impulsivity among people who go on to use ecstasy. To address these concerns, the study set out to replicate and extend a prior pilot investigation using a larger sample while minimising common sources of bias. The research compared neurocognitive performance in ecstasy users versus non-users drawn from the same all-night dance (rave) subculture, imposed strict limits on lifetime exposure to other drugs and alcohol, screened participants for recent substance use by breath, urine, and hair analysis, and adjusted analyses for demographic, family-of-origin, and childhood behavioural variables. The goal was to determine whether residual cognitive deficits attributable to ecstasy use could be detected under these more stringent field-study conditions.

This was a cross-sectional field study recruiting participants from the rave subculture in Salt Lake City, Utah. Recruitment targeted individuals aged 18–45 who were native English speakers and who reported experience in all-night dance parties (at least 10 events staying awake until 4:30 AM). Ecstasy users were defined as reporting at least 17 lifetime episodes of illicit MDMA/ecstasy; non-users reported no lifetime ecstasy use. The investigators excluded people reporting more than 100 lifetime episodes of cannabis or more than 10 episodes of any other class of illicit drug (other than ecstasy), more than 50 lifetime episodes of alcohol intoxication (defined as ≥4 drinks within 4 hours), history of clinically significant head injury or medical illness affecting cognition, and current use of psychoactive medications. Psychiatric disorders were not exclusionary because they might themselves be consequences of ecstasy use. Qualified participants underwent an in-person baseline assessment by a study psychiatrist that included a semi-structured interview for lifetime substance use, a detailed ecstasy-use history (episodes, doses, settings), the Structured Clinical Interview for DSM-IV (SCID) for psychiatric diagnoses, the Hamilton Rating Scales for Depression and Anxiety, the Symptom Checklist-90, and measures of possible premorbid correlates: childhood conduct disorder (DSM-IV items), childhood ADHD (Wender Utah Rating Scale and a modified ADHD rating scale), and family history of substance abuse or psychiatric disorders. A brief neurological examination was also performed. Neuropsychological testing occurred on a separate visit, typically within 4 weeks of baseline. Participants were required to abstain from ecstasy, other illicit drugs, and all-night parties for at least 10 days prior to testing. On arrival they provided a breath alcohol test and a urine dipstick screen for common drugs; a urine aliquot was sent to an external laboratory for MDMA testing, and a hair sample was analysed for residues over the prior 90 days. Participants with positive tests for MDMA or for drugs they had denied were excluded from analyses. The cognitive battery comprised 15 neuropsychological tests, including WAIS-R subtests (Vocabulary, Digit Span, Digit Symbol, Block Design) and other standard measures referenced in the paper; participants also completed the Beck Depression Inventory (BDI). Monetary compensation was provided contingent on passing drug tests. For analysis, ecstasy users were split into "moderate" (17–50 lifetime episodes) and "heavy" (>50 episodes) groups, replicating the pilot study definitions. Primary comparisons were: all ecstasy users versus non-users, and moderate and heavy users versus non-users. Linear regression models adjusted for age, gender, race/ethnicity, four family-of-origin variables (parents' education, parental income while growing up, family history of psychiatric disorder, and family history of substance use disorder), childhood conduct disorder (modeled as presence/absence), and childhood ADHD score (continuous). Models were fitted in Stata 9.2 with two-tailed alpha = 0.05. The authors did not formally correct for multiple comparisons but performed supplementary analyses, including a nonequivalence test using 90% confidence intervals to estimate the maximum effect sizes that could be excluded for comparisons of heavy users versus non-users.

From roughly 1,500 telephone-screened individuals, about 250 qualified for baseline assessment; 116 completed neuropsychological testing, and after excluding five participants with drugs detected in hair or urine, 111 participants were evaluable (52 ecstasy users and 59 non-users). Near the end of recruitment the study slightly relaxed criteria for six participants (two with 15 and 20 lifetime cocaine episodes, three with 11–22 episodes of other hallucinogens, and one reporting 250 lifetime cannabis episodes) and retained them in the sample. Compared with non-users, ecstasy users were more often non-white, had lower parental education, and scored lower on vocabulary. Among ecstasy users, the shortest interval from last use to testing was 25 days; only 3 participants (6%) had abstained for less than 40 days. The user subgroup sizes were 30 moderate and 22 heavy users. Heavy and moderate users did not differ on most baseline variables except that heavy users reported more lifetime raves (median 150 [IQR 70, 238] vs 56 for moderates; P = 0.004) and higher Hamilton Depression Scale scores (median 6.5 [2.5, 10.25] vs 2 [0, 6]; P = 0.02). Comparisons across the full neuropsychological battery revealed few statistically significant differences between ecstasy users and non-users. Some measures—Raven's Progressive Matrices, WAIS-R Digit-Symbol, and WCST total categories—were significantly lower only among moderate users, not heavy users. Heavy users were significantly slower than non-users on the grooved pegboard using the non-dominant hand, with no comparable dominant-hand difference. The most consistent and notable finding was that heavy users showed a significantly lower proportion of "brief" items on the RSAT (a measure of strategic self-regulation), and this effect persisted after additional adjustment for verbal IQ and current BDI scores. Within the ecstasy-user group the RSAT "brief" proportion was inversely associated with lifetime episodes of use (coefficient -1.4, 95% CI -2.3 to -0.4; P = 0.004), using log-transformed lifetime episodes and adjusting for age, gender, and race/ethnicity; inspection of the scatter plot indicated the association was not driven by outliers. Sensitivity checks—re-running analyses with a simplified adjustment set (age, gender, race/ethnicity only) and re-analysing after excluding the six participants who slightly exceeded drug-use criteria—produced results similar to the primary analyses. For outcomes where no significant difference was detected between heavy users and non-users, the nonequivalence calculations indicated the study could exclude medium effects (Cohen's d ≥ 0.5) for many measures and could exclude large effects (Cohen's d ≥ 0.8) for all selected measures at the 0.05 level.

Using stringent field-study methods intended to minimise common confounders, Halpern and colleagues observed few consistent neurocognitive deficits in long-term ecstasy users compared with closely matched non-users from the same rave subculture. The authors interpret lower vocabulary scores among users as likely reflecting premorbid differences in verbal ability rather than ecstasy-induced neurotoxicity, since vocabulary tends to be preserved after neurological insult. Many other comparisons produced no significant differences; where significant findings occurred, they were often concentrated in the moderate-user subgroup rather than in heavy users, which the authors suggest makes a neurotoxic explanation less likely and raises the possibility that some findings represent chance associations owing to multiple comparisons. Their single relatively robust finding was reduced proportion of "brief" items on the RSAT among heavy users, which the authors characterise as poorer strategic self-regulation or increased reflection impulsivity. They emphasise that cross-sectional data cannot determine causality: impulsivity might precede ecstasy use as a risk factor, or it could be an effect of use, and prior studies offer mixed evidence on directionality. The present negative findings contrast with many prior reports (including the authors' own earlier pilot study) that documented cognitive deficits, and the authors discuss possible reasons for divergence: differences in abstinence duration (median abstinence was longer in the present study), unmeasured confounders, sample composition, or chance. They note that some longitudinal and meta-analytic studies find lowered verbal memory associated with ecstasy but that results across studies are inconsistent, with overall small to medium effect sizes and lingering concerns about confounding by other drug use, sleep deprivation, and premorbid traits. Halpern and colleagues conclude that their findings add weight to the view that persistent neurocognitive effects of illicit ecstasy use remain unresolved. They acknowledge two interpretations: either ecstasy produces residual neurotoxicity only in highly exposed individuals or those with vulnerability co-factors (which their sample included few of), or illicit ecstasy use by itself does not generally cause lasting residual neurotoxicity and prior positive findings may have been influenced by confounding. The authors recommend continued caution in attributing neuropsychological deficits to ecstasy exposure and underline that the question is not definitively settled.