Meta-analysis of executive functioning in ecstasy/polydrug users

Ecstasy (MDMA) use is hypothesised to damage serotonergic axons, and this has led researchers to expect deficits in cognitive processes that depend on serotonin-rich prefrontal regions. Despite animal evidence of serotonin neurotoxicity and some molecular and functional imaging findings in humans, the behavioural literature on ecstasy-related cognitive dysfunction is inconsistent. The authors note that part of this inconsistency may stem from task impurity in classic working memory and executive tasks and from treating executive function (EF) as a unitary construct rather than a set of distinct but related components. This study set out to synthesise the behavioural literature by conducting a meta-analysis of function-specific EF tasks in current ecstasy users compared with drug-using controls. Using the unity/diversity framework of executive functions, the analysis focuses on four component EFs—updating, switching, inhibition and access to long-term semantic memory—with the dual aims of testing for an overall executive deficit in ecstasy users and examining whether impairments are specific to particular EF components. The authors frame this as the most comprehensive analysis of EF in ecstasy users to date and as a means of identifying behavioural correlates of potential serotonergic neurotoxicity.

Eligible studies were human behavioural studies of executive function in ecstasy/MDMA users aged 18 years or older, excluding participants with major psychiatric or neurological conditions. Ecstasy groups were required to be current users and not intoxicated at testing; most studies enforced a minimum abstinence period of 7 days (with a few studies using shorter minima of 24–72 hours). Control groups were required to have some recreational drug experience (polydrug controls), except for three studies in which ecstasy users had minimal exposure to other drugs and drug-naive controls were included. Across included studies the mean age of ecstasy users was 23.39 years (47.72% female) and of controls 23.11 years (54.67% female); mean estimated lifetime dose of ecstasy across studies was 346.03 tablets. The meta-analysis targeted tasks mapped to four EFs: updating, inhibitory control, switching and access (semantic retrieval). For each included task the outcome measure judged to most directly index the putative EF was extracted; each task contributed a single outcome. The formal literature search (July 2015) covered PsycINFO, Scopus and Web of Science using the terms 'Ecstasy' OR 'MDMA' AND 'executive function', supplemented by searches combining drug terms with the names of specific tasks and by manual searches of references. One author performed the initial search and extraction with additional authors conducting supplementary searches and cross-checking; the authors contacted study authors when reported data were insufficient, but data requests were not met for five eligible articles. Data handling procedures included computation of composite scores when not reported, weighting and combining gender-split values, and, when a study used multiple tasks for the same EF, entering each task but dividing the group sample size across tasks to avoid double-counting participants. Where studies subdivided user groups (for example, heavy versus light users), the subgroup with the heaviest ecstasy exposure was used. Extracted variables included sample sizes, gender split, age, estimated lifetime dose, time since last use, task and outcome measure means and standard deviations. Statistical synthesis used standardized mean differences (SMDs) to account for heterogeneous outcome metrics, computed via generic inverse variance methods in RevMan 5.2. Subgroup analyses treated each EF as a subgroup; random-effects models were employed because of substantial heterogeneity. A meta-regression (method-of-moments, random-effects) examined whether estimated lifetime dose predicted SMDs, and publication bias was assessed with funnel plots and Egger's test.

The search initially returned 99 Web of Science, 79 Scopus and 386 PsycINFO records; after de-duplication and title/abstract screening, 88 articles underwent full review and 39 studies met inclusion and provided usable data. Five further eligible studies were identified via supplementary searches but five studies that otherwise met inclusion criteria lacked available data, leaving 39 articles for the final meta-analyses. The meta-analysis incorporated 39 studies contributing 89 effect sizes, with data from 1221 current ecstasy users and 1242 drug-using controls. The pooled test for overall effect indicated a small but statistically significant executive performance deficit in ecstasy users (SMD = -0.18, 95% CI -0.26 to -0.11, Z = 5.05, p < 0.001), with high heterogeneity (I2 = 82%). A formal subgroup test showed significant differences across EF components (χ2 = 22.06, df = 3, p < 0.001, I2 = 86.4%), prompting separate analyses by function. Access (semantic retrieval): 13 studies, 13 effect sizes, 483 ecstasy users versus 491 controls. Ecstasy users showed a moderate deficit (SMD = -0.33, 95% CI -0.46 to -0.19, Z = 4.72, p < 0.001, I2 = 74%). Inhibition (inhibitory control): 20 studies, 20 effect sizes, 606 ecstasy users versus 632 controls. No between-group difference was observed (SMD = 0.04, 95% CI -0.07 to 0.15, Z = 0.77, p > 0.05). Switching (mental set shifting): 18 papers, 23 effect sizes, 488 ecstasy users versus 459 controls. A small but significant deficit was found (SMD = -0.19, 95% CI -0.36 to -0.02, Z = 2.16, p < 0.05, I2 = 85%). Updating (information updating/working memory): 24 articles, 33 effect sizes, 872 ecstasy users versus 904 controls. Ecstasy users performed worse (SMD = -0.26, 95% CI -0.37 to -0.15, Z = 4.49, p < 0.001, I2 = 82%). Meta-regression using available estimates of lifetime ecstasy dose (64 comparisons) produced a non-significant association between lifetime dose and SMD (regression coefficient = -0.0001, 95% CI -0.0004 to 0.0002, Z = -0.74, p > 0.05). Separate meta-regressions for each EF were also non-significant. Examination of publication bias showed funnel-plot asymmetry and a significant Egger's test (t88 = -1.96, p = 0.05), but the authors caution interpreting this result because of the high between-study heterogeneity.

Roberts and colleagues interpret the findings as evidence of small overall executive function deficits in current ecstasy users, with clear variability by component EF. The meta-analysis found significant impairments in access, updating and switching, but not in inhibitory control; updating and access produced the more robust effects. The authors view these behavioural deficits as consistent with molecular and functional imaging evidence suggesting serotonergic alterations in prefrontal regions (for example, reduced presynaptic SERT and increased postsynaptic 5-HT2A receptor availability in frontal areas) and with imaging studies that report increased neuronal recruitment in ecstasy users despite similar behavioural performance. Several explanations for the pattern of results are offered. The absence of an association between estimated lifetime dose and effect size could reflect limitations in exposure measurement—many studies did not report lifetime dose and reporting practices vary widely—or it could indicate that other usage variables (recency of use, frequency, high single-night doses, age at onset) are more relevant to cognitive outcome. The authors note that single high doses can be neurotoxic in animal models and that haemodynamic responses have been linked to recency of use in humans. They therefore recommend more standardised and detailed reporting of drug-use histories in future research. The discussion acknowledges important limitations that temper causal inference. Polydrug confounding remains a concern: in many included studies ecstasy users had greater lifetime use, frequency and variety of other illicit drugs than controls, so the unique contribution of ecstasy cannot be fully isolated. Reverse causality—pre-existing EF deficits predisposing to heavier ecstasy use—cannot be ruled out given the cross-sectional nature of the included studies; the authors call for longitudinal designs to address directionality and to assess recovery after abstinence. High heterogeneity across studies and evidence of publication bias further limit certainty. Finally, the authors consider the potential functional implications, noting that deficits in higher-level EF may impair occupational performance and everyday functioning, and they suggest that understanding these processes could have clinical relevance. They conclude that the meta-analysis provides the most comprehensive behavioural evidence to date linking ecstasy use with selective EF impairments, consistent with putative serotonergic neurotoxicity.