Neuroimaging of chronic MDMA (“ecstasy”) effects: A meta-analysis

This meta-analysis (2019) evaluated the adverse of effects chronic MDMA/ecstasy use across neuroimaging studies and found no alterations in cerebral blood flow in the basal ganglia and in neurochemical markers in the occipital and frontal lobes, although the serotonin transporter density was decreased in several regions. Given that there was no association between these alterations and lifetime episodes of MDMA use, other factors – such as doses taken per occasion – might be more important determinants.

Abstract

“Introduction: In this meta-analysis, we aimed to assess the evidence from neuroimaging studies for chronic alterations in the brains of MDMA users.

Methods: The databases PubMed, Embase, and Web of Science were searched for studies published from inception to August 24, 2018, without any language restriction. Sixteen independent studies comprising 356 MDMA users and 311 controls were included. Of these, five studies investigated frontal and occipital N-acetylaspartate/creatine and myo-inositol/creatine ratios, three studies assessed basal ganglia blood flow and ten studies investigated serotonin transporter (SERT) density in various regions.

Results: We found significantly decreased SERT density in eight of 13 investigated regions. Meta-regression indicated a positive association with abstinence, but none with lifetime episodes of use. Therefore, other variables (such as doses taken per occasion) might be more important determinants. Positive associations between time of abstinence and SERT density might indicate that these alterations are reversible to some extent. Furthermore, there were no significant differences between user and control groups in terms of neurochemical ratios in the frontal and occipital lobes and blood flow in the basal ganglia.

Discussion: Overall, MDMA user groups showed heavy use patterns and study quality was poor.”

Authors: Felix Müller, Raphael Brändle, Matthias E.Liechti & Stefan Borgwardt

Summary

ABSTRACT

In this meta-analysis, 16 studies were included comprising 356 MDMA users and 311 controls. The study found significantly decreased SERT density in eight of 13 investigated regions, but no significant differences between user and control groups in terms of neurochemical ratios in the frontal and occipital lobes and blood flow in the basal ganglia.

1. Introduction

MDMA is an amphetamine that primarily acts as a serotonin and norepinephrine releasing agent. There are concerns that MDMA might be neurotoxic in humans, especially to serotonergic neurones, but few attempts have been made to meta-analyze previous findings.

The authors found that MDMA use was associated with reduced SERT availability in 11 of 14 investigated brain regions, and examined possible relationships between alterations in SERT density and lifetime episodes of MDMA consumption using meta-regression.

2.1. Search strategy

We searched PubMed, Embase, and Web of Science Core Collection for studies on mdma, ecstasy, and 3-methylenedioxymethamphetamine using magnetic resonance imaging, perfusion weighted imaging, diffusion tensor imaging, and positron emission tomography.

2.2. Search strategy and selection criteria

The authors removed duplicates, reviewed titles and abstracts of all remaining records, and included studies that used magnetic resonance imaging, positron emission tomography, and single photon emission computed tomography (SPECT).

Studies were classified into whole brain approaches and region of interest approaches, and then further divided into domains, which were further divided into sub-categories.

We conducted meta-analyses of studies reporting results on a whole brain level using Seed-based d mapping (Radua and Mataix-Cols, 2012) and associations between cumulative lifetime doses and time of abstinence and neuroimaging measures. However, neither of these analyses was conducted as there were not enough studies available.

2.3. Recorded variables and data extraction

The study included general information, demographic variables, and information on drug history. The missing values were calculated if necessary, and units were transformed if necessary.

To calculate effect sizes, means and standard deviations of neuroimaging outcomes were extracted for MDMA users and controls. If standard errors of the mean (SE) were reported, values were converted using the formula SD n SE=.

2.4. Standardisation of data on lifetime MDMA use

To allow comparison of different data on lifetime use of MDMA, we standardised the values by estimating the lifetime episodes of MDMA use for each study. We expected less variation in terms of “episodes of use” than lifetime intake of tablets. Eight included studies did not report episodes of MDMA use. A weighted mean of 3.0 tablets per episode was calculated across all included studies, resulting in a higher value than a similar estimate recently reported by our group.

2.5. Statistical analysis

A meta-analysis of region of interests was performed using the R package metafor and OpenMEE. A random effects model was used to calculate the pooled effect size, and a leave-one-out meta-analysis was conducted to assess the robustness of the results.

2.5.1. Moderator analysis

Potential influences of lifetime episodes of MDMA use and time of abstinence from MDMA on neuroimaging measures were assessed using meta-regression. Multiple outcomes within the same study were summarised by calculating combined effect sizes and variances.

2.5.2. Assessment of publication bias

Publication bias was assessed using Rosenberg’s fail-safe N approach, and the target significance level was set at p 0.05.

We identified 934 articles and included 16 studies that measured N-acetylaspartate/creatine (NAA/CR), myo-inositol/creatine (MI/CR), CBF of the basal ganglia (MRI/SPECT) and SERT density in several regions in MDMA users and controls.

The results of neuroimaging studies were assessed with different methodologies and the results were assessed with different methods. This required post hoc decisions on the inclusion and exclusion of different approaches.

3.1. Proton magnetic resonance spectroscopy

Five 1H-MRS studies were included in the analysis of NAA/CR and MI/CR ratios in the occipital lobe and mid-frontal lobe. There were no significant differences between MDMA users and controls in terms of NAA/CR or MI/CR ratios in the investigated regions.

3.2. Cerebral blood flow

Two MRI studies and one SPECT study were conducted to investigate CBF in two regions of interest in MDMA users and controls. The results indicated no significant difference in CBF between the two groups.

3.3. Serotonin transporter density

We included ten studies in our meta-analysis, which included 267 users and 234 controls. The results showed that SERT density was increased in the amygdala in ecstasy users.

Significant reductions in SERT density were observed in the amygdala, anterior cingulate, posterior cingulate, hippocampus, occipital lobe, parietal lobe, temporal lobe, and thalamus, but not in the caudate, frontal lobe, insula, midbrain, and putamen.

Most studies used [11C]DASB to measure SERT density, while two studies used [11C]McN. Removal of one study did not alter the results for most regions, but did result in non-significant results for the amygdala, anterior cingulate, and hippocampus.

3.4. Heterogeneity

Significant heterogeneity was found for assessment of CBF in the globus pallidus and for SERT density in several regions of the brain. However, these findings should be interpreted with caution.

3.5. Moderator analysis

Meta-regression indicated that lifetime episodes of MDMA use had no association with SERT density, but time of abstinence had a positive association. However, the association was largely driven by a single study.

3.6. Publication bias

Rosenberg’s fail-safe N indicated that a high number of unpublished non-significant studies were needed to bring the p value of the effect to > 0.05 for most regions.

3.7. Quality assessment of included studies

The included studies exhibited various sources of bias and were of poor quality. They were also affected by issues inherent to these designs, such as unreliable measures of exposure to MDMA.

4. Discussion

MDMA users had lower SERT density in eight out of 13 investigated regions, which might reflect loss of serotonergic neurons caused by MDMA. However, other reasons are also conceivable, such as MDMA-induced downregulation of SERT in response to serotonergic stimulation. Meta-regression indicated that there was no relationship between lifetime episodes of MDMA use and reductions in SERT density. However, a significant association was found between time of abstinence and SERT density.

Potentially, SERT loss is reversible, as already suspected by several authors. However, exclusion of one study had a relative large impact on the result of the meta-regression, so this finding comes with some uncertainty.

Moderate MDMA use might be a neglected field of research, as heavy MDMA users are only a minority among MDMA users. Therefore, neuroimaging studies might overestimate effects. All studies in this meta-analysis investigated heavy users, and thus might not be representative of the therapeutic use of MDMA in posttraumatic stress disorder. Only one study investigated a comparable low cumulative dose, and found no significant alterations in terms of CBF, MRS, and SERT density.

This meta-analysis includes current as well as former MDMA users, and allows assessment of potential moderators using meta-regression. It is limited by small sample sizes, observational designs, and various possible confounders between users and controls, including pre-existing psychological or biological differences, use of an unknown amount of MDMA, and contamination with other used substances. The included studies exhibit various heterogeneities, which complicate comparisons. A leave-one-out meta-analysis revealed that some results on SERT density were not robust, and that data on episodes of use might be imprecise. The quality of the included studies was poor and the quality assessment tool used was not specifically designed for neuroimaging studies. The included studies explicitly aimed to evaluate neurotoxic effects of MDMA, but the two techniques used did not specifically assess serotonergic neurotoxicity.

5. Conclusion

Although studies have shown that MDMA use is associated with decreased SERT density, no association was found between lifetime episodes of MDMA use and decreased SERT density. It is possible that the reductions in SERT density are due to factors other than MDMA.

Contributions

The authors designed the review, performed the data base search, extracted the data and performed the calculations, and interpreted the results.