Erowid
 
 
Plants - Drugs Mind - Spirit Freedom - Law Arts - Culture Library  
Hand-Crafted Glass Molecules!
Donate $150+ and get an art glass molecule.
(Pick caffeine, DMT, dopamine, ethanol, harmine, MDMA,
mescaline, serotonin, tryptamine, nitrous, THC, or psilocybin)
Does MDMA Cause Brain Damage?
by Matthew Baggott, BA, and John Mendelson, MD
from Ecstasy: The Complete Guide ed. Julie Holland
Spring 2001

This article, written in early 2001, is the best currently available overview of MDMA neurotoxicity. It is from the excellent book Ecstasy: The Complete Guide edited by Julie Holland which contains a number of very interesting articles on the topic of MDMA, its complications, and its potential as a psychiatric medication. Mattew Baggott's article on neurotoxicity was written as he was finishing his truly epic MDMA literature review published by MAPS as part of their work to explore MDMA as a potential therapeutic agent. Erowid had a very small part in helping Matt with his literature review and its dense 376 pages are a must-read for anyone truly dedicated to understanding the scientific complexities of this challenging psychoactive.

This summary of the neurotoxicity issue for Dr Holland's book, however, is much more accessible and should be understandable by anyone with college-level reading skills and interest.



  1. Introduction
  2. MDMA Can Induce Long-term Serotonergic Changes
  3. Serotonergic Changes are Accompanied by Structural Changes to Axons
  4. Are these Serotonergic and Axonal Changes Evidence of Damage?
  5. The Role of Oxidative Stress in MDMA neurotoxicity
  6. Proposed Sources of Oxidative Stress
  7. Extent of Neurotoxicity Depends on Dose, Route of Administration, and Species
  8. Why are such High Doses Used and Can They be Justified?
  9. Extent of Neurotoxicity in Rats is Influenced by Environment, Especially Ambient Temperature
  10. Time Course of Changes and Extent of Recovery
  11. Behavioral and Functional Correlates of MDMA Exposure in Animals
  12. Studies Comparing Ecstasy Users and Nonusers
  13. Personality Differences between Ecstasy users and nonusers
  14. Neurofunctional Differences between ecstasy users and nonusers
  15. Cognitive Differences between ecstasy users and nonusers
  16. Possible Significance of Cognitive Differences and MDMA Neurotoxicity
  17. Findings in Prospective Clinical MDMA Studies
  18. Potential Strategies for Reducing Risk of Neurotoxicity in Clinical Settings
  19. Need for More Research
  20. Summary

Introduction

The acute toxic effects of MDMA are well documented by hundreds of case reports of adverse events in illicit users. Considering how many people use MDMA, serious acute adverse events seem rare. MDMA appears generally similar to psychostimulants such as methamphetamine with respect to the risks of acute toxicity. With trained personnel, properly screened volunteers, and established protocols for monitoring and treating adverse events, these acute risks appear modest and do not present a strong argument against carefully conducted clinical research with MDMA.

On the other hand, the risks associated with possible long-term brain damage are more difficult to assess. Numerous studies in animals have shown that MDMA can produce long-lasting decreases in brain functions involving the neurotransmitter serotonin. It is unclear what these changes mean. Lasting behavioral changes in MDMA-exposed animals have been seldom detected and are fairly subtle when they are found. Though limited in scope, studies of ecstasy users present a strong probability that similar serotonergic changes occur in many humans. Studies comparing ecstasy users and nonusers support an association between modestly-lowered intelligence testing, or cognitive performance tests, and ecstasy use, but clinically significant performance decreases have not been detected. In other words, there is no increased incidence of clinical complaints or findings.

The modest findings in behavioral studies of MDMA neurotoxicity have led some to dismiss concerns about MDMA neurotoxicity as politically-motivated alarmism. It is commonly pointed out that though fenfluramine and methamphetamine produce similar changes, their status as prescription medications was not affected by this finding. [Erowid Note: Although fenfluramine was removed from the U.S. market in 1997, this was due to its likely negative effect on the heart valve, rather than it being related to findings of neurotoxicity.] However, it is reasonable to note that the truly long-term effects of MDMA exposure are unknown. In 15 years of research on MDMA neurotoxicity, no published studies have investigated whether MDMA exposure can cause significant toxicity that only becomes apparent with aging. This fact must be taken into account when considering the risks and benefits of possible clinical studies. Perhaps the single most worrisome issue surrounding MDMA neurotoxicity is that there may be significant toxicity associated with serotonergic changes that is currently undetected. Although millions of people have taken millions of doses of ecstasy, controlled studies of users have not been large enough to detect any but the most common chronic adverse effects. Possible adverse effects such as an increased incidence of affective disorders, like depression, may have gone unnoticed.

Because so little is known about possible long-term clinical implications of MDMA neurotoxicity, we believe it is important to minimize the risks of neurotoxicity in research volunteers. It is hoped that the information presented here may contribute to assessments of, and perhaps reductions in, the risks associated with MDMA use. This chapter will discuss (1) the nature and meaning of MDMA-induced serotonergic changes; (2) the possible mechanisms of these changes; (3) factors influencing the severity of these changes (such as dose, route of administration, species and animal strain, and environment); and (4) the time course of these changes and recovery. The latter part of this chapter will focus on the implications of long-term serotonergic changes by discussing (5) the behavioral and functional effects of MDMA-induced serotonergic changes in animals; (6) studies comparing ecstasy users to nonusers (including personality, cognitive, and functional comparisons); (7) available data from clinical studies in which MDMA was administered; and (8) potential strategies for reducing risk to human volunteers.

Limitations of space unfortunately prevent a full discussion of every important paper and aspect of this complex topic. For a broader sense of the range of views on MDMA neurotoxicity, the reader is therefore advised to consult other review articles (Boot, 2000; Burgess, 2000; Green, 1995; Hegadoren, 1999; McKenna , 1990; Morgan 2000; O'Callaghan, 2001; Seiden, 1996; Sprague, 1998; Steele, 1994), and the issue of Neuropsychobiology (Vol. 42, 2000) dedicated to MDMA neurotoxicity.


MDMA Can Induce Long-term Serotonergic Changes

Before discussing MDMA-induced changes and their meaning, it is necessary to define a few terms. In this chapter, drug doses and dosing patterns used in research that produce these long-term serotonergic changes will be referred to as "neurotoxic regimens." Neurotoxic regimens often consist of four to eight injections of MDMA given over the course of one to four days; however, a single injection of MDMA can also produce these changes. In this chapter, any changes noted at 7 or more days after drug administration will be considered "long-term." Many studies examining the brains of animals at longer time periods (often at 2, 4, or 8 weeks) have established that the MDMA-induced changes at 7 days are primarily long-term in nature.

The term "neurotoxicity" is more difficult to define. Though no universal definition exists, most definitions are broad enough to encompass both short-term alcohol-induced headaches and the permanent nerve cell loss caused by the drug MPTP. A more useful approach to the question of whether MDMA is neurotoxic is to describe the nature and mechanisms of the long-term changes it can cause. In this way, it is evident that some neurotoxic MDMA regimens produce both changes in the serotonergic system and acute damage to the brain by free radicals, and thereby cause a loss of nerve cell axons. This suggests that MDMA neurotoxicity is a type of drug-induced damage, even though the consequences of this damage are unknown.

MDMA does produce long-lasting changes to the serotonergic system at some doses. These long-term changes include decreases in brain concentrations of the neurotransmitter serotonin (5-HT) and its metabolite 5-hydroxyindoleacetic acid (5-HIAA). Levels of tryptophan hydroxylase (TPH), the enzyme that begins the synthesis of 5-HT within the serotonergic nerve cell are decreased. There are also decreases in the density of the serotonin reuptake transporter (SERT), the protein on the membrane of serotonergic neurons that recycles released 5-HT by pulling it back into the cell. Most studies suggest that MDMA primarily causes long-term changes in serotonergic axons that have their cell bodies in an area of the brainstem called the dorsal raphe nucleus.

Long-lasting decreases in these serotonergic markers suggest that either (a) some type of "down regulation" has occurred, meaning the nerve cell is making and maintaining less of the markers, or (b) that serotonergic axons are permanently lost. The question of whether MDMA is truly neurotoxic stems from this issue. Down regulation suggests an active adaptation to drug effects, while axonal loss suggests true damage may have occurred. Determining which actually happens can be difficult. SERT density may change in response to drugs, but this has been difficult to consistently demonstrate (Le Poul, 2000; Ramamoorthy, 1998). Similarly, 5-HT levels can be influenced by diet and other factors. Because MDMA has been shown to rapidly inactivate the enzyme TPH, decreased 5-HT levels would be expected until TPH activity returns to normal. Thus, decreased 5-HT synthesis and subsequent SERT down-regulation initially appear to be a plausible explanation for MDMA-induced serotonergic changes. By examining the structure of serotonergic axons in MDMA-exposed animals, however, it is clear that MDMA can also cause axonal loss.


Serotonergic Changes are Accompanied by Structural Changes to Axons

An important approach to understanding MDMA-induced serotonergic changes involves staining brain slices from MDMA-exposed animals. By attaching a marker to the 5-HT molecule by a process called immunocytochemistry, 5-HT is stained, allowing serotonergic axons and terminals to be seen under a microscope. This technique shows irregular swelling and fragmentation of fine serotonergic axons shortly after a neurotoxic regimen of MDMA or MDA (Kalia, 2000; O'Hearn, 1988; Scallet, 1988). Later measurements taken at 2 or 4 weeks after neurotoxic MDMA regimens also show a persistent decrease in stained axons (O'Hearn, 1988; Scallet, 1988; Slikker, 1988; Wilson, 1989). The initial swelling suggests some type of axonal damage, while the later decrease in stained axons suggests a loss of axons. However, some have argued that immunocytochemistry cannot determine whether or not measured differences in 5-HT are accompanied by changes in the axons themselves. Because of this limitation, it is necessary to confirm the apparent loss of axons using techniques that do not rely on serotonergic markers.

Transport of materials within axons is crucial for maintaining cell structure and function. Lasting reductions in axonal transport suggest a drastic impairment of axonal functioning and, more likely, loss of axons. One can assess axonal transport by measuring the movement of compounds between brain regions that serotonergic axons should connect. For example, if injected into the cortex (the outer layer of the brain) the fluorescent dye Fluoro-Gold should be transported along serotonergic axons into cell bodies in the brainstem. Axonal transport studies have been carried out after neurotoxic MDMA (Ricaurte, 2000) and parachloroamphetamine (PMA) (Fritschy, 1988; Haring, 1992) regimens. Their results suggest that a loss of axons occurs after at least some neurotoxic regimens of MDMA and related drugs.

Another method of assessing loss of nerve terminals involves measuring the vesicular monoamine transporter type II (VMAT2). This is a protein on the storage compartments, or "vesicles," inside the nerve's axon terminals. Because the amount of VMAT2 does not appear to be adjusted in response to drug exposure (Vander Borght, 1995), it is sometimes used as an indirect measure of nerve terminals in research on neurodegenerative disorders such as Parkinson's disease. In other words, decreased VMAT2 would suggest that nerve terminals and axons have been lost. Neurotoxic regimens of MDMA (Ricaurte, 2000) or methamphetamine (Frey, 1997) have been shown to decrease VMAT2. Therefore, at least some neurotoxic regimens of MDMA are associated with structural changes to cells.

The above data consistently indicate that MDMA can cause serotonergic axons to degenerate and that this explains at least some of the MDMA-induced decrease in serotonergic markers. Further evidence of axonal degeneration comes from studies in which recovery from MDMA neurotoxicity is associated with apparent sprouting and regrowth of axons (discussed in more detail below). Why, then, has MDMA neurotoxicity been controversial? One reason is that attempts to measure MDMA-induced cell damage itself yield ambiguous results.


Are these Serotonergic and Axonal Changes Evidence of Damage?

In general, neural cell damage can be detected by two techniques, using silver staining and measuring the expression of glilal fibrillary acidic protein (GFAP). Not all neurotoxic regimens using MDMA are able to demonstrate increased silver staining or GFAP expression. These techniques seem to detect MDMA-induced alterations only at doses higher than those needed to affect serotonergic function (Commins, 1987; O'Callaghan, 1993). Furthermore, the MDMA-induced cell damage detected by silver staining appears to occur in nonserotonergic cells (Commins, 1987; Jensen, 1993) as well as in what are likely serotonergic axons (Scallet, 1988). These inconsistencies are difficult to interpret. Some believe they are evidence that MDMA-induced serotonergic changes result from down regulation of the serotonergic system rather than damage (e.g., O'Callaghan, 2001). Others have argued that the techniques for measuring cell damage are simply insensitive to selective serotonergic damage (Axt, 1994; Bendotti, 1994; Wilson, 1994).

Because studies of axonal transport and VMAT2 changes have provided strong evidence of MDMA-induced axonal damage, it appears that serotonergic down regulation can no longer fully explain the long-term effects of MDMA. Structural changes to serotonergic axons must also be explained. Although we are not aware that this hypothesis has been advanced, one could argue that loss of axons represents a non-neurotoxic form of neuroplasticity, or benign change in the nerve cell in response to drugs. Non-neurotoxic (though not necessarily beneficial) morphological changes can occur in the CNS as the result of alterations in serotonin levels (reviewed in Azmitia 1999). It appears more likely, however, that these changes are, in fact, the result of damage, specifically damage involving oxidative stress.


The Role of Oxidative Stress in MDMA neurotoxicity

Free radicals are highly reactive chemicals that contain one or more unpaired electrons and exist separately. Free radicals can damage neural molecules through reactions called reduction and oxidation, and thereby alter the ability of these molecules to carry out their normal cellular function. Neurotoxic regimens of MDMA increase oxidative stress in the brain. In this chapter, the term "oxidative stress" will be used to refer to both the increase in reactive chemicals, including free radicals, and the burden they place on cellular functioning.

MDMA-induced oxidative stress has been shown in two ways. First, researchers have examined the brains of MDMA-treated animals for thiobarbituric acid reacting substances (Colado,1997a; Jayanthi, 1999; Sprague, 1995b). Increases in these substances suggest that neural lipids, or fat molecules in the brain cells, have been oxidized. Second, researchers have perfused the brains of live animals with either salicylate or d-phenylalanine. These substances react with hydroxyl radicals to form 2,3-dihydroxybenzoic acid and d-tyrosine, respectively. By measuring formation of these compounds, researchers have demonstrated that neurotoxic MDMA regimens increase the amount of extracellular hydroxyl radicals of the striatum (Shankaran, 1999a; b) and hippocampus (Colado, 1999b; 1997b), two areas of the brain involved in movement, and memory, respectively.

There is strong evidence that oxidative stress is involved in the mechanisms of MDMA neurotoxicity. Antioxidants are substances which inactivate free radicals; the antioxidants ascorbate and cysteine each reduce MDMA neurotoxicity in rats without altering levels of MDMA or MDMA-stimulated dopamine release (Gudelsky 1996; Schmidt, 1990). The free radical scavenger N-tert-butyl-alpha-phenylnitrone decreases both MDMA-induced hydroxyl formation and MDMA neurotoxicity in rats; this latter effect, however, may be partially due to an attenuation of MDMA-induced high body temperature, or hyperthermia (Che, 1995; Colado, 1998; 1995; Yeh 1999). Pretreatment with the antioxidant alpha-lipoic acid blocks both MDMA-induced serotonergic neurotoxicity and increased GFAP expression in the rat hippocampus without altering MDMA-induced hyperthermia (Aguirre, 1999). Mice that have been genetically altered to have large amounts of the human antioxidant enzyme, copper/zinc superoxide dismutase, are protected from MDMA-induced dopamine depletions, probably because of the increased trapping of superoxide radicals (Cadet, 1994; Cadet,1995; Jayanthi, 1999). At the same time, these genetically modified mice are protected from the acute inactivation of antioxidant enzymes and free radical changes seen in normal mice after a neurotoxic MDMA regimen (Jayanthi, 1999).

Early evidence that MDMA caused significant oxidative stress came from Stone (1989a) who reactivated TPH which had been inactivated in rats at 3 hours after high dose MDMA by using sulfhydryl reducing conditions. This showed that the acute inactivation of TPH by MDMA was due to intracellular oxidative stress. Intracellular oxidative stress appears to be an effect of MDMA that requires sustained brain concentrations of MDMA (or a centrally formed metabolite). While a single injection of MDMA into the brain had no effect on TPH activity, slow infusion of 1 mg/kg MDMA into the brain over 1 hr produced enough oxidative stress to acutely reduce TPH activity (Schmidt, 1988). The acute decrease in TPH activity is an early effect of MDMA and can be seen 15 minutes after administration (Stone, 1989b). TPH inactivation can also be produced by non-neurotoxic MDMA doses (Schmidt, 1988; Stone, 1989a; 1989b). It therefore appears that MDMA rapidly induces oxidative stress but only produces neurotoxicity when the brain's free radical scavenging systems become overwhelmed.

In summary, MDMA neurotoxicity involves an initial period of oxidative damage, where an increase in free radicals damages neural lipids. This damage seems to be part of the sequence of events producing serotonergic neurotoxicity since treatments that decrease MDMA-induced oxidative stress also decrease the long-term serotonergic changes (Aguirre, 1999). While MDMA can cause loss of axons, some serotonergic down regulation cannot be ruled out. Research on methamphetamine-induced dopaminergic neurotoxicity has led some to conclude that long-term dopaminergic changes can occur without significant axonal loss (Harvey, 2000; Wilson, 1996). Whether this is also the case with MDMA is unknown. For now, it seems reasonable to consider long-term serotonergic alterations after MDMA exposure as indicating that some degree of damage has occurred, while remembering that one is also measuring the response of the serotonergic system to acute drug effects and loss of axons.


Proposed Sources of Oxidative Stress

Several possible sources of neurotoxic oxidative stress have been proposed. First, the sustained effects of MDMA may deplete neuronal energy sources and/or impair energy metabolism within the neuron (Huether, 1997). Second, both MDMA and dopamine can be metabolized to highly reactive quinone-like molecules. Quinones are molecules which are often very reactive, can form free radicals, and are thus potentially damaging to neural molecules. There is not yet conclusive evidence to implicate any of these possible causes and some combination of mechanisms is possible. The possible roles of energy exhaustion or impairment, MDMA metabolites, and dopamine metabolites are discussed below. It has also been proposed that 5-HT metabolites, increased intracellular calcium, nitric oxide, or glutamate may contribute to MDMA neurotoxicity. However, current evidence provides little support for these theories and their discussion will be omitted in the interest of brevity.

Cellular energy exhaustion or impairment may cause MDMA neurotoxicity. Normal activity of the neuron cause a certain degree of oxidative stress. A sustained increase in neuronal activity would therefore be expected to increase oxidative stress. More importantly, increased neuronal activity is accompanied by increased energy consumption that could eventually lead to a depletion of neuronal energy sources. This can impair the energy-requiring mechanisms that maintain and repair neurons. Furthermore, the most important source of cellular energy, mitochondria, can be impaired by oxidative stress (Crompton, 1999). Mitochondria produce adenosine triphosphate (ATP), the source of energy for most cellular processes. Insufficient ATP will lead to cell damage or death.

Whether energy exhaustion or impairment actually plays a role in MDMA neurotoxicity is not yet clear. MDMA does increase activity of the enzyme glycogen phosphorylase (Poblete and Azmitia 1995), which suggests that MDMA could decrease glial stores of glycogen, an important source of energy in the brain. MDMA-induced alterations in mitochondria functioning have been reported (Burrows, 2000), but it is not yet clear if these alterations are sufficient to impair mitochondria and damage cells. One study found that ATP levels were unaltered up to three hours after MDMA administration (Hervias, 2000). Since later times were not examined, it remains possible that ATP is decreased at later time points.

MDMA breakdown products, or metabolites, may also play a role in MDMA neurotoxicity. However, it is difficult to investigate this possible role. Hypothetically, a given metabolite may only be toxic in the presence of MDMA, when the metabolite has high concentrations in the brain for several hours, or when certain acute effects of MDMA have already occurred. In such situations, administering the toxic metabolite on its own would not necessarily lead to toxicity. Thus, it is hard to interpret the many studies in which an MDMA metabolite was administered and no evidence of neurotoxicity was found (Elayan, 1992; Johnson, 1992; McCann, 1991b; Steele, 1991; Zhao, 1992). The MDMA metabolite, alpha-methyl dopamine, may contribute to neurotoxicity as its metabolites that can deplete 5-HT (miller et al 1997).

It has also been suggested that some of the dopamine released by MDMA may be transported by SERT into serotonergic axons (Faraj, 1994) and subsequently oxidized (Nash 1990; Schmidt, 1990; Sprague and Nichols 1995b). The oxidation of dopamine can form hydrogen peroxide which, in turn, may produce hydroxyl radicals. A quinone-like dopamine metabolite may also be formed with potential to generate further free radicals (Cadet and Brannock 1998; Graham, 1978)). Among many other potential toxic effects on cells, dopamine oxidation products have been shown to impair mitochondrial functioning (Berman and Hastings 1999). There is currently little direct evidence to support a role for dopamine metabolites in MDMA neurotoxicity. Some dopaminergic drugs alter MDMA neurotoxicity, but it is not clear that this is due to increasing or decreasing dopamine release. Many dopaminergic drugs are now thought to affect MDMA neurotoxicity through nonspecific mechanisms such as altering body temperature (Colado, 1999a; Malberg, 1996) or scavenging free radicals (Sprague and Nichols 1995a; b; Sprague, 1999). However, dopamine release does seem to play a poorly understood role in MDMA neurotoxicity (Nash and Brodkin 1991; Schmidt, 1990; Shankaran, 1999b; Stone, 1988).


Extent of Neurotoxicity Depends on Dose, Route of Administration, and Species

Extent of neurotoxicity is dose-dependent. Long-term changes occur in rats at doses approximately 5 to 10 times higher than those known to be psychoactive in humans (O'Shea, 1998; Commins, 1987). Most MDMA neurotoxicity studies have used multiple dose regimens. These studies show that "binge" use of MDMA carries greater risk of neurotoxicity than single doses. When administered repeatedly, a non-neurotoxic dose of MDMA can become neurotoxic (Battaglia, 1988; O'Shea, 1998). Multiple dose neurotoxic regimens appear able to produce more profound and possibly more lasting serotonergic changes than single MDMA administration (Battaglia, 1988). The results of multiple dose studies are difficult to compare across species since the same interval between doses can have very different effects in two species with different clearance rates of MDMA.

The effect of the route of MDMA administration in altering long-term serotonergic changes has been investigated. In the rat, subcutaneous injection and oral administration of MDMA produce comparable 5-HT depletions in the hippocampus (Finnegan, 1988). Studies with nonhuman primates have yielded less consistent results. In the squirrel monkey, Ricaurte (1988a) found that repeated oral administration of MDMA resulted in only one-half to two-thirds as much 5-HT depletion as the equivalent subcutaneous dose. In the rhesus monkey, in contrast, Kleven (1989) reported that repeated oral administration of MDMA produced twice the decrease in hippocampal SERT activity as was produced by repeated subcutaneous injection. These apparent differences between nonhuman primate species increase the difficulty of assessing the risk of oral MDMA administration in humans.

Different species differ in sensitivity to MDMA neurotoxicity. In rats, for example, Logan (1988) was unable to detect neurotoxicity when 25 mg/kg MDMA was administered to randomly bred albino rats. In contrast, Dark Agouti rats have a threshold between 4 and 10 mg/kg of injected MDMA for showing 5-HT depletions (O'Shea, 1998). These apparent strain differences may also be influenced by differences in ambient temperature and animal housing (Dafters, 1995; Gordon, 1994).

In comparison to rats, nonhuman primates seem to be more sensitive to MDMA neurotoxicity, suffering more damage at lower doses (Ali, 1993; Fischer, 1995; Insel, 1989; Ricaurte, 1992; Ricaurte, 1992a; but see also De Souza, 1990, for slightly different results). Many MDMA neurotoxicity studies have used squirrel monkeys as subjects. The threshold dose for producing long-term 5-HT depletions in squirrel monkeys is somewhere between 2.5 and 5 mg/kg oral MDMA. Two weeks after a single 5.0 mg/kg oral MDMA dose to this species, 5-HT levels were decreased to 83% of control levels in the hypothalamus and 79% of controls in the thalamus but were not changed in other examined brain regions (Ricaurte, 1988a). In contrast, no long-term serotonergic changes occurred after 2.5 mg/kg MDMA was given orally every two weeks for four months to squirrel monkeys (Ricaurte, unpublished, cited in Vollenweider, 1999). [Editor's note: a therapeutic dose of 125 mg in a 150 pound person would translate to 1.66 mg/kg]

Another commonly studied nonhuman primate species is the rhesus monkey. Determining the threshold dose for 5-HT depletions in this species is difficult since all published studies using rhesus monkeys have employed multiple dose neurotoxic regimens. In one study, 1.25 mg/kg oral MDMA did not produce any long-term serotonergic changes when given twice daily for 4 consecutive days. Similarly repeated doses of 2.5 mg/kg MDMA lowered hippocampal 5-HT (to about 80% of controls) but did not affect levels in 6 other brain regions at post one month (Ali, 1993). In another experiment, Insel (1989) found that 2.5 mg/kg MDMA given intramuscularly twice daily for 4 days to rhesus monkeys produced extensive (possibly short term) 5-HT depletions but did not alter SERT density at 16 to 18 hours after the last drug exposure. Since SERT was unaffected, the researchers concluded that axonal loss had not occurred, despite the 5-HT depletions.

In a study that raises interesting questions about possible tolerance to MDMA neurotoxicity, Frederick (1995) investigated the long-term effects of escalating doses of MDMA. Intramuscular MDMA (0.1-20 mg/kg) was given twice daily for 14 consecutive days at each dose level and followed by three dose-response regimens using single MDMA doses up to 5.6 mg/kg. One month after the final dose-response determination and 21 months after the initial escalating dose regimen, animals were sacrificed. Few significant serotonergic effects were found. MDMA exposure did not produce significant 5-HT depletions in any brain region and decreased SERT to about 60% of control levels only in the hippocampus (and not two other brain regions). Thus, data on rhesus monkeys are complex and perhaps all that can be said with certainty is that the threshold dose for long-term 5-HT depletions appears to be above 1.25 mg/kg oral MDMA in this species.


Why are such High Doses Used and Can They be Justified?

Research on MDMA neurotoxicity has sometimes been criticized for the repeated high dose regimens that are commonly used. Some have questioned whether repeated injections of 20 mg/kg MDMA in rodents can provide useful information about the toxicity of single oral doses of 1.7 to 2.0 mg/kg MDMA in humans. It is true that many of the neurotoxic regimens are not designed to be clinically relevant but were intended to maximize the serotonergic neurotoxicity of MDMA in order to better understand its mechanisms and consequences.

However, comparing dose on the basis on body weight can be misleading. In general, smaller species excrete drugs more quickly and form metabolites in greater amounts than larger species. This is due to many factors including the proportionally larger livers and kidneys and faster blood circulation times in smaller mammals (Lin 1998; Mordenti,1989). As a result of such factors, the time it takes to lower the plasma levels of MDMA by half is about 1.5 hours in a rat (Cho, 1990) and about 8 hours in a human (Mas, 1999). This suggests that small species may require higher doses to achieve drug exposures comparable to those seen in larger species. These considerations at least partially justify the apparently high doses commonly used in rodent toxicity studies. Unfortunately, higher doses tend to alter the character of the drug exposure. While they lengthen the time smaller animals are exposed to the drug, they also tend the produce higher peak blood concentrations of drug and greater acute effects than occur in larger species at lower doses.

A number of techniques have been developed for estimating equivalent drug doses in different species (Ings 1990; Lin 1998; Mahmood 1999; Mordenti, 1989). One of the most commonly used techniques, allometric interspecies scaling, involves administering a drug to different species and measuring resulting blood concentrations of drug. These measurements are then used to determine the relationships between species weight, drug exposure, and dose. Drug exposure in humans can then be estimated from these relationships. In these estimates, equivalent drug exposures are assumed to produce equivalent drug effects, including neurotoxicity. Recently, Ricaurte (2000) estimated that as little as 1.28 mg/kg MDMA may produce long-term 5-HT depletions in humans if interspecies dose conversions for MDMA follow a pattern that is common for drugs that are not extensively metabolized. Estimates of this sort are useful for emphasizing that the MDMA dose required to produce neurotoxicity in humans may be within the range of commonly administered doses, despite the seemingly higher doses used in rodent studies.

However, such estimates require making assumptions about the mechanisms of neurotoxicity. For example, it is necessary to assume that the different species experience comparable drug effects when blood concentrations of drug are the same. This may not be true of neurotoxicity. Several other possible reasons for species differences in MDMA neurotoxicity have already been given. In addition, species may differ in the brain concentration of drug produced by a given blood concentration. It is not known if this is the case with MDMA, although it does seem to be true for fenfluramine (Campbell 1995). Furthermore, if MDMA neurotoxicity is caused by a toxic metabolite, as some have suggested, then the more extensive metabolism of MDMA expected in smaller animals will lead to increased neurotoxicity. Formation of specific drug metabolites in different species is difficult to predict and few data are available on MDMA. Research on species differences in fenfluramine metabolism have led some to conclude that no nonhuman species provides a good model of possible human fenfluramine neurotoxicity (Caccia, 1995; Marchant, 1992). Because current data suggest that both MDMA and metabolite exposure may mediate neurotoxicity, more data are needed from more species before interspecies dose conversions can be made with any confidence.

Data from clinical MDMA studies show that there is a complex relationship between MDMA dose and blood levels of the drug and its metabolites (de la Torre, 2000; Mas, 1999). It appears that MDMA inactivates one of the enzymes in the liver that is important to its metabolism (an enzyme known as cytochrome p450 isozyme 2D6 or 'CYP 2D6') (Brady, 1986; Wu, 1997). As a result, small increases in dose can lead to large increases in drug exposure. When dose was increased from 120 mg to 150 mg, drug exposure almost doubled in human volunteers, as measured by area under the curve of MDMA plasma concentration verses time (de la Torre, 2000). However, formation of some metabolites remained approximately constant. These complex dose-dependent pharmacokinetics in humans further increase the difficulty of estimating dose conversions between species. Nonetheless, these human studies with MDMA do suggest that doses above 120 mg may be associated with unexpectedly increased drug exposure and therefore risks of toxicity.


Extent of Neurotoxicity in Rats is Influenced by Environment, Especially Ambient Temperature

new: Influence of Environment, especially Ambient Temperature, on Neurotoxicity in Rats and Mice
Several studies have explored the relationships between environmental temperature, animal core temperature, and neurotoxicity. In rats, MDMA can dose-dependently impair temperature regulation (Broening, 1995; Colado, 1995; Dafters 1994; 1995; Gordon, 1991), perhaps through alterations in the functioning of the hypothalamus and thermoregulatory behaviors. Resulting changes in animal temperature can alter neurotoxicity; hyperthermia increases and hypothermia decreases serotonergic depletions. Thus, the degree of hyperthermia has been found to correlate with both long-term 5-HT depletions in adult rats (Broening, 1995; Colado, 1993; 1995; Malberg, 1998) and long-term dopamine depletions in mice (Miller, 1994). In addition to the ambient temperature, the degree of hyperthermia is influenced by the thermal conductivity of animal housing and hydration status (Dafters, 1995; Gordon, 1994).

The mechanisms by which temperature affects MDMA neurotoxicity are unclear. Plasma levels of MDMA in rats (Colado, 1995) and brain levels of MDMA in mice (Campbell, 1996) do not appear to be influenced by changes in animal core temperature. MDMA-induced neurotransmitter release may be temperature sensitive (Sabol, 1998), although studies examining the temperature dependence of methamphetamine-induced dopamine release have reported conflicting findings (Bowyer, 1993; LaVoie,1999). It may also be that increased temperature nonspecifically increases the rate of chemical reactions and contributes to oxidative stress, as this does occur in the neurotoxicity which is seen with decreased blood supply (Globus, 1995). Prolonged hyperthermia has been shown to decrease the number or function of mitochondria in some brain regions, suggesting decreased energy stores (Burrows, 1999). However, hyperthermia on its own does not selectively damage the serotonergic system.

Despite the apparent relationship between hyperthermia and MDMA neurotoxicity, it would be a simplification to think that avoiding hyperthermia ensures that humans who have taken MDMA will not undergo long-term serotonergic changes. Inducing hypothermia does not always completely block MDMA neurotoxicity (Broening, 1995). The link between temperature and neurotoxicity has been primarily investigated in rodents but has not been investigated in primates. Hypothermia does protect against methamphetamine-induced dopaminergic neurotoxicity in rodents (Ali et al. 1994; Miller and O'Callaghan 1994). However, the influence of temperature on neurotoxicity remains to be conclusively demonstrated in primates.


Time Course of Changes and Extent of Recovery

High doses of MDMA have a two-phase effect on serotonergic functioning, first causing acute decreases, then partial recovery, then chronic decreases. For example, after a single dose of 10 mg/kg MDMA to a rat, release of 5-HT leads to depletion of tissue levels of 5-HT and its metabolite 5-HIAA within 3 hours of dosing (Schmidt 1987; Stone, 1987b). Approximately 6 hours later, levels begin to return to normal, but this recovery is not sustained. About 24 hours after dosing, 5-HT levels begin a second, sustained decrease and remain significantly lower than baseline 2 weeks later. This sustained decrease is thought to be associated with axonal damage.

The intracellular enzyme TPH follows a similar time course, with decreased activity occurring within 15 minutes of drug administration. However, there is less short-term recovery of TPH activity in comparison to 5-HT. The recovery of TPH activity appears to involve regeneration of enzyme that was inactivated by oxidation rather than synthesis of new enzyme. SERT functioning is also altered. When rats were given 15 mg/kg subcutaneous MDMA and sacrificed an hour later, the uptake of serotonin was decreased by 80% (Fleckenstein, 1999). It should be noted that significant acute 5-HT depletions are not necessarily produced by all active doses of MDMA. Schmidt (1986) reported that 2.5 mg/kg MDMA did not produce an acute decrease in 5-HT or 5HIAA in Sprague Dawley rats at 3 hours after injection. Of note, Kish (2000) did find striatal 5-HT depletions in a chronic ecstasy user who died shortly after ecstasy ingestion. This suggests that at least some of the doses administered by humans are sufficient to produce 5-HT depletions.

The above description focuses on serotonergic changes because these are used to measure toxicity. Many other acute neurochemical changes occur after MDMA exposure. For example, dopamine is released (Stone, 1986) and dopamine transporter reuptake activity is decreased within 1 hr of high dose MDMA (Fleckenstein, 1999; Metzger, 1998). MDMA can also acutely increase dopamine synthesis (Nash, 1990). As noted previously, mice are selectively vulnerable to MDMA-induced dopaminergic neurotoxicity (Logan, 1988; Miller,1994; Stone, 1987a). In some studies, long-term alterations in dopaminergic functioning have been seen in other species (e.g., rats in Commins, 1987).

The time course of damaging events in rats can be seen by administering SSRIs, such as fluoxetine and citalopram, after MDMA. Pretreatment with fluoxetine (Prozac) or citalopram (Celexa) has been shown to block the neurotoxicity of MDMA (Battaglia, 1988; Schmidt 1987; 1990; Shankaran, 1999a), probably by blocking interactions of MDMA with SERT. More interestingly, fluoxetine remains almost fully protective if given 3 or 4 hours after MDMA. By 4 hours, most of the MDMA-induced release of 5-HT and DA has already occurred (Gough, 1991; Hiramatsu, 1990) and increases in extracellular free radicals (Colado, 1997b; Shankaran, 1999a) and lipid peroxidation (the alteration of fat molecules by free radicals) (Colado, 1997a) can be measured. Nevertheless, the administration of fluoxetine at this point decreases subsequent extracellular oxidative stress (Shankaran, 1999a) and long-term 5-HT depletions (Schmidt, 1987; Shankaran, 1999a). Fluoxetine will still be partially protective if given 6 hours after MDMA but has no protective effect 12 hours after administration (Schmidt, 1987). This shows that neurotoxic MDMA regimens initiate a series of events that become increasingly damaging between 3 and 12 hours after drug administration in rats.

Slow recovery of serotonergic functioning can be seen following a neurotoxic dose of MDMA. The extent of recovery is different in different species. In rats, there is extensive recovery of indicators of serotonergic functioning 1 year after drug exposure (Battaglia, 1988; Lew, 1996; Sabol, 1996; Scanzello, 1993), although there is significant variation in recovery between individual animals (Fischer, 1995). In primates, some recovery of serotonergic function occurs but is less extensive than in the rat. Altered serotonergic axon density was still detectable 7 years after MDMA exposure in one study of squirrel monkeys (Hatzidimitriou, 1999). Therefore, despite some recovery, MDMA-induced serotonergic changes are likely permanent in this primate species. This apparent species difference may be partially related to the more severe initial serotonergic damage usually seen in primates compared to rats, but also likely indicates a species difference in regrowth of serotonergic axons.


Behavioral and Functional Correlates of MDMA Exposure in Animals

A number of studies have looked for evidence that MDMA neurotoxicity causes lasting behavioral or functional changes in laboratory animals. These studies are summarized in Table I and are, perhaps, impressive for the limited nature of their behavioral findings. It is clear that neurotoxic MDMA exposure can both alter neurochemical functioning and the response of animals to subsequent drug exposures. However, so far only two published studies suggest that MDMA-exposed animals have behavioral alterations or functional impairments at seven or more days after last MDMA exposure.

Dafters (1998) demonstrated that MDMA-exposed animals have a lasting thermoregulatory impairment. Fourteen weeks after exposure to a neurotoxic MDMA or placebo regimen, rats were placed in a warm environment. MDMA-exposed rats had significantly larger increases in core temperature than control rats. It has been known for many years that individuals who experience heat stroke have increased susceptibility to subsequent episodes for some time (Shapiro, 1979) and it appears possible that the same phenomenon is being detected here.

Another study has suggested that neurotoxic MDMA exposure may cause cognitive impairment in rats. Marston (1999) detected drug-free alterations in performance of a delayed memory task. In contrast, Ricaurte (1993) and Robinson (1993) were unable to demonstrate any long-term effect of MDMA neurotoxicity on spatial navigation memory tasks in rats. However, Robinson did detect short-term residual effects of MDMA on this task when animals were tested 2 days after the last MDMA exposure.

The cautious interpretation of behavioral animal studies of MDMA neurotoxicity is that we should not expect gross behavioral effects of MDMA neurotoxicity in humans, even when extensive serotonergic changes have occurred. It should also be remembered that we poorly understand the role of 5-HT in the brain (reviewed in Lucki, 1998) and that this makes it more difficult to detect 5-HT-related changes. Findings from studies of ecstasy users may allow more focused and hypothesis-driven studies of animals.

Table I: Studies of Long-term Changes after Neurotoxic MDMA Regimens in Nonhuman Animals
Species and Strain MDMA Regimen Significant Differences in MDMA-treated animals Measures showing no significant difference Reference
Rhesus Monkeys 10 mg/kg IM, twice a day, for 4 days Right shift in MDMA and d-fenfluramine dose-response curve for time estimation,  learning task, and motivation tasks at post 1 mo. Baseline performance on all tasks. Frederick et al., 1998
Rhesus Monkeys Escalating doses of 0.10, 0.3, 1.0, 1.75, 3.0, 5.6, 7.5, 10.0, 15.0, and 20 mg/kg, IM, twice daily for 14 consecutive days at each dose. Right shift in MDMA dose-response curve for time estimation, short-term memory, color and position discrimination, and motivation tasks at post 21 mo. Baseline performance on all tasks. Frederick et al., 1995
Rats, Sprague-Dawley 20 mg/kg SC, twice a day for 4 days None, although researchers note that 2 of 8 MDMA-exposed rats failed to acquire lever pressing with 20 sec reinforcement delays during the 8 hr session. Acquisition of and behavior on a lever-press responding task at post 14 days. Byrne, Baker, & Poling 2000
Rats, Sprague-Dawley 10 mg/kg SC, twice a day for 4 days Significant pretreatment x treatment x crossing times interaction, suggesting altered S-MDMA -induced behavioral activation at post 21 days. Drug-free locomotion at 21 days; RU24969-induced behavioral activation at 21 days. Callaway & Geyer 1992
Rats, Wistar 10 mg/kg SC per day for 4 days Increased core temperature when placed in either 22 °C or 28 °C ambient temperature at post 4 or 14 wks. None Dafters & Lynch 1998
Rats, Long-Evans 40 mg/kg SC, twice a day for 4 days None Sexual behaviors at post 10 days; Spontaneous motor activity. Dornan, Katz, & Ricaurte 1991
Rats, Sprague-Dawley 20 mg/kg SC, twice a day for 4 days Decreased electrical-stimulated 5-HT release in DRN at post 2 wks. Electrical-stimulated 5-HT release in MRN or hippocampus at post 2 wks; Number and firing pattern of classical 5-HT neurons and burst-firing neurons in DRN. Gartside, McQuade, & Sharp 1996
Rats, Sprague-Dawley 20 mg/kg SC, twice a day for 4 days None DOI-induced head twitch responses, locomotion, and rearing activity. Granoff & Ashby 1998
Rats, Sprague-Dawley 20 mg/kg SC, twice a day for 4 days Increased conditioned place preference response to cocaine in MDMA group at 2 post wks. Horan, Gardner, & Ashby 2000
Rats, Sprague-Dawley 5 mg/kg sc once per day or 20 mg/kg SC, twice a day for 4 days, followed by 5 mg/kg MDMA 2 days later Increased motor stimulant effects of 5.0 mg/kg SC MDMA in both MDMA-treated groups at post 11 days; Increased motor stimulant effects of 15.0 mg/kg IP cocaine in both MDMA-treated groups at post 11 days; Increased MDMA-stimulated DA release in the nucleus Basal DA in nucleus accumbens at post 2 wks. Kalivas, Duffy, White 1998
Rats, Sprague-Dawley 15 mg/kg IP Loss of rate-dependence of response of nigrostriatal cells to either quipazine or apomorphine at post 1 wk. Basal activity of nigrostriatal DA neurons; Quipazine-induced inhibition of nigrostriatal DA cell firing for all cells at post 1 wk. Kelland, Freeman, & Chiodo 1989
Rats, Sprague-Dawley 6 mg/kg SC, twice a daily for 4 days Left shift in MDMA dose-response curve on DRL task in MDMA group. None Li et al., 1989
Rats, Lister Hooded ascending regimen of 10, 15, and 20 mg/kg IP, each dose given twice daily for one day Decreased performance in operant delayed match to nonsample task. Spontaneous behavior, body temperature, and skilled paw reach ("staircase task"). Marston et al., 1999
Rats, Sprague-Dawley 20 mg/kg SC, twice a day for 4 days Increased cocaine-induced dopamine release in nucleus accumbens in MDMA group at 2 wks after neurotoxic regimen. None Morgan et al., 1997
Rats, Sprague-Dawley 20 mg/kg SC, twice a day for 4 days Increased morphine-induced antinociception (assessed by tail flick test) at post 2 wks. Baseline behavior in tail flick test. Nencini, Woolverton, & Seiden 1988
Rats, Sprague-Dawley 20 mg/kg SC, twice a day for 4 days Decreased inhibitory effects of DA and SKF38393 on glutamate-evoked firing in nucleus accumbens cells at post 9-15 days. Inhibitory effects of GABA on glutamate-evoked firing in nucleus accumbens cells at post 9-15 days. Obradovic, Imel, & White 1998
Rats, Sprague-Dawley 20 mg/kg SC Increased 8-OH-DPAT-induced prolactin release at post 2 weeks.  Decreased 8-OH-DPAT-stimulated ACTH release at 2 weeks. Basal ACTH and prolactin concentrations and ACTH and prolactin response to saline injection. Poland 1990
Rats, Sprague-Dawley 20 mg/kg SC Increased d,l-Fenfluramine-stimulated prolactin release at post 2 and 4 months.  Decreased d,l-Fenfluramine-stimulated ACTH release at 2-8 months. d,l-Fenfluramine-stimulated ACTH at 12 months; d,l-Fenfluramine-stimulated prolactin at 8 and 12 months. Poland et al., 1997
Rats, Sprague-Dawley 20 mg/kg SC, twice a day for 4 days Increased d,l-Fenfluramine-stimulated prolactin release at post 4 and 8 months.  Decreased d,l-Fenfluramine-stimulated ACTH release at post 4, 8, and 12 months. Saline-stimulated ACTH and prolactin release at post 2 weeks; d,l-Fenfluramine-stimulated prolactin release at post 12 months. Poland et al., 1997
Rats, Lister Hooded 20 mg/kg SC, twice a day for 4 days with entire regimen repeated 2 wks later None Performance in a spatial memory task using a T-maze and scopolamine-induced changes in performance on this task. Ricaurte et al., 1993
Rats, Sprague-Dawley 10 mg/kg IP, twice a day for 4 days Increased time to find hidden platfrom in first trial of spatial navigation task at post 2 days. Spatial navigation task after 1st trial, skilled reaching task, place navigation learning-set task, foraging task, with or without atropine pretreatment. Robinson, Castaneda, & Whishaw 1993
Rats, Sprague-Dawley 20 mg/kg SC twice a day for 4 days Decreased discrimination of 1.0 mg/kg MDMA from saline at post 13-15 days. Discrimination of 0.5 or 1.5 mg/kg; Conditioned place preference from MDMA. Schechter 1991
Rats, Sprague-Dawley 20 or 40 mg/kg SC twice a day for 4 days Decreased d-fenfluramine-stimulated 5-HT release in frontal cortex at post 2 wks. None Series, Cowen, & Sharp 1994
Rats, Sprague-Dawley 10 mg/kg IP, every 2 h for 4 injections Decreased behavioral, hyperthermic, and 5-HT-releasing effects of MDMA at 1 wk after neurotoxic regimen. None Shankaran & Gudelsky 1999
Rats, Sprague-Dawley 20 or 40 mg/kg SC twice a day for 4 days Increased cerebral glucose utilization in molecular layer of dentate gyrus and in CA2 and CA3 fields of Ammon's horn in hippocampus at post 14 days. Cerebral glucose utilization in neocortex, raphe nuclei, and some hippocampal areas at post 14 days. Sharkey, McBean, & Kelly 1991
Rats, Sprague-Dawley 5 or 10 mg/kg PO daily for 4 days None Auditory startle, emergence from darkened chamber, complex maze navigation, response to hot plate, FI 90 operant behavioral task at post 2 to 4 weeks.. Slikker et al., 1989
Rat Pups, Sprague-Dawley 10 mg/kg SC every 12 hrs for 4 or 7 injections Decreased rate of ultrasonic vocalization measured up to post 11 days. Behavioral responses to the 5-HT1a agonist 8-OH-DPAT, the 5-HT1b agonist TFMPP, and the 5-HT2 agonist DOI at post 8 days. Winslow & Insel 1990



Studies Comparing Ecstasy Users and Nonusers

Over 35 studies have been published retrospectively comparing illicit ecstasy users to nonusers. Before discussing the findings of these studies (reviewed in Morgan,2000), it is worth discussing their limitations. Retrospective studies are difficult to interpret since it is always possible that there were pre-existing differences between the users and nonusers. It is almost trivial to suggest that frequent users of illicit drugs are different from those who do not use drugs. Thus, one might evaluate studies by considering to what extent they differentiate between typical characteristics of frequent illicit drug users and those specifically associated with ecstasy use. [Editor's note: It is also important to realize that ecstasy users may not be ingesting MDMA alone or sometimes at all, as there is no guarantee that purchased ecstasy contains MDMA. It is possible that polydrug use could contribute to any detected problems.] Other frequent methodological limitations of these studies include: poorly described recruitment and matching of volunteer groups; reliance on self-reports of drug use; failure to separate residual effects of recent drug use from long-term effects; use of the same volunteers in multiple publications; and the difficulty relating serotonergic differences to toxicity. Despite these limitations, some conclusions can be drawn from studies comparing ecstasy users and nonusers. Findings can be grouped into personality, neurofunctional, and cognitive performance differences. These areas are discussed below.

Consistent reports link repeated ecstasy use to depressed mood (Cohen 1995; Curran, 1997; Davison, 1997; Gamma, 2000; Gerra, 2000;1998; Morgan, 1999; Parrott 2000; 1998; Solowij, 1992). Because dysphoric mood is a known residual effect of other psychostimulant drugs (Coffey, 2000), it is likely that ecstasy use plays a causal role in this phenomenon. In a survey of 158 polydrug users, Williamson (1997) found that similar numbers of users reported depression, anxiety, and related adverse effects after cocaine as compared to MDMA. Thus, in some ways, MDMA is very similar to other psychostimulants.

In addition, there are a number of case reports of psychiatric disorders, such as psychosis, depression, and panic attacks in ecstasy users (reviewed in McGuire, 2000). Given that other psychostimulants are associated with psychiatric disorders in illicit users, it would not be surprising if this were also true of MDMA. For example, it is well established that stimulant-induced psychosis can occur in cocaine or methamphetamine users (Angrist, 1994). Reports of MDMA-related psychosis have also been published (Creighton, 1991; McCann, 1991a; McGuire, 2000; 1991). These psychiatric disorders need not be related to the selective neurotoxicity discussed in this chapter. For example, methamphetamine can produce chronic behavioral disturbances resembling psychosis in primates using regimens that are not neurotoxic to dopaminergic or serotonergic systems (Castner, 1999).


Personality Differences between Ecstasy users and nonusers

While ecstasy users have sometimes been found to have different personalities than nonusers, it is not clear that this is an effect of MDMA exposure. Many of the reported personality differences between ecstasy users and nonuser volunteers who do not use illicit drugs likely reflect preexisting differences. Increased novelty-seeking (Gerra, 1998), venturesomeness and impulsivity (Morgan, 1998) have been reported in ecstasy users, but this can be expected in users of illicit drugs compared to nonusers. The possibility of preexisting differences has been pointed out by several authors. For example, Gerra (2000) suggested that the enhanced novelty seeking (measured with the self-report Tridimensional Personality Questionnaire) in ecstasy users undergoing substance abuse treatment reflected a preexisting trait. Similarly, the increased Buss-Durkee Hostility Index (BDHI) direct aggression scores of ecstasy users in substance abuse treatment (Gerra, 2000) and the decreased BDHI indirect hostility scores in untreated ecstasy users (McCann, 1994) may be partially explained by social circumstances and subcultural values, respectively. In order to reduce the influence of traits generally associated with illicit drug use, one could compare ecstasy users with different total ecstasy exposures or compare polydrug users with and withoutecstasy experience. Findings from these comparisons are ambiguous and, at best, such comparisons can provide only limited support for possible MDMA-induced alterations in personality. Studies in which the same individuals are examined at different time points are necessary to properly examine this issue.

There is mixed evidence that MDMA use is associated with increases in self-reported impulsivity. Morgan (1998) reported that a post hoc comparison of more (30+ tablets ingested) and less experienced (20 - 30 tablets ingested) ecstasy users revealed heightened impulsivity (measured with Eysenck's self-report IVE questionnaire) in the more experienced group. Parrott (2000) reported a non-significant trend towards greater IVE impulsivity in polydrug-usingecstasy users with an average of 371 (30-1000) exposures compared to a group of users with an average of 6.8 (1-20) exposures. Tuchtenhagen (2000) found that ecstasy users with an average of 93.4 ± 119.9 (20-500) exposures has significantly higher scores for the nonplanning impulsivity (measured with the self-report Barratt Impulsiveness Scale) compared to controls matched for other drug use. The researchers also noted a trend towards increased experience seeking (measured with the self-report Sensation Seeking Scale) which reached statistical significance only when ecstasy users were compared to nonusers. These findings differ from those of McCann (1994) who compared ecstasy users, with an average of 94.4 +/- 90.6 (25-300) reported ecstasy exposures, to nonusers (without controlling for other drug use). McCann, reported decreased impulsivity (measured as increases in the Control subscale of the Multidimensional Personality Questionnaire) but failed to find significant differences in self-reported impulsivity with a second questionnaire (the self-report Eysenck Personality Questionnaire).

There are less data examining behavioral impulsivity, which is thought to be different from self-reported impulsivity (Evenden, 1999). Gouzoulis-Mayfrank (2000), using the same volunteers as in the Tuchtenhagen (2000) report, did not find evidence of behavioral impulsivity in ecstasy users undergoing a cognitive test battery. In contrast, Morgan (1998) reported that ecstasy users made increased errors in a Matching Familiar Figures task, a difference he interpreted as evidence of increased impulsivity. Morgan suggested his behavioral findings indicated a decreased capacity to cope with high levels of cognitive demands.


Neurofunctional Differences between ecstasy users and nonusers

Studies have also established an association between ecstasy exposure and altered neurofunctioning. Reported neurofunctional differences are summarized in Table II and include putative serotonergic measures as well as more general measures, such as EEG. While retrospective studies cannot, technically speaking, establish causality, many of these user-nonuser differences correlate with extent of ecstasy exposure. Correlations have been reported between ecstasy exposure and measures such as cerebral spinal fluid (CSF) 5HIAA levels (Bolla, 1998), SERT density (McCann, 1998), brain myo-inositol increases (Chang, 1999), and EEG alterations (Dafters, 1999). A primary difficulty in interpreting these studies is that we do not really know what many of these neurofunctional differences mean.

At the minimum, we can certainly conclude that the brains of these ecstasy users are different from those of non-user volunteers. Does this mean that serotonergic neurotoxicity has taken place? This seems the most likely possibility. Several studies have shown differences in measures of serotonergic functioning between users and nonusers. Two groups have reported decreased cortical SERT binding in ecstasy users (McCann, 1998; Semple, 1999), although there is some question about the specificity of the measurement technique (Heinz, 2000; Kuikka, 1999). Three of four studies have found CSF levels of 5HIAA to be lower in users than non-users (decreased in McCann, 1999b; 1994; Ricaurte, 1990; unchanged in Peroutka, 1987). These differences are consistent with animal studies in which neurotoxic MDMA exposure similarly altered these indicators (Insel, 1989; Ricaurte, 1988b; Scheffel, 1998).

Such parallel findings in humans and nonhumans provide some evidence that selective serotonergic neurotoxicity has occurred. However, all published studies in humans have been retrospective. Without knowing what ecstasy users were like before using drugs, we can only guess whether unusual serotonergic functioning is the result of damage. Unfortunately, these serotonergic measures are sufficiently new that we do not know the full range of "normal" values for them. It is therefore difficult to decide whether the values seen in ecstasy users are truly 'abnormal' and indicative of damage. Alternatively, they may be simply 'unusual' for non-drug users but 'usual' for the kind of person who is likely to use ecstasy repeatedly. [Editor's note: There are many complicating factors in measuring 5HT levels in living humans including such as people with depression have been shown to have decreased numbers of SERT in other areas of the brain (Mann, 2000) and decreased levels of 5HT and 5HIAA (Meltzer, 1990) prior to being treated with antidepressants.] In addition, typical indicators of serotonergic function may be affected by influences other than neurotoxicity. Some theories suggest that individuals who abuse psychostimulants are more likely to have unusual serotonergenic functioning (Laviola, 1999; Zuckerman 1996). These interpretive difficulties can be illustrated using studies that investigate the amount of hormone released after serotonergic drug administration in different populations.

Measuring the amount of hormone released in response to a serotonergic drug is one way to test for changes in the serotonergic system. This tactic has uncovered statistically significant user-nonuser differences in 4 of 6 studies (differences detected in (Gerra, 2000; 1998; McCann, 1999a; Verkes, 2000; no significant differences in McCann, 1994; Price, 1989). However, other studies have established that both personality and use of other drugs, such as cocaine, may modulate this serotonergic measure. High sensation-seeking humans have been shown to have blunted hormone response to the partial 5-HT1a agonist, ipsapirone (Netter, 1996). Similarly, the prolactin response to the 5-HT releaser, fenfluramine, in a group of cocaine-dependent individuals was significantly increased between the first and third weeks after discontinuing cocaine use (Buydens-Branchey, 1999), suggesting recovery from cocaine-induced alterations. Therefore, one could argue that factors other than MDMA neurotoxicity still might explain some apparently serotonergic differences between users and nonusers. This issue can only by solved using prospective studies that assess the same individuals at different time points.

One strong argument that MDMA neurotoxicity occurs in many human users is simply that estimated doses ingested by some users exceed those known to produce 5-HT depletions in squirrel monkeys (Ricaurte, 1988a). Given that approximately similar doses are associated with similar changes in serotonergic indices in nonhumans and humans, it seems likely that the same phenomenon is occurring in both species. Furthermore, if one is considering administering MDMA to humans, it may be more important to be conservative in risk assessment than to wait for conclusive scientific proof of neurotoxicity. This is especially important because some individuals may be more susceptible to neurotoxicity than others. Studies comparing ecstasy users to non-users suggest that neurotoxicity may occur with MDMA exposures that are self-administered by humans. MDMA neurotoxicity and its largely unknown possible long-term consequences must therefore be considered when evaluating the risks of clinical MDMA research.


Cognitive Differences between ecstasy users and nonusers

Repeated ecstasy exposure is associated with decreased performance on cognitive tests. Tests of declarative verbal memory have been frequently used to detect this decrease (Gouzoulis-Mayfrank, 2000; Morgan 1999; Parrott, 1998a; 1998b; Reneman, 2000a). However, user-nonuser differences have been detected with a broad range of cognitive tasks (Gouzoulis-Mayfrank, 2000; McCann, 1999b; Rodgers 2000). Some have suggested that specific alterations in executive functioning and working memory may explain the observed differences (Dafters, 1999; Gouzoulis-Mayfrank, 2000; Wareing, 2000), but evidence for this is not yet conclusive.

Perhaps the most thorough study published so far was conducted by Gouzoulis-Mayfrank (2000). In this study, users of both ecstasy and cannabis were compared to cannabis users and drug-free volunteers. Extent of ecstasy use was correlated with decreased performance in a range of tasks. Performance in ecstasy-using volunteers remained, on the average, in the low end of clinically normal functioning. However, this is not particularly reassuring given that these users appeared to have fairly common use patterns (1.4 +/- 0.9 tablets taken 2.4 +/- 1.6 times per month). If modestly decreased cognitive performance is an effect of MDMA, it is likely one experienced by many individuals.

Does ecstasy use cause this poor ognitive performance? The current data are inconclusive but suggest the answer is "yes." Many (but not all, e.g., Morgan 1998) studies have found that repeated ecstasy users perform worse in many cognitive tests than nonusers and that users with more ecstasy exposure perform worse than those with less exposure (Bolla, 1998; Dafters, 1999; Gouzoulis-Mayfrank, 2000; McCann, 1999b).

It is likely that there are differences between ecstasy users and nonusers that predate illicit drug use. Schifano (2000) recently described currently unpublished survey data from high school students in Italy which found that students attending less academic secondary schools were 2.89 times more likely to have used ecstasy than those attending more academic schools. In another survey of 737 Italian ecstasy users, there was evidence of inverse relationships between the tendency to take higher ecstasy doses and both lower schooling level and family income (Schifano, 2000).

The association between ecstasy exposure and lower cognitive performance may also be partially caused by factors correlated with ecstasy exposure, such as repeated sleep and nutrient deprivation associated with attending late-night dance events. Nonetheless, the few scientific studies on these other possible factors (Cho, 2000; Dinges,1991; Kretsch, 1997) would not lead us to expect an effect comparable to what we see in studies of ecstasy users. These other possible factors seem likely to be significant only if the ecstasy-using volunteers in these international studies engage in a particularly 'hard-partying' lifestyle. In the first published study that properly controlled for lifestyle, Verkes (2000) found that 'moderate' ecstasy users (with 73 +/- 68 reported exposures to ecstasy) had lower performance scores than nonusers attending a similar number of 'raves' in the previous 12 months.

Pre-existing differences and effects of lifestyle seem unlikely to fully explain the reported cognitive performance differences. Average performance in immediate declarative verbal memory tasks was decreased by about 0.8 standard deviation units in several studies (Gouzoulis-Mayfrank, 2000; Morgan 1999; Parrott, 1998a; 1998b). This means that the average ecstasy-using volunteer in these studies scored in the bottom 21% of what was expected based on the comparison volunteers. While possible, it seems improbable that primarily the quarter of the population with the worst memory goes on to use ecstasy several times a month (and participates in these studies).

Use of drugs other than MDMA has not always been properly taken into account in studies of ecstasy users. In particular, cannabis use has often been greater in ecstasy-using volunteers than in ecstasy-naïve volunteers. This is significant because chronic cannabis use can cause long-lasting residual decreases in cognitive performance (Pope,1996). Three studies have compared users of both ecstasy and cannabis to users of cannabis alone (Croft, 2000; Gouzoulis-Mayfrank, 2000; Rodgers 2000). Two of these studies have suggested that MDMA is associated with lowered cognitive performance beyond that expected for cannabis (Gouzoulis-Mayfrank, 2000; Rodgers 2000). In contrast, Croft (2000) was unable to detect performance differences between cannabis users and users of both cannabis and ecstasy using a battery of cognitive tests. Furthermore, covariate analysis suggested that performance decreases were more closely related to cannabis than ecstasy use. In another study that attempted to control for the influence of other drugs, Morgan (1999) detected lower memory performance in ecstasy-experienced polydrug users compared to ecstasy-naïve polydrug users. However, matching of drug use between comparison groups was imperfect in this study. It is clear that future studies should control for use of cannabis and that the apparent magnitude of the MDMA-associated cognitive performance decrease is likely exaggerated by cannabis use.

The lower ocognitive performance of ecstasy users may be due to serotonergic neurotoxicity or some other neurochemical alteration. It has been demonstrated that acute serotonergic depletion (by dietary manipulation) can impair declarative verbal memory in healthy volunteers (Riedel, 1999). Two studies of ecstasy users have reported correlations between alterations in serotonergic measures and decreased cognitive performance (Bolla, 1998; Reneman, 2000a; Verkes, 2000). This suggests a relationship between lower cognitive performance and MDMA-induced serotonin depletions or neurotoxicity. On the other hand, if MDMA-induced loss of serotonin or damage to serotonergic axons were sufficient to impair memory to the degree suggested by human studies, one would expect this effect to have been readily detected in prospective animal studies.

It appears possible that the reported lower cognitive performance is related to the volunteers' chronic, repeated patterns of ecstasy use. Because MDMA exposures are limited (usually 4 consecutive days or less) in most animal experiments, this could explain the apparent discrepancy between these studies and ecstasy user studies. Furthermore, it is well established that chronic psychostimulant use lowers cognitive performance (McKetin,1999; Ornstein, 2000). For example, repeated cocaine use is associated with impaired cognitive functioning (Beatty, 1995; Bolla, 1999; O'Malley, 1992), although cocaine use per se does not necessarily produce deficits (Bolla, 1999). Cocaine is not a selective neurotoxin but, like MDMA, can cause both serotonergic (Jacobsen, 2000; Little, 1998) and cerebrovascular (Bartzokis, 1999; Herning, 1999) alterations. Since repeated exposure to other psychostimulants can impair cognitive functioning, it is credible that repeated MDMA use might be associated with cognitive deficits. Suggesting this leaves open the question of whether this effect is due to repeated neurotoxic damage or residual drug effects.

Specific evidence linking the lower cognitive performance of repeated ecstasy users to serotonergic neurotoxicity could come from studies of the time course of these differences. Residual drug effects might be expected to improve more quickly than changes due to serotonergic neurotoxicity. Unfortunately, too few studies have looked for evidence of recovery to draw any conclusions. Morgan (1999) reported that a subset of three ecstasy users who had not taken ecstasy in over 6 months had significantly better immediate and delayed recall (of ideas from stories taken from the Rivermead Behavioral Memory Test) than users with more recent use. In contrast, Wareing (2000) were unable to find evidence of a significant abstinence-related improvement in working memory and executive functioning tasks when 10 current ecstasy users were compared to 10 volunteers who reportedly had not used ecstasy in 6 months. It is therefore not clear if there is recovery from this lower cognitive performance.

In conclusion, repeated ecstasy exposure is associated with lowered cognitive performance. The apparent magnitude of the effect may be exaggerated by limitations in published studies, particularly the confounding effects of cannabis [and perhaps other substances] on performance. There are insufficient data to decide whether there is recovery of performance with abstinence. The question also remains open as to whether this is due to a residual drug effect or a frank neurotoxic change.

Table II: Reported Neurofunctional Differences Between Ecstasy Users and Nonusers
Measure Selective for Serotonergic Differences? Relevant Animal Literature? Correlated with MDMA Exposure? Evidence for Recovery? References
Putative Serotonergic Measures
Decreased CSF 5-HIAA in 3 of 4 studies Yes Decreased up to 2 weeks after MDMA in squirrel monkeys (Ricaurte et al., 1988) and 14 weeks after MDMA in rhesus monkeys (Insel et al., 1989). No No Decreased in McCann et al.,1999, 1994; Ricaurte et al.,1990.  Unchanged in Peroutka et al., 1989
Decreased then Increased 5-HT2a receptor density in 1 of 1 studies Yes Decreased at 24 hr, normal at 21 d after MDMA in rats (Scheffel et al., 1992). Yes Not reported Increased in Reneman et al., 2000a; Decreased then increased in Reneman et al., 2000b
Decreased neuroendocrine response to serotonergic drugs, in 3 of 5 studies Yes Increased at 2 months, normal at 12 months in rats (Poland et al., 1991) Yes in Gerra et al.,  2000 No Decreased in Gerra et al., 2000, 1998; McCann et al., 1999a.  Unchanged in Price et al., 1989; McCann et al., 1994.
Decreased SERT density, estimated with PET, in 2 of 2 studies Disputed - ligand kinetics  may be  altered by other changes (Kuikka & Ahonen 1998). PET measures apparently decreased in one baboon up to 14 weeks after MDMA (Scheffel et al., 1998). Yes, though McCann included controls. Mixed  (Yes in Semple; No in McCann) Semple et al., 1999; McCann et al., 1998.
Increased stimulus dependence for ERP EEG N1/P2 amplitudes in 1 of 1 studies Disputed – 5HT depletion did not change measure in one study (Dierks et al., 1999). Unknown No Not reported Tuchtenhagen et al., 2000
Nonspecific Neurofunctional Measures
Increased brain myo-inositol measured as by 1H MRS in 1 of 1 studies - Unknown Yes Not reported Chang et al., 1999
Decreased total sleep time (non-REM and stage 2 sleep) in 1 of 1 study - Unknown No Not reported Allen et al., 1993
Altered cerebral blood flow or volume in 2 of 4 studies - Increased blood flow 6-9 wks after MDA in rats (McBean et al., 1990) Yes Yes Altered in Reneman et al., 2000b and in Chang et al., 2000 (prospective study).  Unchanged in Chang et al., 2000 (user-nonuser study) and Gamma et al., in press.
Decreased cerebral glucose utilization in 1 of 1 studies - Increased in some hipppocampal areas 2 wks (Sharkey et al., 1991) after MDMA and 6-9 wks after MDA (McBean et al., 1990) in rats No No Obrocki et al., 1999
Increased alpha and beta EEG power in 2 of 2 studies - Unknown Yes Not reported Dafters et al., 1999; Gamma et al., 2000



Possible Significance of Cognitive Differences and MDMA Neurotoxicity

How severe are these cognitive changes? They do not indicate impairment in day-to-day activities. The differences occur in cognitive tests in which young, healthy people perform well. Thus, these differences are generally small in magnitude despite their statistical significance. In fact, neither the investigators nor the ecstasy-using volunteers themselves appear to be aware of any cognitive impairment in these individuals (McCann, 1999b; Rodgers, 2000). These studies raise questions about whether these ecstasy-using volunteers have experienced serotonergic neurotoxicity that might eventually be associated with more severe symptoms. Such symptoms could become prominent as ecstasyusers age. Additionally, larger impairments in specialized areas of functioning may exist but simply have not been discovered yet.

Studies of individual variation in symptoms associated with neurodegenerative disorders have lead to two relevant concepts. First, there is a threshold of damage that must be exceeded in some brain systems before symptoms develop. This has been primarily investigated with dopaminergic cell loss and Parkinson's disease (Brownell, 1999; Calne, 1985; Di Monte, 2000). There are less data on the serotonergic system. In a rat study using the serotonergic neurotoxin, 5,7-DHT, Hall (1999) concluded that a loss of greater than 60% of serotonergic neurons was necessary to decrease extracellular 5-HT levels in the striatum. Alterations in behavior were seen with slightly smaller depletions (51% or more), possibly due to regional variations in neurotoxicity. One might speculate that even smaller depletions may not affect many serotonergic-related behaviors, although the maximal serotonergic response to drugs or other stimuli is likely to be reduced (reduced electrically-stimulated 5-HT release in MDMA-exposed rats was documented by Gartside, 1996).

Second, the concept of cognitive reserve has been developed to explain why greater education, intelligence, or brain size is associated with less severe impairment in conditions such as Alzheimer's disease, AIDS, and normal aging (Alexander, 1997; Coffey, 1999; Graves, 1996; Stern, 1996). This cognitive reserve may be seen as a surplus of processing capacity that protects the individual against loss of functioning when that capacity is decreased. Cognitive reserve could be the result of more extensive functional brain tissue, density of neural connections, or cognitive strategies for problem solving. Individuals with less cognitive reserve could be expected to undergo larger cognitive decreases from MDMA exposure than users with greater cognitive reserve. Support for this possibility comes from Bolla (1998) who reported a significant interaction between dose and vocabulary (measured with the WAIS-R). ecstasy users with lower vocabulary scores showed greater decreases in delayed visual memory performance, while users with higher vocabulary had largely preserved performance. Although the absolute magnitude of performance decrease was small, this study suggests that cognitive reserve could play a role in expression of MDMA neurotoxicity.

Whether symptoms of MDMA neurotoxicity are likely to increase as users age is difficult to predict. Some have speculated that aging ecstasy users might have increased risk of depression and other affective disorders. From a neurochemical perspective, age-related decrease in SERT density appears modest (estimated at 4.3% per decade in one recent study (van Dyck, 2000)), while 5-HT receptors undergo more complex age-related changes (reviewed in (Meltzer, 1998). One would hope that these changes will not cause ecstasy users to exceed a hypothetical threshold for developing symptoms of neurotoxicity. However, we simply do not understand 5-HT or affective disorders sufficiently to make predictions with any confidence. Late onset affective disorders are probably influenced by many nonserotonergic factors, such as social isolation and cerebrovascular disease.

These are serious and legitimate concerns and there is insufficient research to adequately address them. On the other hand, there is no direct evidence to support these concerns. Neurotoxic phenethylamines have been self-administered by humans for over 60 years. In this time, no evidence has been published suggesting that methamphetamine or amphetamine increase risk of Parkinson's disease, despite damaging dopaminergic axons. In contrast, the link between Parkinson's disease and MPTP, a meperidine analogue and dopaminergic neurotoxin that destroys cell bodies, was rapidly discovered (Davis, 1979; Langston, 1983). This suggests that there may be fundamental differences between neurotoxic phenethylamines, which selectively damage a subset of monoaminergic axons but not cell bodies, and other neurotoxins. Similarly, concerns about the selective serotonergic neurotoxicity induced by MDMA and other drugs are not fueled by a toxic syndrome identified in users. Instead, they are motivated by the intuition that the dramatic decreases in indices of serotonergic functioning must have some adverse behavioral consequences.

Some have suggested that MDMA neurotoxicity may be related to its putative therapeutic effects. Although this is technically possible, there are a number of reasons to doubt this hypothesis. The acute intoxication induced by MDMA is unusual. In contrast, similar serotonergic neurotoxicity can be produced by many other drugs. The events associated with MDMA neurotoxicity occur in rats between approximately 3 and 12 hours after drug administration, when subjective effects are decreasing or absent in humans. Thus, the acute intoxication produced by MDMA appears to be fully separable from long-term serotonergic effects. If MDMA proves useful as an adjunct to psychotherapy, it seems more likely that this utility will be associated with the unusual acute intoxication produced by MDMA than with the chronic serotonergic changes produced by many drugs.


Findings in Prospective Clinical MDMA Studies

Few peer-reviewed reports are available that examine volunteers in clinical MDMA studies for evidence of neurotoxic changes. This section therefore significantly relies on unpublished data kindly supplied by researchers who are in the process of preparing reports on their findings. The reader is advised to consider this discussion as preliminary and subject to revision in the more definitive peer-reviewed publications from these researchers.

Preliminary retrospective analysis of data from studies conducted by Dr. Franz Vollenweider and colleagues has reportedly found no evidence that one or two oral exposures of up to 1.7 mg/kg MDMA is associated with lasting cognitive or neurofunctional alterations. Measures in this retrospective analysis include EEG, regional cerebral blood flow, mood, cognitive tests, and indices of information processing such as event related EEG potentials and prepulse inhibition (Dr. Franz Vollenweider, personal communication).

Most importantly, Vollenweider and colleagues conducted a prospective study in which six MDMA-naïve volunteers were administered a single oral dose of 1.5 or 1.7 mg/kg MDMA. PET measures of SERT density (using the same ligand employed by McCann, 1998) were made before and four weeks after MDMA administration. No significant changes were noted. Thus, it would appear that long-term serotonergic changes either do not occur or are too small to measure using this technique after one exposure to up to 1.7 mg/kg MDMA in healthy volunteers. However, these data will need to be replicated with a larger sample size before this conclusion can be made with confidence.

Data collected by Dr. Charles Grob and colleagues are more difficult to interpret. These researchers administered two doses (separated by two weeks or more) of up to 2.5 mg/kg MDMA to ecstasy-experienced volunteers, carrying out cognitive testing in 14 of these individuals before and approximately two weeks after study participation. No alterations in cognitive performance were detected (Dr. Charles Grob, personal communication). However, MDMA-induced decreases in regional cerebral blood flow occurred in a subset of eight volunteers assessed 10 to 21 days after last MDMA exposure (Chang, 2000). Cerebral blood flow was measured using [99mTc]-HMPAO SPECT co-registered with MRI and significant decreases were found bilaterally in the visual cortex, caudate, superior parietal, and dorsolateral frontal regions. Therefore, doses as low as 1.25 mg/kg MDMA may decrease cerebral blood flow at 2 or 3 weeks after drug exposure.

How long do these decreases last? This is not clear. Two volunteers who underwent repeated SPECT scans showed evidence of possibly increased cerebral blood flow at later time points (43 and 80 days after MDMA, respectively). This suggests that the decreased cerebral blood flow is either a subacute drug effect of limited duration or part of a lasting biphasic effect (with decreases followed by increases). Chang states that decreased regional cerebral blood flow was generally less in volunteers with greater time from last MDMA exposure, providing evidence of recovery. In addition, the authors did not find differences in cerebral blood flow when 21 ecstasy-experienced volunteers were compared to 21 nonusers (in press). Similarly, Gamma (2001) saw no significant differences between 16 ecstasy users (most of whom had used ecstasy at least 100 times) and 17 nonusers when regional cerebral blood flow was measured during a vigilance task using [H2 15O]-PET. Finally, it should also be pointed out that Vollenweider and colleagues reportedly did not detect changes in regional cerebral blood flow using [H2 15O]-PET in a retrospective analysis of a study in which volunteers received 1.7 mg/kg MDMA (Dr. Alex Gamma, personal communication).

One possible mechanism for subacute alterations in regional cerebral blood flow is suggested by two preliminary reports of a study by Dr. Liesbeth Reneman and colleagues (Reneman, 2000a; 2000b). These researchers used [123I] R91150 SPECT to measure cortical 5-HT2a receptors and found evidence of decreased 5-HT2a receptors in ten ecstasy users with 7 ± 5 weeks from last ecstasy exposure. In contrast, a group of five ecstasy users with at least two months from reported last exposure (18 +/- 15 weeks) showed a trend toward increased cortical 5-HT2a binding which did not reach statistical significance. Reneman, suggest that MDMA-induced 5-HT release may have led to a downregulation of 5-HT2a receptors. Indeed, Scheffel (1992) reported a transient downregulation of these receptors in rats after a neurotoxic regimen of MDMA. Changes in 5-HT2a receptors are thought to play a role in regulation of cerebral blood vessel constriction (Nobler, 1999). Consistent with this idea, Reneman (2000b) reported correlations between apparent 5-HT2a density and regional cerebral blood volume in the occipital cortex and globus pallidus of a subset of five ecstasy users in whom cerebral blood volume was measured using MRI. Thus, the decreased cerebral blood flow/volume seen in Grob's volunteers and Reneman's ecstasy users may be the result of transient 5-HT2a downregulation due to MDMA-induced 5-HT release.

This hypothesis does not, however, explain the trends toward increased cerebral blood flow or volume seen by both groups at later time points. Given that this trend occurred in very few volunteers, it must be interpreted with caution until confirmed in a more detailed study. Nonetheless, the duration of this possible increase provides cause for concern. A rodent study by McBean (1990) found that a neurotoxic regimen of MDA increased regional cerebral blood flow in rats at six to nine weeks after drug exposure. Thus, although one could hypothesize non-neurotoxic mechanisms, long-term increases in cerebral blood flow have been associated with serotonergic neurotoxicity. Therefore, it appears possible that one or more of the doses received by these two volunteers is sufficient to produce neurotoxicity. The two volunteers each received 2.0 mg/kg MDMA in one session and either 1.75 or 2.25 mg/kg in another session. Increases in cerebral blood flow after MDMA may not be permanent since the MDMA-experienced volunteers in Grob's study did not have increased cerebral blood flow in comparison to non-users (Chang, 2000). Nonetheless, researchers may wish to consider carefully the risks and benefits of exposing volunteers to a drug that may have detectable effects 80 days later.

Overall, preliminary findings from clinical studies suggest that cognitive functioning is not likely to be significantly altered by one or two exposures to MDMA in a clinical setting. However, changes in cerebral blood flow lasting several weeks or longer may occur. Although the mechanism of these changes has not been directly investigated, MDMA neurotoxicity cannot be ruled out as a possible explanation for any changes lasting several months. Of course, it must be again noted that the numbers of volunteers in the studies described in this section are small and any conclusions must be tentative. Further research is necessary.


Potential Strategies for Reducing Risk of Neurotoxicity in Clinical Settings

Because the range of psychoactive but non-neurotoxic MDMA doses appears narrow in most species and the possible long-term consequences of neurotoxicity are unknown, researchers and therapists may wish to consider strategies for reducing risk of neurotoxicity. For example, although high ambient temperature and humidity are unlikely in clinical settings, it is probably worth noting that these factors may increase body temperature, which is associated with increased MDMA neurotoxicity in rats.

In addition, MDMA dose and frequency of administration could be kept to the minimum required. Even if they do not produce measurable neurotoxicity, all active doses of MDMA likely cause some degree of oxidative stress in the brain. Furthermore, the nonlinear pharmacokinetics of MDMA suggest that small increases in dose may lead to large increases in plasma MDMA levels (de la Torre, 2000; Mas, 1999) and, possibly, risk of neurotoxicity. Administration of a small 'booster' dose to lengthen MDMA intoxication may also increase risk, given the apparently saturable metabolism of MDMA. The possible benefits of such 'booster' doses should be carefully weighted against this risk.

Anecdotal reports suggest that many ecstasy users are already employing pharmacological interventions that have been found to be neuroprotective in rodent MDMA studies. These interventions include antioxidants, SSRIs, and 5-hydroxytryptophan. Because rodent studies demonstrating neuroprotection have almost exclusively used multiple injections of high doses of neuroprotective agents, it is not always clear that humans can achieve comparable neuroprotection with oral dosing. This is particularly in question for vitamin C, which has saturable absorption and increased clearance after high doses in humans (Blanchard, 1997; Graumlich, 1997). Similarly, rodent studies using 5-hydroxytryptophan typically co-administer another drug to decrease peripheral metabolism of 5-hydroxytryptophan, allowing more to reach the brain. Oral administration of 5-hydroxytryptophan in humans therefore may not achieve adequate brain levels.

Keeping in mind the almost complete lack of controlled studies examining these interventions in humans, some seem sufficiently promising to warrant further consideration when designing protocols. Administering an SSRI when subjective MDMA effects have become minimal could be considered. Liechti (2000) has shown that pretreatment with 40 mg of the SSRI, citalopram, decreases the effects of l.5 mg/kg MDMA. This study demonstrates that these drugs can be co-administered safely in a clinical setting, despite a previous case report describing a possible adverse interaction between these compounds (Lauerma, 1998). Thus, giving an SSRI at 3 or 4 hours after MDMA administration could be considered if MDMA pharmacokinetics are not being measured and SSRIs are not otherwise contraindicated in the relevant patient or volunteer population.

Antioxidant supplements may also prove useful. Aguirre (1999) reported that twice daily administration of high dose alpha-lipoic acid for two days completely blocked the neurotoxicity of a subsequent single dose of MDMA in rats. Because the acute inactivation of TPH can occur after non-neurotoxic MDMA doses and is due to oxidative stress, it is plausible that antioxidants may also enhance recovery from even low MDMA doses. One consideration with antioxidants is that high doses of some acids, such as ascorbic acid (vitamin C), may alter urinary pH and thus affect excretion of MDMA. Aside from such possible pharmacokinetic interactions, doses of antioxidants that are known to be well tolerated appear unlikely to increase risk of adverse events and may decrease risks of chronic toxicity.

Although these potentially neuroprotective strategies are worth considering, appropriate doses and timing are largely unknown. It would furthermore be technically and ethically difficult to establish whether a given intervention has been successful in reducing MDMA neurotoxicity. These interventions should therefore be considered experimental and not be used to reassure potential volunteers that the risks of MDMA neurotoxicity are reduced.


Need for More Research

Individuals who have managed to read this far likely need little convincing that more research is needed. Further animal studies are needed to investigate possible symptoms of MDMA neurotoxicity and whether the aging process influences these symptoms. Animal studies are also needed to better characterize neurotoxic MDMA exposure. Measuring blood and brain levels of MDMA and metabolites after neurotoxic and non-neurotoxic exposures in different species would allow us to better predict what MDMA dose might be neurotoxic in humans. There is a particular need for primates studies that establish threshold doses for neurotoxicity and measure MDMA pharmacokinetics. Now that MDMA pharmacokinetics have been characterized in humans and rats, it would be possible to design repeated dose drug regimens that expose rats to the same MDMA plasma concentration versus time curves that occur in humans after commonly used MDMA doses. Once nonhuman primate pharmacokinetics are established, similar studies could be carried out in those species. Such studies would make great advances in our understanding of the risks of MDMA neurotoxicity to humans.

Long-term follow-up studies in human should investigate whether MDMA exposure is associated with clinically significant symptoms, such as increased risk of affective disorders. The hundreds of patients who underwent MDMA-assisted psychotherapy in the 1970s and early 1980s are one important population who could be assessed. Prospective studies of ecstasy users are needed to definitively establish the extent to which MDMA decreases cognitive performance and whether abstinence from MDMA is associated with recovery. Both human and animal studies should investigate the time course of MDMA-induced changes in regional cerebral blood flow and its relationship to serotonergic functioning.


Summary

High or repeated-dose MDMA regimens can produce long-term changes in indices of serotonergic and axonal functioning in animals. Increasing evidence supports the view that these changes are at least partially the result of damage. The magnitude of these serotonergic changes varies with dose, species, and route of administration. Rodent studies have shown that changes in the core temperature of animals can increase or decrease MDMA neurotoxicity. While some recovery does occur, a study in squirrel monkeys suggests that there may be permanent changes in axonal distribution in some areas of the brain. Oxidative stress appears to play an important role in MDMA neurotoxicity, although the exact mechanisms are poorly understood. The sustained acute pharmacological effects of MDMA may exhaust neuronal energy sources and antioxidant defenses, leading to damage. Metabolites of MDMA are another possible source of oxidative stress. Very few behavioral correlates of MDMA exposure have been found in drug-free laboratory animals, despite dramatic serotonergic changes, alterations in neurofunctioning, and changes in response to drugs. A growing number of studies describe differences between ecstasy users and nonusers. These studies have serious limitations, but suggest that some ecstasy users experience serotonergic changes and cognitive alterations. In contrast to studies of illicit users, the few controlled clinical trials with MDMA in healthy volunteers have reportedly not found evidence of cognitive changes, despite cerebral blood flow alterations in one study. The possible risks of neurotoxicity must be considered when assessing the potential administration of MDMA to humans.