Chronic cognitive impairment in users of 'ecstasy' and cannabis

EARLY STUDIES ON MDMA AND COGNITION

Taking MDMA leads to an acute massive neuronal release of serotonin (5-HT), followed by a period of depletion before levels return to normal. More chronic toxic effects appear also to involve primarily the 5-HT system, with the demonstration of serotonergic degeneration in several animal species, including non-human primates (3). 5- HT is thought to play a prominent role in memory function and marked toxic effects of MDMA have been observed in the hippocampus and frontal cortex - areas crucial to memory and other cognitive functions (4). This suggests that MDMA may have long-term effects on cognition.

Important difficulties are faced by researchers investigating the effects of MDMA on humans. First and foremost, there are ethical and legal proscriptions to repeatedly administering an illicit and potentially toxic drug in controlled laboratory conditions. This means that studies are mostly restricted to recreational users, with the possible consequence of inaccurate self-report of the amounts of MDMA used. Furthermore, 'ecstasy' tablets bought on the black market contain a variable amount of MDMA (5), and may contain a variety of other substances, including the closely related 3,4-methylenedioxyamphetamine (MDA) and 3,4-methylenedioxyethylamphetamine (MDEA), caffeine, ephedrine, selegiline, amphetamine, ketamine, and LSD (6,7). Thus, amounts of MDMA consumed by experimental subjects are estimates and extraneous substances may affect findings. Then there are difficulties in finding suitable MDMA naïve control groups to match user groups. MDMA is primarily used in the dance/'rave' scene, and there are very few individuals who share this lifestyle - including for example the sleep deprivation associated with all night raves - who have not taken MDMA. Premorbid differences between users and controls in intellectual, cognitive or psychological functions also need to be addressed. Furthermore, most MDMA users take other psychoactive drugs, including alcohol and illicit drugs, and these may have their own effects on cognition. Importantly, acute and chronic effects are often masked by too short an abstinence period for both MDMA and cannabis, especially the latter when the focus is on MDMA.

One of the first indications of chronic cognitive deficits resulting from MDMA came in 1992 (8). Impairment of initial and delayed paragraph recall was described in 9 subjects compared to age-matched normative values. Subjects had taken an average of approximately 130 'ecstasy' tablets. They reported abstinence from MDMA for an average of about 66 days prior to testing and from all psychoactive drugs for 3 weeks. Performance on other cognitive tests was unimpaired. Results were, however, limited by the small number of subjects, the past use of other psychoactive substances, the fact that they were given a tryptophan infusion prior to testing and the lack of a control group. Furthermore, some of the subjects had psychiatric histories.

Parrott et al (9) then compared three groups: 10 regular MDMA users who had taken MDMA on more than ten occasions, 10 novice MDMA users (< ten occasions), and 10 age-matched control subjects who had never taken MDMA. The period of abstinence was undefined, so that possible lingering effects following recent MDMA or other drug taking were not excluded. The MDMA groups were significantly impaired on immediate and delayed word recall, but not on other cognitive tests (information processing speed, sustained attention). An important flaw was that details of other illicit drugs were not recorded. This study was noteworthy for finding deficits in light users.

A study looking at slightly heavier consumption compared 24 MDMA users (> 25 times) and 24 MDMA naïve controls after an abstinence period from all psychoactive drugs of 2 weeks (10). No significant differences in memory function were found between the two groups. However, impairment in immediate verbal and delayed visual memory was found in the heavier MDMA users when estimated monthly dose of MDMA was included in regression models. Both subjects and control groups had taken other illicit substances, thus achieving some control for this potential confounder. However, levels of use were much greater in the MDMA group.

Shortly after the above studies, in 1999, we instigated a study of MDMA recruiting participants via advertisements in the popular press (11). The sample size of 36 was relatively large for the field and MDMA use was greater than in most preceding studies (estimated mean 235 tablets). Subjects were identified for whom MDMA was the primary drug and who reported using other drugs only irregularly. They had not consumed any illicit drug for a mean of 78 days prior to the testing. Thus, impairments were unlikely to represent either acute MDMA effects, which last approximately 3-5 hours after a dose (12), or lingering effects which may last several days and may reflect the time course of regeneration of 5-HT.

We found discrete deficits in memory and learning compared to drug naïve controls, which were both verbal (immediate recall of words from a list, digit sequence learning) and non-verbal (face recognition and spatial associative learning). Other cognitive functions were intact, including executive/frontal lobe functions, attention and verbal and non-verbal working memory. Both sides of the brain were implicated and the patterned deficit profile excluded explanations based on non-specific factors such as subjects' motivation. Mood, as assessed by the Beck Depression Inventory, also did not affect results. Cannabis was clearly identified as the next most commonly used illicit substance; however, none of the deficits showed any correlation with frequency or amount of cannabis used. Nevertheless, the difficulty in identifying a pattern to the MDMA deficits prompted a second study looking more closely at the role played by cannabis.

In this second study (13), a group taking cannabis but not MDMA (n=18) was introduced and compared with a group that took both cannabis and MDMA (n=11) and controls who had not taken MDMA or cannabis (n=31). Both user groups performed more poorly than controls on tests of verbal and visual memory, verbal fluency, speed of processing and manual dexterity. However, there was little difference between user groups on any of the tests. Thus, deficits appeared more closely related to cannabis than MDMA. It should be noted, however, that the size of the cannabis+MDMA group was comparatively small and the mean MDMA usage was about 42 tablets, far lower than in our previous study, whereas cannabis use was heavy, with mean estimates of about 11000 and 7800 total joints in the user groups. Recent cannabis use may have affected findings, as the requested abstinence period from cannabis was 48 hours and a number of subjects reported even less than this.

Notwithstanding these findings, MDMA has continued to be implicated in contemporaneous research, and further work in our department with electrophysiological measures highlighted residual effects that were independent of cannabis. Croft et al (14) focussed on the suspected neurophysiological correlate of MDMA-related cognitive impairment, namely depressed 5-HT function. 5-HT in the primary auditory cortex is thought to operate as a protective mechanism by attenuating the cortical response to loud auditory stimuli, and this was employed as an electrophysiological index of 5-HT function. It was found that long-term predominantly MDMA users exhibited 5-HT dysfunction relative to both predominantly cannabis users and drug-naïve controls. The 5-HT impairment was strongly related to total MDMA consumption, but was independent of frequency of MDMA use, suggesting a causal relationship, i.e. that MDMA caused the 5-HT impairment. This follows the argument that, if the relationship were reversed, such that impaired 5-HT predisposed an individual to 'risky behaviour' including MDMA use, then frequency of MDMA use and not total tablets consumed would be more closely related to the 5-HT deficit. The total number of tablets consumed would be determined by the stage of the subject's MDMA 'career' when recruited to the study. This study also found that cannabis was not related to the 5-HT deficit.

Now in the new millennium there have been a plethora of studies on the cognitive effects of chronic MDMA use. Before reviewing these, the cognitive effects of cannabis will be considered. Cannabis is a popular drug of choice in the rave subculture and is also taken by MDMA users to ameliorate the low mood and irritability of the 'coming down' period after MDMA use.

CANNABIS AND COGNITION

The neuropsychological effects of cannabis can be divided into acute and residual (15). Acute effects are those associated with intoxication. Residual effects may be 'drug residue' effects from cannabis accumulation in the CNS in the hours/days after acute intoxication, or a chronic residual toxicity that persists after the drug has left the body representing CNS alteration effects.

Considering first acute influences, cross-sectional studies experimentally administering cannabis to subjects have indeed supported the existence of a syndrome of acute intoxication, with mood change, perceptual changes and characteristic cognitive impairments in memory and attention, motor skills, reaction time and skilled activities (16). After the period of acute intoxication, it takes some time for performance to return to pre-dose levels, suggesting a drug residue effect. For example, performance on a flight simulator was impaired for up to 24 hours in aircraft pilots after smoking cannabis (17). In light users there has been evidence of a drug residue effect for 12-24 hours following a dose, but no clear evidence of a deficit persisting for more than 48 hours.

Investigation into the chronic residual/toxic effects of cannabis has focussed on 'naturalistic studies' of individuals who have taken the drug more heavily or over longer periods than could be ethically duplicated in the laboratory. Users have been tested after a period of abstinence to exclude acute or drug residue effects, although the length of this period has varied from study to study. In the 1970s and 1980s studies were roughly equally divided with respect to positive and negative results. Most had serious methodological flaws. In general abstinence periods were less than 24 hours or unrecorded, thus drug residue effects may easily have explained findings.

Two early reports of chronic deficits resulting from cannabis were subject to limitations. In groups that were small and not well matched for gender and use of other drugs, Schwartz et al (18) compared 10 cannabis dependent users after 6 weeks of supervised abstinence with control groups of polydrug users and non-users. Visual retention and verbal memory for prose passages were found to be impaired. A larger scale study importantly matched 144 users and 72 non-using controls in intellectual ability prior to onset of cannabis use (19). Heavy cannabis users (7+ times weekly) showed modest deficiencies in mathematical reasoning and verbal expression as well as selective impairments in the memory retrieval process. Light and intermediate use was not associated with deficits. The user group had, however, experienced other drugs far more extensively and there was only a 24 hour unsupervised abstinence period.

In more rigorous studies, a consistent theme has been attention impairment as a result of heavy or long-term use, as well as memory impairment. A study comparing groups of heavy and light cannabis users with a supervised abstinence period of 19 hours before testing found impairments in mental flexibility (attentional set shifting) in heavy users (20). Learning of word lists was also impaired. The group differences remained after some statistical control of potential confounding variables, such as estimated levels of premorbid cognitive functioning and use of alcohol and other substances. In a Costa Rican sample having an abstinence period of 72 hours verified by urinalysis, subtle deficits in selective and divided attention associated with working memory and in verbal memory were disclosed that were specific to long-term users (21).

The most recent report (22) found that long-term cannabis takers (mean of 24 years use) performed significantly less well on verbal memory than shorter-term (mean 10 years) and non-user controls. Impairments were not severe, but learning, retention and retrieval were all affected following a median self-reported abstinence period of 17 hours. There was some suggestion of attention impairment, although executive tasks were generally unimpaired. Problems with this study include the fact that the cannabis users recruited were seeking treatment for dependence due to concerns including that of subjective cognitive impairment and may thus not be representative of cannabis users in general. The abstinence period was short and groups differed in use of other drugs; however, attempts were made to control these factors statistically. Performance measures on several tests correlated significantly with the duration of cannabis use, suggesting a chronic neurotoxic effect. Interestingly, apart from a time estimation task, no deficits were shown in shorter-term users with a mean use of 10 years and who were using cannabis daily.

Although the studies described so far used abstinence periods of 12-72 hours, it is still not clear if the deficits found were neurotoxic effects or whether they were due to a residue of cannabinoids in the brain. The principal active component of cannabis, delta-9-tetrahydrocannabinol, accumulates in fatty tissue, and has a tissue elimination half-life of about 7 days (23). In heavy users cannabinoids may remain in the body for more than two months after cessation (24). This prompted a study by Pope et al (25) in 108 long-term heavy users of cannabis throughout 28 days of monitored abstinence confirmed by urinalysis. Deficits were found on memory of word lists detectable at least 7 days after discontinuation of the drug and which were related to initial urinary cannabinoid concentration. These appeared to be reversible phenomena associated with recent drug exposure - attributable either to cannabinoids lingering in the CNS or to withdrawal from abruptly stopping use. An association between cumulative lifetime use of cannabis and cognitive deterioration was not found. After 28 days of abstinence, users showed virtually no differences from control subjects on a battery of 10 neuropsychological tests. Thus, impairments were related to recent cannabis exposure and were reversible with abstention. These findings are compatible with studies that found no significant long-term effects on IQ measures and the Mini Mental State Examination (26,27). A metaanalysis of 13 of the more methodologically sound studies also found no significant evidence for chronic deficits in 7 out of 8 neuropsychological ability zones, with a possible small effect in only one domain, namely learning (28).

Finally, looking at electrophysiological evidence of attention impairment, Solowij et al (29,30) examined cortical event related potentials for anomalies of processing negativity in an auditory detection task. The ability to focus attention and filter out irrelevant information was impaired progressively with the number of years of cannabis use and only partial recovery was evident in former heavy users after a mean of approximately 2 years abstinence.

In summary, there is clear evidence that cannabis has a residual effect on cognition that lasts 12-24 hours after even a single episode of smoking. This needs to be considered when investigating effects of MDMA. Frequent smoking may lead to an accumulation of CNS cannabinoids and a drug residue effect may persist for much longer and may be continuously present. It is likely to include impaired focussed attention and working memory as well as impaired visual and verbal memory. There is less evidence for chronic neurotoxic effects following cannabis; however, there may be more subtle deficits influencing attention, such as those demonstrated electrophysiologically which appear to be related to duration of cannabis use. Studies into MDMA should aim for periods of abstinence from cannabis of some weeks rather than days, to avoid or minimize residual effects, particularly if cannabis is taken frequently. It also has to be borne in mind that withdrawal effects - characterised by insomnia, restlessness and irritability - in heavy cannabis users may affect cognitive performance. However, a more complex effect with MDMA and cannabis potentiating or diminishing the impact of the other on cognition cannot be ruled out.

RECENT RESEARCH INTO CHRONIC EFFECTS OF MDMA ON COGNITION

In an attempt to address the fact that past use of other psychoactive drugs may confound MDMA results, Morgan (31,32) compared polydrug taking controls who had never taken MDMA with MDMA+polydrug users and nonusers. The MDMA group had consumed an average of 50 tablets and reported being MDMA free for an average of 9 weeks. Participants were only required to abstain from taking other psychoactive drugs on the day of the study, so that effects from their recent use were not excluded. The MDMA group had impaired immediate and delayed prose recall compared to both control groups. Performance on other cognitive tests, including executive function and working memory, was unimpaired.

Focussing more specifically on controlling for past cannabis taking, Gouzoulis-Mayfrank et al (33) compared groups of non-drug controls with cannabis and cannabis+MDMA users (average about 94 tablets). Abstinence from MDMA was reported to be more than 21 days; however, subjects were asked to refrain from cannabis only on the day of the study. The MDMA+cannabis group performed significantly worse than both other groups on tests involving complex attention (but not basic alertness), visual learning, problem solving and strategic planning, suggesting that MDMA was primarily responsible. No significant deficits were associated with the cannabis groups, but of note was that in the MDMA group heavier use of cannabis was associated with stronger cognitive deficits. A similar design was used by Rodgers (34) to investigate lighter MDMA users (average of 20 times) after 2 months abstinence from MDMA and 1 month abstinence from cannabis, although this was not confirmed by drug screening. Deficits associated with MDMA were found on delayed recall of visual and verbal paired associates.

Then in order to minimise differences in psychosocial factors, Verkes et al (35) used MDMA naïve controls who were also rave party visitors, and compared them with 'moderate' frequency MDMA users (mean about 169 tablets), and 'heavy' frequency MDMA users (741 tablets). Abstinence from all psychoactive agents for one week was checked by urine drug screening. MDMA groups were impaired on memory span and word and figure recognition. The MDMA groups had consumed more cannabis, cocaine and amphetamines; however, deficits remained after some statistical control for this had been introduced.

Thus, while there is a growing consensus that memory deficits remain when past use of other drugs and psychosocial factors are considered, none of the recent studies are free from methodological confounds. MDMA has inevitably been taken with other drugs, and while our review suggests that cannabis may not produce memory deficits as a long-term residual effect, synergistic effects cannot be ruled out, but even more pertinently abstinence periods have been poorly controlled. For example, with an abstinence period from all drugs of only 24 hours, Bhattachary and Powell (36) found impaired immediate and delayed prose recall and executive function (verbal fluency). Similarly, Fox et al (37), although requesting participants to be free of MDMA and other illicit drugs for at least 2 weeks, only requested them to be free of cannabis for 1 day prior to testing. They reported deficits in verbal learning that were more related to storage and/or retrieval than problems associated with capacity. Long-term (>8 years) users showed some evidence of selective executive problems. Then with an abstinence period from all drugs of 2 weeks confirmed by drug screening, but with no control group and with polydrug use in the MDMA group, a longitudinal study found that continued use of MDMA over one year was associated with decline in immediate and delayed prose recall (38).

Consensus is also emerging for deficits in working memory and central executive functions in heavier MDMA users, though with less consistency across reports compared with deficits of long-term memory. Furthermore, these studies remain fraught with the same methodological confounds. With no defined abstinence period for other drugs and limited statistical control for their past use and with no definition of an MDMA-free period in current heavy MDMA takers (about 1200 tablets), Wareing et al (39) found impairment in aspects of central executive functioning (random letter generation) along with impaired accuracy of information processing. A study designed to avoid effects of recent psychoactive drug taking included abstinence periods from all recreational drugs of at least three weeks with confirmation of drugfree status by drug screening (40). Heavy MDMA polydrug users (mean of about 584 tablets) were subtly impaired on computerised tests of sustained attention, complex attention/ working memory and logical reasoning as well as delayed memory compared to MDMA naïve polydrug controls. However, MDMA subjects had used more recreational drugs than controls and this may have contributed to cognitive impairments.

Requiring participants not to have smoked cannabis for three days and not to have consumed other drugs for 24 hours, both relatively short abstinence periods, Heffernan et al (41) investigated the central executive functions involved in prospective memory (i.e. remembering to do something in the future). Heavy MDMA users were impaired in self-rated prospective memory, as well as in verbal and semantic fluency tasks compared to a MDMA naïve polydrug group. Impairments remained significant when a measure of statistical control was used for consumption of cannabis and other drugs; nevertheless use of other drugs was much greater in MDMA groups. This limitation also applied to the carefully designed studies of Fox et al (42,43), both of which used MDMA naïve polydrug controls. In the first study, subjects with subjective problems attributed to MDMA (mean of about 372 tablets) were compared with MDMA users without these problems (mean of about 357 tablets) and controls (42). Deficits were disclosed on executive planning and spatial working memory which related to total MDMA consumption and were not related to subjective problems, suggesting that damage may occur without conscious appraisal. However, the second study (43) provided somewhat conflicting results whereby participants were unimpaired on most measures of prefrontal functioning. Here less heavy MDMA polydrug users (mean of about 172 tablets) were compared with controls. The MDMA group showed deficits in memory tasks sensitive to involvement of temporal structures or in a manner similar to patients with temporal pathology. Results remained significant when measures of drug use which differed between subjects and controls were used as covariates. In both these studies, subjects were requested to be free of psychoactive drugs for 2 weeks, but this was not tested. The discrepancy regarding executive functions was attributed to lower total doses and shorter duration of MDMA taking in the second study.

Executive function was subtly impaired in a recent study of heavy users that looked more closely at duration of toxic effect with abstinence (44). Subjects had consumed averages of 93-577 tablets across the different groups and included a group of ex-MDMA users who had been abstinent for at least 6 months. Participants were required to abstain from cannabis for 24 hours and other illicit drugs for 1 week. Deficits in the MDMA groups which persisted when measures of other illicit drugs were used as covariates involved story recall and working memory/ executive function. Subtle executive impairment was again described in what was predominantly an EEG study (45). In subjects abstinent from all drugs for at least a week, a test of executive function (rule shift card test) correlated negatively with MDMA use in the past year. Interestingly, MDMA use also correlated with EEG coherence, a measure of synchronisation of firing, in posterior brain sites overlying visual association pathways, possibly indicating dysfunctional local connectivity.

Correlations between reported total MDMA exposure and measures of cognitive impairment suggest that deficits are primarily associated with MDMA and support evidence of chronic toxicity in heavier users. Estimated lifetime consumption of MDMA has been found to correlate with verbal memory functions (33,36,37,44), and with working memory and executive function (33,36). However, the fact that MDMA exposure calculations are based on self report and that the amount of MDMA in a tablet may vary considerably make such calculations problematic. Furthermore, the fact that individuals with higher lifetime MDMA consumption tend also to have higher levels of consumption of other drugs needs consideration when assessing regression or correlation analyses. It also has to be considered that MDMA induced neurotoxicity may not simply relate to cumulative dose, but that patterns of use may be important and single high doses of MDMA may represent a particular risk, although this has not been widely studied in humans.

Morgan et al (44) suggested that cognitive impairment resulting from MDMA might be a long-lasting phenomenon. Deficits in verbal recall and working memory persisted for at least 6 months and for an average of 2 years after cessation of use. Wareing et al (39) also found that working memory impairment persisted for at least 6 months and Reneman et al (46) reported that verbal recall remained affected at 1 year after stopping use. There has been, however, tentative evidence of some recovery of memory in a small group of 3 MDMA users who had abstained for more than 6 months (32).

Some of the studies mentioned have provided evidence of an association between cognition and indirect biological evidence of 5-HT disruption. Decreased cerebrospinal fluid (CSF) 5-hydroxyindoleacetic acid (5-HIAA - a metabolite of 5-HT) has been positively correlated with memory decline (10). Memory span has been found correlated with cortisol response to a serotonergic challenge (35). On the other hand, although finding significant decrements in CSF 5-HIAA in MDMA users, a correlation with cognitive deficits was not found by McCann et al (40).

In a single photon emission computed tomography (SPECT) study, Reneman et al (47) found post-synaptic cortical 5HT2A receptor binding positively correlated with verbal recall in a group of 5 heavy MDMA users. Reneman et al (48) found impairment in verbal delayed recall to be associated with prefrontal cortex neuronal loss or dysfunction as indicated by altered N-acetylaspartate to creatinine ratio on magnetic resonance spectroscopy. Subjects were 8 heavy MDMA users who had been abstinent for a week. These studies were supportive of relationships between 5-HT disruption and cognitive dysfunction, but were small and not controlled for use of other substances.

A larger SPECT study by Reneman et al (46) did not, however, find brain pre-synaptic 5-HT transporter densities in the cortex to be associated with cognitive measures. Receptor density was significantly lower in MDMA users compared to controls, but not in ex-MDMA users who had stopped at least one year before the study, suggesting that the effect was reversible. By contrast, verbal recall remained impaired in ex-MDMA users. Similarly, a SPECT study by Semple et al (49) found reduction in 5- HT transporter densities in the cortex of heavy MDMA users after about 18 days abstention. They did not report a correlation between receptor binding and neuropsychological measures. Transporter reduction in many regions was inversely correlated with time since last dose, suggesting again that this may represent a reversible effect.

CONCLUSIONS

MDMA is almost never taken without other drugs, most notably cannabis. Studies of cannabis without MDMA have provided clear evidence of acute and shortterm effects on attention and memory with possibly the only long-term effects being on frontal attentional networks. In contrast, there is evidence that MDMA does chronically impair complex memory tasks (hippocampus) and, in heavier users, higher executive information processing (frontal cortex) as well. Much of the evidence is compromised by the fact that short-term residual effects of cannabis and sometimes even acute effects cannot be ruled out. Furthermore, synergistic effects of cannabis and MDMA may explain effects on cognition. Circumstantial evidence correlating total lifetime dose of MDMA with cognitive measures does imply that MDMA is more likely to be the culprit than interaction with cannabis, but in the absence of evidence from pure groups this cannot be known for certain.

The cognitive impairment associated with MDMA and cannabis is not large relative to normal cognitive variability among individuals. Nevertheless, it may be sufficient to affect scholastic performance or those embarking on intellectually challenging careers and may become more manifest as neuronal reserve diminishes with age. Conversely, impairments may resolve with prolonged abstinence. If deficits in MDMA users are indeed a clinical manifestation of serotonergic dysfunction, as some biological studies would suggest, this is cause for concern that users may have increased risk of other psychiatric conditions with strong serotonergic aspects, including depression, schizophrenia, anxiety, impulsivity, aggression, and obsessivecompulsive disorder.