Purpose: Cognitive function (memory); to assess potential differences in verbal memory (immediate, delayed, interference) in long-term ecstasy users, short-term ecstasy users and polydrug users who used little or no ecstasy, and to examine the relationship, if any, that cannabis use played in altering memory function. Specific hypotheses tested - That both long-term and short-term ecstasy users will exhibit reduced learning in an auditory verbal learning task due either to deficits in working memory or in short-term verbal memory.

Design: Non-experimental (retrospective) 3-group between subjects/within subjects design, with drug use (long term ecstasy user, short-term ecstasy user, polydrug-no ecstasy user) serving as a between subjects factor, and with trial (Trials 1 through 7) as a within-subjects factor. Presence and extent of use of drugs other than ecstasy served as covariates in some analyses. All participants completed the RAVLT.

Subjects: 14 long-term ecstasy users, 14 short-term ecstasy users and 14 non-user controls probably residing in the London (England) area, recruited via advertisements in local (London-area) magazines or through snowball technique. Matching - Groups matched on age, estimated verbal IQ (assessed via NART), and approximately matched on use of drugs other than ecstasy.

Criteria for Inclusion, Long Term Ecstasy Users - Having reported use of ecstasy for a duration of 8 years or longer. Short-Term Ecstasy Users - Having reported use of ecstasy for no longer than 5 years. Polydrug-no ecstasy users - Having reported little (lifetime use of 2 tablets or less) or no ecstasy use over a lifetime, but use of other drugs permitted. All Groups - Absence of major psychiatric or neurological illness, or alcohol dependence, with presence determined through self-report in medical history questionnaire, and abstinence from all psychoactives save cannabis for 2 weeks prior to study day and abstinent from cannabis for at least 24 h prior to study day, with abstinence verified through self-report only.

Drug Use Parameters - Long-term ecstasy users reported taking 811.5 +/- 981.8 tablets across a lifetime, with average dose per use reported as 2.1 +/- 1.3 tablets. Duration of ecstasy use, for long-term users, reported in months, was 130.8 +/- 9.6 months. No information provided on frequency of ecstasy use per month. Time since last use, in days, was 123 +/- 153 days. Maximum dose ever used on one occasion was reported as 5.3 +/- 3.6 tablets. Short-term ecstasy users reported having taken, on average, 223.9 +/- 387.3 tablets over a lifetime, with average dose per use reported as 2.5 +/- 1.9 tablets per use. Duration of use, in months, was reported as 46.8 +/- 21.6 months. No information is provided about frequency of use. Time since last use, in days, reported as 498 +/- 831 days. Non-ecstasy users reported using 0.6 +/- 0.9 tablets in a lifetime. Long-term ecstasy users took significantly more ecstasy tablets over a lifetime than short-term ecstasy users. Time since last use appears to be over a year for short-term ecstasy users, making them similar to "abstinent" users in other studies; however, they did not differ from long-term users in time since last use. Other Drugs Use - Either reported as lifetime use or as use in (unspecified) time span. Cannabis use, reported in joints. Long-term users = 10,306.8 +/- 22,119.5 joints, in short-term users = 1617.3 +/- 2898.4 joints, in non-ecstasy users = 447.3 +/- 629.2 joints. Amphetamine use, in long-term users = 157.5 +/- 222.3 occasions, short term users = 127.1 +/- 214.6 occasions, and in non-ecstasy users = 360.9 +/- 1160.2 occasions. Cocaine use, in occasions, for long-term users = 160 +/- 254.5 times, for short-term users = 123.1 +/- 244.5 times, and in non-ecstasy users = 45.8 +/- 132.5 times. LSD use, in trips or times, for long-term users = 75.4 +/- 133.1 times, in short-term users = 8.9 times, in non-ecstasy users = 30.6 +/- 105.5 times. Barbiturate use, in times/pills, for long-term users = 45 +/- 155.6 times/pills, in short-term users = no use, and in no-ecstasy users = no use. Opiate use, in long-term ecstasy users = 124.2 +/- 313.1 times/pills, in short-term users = 1.1 +/- 2.2 times, and in no-ecstasy users = 11.4 +/- 42.8 times/pills. Psilocybin [magic mushroom] use, in doses, for long-term users = 34.4 +/- 83.2 doses, for short-term users = 14.4 +/- 39.8, and for no-ecstasy users = 0.8 +/- 2.7 doses. Solvents, in times/doses, for long-term users = 262.4 +/- 975 times/doses, for short-term users = no use, and for no-ecstasy users = 14.3 +/- 53.5 times/doses. Nicotine, prob. in cigarettes per day (units not indicated), for long-term users = 8.8 +/- 8.3 cigarettes per day, in short-term users = 13.1 +/- 13.5 cigarettes per day, and in non-ecstasy users = 9.6 +/- 11.2 cigarettes. Alcohol, in units per week? (units not provided), for long-term ecstasy users = 13.1 +/- 9.9 units per week, in short-term ecstasy users = 17.9 +/- 22.1 units per week, and in non-ecstasy users = 10.6 +/- 9.1 units per week. Significant differences were found in extent of cannabis user (long term > short-term > no-ecstasy), cocaine (long-term > short-term > no-ecstasy), LSD (short-term > no-ecstasy > long-term), and psilocybin [magic] mushrooms (long-term users > short-term users > no-ecstasy). Note also that only long-term ecstasy users reported using some barbiturates, and that no-ecstasy users used more (but not significantly more) amphetamine than either long-term or short-term ecstasy users.

Group Demographics and Matched Variables - The authors apparently matched samples for age and estimated verbal IQ, and for extent of drug use (other than ecstasy). Gender, as M/F ratio - Long-term ecstasy users = 11/3, short-term ecstasy users = 7/7, no-ecstasy users = 4/10. Age - Average age was 30.7 +/- 4.1 years for long-term ecstasy users, 26.4 +/- 5.9 years for short-term ecstasy users, and 29.1 +/- 10.9 for no-ecstasy users. Estimated Verbal IQ - As assessed through National Adult Reading Test (NART) score; long-term ecstasy users = 114.1 +/- 3.8, for short-term ecstasy users = 114.1 +/- 6.5, and for no-ecstasy users = 113.6 +/- 5.2. There were significantly more women in the no-ecstasy use group and significantly more men in the long-term ecstasy use group.

Measures: Verbal learning and memory assessed via RAVLT. Immediate memory assessed through Trials 1 to 5. List B served as "interference" (novel word) list. Trail 6 measured memory for first list without prior presentation. Delayed memory assessed via recall of first list (List A) 30 min after initial presentation. Errors scored separately from number of words recalled; errors scored as "intrusion" (from A to B or vice versa), "confabulation" (false recall for unrelated words not appearing on list) and "association" (false positive for semantically or phonetically similar words).

Analyses: An analysis of variance was performed to compare data across groups, with drug-use (long-term ecstasy use, short-term ecstasy use and no ecstasy use) serving as between subjects factor. Either separate analyses were conducted on immediate recall, delayed recall, intrusions, confabulations, associations and learning, or each was used as Trial (within-subjects variable, with Trials 1 to 7 as the within-subjects factor. Post-hoc comparisons were made via Duncan's range statistic, with probability value apparently set at p = 0.05. Chi-square "goodness of fit" test was used to examine group variations in first five trials. Because chi-square assumptions violated for error data for novel list (List B) or for delayed recall, no inferential analyses performed. Drug Use Parameters and RAVLT Scores - Associations between RAVLT performance and extent of ecstasy use (lifetime dose, average dose per use, duration of use, time since last use and maximum dose per use) was examined via Pearson's r. Correlations were also performed to examine possible associations between RAVLT performance and extent of drug use for all other drugs where use patterns were significantly different across group (cannabis, cocaine, LSD and mushrooms).

Analysis of Covariance - RAVLT scores were also compared across drug use groups (long-term ecstasy use, short-term ecstasy use and no ecstasy use) via analysis of covariance (ANCOVA) in order to control for consumption of drugs that were used differentially across groups.

Potential Gender Effects - Because drug use groups were not matched for gender composition, a 2-way ANOVA with gender (male, female) serving as 1 between-group factor and drug use (long-term ecstasy user, short term ecstasy user, no-ecstasy) serving as the other between-subjects factor, and with each RAVLT score either analyzed independently or serving as within-subjects factors.

Results - Significant Differences Found: Number of words recalled in trials 1 and 2 (immediate recall) compared to no ecstasy group (long-term = short-term < no ecstasy). Long-term users also recalled fewer words on Trial 2 than short-term ecstasy users. Number of words recalled on delayed memory (Trial 7) were reduced in both long-term and short-term ecstasy users (long-term = short-term < no ecstasy). A significant polynomial linear function was found for 1, 2, 3, List B and delayed recall (presumably increased recall for all 3 groups across trials), with List B producing lower recall via interference. Long-term ecstasy users made significantly greater errors than non-user controls (goodness of fit). Short-term ecstasy users made fewer errors than long-term users, but a greater number than controls. Though no formal analyses performed (so not clear if significantly different), long-term ecstasy users made more intrusion and association errors on List B than short-term ecstasy users or non-users. Short term users made more intrusion errors on delayed recall task than either long-term users or non-users.

Potential Gender Effects - Women participants in all 3 groups scored significantly higher in trials 4 and 5 (learning / immediate recall) than men in all 3 groups.

Drug Use Parameters and RAVLT Scores - Total lifetime ecstasy consumption (number of tablets taken) was negatively correlated with scores on trials 1, 2 and 3 (immediate recall/learning), with greater number of ecstasy tablets taken associated with fewer words recalled on these trials. Number of words recalled on delayed memory trial (Trial 7) were negatively correlated with average dose per use (usual dose) and largest dose per use, with larger average dose per use and larger maximum dose per use associated with fewer words recalled on delayed memory trial. A significant positive association between time since last use of ecstasy and number of words recalled on delayed recall trial was found, with number of words recalled increasing with increasingly long abstinence from ecstasy use. Number of words recalled on Trial 2 (immediate recall/learning) was negatively associated with lifetime consumption of cannabis and "magic mushrooms," with number of words recalled decreasing as lifetime use for these drugs increased.

Analysis of Covariance - After performing ANCOVA with use of cannabis, cocaine, LSD and mushrooms serving as covariates, differences between ecstasy users and non-ecstasy users on Trials 1, 2, 3 and delayed recall remained. Results - No Significant Differences: There were no differences in number of words recalled for any of the 3 groups (long-term ecstasy user, short-term ecstasy user, no ecstasy user) on trials 4 and 5, 6 (learning/immediate recall) or on List B (novel list). Though formal analyses were not conducted on errors for List B or delayed recall, the 3 groups made similar numbers of association errors on delayed recall, and long-term ecstasy users and non-users had similar numbers of intrusion errors on delayed recall (Trial 7).

Potential Gender Effects - Women and men did not perform differently from each other in RAVLT trials 1, 2, 3, 6, List B or Trial 7.

Drug Use Parameters and RAVLT scores - Average dose per use, maximum dose per use, and time since last use were not significantly associated (negatively or positively) with immediate memory trials. Total lifetime consumption (number of tablets taken) was not significantly correlated with number of words recalled on delayed recall trial (7).

Overall Effects: Immediate and delayed recall, assessed by RAVLT, was impaired in long-term and short-term ecstasy users compared with polydrug users who used no (or no more than 2 tablets) of ecstasy. Long-term ecstasy users also recalled fewer words on one immediate recall trial than short-term users. Long-term ecstasy users made the highest number of errors and non-users the fewest, with number of errors for short-term users in between those of long-term users and non-users. There were some differences in intrusion versus association errors after List B for long-term and short-term users, but these were not formally analyzed. These included greater intrusion errors after list B for long-term users, and short term users making more intrusion errors on delayed recall than either non-users or long-term users. Greater total lifetime ecstasy consumption was associated with greater impairment on immediate recall, but not delayed recall. Size of average dose per use and largest (maximum) dose per use were associated with impaired recall on delayed recall, but not immediate recall, and a longer interval since last use was associated with better performance on delayed recall, but not immediate recall. Gender effects were seen, but only in trials where no differences between drug-use based groups were seen. Differences in recall related to ecstasy use still appeared after effects from use of other drugs (specifically cannabis, cocaine, LSD and mushrooms) were statistically controlled for. The authors' hypotheses were partially confirmed; both long-term and short-term ecstasy users demonstrated impaired immediate and delayed recall compared with non-users, with perhaps some impairment in list-learning as well. However, because they postulated that difficulties would either be related to deficits in short-term memory or working memory or both, it is difficult to falsify this aspect of the hypothesis.

Comments: This paper is one of the latest publications to use the RAVLT, or similar list-learning tasks, to compare immediate and delayed recall in ecstasy users and in polydrug users who have not used ecstasy, and results here can be compared with those found by others (Bolla et al. 1998; Gouzoulis-Mayfrank et al. 2000; Reneman et al. 2000). Most researchers who have assessed verbal memory have found impairments associated with ecstasy use. As was the case in the sample studied by Gouzoulis-Mayfrank and colleagues, past cannabis use, or use of other drugs, was not related to reduced verbal memory performance. It is thus surprising that another publication authored by Fox and colleagues, but with an apparently different sample, failed to find differences between ecstasy users and controls on tests of verbal memory; instead, they showed impairments on a test of spatial working memory. Association between time since last use and delayed recall performance suggests that this impairment is not permanent, or that people possess sufficient compensatory mechanisms to overcome this deficit over time. Associations between drug use parameters and RAVLT scores also hint at two at least partially separate processes, one more sensitive to dose per use and related to delayed recall, and one more sensitive to total lifetime consumption and related to immediate recall. This paper shares the limitations inherent in retrospective paper, though better matching appears in this study than in other studies. Reliance on self-report as verification of abstinence from drugs before study day could provide a misleading measure of actual time since last use.