Sleep. )6(6):560-564
1993 American Sleep Disorders Association and Sleep Research Society

Pharmacology and Sleep

Persistent Effects of (+)3,4-Methylenedioxymethamphetamine

(MDMA, "Ecstasy") on Human Sleep

*Richard P. Allen, tUna D. McCann and *George A. Ricaurte

*Department of Neurology, Johns Hopkins University, School of Medicine,

Francis Scott Key Medical Center, 5501 Bayview Drive, Baltimore, Maryland, U.S.A.; and
tSection onAnxiety and Affective Disorders, Biological Psychiatry Branch, AVAIH, Building 10,
Room 3S-239, 9000 Rockville Pike, Bethesda, Maryland, US.A.

Summary: (±) 3,4-meth ylenedioxymethampheta mine (MDMA) is a recreational drug of abuse which damages serotonin neurons in animals. It is not known whether MDMA is also neurotoxic in humans, and if so, whether there are functional consequences. Given the putative role of serotonin in sleep, it was hypothesized that one manifestation ofserotonin neurotoxicity in humans might be disturbances ofsleep. To determine whether MDMA use has effects on sleep, all-night polysomnograms of 23 MDMA users were compared to those of 22 age- and sexmatched controls. On average, MDMA users had 19 minutes less total sleep and 23.2 minutes less non-REM (NREM) sleep than controls. These statistically significant differences in NREM sleep were due primarily to an average of 37 minutes less stage 2 sleep, with no significant differences noted in stages 1, 3 or 4. Although it is not known whether the alterations in sleep observed in MDMA users are due to serotonin neurotoxicity, the present findings suggest that MDMA use can-lead to persistent changes in CNS structures involved in human sleep generation. Key Words: MDMA - Scrotonin - Neurotoxicity -Amphetamines -Sleep.

Recreational use of (±)3,4-methylenedioxymethamphetamine (MDMA, "Ecstasy"), a ring-substituted amphetamine derivative that acts indirectly by stimulating the release of brain monoamines (1), has become increasingly popular in the United States and Europe (2,3). In animals, MDMA has been shown to damage brain serotonin neurons (4-9). Because the neurotoxic dose of MDMA in nonhuman primates approximates the typical recreational dose for humans (10), there is concern that humans using MDMA may also incur selective neurotoxic damage to brain serotonerg1c systems (11). While the specific role of serotonin in human behavior is not known, it has been implicated in the regulation of sleep (12), mood (13), anxiety (14), pain (15), aggression (16), memory (17) and appetite (18).

Although the specific mechanisms by which serotonin influences sleep are not known, one of the more Subjects widely held views is that serotonin acts indirectly by modulating the role of other neurotransmitters involved in sleep (19,20). It also has been proposed that serotonin may act indirectly by influencing the accumulation of hypnogenic factors (21,22). Regardless of exactly how scrotonin influences sleep, it seems clear that serotonin is involved in normal sleep-wake mechanisms (23). Thus, if individuals previously exposed to significant amounts of MDMA incur serotonergic damage, abnormalities in sleep might be anticipated. Specifically, based on preclinical and clinical data (2430), we hypothesized that MDMA users would have decreased total sleep time (TST), and that the decrease in TST would be related to reductions in both REM and NREM sleep.

METHODS

Subjects
Fifty-three volunteers participated in the study. Experimental subjects all reported that they had used MDMA on more than 25 separate occasions. Control subjects had no prior history, of NIDMA use. All subjects agreed to refrain from use of any psychoactive substances for at least two weeks before participating in the study, and understood that abstinence would be confirmed by a urine and blood drug screen upon presentation at the clinical research center. Prior to study participation, all subjects were screened by telephone interview, and when appropriate, were invited to the clinical research center for further evaluation. At that time, after giving informed consent, subjects also underwent a physical examination and psychiatric interview using the Scheduled Interview for DSM-111-R (SCID-R). Each experimental subject estimited the number of times s/he had used MDMA. Exclusion criteria for all subjects included past or present major medical illness, history of psychosis, current major depressive illness, current alcohol dependence or history of sleep disorder. Sleep study analyses were limited to subjects between 18 and 40 years of age, because suitable age and sex matched controls were not available for older subjects. Ofthe 47 subjectswho met the above criteria, two control subjects failed to complete the protocol and were excluded from the data analyses, leaving 22 controls and 23 experimental subjects. All subjects entered the clinical research unit I day before the sleep study. Diet and activity of the subjects were controlled during the study.

Polysomnogram recordings and scoring

Subjects came to the sleep center for 2 consecutive nights of polysomnographic recordings (PSGs). On both nights, routine sleep EEG recordings were obtained using two channels for eye movements, two EEG channels (C3A 1, C4A2). and one submental EMG channel. On the first night EMGkA,as recorded from the left and right anterior tibialis muscles. Respiratory activity was recorded using oximetry, airflow from the mouth and nose, and respiratory effort was measured by abdominal and thoracic strain gauges. Sleep time was from 11:00 p.m. to spontaneous awakening after 6:00 a.m. with a maximum of 8 hours permitted before ending sleep. The PSGs from the first night were scored for apnea and periodic limb movements in sleep. EEG data from the first night were considered "adaptation sleep", and were not used for experimental analyses. Polysomnograms from the second night were visually scored by two independent raters for all-sleep stages following the standard Rechtschaffen and Kales procedures (3 1). When significant differences were noted in a particular stage (e.g. stage 2), sleep records were further analyzed for differences within that stage (e.g. spindle or K-complex density). Technicians responsible for recording and scoring polysomnograms were unaware whether subjects bad been exposed to MDMA or were controls.

Statistics

Minutes of TST (all sleep stages other than wake), stages 2, 3 and 4 sleep, and REM sleep were compared between the groups using a standard analysis of variance (ANOVA). Independent factors in the ANOVA in addition to MDMA use were time zone of origin (ordered from cast to west), age and presence or absence of DSM111-R diagnoses which might affect sleep structure (current alcohol abuse or dependence, mood or anxiety disorder).

RESULTS

Subjects

The 23 subjects with significant histories of MDMA use and 22 control subjects completing the protocol were fairly well matched for sex (65% and 77% males, respectively) and for age (averages of 26.7 and 26.1 years, respectively) (Table 1). More MDMA subjects were from the Mountain and Pacific time zones (47.8% MDMA subjects versus 22.7% control subjects). There were no major difrerences between groups in the number of DSM-III-R diagnoses which might afrect sleep (six MDMA subjects versus four control subjects). The MDMA subjects reported using MDMA on average 79.4 times (range of 25 to 300 times).

Polysomnogram data for primary hypotheses

MDMA subjects had significantly less TST than control subjects (mean ± SEM of 384.4 ± 8.7 and 403.8 -+- 6.1 minutes, respectively; p = 0.038), and significantly less NREM time (mean ± SEM of 301.4 ± 7.4 and 324.6 ± 6.0 minutes, respectively; p = 0.006) (Table 2). There was no significant differences between MDMA and control subjects for the amount of AEM sleep (mean ± SEM of 82.4 ± 4.7 and 79.1 ± 4.5 minutes, respectively; p < 0. 10) (Table 2).

Within NREM sleep, MDMA subjects had significantly less stage 2 sleep (mean ±- SEM of 185.1 ± 8.3 minutes for MDMA subjects versus 222.4 ± 6.5 minutes for controls, p = 0.002). TherekA,ere no differences between the two groups for slow-wave sleep (SWS) [(stages 3 + 4) mean ± SEM of 86.5 ± 7.2 minutes for MDMA subjects versus 72.6 ± 8.0 minutes for controls, p > 0.1] (Table 2). No differences between groups were noted for total duration of stage I (mean ± SEM of 29.8 ± 2.8 minutes for MDMA subjects versus 29.6 ± 3.6 minutes for control subjects, p > 0.1) (Table 2).

Increased age was associated with a decline in SWS (stage 3 + 4) time (df = 1,40~ F = 4.99, p = 0.03), with no apparent effects on stages 2 or 1. Psychiatric diagnoses or time zone of origin were not associated with changes in SWS Q + 4),stage 2 or stage I sleep. When data were analyzed after removing the I I subjects with one of the DSM-111-R diagnoses identified as possibly altering sleep, the same statistically significant results were observed (i.e. MIDNIA subjects had less TST, NREM sleep and stage 2 sleep than control subjects, with no differences between groups for the amount of time spent in REM sleep, stage I sleep or SWS). These same results were obtained when the analysis was performed for these directional hypotheses with the limited subset of data for the 28 subjects from only the Eastern and Central time zones.

Other polysomnographic variables

The descriptive comparison of other usual sleep measurements is presented in Table 2 and includes: sleep latency, minutes of stage 3 (S3), minutes of stage 4 (S4), wake minutes after sleep onset, minutes of final awakening before leaving bed, minutes from sleep onset to the first REM, wake minutes after sleep onset to the first REM, number of REM periods, movement arousals per hour of sleep and sleep efficiency. Advanced age was found to be associated with decreased stage 4 (df = 1,4 1, F = 12.98, p < 0.0 1). There were marginally significant differences (p < 0. 10) for the two treatment groups for any of these measurements (Table 2). The pattern of occurrence of SWS in relation to REM periods was reviewed for each patient, and all showed the usual pattern with most of the SWS occurring before the second REM period with relatively increasing amounts of stage 2 and stage I sleep occurring later. No diffierences between the two groups were noted for the pattern of sleep during the night.

Patients with one of the psychiatric diagnoses that might alter sleep had significantly greater number of movement arousals per hour than those without such a diagnosis (df = 1,40, F = 4.84, p = 0.03). There was a significant age effect for minutes of stage 4 sleep. There were no other significant effects for age, time zone of origin or psychiatric diagnoses.

Relationship between extent of MDMA use and sleep abnormalities

The correlation between sleep abnormalities and the number of times individuals had used MDMA did not achieve statistical significance (p > 0.50).

DISCUSSION

The results of the present study indicate that individuals with significant prior exposure to MDMA have less TST than age and sex matched controls. Analysis of REM and NREM sleep shows that this difference is due to decreased NREM sleep, with no differences between groups for REM sleep. Within NREM sleep, comparison of MDMA-exposed and control subjects reveals that MDMA subjects have less stage 2 sleep, with no apparent differences in stage I sleep or SWS (stage 3 + 4) (Table 2)..

The finding that the primary difference between MDMA subjects and controls is the amount of stage 2 sleep is somewhat surprising. Preclinical studies involving pharmacological depletion or anatomic lesions of serotonin neurons generally have shown Chang~es. in REM and all of deeper NREM sleep t tages 2, 3 and 4) (24,25). Although studies involving serotonin depletion in humans are few, humans who have been treated with the serotonin synthesis inhibitor p-chlorophenylalamne (PCPA) have been found to have decreased REM sleep, with no changes reported in stage 2 (26). However, most of these preclinical studies involved massive depletions of serotonin with sleep studies generally performed shortly after lesioning. Also, sleep in PCPA-treated humans was evaluated while individuals were still taking PCPA. In contrast, individuals in the present study had taken varying amounts of MDMA in the remote past; the persisting effects may differ from acute effects. It is possible that if studies were performed acutely following MDMA exposure, changes in REM and SWS would be observed.

It should be noted that changes in stage 2 sleep were not due to structural changes in stage 2 itself Specifically, when individual records were examined in greater detail, no changes in spindle number or density, or differences in K-complexes were detected. Indeed, no gross abnormalities in the sleep records of MDMAexposed individuals were apparent, suggesting that sleep-generating mechanisms remained largely intact. Even the occurrence of stage 3 and 4 predominantly in the first two NREM cycles showed the same normal pattern for both study groups.

Two factors should be addressed when considering these data. First, because some of the subjects lived in diffierent time zones, polysomnographic recordings might have been affected by recent trans-meridian jet travel. However, 3 days of adjustment generally is adequate for individuals in the age range evaluated in this study. Further, because the expected efrects of a west to cast time zone change (i.e. increased wake time and longer sleep latency) (32) were not the observed diffierences between experimental groups, this possibility is less concerning. Time zone of origin also was included in the basic statistical analysis as an independent variable, and no etTects were found for any of the sleep variables. As a final check, an analysis of the subset of data from only the Eastern and Central time zones showed essentially the same results as the total sample. A time zone shift of only 1 hour would not be expected to produce the significant etTects noted in this study, especially after 3 days of accommodation to the new time zone. A second potential complication of the present findings is the presence of psychiatric disorders in some of the control and MDMA subjects. Again, presence of relevant psychiatric diagnoses was included as an independent variable in the basic statistical analysis and no significant efrects were found. Indeed, even when data from these 10 subjects were removed from the analyses, the same statistically significant effects on TST and stage 2 were observed.

It deserves note that despite the heterogeneity of the study population (with some subjects abstaining from MDMA for more than a year), the observed ef[ects on total and stage 2 sleep times are robust. Although the clinical significance of the sleep loss remains unclear, chronic loss of sleep, even in small amounts, might lead to decreased alertness and-difficulty with tasks requinng vigilance. Further, individuals with a chronic sleep debt would be expected to be more susceptible to the detrimental efrects of sleep deprivation in jobrelated or social situations. Problems that might be anticipated include errors of omission, diminished ability to retain new information and increased risk for accidents while performing tasks requiring sustained alertness.

In summary, the present results are consistent with the view that scrotonin is a modulator of sleep, although the specific mechanism by which serotonin influences sleep remains to be elucidated. The present findings are also consistent v,'ith the view that recrealional use of MDNIA in substantial cumulative doses may induce lasting CNS serotonergic damage. It may be that by challenging MDMA-exposed individuals with sleep deprivation or with drugs that interfere with serotonin neurotransmission, serotonin dependent functional deficits may become apparent. Such studies hold promise for helping delineate the functional role of serotonin in human behavior. Finally, given the relative paucity of known functional consequences of MDMA use, sleep changes may provide a sensitive measure for detecting subclinical CNS damage secondary to MDMA and related synthetic amphetamine analogs.

Acknowledgements: This work was supported by a grant 19. from the U.S. Public Health Service (NIDA grant #DA05707).

REFERENCES

1. Schmidt CJ, Levin JA, Lovenberg W. In vitro and in vivo neurochernical effms of rneth%,Ienediox~lmetbamphetainine on striatal monciamine systems in the rat brain. Biochein Phar macol 1987;36:747-55.

2. Peroutka SJ. Incidence of recreational use of 3,4-methylenedimethoxymethampbetamine (MDMA, "Ecstacy") on an undergraduate campus. New Engl J3fed 1987;317:1542-43.

3. Henry JA, Jeffreys KJ, Dawling S. Toxicity and deaths from 3,4-methylenediox)'Metham pheta mine ("ecsiacy") Lancet 1992; 340:384-87.

4. Stone DM, Stahl DS, Hanson GL, Gibb JW. The effects of 3,4methylenedioxymethamphetamine (MDMA) and 3,4-methylenedioxyamphetamine on monoaminergic systems in the rat brain. Eur J Pharinacol 1986; 128:41-8.

5. Commins DL, Vosmer G, Virus R, Woolverion W, Schuster C, Seiden L. Biochemical and histological evidence that methylenedioxymeth-amphetamine (MDMA) is toxic to neurons in the rat brain. J Pharinacol Exp Ther 1987;241:338-45.

6 Schmidt CJ. Neurotoxicity of the psychedelic amplfttamine, methvienedioxyMethamphetamine.JPharinaco/Exp Aer 1987; 240: f-7.

7. Battaglia G, Yeh SY, O'Hearn E, Molliver ME, Kuhar MJ, DeSouza EB. 3,4-Methylcnedioxymethamphetamine and 3,4methylenedioxyamphetamine destroy serotonin neurons in the rat brain: quantification ofneurcidegeneration by measurement of [~Hlparoxetine-labelled serotonin uptake sites. J PharmacolExp Ther 1 987;242:911-6.

8. Ricaurte GA, Finnegan KT, DeLanney LE, Irwin 1, Langston JW. Aminergic metabolites in cerebrospinal fluid of humans previously exposed to MDMA: prelirninar~y observations. A'1AS 1990;600:699710.

9. O'Hearn E, Battalgia G, DeSouza EB, Kuhar MJ, Molliver ME. Methylenedioxyamphetarnine (MDA) and methylenedioxymethamphetamine (MDMA) cause ablation of serotonergic axon terminals in forebrain: immunoc~ tochemical evidence. J Aeu rosci 1988:8:2788-803.

10. Ricaurie GA, Forno LS. Wilson MA, DeLanney LE, Molliver ME, Langston JW. (--)3,4-Meth)-lenediox~-methamphetamine (MDMA) selectively damages central serolonergic neurons in nonhuman primates. J.4.114 1988;260:51-5.

11. Ricaurte G, DeLarine% LE, Irwin 1, Langston JW. Toxic effects of MDMA on central seroionergic neurons in the primate: importance of route and firequence of drug administration. Brain Res 1988;446:165-8. Sleep, Vol. 16, A'o. 6, 19~13

12. Oswald 1, Berger RJ, Evans J 1, Thacore VR. Effects of L-tryptoPhan upon human sleep. BrJ Psichiatri- 1964:109:66-78.

13. Meltzer HY. Role of scrotonin in depression. NYAS 1990;600:

14. Charney DS, Woods SW, Goodman WK, Heninger GR. Serotonin function in anxiety. 11. Effects of the serotonin agonist mCPP in panic disorder patients and health~ subjects. Psychop~arinacolon, 1 988;92:14-24.

15. Messing RB, Lyfle LD. Serotonin containing neurons: their possible role in pain and analgesia. Pain 1977:4:1-22.

16. Coccaro EF. Central serotonin and impulsive aggression. Br J Psychiatry 1989;155:52-62.

17. Weingartner H, Rudoerfer MV, Buchsbaum INIS, Linnoila M. Effects of serotonin on memory impairment produced by ethanol. Science1983;221:472-4.

18. Fernstrom JD. Food-induced changes in brain serotonin synthesis: is there a relationship to appetite for specific macronutrients? Appeiiie 1987;8:163-82.

19 Jones BE. Basic mechanisms of sleep-wake states. In: Kryger MH, Roth T, and Demerit WC, eds. Principles andpractice of sleep medicine. Philadelphia: W. B. Saunders, 1989:121-38.

20. Jacobs BL. Overview of the activity of brain monoaminergic neurons across the sleep-wake cycle. in: Wauquier JM, Gaillard JM, Monti JM, Radulovacki M, eds. Sleep: neurotransminers and neuromodulators. New York: Raven Press, 1985:1-14.

2 1. Jouvet M. Neuromediateurs et facteurs hypnogenes. Rev Neurol (Paris) 1984;140:389-400.

22. Jouvet M, Denoyer M, Kitaharna K, Sallanon M. Slow wave sleep and indolamines: a hypothalamic target. In: Wauquier A, Dugovic C, Radulovacki M, eds. Slow wavc sleep: physiology, palhoplij,siolog), andfunctional aspects. New York: Raven Press, 1989:91-108.

23. Hobson JA, Lyclic R, Baghodoyan HA. Evolving concepts of sleep cycle generation: from brain centers to neuronal populations. Behav Brain Sci 1986;9:371-448.

24, Koella WP, Felstein A, Cziman JS. The effects of parachlorophenylalanine on the sleep ofcats. EleciroencephalogrClin Aleurophysiol 1968;25:481-90.

25. Borbely AA, Neuhaus HU, Tobler 1. Effect of p-chlorophenylalanine and tryptophan on sleep, EEG, and motor activity in the rat. Behai, Brain Res 198 1;2:1-22.

26. Wyatt RJ, Chase TN, Engelman K, Kupfer DJ. Scott J, Sjoerdsma A, Snyder F. Reversal of parachlorophenylalanine (PCPA) REM suppression in man by 5-hydroxytrypiophan (5HTP). Psychoph ' isiolog.v 1970;7:318-19.

27. Wyatt RJ, Zarcone V, Engelman K, Demerit WC, Snyder F, Sjoerdsma A. Effects of 5-hydroxytryptophan on the sleep of normal human subjects. Eleciroencephalogr Clin Neurophysiol 197 1;30:505-9.

28. Quattrochi JJ, Mamelak A, Binder DK, Williams J. Rittenhouse C, Hobson )A. Dynamic suppression of REM sleep by parental administration of the serotonin-I agonist eltoprazine. Sleep 1992; 15:125-32.

29. Dugovic C, Wauquier A, Leysen JE, Janssen PAJ. Role of serotonin-S2 receptors in the control of sleep-%% akefulness states in the rat. In: Wauquier A, Dugovic C, Radulo%acki M, eds. Slow waie sleep: physiology, pathophysiology and.functional aspects. New York: Raven Press, 1984:183-96.

30. Stutzman J-M, Eon B, Lucas M. Blanchard J-C, Laduron PM. RP 62203, a 5-hydroxyti-yptamine. antagonist enhances deep NREM sleep in rats. Sleep 1992:15:199-24.

31. Rechtschaffen A, Kales A, eds. .4 manual c?f siandardi:ed terrninology, techniques, and scoring SJ'S[Cln for sleep stages in human subjects. Los Angeles: UCLA Brain Information Service/ Brain Research Institute, 1968,

32. Wegmann HM, Klein KE. Sleep and air traN el. In: Wheatley D, ed. Psichopharmacologyqfsleep. NeA York: RaN en Press: 198 1: 95-1 f6.