A Critical Evaluation of the Hypnotic Efficacy of Melatonin

7 downloads 0 Views 314KB Size Report
Summary: This paper reviews the available literature on the use of melatonin as a hypnotic in normal volunteers and in patients with noncircadian insomnias.
Sleep. 20(10):916-919 © 1997 American Sleep Disorders Association and Sleep Research Society

A Critical Evaluation of the Hypnotic Efficacy of Melatonin Wallace B. Mendelson Sleep Research Laboratory, The University of Chicago, Chicago, Illinois, U.S.A.

Summary: This paper reviews the available literature on the use of melatonin as a hypnotic in normal volunteers and in patients with noncircadian insomnias. Data are ordered in terms of studies of subject self-reports and polygraphic data and are seen in the context of generally accepted criteria for assessment of clinically used hypnotics. It is concluded that, at the present time, there is very little systematic evidence to suggest that melatonin has significant hypnotic efficacy in these populations. Key Words: Melatonin-Hypnotics-Sleep-Insomnia.

My view on the conflicting viewpoints about melatonin is that a great deal of the difficulty stems from three phenomena. These include: 1. Very broad-based claims in the lay press regarding multiple therapeutic goals; 2. A higher than usual proportion of studies that use atypical methodology or report only partial sleep data; and 3. Inconsistency inside our own field as to what body of data is required to establish an agent as a hypnotic. In this brief review, I will focus on studies with a single therapeutic goal: the use of melatonin as a hypnotic in those types of insomnia that are not thought to be due to circadian dysregulation. First, there will be a few general comments about determining hypnotic efficacy, then a summary of the available data; a more detailed analysis is provided in a recent review (1). I will conclude with some speculations as to why there is such great interest in this substance. An analysis of exactly what criteria need to be met is beyond the range of this article, but a few general comments are in order. Certainly, both self-report and objective (polygraphic) data are needed to establish the efficacy of any sleep-inducing substance. The two provide complementary types of evidence. Subjective reports address whether the compound is actually providing relief for the clinical complaint. Polygraphic data, by contrast, have the advantage of being more quantitative and provide assurance that patients are not selecting a compound for some non-hypnotic reason (e.g. for euphoriant qualities). Secondly, since sleep Accepted for publication September 1997. Address correspondence and reprint requests to Dr. W. Mendelson, The Sleep Research Laboratory, The University of Chicago, 5743 South Drexel Avenue, Chicago, IL 60637, U.S.A.

complaints are so varied, and since many insomniacs have multiple complaints (2,3), it seems to me that a new hypnotic would have to be shown to address a range of difficulties, including sleep initiation and sleep maintenance problems, unless it is proposed to be used in a very specialized group of patients. Thus, it seems crucial that reports of efficacy studies include not only measures of sleep latency, but also total sleep time and variables that reflect continuity of sleep (e.g. wake time after sleep onset). Moreover, the benefits a compound does show should not only be statistically significant, but also quantitatively large enough to be clinically significant. A classical example of this issue comes from one of the studies most frequently cited to support the use of flurazepam during its early years on the market; in this report, 30 mg of flurazepam given for 28 days increased total sleep by only 6--8%, shortened waking time by only 14-17 minutes, and significantly reduced sleep latency only on nights lIB (4). Such data beg the question as to whether patients in the study would recognize an increase in sleep time of 6-8% as an improvement in their symptomatology, or even be aware of such a small change. Finally, although data showing enhancement of sleep in the daytime are suggestive, it seems important to demonstrate that the agent improves nighttime sleep, and that it does so not only in normal volunteers, but also in the relevant group of patients (insomniacs). When insomniacs are studied, data are difficult to interpret unless their sleep disorder is characterized both in the sense of ruling out sleep disturbance due to specific pathophysiologies (e.g. sleep apnea) and in providing data on baseline sleep parameters. While I realize that each of these points is subject to debate, they are presented here as the framework within which I will argue that there is not at present an available body of data

916

'," 917

THE HYPNOTIC EFFICACY OF MELATONIN suggesting that melatonin is an effective hypnotic. To begin, we will examine data on subjective reports, then look at polygraphic data in normal volunteers and insomniacs. Some studies have included data on both subjective and polygraphic data, and these different aspects of their data will thus appear in both sections.

Patient self-reports There have been several studies of physiologic or low pharmacological doses given to normal volunteers but only three presenting data (which were largely negative) in insomniacs. Among the studies of normal volunteers was that of James et al. (5), in which 1 and 5 mg of melatonin and placebo were given for a single night each in random order at 10:45 p.m. There were no significant changes in self-report of sleep with either dose of melatonin. In a complex study in which melatonin 0.3 and 1 mg and placebo were given to normal volunteers at three time points, Zhdanova et al. (6) reported that subjects could retrospectively distinguish melatonin from placebo after it had been given at 6:00 p.m. and 8:00 p.m.; it was not reported whether they could do so after the 9:00 p.m. dosing. Nave et al. (7) reported that when 3 and 6 mg of melatonin was given to 12 normal volunteers 120 and 30 minutes before an early evening nap opportunity, there was a nonsignificant trend for greater reported sleep depth, although other variables such as mood or tiredness were unchanged. There have also been three studies of very high doses of melatonin in normal volunteers. Ferini-Strambi and Zucconi (8) commented that when melatonin 100 mg and triazolam 0.125 mg were given alone and in combination, all drug treatments improved subjective sleep depth or quality, though no data were presented. A study of melatonin 50 mg IV also found no changes in reported sleep quality (9). When melatonin 80 mg was given at 9:00 p.m. and sleep recordings were started at 10:30 p.m. in normal volunteers exposed to tape recordings of traffic noise, there were no significant changes in subjective measures of sleep and awakening quality, though there was a trend in the direction of improvement (10). Among the studies of insomniacs is that of James et al. (11), in which melatonin I and 5 mg and placebo were given at 10:45 p.m. to polysomnographically prescreened insomniacs. The patients reported improved sleep quality, though they were not more rested in the morning and believed that their total sleep time had been shorter on melatonin nights. When melatonin 5 mg was given at 8:00 p.m. for 1 week to patients with psychophysiological insomnia, there was no reported change in sleep quality; 8 of the 15 were unable to distinguish the period of active melatonin treatment

,

(12). Hughes et al. (unpublished) gave immediate release and controlled release preparations of melatonin 0.5 mg 30 minutes before sleep, and an immediate release preparation of 0.5 mg halfway through the night, to polysomnographically prescreened elderly patients with sleep maintenance insomnia. They found no melatonin-induced changes in reported sleep quality or daytime measures of mood and alertness. In a letter to the editor, Wurtman and Zhdanova (13) summarized a study in which 0.3 mg of melatonin was given 30 minutes before bedtime to elderly insomniacs. Melatonin plasma concentrations were not presented. It was stated that sleep quality was better, although there was no presentation of the data. In summary, there is very little systematically presented data supporting the view that melatonin might improve self-rated sleep quality in normal volunteers or insomniacs. Given the widespread use of melatonin purchased at health food stores, it is not surprising that the Consumer's Union performed a survey of satisfaction with its effectiveness (14). They reported that approximately one quarter of respondents said that it was "very helpful", roughly half the rate of such favorable responses for prescription hypnotics. Melatonin was found to have the highest frequency for any treatment of being considered "not at all helpful". Of course, the persons surveyed were taking a variety of doses under a multitude of conditions, but this gives an indication of the limited consumer satisfaction with melatonin as it is used in the community.

Polygraphic studies of normal volunteers James et al. (5), as described earlier, administered 1 and 5 mg of melatonin and placebo at 10:45 p.m. to 10 volunteers and began sleep recordings 15 minutes later. To obtain a sense of blood concentrations achieved by doses in this range, a parallel study in the laboratory found that after a 2-mg oral dose given to normal volunteers in the evening, plasma concentrations rose to a mean of 69 pg/ml within 30 minutes. James et al. (5) found no significant changes in traditional measures of hypnotic efficacy such as sleep latency, wake time after sleep onset, or total sleep time. A single variable was significantly altered: rapid eye movement (REM) latency was increased by the 5-mg dose. Zhdanova et al. (15) administered melatonin 0.3 and 1 mg and placebo to six subjects at 6:00, 8:00, and 9:00 p.m. Melatonin blood concentrations were not reported. The only polygraphic measures reported were latencies until sleep onset (defined as 1.5 minutes of stage 1), stage 2 sleep, and REM sleep. It was found that both doses of melatonin reduced latencies until sleep onset and stage 2. Measures such as total sleep time, wake time after sleep onset, and sleep efficiency Sleep, Vol. 20, No. 10, 1997

918

W. B. MENDELSON

were not presented. REM latency decreased after melatonin 0.3 rug in half of the subjects, but this did not reach statistical significance. Zhdanova et a1. (15) gave 0.3 and 1 rug of melatonin and placebo to 12 volunteers at 9:00 p.m. and began sleep recordings after 1 hour. Latencies to sleep onset and stage 2 were decreased, resulting in an increase in sleep efficiency. There was a nonsignificant tendency toward decreasing REM latency at the lower dose. Wake time after sleep onset, number of awakenings, and percentage of sleep stages were not altered. Data on total sleep time were not presented. Among low dose studies using shorter or specialized recording conditions is that of Tzichinsky and Lavie (16), who examined the sleep propensity function in a paradigm in which subjects are given a 7-minute opportunity to sleep, then are kept awake for 13 minutes in a repeating pattern. In this model, administration of melatonin 5 mg at 5:00 and 7:00 p.m. induced increased sleepiness 2 hours later; a more rapid onset of action was seen following administration at 9:00 p.m. In a study of 3 and 6 mg of melatonin given 120 and 30 minutes before a 6:00 p.m. nap opportunity, Nave et a1. (7) found that melatonin given at both times reduced sleep latency and raised total sleep time. Studies of higher doses include that of Ferini-Strambi and Zucconi (8), described above, in which melatonin 100 mg was given alone and in combination with triazolam 0.125 mg. Melatonin did not alter traditional measures of hypnotic efficacy such as total sleep time, sleep latency, wake time after sleep onset, or sleep efficiency, although it was reported to change a nontraditional variable, the cyclic alternating pattern. In a study of melatonin 80 mg given to normal volunteers exposed to recordings of traffic noise, Waldhauser et aI. (10) reported decreases in sleep latency and number of awakenings, and an increase in sleep efficiency. In contrast, melatonin did not alter total sleep time or total wake time. When Cramer et a1. (9) administered melatonin 50 mg intravenously to normal volunteers at 9:30 p.m., sleep latency was decreased. Total sleep time was unchanged, and there was an increase in awakenings and stage 1 sleep. Polygraphic studies of young and elderly insomniacs Two studies have examined low dose melatonin treatment in insomniacs whose sleep was characterized by prestudy polysomnography. As described earlier, James et a1. (11) administered melatonin 1 and 5 mg and placebo for one night each to 10 insomniacs with a mean age of 33.4 years. Medication was given at 10: 45 p.m., and sleep recordings were begun 15 minutes later. Melatonin did not alter sleep latency, sleep effiSleep, Vol. 20, No. 10, 1997

ciency, total sleep time, or wake time after sleep onset. The only significantly altered variable was REM latency, which increased with the I-mg dose. Hughes et aI., whose study of elderly insomniacs with sleep maintenance insomnia is described above, reported that 0.5 rug of melatonin in both immediate and controlled release preparations reduced sleep latency but did not alter wake time after sleep onset (an important variable in sleep maintenance insomnia) or total sleep time. Other studies include that of Rogers et a1. (17), in which elderly sleep maintenance insomniacs were given transbuccal melatonin 0.5 mg or placebo. It was found that melatonin increased the length of awakenings during the first half of the night and decreased it in the second half. All other variables such as sleep latency, total sleep time, waking time after sleep onset, and sleep efficiency were unaffected. In a letter to the editor, Wurtman and Zhdanova (13) summarized a study of administration of 0.3 mg of melatonin given 30 minutes before bedtime to nine elderly insomniacs. It was said that sleep onset latency, awakenings, and movements were improved; the effects on total sleep were not stated. CONCLUSIONS As we consider these rather mixed results, it seems important to also consider that the effects of melatonin on sleep may not be completely benign. Middleton et a1. (18), in an actigraphic study, have raised the possibility that administration of melatonin ruay result in irregular sleep-wake cycles in some volunteers under the paradigm they employed. In the intravenous administration study of Cramer et a1. (9), wake time was increased, a finding similar to the increased length of arousals noted during the first half of the night in the intrabuccal administration study of Rogers et a1. (17). Maksoud et a1. (19), in a study of 12 obstructive sleep apnea patients receiving continuous positive airway pressure (CPAP), found that 6 rug of timed release melatonin was associated with worsening of apneahypopnea index, mean apnea duration, and longest apnea duration. If the latter data are supported by further studies, the implication would be that it is ill-advised for patients with sleep disturbance to take melatonin without first having ruled out sleep apnea, and that melatonin might be contraindicated in patients with known sleep apnea who wish medication for sleep in addition to their CPAP. Although it is beyond the area of this review, it should also be noted that studies of melatonin's possible cognitive actions need to be considered. One recent report of 5 mg given in the daytime indicated increases in visual decision and response time latencies

THE HYPNOTIC EFFICACY OF MELATONIN

919

for up to 7 hours, a duration close to that of a normal suggest that it has any significant clinical hypnotic efnocturnal sleep period (20). Other studies of lower ficacy. doses (e.g. 15) have not noted effects on simple or four-choice reaction time when tested 10 hours later REFERENCES the following morning. The former study, however, I. Mendelson WB. Efficacy of melatonin as a hypnotic agent. J does raise the possibility that at 5-mg or higher doses, Biol Rhythms, 1997 (in press), patients getting up in the morning might still be ex2. Gallup Organization. Sleep in America. Washington, DC: National Sleep Foundation, 1990. periencing cognitive effects of melatonin. Certainly, 3. Mendelson WB. Long-term follow-up of chronic insomnia. this would have to be explored before recommending Sleep 1995;18:698-701. melatonin for widespread use as a hypnotic. 4. Kales A, Kales JD, Bixler EO, Scharf MB. Effectiveness of hypnotic drugs with prolonged use: flurazepam and pentobarAnother possibility is that the timing of melatonin bital. Clin Pharmacol Ther 1975;18:356-63. administration may be crucial, as evidenced by some 5. James SP, Mendelson WB, Sack DA, Rosenthal NE, Wehr TA. data that improved sleep latency is seen only with adThe effect of melatonin on normal sleep. Neuropsychopharmacology 1987;1:41-4. ministration 1 or 2 hours before bedtime (e.g. 6,7). It 6. Zhdanova IV, Wurtman RJ, Lynch HL, et al. Sleep-inducing may turn out, for instance, that its effectiveness will effects of low doses of melatonin ingested in the evening. Clin Pharmacol Ther 1995;57:552-8. be manifest only when given following careful labo7. Nave R, Peled R, Lavie P. Melatonin improves evening napping. ratory studies to characterize the patient's circadian Eur J Pharmacol 1995;275:213-6. rhythms of sleep and temperature. If that should be the 8. Ferini-Strambi L, Zucconi ME. Effects of melatonin on sleep microstructure: preliminary results in healthy subjects. Sleep case, then it would become part of the specialist's ar1993;16:744-7. mamentarium, but hardly a useful tool for widespread 9. Cramer H, Rudolph J, Consbruch U, Kendel K. On the effects use in the primary care clinic or when bought over the of melatonin on sleep and behavior in man. Adv Biochem Psychopharmacol 1974; 11: 187-91. counter. This might explain the relatively disappoint10. Waldhauser F, Sa1etu B, Trinchard-Lugan I. Sleep laboratory ing results currently reported in "real world" use (14). investigations on hypnotic properties of melatonin. PsychoAs the above review indicates, the strongest part of pharmacology 1990; I 00:222-6. a relatively weak case for melatonin's efficacy seems II. James Sp, Sack DA, Rosenthal NE, Mendelson WB. Melatonin administration in insomnia. Neuropsychopharmacology 1989;3: to be that it might be useful primanly for aiding sleep 19-23. onset latency. Indeed, measures of wake time after 12. Ellis CM, Lemmens G, Parkes JD. Melatonin and insomnia. J Sleep Res 1996;5:61-5. sleep onset in elderly sleep maintenance insomnia have 13. Wurtman RJ, Zhdanova I. Improvement of sleep quality by melbeen disappointing [(18), Hughes et al. (unpublished)]. atonin (Letter to the Editor). Lancet 1995;346:1491. If this is the case, one would have to address the com- 14. Consumer's Union. Overcoming insomnia. Consumer Rep 1997; mon observation that many insomniacs have mixed 15. 62:10-3. Zhdanova IV, Wurtman RJ, Morabito C, Piotrovska VR, Lynch complaints of difficulty with both sleep onset and HJ. Effects of low oral doses of melatonin given 2-4 hours before habitual bedtime, on sleep in normal young humans. maintenance (e.g. 2,3). Thus, if efficacy were to be Sleep 1996;19:423-31. established only for sleep induction, the proportion of 16. Tzichinsky 0, Lavie P. Melatonin possesses time-dependent hypnotic effects. Sleep 1994; 17 :638-45. patients who might benefit might be relatively small. It would be important not to lose track of one aspect 17. Rogers NL, van den Heuvel CJ, Lushington K, Kennaway D, Dawson D. Effect of transbuccal melatonin administration on of the media attention, however: even though (in my sleep in elderly insomniacs. Sleep Res 1997;26:482. view) melatonin may not live up to expectations, there 18. Middleton EA, Stone BM, Arendt J. Melatonin and fragmented sleep patterns. Lancet 1996;348:551-2. is a very great interest in, and demand for, a successful 19. Maksoud A, Moore CA, Harshkowitz M. The effect of melahypnotic agent. Both clinical experience and laboratonin administration on patients with sleep apnea. Sleep Res 1997;26:114. tory studies of pill self-selection (e.g. 21) indicate that 20. Rogers NL, Phan 0, Kennaway D, Dawson D. Effect of oral insomniacs are a remarkably treatment-seeking group. daytime melatonin administration on cognitive psychomotor performance in humans. Sleep Res 1997;26:212. There is also a certain intuitive appeal to seeking a T, Pedrosi B, Rosenthal L, Roth T. Hypnotic self adnatural substance to assuage one's ills. Nonetheless, in 21. Roehrs ministration and dose escalation. Psychopharmacology 1996; the case of melatonin, I believe it very premature to 127:150-4.

Sleep, Vol. 20, No. 10, 1997