Periodic Leg 1vIovements (PLM): Their Relationship to Sleep Stages

Transcription

1 Sleep. 16(6): American Sleep Disorders Association and Sleep Research Socit,ty Periodic Leg 1vIovements (PLM): Their Relationship to Sleep Stages *Thomas Pollmacher and thartmut Schulz * Max Planck Institute of Psychiatry. Clinical Institute, Department of Psychiatry, Munich, F.R. G.; and t Free University of Berlin, University Clinic, Department of Psychiatry, Berlin, F.R.G. Summary: We investigated the characteristics of periodic leg movements (PLM) during nocturnal sleep and wakefulness in 13 drug-free patients presenting with the restless legs syndrome (RLS, n = 9) or with isolated PLM (n = 4). Eight-hour polygraphic sleep recordings included the electromyogram (EMG) of both tibialis anterior muscles. Scoring ofleg movements was done according to established criteria for periodic movements in sleep, but movements occurring during episodes of wakefulness were scored as well. Twelve out of 13 patients had PLM during wakefulness, including three subjects not affected by RLS. The frequency of periodic movements in sleep (PMS) per hour of total sleep time was significantly lower than the frequency ofplm (including movements during wakefulness) per hour of polygraphic recording. Movement indices based on PMS alone underestimated the relative frequency of PLM particularly in patients with high amounts of wakefulness (> 20%). All features of PLM clearly differed between sleep stages. Relative frequency of movements, their duration and their arousing effect decreased along the nonrapid eye movement (NREM) sleep stages, whereas the intermovement interval increased. During rapid eye movement (REM) sleep the duration of movements was shortest and the intermovement interval was longest. The results presented suggest that the processes underlying PLM are most active at the transition from wakefulness to sleep and considerably attenuated duiing deep NREM sleep and even more during REM sleep. We suggest including movements during wakefulness in routine PLM scoring to get a more complete picture of the disturbance. Key Words: Sleep-Periodic leg movements (PLM)-Restless legs syndrome. Periodic leg movements (PLM) in sleep were first described by Symonds (1), who named them "nocturnal myoclonus". They were first recorded polygraphically by Lugaresi et al. in 1965 (2). PLM occur in various sleep disorders, including sleep apnea (3) and narcolepsy (4), but they may also be seen without any other evidence of a sleep-related disorder (5). In patients suffering from the restless legs syndrome (RLS; 6) PLM are almost always present during night sleep (7). PLM are short-lasting movements of the lower limbs. The movements repeat quite regularly in intervals of seconds (7). They may lead to partial or full arousals, and the degree of sleep fragmentation is highly variable (8). Subjective complaints include insomnia or excessive daytime sleepiness (9). The pathophysiological basis of PLM is unknown. Because of their resemblance to the Babinski response, some investigators attribute them to a suppression of Accepted for publication April Address correspondence and reprint requests to Dr. Thomas Pollmacher, Max Planck Institute of Psychiatry, Clinical Institute, Department of Psychiatry, Kraepelinstrasse 10, W-8000 Munich 40, Federal Republic of Germany. 572 supraspinal inhibiting pathways (10,11). An involvement of either the dopaminergic or the opioid system is proposed on the basis of findings indicating therapeutic effectiveness of opioids (12-14) and dopaminergic drugs (15,16). It has been hypothesized that PLM may result from rhythmical fluctuations of reticular excitability (17), and electrophysiological studies support the hypothesis that the underlying mechanisms are operative at the pontine level or rostral to it (18). However, studies in patients with myelopathy strongly suggest that PLM can be generated in the spinal cord (19). On the other hand, an involvement of the peripheral nervous system has been suggested on the basis of neurophysiological findings (20,21), and even alterations in blood flow to the legs have been introduced as a possible pathophysiologically relevant factor (22). Traditionally, the quantification of nocturnal PLM is based exclusively on those events occurring during sleep (periodic movements in sleep, PMS). It has been shown that PMS are most frequent during nonrapid eye movement (NREM) sleep stages 1 and 2, less frequent during NREM sleep stages 3 and 4 (slow wave sleep) and almost absent during rapid eye movement

2 PLM AND SLEEP STAGES 573 ~ (REM) sleep (7). However, it has also been reported that in patients in whom PMS are found during sleep, PLM of the same type also occur during quiet wakefulness, i.e. before sleep onset or in the course of nocturnal waking episodes (2,14). In a single subject, the intervals between PLM were shown to be shortest during wakefulness and to increase with the depth of sleep (23). However, in seven patients investigated in our laboratory (24), only subjects who had a relatively short intermovement interval in waking demonstrated a lengthening of this interval with the transition from wakefulness to sleep. The aim of the present study was to investigate the relationship of the characteristic features of PLM to the state of vigilance in more detail, including some aspects that have not been addressed until now, namely the relationship of the duration of movements and their arousing effect to sleep stages. METHODS Patients and polygraphic recording Diagnostic sleep recordings of 13 unrelated patients (eight men, five women; mean age: 42.2 years; range years) were included in the present analysis. Nine patients had been diagnosed clinically as suffering from the restless legs syndrome (familial RLS in six cases). In the remaining four subjects, PLM in sleep were found without any evidence of another definite sleep disorder like sleep apnea or narcolepsy. Subjects did not take specific medication (benzodiazepines, carbamazepine, opioids, dopamine agonists or related drugs) for at least 2 weeks prior to the investigation. In two subjects taking nifedipine (20 and 30 mg, respectively) for the control of mild hypertension, this medication was not discontinued. In all patients, an 8-hour polygraphic sleep recording starting at approximately 2300 hours was performed, including the registration of the electroencephalogram (EEG) (C3-A2, C4-A 1), electrooculogram (EOG) and mental as well as submental electromyogram (EMG) according to Rechtschaffen and Kales (25). The surface EMG of both tibialis anterior muscles was recorded as described by Coleman (7). Sleep staging followed the recommendations of Rechtschaffen and Kales (25). Scoring of periodic leg movements Leg movements were scored according to the following criteria: 1) amplitude of EMG pen deflection had to be at least twice the pre sleep waking level; 2) critical duration of leg movements was between 0.5 and 5 seconds. Movements were scored irrespective of whether they occurred in one or in both legs. The duration of each movement was measured and identified according to whether it led to partial (duration < 15 seconds) or full arousal (duration;::: 15 seconds). To assess the reliability of the scoring of PLM duration and of their arousing effect, a random subset of three recordings was evaluated by two scorers independently. The interrater correlation coefficient for the duration of movements was 0.79 (n = 587, p < 0.001). With respect to the scoring of arousals (partial and full ones combined), the interrater agreement was 89.0%. The interval between the onset of successive movements was calculated. Movements were defined as periodic if they occurred in a series of at least three movements seconds apart. According to the criteria of Coleman (7), excluding movements during wakefulness, the number of periodic movements per hour of total sleep time (PMS index) and the number of movements with partial or full arousal per hour oftotal sleep time (PMS-arousal index) were calculated. Similar indices were computed in relation to time in bed including PLM during both sleep and wakefulness: the number of movements per hour of time in bed (PLM index) and the number of movements with partial or full arousal per hour of time in bed (PLM-arousal index). Although rules for scoring arousals (7) cannot properly be applied to PLM during wakefulness, they all were included when calculating the PLM-arousal index because they may prevent subjects from going or returning to sleep. Data analysis and statistics Stages 3 and 4 were pooled as slow wave sleep (SWS) because PLM were relatively infrequent during these sleep stages. For parts of the data analysis, stage 2 and SWS were pooled (stages ). For each subject and each sleep stage, the PLM index and PLM -arousal index as well as the relative arousal index (PLM-arousal index/plm index), the mean duration of movements and the mean intermovement interval were calculated. Statistical analysis of the differences between sleep stages with respect to movement characteristics was done by Friedman analysis of variance. The Wilcoxon matched-pairs signed-ranks test was used for post-hoc pairwise comparison. To evaluate the relationship of the duration of movements to their arousing effect, movements were grouped into classes according to their duration and the relative distribution of those movements with and without arousal was compared for the different sleep stages by the chi-square test. The level of significance was set to 5%. Two-tailed p values are reported and all figures in the text are expressed as mean ± standard deviation, if not stated otherwise. Sleep, Vol. 16, No.6, 1993

3 574 T. POLLMAcHER AND H. SCHULZ TABLE 1. I nterindividual variability of periodic leg movements (PLM) movements per hour 100, , Mean ± SD Range n ofplm 227 ± % movements occuning during sleep (PMS) 70.0 ± PLM index 28.9 ± PMS index 23.5 ± PLM-arousal index 22.5 ± PMS-arousal index 16.1 ± n of PMS series 10.9 ± n of movements per series 17.3 ± Abbreviations: PLM index = periodic movements per hour of time in bed (TlB), PLM-arousal index = periodic movements with partial or full arousal per hour oftlb, PMS index = periodic movements in sleep per hour of total sleep time (TST), PMS-arousal index = periodic movements with partial or full arousal per hour of TST. RESULTS Frequency of leg movements (see Table 1) In the 13 polygraphic recordings there were 2,954 periodic leg movements, and 70.0 ± 28.0% of these movements occurred during sleep. The total number ofplm varied between 36 and 892 per sleep recording; % of those movements occurred during sleep. The PMS index ranged from /hour and the PLM index from /hour. The PMS-arousal index varied between 0.7 and 64.3/hour and the PLMarousal index between 2.1 and 91.3/hour. Both the PMS index and the PMS-arousal index were significantly lower than the PLM index (23.5 ± 24.4 vs ± 26.6; p < 0.05) and the PLM-arousal index (16.1 ± 16.2 vs ± 22.0; p < 0.05), respectively. This indicates that excluding movements during wakefulness results in an underestimation of the relative fn: quency of leg movements during bedtime. The mean difference between the PMS index and the PLM index was -6.4 ± 9.1, and it was -5.4 ± 7.3 between the PMS-arousal index and the PLM -arousal index. As illustrated in Fig. 1, the underestimation of the relative frequency of periodic movements by the indices based exclusively on PMS was most prominent in those subjects with high amounts of wakefulness (> 20% oftime in bed). Because PLM indices were calculated relative to time in bed and PMS indices relative to total sleep time, this finding cannot be explained simply by the higher number of PLM as compared to PMS. Movement indices by sleep stages (see Table 2) PLM were observed in seven out of 13 subjects during SWS and in nine out of 13 during REM sleep. During wakefulness, 12 out of 13 subjects had periodic leg movements. Three of these subjects did not suffer from RLS. The mean PLM-index during wakefulness ~ ~ % awake "PLM PMS FIG. 1. PLM and PMS-arousal index as well as the difference between both indices in relationship to the percentage of time awake relative to time in bed. Each pair of a hatched and a black column represents one subject. for these patients was somewhat lower (39.0 ± 31.5; range ) as compared to those subjects suffering from RLS (47.1 ± 39.2; range ). The two subsamples, however, were too small for statistical comparison. For the total group of subjects, the PLM index did not differ between stages W (wake), 1 and 2, but was significantly lower in SWS and REM sleep. In contrast, the PLM-arousal index was significantly lower in stage 2 than in stage 1 and even lower in SWS and REM sleep. Duration of movements and intermovement interval (see Table 3) The mean duration of PLM was 2.6 ± 1.2 seconds. The duration was longest during stage W. It was significantly shorter during stages and shortest during REM sleep. The mean intermovement interval was 34.1 ± 22.9 seconds. Although the mean intermovement interval did not differ between stage Wand stage 1, it was significantly longer during stages and longest during REM sleep (p = 0.05). As illustrated by Fig. 2, there was a somewhat greater interindividual variability of intermovement intervals TABLE 2. Movement indices for the different sleep stages (mean number of movements per hour ± SD) Stage W Stage I Stage 2 SWS REM sleep PLM index 41.6 ± ± ± ± ± 21.3 PLM-arousal index 41.6 ± ± ± 17.7h 4.5 ± II.4 b 4.9 ± 9.7 Abbreviations: PLM index = periodic movements per hour, PLMarousal index = periodic movements with partial or full arousal per hour. For both indices Friedman two-way ANOV A indicated a significant difference between sleep stages (p < 0.05). o p < 0.05 vs. next upper column. b p < 0.01 vs. next upper column. Sleep. Vol. 16. No

4 PLM AND SLEEP STAGES 575 sec 70, , ~~><~ ~~~~~~~:;; OL ~ Awake Stage 1 FIG. 2. Mean PLM intermovement intervals during waking and during stage I for the 12 subjects showing PLM during both sleep stages. The mean of all patients is marked by an arrow. during stage W as compared to stage 1. Subjects showing long intervals during this stage tended to have shorter ones during stage 1, whereas most of those with short intervals during the waking state showed longer ones during stage 1. Arousing effect (see Tables 3, 4, and 5) The PLM index and the PLM-arousal index both were lowest during SWS and REM sleep. From this, it cannot be concluded that the probability that movements have an arousing effect is lower during SWS and REM sleep as compared to the other sleep stages. We therefore calculated the relative arousal index by dividing the PLM-arousal indices by the PLM indices. Table 3 shows that this relative index was highest (1.0 for all subjects) during stage W, because all movements occurring during wakefulness were included when calculating the PLM-arousal index (see Methods). During stage 1, the relative arousal index was significantly higher as compared to the combined stages and REM sleep. Finally, the hypothesis was tested that the arousing effect of movements is related to their duration. Because the arousing effect differed between sleep stages, only PLM during stage 2 were included for this part of the analysis. Table 4 illustrates that the percentage of movements with arousal increased with duration of PLM. This effect was especially pronounced for movements of > 3 seconds in duration. However, the difference between sleep stages with respect to the arousing effect of PLM cannot be exclusively accounted for by the difference in the duration of movements. The data in Table 5 show that the arousing effect differs significantly between sleep stages, even if only movements in a narrow range of duration (1-2 seconds) are selected. TABLE 3. Duration of PLM, intermovement interval and relative arousal index for the different sleep stages (mean ± SD) Stage W Stage I Stages REM sleep Duration a 2.9 ± ± 0.6' 2.3 ± 0.5 d 1.8 ± 0,7d Intermovement interval" 33.3 ± ± ± 5.6 d 53.2 ± 18.3' Relative arousal index h 1.00 (by definition) 0.88 ± ± 0.24' 0.65 ± 0.31 For all variables Friedman two-way ANOV A indicated a significant difference between sleep stages (p < 0.05). a Measurements given in seconds. h Relative arousal index = PLM-arousal index/plm index. c p < 0.09 vs. next upper column. d p < 0.05 vs. next upper column., p < 0.01 vs. next upper column. fp = DISCUSSION The present study confirms earlier reports indicating that PLM do not occur exclusively during sleep, but also during wakefulness (2,14,23,24). The results presented here extend the knowledge on PLM during wakefulness with respect to the following points: 1) in our series of 13 unrelated patients, all but one presented with PLM during wakefulness, suggesting that their occurrence during the waking state is the rule rather than the exception; 2) PLM during wakefulness were not restricted to patients suffering from RLS, as already suggested by Lugaresi et al. (26) and 3) excluding movements during wakefulness from the calculation of movement indices may lead to an underestimation of the relative frequency of PLM, which is especially pronounced in those patients who spend a considerable time awake. This is in line with the finding that movements were significantly more frequent during wakefulness as compared to sleep stages 2, SWS and REM. Because up to 82.8% ofplm occurred during wakefulness, our results strongly support the recommendation of Montplaisir et al. (27) to score PLM during both wakefulness and sleep in patients suffering from RLS in order to get a more complete picture of the motor disturbance. The finding that three out of four subjects without RLS as well had PLM during wakefulness suggests that this recommendation should even be generalized to all patients with PLM. Further research, however, is needed to clarify the clinical usefulness of PLM scoring during wakefulness for diagnosis and treatment evaluation. Regarding the relationship of specific features ofplm to the state of vigilance, the results confirm earlier reports that movements become less frequent and the interval between movements becomes longer as sleep deepens along the NREM sleep stages (7,23). However, Sleep. Vol. 16. No

5 576 T. POLLMACHER AND H. SCHULZ TABLE 4. The relationship between the duration of PLM and their arousing effect TABLE 5. The arousing effect of PLM with a duration of 1-2 seconds during different sleep stages Movements with arousal Duration n of ofplmh movements n % > > > > Total 1, Chi-square = 48.09, df= 4, p < a The analysis is based exclusively on PLM occurring during sleep stage 2. h Measurements given in seconds. we could not confirm the finding of Montplaisir et a1. (23) that the intermovement interval is longer during stage I as compared to wakefulness. In the present study, mean intervals did not differ significantly between these stages, which is in accordance with a previous report from our laboratory (24). Variability was, however, lower during stage I, suggesting that processes underlying PLM become more stable at the onset of sleep. Our results support the assumption that the duration of movements decreases along the NREM sleep stages. The processes underlying PLM seem to be maximally suppressed during REM sleep, when the relative frequency was as low as during SWS, movement duration was even shorter and the intermovement intervals tended to be longer. The relationship ofthe arousing effect ofplm to the state of vigilance has not been investigated so far. The present results show that the probability that a movement occurring during stage I leads to an arousal is significantly higher than during the other NREM sleep stages and REM sleep. However, the arousing effect of PLM does not only depend on the state of vigilance but also on the duration of the movements. This is suggested by the finding that the percentage of PLM followed by EEG arousal increases with increasing duration of the movements. Based on the assumption that the arousing effect of the movements is probably crucial for the induction of sleep disturbances, the duration of PLM may prove to be an important parameter when studying the influence of drugs on PLM in the future. In summary, the results of the present study support the assumption that the processes underlying PLM are most active at the transition from wakefulness to sleep and considerably suppressed during NREM and even more during REM sleep. Because the traditional restriction of PLM scoring does not yield a complete picture of the disturbance, we suggest including PLM during wakefulness in routine PLM scoring. Since the arousing effect of movements is related to their du- Sleep, Vol. 16, No.6, 1993 Movements with arousal n of Sleep stage movements n % Stage Stage SWS REM sleep Total Chi-square = 76.5, df = 3, p < ration, the latter may prove to be a useful additional parameter, especially when studying drug effects on PLM. REFERENCES I. Symonds CPo Nocturnal myoclonus. J Neurol Neurosurg Psychiatry 1953;16: Lugaresi E, Tassinari CA, Coccagna G, Ambrosetto C. Particularites cliniques et polygraphiques du syndrome d'impatience des membres inrerieurs. Rev Neurol (Paris) 1965; 113: Ancoli-Israel S, Kripke DF, Mason W, Kaplan OJ. Sleep apnea and periodic movements in an aging sample. J GT!r()J1tol 40; 1985: Bedard MA, Montplaisir J, Godbout R. Effect of I-dopa on periodic movements in sleep in narcolepsy. Eur Neurol 1987; 27: Coleman RM, Pollak CP, Weitzman ED. Periodic movements in sleep (nocturnal myoclonus): relation to sleep disorders. Ann NeuroI1980;8: Ekbom KA. Restless legs syndrome. Neurology (Minneap) 1960; 10: Coleman RM. Periodic movements in sleep (nocturnal myoclonus) and restless legs syndrome. In: Guilleminault C, ed. Sleeping and waking disorders. Menlo Park, CA: Addison-Wesley, 1982: Rosenthal L, Roehrs T, Sicklesteel J, Zorick F, Wittig R, Roth T. Periodic movements during sleep, sleep fragmentation, and sleep-wake complaints. Sleep 1984;7: Sa skin P, Moldofsky H, Lue FA. Periodic movements in sleep and sleep-wake complaint. Sleep 1985;8: Smith RC. Relationship of periodic movements in sleep (nocturnal myoclonus) and the Babinski sign. Sleep 1985;8: II. Mosko SS, Nudleman KL. Somatosensory and brainstem auditory evoked responses in sleep-related periodic leg movements. Sleep 1986;9: Trzepacz PT, Violette EJ, Sateia MJ. Response to opioids in three patients with restless legs syndrome. Am J Psychiatry 1984; 141: Hening WA, Walters A, Kavey N, Gidro-Frank S, Cote L, Fahn S. Dyskinesias while awake and periodic movements in sleep in restless legs syndrome: treatment with opioids. Neurology 1986;36: Walters A, Hening W, Cote L, Fahn S. Dominantly inherited restless legs with myoclonus and periodic movements of sleep: a syndrome related to the endogenous opiates? In: Fahn S, Marsden CD, Woest VMC, eds. Advances in neurology, Vol. 43: Myoclonus. New York: Raven Press, 1986: Akpinar S. Restless legs syndrome treatment with dopaminergic drugs. Clin Neuropharmacol1987; I 0: Montplaisir J, Godbout R, Poirier G, Bedard MA. Restless legs syndrome and periodic movements in sleep: physiopathology and treatment with I-dopa. Clin NeuropharmacoI1986;9: Lugaresi E, Coccagna G, Mantovani M, Lebrun R. Some pe-

6 PLM AND SLEEP STAGES 577 riodic phenomena arising during drowsiness and sleep in man. Electroencephalogr Clin Neurophysio/1972;32: Wechsler LR, Stakes JW, Shahani BT, Busis NA. Periodic leg movements of sleep (nocturnal myoclonus): an electrophysiological study. Ann Neurol 1986; 19: Yokota T, Hirose K, Tanabe H, Tsukagoshi H. Sleep-related periodic leg movements (nocturnal myoclonus) due to spinal cord lesion. J Neurol Sci 1991;104: Bliwise DL, Ingham RH, Date ES, Dement We. Nerve conduction and creatinine clearance in aged subjects with periodic movements in sleep. J Geronto/1989;44:M Salvi F, Montagna P, PI as mati R, et al. Restless legs syndrome and nocturnal myoclonus: initial clinical manifestation of familial amyloid polyneuropathy. J Neurol Neurosurg Psychiatry 1990;53: Watanabe S, Ono A, Naito H. Periodic leg movements during either epidural or spinal anesthesia in an elderly man without sleep-related (nocturnal) myoclonus. Sleep 1990; 13: Montplaisir J, Godbout R, Boghen D, DeChamplain J, Young SN, Lapierre G. Familial restless legs with periodic movements in sleep: electro physiologic, biochemical, and pharmacologic study. Neurology 1985;35: Schulz H, Kiss E. Sleep disturbances caused by rest-dependent muscular and motor dysfunction. In: Peter JH et ai., eds. Sleep and health risk. Berlin: Springer, 1991: Rechtschaffen A, Kales A, eds. A manual of standardized terminology, techniques and scoring system for sleep stages of human subjects. U.S. Department of Health, Education, and Welfare, Public Health Service-National Institutes of Health, National Institute of Neurological Diseases and Blindness, Neurological Information Network, Bethesda, MD, Lugaresi E, Cirignotta F, Coccagna G, Montagna P. Nocturnal myoclonus and restless legs syndrome. In: Fahn S, Marsden CD, Woest VMC, eds. Advances in neurology, Vol. 43: Myoclonus. New York: Raven Press, 1986: Montplaisir J, Lapierre 0, Warnes H, PelletierG. The treatment of the restless legs syndrome with or without periodic leg movements in sleep. Sleep 1992; 15: Sleep. Vol. 16. No