Circadian rhythm of restless legs syndrome Relationship with biological markers.код для вставкиСкачать
Circadian Rhythm of Restless Legs Syndrome: Relationship with Biological Markers Martin Michaud, PhD,1,2 Marie Dumont, PhD,2,3 Brahim Selmaoui, PhD,3 Jean Paquet, PhD,1,3 Maria Livia Fantini, MD, MSc,1,2 and Jacques Montplaisir, MD, PhD1,2 Recently, it was suggested that the intensity of restless legs syndrome (RLS) symptoms may be modulated by a circadian factor. The objective of this study was to evaluate, during a 28-hour modified constant routine, the nycthemeral or circadian variations in subjective leg discomfort and periodic leg movements (PLMs) and to parallel these changes with those of subjective vigilance, core body temperature, and salivary melatonin. Seven patients with primary RLS and seven healthy subjects matched for sex and age entered this study. Although the symptoms were more severe in patients than in controls, a significant circadian variation in leg discomfort and PLM ( p < 0.01) was found for both groups. In both groups, the profiles of leg discomfort and PLM were significantly correlated with those of subjective vigilance, core body temperature, and salivary melatonin. However, among these variables, the changes in melatonin secretion were the only ones that preceded the increase in sensory and motor symptoms in RLS patients. This result and those of others studies showing that melatonin exerts an inhibitory effect on central dopamine secretion suggest that melatonin might be implicated in the worsening of RLS symptoms in the evening and during the night. Ann Neurol 2004;55:372–380 Restless legs syndrome (RLS) is a sensorimotor disorder found in approximately 10% of the white population.1–3 Its diagnosis relies on the presence of four mandatory clinical features, namely, (1) an urge to move the legs usually accompanied by unpleasant sensations felt deeply at the level of the lower limbs, (2) a beginning or a worsening of the symptoms during periods of rest or inactivity, (3) a partial or total relieving of the symptoms by movements, and (4) a worsening of the symptoms in the evening or during the night.4 Another feature of RLS seen in most patients is the presence of stereotyped and recurring movements of the lower limbs. These so-called periodic leg movements (PLMs) occur during both sleep and wakefulness and are characterized by the extension of the big toe and dorsiflexion of the ankle, with occasional flexion of the knee and hip. Several factors may contribute to the worsening of RLS symptoms in the evening and during the night. One factor is the increase of sleepiness in the evening compared with the daytime. Indeed, it is quite frequent for patients with RLS to report their symptoms to be worst when they are excessively tired or sleep deprived. Another contributing factor is the decrease of motor activity in the evening compared with the daytime.5 Because RLS symptoms are known to be worsened by immobility,6 it is likely that a decrease in motor activity in the evening could be responsible for the increase in RLS symptoms severity. A third factor would be that the worsening of symptoms is the manifestation of an intrinsic circadian rhythm in RLS symptomatology. Recently, two studies used modified constant routine protocols7 to investigate the circadian pattern in the occurrence of RLS symptoms.8,9 The symptoms of RLS (leg discomfort and PLM or motor restlessness) were quantified using a modified version of the “Suggested Immobilization Test” (SIT),6 which was administered every 3 or 4 hours during the protocols. Both studies showed that the severity of leg discomfort followed a circadian rhythm with a maximum occurring shortly after midnight. They also showed that the peak intensity of symptoms occurs on the falling limb of the core body temperature rhythm. Together, these two From the 1Sleep Disorders Center, Sacré-Cœur Hospital; 2Faculty of Medicine, University of Montreal; and 3Chronobiology Laboratory, Sacré-Cœur Hospital, Montreal, Quebec, Canada. Address correspondence to Dr Montplaisir, Centre d’étude du sommeil et des rythmes biologiques, Hôpital du Sacré-Coeur de Montréal, 5400 boulevard, Gouin Ouest, Montréal, Québec, Canada H4J 1C5. E-mail: firstname.lastname@example.org Received Jul 11, 2003, and in revised form Oct 16. Accepted for publication Oct 16, 2003. 372 © 2004 American Neurological Association Published by Wiley-Liss, Inc., through Wiley Subscription Services studies suggest that RLS symptoms intensity may be modulated by a circadian factor. However, several methodological limitations characterize these studies. First, neither of them has evaluated the vigilance of the patients during the constant routine protocol. The study of vigilance is of major interest because increased sleepiness may contribute to the nocturnal worsening of RLS symptoms. A second limitation is that the SIT was administered every 3 or 4 hours, yielding only seven or eight measures over the 24-hour protocol and therefore decreasing the temporal resolution in rhythm parameters. A third limitation is the scarcity of data on biological markers of endogenous circadian rhythmicity. Only core body temperature was measured, and in one study8 temperature data were reported in only four of the eight patients. Finally, neither of the studies included a control group. The use of control group is necessary to determine whether the occurrence of leg discomfort and motor activity at night is specific to patients with RLS or whether it is also present in normal subjects studied in the same experimental conditions. The aim of this study was to evaluate nycthemeral or circadian variations in both leg discomfort and PLM in patients with RLS and in matched healthy controls. These variations then were compared with those of two well-defined markers of endogenous circadian rhythmicity, namely, core body temperature and melatonin secretion rhythms. Vigilance measures also were included to verify whether the fluctuations in vigilance levels were directly associated with similar fluctuations in RLS symptoms severity. A modified constant routine procedure was used to collect the data to control for any masking effect due to variations in activity or posture. Subjects and Methods Population Seven patients (three men and four women; mean age ⫾ standard deviation [SD], 43.9 ⫾ 15.7 years; range, 25– 63 years) with diagnosed primary RLS participated in the study. They fulfilled the four mandatory criteria for RLS, as described previously. Exclusion criteria were the presence of medical conditions known to be associated with RLS such as anemia and renal failure. These conditions were ruled out by clinical evaluation and appropriate laboratory tests. In addition, none of the patients had clinical signs or history of psychiatric or neurological disorders, and none had experienced transmeridian travel or night work within the 6 months preceding the study. Patients were also free of any medication or drug known to affect sleep, sensory functions, or melatonin secretion (␤-blockers or nonsteroidal antiinflammatory drugs) for at least 1 month before entering the study. In four patients, RLS was treated with the dopaminergic agonist pramipexole (0.5–1mg given at bedtime). This treatment was discontinued for 7 to 14 days before the study. The three other patients had never been treated for RLS. Finally, all the patients were nonsmokers and were asked to refrain from using caffeine for the duration of the study. Premenopausal women not taking hormonal contraception were studied during the follicular phase of their menstrual cycle. The control group included seven subjects matched for sex and age (mean age ⫾ SD, 42.4 ⫾ 14.9 years; range, 24 – 60 years). The exclusion criteria were the same as for RLS patients. On a screening nocturnal polysomnographic (PSG) recording, all the patients with RLS and none of the control subjects showed a PLMS index greater than 5. Subjects of both groups were excluded if they had an index of respiratory events (apnea and hypopnea) greater than 5. This research was approved by the university/hospital ethics committee. Each RLS patient and control subject signed a consent form for participating in the study and received financial compensation for their participation. Procedures Before the study, each subject was instructed to maintain a regular (⫾30 minutes) sleep schedule for at least 1 week, which was confirmed by sleep diaries. A French version of the Horne and Ostberg questionnaire10 was also completed by all subjects (except one patient) to characterize “morningness-eveningness” typology. The experimental protocol consisted of an 8-hour PSG recording starting at 23:00 followed by a 28-hour modified constant routine procedure. After the night of PSG recording, subjects were awakened around 07:15. The modified constant routine started at 08:00 and terminated at noon on the next day. Because it was not possible to maintain RLS patients in bed for 28 consecutive hours, the constant routine was divided into 14 episodes of 2 hours. Each 2-hour episode was identical. During the first 20 minutes, subjects had to first evaluate their subjective vigilance, rinse their mouth, and provide a salivary sample (4 minutes), and then they were free to walk around and to go to the bathroom (free-moving period). Then, they were confined to a reclining chair for 40 minutes (resting period) during which they were allowed to watch TV or video movies and were offered a light snack. This resting period was followed by another 20-minute period and then by a 40-minute SIT. For the SIT, subjects were reclined at a 45-degree angle with their legs outstretched. They were instructed to avoid voluntary movements for the duration of the test. They also were instructed to quantify their level of leg discomfort every 5 minutes on a visual analog scale. Based on results of a previous study,6 SIT duration was set at 40 minutes. At all times, ambient lighting was kept below 15 lux at eye level, and a member of the research team was present in the subject room to prevent the subject from falling asleep. Measures SUGGESTED IMMOBILIZATION TESTS. During the SIT, surface electromyograms from right and left anterior tibialis muscles were recorded to score PLM according to the criteria developed recently in our laboratory.11 Based on these criteria were quantified only movements occurring in series of four or more, lasting 0.5 to 10 seconds and occurring at intervals of 4 to 90 seconds. PLMs were scored by an expe- Michaud et al: Circadian Rhythm of RLS Symptoms 373 rienced PSG technician who was blind to group assignation. Moreover, every 5 minutes during the SIT, an auditory signal was given to the subjects, at which time they had to estimate their level of leg discomfort on a 100mm horizontal visual analog scale.6 The descriptors “no discomfort” and “extreme discomfort” were used at the left and right end points of the visual analog scale, respectively. The scoring was done by converting these measures on a numerical scale from 0 to 100. Two parameters were derived from the SIT, namely, (1) the SIT PLM index representing the number of PLM per 40 minutes of immobility, and (2) the SIT mean discomfort score (MDS) representing the averaged value of the 8 measures (one every 5 minutes for 40 minutes) taken during the test. For each subject, data were transformed in percentage of the mean. The 28-hour profiles of the SIT MDS and SIT PLM raw data were calculated for each individual and then averaged for each group. SUBJECTIVE VIGILANCE. Subjective vigilance was assessed by the subjects every hour on a 100mm horizontal visual analog scale. The descriptors “very sleepy” and “very alert” were used at the left and right end points of the visual analog scale, respectively. The measures were converted on a numerical scale from 0 to 100. For each subject, data were transformed in Z-scores. The 28-hour profile of subjective vigilance was calculated for each individual and then averaged for each group. Core body temperature was measured throughout the modified constant routine using a disposable rectal thermistor (Yellowsprings Instruments, Yellow Springs, OH) and was recorded every minute by a MiniLogger monitor (Mini-Mitter, Bend, OR). For each subject, data were transformed in percentage of the mean. The 28hour profile was calculated for each individual and then averaged for each group. CORE BODY TEMPERATURE. Twenty-eight saliva samples (one every hour during the modified constant routine) were collected in all subjects using Salivettes (Sarstedt, Newton, NC). Melatonin concentrations were determined by radioimmunoassay with a 125-iodine–labeled tracer (Bühlmann Laboratories, Basel, Switzerland). With this method, the functional least detectable dose is 0.65pg/ml. Samples were assayed in duplicate and all samples from a given subject were assayed in the same run. The intra-assay coefficients of variation for control samples of 1.49 and 13.30pg/ml were 12.5% and 7.0%, respectively, and the interassay coefficient of variation for a value of 17.0pg/ml was 12.6%. The salivary melatonin data from one control subject were lost because of inadequate freezing. For each subject, melatonin concentrations were transformed as percentage of the mean. The 28-hour profile of salivary melatonin was calculated for each individual and then averaged for each group. The circadian phase of the melatonin secretion was defined as the time of the melatonin secretion onset (DLMO, for dim light melatonin onset).12 DLMO was calculated by interpolation using a threshold of 1.3pg/ml, which represents twice the functional least detectable dose.13 However, in the few individuals (three RLS paSALIVARY MELATONIN. 374 Annals of Neurology Vol 55 No 3 March 2004 tients) with diurnal melatonin levels over 0.65pg/ml, the threshold was calculated by taking twice the average of the samples collected at 14:00, 15:00, and 16:00.14 The melatonin secretion offset (DLMOff) was also calculated by interpolation using the same threshold as for the DLMO. The melatonin synthesis offset (SynOff) was assessed using semilog plots and was defined as the last point after which levels began to decrease, taking into account the point-to-point variability. The total duration of the melatonin episode was calculated as the time from DLMO to DLMOff, whereas the melatonin secretion duration represented the time elapsed from DLMO to SynOff. The subjects’ level of activity was monitored using an actigraph (Ambulatory Monitoring, Ardsley, NY). A piezoelectric accelerometer with a lower limit of sensitivity of 0.1gm translates movements into electrical signals. Data acquisition was performed at a 10Hz sampling rate and the epoch length for analysis was 30 seconds. ACTIGRAPHY MEASURES. Data Analysis Circadian parameters (acrophase and amplitude) for SIT MDS, SIT PLM index, subjective vigilance, core body temperature, and salivary melatonin were estimated with cosinor analyses.15 The acrophase represents the clock time of the maximal value of the curve, whereas the amplitude corresponds to half of the difference between the maximal and minimal value of the curve. Analyses were conducted on both individual and group profiles. The area under the curve of individual raw data also was calculated for each variable. Between-group differences for demographic and chronobiological characteristics were assessed either by Student’s t-tests for independent samples or by 2 tests. Two-way analyses of variance (ANOVAs) with one independent factor (group) and one repeated measure (time) were performed to evaluate between-group differences across the modified constant routine for SIT MDS, SIT PLM index, subjective vigilance, salivary melatonin, core body temperature, and activity level via actigraphy. Simple effect analyses were performed to decompose interaction effects, whereas post hoc comparisons were done to further evaluate mean differences for the main effects using Tukey honest significance difference. The significance level for repeated measures with more than two levels was adjusted with Huynh–Feldt correction, but the original degrees of freedom are reported. Temporal relationships between the profiles of body temperature, salivary melatonin, subjective vigilance, and those of SIT MDS and SIT PLM index were assessed by cross-correlation analyses. Statistical significance was defined as p value less than 0.05. Between-group comparisons were corrected using Bonferroni adjustment for multiple testing (significance set at 0.005). Results Demographic Characteristics The demographic characteristics of both patients with RLS and control subjects are summarized in the Table. No between-group difference was seen for sex, age, “morningness-eveningness” score, or habitual bedtime. Table. Demographic Characteristics of Patients with RLS and Healthy Control Subjects RLS Patients (n ⫽ 7) Control Subjects (n ⫽ 7) pa 43.9 ⫾ 15.7 (25–63) 3/4 59.0 ⫾ 11.6 (43–72) 22:53 ⫾ 01:02 (21:33–23:48) 42.4 ⫾ 14.9 (24–60) 3/4 54.9 ⫾ 8.6 (38–64) 23:39 ⫾ 00:36 (23:06–00:41) NS NS NS NS Characteristic Age (yr), mean ⫾ SD Sex (male/female) Chronotype scoreb, mean ⫾ SD Habitual bedtime (hr),c mean ⫾ SD Student’s t tests (except for sex for which 2 tests were used). Defined according to the Horne and Ostberg questionnaire (score from 16 to 86); the data for one RLS patient are missing. c Assessed with sleep diaries. a b RLS ⫽ restless legs syndrome; NS ⫽ not significant. Leg Discomfort and Periodic Leg Movements Index Profiles Figure 1 illustrates the average ⫾ standard error of the mean (SEM) profiles of the SIT MDS and SIT PLM index for patients and controls. Cosinor analyses performed on group data showed a significant circadian variation in the SIT MDS profiles ( p ⬍ 0.01) in both groups, with an acrophase occurring at 03:26 in the patients and at 04:20 in the controls. No significant between-group difference was found for the acrophase of these rhythms (95% confidence intervals). Taken individually, all patients but one showed a significant circadian variation in their SIT MDS profiles. In comparison, only three control subjects of seven showed a significant circadian rhythm. A two-way ANOVA was performed to evaluate the between-group difference for SIT MDS profiles and found a significant Group by Time interaction (F[13,156] ⫽ 2.78; p ⫽ 0.0013). Simple effect analyses showed that the patients’ SIT MDS scores were higher than those of the controls for the entire duration of the protocol, except for the last two SIT. For the SIT PLM index profile, cosinor analyses performed on group data showed a significant circadian pattern ( p ⬍ 0.005) in both groups. In patients, the acrophase occurred at 03:05 compared with 06:29 for controls. This rather large between-group difference in the acrophase did not reach statistical significance (95% confidence intervals). Taken individually, only two patients and one control showed a significant circadian variation of their SIT PLM index. However, note here that in three patients, no PLM were quantified during any of the 14 SITs of the protocol. The repeated-measures ANOVA did not show an interaction effect but rather a main Time effect (F[13,156] ⫽ 2.43; p ⬍ 0.005). Core Body Temperature, Salivary Melatonin, and Subjective Vigilance Profiles Figure 2 shows the average ⫾ SEM profiles of core body temperature, salivary melatonin, and subjective vigilance across the 28-hour modified constant routine for each group. When assessed individually, all profiles of salivary melatonin and core body temperature for patients and control subjects showed a significant circadian rhythm. The average acrophase of the core body temperature and salivary melatonin curves occurred in patients, respectively, at 16:28 ⫾ 1:24 and 01:29 ⫾ 1:19, and in controls it occurred at 16:24 ⫾ 1:45 and 02:30 ⫾ 0:54. No significant between-group difference was seen for either the amplitude, the acrophase, or the area under the curve of either salivary melatonin or core body temperature rhythms. However, patients showed a significantly shorter melatonin secretion duration than controls (6.0 ⫾ 1.9 hours vs 8.9 ⫾ 0.8 Fig 1. The raw data average ⫾ SEM profiles of the Suggested Immobilization Test (SIT) mean leg discomfort score (MDS) and SIT periodic leg movements (PLM) index across the 28hour modified constant routine for patients with restless legs syndrome and healthy control subjects. Michaud et al: Circadian Rhythm of RLS Symptoms 375 Fig 2. The average ⫾ SEM profiles of core body temperature, salivary melatonin, and subjective vigilance across the 28-hour modified constant routine for patients with restless legs syndrome and healthy control subjects. hours; p ⫽ 0.004) and trends for an earlier melatonin SynOff (02:52 ⫾ 2:40 vs 05:30 ⫾ 2:29; p ⫽ 0.031) and shorter total duration of melatonin episode (8.2 ⫾ 2.0 hours vs 10.1 ⫾ 0.6 hours; p ⫽ 0.045). The twoway ANOVAs performed on each salivary melatonin and core body temperature data both showed a main effect for Time (F[27,297] ⫽ 59.67 and F[28,308] ⫽ 21.44, respectively; p ⬍ 0.0001). No interaction effect or main effect for Group was found. For subjective vigilance, cosinor analyses performed on group data showed a significant circadian variation for each group. In patients, the acrophase of the sub- 376 Annals of Neurology Vol 55 No 3 March 2004 jective vigilance occurred at 16:43 compared with 15:29 for controls. No significant between-group difference was found for the acrophase (95% confidence intervals). Individually, a significant circadian variation was observed in all the controls but in only five of the seven patients. The two-way ANOVA showed a main effect for Time (F[13,156] ⫽ 30.93; p ⬍ 0.0001) and Group (F[1,12] ⫽ 4.73; p ⫽ 0.05). Thus, a higher visual analog scale score was observed in the patients, suggesting that they were less sleepy than the controls. Finally, there was no Group effect Fig 3. Relationship between the profile of Suggested Immobilization Test (SIT) mean leg discomfort score (MDS) (squares: average ⫾ SEM) and those of core body temperature, salivary melatonin, and subjective vigilance for patients with restless legs syndrome and healthy control subjects. Cross-correlation function (CCF) and lag value are reported for each correlation. or Group by Time interaction for the activity level as measured by actigraphy. Cross-correlation and Correlation between Variables Figure 3 illustrates for both patients and controls the average ⫾ SEM profile of the SIT MDS in relation to salivary melatonin, subjective vigilance, and core body temperature profiles. In patients as well as in controls, the SIT MDS group profiles were strongly correlated with salivary melatonin, subjective vigilance, and core body temperature profiles (all p values ⬍0.003; see Fig 3 for cross-correlation function and lag value). For the SIT PLM index profile (see Fig 4 for cross-correlation function and lag value), significant cross-correlations (all p values ⬍0.003) were also found with salivary melatonin, subjective vigilance, and core body temperature profiles both for patients and controls. As to the phase relationship between patients’ variables, we found that the acrophase of melatonin profile preceded the SIT MDS and SIT PLM acrophase by approximately 2 hours. On the other hand, the decrease in core body temperature occurred at the same time as the increase in SIT MDS, whereas the acrophase of subjective vigilance and core body temperature was delayed by approximately 2 hours compared with that of the SIT PLM. Michaud et al: Circadian Rhythm of RLS Symptoms 377 Fig 4. Relationship between the profile of Suggested Immobilization Test (SIT) periodic leg movements (PLMs) index (squares: average ⫾ SEM) and those of core body temperature, salivary melatonin, and subjective vigilance for patients with RLS and healthy control subjects. Cross-correlation function (CCF) and lag value are reported for each correlation. Discussion Circadian Rhythms of Leg Discomfort and Periodic Leg Movement The results of this study clearly show the presence of a circadian rhythm of leg discomfort and PLM in the group of RLS patients. Both sensory and motor symptoms of RLS showed an acrophase at approximately 03: 00. Group data for control subjects also showed a circadian variation in the level of leg discomfort and the number of PLM, although these manifestations were not as severe as those seen in patients. In control subjects, the acrophases of sensory and motor rhythms 378 Annals of Neurology Vol 55 No 3 March 2004 were seen at 04:00 and at 06:00, respectively. Overall, these results are in agreement with those of previous studies of patients with RLS showing a circadian variation of leg discomfort and of PLM, with acrophases occurring between midnight and 04:00.8,9 Moreover, the results obtained on individual data show that the measure of leg discomfort showed a more consistent circadian variation in patients than in controls. Indeed, six of the seven patients showed a significant circadian variation in the SIT MDS compared with only one in seven controls. For the SIT PLM index, a significant circadian variation was found in only one control and two patients. Because three patients did not show PLM during any of the SIT, half of the patients experiencing PLM (two of four) showed a significant circadian variation of their PLM. The complete absence of PLM during the SIT in three patients parallels the findings of a recent study conducted on a large sample of patients with RLS, showing that a significant number of patients do not experience PLM during the SIT.16 One possibility may be that they are able to restrain themselves from moving (according to the test instructions), although experiencing high levels of leg discomfort. This finding and the absence of significant betweengroup difference for the SIT PLM index across the modified constant routine support the hypothesis that the SIT MDS may be a better marker of RLS symptoms than the SIT PLM index.16 Relationships between Restless Legs Syndrome Symptoms and Biological Variables THE ROLE OF VIGILANCE IN THE WORSENING OF RESTLESS LEGS SYNDROME SENSORY AND MOTOR SYMPTOMS. In both patients and control subjects, the level of subjective vigilance across the 28-hour modified constant routine was negatively correlated with leg discomfort and PLM. In other words, there was a progressive increase in RLS sensory and motor symptoms during the night when subjective vigilance decreased. The association with subjective vigilance is especially striking for SIT MDS in control subjects, for whom a high level of correlation was found at lag 0. However, in RLS patients, the circadian rhythm of subjective vigilance was delayed compared with both SIT MDS and SIT PLM index profiles. This suggests that the decrease in vigilance is unlikely the only cause for the worsening of symptoms at night, at least in RLS patients. The role of vigilance as a contributing factor for the increase in RLS symptoms severity needs to be further clarified using, for example, quantitative electroencephalogram measures. THE ROLE OF CORE BODY TEMPERATURE IN THE WORSENING OF RESTLESS LEGS SYNDROME SENSORY AND MOTOR SYMPTOMS. Strong negative correlations were found between core body temperature and both leg discomfort and PLM in patients and in controls. In patients, the leg discomfort profile was almost the mirror image of the core body temperature profile (lag of 0 hour). Thus, the peak in RLS symptoms severity occurred almost in conjunction with the nadir of the core body temperature. This result is slightly different from those of previous studies showing the occurrence of symptoms peak intensity during the falling limb of the patients’ core body temperature.8,9 This small discrepancy between studies is most likely attributable to the difference in the temporal resolution of symptoms as- sessment (2 hours in this study compared with 3 or 4 hours in the two aforementioned studies). THE ROLE OF MELATONIN IN THE WORSENING OF RESTLESS LEGS SYNDROME SENSORY AND MOTOR SYMPTOMS. This study measured for the first time to our knowledge the circadian rhythm of melatonin secretion in patients with RLS. Cross-correlation analyses showed a strong relationship between symptoms of RLS and salivary melatonin level, the changes in melatonin concentration preceding (by a lag of 2 hours) the increase in leg discomfort and PLM in the patients group. The results showed that, although the symptoms start to worsen at the time of onset of melatonin secretion, the acrophase of RLS symptoms is reached approximately 2 hours after the peak of melatonin secretion. This is congruent with the hypothesis that melatonin secretion may be driving the increase of RLS symptoms in the evening and during the night. The association noted between RLS symptoms and melatonin secretion raises the possibility that melatonin would play a direct role in the pathophysiology of RLS. Actually, little is known about the physiological mechanisms underlying RLS. There are several evidences that central dopaminergic (DA) systems may be involved, coming both from imaging17–19 and pharmacological studies.20 –24 In the last decade, it has been shown that physiological concentrations of melatonin exert an inhibitory effect on DA secretion in several areas of the mammalian central nervous system.25 This inhibition of DA release by melatonin appears to be mediated by membranal, low-affinity melatonin binding sites via the suppression of calcium influx into the stimulated nerve endings. Besides its presynaptic effect, melatonin is also known to suppress postsynaptic N-methyl-D-aspartate receptor–mediated excitatory responses of striatal neurons to glutamate. On the basis of these latter findings, it is reasonable to think that the increase in melatonin secretion in the evening may facilitate the occurrence of RLS symptoms by decreasing the activity of the central DA systems. Although the results of this study strongly support the hypothesis of an intrinsic circadian rhythm for RLS symptoms, further studies will be necessary to confirm this hypothesis. For example, the role of melatonin in the pathophysiology of RLS could be clarified by the administration of exogenous melatonin in patients with and without RLS or by melatonin suppression with light. In addition, the phase-shifting properties of bright-light exposure could be used to change the endogenous circadian phase within a fixed sleep-wake schedule to verify whether a delay in the phase of the melatonin rhythm would be accompanied by a delay in the increase of RLS symptoms. Michaud et al: Circadian Rhythm of RLS Symptoms 379 This research program has been supported by the Canadian Institutes of Health Research (J.M., M.D.; and studentship to M.M.). M.M. is currently supported by a studentship from the faculty of medicine (COPSE) of the University of Montreal. We thank B. Adam and M. Ruffiange, for their technical help, and D. Petit, for reviewing the manuscript. References 1. Lavigne GJ, Montplaisir JY. Restless legs syndrome and sleep bruxism: prevalence and association among canadians. Sleep 1994;17:739 –743. 2. Rothdach AJ, Trenkwalder C, Haberstock J et al. Prevalence and risk factors of RLS in an elderly population: the MEMO study. Memory and Morbidity in Augsburg Elderly. Neurology 2000;54:1064 –1068. 3. Phillips B, Young T, Finn L, et al. Epidemiology of restless legs symptoms in adults. Arch Intern Med 2000;160:2137–2141. 4. Allen RP, Picchietti D, Hening WA, et al. Restless legs syndrome: diagnostic criteria, special considerations, and epidemiology. A report from the restless legs syndrome diagnosis and epidemiology workshop at the National Institutes of Health. Sleep Med 2003;4:101–119. 5. Ancoli-Israel S, Cole R, Alessi C, et al. The role of actigraphy in the study of sleep and circadian rhythms. Sleep 2003;26: 342–392. 6. Michaud M, Lavigne G, Desautels A, et al. Effects of immobility on sensory and motor symptoms of restless legs syndrome. Mov Disord 2002;17:112–115. 7. Mills JN, Minors DS, Waterhouse JM. Adaptation to abrupt time shifts of the oscillator(s) controlling human circadian rhythms. J Physiol 1978;285:455– 470. 8. Trenkwalder C, Hening W, Walters AS, et al. Circadian rhythm of periodic limb movements and sensory symptoms of restless legs syndrome. Mov Disord 1999;14:102–110. 9. Hening WA, Walters AS, Wagner M, et al. Circadian rhythm of motor restlessness and sensory symptoms in the idiopathic restless legs syndrome. Sleep 1999;22:901–912. 10. Horne JA, Ostberg O. A self-assessment questionnaire to determine morningness-eveningness in human circadian rhythms. Int J Chronobiol 1976;4:97–110. 380 Annals of Neurology Vol 55 No 3 March 2004 11. Michaud M, Poirier G, Lavigne G, et al. Restless legs syndrome: scoring criteria for leg movements recorded during the Suggested Immobilization Test. Sleep Med 2001;2: 317–321. 12. Lewy AJ, Cutler NL, Sack RL. The endogenous melatonin profile as a marker for circadian phase position. J Biol Rhythms 1999;14:227–236. 13. Deacon SJ, Arendt J. Phase-shifts in melatonin, 6-sulphatoxymelatonin and alertness rhythms after treatment with moderately bright light at night. Clin Endocrinol 1994; 40:413– 420. 14. Voultsios A, Kennaway DJ, Dawson D. Salivary melatonin as a circadian phase marker: validation and comparison to plasma melatonin. J Biol Rhythms 1997;12:457– 466. 15. Nelson W, Tong YL, Lee JK, et al. Methods for cosinorrhythmometry. Chronobiologia 1979;6:305–323. 16. Michaud M, Paquet J, Lavigne G, et al. Sleep laboratory diagnosis of restless legs syndrome. Eur Neurol 2002;48:108 –113. 17. Staedt J, Stoppe G, Kogler A, et al. Dopamine D2 receptor alteration in patients with periodic movements in sleep (nocturnal myoclonus). J Neural Transm 1993;93:71–74. 18. Turjanski N, Lees AJ, Brooks DJ. Striatal dopaminergic function in restless legs syndrome. Neurology 1999;52:932–937. 19. Michaud M, Soucy JP, Chabli A, et al. SPECT imaging of striatal pre- and postsynaptic dopaminergic status in restless legs syndrome with periodic leg movements in sleep. J Neurol 2002; 249:164 –170. 20. Akpinar S. Treatment of restless legs syndrome with levodopa plus benserazide. Arch Neurol 1982;39:739. 21. Brodeur C, Montplaisir J, Godbout R, et al. Treatment of restless legs syndrome and periodic movements during sleep with L-dopa: a double-blind, controlled study. Neurology 1988;38: 1845–1848. 22. Earley CJ, Yaffee JB, Allen RP. Randomized, double-blind, placebo-controlled trial of pergolide in restless legs syndrome. Neurology 1998;51:1599 –1602. 23. Wetter TC, Stiasny K, Winkelmann J, et al. A randomized controlled study of pergolide in patients with restless legs syndrome. Neurology 1999;52:944 –950. 24. Montplaisir J, Nicolas A, Denesle R, et al. Restless legs syndrome improved by pramipexole: a double-blind randomized trial. Neurology 1999;52:938 –943. 25. Zisapel N. Melatonin-dopamine interactions: from basic neurochemistry to a clinical setting. Cell Mol Neurobiol 2001;21: 605– 616.