Motor disturbances during non-REM and REM sleep in narcolepsy–cataplexy: a video-polysomnographic analysis

Authors


Birgit Högl, MD, Department of Neurology, Innsbruck Medical University, Anichstrasse 35, A-6020 Innsbruck, Austria. Tel.: +43 512 504 23811; fax: +43 512 504 23842; e-mail: birgit.ho@i-med.ac.at

Summary

Motor events during sleep can be frequently observed in patients with narcolepsy–cataplexy. We hypothesized that increased motor events and related arousals contribute to sleep fragmentation in this disease. We aimed to perform a detailed whole-night video-polysomnographic analysis of all motor events during non-rapid eye movement and rapid eye movement sleep in a group of narcolepsy–cataplexy patients and matched controls, and to assess the association with arousals. Video-polysomnographic registrations of six narcolepsy–cataplexy patients and six sex- and age-matched controls were analysed. Each motor event in the video was classified according to topography, number of involved body parts, duration and its association with arousals. The mean motor activity index was 59.9 ± 23.0 h−1 in patients with narcolepsy–cataplexy compared with 15.4 ± 9.2 h−1 in controls (P = 0.004). Distribution of motor events was similar in non-rapid eye movement and rapid eye movement sleep in the patient group (P = 0.219). In narcolepsy–cataplexy, motor events involved significantly more body parts (≥ 2 body regions: 38.2 ± 15.6 versus 14.9 ± 10.0; P = 0.011). In addition, the proportion of motor events lasting longer than 1 s was higher in patients than controls (88% versus 44.4%; P < 0.001). Both total and motor activity-related arousal indices were increased in narcolepsy–cataplexy (total arousal index: 21.6 ± 9.0 versus 8.7 ± 3.5; P = 0.004; motor activity-related arousal index: 17.6 ± 9.8 versus 5.9 ± 2.3; P = 0.002). Motor activity and motor activity-related arousal indices are increased in both non-rapid eye movement and rapid eye movement sleep in narcolepsy–cataplexy compared with controls. This supports the concept of a general sleep motor dysregulation in narcolepsy–cataplexy, which potentially contributes to or even underlies sleep fragmentation in this disease.

Introduction

Narcolepsy–cataplexy is characterized by excessive daytime sleepiness, cataplexy, sleep paralysis, hypnagogic hallucinations and disturbed nocturnal sleep (Guilleminault and Fromherz, 2005). The International Classification of Sleep Disorders-2 (ICSD-2) highlights nocturnal sleep fragmentation as one of the polysomnographic (PSG) hallmarks of the disease (International Classification of Sleep Disorders, 2005). Abnormal motor activity in rapid eye movement (REM) sleep has been described early in narcolepsy. The frequency of REM sleep behaviour disorder has been found to be increased (Ferri et al., 2008; Nightingale et al., 2005; Santamaria et al., 2008; Schenck and Mahowald, 1992), and REM sleep behaviour disorder itself has been reported to be at least in some patients the presenting symptom of the disease (Bonakis et al., 2008; Nevsimalova et al., 2007). In contrast to REM sleep-related motor disturbances, only a few studies took a closer look at whether non-REM parasomnias are also increased in the disease (Knudsen et al., 2010; Mayer and Meier-Ewert, 1993). Mayer et al. examined 27 patients with narcolepsy. They found that patients with narcolepsy with REM sleep behaviour disorder had more frequently non-REM parasomnias than patients with narcolepsy without REM sleep behaviour disorder. The authors concluded that narcolepsy is associated with a disturbed motor control in both non-REM and REM sleep (Mayer and Meier-Ewert, 1993). In line with this study, Knudsen et al. (2010) only recently reported that hypocretin deficiency is independently associated with increased short and long electromyographic (EMG) muscle activity in both non-REM and REM sleep.

Based on our own observation in sleep recordings in narcolepsy–cataplexy, we noticed that motor events are also increased during non-REM sleep and contribute to sleep fragmentation itself. Therefore, we performed a systematic video-PSG study with analysis of all motor events during both non-REM and REM sleep in narcolepsy–cataplexy compared with a sex- and age-matched control group, and assessed potential associations with arousals.

Materials and methods

Patients

Six consecutive patients with narcolepsy–cataplexy according to ICSD-2 criteria (International Classification of Sleep Disorders, 2005) were included in this study between October 2002 and October 2003. All patients were referred for routine PSG for at least two consecutive nights, and multiple sleep latency testing because of clinically suspected narcolepsy. Patients in whom a clinically relevant sleep-related breathing disorder (apnoea–hypopnoea index > 10 h−1) was confirmed by PSG were excluded from further analysis. Mean cataplexy severity score, measured on an eight-point rating scale, ranging from 0 indicating no cataplexy to eight indicating very severe and very frequent cataplexy, was 4.8 ± 1.0 (3–6). No patient had manifest non-REM or REM parasomnia, or clinically relevant restless legs syndrome. No patient was on wake-enhancing medication at the time of the investigation. One patient (patient 4) was treated for cataplexy with 40 mg paroxetine. All patients gave written informed consent. A sex- and age-matched control group was established based on subjects who were examined for suspected sleep-related breathing disorders. Subjects were eligible for inclusion in the control group if a sleep-related breathing disorder was not confirmed by PSG (a diagnosis of primary snoring was not considered an exclusion criterion), and they had no further history or PSG finding of any clinically relevant sleep disorder (e.g. restless legs syndrome, periodic limb movement disorder, insomnia, etc.). Periodic leg movements in sleep without restless legs syndrome or outside the context of periodic leg movement disorder were not considered an exclusion criterion. Epworth sleepiness score was below 10 in all control subjects. None of the controls was on CNS active medication at the time of investigation.

Video-PSG procedures

Polysomnographies for this study were recorded with a digital polygraph (Brainlab, Schwarzer, Munich, software version 3.0/4.0, Germany), and included electroencephalography (EEG; C3, C4, O1, O2, A1 and A2 electrodes), electrooculography (vertical and horizontal eye movements), EMG (mental, submental and both tibialis anterior muscles) and cardiorespiratory recording [single-channel electrocardiography, nasal airflow (thermocouple), tracheal microphone, thoracic and abdominal respiratory movements (piezo), transcutaneous oxygen saturation]. Sleep stage scoring was performed visually in 30-s epochs according to Rechtschaffen and Kales (1968), as PSGs of patients for this study were recorded before the introduction of the new AASM 2007 criteria (Iber et al., 2007). REM sleep was scored according to standard criteria, with allowance to score REM sleep despite persistence of tonic or phasic muscle activity in the case of REM sleep without atonia (Mahowald and Schenck, 2005). Arousals were defined according to ASDA 1992 criteria (Bonnet et al., 1992). The video was recorded by an infrared camera (Elbex, EX series, Regensburg, Germany) and digitally stored (data rate: 1 500 000 bits s−1). The screen resolution for video analysis was 1280 × 960 pixels. The whole-night videography was carefully looked through for movements. This was done in a systematic way in several runs (first run: focusing on movements of the head, neck, trunk; second run: focusing on movements of the upper extremities; third run: focusing on movements of the lower extremities; fourth run: if needed to clarify arbitrary decisions) by trained raters (L.E., I.T., S.V.S.). Doubtful cases were supervised by board-certified neurologists with specialization in sleep medicine (B.F., B.H.). All patients and controls underwent two consecutive nights of PSG. The first night was used for adaptation; the second night was used for analysis.

Classification of motor events

All visible movements were described as motor events. Of note, no attempt was made to exclude periodic leg movements from counting, because due to the definition criteria with an intermovement interval from 5 to 90 s, all motor events occurring within this range are considered periodic leg movements irrespective of whether they are classical periodic leg movements or not. Topographic appearance of motor events was classified into ‘involved body parts’ (head, neck, trunk, upper extremity, lower extremity). ‘Laterality’ was described as unilateral or bilateral. In addition, the number of involved body parts (head, neck, trunk, upper extremity, lower extremity; maximum: 5) was provided for every motor event. The ‘duration’ of each motor event was determined in seconds with a hand-held chronograph. Voluntary movements following prolonged arousals or awakenings were not included in the analysis.

Derived measures

All identified motor events were tabulated in an excel spreadsheet, and associated to sleep stage and arousal. The motor activity index was defined as total number of motor events observed in the video per hour of sleep stage. It does not refer to EMG-related muscle activity in the chin or tibialis anterior channels. The exact onset and offset of motor events was observed in the video and linked to the corresponding mini-epochs. Motor events shorter than 1 s were presented together because a reliable measurement of duration in the video was not possible for a time span shorter than 1 s. Total arousal and motor activity-related arousal indices represent total arousal and motor activity-related arousal counts per hour. An arousal was scored as related when the arousal was either preceding or following the movement by ≤ 0.5 s. The time interval of ≤ 0.5 s was chosen according to scoring rules for periodic leg movement in sleep-related arousals (Zucconi et al., 2006). In case of any motor event, the PSG registration was re-checked for accompanying movement artefacts in the EEG channels and the presence of EMG activity in the chin or tibialis anterior muscles irrespective of amplitude and duration.

Statistics

SPSS 16.0 for Macintosh (SPSS, Chicago, IL, USA) was used for all statistical analyses. Descriptive statistics are calculated for patients and controls, and given as means ± standard deviations. Data were tested for normal distribution using the Shapiro Wilk test. All statistical tests were calculated for groups with n = 6 using aggregated percentages or mean values per individual patient. In case of normal distribution, simple t-tests were calculated. In case of non-normal distribution, non-parametric statistics were applied (Mann–Whitney U-test for two unrelated variables, Wilcoxon test for two related variables, Friedman test for multiple related variables). For categorical variables, the Chi square test was performed. Correlations were calculated using Spearman’s correlation coefficient. P-values below 0.05 were considered significant. Due to the exploratory character of the study, no corrections for multiple comparisons were applied.

Results

Demographic and clinical characteristics

Night two of six right-handed patients with narcolepsy–cataplexy (five men and one woman) was analysed. The mean age of the patient group was 29.2 ± 13.5 years (range: 13–53 years). The mean duration of excessive daytime sleepiness was 9.9 ± 15.3 years (range: 0.75–40 years). All patients were HLA DRB 1*1501 and HLA DQB1*0602 positive. Individual data and additional information are given in Table 1.

Table 1.   Characterization of the NC patient group
 Pat 1Pat 2Pat 3Pat 4Pat 5Pat 6
  1. MSLT, multiple sleep latency testing; PSG, polysomnography; REM, rapid eye movement.

Demographics
 Gendermmmmmf
 Age at investigation, years162435341353
 Disease duration, years0.75123.52140
Clinical features
 Epworth sleepiness score171719201222
 Cataplexy severity score535655
 Hypnagogic hallucinations++
 Sleep paralysis++
 Disturbed nocturnal sleep++
PSG findings
 Sleep latency in MSLT, min320, 533, 51, 5
 Sleep-onset REM episodes in MSLT, n4/54/54/54/54/55/5
 REM sleep without atonia++++
Additional sleep diagnoses
 Restless legs syndromeSporadicMildMild
 Periodic leg movements in sleep > 15 h−1++++
 Non-REM parasomnia
 REM sleep behaviour disorder
 Sleep-related breathing disorderMild

The control group consisted of six age- and sex-matched subjects [five men, one woman; mean age, 28.3 ± 11.4 years (range: 15–49 years)].

PSG results

Patients with narcolepsy–cataplexy tended to have a shorter sleep latency in the second night of PSG recording than controls (9.8 ± 8.3 versus 26.3 ± 18.3 min; P = 0.092). Sleep onset REM periods were present in nocturnal sleep in three out of six narcolepsy–cataplexy patients, but not in any control (P = 0.045). In addition, narcolepsy–cataplexy patients had fragmented nocturnal sleep, with an increased percentage of stage 1 sleep (narcolepsy–cataplexy patients versus controls, 20.0 ± 13.2 versus 9.1 ± 5.7 min; P = 0.071), a decreased percentage of S2 sleep (narcolepsy–cataplexy patients versus controls, 35.5 ± 10.7 versus 51.5 ± 8.5 min; P = 0.022) and an increased total arousal index (narcolepsy–cataplexy patients versus controls, 21.6 ± 9.0 versus 8.7 ± 3.5; P = 0.004). There was no difference in sleep efficiency between both groups (narcolepsy–cataplexy patients versus controls, 90.8 ± 4.4 versus 92.0 ± 6.6%; P = 0.725). For further sleep variables, see Table 2.

Table 2.   Night sleep and respiratory parameters of patients and controls
 NC patients (mean ± SD)Controls (mean ± SD)P-value
  1. REM, rapid eye movement; SoREMs, sleep-onset REMs; WASO, wake after sleep onset.

Time in bed (min)466.6 ± 24.5484.3 ± 18.20.185
Total sleep time (min)415.9 ± 41.8425.3 ± 38.10.692
Sleep period (SPT; min)457.2 ± 28.8458.4 ± 20.40.818
Sleep efficiency (% SPT)90.8 ± 4.492.0 ± 6.60.725
WASO (min)41.3 ± 17.633.1 ± 26.70.545
S1 (% SPT)20.0 ± 13.29.1 ± 5.70.092
S2 (% SPT)35.5 ± 10.751.5 ± 8.50.016
S3/S4 (% SPT)9.8 ± 9.914.0 ± 8.10.441
Stage REM (% SPT)20.0 ± 6.617.6 ± 3.60.446
Sleep latency (S2), (min)9.8 ± 8.326.3 ± 18.30.071
REM latency (min)42.8 ± 48.5109.7 ± 36.00.022
SoREMs (n patients)3/60/60.045
Apnoea–hypopnoea index3.5 ± 4.01.7 ± 1.90.589
Oxygen desaturation index2.2 ± 3.51.0 ± 1.40.818

Descriptive analysis of motor events

The mean motor activity index in patients with narcolepsy–cataplexy was 59.9 ± 23.0 h−1 of sleep compared with 15.4 ± 9.2 h−1 of sleep in controls (P = 0.004). Sleep stage-specific motor event indices were consistently higher in patients with narcolepsy–cataplexy than in controls, with highest indices during S1 sleep followed by REM sleep, S2 sleep and S3/S4 sleep (for details, see Table 3). This is illustrated in Fig. 1.

Table 3.   Motor activity indices and their distribution across sleep stages
 Pat 1Pat 2Pat 3Pat 4Pat 5Pat 6NC patients, mean ± SDControls mean ± SDP
  1. REM, rapid eye movement.

Total motor activity index27.142.072.880.384.552.959.9 ± 23.015.4 ± 9.20.004
Motor activity index
 Non-REM sleep18.637.264.879.286.454.656.8 ± 25.711.8 ± 7.70.009
  S159.0328.3108.184.6129.281.0131.7 ± 99.338.1 ± 20.10.004
  S212.627.519.175.375.850.743.5 ± 28.08.3 ± 4.40.004
  S3/S42.422.0N.A.6084.54.934.7 ± 36.14.1 ± 3.10.082
 REM sleep6568.787.682.873.744.870.4 ± 15.229.9 ± 17.00.001
Figure 1.

 Distribution of motor activity indices across sleep stages. Bars represent mean values, whiskers represent the standard error of the mean. REM, rapid eye movement.

In patients with narcolepsy–cataplexy motor activity indices were similar during REM sleep and non-REM sleep (P = 0.219), whereas in controls motor activity indices were higher in REM sleep compared with non-REM sleep (P = 0.028). There was no significant correlation between motor activity index and cataplexy severity score (Spearman’s rho = 0.507; P = 0.305).

Topographic distribution

In narcolepsy–cataplexy patients, motor events involved most frequently the extremities (lower extremities: 32.4 ± 24.8 h−1; upper extremities: 25.7 ± 12.0 h−1), followed by the neck/face (neck: 14.9 ± 10.7 h−1; face: 12.9 ± 11.7 h−1) and the trunk (3.5 ± 2.5 h−1; P = 0.01). The same distribution of involved body parts was found for the control group (lower extremities: 6.3 ± 4.8 h−1; upper extremities: 7.8 ± 6.9 h−1; neck: 4.4 ± 8.0 h−1; face: 1.9 ± 1.9 h−1; trunk: 0.4 ± 0.5 h−1; P = 0.015). Vocalizations were very rare in narcolepsy (1.1 ± 1.6 h−1), and absent in the control group (P = 0.002). Abnormal behavioural manifestations such as scenic/complex behaviours and violent motor events were not observed. In both patients and controls, most motor events involved only one body region. The percentage of motor events involving only one body region was more frequent in the control than the patient group (86.5 ± 7.0 versus 65.6 ± 12.8; P = 0.006), whereas the percentage of motor events involving more than one body region was more frequent in patients than in controls (≥ two body regions: 38.2 ± 15.6 versus 14.9 ± 10.0; P = 0.011; Table 4). In narcolepsy–cataplexy, 22.5 ± 11.9 motor events per hour were unilateral, which were evenly distributed between the left (10.7 ± 8.3 h−1) and right body side (12.0 ± 5.4 h−1; P = 0.463). Individual information is given in Table 4.

Table 4.   Topographical distribution of motor events
 Pat 1Pat 2Pat 3Pat 4Pat 5Pat 6N patients, mean ± SDControls mean ± SDP
Motor activity index27.142.072.880.384.552.959.9 ± 23.015.4 ± 9.20.004
 Face3.213.714.334.69.12.812.9 ± 11.71.9 ± 1.90.009
 Neck11.56.217.135.37.211.914.9 ± 10.74.4 ± 8.00.041
 Trunk1.62.03.83.81.78.13.5 ± 2.50.4 ± 0.50.002
 Upper extremity16.815.244.936.220.620.725.7 ± 12.07.8 ± 6.90.013
 Lower extremity5.516.517.948.672.633.432.4 ± 24.86.3 ± 4.80.015
 Vocalization0.10.11.44.20.60.21.1 ± 1.600.002
 Unilateral14.711.944.124.914.724.822.5 ± 11.99.2 ± 7.40.048
Simultaneous involvement of body regions
 1 body region (%)6670.572.641.077.266.365.6 ± 12.886.5 ± 7.00.006
 2 body regions (%)24.117.620.628.413.824.121.4 ± 5.211.2 ± 6.50.013
 3 body regions (%)7.97.25.122.26.17.79.4 ± 6.42.3 ± 2.30.009
 4 body regions (%)2.04.41.68.22.91.73.5 ± 2.500.002
 5 body regions (%)00.30.20.200.30.2 ± 0.100.065

In narcolepsy, unilateral movements were more frequent during REM than during non-REM sleep (52.5 ± 25.8% versus 36.1 ± 15.4%; P = 0.028). The type and number of involved body parts and motor events did not differ between non-REM and REM sleep (P > 0.05).

Duration

Patients with narcolepsy–cataplexy had less short motor events < 1 s (22.0% versus 55.6%; P < 0.001) and more long motor events > 1 s (78.0% versus 44.4%; P < 0.001) compared with the control group. The mean duration of long motor events > 1 s was 8.2 ± 11.4 s in the narcolepsy–cataplexy group versus 4.4 ± 4.6 s in the control group (P < 0.001).

Arousal and motor activity-related arousal indices

Arousals were more frequent in patients with narcolepsy–cataplexy than in controls (mean total arousal index, 21.6 ± 9.0 h−1 versus 8.7 ± 3.5 h−1 of sleep; P = 0.004). In addition, arousals related to motor events were more frequent (motor activity-related arousal index, 17.6 ± 9.8 h−1 versus 5.9 ± 2.3 h−1 of sleep; P = 0.002). For individual arousal and motor activity-related arousal indices given per sleep stage, see Table 5.

Table 5.   Arousal indices and their relation with motor activity
 Pat 1Pat 2Pat 3Pat 4Pat 5Pat 6Patients, mean ± SDControls, mean ± SDP
  1. REM, rapid eye movement.

Total arousal index18.021.918.339.213.918.121.6 ± 9.08.7 ± 3.50.004
 Non-REM sleep17.327.021.840.814.516.423.0 ± 9.88.3 ± 4.30.013
  S152.1246.235.352.451.937.679.3 ± 82.228.1 ± 13.80.015
  S212.49.37.017.911.47.210.0 ± 4.15.8 ± 3.10.036
  S3/S44.010.4N.A.N.A.2.62.54.8 ± 3.82.8 ± 1.90.365
 REM sleep22.129.713.235.06.623.821.8 ± 10.510.6 ± 3.80.047
Motor activity related arousal index13.914.814.837.310.614.217.6 ± 9.85.9 ± 2.20.002
 Non-REM sleep12.717.017.340.211.713.218.7 ± 10.85.0 ± 2.70.026
  S146.2161.328.951.542.729.860.1 ± 50.417.5 ± 5.90.002
  S26.45.14.517.98.16.18.0 ± 5.03.3 ± 1.60.070
  S3/S44.07.0N.A.N.A.1.02.53.6 ± 2.51.7 ± 2.10.254
 REM sleep20.026.312.431.35.020.019.2 ± 9.57.3 ± 4.00.027

Movement artefacts in the EEG channels/EMG activity in chin or tibialis anterior muscles

In both groups, the majority of motor events (patients: 92.9 ± 12.2; controls: 87.7 ± 13.7) were accompanied by either movement artefacts in the EEG channels or EMG activity in chin or tibialis anterior muscles. Detailed results and subanalyses are given in Table 6.

Table 6.   Motor events and their relation to EEG movement artefacts and EMG activity
 Pat 1Pat 2Pat 3Pat 4Pat 5Pat 6Patients, mean ± SDControls, mean ± SDP
  1. Percentages of motor events accompanied by EEG artefacts and/or EMG activity in either chin or tibialis anterior muscles are provided. Note that both in patients and controls most motor events were accompanied by either EEG artefacts or EMG activity in the EMG channels.

  2. EEG, electroencephalography; EMG, electromyography.

EEG artefact (%)92.183.748.489.857.178.975.0 ± 18.140.8 ± 15.70.006
EMG activity %
 Chin, (%)92.678.147.599.150.386.075.6 ± 21.867.7 ± 24.80.573
 Tibialis anterior (%)73.460.854.179.491.268.971.3 ± 13.357.2 ± 8.40.053
EEG/EMG (%)98.596.268.199.897.896.792.9 ± 12.287.7 ± 13.70.818

Discussion

The present study addressed motor disturbances and their association with arousals in narcolepsy–cataplexy by performing a systematic video-PSG analysis of all motor events during sleep compared with a sex- and age-matched control group.

Our main finding is that patients with narcolepsy–cataplexy exhibited increased rates of motor events during sleep compared with controls. This is in line with the hypothesis of a sleep motor dysregulation in narcolepsy (Mayer and Meier-Ewert, 1993). The hypocretinergic system, which has an important role in narcolepsy with cataplexy (Thannickal et al., 2000), does not only stabilize and coordinate sleep–wake switches (Saper et al., 2001), but is also a key player in motor control regulation (Brown et al., 2008; Yamuy et al., 2004) by hypocretinergic efferent projections to the subcoeruleus nucleus (= sublaterodorsal nucleus) in the lower pons (Brown et al., 2008) and by directly projecting to lumbar motor neurons (Yamuy et al., 2004). The sublaterodorsal nucleus was shown in the rat model to be necessary for maintaining muscle atonia (Boissard et al., 2002; Lu et al., 2006). In a recent study in human narcolepsy, hypocretin deficiency was shown to be independently associated with disturbed motor control during wakefulness (cataplexy), non-REM sleep and REM sleep (periodic leg movements in sleep, short and long muscle activations; Knudsen et al., 2010). The authors therefore concluded that this might support a general instability of motor regulation itself (Knudsen et al., 2010).

Interestingly, the motor activity index did not differ between non-REM and REM sleep in narcolepsy–cataplexy in contrast to controls. While a number of studies addressed the REM sleep motor disturbance in narcolepsy–cataplexy, the motor disturbance of non-REM sleep was so far a rather underestimated feature. Nevertheless, as early as 1993, Mayer and Meier-Ewert reported on motor disturbance of non-REM sleep in patients with narcolepsy–cataplexy, and suggested that the motor dyscontrol starts in non-REM sleep with the onset of narcolepsy as the turning point for its intrusion into REM sleep (Mayer and Meier-Ewert, 1993). Several other authors consistently reported on increased periodic leg movement indices in both non-REM and REM sleep in narcolepsy (Dauvilliers et al., 2007; Ferri et al., 2008; Mattarozzi et al., 2008), pointing to a general disturbed sleep motor regulation related to the disease itself.

Of note, not only motor activity itself but also arousal rates that were predominantly related to motor activity were increased in narcolepsy–cataplexy. Sleep fragmentation in narcolepsy has been attributed to a disturbance of sleep homeostasis, which has been postulated by several authors (Mukai et al., 2003; Nobili et al., 1995; Tafti et al., 1992). In line with this concept, Khatami et al., (2007, 2008) suggested that insufficient non-REM sleep intensity is associated with non-consolidated nocturnal sleep in narcolepsy. Further known reasons for sleep fragmentation consist of other sleep comorbidities in narcolepsy, for example restless legs syndrome, periodic leg movements in sleep and sleep-related breathing disorders (Lackner et al., 2008).

In our study, four patients had REM sleep without atonia, but none of the investigated patients had a history of REM sleep behaviour disorder. The finding that patients with narcolepsy had a higher percentage of motor events lasting more than 1 s compared with the control group is interesting, given that in a recent study in REM sleep behaviour disorder, which differentiated between short- and long-lasting muscle activity, long-lasting muscle activity was associated with a higher probability of having REM sleep behaviour disorder (Mayer et al., 2008). This association raises the question, if patients with narcolepsy with REM sleep without atonia will eventually develop REM sleep behaviour disorder in the course of the disease.

Concerning motor activity indices, one could only speculate if narcolepsy–cataplexy patients with full-blown REM sleep behaviour disorder have even higher rates of motor events during REM sleep, similar to that in patients with Parkinson’s disease with REM sleep behaviour disorder (Frauscher et al., 2007), or if alternatively there is only a mild difference in the rate of motor events between patients with narcolepsy with and without REM sleep behaviour disorder. The latter is at least to some extent supported by the existing literature giving evidence that REM sleep behaviour disorder in narcolepsy is of much milder intensity (Tafti et al., 1992) and that it is often polysomnographically not different from narcolepsy without REM sleep behaviour disorder (Ferri et al., 2008).

Increased periodic leg movements in sleep indices > 15 h−1 in patients 5 and 6 may have contributed to the higher motor activity indices in these patients. However, not only motor events of the lower extremities, but also of several other body parts were increased. Moreover, in patient 4, who had no periodic leg movements in sleep, motor activity was markedly increased. Therefore, we do not think that periodic leg movements are a relevant confounding factor of our findings.

The majority of the motor events were accompanied by either movement artefacts in the EEG channels or EMG activity in chin or tibialis anterior muscles. From a practical point of view, this means that selective screening of the video only in case of movement artefacts in the EEG or in case of an increased EMG activity in the mentalis or tibialis muscle is sufficient to capture most motor activity in the video. This is of relevance, as time constraints in clinical routine make it impossible to look through the whole-night videography in real time. Concerning movement artefacts in the EEG channels, it is noteworthy to mention that controls had significantly less artefacts than patients. This might be explained by the fact that motor events in controls were of shorter duration and involved less body parts than in patients.

One potential limitation of our study is that our patient sample with narcolepsy–cataplexy was relatively small and heterogeneous concerning age at investigation, disease duration and severity of excessive daytime sleepiness, which might have influenced our findings. For example, patients 3 and 4 had demographic features that were very different from the rest of the patient group. On the other hand, one might argue that from a clinical point of view this heterogeneity reflects the broad clinical spectrum of narcolepsy–cataplexy, and that results of our study are of even more value as they apply not only for a patient subpopulation of a certain age or a certain disease duration, but for the disease in general. Moreover, the issues of sample size and clinical heterogeneity seem more likely to render statistical results non-significant than to create false-positive associations.

In conclusion, patients with narcolepsy–cataplexy exhibit high numbers of motor events during non-REM and REM sleep. This adds further evidence to the previously suggested concept of an instable or dysregulated motor control in this disease. In addition, increased arousal rates in association with motor events contribute to sleep fragmentation in narcolepsy.

Acknowledgements

The authors thank Heinz Hackner for the always-excellent technical realization of video-polysomnographies. Tina Falkenstetter was supported by the National Bank of Austria (Anniversary fund 12594).

Conflicts of interests

All authors have neither personal nor financial disclosures in the subject matter of the paper, nor are they involved with organizations with financial interest in the subject matter of the paper.

Ancillary