Abstract
In amyotrophic lateral sclerosis (ALS), the progressive loss of upper and lower motor neurons leads to respiratory failure, often with predominant diaphragm dysfunction, and death. Because the diaphragm is the only active inspiratory muscle during rapid eye movement (REM) sleep, there is a high theoretical risk of respiratory disorders during REM sleep in patients with ALS. To assess this hypothesis, we studied sleep characteristics (polysomnography) in 21 patients with ALS, stratified according to the presence or absence of diaphragmatic dysfunction. Diaphragmatic dysfunction was defined as an absent or delayed diaphragm response to cervical or cortical magnetic stimulation, abdominal paradox, or respiratory pulse (Group 1, 13 patients). These patients did not differ in age, clinical course, or form (bulbar or spinal) from the eight others, who did not have diaphragmatic dysfunction (Group 2). REM sleep was reduced in Group 1 (7 ± 7% of total sleep time; mean ± SD) and normal in Group 2 (18 ± 6%, p = 0.004). Apneas or hypopneas were rare in both groups. In Group 1, REM sleep was absent or minimal (less than 3 min) in five patients. An unusual and remarkable preservation of phasic inspiratory sternomastoid activation during REM was associated with longer REM sleep duration in six of the other patients with diaphragmatic dysfunction. Median survival time was dramatically shorter (217 d) in Group 1 than in Group 2 (619 d, p = 0.015). Arnulf I, Similowski T, Salachas F, Garma L, Mehiri S, Attali V, Behin-Bellhesen V, Meininger V, Derenne J-P. Sleep disorders and diaphragmatic function in patients with amyotrophic lateral sclerosis.
Amyotrophic lateral sclerosis (ALS) is a neurodegenerative disorder characterized by a progressive loss of lower and upper motor neurons (1). The course of the disease is inexorably progressive. Median survival time varies from 36 to 48 mo, related with various factors including vital capacity and a bulbar site of onset (2, 3). During the course of the disease, the involvement of various respiratory muscle groups leads to a restrictive ventilatory defect and ultimately to hypercapnic respiratory insufficiency. In the majority of cases, death is related to respiratory events (4). Among the mechanisms of respiratory impairment in ALS, diaphragmatic dysfunction probably plays an important role (5). We have recently reported that patients with ALS exhibited features typical of diaphragm paralysis with compensatory use of inspiratory neck muscles, and we suggested that diaphragm dysfunction is a major determinant of dyspnea in ALS (6, 7).
Such predominant diaphragm dysfunction implies a high probability of major respiratory abnormalities during sleep. During rapid eye movement (REM) sleep, the maintenance of ventilation is critically dependent on diaphragm contraction because of the inhibition of intercostal and accessory inspiratory muscles (8). Diseases that affect diaphragm function, either through an intrinsic neuromuscular process (9) or by putting the diaphragm at mechanical disadvantage (e.g., chronic obstructive pulmonary disease [COPD]) are associated with sleep-disordered breathing (10). In the present study, we have tried to verify the hypothesis that diaphragm dysfunction in ALS is associated with changes in sleep architecture. Our major finding is that diaphragmatic paresis or palsy is associated with a reduction in REM sleep, and that this seems to have an impact on survival.
Twenty-one patients with definite ALS according to the standard criteria (11) were analyzed prospectively from March 1995 to June 1996. They were 15 men and six women, ages 38 to 80 yr, with a mean age of 59 yr. Weight ranged from 53 to 96 kg and height ranged from 154 to 187 cm. Mean disease duration was 17 mo (range 3 to 54 mo). Exclusion criteria were contraindications to magnetic stimulation (namely, a history of neurosurgery of the head, epilepsy, cardiac pacemaker). The study involved only noninvasive techniques, routinely used in pulmonary function and clinical neurophysiology laboratories. All patients were informed of the purpose of the study and the methods used, and had given consent to participate. The cohort was followed until February 1999, with quarterly visits to the center. Date of death was recorded, and no patient was lost to follow-up.
Neurologic assessment, clinical respiratory assessment, pulmonary function tests, electrophysiological diaphragm studies, and sleep studies were performed and analyzed separately by investigators unaware of the other results. In each patient, all data were gathered over a 1-wk period.
The data gathered were as follows:
1. Severity of the disease, using different scores (limb and bulbar functional testing, manual muscular testing) (12). Disease duration at the time of the study was recorded. The use of drugs known to interfere with sleep architecture was systematically noted.
2. Detailed clinical respiratory examination, with emphasis on diaphragm dysfunction (13).
3. Standard pulmonary function tests (21 patients), including spirometry and flow–volume curves (Pulmonet III; Gould, Cleveland, OH). Predicted values were taken from the 1993 European recommendations for pulmonary function tests (14). Arterial blood gases were not performed systematically for this study, to keep it as noninvasive as possible, in line with the policy of the department of neurology. Because of their very low sensitivity to stage the respiratory impact of ALS (15, 16), it was not felt that the quantum of additional information would be significant.
4. Static maximal pressures (15 patients), inspiratory (Pi max) and expiratory (Pe max), were measured at the mouth at functional residual capacity (FRC) using an occluded mouthpiece with a small leak using a Validyne DP-45 pressure transducer linear within the range ± 150 cm H2O (Validyne, Northridge, CA); predicted values were taken from Black and Hyatt (17); a similar transducer was used to record mouth pressure twitches (Pm,t) during phrenic stimulation. This information is available in only 15 of the 21 patients; in the other six (three from each group), static pressures could not be measured because of difficulties with the mouthpiece or because of insufficient coordination of respiratory movements.
5. Diaphragm electromyogram, in response to cervical and cortical (18, 19) magnetic stimulation (CMS and CxMS, respectively).
Technique. Surface recordings of the right and left diaphragmatic electromyograms (EMGdi) were obtained using disposable skin-tapered silver electrodes placed in the sixth to eighth intercostal spaces and connected to a Neuropack Sigma electromyograph (Nihon Kohden, Tokyo, Japan). Magnetic stimulation was performed with a Magstim 200 upgraded stimulator (Magstim, Sheffield, UK; 90-mm doughnut-shaped coil, maximal output 2.5 T). All stimulations were performed at FRC. The minimal interval between two stimulations was 2 min. Surface electromyograms of the right and left abductor pollicis brevis muscles (APB) (EMGapb) were simultaneously recorded to serve as control. EMG signals were band pass filtered between 10 Hz and 2 kHz.
Data handling. The latency of an EMG response to CMS (henceforth termed M-wave) or to CxMS (henceforth termed motor-evoked potential [MEP]) was defined as the time elapsed between stimulation and the first departure of the EMG signal from the baseline. For CxMS, the shortest latency measured in three repeated trials was kept for analysis (20).
Abnormal responses to CMS or CxMS consisted of absent or delayed M-wave and MEP. Cortex-to-diaphragm (CDCT), cortex to APB (CACT), phrenic nerve (PNCT), and peripheral conduction time to APB (ACT) were measured.
The patients were stratified a posteriori according to the presence or absence of diaphragmatic dysfunction, defined as the presence of one or more of the following: (1) paradoxical inward movement of the abdominal wall during quiet inspiration (AB paradox); (2) respiratory pulse, defined as a visible inspiratory contraction of inspiratory neck muscles (13) (we have previously shown that ALS patients with diaphragmatic dysfunction can have abnormal inspiratory neck muscles, recruited during tidal volume despite the absence of gross respiratory mechanics abnormalities, and atypically capable of producing inspiratory pressures in response to CMS [7]); (3) absence of diaphragm M-wave in response to CMS, or a PNCT above 8 ms (namely, at least 40% longer than what would be considered normal with this technique [21]); and/or (4) absence of diaphragm MEP in response to CxMS, or a cortex-to-diaphragm latency superior to 18 ms (19, 21).
Night polysomnography was performed in the 21 patients studied using a 16-channel analogic polygraph (ECEM alpha; Oxford Instrument, Orsay, France, n = 21) with a paper speed set at 15 mm/s, an infrared pulse oximeter (Biox 3740; Ohmeda Ltd, Denver, CO). The recordings included electroencephalogram (EEG) (leads Fp2/Cz, Fpl/ Cz, O2/Cz, O1/Cz), electro-oculogram (EOG), EMG of the levator menti and tibialis anterior for sleep stage assessment, naso-oral flux through thermistors, abdominal and thoracic displacements, and electrocardiogram. Phase angle between abdominal and thoracic displacements was measured in addition to the above recordings in 11 patients (seven in Group 1, four in Group 2), using an inductance plethysmography based respiratory monitoring device (DMS 100; Densa Ltd, Flint, UK). Thoracoabdominal phase angle and pulsed oximetry were not displayed on the polysomnograph paper recordings, but were time-synchronized with it. Sleep stages were scored according to standard criteria (22), microarousals with the American Sleep Disorders Association (ASDA) criteria (23), and leg movements with Coleman's criteria (24). Microarousal index represented the number of microarousals per sleep hour. In addition, REM sleep was divided into phasic and tonic stages, according to the presence or the absence of REM bursts, respectively. Obstructive apneas were defined as complete cessation of naso-oral flux for ⩾ 10 s with persistence of thoracic and/or abdominal movements. Central apneas were defined as complete cessation of both naso-oral flux and thoracoabdominal movements. Mixed apneas were defined as a succession of central and obstructive apneas. Hypopneas were defined as 50% drops in naso-oral flux during at least 10 s followed by an arousal. Sleep was recorded from light off (11:00 p.m.) to light on (7:00 a.m.). Daytime sleepiness was assessed using the Epworth Sleepiness Scale (25).
Surface EMG of the genioglossus (GG) and sternomastoid (SM) muscles were recorded on paper, with the polysomnographic signals. SM activity was recorded using two surface silver chloride monopolar electrodes taped on the skin over the SM muscle mass, with a 3-cm interelectrode distance. The quality of SM recordings was checked during opposed rotation of the head toward the contralateral side and during deep inspirations. To record the GG signals, two similar electrodes were taped to the skin, approximately 3 cm behind the anterior part of the lower mandibular bone, with a 2-cm interelectrode distance. The exact position of the electrodes was chosen according to palpation of the area during protraction of the tongue. The quality of the GG recordings was checked during protraction of the tongue and during swallowing efforts. In our experience, there is a good correlation between the EMG of the GG recorded with this technique or with intramuscular needles. The EMG signals were fed to an AC amplifier (ECEM alpha; Oxford Instrument) with low–high band-pass filters set at 30 and 1,000 Hz, respectively.
Right-to-left comparisons of physiological data (central and peripheral conduction times) were performed using a two-tailed Student's paired t test. When no statistical difference was detected, right and left values were pooled.
Differences between patients stratified according to presence of diaphragm dysfunction (Group 1, diaphragm dysfunction; Group 2, no diaphragm dysfunction) were assessed using the Mann-Whitney test for continuous variables, and a chi-square test for dichotomous variables. The usual p < 0.05 threshold for significance was retained for all analysis.
Survival was compared between Group 1 and Group 2 using a Kaplan-Meier analysis and the log rank tests and the Wilcoxon test to account for the high likeliness of marked early differences (26).
The statistical analysis was performed using StatView 5.0 software and the SAS 6.12 package (SAS Institute, Inc., San Francisco, CA).
General characteristics. According to the above definition, diaphragm dysfunction was present in 13 patients (Group 1) and absent in eight patients (Group 2). Demographic and clinical characteristics of the patients are shown in Table 1. Patients in both groups were matched for these variables.
| Group 1 Diaphragmatic Dysfunction (n = 13) | Group 2 Diaphragmatic Dysfunction (n = 8) | |||
|---|---|---|---|---|
| Age, yr | 60 ± 12 (38–80) | 56 ± 9 (45–75) | ||
| Male/female | 10/3 | 5/3 | ||
| Weight, kg | 67 ± 13 (53–96) | 64 ± 6 (54–73) | ||
| Height, cm | 167 ± 10 (154–187) | 168 ± 9 (155–178) | ||
| Disease duration, mo | 19 ± 14 (7–54) | 14 ± 9 (3–26) | ||
| Limb functional testing† | 45 ± 17 (5–63) | 53 ± 3 (47–58) | ||
| Bulbar functional testing† | 27 ± 11 (7–38) | 33 ± 5 (24–39) | ||
| Manual muscle testing† | 117 ± 16 (87–136) | 123 ± 24 (97–147) |
Pulmonary function tests. The number of dyspneic patients (four patients in Group 1 versus three in Group 2) and the intensity of dyspnea (1.3 ± 2.6 in Group 1 versus 1.1 ± 1.8 in Group 2) were not significantly different. By definition, no patient had abdominal paradox in Group 2 (versus seven in Group 1). Respiratory pulse was constantly absent in Group 2 and present in seven patients in Group 1.
Table 2 summarizes the results of pulmonary function tests. Spirometric data were not significantly different between both groups, except for vital capacity, which was significantly reduced in Group 1. Forced expiratory volume in one second (FEV1) tended to be smaller in Group 1, but the difference did not reach significance.
| Group 1 Diaphragmatic Dysfunction | Group 2 No Diaphragmatic Dysfunction | |||
|---|---|---|---|---|
| VC, % pred | 60.7 ± 24 (38–108) | 85.8 ± 23.1† (56–119) | ||
| Median = 55 | Median = 90 | |||
| FRC, % pred | 95.3 ± 46.9 (58–148) | 114.5 ± 24.7 (97–132) | ||
| TLC, % pred | 84.3 ± 37.2 (45–119) | 107.3 ± 18.5 (86–118) | ||
| RV, % pred | 97.7 ± 33 (61–125) | 115.7 ± 27.1 (86–139) | ||
| FEV1, % pred | 63.3 ± 21.5 (36–88) | 89.8 ± 24.9 (49–119) | ||
| Pi max at FRC, % pred | 33.3 ± 26.3 (6–88) | 51.6 ± 19.1 (31–82) | ||
| Pe max at FRC, % pred | 39.7 ± 18.3 (22–70) | 54.4 ± 33.5 (16–108) |
Transcutaneous oxygen saturation during room air breathing was normal for both groups (wake SpO2 96 ± 2% and 99 ± 1% in Group 1 and 2, respectively).
Electrophysiologic data. An EMG response to CMS was present in all patients in Group 2, with a PNCT of 7.3 ± 1.1 ms on the right side and 6.7 ± 1.8 ms on the left side (not significant [NS]). In one patient in Group 1, CMS did not evoke any motor response on either side in spite of satisfactory recording conditions. In the other patients, PNCT was 7.7 ± 1.4 ms on the right side, and 8.1 ± 2.5 ms on the left side (NS). In spite of the fact that the longest PNCTs were found in Group 1, there was no statistically significant difference between groups (7.9 ± 2.0 ms versus 7.0 ± 1.5, p = 0.32). The pattern of response of the APB to CMS was similar to that of the diaphragm, with ACT not significantly different between sides or groups. Nevertheless, the values were more scattered in Group 1 (right 18.8 ± 3.4 ms, left 19.1 ± 5.2 ms) than in Group 2 (right 14.0 ± 0.7 ms, left 17.3 ± 1.5 ms).
CxMS consistently elicited bilateral MEPs in Group 2 patients with latencies ranging from 12.4 to 18 ms. In Group 1, surface diaphragm EMG could not be recorded after CxMS in two patients considered as having diaphragm dysfunction, in spite of a persistent response to CMS. This was also the case in the patient of the same group who did not respond to CMS. In the remaining 10 patients, MEP latencies ranged from 18.2 to 23.5 ms. CDCT (pooled right and left values) was 19.6 ± 1.8 ms in Group 1, versus 16.8 ± 1.5 ms in Group 2 (p = 0.003). CACT was not significantly different between both groups (Group 1: 27.7 ± 4.9 ms; Group 2: 24.0 ± 3.9 ms, p = 0.29).
Daytime sleepiness. Epworth sleepiness scores were normal and similar in both groups (4.8 ± 2.9 versus 5.7 ± 4, NS).
Sleep architecture. Table 3 summarizes the results of sleep data. Sleep latency was relatively long in both groups. Total sleep time (TST) tended to be shorter in Group 1 than in Group 2, but the difference was not significant. Nevertheless, the frequency of awakenings after sleep onset was significantly higher in Group 1 than in Group 2. The time spent in sleep stages 1–2 and 3–4 slow wave sleep (SWS) was comparable in both groups and within usual values (27, 28). Therefore, the reduction in TST is likely the result of a reduction in REM sleep. Indeed, REM sleep duration was normal in Group 2 (59 ± 24 min, 18 ± 6% of TST), but was dramatically reduced in Group 1 (Figure 1) (20.9 ± 20.9 min, 7 ± 7% TST), this difference being significant with REM sleep expressed in minutes (p = 0.001) or in percentage of TST (p = 0.004). Among patients from Group 1, only one had a normal REM sleep duration (61 min). No REM sleep was observed in two patients. In three, REM sleep was reduced to a short bout of approximately 60 s duration, consistently terminated by a prolonged awakening. In the remaining two patients, REM sleep duration was abnormally low (13 and 48 min, respectively). In summary, all patients in Group 1 but one had an abnormally low REM sleep duration for their age.
| Group 1 Diaphragmatic Dysfunction | Group 2 No Diaphragmatic Dysfunction | |||
|---|---|---|---|---|
| Total sleep period, min | 388 ± 100 (122–557) | 402 ± 60 (264–454) | ||
| Total sleep time (TST), min | 282 ± 90 (70–432) | 339 ± 56 (234–417) | ||
| Wake after sleep onset | 106 ± 10 (52–125) | 63 ± 4† (30–151) | ||
| Sleep latency, min | 77 ± 106 (18–398) | 49 ± 39 (9–122) | ||
| Arousal index | 26 ± 15 (7–56) | 20 ± 9 (9–37) | ||
| Stages 1–2 duration, % TST | 68 ± 11 (48–86) | 63 ± 8 (54–77) | ||
| Slow wave sleep duration, % TST | 24 ± 9 (14–42) | 20 ± 5 (15–29) | ||
| REM sleep duration, % TST | 7 ± 7 (0–22) | 18 ± 6† (8–24) | ||
| Apnea–hypopnea index | 3.4 ± 5.3 (0–19) | 1.4 ± 1.9 (0–5) | ||
| SaO2 per min, % | 80.5 ± 11 (46–90) | 88 ± 5 (79–94) |
Fig. 1. REM sleep duration (in percentage of TST) in Group 1 patients (diaphragmatic dysfunction) and in Group 2 patients (no diaphragmatic dysfunction). The horizontal line indicates the mean value for each group.
[More] [Minimize]Sleep fragmentation. Periodic leg movements sequences were observed in eight patients (five in Group 1 and three in Group 2), not associated with arousal. Microarousal indexes were not different between groups. They ranged from 5 to 54 in Group 1 (mean: 22 ± 14) and from 7 to 34 (mean: 17 ± 9) in Group 2. Awakenings lasting more than 15 s were rare and of similar frequency between groups (mean: 4 ± 3 per hour in Group 1 and 2 ± 1 in Group 2). The fragmentation indexes including microarousals and awakenings were not different between groups (Table 3).
Polygraphic respiratory parameters. In Group 1, 11 patients exhibited tonic SM EMG activity while awake, with inspiratory bursts of SM EMG activity superimposed in eight of them (Figure 2).
Fig. 2. Inspiratory neck muscles EMG in one patient from Group 1 (diaphragmatic dysfunction). Inspiratory neck muscles typically exhibited tonic and phasic activity during wakefulness (left panel ), inspiratory bursts of activity during SWS (middle panel ), and no activity during REM sleep (right panel ). EEG = electroencephalogram; EOG = electro-oculogram; EMG = electromyogram; SM = sternomastoid; RC = rib cage movements; ABD = abdominal movements.
[More] [Minimize]During non–rapid eye movement (NREM) sleep, the inspiratory bursts of SM EMG activity persisted in all cases, but the tonic component was always abolished (Figure 2). A phasic inspiratory SM pattern was observed during NREM sleep in four patients of Group 2. The transition between sleep and wakefulness was synchronous with the reappearance of the continuous tonic SM contraction.
During REM sleep, the EMG of the sternomastoid was silent or exhibited an erratic unorganized single fiber activity in Group 2 patients. During the single 60-s bout of REM sleep recorded in three patients in Group 1, respiratory rate was increased, the amplitude of naso-oral flux and of thoracic and abdominal bands decreased, and no SM activation was seen. A typical record, obtained during the transition from NREM to REM sleep and from REM to wake, is shown in Figure 3. In Group 1, the three patients with the longest REM sleep duration exhibited unexpected and consistent inspiratory bursts of SM EMB during this stage (Figure 4, left panel ). In three other patients, a similar pattern of SM activity was restricted to phasic REM sleep, synchronous with bursts of rapid eye movements (Figure 4, right panel ). These patterns were not observed in the Group 1 patients with the shortest REM duration (Figure 3), nor in any of the patients in Group 2.
Fig. 3. Example of a polysomnographic pattern that was observed solely in three patients from Group 1 (diaphragmatic dysfunction). During a REM sleep epoch, rapid shallow breathing suggesting hypoventilation occurs, followed by an arousal. Inspiratory neck muscles phasic activity, sustained during the occurrence of REM sleep, disappears during REM sleep and reappears after the arousal. SM = sternomastoid; GG = genioglossus; RC = rib cage movements; ABD = abdominal movements.
[More] [Minimize]
Fig. 4. Two REM sleep breathing patterns observed in those Group 1 patients (diaphragmatic dysfunction) who had subnormal REM sleep duration. In the left panel, inspiratory sternomastoid activity persists during the whole REM sleep epoch. There is a sudden decrease in respiratory amplitude with a concomitant increase in respiratory rate during phasic REM sleep, as is commonly the case in normal subjects. In the right panel (corresponding to another patient), sternomastoid phasic inspiratory activation appears restricted to bursts of REM, together with an increase in respiratory movements and a decrease in respiratory rate.
[More] [Minimize]The apnea–hypopnea index (AHI) was within the normal range in both groups (3.4 ± 5.2 in Group 1, 1.4 ± 1.9 in Group 2, p = 0.30). There was no episode of central apnea or hypopnea during REM sleep, but two patients in Group 1 exhibited a few episodes of obstructive apnea during REM sleep. Some central hypopneas were observed during NREM sleep. One patient from Group 1 who exhibited no REM sleep at all had an AHI of 19, with exclusively central hypopneas. The minimum SpO2 recorded during sleep tended to be smaller in Group 1 (80% ± 12%, range 46 to 90%) than in Group 2 (88 ± 5%, range 79 to 94%, p = 0.09), but the difference did not reach significance. The lowest SpO2 recorded was 46%, in one patient from Group 1, and this was during NREM sleep.
The recording of abdominal and thoracic movements confirmed the clinical diagnosis of abdominal paradox in five patients from Group 1 where it was performed. The 90–180° phase opposition between the thorax and the abdomen persisted during every sleep stage.
None of the patients was lost to follow-up. Follow-up duration for surviving patients ranged from 974 to 1,438 d. Survival curves were significantly different between groups (p = 0.015; Figure 5). Survival median was 217 d for Group 1 (95% confidence interval [CI]: 177 to 506) and 619 d for Group 2 (95% CI: 530—nonestimable). The median survival time in Group 1 was 191 d (95% CI: 99 to 367) when REM sleep was quasi- absent and 243 d (95% CI: 193—not estimable) when REM sleep was present (p = 0.13).
Fig. 5. Kaplan-Meier analysis of survival, after stratification of the study population according to the presence (Group 1, n = 12) or absence (Group 2, n = 9) of diaphragmatic dysfunction. The median survival time in Group 1 patients was three times shorter than in Group 2 (p < 0.02). Vertical bars indicate the survival function standard errors.
[More] [Minimize]The main finding of this study is that diaphragmatic dysfunction in patients with ALS is associated with a dramatic reduction of REM sleep duration, and has impact on survival.
ALS is a progressive and lethal disease that affects both the upper and lower motor neuron pathways of the nervous system. Respiratory symptoms are rare at the time of diagnosis despite the early functional respiratory abnormalities (29, 30). With the progression of the disease, dyspnea becomes more frequent and respiratory muscle force declines, leading to respiratory failure and death (4, 29, 31). Indeed patients with ALS usually die from respiratory problems including alveolar hypoventilation, aspiration pneumonia, and pulmonary embolism (30, 32, 33). Death commonly occurs during sleep (34, 35).
Spirometric measurements show that in a majority of cases, regardless of the pattern of motor neuron impairment, forced vital capacity (FVC) decreases with disease progression and is correlated with survival (17, 29, 30, 32).
Static pressures, inspiratory and expiratory, are also reduced, and decline with the course of the disease (15, 36). All available data concur with the idea that the relationship between spirometry or muscle pressure and symptoms is loose.
More recently we reported that dyspnea was generally associated with diaphragmatic dysfunction (7), defined as the presence of abdominal paradox, respiratory pulse, and delayed or abolished responses of the diaphragm to phrenic nerve stimulation or cortical stimulation. These findings were consistent with previous neuropathological observations of predominant phrenic nerve abnormalities in ALS patients (37).
Rationale. Evidencing severe diaphragm dysfunction in ALS naturally leads to the question of sleep-disordered breathing. Indeed, the diaphragm, normally the main agonist of inspiration during wakefulness, is the only active inspiratory muscle during REM sleep. If diaphragm dysfunction, whatever its mechanisms, implies the recruitment of other inspiratory muscles to maintain ventilation during wakefulness, then hypoventilation is expected during REM sleep. In six patients with severe COPD exhibiting an increased participation of extradiaphragmatic muscles (namely the scalene, sternomastoid, and external intercostal muscles) to inspiration during wakefulness Johnson and Remmers (10) observed an abrupt and complete loss of activity of these muscles during REM sleep. This caused chest wall distortion with the resulting deterioration in ventilation and oxyhemoglobin desaturation. In spite of this increased respiratory vulnerability, REM sleep duration was not significantly reduced in this subset of patients.
In ALS patients with diaphragmatic dysfunction severe enough to cause inspiratory abdominal paradox, ventilation has to be produced by extradiaphragmatic muscles. In such patients, we have previously found arguments suggesting compensation for diaphragm paralysis at the level of inspiratory neck muscles (7). Therefore, the expected REM sleep-related loss of activity of neck muscles should compromise ventilatory stability in ALS patients with diaphragm dysfunction, hence our study.
Available data. Earlier sleep studies in ALS have yielded inhomogeneous results. Minz and coworkers (38) studied 12 patients with variable clinical presentation and severity. They did not observe specific polygraphic aspects, although they reported an absence of REM sleep in one patient who died soon thereafter. Gay and coworkers (16), in a prospective study of 21 patients described disordered breathing and nocturnal desaturation in a majority of them. They attributed these episodes to hypoventilation. Carre and coworkers (39) published a case of obstructive sleep apnea syndrome in an obese patient, subsequently leading to the diagnosis of ALS. More recently, Ferguson and coworkers (40) reported decreased sleep efficiency and increased sleep fragmentation in 18 patients with bulbar form of ALS during the first night of polysomnography, which improved during the second recording night. Sleep-disordered breathing was observed in eight of the 18 patients and consisted of REM-related nonobstructive hypopneas or central apneas, resulting in hypoventilation and oxyhemoglobin desaturation.
Results of the present study. We found a significant reduction of REM sleep duration in ALS patients with diaphragmatic dysfunction. This reduction was not related to aging, nor to the clinical course of ALS, and was not influenced by the clinical form (spinal or bulbar) of the disease. To our knowledge, this if the first time that a relationship is shown between evidence of diaphragm dysfunction and REM sleep disruption in a controlled study.
Possible confounders. The patients did not take drugs known to suppress REM sleep. A first night effect (41), or stress, could explain REM sleep reduction. However, SWS duration and the arousal index were normal and similar in either group. Furthermore, there is no reason to think that a first night effect, if present, would have been different between groups. REM sleep depends on homeostatic [REM sleep increases with the duration of prior NREM sleep (42)] and circadian [REM sleep epochs are longer at the trough of body temperature (43)] factors. Because in our patients both absolute and relative REM sleep duration were decreased, reduced TST or NREM sleep duration could not account for a “homeostatic” reduction in REM sleep, except in one subject from Group 1 who had an extremely short TST. Of note, SWS was abundant in patients of either group.
Because sleep recordings were performed with the same schedule in both groups, in inpatients routinely used to wake around 7:00 a.m., it is unlikely that a circadian factor could have explained the difference between groups regarding REM sleep. The awakening and microarousal indexes were similar in both groups, and therefore could not be called on to explain the REM sleep reduction in Group 1.
Finally, apneas occurring during REM sleep can reduce its duration. Obstructive apneas could have been expected in those of our patients with a bulbar form of ALS, because of upper airway dilator muscles weakness (39, 40). Yet this was not observed, and the AHI was normal and similar in both groups. This could be explained by various reasons, such as increased airway stability due to major weight loss, inability of the diaphragm to generate negative pressures large enough to reach upper airway critical closing pressure (40), and stiffening of the upper airways caused by upper motor neuron pharyngeal spasticity.
All in all, we believe that REM sleep reduction was a genuine feature of the patients with ALS and diaphragm dysfunction in this study.
Methodologically speaking, we relied on thermistor signals to study the ventilatory pattern in our patients because we wished to avoid the use of esophageal probes and because current tools (nasal cannula, pulse transit time) to assess the changes in respiratory efforts were not validated when the study was conducted. In future work, these tools could help refine the analysis of respiratory events in similar populations.
Among the patients constituting Group 1, it seems possible to identify two main patterns. In some patients, REM sleep was absent or almost so. In others, REM sleep was reduced quantitatively but still relatively abundant. What seems to distinguish the two patterns is the absence or presence of a significant extradiaphragmatic inspiratory activity during REM sleep, respectively.
Changes in the control of extradiaphragmatic inspiratory muscles during REM sleep. The patients with diaphragm dysfunction in whom REM sleep significantly persisted (as opposed to the five in whom it was absent or limited to less than a minute) had an unusual and remarkable preservation of phasic inspiratory SM and GG activation during REM sleep, limited to the bursts of REM in three cases. In the three patients with diaphragmatic dysfunction who had the longest REM sleep durations, tonic SM and GG activity also persisted. The persistence of muscle tone during REM sleep has been described in various neurological diseases (44) and in ALS (38). It could be related to functional changes occurring within the ponto-spinal REM–antonia pathway (45, 46). In our patients, the persistence of muscle tone was restricted to SM and GG and did not occur in nonrespiratory muscles, i.e., tibialis anterior and levator menti. The persistence of phasic activity of extradiaphragmatic inspiratory muscles after phrenicotomy in rat is associated with a near-normal amount of REM sleep (47). It thus can be viewed as a compensatory plasticity allowing the central nervous system to preserve both REM sleep and adequate ventilation during it, in spite of diaphragm paralysis. Such a mechanism may have been present, although perhaps imperfectly, in the subset of patients from Group 1 in whom REM sleep persisted and was accompanied by phasic inspiratory SM contractions.
Reduction of REM sleep. REM sleep was absent in two patients and seemingly so in three others. Complete suppression of REM sleep is an uncommon phenomenon. REM sleep– related respiratory abnormalities are not usually associated with major reductions in REM (e.g., kyphoscoliosis or neuromuscular patients). In this regard, it must be kept in mind that ALS is a unique neurologic disease, characterized by degeneration not only of the lower motor neuron (as is the case, for instance, in poliomyelitis and spinal atrophy) but also of the upper motor neuron. In the three patients among these five who displayed a single, extremely short, REM sleep episode, there was no evidence of SM activation. Expectedly in the absence of accessory inspiratory muscles to assist the diseased diaphragm, REM sleep was accompanied by rapid shallow breathing suggestive of hypoventilation (Figure 3), the occurrence of which was rapidly followed by sustained awakening. Contrary to what has been reported in patients with bilateral or unilateral diaphragmatic weakness in other diseases (8, 48– 50), severe alveolar hypoventilation was not observed during REM sleep in our patients, where no deep oxyhemoglobin desaturations were present (see Results). This may suggest that awakening from REM sleep occurred before the time necessary for hypoxemia severe enough to develop had elapsed, and hence was triggered by other events. It may be hypothesized that the changing breathing pattern and in particular the increasing respiratory rate during hypoventilation episodes were perceived precociously as “alarm signals.” In support of this contention is the fact that we did not observe recurrent central apneas during REM sleep, but rather episodes suggestive of hypoventilation. We are nevertheless aware that the notion of hypoventilation in this study has to be taken with extreme caution in the absence of direct measurement of flow or esophageal pressures. This is a limit to the study, but was a deliberate choice motivated by the difficulties for patients with ALS to tolerate orofacial devices and to swallow probes. Restricting the study to patients capable of standing these constraints would have dramatically biased it.
In summary, we end up with a possible scenario that it is largely hypothetical and needs further validation, but has the merit to reconcile the various findings made in our patients. Diaphragm dysfunction in ALS could be first compensated during REM sleep (or partly compensated) by some degree of plasticity of the respiratory control system leading to the persistence of an inspiratory activity of inspiratory neck muscles during this stage of sleep. In patients failing to establish this compensation, or losing it with progression of the disease, REM sleep would first be reduced in duration because of hypoventilation-related cessation of REM sleep, then completely disappear, possibly as a “protective” mechanism against hypoventilation. Our survival data may be integrated in such a theory. Diaphragmatic dysfunction in our series of patients was associated not only with a decrease or suppression of REM sleep, but also with a survival time reduced threefold. In addition, within Group 1, survival tended to be shorter in patients with the smallest durations of REM sleep. Thus it is conceivable that patients with ALS surviving despite severe pulmonary function impairment do so because of SM activation during REM sleep (and die when they ultimately lose this adaptation), whereas others succumb to respiratory failure with larger and presumably adequate functional reserve (34). A longitudinal study is needed to support this hypothesis.
In conclusion, ALS-related diaphragmatic dysfunction can be associated with a dramatic reduction in REM sleep and a reduced survival. The features of the REM sleep reduction that we observed, which have not been described in other neuromuscular diseases, suggest that ALS could provide a peculiar model of neurological compensation of the respiratory vulnerability owing to diaphragm dysfunction. The finding of an impaired sleep in ALS has several consequences. Pathophysiologically, it suggests that nocturnal ventilation could be beneficial not only in preventing episodes of hypoventilation, but also in preserving daytime inspiratory muscle function. Indeed, a negative impact of sleep deprivation on inspiratory muscles endurance has been shown in normals (51). Clinically, our results lend support to a strategy for nocturnal ventilatory assistance in ALS based on diaphragm function studies or on polysomnographic studies that could provide indexes more sensitive than the currently available ones.
The authors thank Alain Mallet, from the Department of Biostatistics, Groupe Epidemiologie et Recherche Clinique, Hopital Pitie-Salpetriere, for his help in statistical analysis.
Supported in part by a grant from the French “Comite National contre les Maladies Respiratoires et la Tuberculose” (CNMRT).
| 1. | Charcot J., Joffroy A.Deux cas d'atrophie musculaire progressive avec lesions de la substance grise et des faisceaux anterolateraux de la moelle epiniere. Arch. Physiol.21869354 |
| 2. | Jablecki C. K., Berry C., Leach J.Survival prediction in amyotrophic lateral sclerosis. Muscle Nerve121989833841 |
| 3. | Bensimon G., Lacomblez L., Meininger V.Amyotrophic Lateral Sclerosis Rilizole Study GroupA controlled trial of riluzole in amyotrophic lateral sclerosis. N. Engl. J. Med.3301994585591 |
| 4. | Caroscio J. T., Mulvihill M. N., Sterling R., Abrams B.Amyotrophic lateral sclerosis. Neurol. Clin.519871 |
| 5. | Serisier D. E., Mastaglia P. L., Gibson G. J.Respiratory muscle function and ventilatory control: I. In patients with motor neurone disease. II. In patients with myotonic dystrophy. Q. J. Med.2021982205226 |
| 6. | Similowski, T., V. Attali, G. Bensimon, F. Salachas, S. Mehiri, I. Arnulf, L. Lacomblezs, M. Zeltter, V. Meininger, and J.-P. Derenne. 1999. Diaphragmatic dysfunction and dyspnoea in amyotrophic lateral sclerosis. Eur. Respir. J. (In press) |
| 7. | Attali V., Mehiri S., Straus C., Salachas F., Arnulf I., Meininger V., Derenne J.-P., Similowski T.Influence of neck muscles on mouth pressure response to cervical magnetic stimulation. Am. J. Respir. Crit. Care Med.1561997509514 |
| 8. | Stradling J. R., Kozar L. F., Dark J., Kirby T., Andrey S. M., Phillipson E. A.Effect of acute diaphragm paralysis on ventilation in awake and sleeping dogs. Am. Rev. Respir. Dis.1361987633637 |
| 9. | Bye P. T., Ellis E. R., Issa F. G., Donnelly P. M., Sullivan C. E.Respiratory failure and sleep in neuromuscular disease. Thorax451990241247 |
| 10. | Johnson M. W., Remmers J. E.Accessory muscle activity during sleep in chronic obstructive pulmonary disease. J. Appl. Physiol.57198410111017 |
| 11. | Brooks B. R.El Escorial World Federation of Neurology criteria for the diagnosis of amyotrophic lateral sclerosis. Subcommittee on Motor Neuron Diseases/Amyotrophic Lateral Sclerosis of the World Federation of Neurology Research Group on Neuromuscular Diseases and the El Escorial “Clinical limits of amyotrophic lateral sclerosis” workshop contributors. J. Neurol. Sci.1241994S96S107 |
| 12. | Lacomblez L., Bensimon G., Leigh P. N., Guillet P., Meininger V.Dose-ranging study of riluzole in amyotrophic lateral sclerosis. Amyotrophic Lateral Sclerosis/Riluzole Study Group II. Lancet347199614251431 |
| 13. | Maitre B., Similowski T., Derenne J. P.Physical examination of the adult patient with respiratory diseases: inspection and palpation. Eur. Respir. J.8199515841593 |
| 14. | European Respiratory Society. 1994. Edition Française des Recommandations Européennes pour les Explorations Fonctionnelles Respiratoires. Rev. Mal. Resp. 11(Suppl. 3):5–33. |
| 15. | Vitacca M., Clini E., Facchetti D., Pagani M., Poloni M., Porta R., Ambrosino N.Breathing pattern and respiratory mechanics in patients with amyotrophic lateral sclerosis. Eur. Respir. J.10199716141621 |
| 16. | Gay P. C., Westbrook P. R., Daube J. R., Litchy W. J., Windebank A. J., Iverson R.Effects of alterations in pulmonary function and sleep variables on survival in patients with amyotrophic lateral sclerosis. Mayo Clin. Proc.661991686694 |
| 17. | Black L., Hyatt R.Maximal respiratory pressures: normal values and relationship to age and sex. Am. Rev. Respir. Dis.991969696702 |
| 18. | Similowski T., Fleury B., Launois S., Cathala H. P., Bouche P., Derenne J. P.Cervical magnetic stimulation: a new painless method for bilateral phrenic nerve stimulation in conscious humans. J. Appl. Physiol.67198913111318 |
| 19. | Similowski T., Straus C., Coı̈c L., Derenne J. P.Facilitation-independent response of the diaphragm to cortical magnetic stimulation in conscious man. Am. J. Respir. Crit. Care Med.154199617711777 |
| 20. | Similowski T., Catala M., Rancurel G., Derenne J.-P.Impairment of central motor conduction to the diaphragm in stroke. Am. J. Respir. Crit. Care Med.1541996436441 |
| 21. | Similowski T., Mehiri S., Attali V., Duguet A., Straus C., Derenne J.-P.Comparison of magnetic and electrical phrenic nerve stimulation in assessment of phrenic nerve conduction time. J. Appl. Physiol.82199711901199 |
| 22. | Rechstchaffen, A., and A. Kales. 1968. A Manual of Standardized Terminology, Techniques and Scoring System for Sleep Stages of Human Subjects. UCLA Brain Information Service/Brain Research Institute, Los Angeles. 1–12. |
| 23. | American Sleep Disorders AssociationEEG arousals: scoring rules and examples. A preliminary report from the Sleep Disorders Atlas Task Force of the American Sleep Disorders Association. Sleep151992174184 |
| 24. | Coleman, R. M. 1982. Periodic movements in sleep (nocturnal myoclonus) and restless leg syndrome. In C. Guilleminault, editor. Sleeping and Waking Disorders. Menlo Park, CA. 265–295. |
| 25. | Johns M. H.A new method for measuring daytime sleepiness: the Epworth Sleepiness Scale. Sleep141991540545 |
| 26. | Kalbfleish, R., and R. L. Prentice. 1980. The Statistical Analysis of Failure Time Data. John Wiley, New York. 144–147. |
| 27. | Feinberg I.Changes in sleep patterns with age. J. Psychiat. Res.101974283306 |
| 28. | Bixler E. O., Kales A., Jacoby J. A., Soldatos C. R., Vela-Bueno A.Nocturnal sleep and wakefulness: effects of age and sex in normal sleepers. Int. J. Neurosci.2319843342 |
| 29. | Schiffman P. L., Belsch J. M.Effect of inspiratory resistance and theophylline on respiratory muscle strength in patients with amyotrophic lateral sclerosis. Am. Rev. Respir. Dis.139198914181423 |
| 30. | Fallat R. J., Jewitt B., Bass M., Kamm B., Norris F. H.Spirometry in amyotrophic lateral sclerosis. Arch. Neurol.3619797480 |
| 31. | Tandon R., Bradley W. G.Amyotrophic lateral sclerosis: part 1. Clinical features, pathology and ethical issues in management. Ann. Neurol.181985271280 |
| 32. | Nakano K. K., Blass H., Tyler H. R., Carmel R. J.Amyotrophic lateral sclerosis: a study of pulmonary function. Dis. Nerv. Syst.3719763235 |
| 33. | Williams D. B., Wiredebank A. J.Motor neuron disease (amyotrophic lateral sclerosis). Mayo Clin. Proc.6619915482 |
| 34. | Miller, A. 1986. Pulmonary function and respiratory failure in neuromuscular diseases with reference to amyotrophic lateral sclerosis. In J. T. Caroscio, editor. Amyotrophic Lateral Sclerosis: A Guide to Patient Care. Thieme, New York. 61–77. |
| 35. | Shimizu T., Hayashi H., Kato S., Hayashi M., Tanabe H., Oda M.Circulatory collapse and sudden death in respirator-dependent amyotrophic lateral sclerosis. J. Neurol. Sci.12419944555 |
| 36. | Polkey M. I., Lyall R. A., Green M., Leigh, Nigel P., Moxham J.Expiratory muscle function in amyotrophic lateral sclerosis. Am. J. Respir. Crit. Care Med.1581998734741 |
| 37. | Bradley W. G., Good P., Rasool C. G., Adelman L. S.Morphometric and biochemical studies of peripheral nerves in amyotrophic lateral sclerosis. Ann. Neurol.141983267277 |
| 38. | Minz M. A., Autret A., Laffont F. T., Beillevaire T., Cathala H. P., Castaigne P.A study on sleep in amyotrophic lateral sclerosis. Biomedicine3019794046 |
| 39. | Carre P. C., Didier A. P., Tiberge Y. M., Arbus L. J., Leophonte P. J.Amyotrophic lateral sclerosis presenting with sleep hypopnea syndrome. Chest93198813091312 |
| 40. | Ferguson K. A., Strong M. J., Ahmad D., George C. F. P.Sleep-disordered breathing in amyotrophic lateral sclerosis. Chest1101996664669 |
| 41. | Agnew H. W. J., Webb W. B., Williams R. L.The first night effect: an EEG study of sleep. Psychophysiology21966263266 |
| 42. | Benoit O., Foret J., Bouard G.The time course of slow wave sleep and REM sleep in habitual long and short sleepers: effect of prior wakefulness. Human Neurobiol.219839196 |
| 43. | Czeisler C. A., Zimmerman J. C., Ronda J. M., Moore-Ede M. C., Weitzman E. D.Timing of REM sleep is coupled to the circadian rhythm of temperature. Sleep21980329346 |
| 44. | Schenck C. H., Bundlie S. R., Ettinger M. G., Mahowald M. W.Chronic behavioral disorders of human REM sleep: a new category of parasomnia. Sleep91986293308 |
| 45. | Jouvet M., Delorme F.Locus coeruleus et sommeil paradoxal. C. R. Soc. Biol.1591965895899 |
| 46. | Lai Y. Y., Siegel J. M.Pontomedullary glutamate receptors mediating locomotion and muscle tone suppression. J. Neurosci.11199129312937 |
| 47. | Sherrey J. H., Megirian D.After phrenicotomy the rat alters the output of the remaining respiratory muscles without changing its sleep–waking pattern. Respir. Physiol.811990213226 |
| 48. | NewsomDavis J., M. Goldman, L. Loh, and M. CassonDiaphragm function and alveolar hypoventilation. Q. J. Med.177197687100 |
| 49. | Skatrud J., Iber C., McHugh W., Rasmussen H., Nichols D.Determinants of hypoventilation during wakefulness and sleep in diaphragmatic paralysis. Am. Rev. Respir. Dis.1211980587593 |
| 50. | Patakas D., Tsara V., Zoglopitis F., Daskalopoulou E., Argyropoulou E., Maniki E.Nocturnal hypoxia in unilateral diaphragmatic paralysis. Respiration5819919599 |
| 51. | Chen H. I., Tang Y. R.Sleep loss impairs inspiratory muscle endurance. Am. Rev. Respir. Dis.1401989907909 |
