Sleep and neuromuscular disease

Patient selection and population.

The study was performed over a 10-day period at the New Mexico Neuromuscular Disorders Clinic, and all patients who were scheduled to visit the clinic during that time were asked to participate. Prior to the study, information about the study had been sent to the entire clinic population of 304 patients with the quarterly newsletter. Patients attending the NMD clinic during the 3 weeks prior to the study were reminded about the study on the chance that they might schedule a follow-up visit during the study period. All subjects who were scheduled to come to the clinic during the period when the investigation occurred knew in advance about the study. All but three patients attending the clinic over these 10 days agreed to participate.

Fifty-four adults and 10 children signed informed consent forms and were assigned testing dates. Sixty subjects completed the study. Three patients withdrew before testing and one died in his sleep 2 days before the test date. The NMD clinic population was chosen because of its diversity of neuromuscular disorders. The special benefits of free care, including neurologic, orthopedic, and physical therapy services, assure that virtually all patients with NMD in the entire state are referred there. The distribution of diagnoses among the participants was similar to the range of diagnoses in the entire clinic population.

Study design and observations.

All patient charts were reviewed for diagnosis and diagnostic data. We accepted the diagnosis made by the neurologists of the NMD clinic, which was based on clinical evaluation, clinical neurophysiologic testing, and, if needed, muscle biopsies. On the evening of the study, subjects completed a questionnaire concerning their sleep and daytime functioning, including an Epworth Sleepiness Scale evaluating subjective daytime sleepiness. ^[9-11] Other measurements were age, height, weight, blood pressure, and heart rate. Spirometry and maximum inspiratory and expiratory force were recorded in both the upright and supine positions on the same day as the sleep recording. Body mass index (BMI) was calculated. Patients were also examined by the investigators and determined not to have an acute illness. A standardized disability index, based on the clinical examination, was used to classify each patient Table 1. All observations were recorded at an altitude of 1,500 meters (Albuquerque, NM).

The subjects underwent a full-night sleep study using the multiparameter EdenTrace II recorder. This monitor was chosen for its ease of operation, its accuracy in testing for SDB, and its ability to study the patient in the home environment. Nasal oral airflow (thermistry), oxygen saturation (pulse oximetry SpO₂), chest effort (thoracic impedance), heart rate (R-R interval), body position, and tracheal sounds were recorded. Most patients were studied at home while out-of-town patients were recorded in the University Hospital Clinical Research Center. In the morning, data from the EdenTrace were downloaded and printed. Patients filled out a morning questionnaire concerning the prior night's sleep. The study was scored based upon prior published criteria ^[12,13] and those obtained during a prior validation study of the equipment performed at the Stanford University Sleep Disorders Center (see below).

The total number of central and obstructive apneas; hypopneas lateral and supine; mean, baseline, and lowest O₂ saturation awake and asleep; and RDI (average number of apneas and hypopneas per hour of sleep) were recorded. Total sleep time was estimated by using the morning questionnaires and deleting time where the record showed persistent motion artifact. Events were recorded as obstructive apnea (complete cessation of airflow with continuation of respiratory effort and O₂ desaturation greater than 3%), central apnea (complete cessation of airflow and respiratory effort with O₂ desaturation of greater than 3%), and hypopnea (decrease in airflow by one-third of baseline with continuation of normal effort and O₂ desaturation more than 3% or decrease in both airflow and respiratory effort by at least one-third from reference amplitude and O₂ desaturation of more than 3%).

Validation of a portable sleep apnea recording system.

The validity of measurements made with the EdenTrace II was assessed at the Stanford Sleep Disorders Clinic in volunteers who were recruited through newspaper advertisements. All subjects were simultaneously monitored with the portable recorder (EdenTrace II; variables listed above) and full nocturnal polysomnography. Polygraphic recording included EEG (C3/A2, C4/A2), electrooculogram, chin and leg electromyogram, and ECG (modified V2 lead). Respiration was measured by oronasal thermistors (different from EdenTrace), uncalibrated inductive plethysmography indicating thoracic and abdominal efforts, intercostal EMG, finger pulse oximetry (Nellcor), and snoring sounds through a calibrated microphone. Subjects were also under continuous surveillance through an infrared camera. All individuals completed questionnaires evaluating daytime functioning and nocturnal sleep. They were also asked to complete a sleep log indicating their estimated total sleep time, lights-out and lights-on times, number of awakenings, and amount of time spent awake during the night.

Nocturnal polygraphic recordings were scored blind by two technicians. Following the internationally accepted criteria, ^[12-14] respiratory events were classified as apnea or hypopnea and were subdivided into "central," "mixed," and "obstructive." EdenTrace II recordings were scored blind by an independent team. Apneas and hypopneas and types of events (central, mixed, and obstructive) were scored based on results of airflow and impedance recordings. The following variables were considered: number of events, number of oxygen desaturations (based on a drop of 3% from just prior testing), lowest SpO₂, RDI, body movement, and heart rate. Tachypnea, tachycardia, and abnormal rhythm patterns were also noted. Descriptive statistics, including two-tailed t tests, were used for statistical comparison.

Fifty-five subjects (14 women and 41 men, mean age 39 +/- 14 years, range 19 to 72 years, mean BMI 25 +/- 7.14 kg/m², range 19 to 35 kg/m²) participated in the study. Only one recording was a failure, due to loss of impedance signal during the night, and was subsequently repeated.

Polygraphic monitoring (PSG) identified 20 subjects with an RDI > 5 (mean age = 41.5 +/- 12.4 years; mean BMI = 26.9 +/- 6.2 kg/m²). Comparison of selected parameters of sleep and breathing are presented in Table 2. The results indicated that the EdenTrace II always classified subjects appropriately in the above or below 5 breathing events per hour of sleep group (i.e., 100% sensitivity and specificity in this study for group subdivision). There were some discrepancies with EdenTrace II in the number of respiratory events, particularly when RDI was calculated: the absence of EEG always limits the accuracy of determination of total sleep time (TST). The errors, however, were always limited and as all subjects were classified appropriately in the "normal" or "abnormal" breathing during sleep groups.

In the RDI > 5 group, the largest discrepancy between the two measures was 12 events, all hypopneas, overestimated by the EdenTrace II in one subject. In the RDI < 5 group of 35 subjects, similar results were found. In both groups, the largest discrepancies were in the scoring of central respiratory events, with the maximum discrepancy of 7 central events in one subject classified as obstructive with the EdenTrace II.

A second investigation with the same protocol was performed on 10 children (age 10 to 14) referred for observation of snoring during sleep. In addition to the laboratory investigation, all the children had nonattended, ambulatory home monitoring, five subjects the day before and five the day after. The data from the EdenTrace II were scored blind by two different technicians. The laboratory study demonstrated the largest discrepancy of 14 events (range 7 to 14, mean = 10 +/- 3) - all hypopneas - between the two types of recordings. The EdenTrace II usually missed these hypopneas. There was a maximum difference of 25 events between home and laboratory recordings. The discrepancies were both on the positive and negative side, with six subjects presenting less events at home than on the polygraphic recording (range 14 to 25, mean = 21 +/- 2.5) and four subjects presenting more events at home with a maximum difference of 21 events (range 17 to 21, mean = 19 +/- 1.5). The EdenTrace equipment also was accurate in determining the total number of sleep-disordered respiratory events and appropriately subdividing them into central, mixed, or obstructive apneas and hypopneas, with the greatest margin of error seen with central hypopneas. It was thus determined to be an appropriate tool for the planned NMD clinic population survey.

Analyses performed on the NMD population.

Data were analyzed following the criteria and rules outlined in the "validation study." Descriptive statistics, multiple regression analyses, and analysis of variance were applied to the collected data. Statistical significance was inferred when p was <0.05. In Spearman rank correlations "rho" was required to be at least 0.30 to establish validity since any lower results would account for less than 9% of the variance.

Patients.

The 60 participants included 50 adults and 10 children Table 2. The age of the adults was 20 to 74 years, mean 48.2 +/- 16.6 (SD), and the age of the children was 7 to 17 years, mean 13.2 +/- 3.1 years. Twenty-six (52%) adults and 7 (70%) children were male. Mean BMI of the adults was 25.6 +/- 7.7 kg/m² and of the children was 18.8 +/- 3.9 kg/m². Both means were within the normal range. Obesity, defined in adults as BMI of 30 kg/m² or more and in children based on appropriate age-related tables, was present in 11 adults (22%) and none of the children. One subject had been previously investigated for a breathing disorder during sleep and had received treatment with a dental appliance for obstructive sleep apnea but was no longer using it. None of the others had ever been investigated for SDB. Three patients received oxygen by nasal prongs for temporary relief of shortness of breath during wakefulness. One of the patients was using O₂ therapy during sleep and one used assisted ventilation through a tracheostomy during sleep. These therapeutic decisions had been made based upon daytime wake pulmonary studies. The degree of disability was variable, but only 43% of patients were able to ambulate without much assistance, and 20% only scored a "1" (can climb stairs without support) on the disability scale.

Adults versus children.

We examined our data for significant differences between adults and children in questionnaire responses, pulmonary function, and sleep and breathing measurements. Of 36 analyses, the only significant differences were that adults had a greater BMI (despite the fact that it was, overall, within the considered normal range), more sleep apneas or hypopneas in the supine position (but not in other positions), and a lower minimum O₂ saturation. Because these differences could occur by chance given the size of the children's group, we conclude that the systematic differences between adults and children in the dimensions evaluated in this study are small. For this reason, we present the rest of the data as a single group of observations.

Sleep complaints.

In describing their typical sleep, patients reported a mean sleep onset latency of 30 minutes (range: 0 to 360 min); a mean of 5 awakenings during sleep (range: 0 to 10); and an estimated mean total sleep time of 7.8 hours (range: 3 to 12). Despite these relatively normal sleep parameters, symptoms of disturbed sleep were relatively prevalent. Fatigue (83%) and sleepiness (63%) were reported by the majority of patients. The mean Epworth Sleepiness Scale score was 7.5 (range: 0 to 21); 34 patients (74%) reported a score of 6 or greater (upper limit of normalcy = 10 with a mean score of 6, as indicated by Johns ^[9-11]). Snoring and/or restless legs and reports of leg jerks were present in nearly two-thirds of the population. In addition, 22 patients typically napped during the day for a mean of 45 minutes (range 0 to 120). The frequency of exertional dyspnea (78%) and orthopnea (57%) is noteworthy as an indicator of respiratory insufficiency, and choking (51%) as an indicator of upper airway weakness Table 3.

Breathing during sleep.

The EdenTrace monitor was used during the patients' usual sleep time Table 4. Sleep efficiency was variable, but sleep occurred in the majority of the recording period in each case. The total sleep time derived from the monitor was a mean of 429 +/- 96 minutes.

Fifty patients (83%) had an RDI of 5 or more, 38 (63%) had an RDI of 10 or more, and 25 (42%) had an RDI of 15 or more. Over 90% of the respiratory events were scored as hypopneas; only 1.5% of all events were apneas. Sixty-eight percent of the apneas were central, or nonobstructive-type, apneas Table 4. The majority (61%) of apneas and hypopneas occurred in a position other than supine.

Oxygenation awake and seated was above 92% in all but three patients Table 4. The mean saturation was 94.7%, which is within the normal range at the altitude of 1,500 m. The mean saturation during sleep was 90.8%, which is considered to be the lower limit of normal (>90%) at this altitude. Using these criteria, 14 patients (23%) had a reduced mean saturation. Thirty-five patients (58%) had a minimum O₂ saturation value less than 85% at least once during the night. A minimum saturation below 60% was observed in only one patient. Apneas and hypopneas were closely correlated with O₂ desaturations (Spearman rank correlation between RDI and ODI r = 0.9, p < 0.01).

Spirometry and ventilatory forces awake.

Spirometry and maximum inspiratory and expiratory forces awake seated and supine could only be assessed appropriately in 55 patients. The remaining five patients were too weak to perform the required movements while supine. Table 5 presents results obtained when the group was divided in RDI <5 and >or=to5, and RDI <10 and >or=to10 events per hour of sleep. Overall, compared with predicted values, there were reduced spirometric and ventilatory force values in our patients. The awake findings were consistent with a restrictive impairment due to weak respiratory muscles, but reduction of vital capacity with supine posture, a sign of diaphragmatic weakness, was not consistently recorded.

Thirty-seven subjects were considered to have abnormal spirometric results. All but three subjects showed a typical restrictive defect; the last three individuals had an obstructive defect. The severity of the ventilatory impairment was mild (vital capacity 80, 65% of predicted) in 14 subjects (38%), moderate (64, 50% of predicted) in 13 subjects (35%), and severe (below 50% of predicted) in 10 subjects (27%).

Maximum inspiratory forces were abnormally reduced in 21 (40%) of the subjects and maximum expiratory forces in 22 (41%). If we considered higher SDB cut-off points (RDI of 5, 10, 15, or 20), consistent findings were noted: the mean (or median) values compared with predicted were consistently lower in patients with increased RDI, but none of the comparisons was statistically significant due to the variability of findings between subjects (see below).

Predictors of abnormal breathing during sleep.

First, we used RDIs to evaluate pulmonary function. We found (see Table 5) that none of the considered variables could be used to differentiate the patient groups clearly. We then considered the clinical data for association that could predict an increased RDI or oxygen desaturation during sleep. Table 6 presents associations of interest. Continuous variables were examined with the Spearman rank correlation test. Table 6b indicates the significant correlations.

RDI had no correlation with any of the variables (SaO₂ awake had a rho of -0.29 and we required a minimum rho of -0.30). Mean and minimum oxygen saturation during sleep were correlated with age, obesity, and waking lung function, but the strength of association was again weak. The effect of gender, obesity, sleepiness, and disability rank were examined with the Kruskal-Wallis test. Association with RDI or oxygenation during sleep was not significant in any of these analyses. The between-subjects variability was the explanation for the absence of significance. Similar analyses with symptoms of sleep disturbance, i.e., complaint of daytime sleepiness, fatigue, nocturnal sleep disruption, and snoring, also failed to demonstrate a significant association.

Neuromuscular disorder and obstructive sleep apnea.

Twelve adult patients were identified who had a predominantly obstructive upper airway breathing disorder during sleep. Their NMDs included predominantly hereditary motor-sensory neuropathy, myotonic dystrophy, or other myopathy Table 1. These etiologies, however, were also noted in patients with predominantly "central" sleep-disordered breathing. Patients with predominant obstructive sleep apnea reported frequent and clearly audible snoring, which was confirmed by the EdenTrace recording. All of these patients had an RDI above 10 events per hour of sleep. Compared with the other subjects with predominantly central SDB, these 12 subjects demonstrated specific trends, but none of the results was statistically significant. Compared with the remaining subjects, they had a higher mean BMI (28.0 versus 23.5 kg/m²), a higher mean Epworth Sleepiness Scale score (9.0 versus 7.1), a lower mean oxygen saturation awake (93.9 versus 95.4%), a lower mean O₂ saturation while asleep (89 versus 91.3%), and a lower mean lowest SaO₂ during sleep (76.3 versus 80%).

Monitoring of death during the study.

A patient with advanced amyotrophic lateral sclerosis died during the study recording night. He was clearly disabled (i.e., not ambulatory) and had complained of recent progressive deterioration in his condition (dyspnea, dysphagia, fatigue) in the previous several weeks. On the day of the study, however, awake oxygen saturation was >90% and there was no indication of respiratory distress or acute decompensation. In fact, 10 subjects had worse wake pulmonary function tests than this patient at the time of the study. With sleep onset, oxygen saturation fell to 85%. Over the next 4.5 hours, SpO₂ progressively dropped to <60%, with progressive worsening of hypoventilation as indicated by continuous SaO₂ recording during sleep. Heart rate rose to >100 beats/min, but breathing remained regular in rhythm although increased in rate. During a subsequent arousal, the electrodes which monitored heart rate and breathing effort became disconnected. With SpO₂ consistently below 60%, respiratory pauses of irregular length developed, and after 6 minutes, breathing ceased.

Study limitations.

We performed this investigation in an NMD clinic at a university hospital in the United States, on a population seen during a very specific time period. We did not attempt an in-depth investigation of a specific disorder and, therefore, no single etiologic subgroup was large enough to draw definitive conclusions by subgroup. The clinic was located at an altitude of 1,500 m and this may have added to the degree of severity of the breathing disorders seen in our population. Many western states have large populations of NMD patients who live above this altitude, for whom these findings will be relevant. Finally, our study was performed with an ambulatory monitoring system that cannot fully appreciate the degree of sleep fragmentation (related to breathing disorders or other causes, such as pain and/or inability to change position) and breathing disorders without long hypopneas (>or=to10 seconds). As a result, we may have missed certain patients with mild disordered breathing during sleep.

Neuromuscular disorders and sleep.

The most common complaints reported by our patients were fatigue, exertional dyspnea, and excessive daytime sleepiness. Sleep disruption in this group may result from the inability to change position during sleep, muscle twitches and leg jerks during sleep, and from sleep-disordered breathing. We found a very high frequency of SDB in this population compared with the findings of recent population-based studies ^[15,16] of working adults Table 2. Subjects in general population studies, such as that of Young et al., ^[15] had predominantly obstructive sleep apneas and hypopneas, which were associated with well-known predictors of this syndrome (e.g., male gender, obesity, snoring). Twenty percent (12 patients) of our subjects had the usual symptoms associated with upper airway obstruction during sleep, i.e., loud regular snoring, high BMI, daytime sleepiness, but in our patient population, regardless of the type of apnea, the gender difference was marginal and, overall, RDI was not associated significantly with age, gender, or obesity (see Table 1).

Our survey also failed to show that monitoring of awake patients can predict SDB. There was a relationship between the most severe wake pulmonary function test and the greatest sleep-related O₂ desaturations, but one of our daytime studies, performed hours before the sleep-related death of a subject, did not identify his very severe sleep-related hypoventilation.

The EdenTrace monitor, a simple device that can be used in the home environment and that is easy to interpret, is quite sensitive and provides specific information for detecting SDB. Redline et al., ^[17] for example, compared in-hospital recording with at-home recording and still found a high correlation (r = 0.96) between the RDI found at home and the one obtained in the hospital. There are several other commercially available devices that may perform as well, and the most modern devices even monitor sleep EEG in the home environment. This is particularly important as there are simple, nonsurgical treatment options that can greatly improve quality of life, and may help to prevent premature death. Devices such as bi-level positive airway pressure (BiPAP, Respironics Inc., Murrayville, PA) or nasal intermittent positive pressure ventilation (nasal-IPPV) are very successful treatments. Nasal-IPPV in particular is an effective treatment in patients with severe hypoventilation during sleep. ^[18-23] Bi-level positive airway pressure has received much less attention in the literature, with use reported mostly anecdotally ^[24] and predominantly in the pediatric literature. ^[25] In our own experience, involving mostly adult patients, we find bi-level positive airway pressure a less costly treatment modality than nasal-IPPV. It is easily titrated in a sleep laboratory during one night and, when not used on the ventilation mode, it responds spontaneously to mild inspiratory efforts.

Before monitoring NMD patients for SDB, the physician should question all patients about their sleep and daytime function. "Tiredness" and "fatigue" are common terms used to indicate the presence of sleepiness, and physicians should not automatically assume that these complaints are related solely to the NMD. Frequency of napping, rest periods during the day, and symptoms known to be related to daytime sleepiness (decreased memory, increased errors on the job, decreased intellectual performance, decreased attention span, etc.) should be considered in these patients. Simple pulmonary function tests may indicate the presence of mild problems, but these tests as well as measurement of blood gases must be done, not only on seated subjects, but on subjects awake in the supine position for 15 minutes. Why daytime wake studies cannot accurately predict SDB relates to the physiology of sleep being different from that of wakefulness. The regular occurrence of REM sleep (20% of total sleep time in an adult) and the phenomenon of REM-related muscle atonia means that the intercostal muscles and accessory respiratory muscles will not be active during REM sleep, which will undoubtedly lead to much worse results than those obtained during supine wakefulness. Also during slow-wave sleep (stage 3 to 4, NREM sleep) the ventilatory control system is only dependent of chemical control. Finally, sleep fragmentation and the resulting sleep debt suppress the arousal response, the last defense against respiratory failure during sleep. The absence of the arousal response to worsening hypoventilation during sleep was most probably responsible for the unexpected death during sleep of our patient. Systematic identification of sleep-related respiratory pathology will be the only way to provide adequate treatment to NMD patients.

From the data obtained, one might conclude that clinical criteria alone, particularly those obtained in awake patients, should not be used to exclude particular NMD patients from sleep investigations. Ambulatory studies with simple equipment may be very helpful as these devices are cost-effective and the most recent models, such as the Minisomno (Nellcor, Oakland, CA) even allow home polysomnography, limiting the need for nocturnal sleep laboratory investigations of difficult cases.

Only three subjects in our total sample (5%) had been treated for their breathing disorders at the time of the survey and only one had been treated with specific recognition of an SDB problem. At present, no studies show that early diagnosis and treatment will slow progression of the disease or prevent respiratory failure and other related complications. Preliminary data indicate, however, that treatment of SDB improves the quality of life in these patients. Further outcome studies in this area would be useful.