Published ahead of print on July 11, 2003, doi:10.1164/rccm.200301-092OC
© 2003 American Thoracic Society
Genioglossus Activity in Children with Obstructive Sleep Apnea during Wakefulness and Sleep OnsetDivision of Respiratory Diseases, Department of Medicine, Children's Hospital Boston; and Division of Sleep Medicine, Department of Medicine, Brigham and Women's Hospital, Harvard Medical School, Boston, Massachusetts Correspondence and requests for reprints should be addressed to Eliot S. Katz, M.D., Division of Respiratory Diseases, Mailstop 208, Children's Hospital, Boston 300 Longwood Avenue, Boston, MA 02115. E-mail: eliot.katz{at}tch.harvard.edu
A prominent role for upper airway neuromuscular control mechanisms in the pathophysiology of pediatric obstructive sleep apnea syndrome (OSAS) is suggested by the observation that obstruction does not occur during wakefulness and is infrequently seen during non-REM sleep. Using a custom intraoral surface electrode to record genioglossal activity (genioglossal electromyography [EMGgg]), normalized with a maximal maneuver, we studied 10 children with OSAS and 6 normal control subjects to determine EMGgg activity during (1) wakefulness, (2) the sleep onset period, and (3) stable non-REM sleep. We observed that the EMGgg activity in patients with OSAS compared with control subjects was significantly greater during wakefulness (3.6 ± 1.8 vs. 1.6 ± 1.8% maximum, p < 0.05) and had a greater decline during the early and late sleep onset period (p < 0.05). During stable non-REM sleep, EMGgg remained below the wakeful baseline in all normal control subjects but increased above the baseline in four of the patients with OSAS. We speculate that the increased EMGgg activity during wakefulness represents a reflex-driven neuromuscular compensation for an anatomically compromised airway. Furthermore, the larger decline in EMGgg at sleep onset observed in patients with OSAS is consistent with the relative loss of this reflex. Finally, the return of EMGgg activity above baseline in patients with severe OSAS suggests that some chemical or mechanical compensatory mechanisms remain active during stable non-REM sleep in children.
Key Words: intraoral surface electrode • sleep apnea • genioglossus EMG The upper airway size is determined by the balance between static pharyngeal mechanics, neuromuscular activity, and luminal pressure (1). The phasic inspiratory activation of upper airway muscles results in a net airway dilating force (2). Muscle activation affects both the luminal size and stiffness of the airway, thus determining collapsibility (3). The importance of airway neuromuscular mechanisms in children is suggested by both physiologic and imaging data. Anesthetized children with the obstructive sleep apnea syndrome (OSAS) have narrower and more collapsible airways compared with normal control subjects (4). However, there is considerable overlap in the pharyngeal luminal diameter of awake, sedated children with and without OSAS (5), suggesting a role for neuromuscular compensation. An active airway neuromuscular mechanism is further evidenced by the lack of obstruction during wakefulness, the REM predominance of apnea (6), and the tendency for sedatives to induce apnea. The neuromuscular control mechanisms active in maintaining airway patency in children are not well understood. Genioglossal activity produces forward movement of the tongue, increasing the oropharyngeal airway (7). Although the primary site of obstruction in OSAS is in the velopharyngeal region, activation of the hypoglossal motor nerve decreases airway collapsibility (8). Phasic inspiratory genioglossal activity (genioglossal electromyography [EMGgg]) is observed with both intramuscular and surface electrodes in the majority of adults, during wakefulness and sleep (9, 10). Furthermore, in adults, the waking tonic and phasic EMGgg activity is greater in patients with OSAS than in normal control subjects (9). Similarly, phasic EMGgg was observed from external surface recordings in children with OSAS, but not in control subjects or after an adenotonsillectomy for OSAS (11). This is consistent with neuromuscular compensation for an anatomically compromised airway. Data comparing intraoral surface EMGgg in children with and without OSAS has not been reported. At the wake–sleep transition, there is a greater decline in EMGgg activity in adult patients with OSAS than in control subjects (12), likely indicating loss of compensatory mechanisms. This sleep-dependent decrement in muscle activity leads to a reduction in pharyngeal area in the anatomically predisposed airway, and initiates the cycle of airway obstruction/arousal. In normal adults, EMGgg activity during stable sleep may equal, or actually exceed, that during wakefulness (13, 14). The mechanisms of this increased activity have not been established. Thus, the primary abnormality in OSAS is deficient airway anatomy, and the behavioral-state dependence of neuromuscular control mechanisms account for the temporal occurrence of airway collapse. In children, the state-dependent changes in EMGgg have not been defined in normal control subjects or patients with OSAS. The objectives of the present study were to evaluate EMGgg using an intraoral surface electrode in children with OSAS versus control subjects: (1) during wakefulness, (2) throughout the sleep onset period, and (3) during stable non-REM sleep. The preliminary data from this work has appeared in abstract form (15, 16). We hypothesize that children with OSAS will have evidence of increased waking EMGgg as a compensation for their deficient pharyngeal airway anatomy. This augmented muscle activation will likely be diminished at sleep onset. Whether muscle activity will recover during stable non-REM sleep as a result of mechanical compensatory mechanisms is unclear at this time.
Two groups of children with symptoms of sleep-disordered breathing (n = 11) were recruited from the Pulmonary Clinic at the Children's Hospital Boston: (1) children with a previous overnight polysomnogram demonstrating OSAS, who agreed to repeat polysomnography with an intraoral surface electrode (n = 6) and (2) children who were suspected to have OSAS based on clinical history and physical examination (n = 5), that agreed to include an intraoral surface electrode as part of their initial polysomnogram. Symptomatic patients having an apnea/hypopnea index (AHI) greater than 1 on the experimental night were included in the analysis as patients with OSAS. Two of the patients with OSAS had previously (before our polysomnogram) had an adenotonsillectomy, but none had an adenotonsillectomy between their baseline polysomnogram and the experimental study. A control group of nonsnoring, asymptomatic subjects were recruited from the community (n = 6). All subjects underwent an overnight polysomnogram including continuous measurement of EMGgg using an intraoral surface electrode. Our polysomnographic montage, event definitions, and clinical classification are described in detail in the online supplement. Signed, informed consent was obtained from the parent, and assent from the children. The study was approved by the institutional review board at Children's Hospital Boston.
Intraoral Surface Electrode
Definition of the Sleep Onset Period
For each subject, two distinct reference intervals were evaluated: (1) a 5- to 10-minute period of eyes-closed waking
A schematic diagram of the sleep onset period is depicted in Figure 1A
. The
Protocol For each subject, the mean peak phasic EMGgg was determined supine during (1) 5 to 10 minutes of baseline breathing during eyes-closed wakefulness (  100 breaths), (2) early sleep onset period (5 breaths), (3) late sleep onset period (5 breaths), and (4) Stage 2 sleep (20 breaths). The change in SpO2 and PETCO2 (end tidal [partial] carbon dioxide pressure) between the wakeful baseline and stable Stage 2 sleep was measured for each subject. Data containing arousals were excluded from analysis.
Statistical Analysis
All children recruited as control subjects (n = 6) had no polysomnographic evidence of OSAS. All of the children (n = 6) with previously documented OSAS also demonstrated OSAS on the night of the experimental study with the intraoral surface electrode. Four out of five of the symptomatic children, without a previous overnight polysomnogram who were enrolled in the study, demonstrated OSAS on the night of study. Data from the symptomatic patient without documented OSAS was excluded from further analysis. One child with OSAS could not fall asleep with the intraoral surface electrode in place, but was able to complete the EMGgg measurement during wakefulness. Demographic and polysomnographic data from the 10 patients with OSAS and 6 control subjects in whom usable data was obtained are summarized in Table 1 .
EMGgg Activity during Wakefulness The mean coefficient of variation for the maximal EMGgg activity during repeated measures awake for all subjects was 16.2 ± 6% (range 6–27%). The geometric mean EMGgg activity during resting wakefulness was significantly higher in patients with OSAS versus control subjects (3.6 ± 1.8 vs. 1.6 ± 1.8, p < 0.05) (see Figure 2) . There was, however, some overlap between the two groups. The correlation coefficient between EMGgg and AHI in patients with OSAS was 0.73 (p < 0.05). The correlation coefficient between EMGgg and the age of patients with OSAS was 0.26 (not significant), whereas the correlation between EMGgg and body mass index in patients with OSAS was 0.40 (not significant). Phasic EMGgg activity was present in five patients with OSAS, but was not observed in any control subjects. The mean AHI for patients with OSAS with and without phasic activity was 11.1 ± 8 and 4.7 ± 3 per hour, respectively (p = 0.11). An epoch of raw data demonstrating phasic activity during the sleep onset period is shown in Figure 3 .
EMGgg Activity Measurement Reproducibility The intrasubject variability data from the two patients with OSAS and three control subjects that were studied on two occasions (with different intraoral electrodes) are shown in Figure E3 in the online supplement. Patients were restudied after 41 ± 43 days (range 1–104 days). The mean wakeful EMGgg for these five subjects between the first and second measurement was not significantly different (4.0 ± 3.5 vs. 3.1 ± 2.4, p = 0.66). The intraclass correlation coefficient was 0.91.
EMGgg Activity during the Sleep Onset Period
EMGgg Activity during Stable Non-REM Sleep In all normal subjects, the EMGgg activity during Stage 2 sleep remained significantly below the wakeful level (p < 0.05, see Figure 4 and online supplement Figures E4a and E10). For the patients with OSAS, compared with wakefulness, the mean peak phasic EMGgg activity was not significantly different in the stable non-REM sleep (p = 0.12). As a percentage of the baseline wakeful activity, the patients with OSAS were significantly higher than control subjects during stable non-REM sleep (p < 0.05) (see online supplement Figure E5). Four of the nine patients with OSAS able to sleep with the intraoral surface electrode had Stage 2 EMGgg activity levels that were above the wakeful baseline (see Figures 4 and 6). The number of breaths after the midpoint of the sleep onset period that it took these four patients to surpass the baseline wakefulness EMGgg activity was 8, 16, 20, and 26 breaths, respectively. The mean AHI of the four patients with OSAS with increased Stage 2 EMGgg activity was 11.4 ± 9.5/hour, compared with only 6.5 ± 2.7/hour for the patients with OSAS with Stage 2 EMGgg activity below the wakeful baseline (p = 0.30).
Oxygen Saturation and PETCO2
The major findings of this study are as follows. (1) During wakefulness, the mean peak phasic EMGgg activity as a percentage of maximum is greater in children with OSAS than in control subjects, although there is some overlap between the groups. (2) The mean peak phasic EMGgg activity decreased significantly relative to wakefulness by the late sleep onset period, in both patients with OSAS and normal control subjects. (3) The absolute reduction in EMGgg activity between baseline wakefulness and the early and late sleep onset periods were greater in patients with OSAS than in control subjects. (4) During stable non-REM sleep, the mean peak phasic EMGgg activity remained at low levels in all control subjects, but increased in some patients with OSAS to levels above wakefulness. Phasic inspiratory EMGgg activity is observed with intramuscular electrodes in the majority of adults, during wakefulness and sleep (9). Surface electrodes have also been reported to record phasic EMGgg activity (10), particularly in patients with OSAS, and when augmented with hypercapnia (19) or negative pressure stimuli (20). Intraoral surface electrodes have been used extensively in the adult literature (10, 17, 21) and sparingly in infants (22, 23). Using external surface electrodes, phasic inspiratory EMGgg and increased EMGgg during respiratory arousal was reported in children with OSAS, but not in normal control subjects (11, 24). However, submental electrode placement has the disadvantage of recording from the geniohyoid, mylohyoid, and platysma, as well as the genioglossus, thus complicating interpretation of the signal. Furthermore, the EMGgg in these studies was not expressed as a percentage of maximal, and therefore the relative EMGgg activity between children with OSAS and control subjects could not be determined. Using an intraoral surface electrode in the present study, we also observed phasic activity in many patients with OSAS, but not in normal control subjects. There was also a trend toward increasing AHI in children with OSAS with phasic activity, but this did not reach statistical significance. This supports the hypothesis that increased genioglossal activity may be an important compensatory mechanism during wakefulness and sleep in children with OSAS. Consistent with our findings, adults with OSAS have greater EMGgg activity during wakefulness than normal control subjects (9, 25). Furthermore, at the wake–sleep transition, there is also a greater decline in EMGgg in adults with OSAS than in control subjects (12), likely indicating a loss of compensatory mechanisms (18). It is hypothesized that during the wake–sleep transition in OSAS, the pharyngeal muscle activation resulting from mechanoreceptors responding to negative luminal pressure is diminished (26, 27). This sleep-dependent decrement in muscle activity leads to a reduction in pharyngeal area and, in the anatomically predisposed airway, initiates the cycle of airway obstruction/arousal. Our data in children with OSAS also indicated a greater decline in the EMGgg activity during the sleep-onset period. On the other hand, normal adults and children are less reliant on mechanoreceptor input to the upper airway and therefore have smaller changes in EMGgg activity at sleep onset. In contrast to the present findings in children, previous reports in normal adults have demonstrated an increase in EMGgg during stable NREM sleep compared to wakefulness (18). This is suggestive of a reflex activation of airway muscles likely resulting from an increase in airway resistance and therefore transmural pressure. We observed that, in normal children, the EMGgg remained at levels well below wakefulness during stable NREM sleep. It is likely that the greater stability of the normal pediatric airway, as has been demonstrated in both anesthetized (4) and spontaneously sleeping children (28), accounts for the ability of normal children to sustain unobstructed breathing after sleep onset in the absence of augmented airway muscle activity. Thus, neuromuscular compensatory mechanisms during stable NREM sleep are less active or not required in normal children compared with normal adults. By contrast, approximately half of the children with OSAS demonstrated an elevation in EMGgg during stable NREM sleep compared to wakefulness. Those OSAS patients with increased EMGgg activity during stable NREM trended towards a higher AHI than the patients with OSAS in whom the EMGgg decreased from wakefulness to NREM sleep. We speculate that children with more severe OSAS generate larger negative inspiratory intraluminal pressures sufficient to stimulate the airway musculature. We further speculate that patients with a highly collapsible airway may be unable to maintain pharyngeal patency despite reflex activation of dilator muscles, thus leading to pharyngeal collapse and sleep apnea. The stimulus to augmented EMGgg during stable sleep is unclear, but may arise from chemoreceptors and/or mechanoreceptors (29). During wakefulness, both hypoxia and hypercapnia lead to increased activation of upper airway dilator muscles, although this responsiveness seems to be reduced or absent during sleep (30–32). Similarly, negative airway pressure is a potent stimulus to pharyngeal muscle activation awake and almost certainly accounts for the increased muscle activity observed in both adults and children with sleep apnea (33, 34). However, this responsiveness is also substantially reduced during sleep. Thus, the stimulus leading to increased EMGgg in some subjects with OSAS in this study, and normal adults in other studies, remains unclear. We speculate that there is some residual responsiveness to negative pressure during sleep that allows the genioglossus to activate as resistance increases. However, this has not been carefully studied to date.
There are a number of limitations of the present study. First, the between-group comparison of the EMGgg activity requires that each individual be normalized to a reproducible EMGgg value. The use of a nonrespiratory maneuver, forced protrusion against the alveolar ridge, to ascertain maximal EMGgg activity could be criticized. However, this procedure has been used with success in adults (9) and was quite reproducible in children tested with different intraoral electrodes on separate days. Other voluntary maneuvers, including maximal inspiratory efforts and lingual force measurements, are too variable in children to be suitable for between-group comparisons. In addition, chemostimulation (carbon dioxide) does not produce muscle activation in the maximal range. Thus, we believe the methodology used to be the best approach. Second, we recognize that the relationship between EMG activity and force is nonlinear and complex. However, EMGgg activity is widely regarded as a reliable measure of neural input, and thus our conclusions that neuromuscular compensation is more robust in OSAS are plausible. Third, choosing an electroencephalographic definition of sleep onset by necessity relies on arbitrary criteria (35). We chose to evaluate sleep onset as a gradual period rather than a single point of time. This is consistent with incremental changes at sleep onset in other physiologic systems such as blood pressure, evoked potentials, and sensory thresholds. Our approach consistently placed the midpoint of the sleep onset period visibly in the center of the In summary, we observed an increase in the EMGgg activity during wakefulness in children with OSAS versus control subjects. We speculate that this represents neuromuscular compensation mediated via mechanoreceptors in an anatomically compromised airway. During the process of sleep onset, this reflex mechanism is gradually diminished in both patients with apnea and control subjects, having a greater effect on the patients with apnea. Insofar as normal children have a relatively stable upper airway in the absence of muscle activation, there is no increase in EMGgg as stable non-REM sleep progresses. By contrast, the EMGgg in some children with severe OSAS rebounds after sleep onset, consistent with the reactivation of reflex mechanisms in the setting of rising airflow resistance. We speculate that a progressive increase in negative intrapharyngeal pressure occurs during the sleep onset period that is sufficient to activate pharyngeal dilator muscles. Further studies will be required to evaluate the mechanism of the increase in EMGgg during sleep in some children with OSAS.
Supported in part by grant MO1 RR02172 to the Children's Hospital, Boston General Clinical Research Center by the National Center for Research Resources, NIH. E.S.K and D.P.W. were supported by NIH/NHLBI 1 P50 HL60292. This article has an online supplement, which is accessible from this issue's table of contents online at www.atsjournals.org Conflict of Interest Statement: E.S.K. has no declared conflict of interest; D.P.W. has participated as a speaker in scientific meetings or courses organized and financed by various sleep technology companies (Itimar, Respironics) and received research grants from Respironics Inc., Itamar Medical LLC, Alfred E. Mann Foundation, and Widemed. Received in original form January 21, 2003; accepted in final form June 3, 2003
This article has been cited by other articles:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
METHODS
or 
according to the technique adapted from Worsnop and coworkers (
0.5). (B) Eight-year-old with obstructive sleep apnea syndrome (OSAS). Plot of the breath-by-breath 
