INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Pharyngeal airway closure during sleep in patients with obstructive sleep apnea (OSA) is generally felt to be due to the state-related loss of motor output to skeletal muscles surrounding the pharynx that dilate and stiffen this potentially collapsible upper-airway segment (1, 2). However, it has long been speculated that upper-airway closure during sleep may be due to activation of muscles that constrict the pharyngeal airway, notably the tongue retractors (hyoglossus and styloglossus) and pharyngeal constrictors (PC) (3). The superior, middle, and inferior PC are saillike muscles forming the lateral and posterior walls of the pharyngeal airway. The constricting effect of these muscles is supported by their activation during the pharyngeal-closure phase of swallowing (6). On the basis of fiberoptic observations of the upper airway during sleep in OSA subjects, Borowiecki and colleagues (4) suggested that pharyngeal airway closure during sleep may result from activation of the PC muscles. Sanders and coworkers (5) speculated that activation of upper-airway muscles with phasic expiratory activity might account for the increased expiratory airway resistance in the breaths preceding an apneic episode in OSA subjects. Badr and associates (3) proposed that pharyngeal airway closure during an induced central apnea in non-rapid-eye-movement (NREM) sleep in normal and OSA subjects might be an active phenomenon, resulting from activation of the PC muscles. Although the PC muscles can exhibit phasic expiratory activity in spontaneously breathing animals and awake normal adult humans (8, 11, 12), Guilleminault and colleagues (14) recorded PC electromyographic (EMG) activity during apneic episodes in OSA patients, and reported a decrease or complete disappearance of PC-muscle activity during the obstructive apnea and PC muscle activation with airway reopening, a pattern similar to that of upper-airway dilating muscles.

The purpose of the present study was to further characterize the respiratory-related activity of the superior pharyngeal constrictor (SPC) muscle in OSA subjects during quiet breathing in wakefulness and during spontaneous apneas in NREM sleep. The study also examined whether or not it is possible to induce upper-airway closure during sleep in OSA subjects in the absence of SPC-muscle activation. The results indicate that the SPC muscle in OSA subjects can exhibit respiratory-related activity during wakefulness and has an activation pattern similar to that of pharyngeal dilator muscles during spontaneous and induced apneas in NREM sleep. SPC muscle activation was not necessary to induce upper-airway closure in NREM sleep.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Experiments were performed with 18 subjects (12 males and six females) with untreated OSA. The subjects' mean age was 42 ± 11 yr (mean ± SD), their mean weight was 130 ± 20 kg, and their mean height was 171 ± 4 cm. The diagnosis of OSA was established on routine polysomnography. All subjects had moderate to severe OSA. The mean apnea/hypopnea index (AHI) was 70 ± 25 events/h, with a range of 38 to 107 events/h. All subjects presented with symptoms of excessive daytime hypersomnolence. The protocol used in the study was approved by the Institutional Review Board of The University of Texas Medical Branch at Galveston, and signed informed consent was obtained from all subjects.

A nighttime polysomnogram was recorded from each subject, using previously described techniques (15, 16). Recordings were obtained with the subjects in the supine position. The following parameters were recorded on the polygraph (Gould, Cleveland, OH) and magnetic tape (Neurodata, New York, NY): C3A2 and O2A1 electroencephalogram (EEG), bilateral electrooculograms (EOGs), esophageal pressure, arterial oxygen saturation (Sa_O2) (Ohmeda, Louisville, CO), presence or absence of tracheal breath sounds, and the EMG of the SPC and genioglossus (GG) or alae nasi (AN) muscles. The EMG signals were also continuously monitored on an oscilloscope.

The AN-muscle EMG was recorded with surface electrodes (17). A pair of 38-gauge hooked-wire electrodes (Belden, Chicago, IL) were implanted into the GG and SPC muscles. The GG electrodes were implanted through the peroral approach of Basmajian and Stecko (18). Hooked-wire electrodes were implanted into the SPC muscle via a nasopharyngoscope (11, 15, 16). This transnasal technique was used instead of the peroral technique of Hairston and Sauerland (8) because it was not possible to visualize the posterior oropharyngeal wall through the mouth in most of the OSA subjects. Both nasal passages were sprayed with the long-acting nasal decongestant oxymetazoline and with 2% lidocaine. A nasopharyngoscope (Pentax, Orangeburg, NY) was passed transnasally into the nasopharynx. The electrode wires (Belden Wire and Cable, Chicago, IL) were threaded into a retractable 21-gauge needle catheter (Microvasive, Watertown, MA) which, after gas sterilization, was passed through the fiberoptic scope's suction channel. The electrode wires were implanted under direct vision near the midline of the dorsal wall of the pharyngeal airway, at the level of the rim of the soft palate, by advancing the needle through the mucosa into the underlying muscle (Figure 1). The needle catheter and fiberoptic scope were then withdrawn, leaving the electrode wires exiting the naris. Several minutes after withdrawal of the fiberoptic scope, the subjects were unaware of the presence of the wires.

View larger version (42K): [in this window] [in a new window]   Figure 1.   Fiberoptic technique used to implant hooked-wire electrodes into the superior pharyngeal constrictor muscle.

Correct placement of the EMG electrodes was confirmed at the beginning and end of the experiments by having the subjects perform voluntary maneuvers. The criteria for SPC-muscle recordings were activation during swallows and repetitive "pa" sounds. The plosive portion of the "pa" sound is associated with velopharyngeal closure and SPC muscle activation (11). The criterion for GG-muscle recordings was activation with tongue protrusion against the upper incisors. The criterion for AN-muscle recordings was activation on a sniff. The EMG signals were amplified, filtered between 30 and 3,000 Hz, rectified, and passed through a Paynter filter with a 100 ms time constant to obtain a moving average signal.

Following placement of the EMG electrodes, a size 5-French catheter with a pressure transducer at its distal end (Millar Instruments, Houston, TX) was advanced through the opposite nasal passage into the oropharynx or esophagus. Output from a microphone taped just rostral to the suprasternal notch was used to determine the presence or absence of tracheal breath sounds.

Recordings were obtained with subjects in the supine position under several different conditions: quiet breathing in wakefulness prior to sleep onset, obstructive apneic episodes during NREM sleep, quiet breathing during NREM and REM sleep on nasal continuous positive airway pressure (CPAP), and airway occlusions induced by sudden withdrawal of nasal CPAP during NREM sleep. A minimum of 10 min of quiet breathing in wakefulness and a minimum of 30 apneic episodes in NREM sleep were recorded in each subject. A mean of 6.5 ± 4.7 induced occlusions were elicited (range: 1 to 19 induced occlusions). After recordings were obtained during sleep to determine EMG activity in association with spontaneous apneic episodes, the subjects were given nasal CPAP (Respironics, Murrysville, PA) and allowed to go back to sleep. During this phase of the study, nose-mask pressure (Statham, Hato Rey, PR) was recorded along with the other parameters listed earlier. Mask and esophageal pressures were calibrated in cm H₂O, using a manometer. The nasal CPAP was set at a pressure that eliminated upper-airway closures during sleep and restored a stable sleep pattern. Intermittently during expiration, nose-mask pressure was abruptly lowered to atmospheric pressure. The presence of upper-airway occlusion was based on the absence of respiratory fluctuations in mask pressure despite the presence of respiratory efforts. Nose-mask pressure was restored to the therapeutic level following spontaneous reopening of the airway associated with arousal, or when the airway was still occluded prior to arousal.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

PC-muscle Activation in Wakefulness

During quiet breathing in wakefulness, phasic expiratory SPC muscle activity was consistently present in six subjects and intermittently present in 12 subjects, particularly following a swallow. Phasic activation began in late inspiration and continued with a plateaulike discharge pattern throughout expiration. No differences were apparent to distinguish subjects in whom respiratory-related SPC-muscle activity was usually present in wakefulness from those who exhibited transient respiratory-related activity.

PC-muscle Activation During Spontaneous Apneic Episodes

The two most common patterns of SPC-muscle activation during spontaneous obstructive apneic episodes were: (1) activation upon airway reopening, with or without waning phasic or tonic activation in the first half of the ensuing apnea (n = 13); and (2) absence of SPC-muscle activation during arousals or apneas (n = 5). In the 13 subjects with SPC-muscle activation upon airway reopening, both the presence and absence of waning activation in the first halves of apneas were usually evident within a given subject. The absence of SPC-muscle activation during arousals or apneas was a uniform finding in five subjects. This pattern was not evident on any of the apneas in the other 13 subjects. There were no apparent differences between the subjects to explain the presence of a particular activation pattern. At the end of each study, SPC-muscle activation was present during a voluntary swallow in all subjects, including the five subjects without SPC-muscle activation associated with spontaneous obstructive apneas.

Figures 2-4 show SPC-muscle activation patterns during repetitive obstructive apneas in three of the subjects. In Figure 2, SPC-muscle activation occurs with apnea termination coincident with GG-muscle activation. Both the SPC and GG muscles remain essentially inactive during the actual apneic episode. Figure 3 compares SPC- and AN-muscle activation during an obstructive apnea. Phasic AN-muscle activity appears in the latter part of the apnea when SPC-muscle activity is absent. SPC-muscle activation occurs concurrently with airway reopening, as indicated by the less negative esophageal pressure on inspiration. The bursts of SPC-muscle activation occur during inspiration and are nearly coincident with phasic AN-muscle activation. Tonic SPC-muscle activity gradually disappears during the first half of the apnea. In Figure 4, SPC-muscle activation occurs at apnea termination, followed by a progressively waning phasic expiratory activity during the first half of the subsequent apnea. Swallows are evident during the two arousal periods in this particular example, as evidenced by the large SPC-muscle discharge followed by the sharp positive deflection in esophageal pressure. As shown in Figures 2-4, SPC-muscle activity was uniformly absent in the latter half of the apnea. This contrasts with the activation pattern of the GG and AN muscles, which frequently showed a progressive increase in phasic or tonic activity in the latter half of the apneic episode.

View larger version (16K): [in this window] [in a new window]   Figure 2.   Superior pharyngeal constrictor and genioglossus electromyographic activity during obstructive apneas in an OSA subject. Bursts of pharyngeal constrictor activity are associated with airway reopening and are coincident with genioglossus activation.

View larger version (27K): [in this window] [in a new window]   Figure 3.   Superior pharyngeal constrictor and alae nasi activity during an obstructive apnea in an OSA subject. Sudden pharyngeal constrictor activation during inspiration at the end of the apnea is associated with a less negative esophageal pressure.

View larger version (28K): [in this window] [in a new window]   Figure 4.   Superior pharyngeal constrictor and genioglossus activity during an obstructive apnea in an OSA subject. A waning phasic expiratory pharyngeal constrictor activity is present in the initial part of the apnea. Examination of the EMG signal on an oscilloscope revealed that tonic pharyngeal constrictor activity was absent in the latter part of the apneic period.

PC-muscle Activation During Quiet Breathing in NREM and REM Sleep during Nasal CPAP

At nasal CPAP levels that eliminated apneic episodes and restored a normal sleep pattern, the PC muscle was electrically silent in NREM sleep in all but two subjects, who had persistent phasic expiratory activity. In all subjects in REM sleep during nasal CPAP, the SPC muscle exhibited transient, sporadic, non-respiratory-related bursts of activity similar to the activation patterns described in other upper-airway muscles (15, 16).

PC-muscle Activation During Induced Apneas

Regardless of the presence or absence of SPC-muscle activity during NREM sleep during nasal CPAP, upper-airway closure was present on the first respiratory effort following sudden withdrawal of nasal CPAP. In the 16 subjects in whom SPC-muscle activity was absent during nasal CPAP in NREM sleep, no activation was present during the induced apnea (Figures 5 and 6). In the two subjects in whom phasic expiratory SPC-muscle activity was present during nasal CPAP in NREM sleep, the level of phasic activity either did not change or was reduced during the induced apnea (Figure 7).

View larger version (28K): [in this window] [in a new window]   Figure 5.   Superior pharyngeal constrictor activity during an induced occlusion in an OSA subject. Sudden withdrawal of nasal CPAP returned mask pressure to atmospheric pressure. The absence of respiratory fluctuations in mask pressure and the increase in esophageal pressure indicate the presence of airway occlusion. Airway reopening, indicated by the appearance of rapid fluctuations in mask pressure just prior restoration of nasal CPAP, is associated with pharyngeal constrictor activation.

View larger version (25K): [in this window] [in a new window]   Figure 6.   Superior pharyngeal constrictor activity during an induced occlusion in an OSA subject. Prior to restoration of nasal CPAP, the airway remained closed despite phasic genioglossus activation. Airway reopening following restoration of nasal CPAP is associated with pharyngeal constrictor activation on inspiration. The burst of pharyngeal constrictor activity upon withdrawal of nasal CPAP in this example was very unusual.

View larger version (25K): [in this window] [in a new window]   Figure 7.   Superior pharyngeal constrictor and genioglossus activity during an induced occlusion in an OSA subject. The pharyngeal constrictor and genioglossus continued to have phasic expiratory activity on nasal CPAP. Sudden return of mask pressure to atmospheric pressure induced an airway occlusion without a change in EMG activity.

Restoration of nasal CPAP prior to arousal reopened the airway and was not associated with SPC-muscle activation. SPC-muscle activation upon arousal prior to restoration of nasal CPAP was associated with airway reopening (Figure 5). Arousal immediately following restoration of nasal CPAP and reopening of the airway was also associated with SPC-muscle activation (Figure 6). SPC-muscle activation upon arousal and airway reopening occurred on inspiration in six subjects and on expiration in seven subjects.

In almost all subjects, phasic GG- and AN-muscle activity was absent during nasal CPAP. As shown in Figure 6, phasic GG- and AN-muscle activity could appear during the latter part of the induced occlusion prior to arousal. Arousal prior to or immediately after restoration of nasal CPAP was associated with large increases in phasic inspiratory GG- and AN-muscle activity.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

This EMG study investigated respiratory-related SPC-muscle activity in OSA subjects during wakefulness and sleep. During quiet breathing in wakefulness, the SPC muscle frequently exhibited respiratory-related activity, especially following a swallow. During spontaneous apneas in NREM sleep, the most common EMG pattern was SPC-muscle activation upon reopening of the airway on arousal, with or without a progressive decrement in phasic and/or tonic activity in the first half of the subsequent apnea. With elimination of spontaneous apneas through application of nasal CPAP, SPC-muscle activity was generally absent in NREM sleep and occurred in sporadic, transient, non-respiratory-related bursts in REM sleep. Sudden withdrawal of nasal CPAP induced upper-airway obstruction in the absence of SPC-muscle activation. Spontaneous reopening of the airway on arousal following the induced occlusion was frequently associated with SPC-muscle activation during inspiration.

Quiet breathing in wakefulness was recorded only prior to sleep onset, and was not systematically studied. SPC-muscle activity during wakefulness was recorded at the beginning of the experiment, when muscle irritation from electrode insertion was at its peak. Given these caveats, it was our impression that phasic SPC-muscle activation in expiration during quiet breathing was more commonly observed in the OSA subjects than in a group of normal subjects previously studied in our laboratory (11). Phasic expiratory activity was consistently present in six OSA subjects during quiet breathing in wakefulness. In contrast, only sporadic, transient respiratory-related SPC-muscle activity was observed during quiet breathing in awake normal adults (11).

The pattern of SPC-muscle activation following spontaneous apneas in NREM sleep was very similar to that of the GG and AN muscles, two upper-airway-dilating muscles. Like the activity of the GG and AN muscles, SPC-muscle activity exhibited large bursts with airway reopening in the majority of subjects. Similar findings have been reported by Guilleminault and associates (14). During quiet breathing in wakefulness, the GG and AN muscles exhibited phasic activity during inspiration, whereas the SPC muscle exhibited phasic expiratory activity. Phasic expiratory SPC-muscle activity was also evident in some subjects during the initial portion of the obstructive apnea. However, the bursts of SPC-muscle activation associated with airway reopening at the termination of a spontaneous apnea usually occurred during inspiration, coincident with activation of the upper-airway-dilating muscles. In addition, activation of the SPC-muscle upon arousal during induced occlusion often occurred during inspiration. Thus, the SPC muscle could exhibit phasic activation on inspiration or expiration under different conditions in the same subject. Previous studies support this finding. Grélot and colleagues (13) recorded the discharge of motor axons supplying the PC muscles in decerebrate, paralyzed, artificially ventilated cats. Most units fired only in expiration and exhibited a steady, a decreasing, or a late-augmenting discharge pattern; however, inspiratory units with a steady, late-augmenting, or tonic discharge pattern were also present. Kawasaki and colleagues (9, 10) have reported phasic PC-muscle activation in inspiration or expiration in dogs. In awake normal adult humans, PC-muscle activation has been noted during the inspiratory phase of voluntary cough, and biphasic PC-muscle activity can be present under hypercapnic conditions (11). Additionally, Guilleminault and coworkers (14) found phasic inspiratory PC activation upon airway reopening at apnea termination, resembling our findings in OSA subjects.

In 16 of the 18 OSA subjects, SPC-muscle activity in NREM sleep was absent during nasal CPAP, and the muscle was not activated during the induced apnea until arousal and reopening of the airway. The ability to induce upper-airway closure while the SPC muscle is electrically silent suggests that SPC-muscle activation is not necessary to induce obstructive apnea. However, no attempt was made in this study to determine the site of pharyngeal airway closure during obstructive apnea. Previous investigators have shown that the airway closure in OSA subjects during sleep can occur in different segments of the pharyngeal airway (19, 20). The induced apneas in our subjects may have been due to airway closure in the hypopharynx, an area that would be affected by activation of the middle and inferior PC muscles but not the SPC muscle. However, it is likely that most of our subjects' airways were closing at the level of the soft palate, since the aforementioned studies indicate that pharyngeal airway closure occurs at the level of the soft palate in the majority of subjects. Although the middle and inferior PC muscles were not recorded in our study, Guilleminault and coworkers (14) found very similar patterns of activation among the three PC muscles during spontaneous apneas in OSA subjects. In addition, simultaneous recordings of the three PC muscles in normal adult humans under a variety of circumstances reveal a remarkably uniform pattern of activation (11). On the basis of these latter two studies and the EMG recordings of the SPC muscle in the current study, it is probable that airway closure in the lower pharynx in OSA subjects is also not due to PC-muscle activation. The uniform ability to induce upper-airway closure during sleep without PC-muscle activation is explained by the work of previous investigators, who found a positive pharyngeal closing pressure in most OSA subjects (i.e., in the absence of upper-airway muscle activation, it is necessary to have a positive intraluminal pressure in most OSA subjects in order to maintain pharyngeal airway patency) (21, 22).

Previous studies have speculated that upper-airway closure during sleep may be an active phenomenon due to the contraction of the PC muscles (3). However, our EMG recordings offer circumstantial evidence that SPC-muscle activation is not necessary to cause apneas, as follows: SPC-muscle activity decreases with the onset of obstructive apnea, airway closure can be induced without SPC-muscle activation, airway reopening following spontaneous and induced apneas is associated with SPC-muscle activation, and SPC-muscle activation can occur on inspiration.

Although it is well accepted that SPC-muscle activation promotes pharyngeal closure during a swallow, the current EMG data do not allow any conclusions about the mechanical effects of respiratory-related SPC-muscle activation. One can speculate that the electrical activity of the SPC muscle during respiration has no mechanical corollary. Alternatively, SPC-muscle activity may increase elastance with or without an increase or decrease in upper-airway caliber. The latter speculation may help explain the frequently observed respiratory-related SPC-muscle activation that appears transiently after a swallow in wakefulness and sleep. It is possible that respiratory-related SPC-muscle activation when the airway is closed or very narrow may help restore airway patency by stiffening and/or dilating the airway. This proposed functional duality of the SPC muscle as a constrictor during swallows that also dilates or stiffens the airway would not be unique to the SPC muscle. Although internal intercostal muscles generally promote exhalation, their contraction at very low lung volumes facilitates inspiration (23).

In summary, our EMG study in OSA subjects during wakefulness and sleep reveals that the SPC muscle, although normally exhibiting phasic expiratory activity, can have phasic activation on inspiration, particularly during conditions associated with pharyngeal airway reopening. The results in OSA subjects indicate that SPC-muscle activation is similar to that of pharyngeal dilator muscles during spontaneous and induced apneas, and is not necessary to induce upper-airway closure during NREM sleep in OSA subjects.