INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Keywords: genioglossus; sleep apnea; pharyngeal muscles

Obstructive sleep apnea (OSA) is a common disorder affecting 2 to 4% of the middle-aged population (1). This disorder is characterized by repetitive collapse of the pharyngeal airway during sleep (2, 3) and is associated with important adverse consequences for afflicted individuals (4). Numerous studies have demonstrated that the pharyngeal airway of the patients with apnea is anatomically small when compared with that of control subjects, and is thus potentially more vulnerable to collapse (9).

There is substantial evidence in both animals and humans that upper airway dilator muscles play an important role in maintaining airway patency (13). Many of the pharyngeal dilator muscles are known to demonstrate inspiratory phasic activity, the onset of which precedes diaphragmatic activity, thus "preparing" the pharyngeal airway for the development of negative pressure during inspiration. We have previously shown in patients with OSA during wakefulness that there is augmented activity of the genioglossus (GG) muscle as well as other pharyngeal dilator muscles when compared with healthy control subjects (14). This activity is thought to represent a neuromuscular compensatory mechanism for an anatomically small and more collapsible pharyngeal airway. This augmented upper airway dilator muscle activity is lost at sleep onset and is associated with pharyngeal collapse (15). Thus the mechanisms controlling pharyngeal muscle activation are important in understanding disease pathogenesis.

The activity of the pharyngeal dilator muscles is influenced by numerous variables, including blood gases (Pa_O2 and Pa_CO2), sleep-wake state, gender-specific hormones, blood pressure, temperature, lung inflation, pharyngeal airflow, and intrapharyngeal negative pressure (16). However, most data suggest that intrapharyngeal pressure is the primary stimulus to phasic pharyngeal dilator muscle activation. First, it is well known that the application of negative pressure to the pharyngeal airway in animals and in humans leads to a substantial increase in the activity of the genioglossus as well as other upper airway muscles (21, 22). The time course of this response (maximal response within 200 ms) suggests that it is a neural reflex. Second, we have recently shown that peak phasic GG activity correlates closely with the peak negative epiglottic pressure generated during inspiratory resistive loading in healthy control subjects (23). Finally, using an iron-lung model of passive ventilation, data from our laboratory have demonstrated that during inspiration, there is an extremely tight correlation between epiglottic negative pressure (Pepi) and GG electromyogram (GGEMG) (24). The responsiveness of the GGEMG (slope of GGEMG/Pepi) was found to be remarkably constant in a given individual over a range of epiglottic pressures and was not independently influenced by changes in PO₂, PCO₂, or airflow (Malhotra, unpublished observation).

The mechanisms driving the increased genioglossal muscle activation seen in patients with OSA during wakefulness has not been carefully investigated. We hypothesized that the relationship between GG activation and pharyngeal negative pressure would be augmented in the patient with apnea, representing a neural plasticity evolved over years of reflex muscle activation. In order to test this hypothesis we measured the sensitivity of inspiratory GG activation to pharyngeal pressure changes under a wide range of breathing conditions, including basal breathing, increased negative epiglottic pressure generation (iron-lung ventilation and inspiratory resistive loading), and reduced negative epiglottic pressure generation (heliox breathing). The advantages of using the iron-lung was twofold. First, as we have previously shown, it allows us to substantially attenuate spontaneous ventilation. Thus, during relatively "passive" ventilation we can evaluate the relation between genioglossal muscle activation and epiglottic pressure changes across inspiration with a minimum of premotor input. Second, the iron-lung allows us to substantially increase the range of negative epiglottic pressures over which we can evaluate this relationship. Thus, we could compare the slope of this relationship (Pepi versus GGEMG) in patients with OSA versus control subjects.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects

Ten men with moderate to severe Obstructive Sleep Apnea Syndrome (AHI  25/h of sleep) were studied. These participants were recruited from the sleep laboratory of the Brigham and Women's Hospital. Control subjects were 12 healthy male volunteers who had no sleep complaints. All control subjects had a body mass index < 25 kg/m² and had no history of snoring. Informed consent was obtained from each participant, with the protocol having the prior approval of the Human Subjects Committee of the Brigham and Women's Hospital.

Equipment and Techniques

All studies were performed during wakefulness in the supine posture. Subjects lay within a negative pressure ventilator (Series J; Iron Lung, Emerson, MA). Inspiratory flow was determined with a calibrated pneumotachometer (Fleish, Inc., Lausanne, Switzerland) and differential pressure transducer (Validyne Corp., Northridge, CA). For flow measurements during Heliox (80% helium/20% oxygen) conditions a correction quotient of 1.09 was used. This value is based on published gas viscosity correction factors making the assumption that flow through the pneumotachometer is largely laminar (25). The standard techniques of our laboratory were used to measure end-tidal CO₂ (PET_co2), mask leak, mask pressure (Pmask), choanal pressure (Pcho), epiglottic pressure (Pepi), and intramuscular genioglossus electromyography (GGEMG, as a percent of maximum activity) (14, 23).

Passive ventilation ("Iron lung").Subjects were studied while supine with the head outside and the body within a negative pressure ventilator. The ventilator was switched on only for specific parts of the experiment. The iron lung could be adjusted to achieve the desired upper airway pressure and breathing frequency, so that passivity (or completely synchronous ventilation) was achieved. The iron lung excursions required to achieve passivity varied between subjects. All subjects required some initial coaching to enable passive mechanical ventilation. This involved asking the subjects to remain completely relaxed while the investigators provided feedback on a breath-by-breath basis to achieve consistent timing and shape of the pressure and flow traces. Recordings were stopped when there was departure from this passive pattern until adequate passivity could be achieved, or the experiment was terminated. Measurements were recorded only during steady-state conditions (24).

Heliox administration.Heliox was used as a condition in which pressure and flow would change in opposite directions (pressure decreases and flow increases). A 50-liter meteorological balloon was attached to the inspiratory line of the breathing apparatus connected by a three-way stopcock. When breathing room air, the device was open to atmosphere. While breathing Heliox, the valve was open to the 50-liter balloon, which had been previously filled with a mixture of helium (80%) and oxygen (20%). After each change in inspired gases, 3 to 5 min passed before data collection to allow full equilibration.

Inspiratory resistive loading (IRL).IRL leads to an increase in pharyngeal negative pressure, which was generated by the subject/ patient (rather than passively in the iron lung). Resistance was added to inspiration using a specially designed variable resistance device. Inspiration could be loaded to any desired level by varying the effective caliber of the inspiratory pathway (26). In this experiment, two levels of load were used (10 and 25 cm H₂O/L/s). Each was applied for three consecutive breaths and was repeated at least three times (nine breaths per load).

Protocol

After obtaining informed consent, the equipment was attached and calibrations were performed. In random order, each subject was recorded under a variety of conditions: (1) during spontaneous basal breathing (air and heliox), (2) during passive (iron lung) ventilation, and (3) during inspiratory resistive loading. Carbon dioxide was added to the inspiratory line during iron-lung ventilation to maintain eucapnia.

Data Recording and Analyses

All signals were recorded on a 16-channel Grass model 78 polygraph (Grass Instruments, Astro-Med, Inc., West Warwick, RI). Certain signals (GGEMG MTA, airway pressures, inspiratory flow, ET_CO2) were also recorded onto a computer using signal-averaging software (Spike 2; Cambridge Electronic Design, Ltd, Cambridge, UK).

For each condition (except IRL), one buffered breath was generated by signal averaging all breaths in that particular condition. During IRL, each of the three loaded breaths was signal-averaged. For each buffered breath the following variables were determined: peak negative pressure (at the levels of choanae and epiglottis), peak flow, tonic GGEMG, and peak phasic GGEMG (peak activation during inspiration). Phasic EMG was defined as the difference between peak phasic and tonic EMG. Pharyngeal resistance (Rpha) and supraglottic resistance (Rsg) were calculated at peak inspiratory flow. Data were analyzed both within breaths to generate regression relationships between epiglottic pressure and GGEMG in a given condition, as well as between breaths by comparing the mean nadir epiglottic pressure in each condition with mean peak GGEMG.

Statistical Analyses

All statistical analyses were performed with commercially available software (SigmaStat + Sigmaplot, SPSS, Chicago, IL). Two-tailed independent t tests were performed to compare values between apneic patients and control subjects (or nonparametric methods when appropriate). Alpha was set at 0.05. ANOVA for repeated measures (or ANOVA on ranks when appropriate) was used to compare slopes and correlations between conditions within each group. Results are presented as mean ± SEM.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Complete data sets were obtained in 20 of 22 subjects. In two subjects (one control, 1 OSA), IRL could not be performed because of repeated swallowing when the load was applied. Otherwise their data sets were complete as well. The mean age and BMI of the patients with OSA were significantly higher than that of the control group (Age: 46.7 ± 3.2 versus 29.9 ± 3.7 yr, p < 0.05; and BMI: 32.00 ± 1.26 versus 22.43 ± 0.54 kg/m², p < 0.05). The patients with OSA had moderate-severe apnea as demonstrated by a mean RDI of 71.53 ± 11.73. Five of the 10 subjects were using nasal CPAP as therapy for OSA, the rest were not currently receiving therapy.

GG Activation and Pharyngeal Mechanics

Across all breathing conditions, GG muscle activity was greater in patients with OSA than in control subjects (Table 1). This was true for tonic, peak phasic, and phasic GGEMG (p values < 0.05). During spontaneous breathing with heliox, these differences approached, but did not reach, statistical significance (p < 0.10). Peak phasic GGEMG was greater in both groups during iron lung ventilation than during spontaneous breathing. Pharyngeal pressures were also more negative during iron-lung ventilation than when encountered during spontaneous breathing (p < 0.05 for both control subjects and patients with OSA, spontaneous breathing versus iron-lung, Table 1). There was no significant change in tonic GGEMG between these conditions (p > 0.5 for all comparisons).

During all conditions, peak epiglottic negative pressure and both pharyngeal plus supraglottic resistance were greater in the apneic patients than in the control subjects (Table 1). These differences were most marked during the high flow conditions seen during iron-lung ventilation. There were no significant differences in peak airflow rates between the patients with sleep apnea and the control subjects (all p > 0.10, data not shown).

GGEMG/Epiglottic Pressure Relationship

Across all conditions, the genioglossal/inspiratory negative pressure (GG/Pepi) relationship within breaths remained robust with statistically significant r values across all conditions (r  0.69, p < 0.05). An example of the consistency of this relationship is shown in Figure 1 for one patient with OSA and one control subject during basal breathing and in the iron lung. Furthermore, in both patients with OSA and control subjects, the slope of the within-breath GG/Pepi relationship was not statistically different between conditions, regardless of the epiglottic pressure generated, resistance, airflow, or gas density (ANOVA: control subjects, p = 0.80; patients with OSA: p = 0.71; Table 1). Thus there was a constant relationship (slope) between GGEMG and epiglottic negative pressure across all conditions in both groups.

View larger version (23K): [in this window] [in a new window]   Figure 1.   The relationship between inspiratory epiglottic pressure and genioglossal muscle activation in one patient with sleep apnea (top panels) and one control subject (bottom panels). The conditions are (1) spontaneous breathing, and (2) iron-lung breathing. The consistent relationship between epiglottic pressure and genioglossal activation across these conditions can be seen.

Contrary to our initial hypothesis, there was no significant difference in the mean GGEMG/Pepi relationship between patients with OSA and healthy control subjects (Figure 2 and Table 1). In fact, the slope of this relationship in these two groups was nearly identical under all conditions. Thus, the higher peak phasic GGEMG seen in the apneic patients was not due to an increased responsiveness to pharyngeal negative pressure, but rather to the combination of a higher tonic GGEMG and a greater phasic EMG as well with the increased phasic muscle activity being a product of greater intrapharyngeal negative pressure (higher resistance airway) with a similar GGEMG/Pepi slope.

View larger version (8K): [in this window] [in a new window]   Figure 2.   The individual slopes of the GGEMG/Pepi within breath relationship as well as the mean slope (± SEM) during spontaneous breathing is shown for patients with OSA and control subjects. There were no statistically significant differences between patients with OSA and control subjects, with considerable overlap between these two groups.

In both groups, mean peak epiglottic pressure across all conditions (basal-breathing, heliox, iron lung ventilation, and both levels of IRL) was strongly correlated (between-breath analysis) with both phasic GGEMG and peak phasic GGEMG in both patients with OSA and control subjects (Figure 3, between-breath relationships).

View larger version (16K): [in this window] [in a new window]   Figure 3.   The between-breath relationship between mean peak epiglottic pressure and mean peak phasic GGEMG for all conditions is shown for patients with OSA (A) and control subjects (B). Each point represents the mean (± SEM) GGEMG and epiglottic pressure for a given condition. In both groups there is a tight correlation between these variables.

Air versus Heliox Breathing

During spontaneous breathing, there was a significant decrement in peak negative epiglottic pressure during heliox breathing in the group as a whole when compared with air breathing (2.56 ± 0.28 cm H₂O versus 2.08 ± 0.23 cm H₂O, p = 0.008) (see online data supplement). This decreased epiglottic negative pressure was associated with an increase in peak inspiratory flow and a decrease in supraglottic resistance as expected. In addition, a significant decrement in peak phasic and phasic GGEMG were seen during heliox breathing (see online data supplement). The greatest effect on upper airway resistance during heliox was seen in supraglottic resistance, a finding that is likely caused by a decrease in turbulent airflow in the nose while breathing heliox.

Inspiratory Resistive Loading

During IRL, between-breath analysis revealed that in 14 of 18 subjects in whom data sets were available, there was a significant correlation between peak epiglottic pressure and peak GGEMG (mean r values control subjects, 0.76 ± 0.06; patients with OSA, 0.74 ± 0.07; p = NS between groups). Representative examples of the relationship between epiglottic pressure and peak GGEMG are shown in Figure 4. At an externally applied load of 25 cm H₂O/L/s, tonic, peak phasic, and phasic GGEMG were all greater in the patients with sleep apnea (Table 2). Induced pharyngeal pressure was also greater during loading in the apneic patients, as was pharyngeal and supraglottic resistance, although there was no difference in peak inspiratory flow during loading (Table 2). There was no significant difference in the slope of the relationship between peak epiglottic pressure and peak GGEMG between the two groups (control subjects: slope, 0.72 ± 0.19% max/cm H₂O; patients with OSA: slope, 0.90 ± 0.21% max/cm H₂O; p = 0.54).

View larger version (17K): [in this window] [in a new window]   Figure 4.   The between-breath correlation between peak epiglottic pressure and peak phasic GGEMG during inspiratory resistive loading in two representative patients with sleep apnea (top panels) and two control subjects (bottom panels) is shown. Each point represents the averaged epiglottic pressure and GGEMG of the three trials of each load, as well the three breaths immediately preceding the load (the baseline).

Age-matched Data

As the patients with sleep apnea were significantly older than our control subjects, we performed a subgroup analysis in five patients with OSA and five age-matched control subjects (within 3 yr). The results were identical in this subgroup, and they are shown for the basal breathing and the iron-lung conditions in Table 3. As for the group as a whole, tonic, phasic, and peak phasic GGEMG were greater in the apneic patients than in the aged-matched control subjects, as was peak epiglottic pressure and pharyngeal resistance. Although the magnitude of the difference was the same in this subgroup, the findings were occasionally no longer statistically significant, likely because of the small sample size (Table 3). Again, there was no difference in the GGEMG/Pepi relationship between the two groups (Table 3).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The findings of this study improve our understanding of pharyngeal motor control in normal persons and in patients with OSA. First, we confirmed the finding that genioglossal muscle activation is greater in the apneic patient during spontaneous breathing than in the healthy control subject, and this increased activation is seen across a wide range of epiglottic pressures and gas densities. This finding is consistent with the observation that the pharyngeal airway of the apneic patient is smaller and more collapsible, thus requiring higher dilator muscle activation to maintain patency during wakefulness.

Contrary to our initial hypothesis, this increased dilator muscle activity in the patient with sleep apnea is not due to an increased responsiveness of the genioglossus to changes in pharyngeal pressure during inspiration. In fact, under all conditions, the slope of the within-breath GGEMG/Pepi relationship was not different in patients with OSA from that in control subjects. This was true over a substantial range of pharyngeal negative pressures, both when this relationship was analyzed within breaths and between breaths. It thus appears that the higher peak phasic GGEMG seen in the apneic patient is due to a combination of two factors. First, under all conditions, tonic GGEMG was higher in the apneic patients. Second, as intrinsic upper airway resistance was higher in the apneic patients, they generated more negative intrapharyngeal pressures under all conditions, which, in the face of an identical GGEMG/Pepi slope, yielded increased phasic GGEMG. In both groups, we found that across all conditions, phasic GGEMG was best correlated with peak epiglottic pressure. Phasic genioglossal activity thus appears to be regulated during inspiration largely by intrapharyngeal pressure on a moment-to-moment basis in both patients with sleep apnea and control subjects, and this control mechanism is not intrinsically different in the apneic patient. All of these findings would be consistent with the hypothesis that an anatomic abnormality (tissue characteristics, obesity, etc.) is the primary determinant of apnea pathogenesis, rather than altered pharyngeal motor control.

Although inspiratory genioglossal activation was higher in the patients with sleep apnea under all conditions, pharyngeal and supraglottic resistance were higher as well. This finding suggests that neuromuscular compensation is not complete. However, this is not surprising and would be expected in any physiologic control system. To compensate completely would likely produce an unstable situation.

The increased tonic GGEMG seen in the apneic patients is also of interest, although the mechanisms controlling tonic GGEMG have not been well studied. In this study, tonic GGEMG was constant across all conditions in both groups, suggesting that it does not respond quickly to changes in pharyngeal pressure or airflow. However, it is possible that the increased tonic GGEMG seen in the apneic patient is a response to chronically elevated upper airway pressures or resistance, thus representing neural plasticity, which prevents complete pharyngeal collapse during expiration. Other examples of neural plasticity clearly exist within the respiratory control system. For example, chronic intermittent hypoxia leads to an upregulation of the hypoxic ventilatory response, a phenomenon termed long-term facilitation (27). Neural plasticity thus represents permanent functional transformations that occur in systems of neurons as a result of a chronic stimulus. In the setting of obstructive apnea, repetitive airway collapse or increased resistance could represent the stimulus for long-term augmentation of tonic EMG. We have demonstrated previously that the EMG activity of a purely tonic pharyngeal muscle, the tensor palatini (TP), is increased in patients with sleep apnea as well (28). If this increased tonic activity is the primary manifestation of neuromuscular compensation, it may have important implications in the pathophysiology of OSA. We have previously demonstrated a large decrement in the activity of a tonic upper airway dilator muscle (the tensor palatini) during sleep in normal subjects and a consistent, even greater, rapid loss of activity in this muscle in apneic patients (15). This is consistent with the observations of Orem and colleagues (29) who observed larger sleep-induced decrements in the activity of respiratory neurons with less phasic activation patterns when compared with those with more phasic ones. Thus, loss of basal tonic muscle activation plus the tonic neuromuscular compensation may importantly affect muscle performance during sleep.

Our study has several potential limitations that should be recognized. First, all subjects may not have been completely passive during iron-lung ventilation. However, we have previously shown that iron-lung ventilation leads to a loss of dilator muscle preactivation prior to the onset of airflow in most subjects, and a decrement in surface diaphragmatic EMG, both of which are consistent with the idea that brainstem premotor output is decreased (24). However, even if subjects were not passive in the iron lung, but rather were breathing synchronously with the ventilator, it would not substantially alter the conclusions of this investigation, as the primary goal of iron- lung ventilation was to increase the range of negative pressures that could be generated in the upper airway, rather than decrease central drive. Second, in order to measure airflow with variable gas densities, we have used the assumption that flow through the pneumotachometer is purely laminar. Although some turbulence may exist in the Fleisch pneumotachometer within the range of flows measured in the present study, we believe that any errors introduced by this assumption are small and unlikely to influence the conclusions of the present study. Third, the methods we used to accomplish intersubject comparisons of EMG (% of maximum) could certainly be challenged. However, we have used this methodology for many years now, and have found the technique to be reproducible within a given subject (14).

Finally we did not match our apneic patients and control subjects for either BMI or age. To attempt a match for BMI would have been difficult. Almost all middle-aged to older men with a BMI > 30 kg/m² will snore or have some degree of sleep apnea. Thus this was not attempted. To match for age would have been desirable. For several reasons, we doubt that the difference in age between patients and control subjects influenced our observations. First, in several previous studies we have not observed an independent effect of age on either waking pharyngeal resistance or genioglossal activation/responsivity (30, 31). Second, in the subgroup of age-matched control subjects and apneic patients in this study, the results were essentially identical to that for the groups as a whole. As a result, we doubt that age differences influenced our observations, although this possibility cannot be completely excluded.

In conclusion, we observed the relationship between pharyngeal negative pressure and inspiratory phasic genioglossal muscle activation to be a tight one in both apneic patients and normal control subjects across a wide range of breathing conditions both within and between breaths. In addition, the slope of this relationship was identical between patients and control subjects. The increased genioglossal activity seen in the apneic patient therefore is a product of increased tonic genioglossal activity and an increase in pharyngeal negative pressure during inspiration. We hypothesize that the increased tonic GGEMG represents long-term neuromuscular compensation, although further investigation will be required to elucidate the mechanisms controlling this tonic genioglossal activity. Finally, loss of the tonic EMG during sleep combined with reduced reflex responsivity of the muscles may explain the decrement in pharyngeal dilator muscle activation seen in patients with sleep apnea at sleep onset, leading to airway collapse.