INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

During wakefulness, a rapid negative pressure pulse applied to the human upper airway induces reflex augmentation of the activity of the genioglossus muscle (1, 2). This response is reduced by topical upper airway anesthesia, suggesting that mechanoreceptors in this location form the afferent limb of this reflex (3). Although this reflex appears to be much diminished during non-rapid-eye-movement (NREM) sleep (2), upper airway mechanoreceptor information may still influence upper-airway muscle activity during sleep. Okabe and coworkers (4) compared the relationship of the peak moving-time average (MTA) of the genioglossus electromyogram (EMG-GG) and intercostal muscle EMG (EMG-IM) activity during awake breathing under hypercapnic or hypoxic conditions with that during obstructive apnea. During apnea, the slope of the EMG-GG/EMG-IM relationship was greater than during awake stimulated breathing (hypercapnia or hypoxia). Okabe and coworkers hypothesized that the increase in EMG-GG activity for a given amount of EMG-IM was due to nonchemical stimuli from mechanoreceptors activated by upper airway obstruction.

We reasoned that the upper airway is one site of the mechanoreceptors mediating the augmentation of genioglossus muscle activity during obstructive apnea. It is known that the laryngeal area (which is below the site of airway obstruction) is rich in mechanoreceptors that respond to pressure (5). Therefore, negative pressure in the upper airway during obstructed respiratory efforts should stimulate upper airway mechanoreceptors. Information from these mechanoreceptors could be communicated to the central nervous system (CNS) and result in augmentation of genioglossus activity. If so, then upper airway anesthesia (UAA) (decreased mechanoreceptor output to the CNS for a given negative pressure) should reduce genioglossus muscle activity during obstructive apnea. During obstructive apnea (no flow), esophageal pressure deflections should be equivalent to the pressure deflections in the upper airway below the site of obstruction (6). We therefore studied the relationship between the esophageal pressure deflections and the phasic activity of the EMG-GG during the terminal phase of obstructive apnea before and after application of topical UAA.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Six male subjects with severe obstructive sleep apnea (OSA) were studied in the sleep laboratory. The subjects had more than 60 apneas + hypopneas per hour in a previous study. The current study was approved by the human studies subcommittee of our hospital. Written informed consent was obtained from all subjects before they participated in the study.

Sleep was monitored with two pairs of electrocencephalographic (EEG) leads (C4-A1, O2-A1), two pairs of electrooculographic (EOG) leads, and chin EMG leads using standard techniques (7). An electrocardiographic (ECG) lead was also monitored. The subjects' arterial oxygen saturation (Sa_O2) was measured continuously using pulse oximetry (Model 3700; Ohmeda, Boulder, CO). Airflow was detected with a mask having a built-in pneumotachograph that was worn over the nose and mouth and kept in place with head straps.

The deflection in esophageal pressure during occluded breathing efforts (DP) was determined with a soft, fluid-filled catheter (8). The catheter, a size 5 French polyurethane pediatric feeding tube (Sherwood Medical, St. Louis, MO), was inserted through one nostril and swallowed into the esophagus. The tip was positioned 34 to 36 cm from the nares to obtain tracings with the smallest amount of cardiac artifact. Prior to insertion of the catheter, 1 ml of 4% lidocaine was dripped into one nostril to minimize discomfort during catheter insertion. This occurred 1 h before data were acquired.

The EMG-GG was measured with a pair of unipolar intramuscular electrodes referenced to a single ground (forehead), thus producing a bipolar recording. The fine-wire hook electrodes were constructed from 30-gauge, Teflon-coated stainless steel wire placed within 25-gauge needles with a 2 to 4-mm wire bend (hook) around the needle bevel. After the administration of localized topical anesthesia (4% lidocaine jelly under the tongue), the electrode-containing needles were inserted perorally about 2 cm into the body of the tongue at points 3 mm lateral to the frenulum and midway between the first mandibular incisor and the sublingual fold. Once inserted, the needles were quickly removed, leaving the wires in place. The EMG signals were then amplified, bandpass filtered between 50 and 1,000 Hz, rectified, and electronically integrated on a moving-time average (MTA) basis with a time constant of 100 ms (BMA-830 and MA821RSP; CWE, Ardmore, PA). The phasic EMG-GG activity was defined as the maximum value of the MTA of the genioglossus muscle during an inspiratory effort, minus the tonic value (value at end expiration). The gain was adjusted to provide sufficient deflections during apnea, and was then left unchanged for the remainder of the night. Consequently, EMG-GG measurements in each subject had their own arbitrary units (au).

Upper Airway Anesthesia

A customized nasal cannula was constructed from a CO₂ monitoring cannula by lengthening both of its prongs by inserting 2- to 3-cm segments of polyurethane tubing (size 5-French). The cannula was inserted before the mask was placed on each subject. After lights out, monitoring was performed for 120 min. The subjects were then awakened and lidocaine (5 ml of a 4% solution) was slowly dripped deep into both nostrils through the cannula over a period of 10 min while the subjects were supine. The bulk of the lidocaine subsequently dripped into the hypopharynx-laryngeal area, which was the target site for UAA. All subjects reported having a "lump in the throat." A second 120-min monitoring period was then begun. During this period, another 100 mg of 4% lidocaine (2.5 ml) was slowly dripped into the nose through the customized nasal cannula. The extra length of the nasal prongs ensured that the anesthesia fluid would run into the nasopharynx and not out the anterior nares. During the entire study, subjects slept only in the supine position. The head was kept in a neutral position on a special pillow with an indentation. Subject position was verified by observation with a near-infrared camera and monitor.

Data Analysis

Sleep was staged using standard criteria (7). Apnea was defined as an absence of airflow for 10 s or longer. Hypopnea was defined as a reduction in airflow to less than 50% of baseline for 10 s or longer. The amount of apnea and hypopnea was determined and the apnea + hypopnea index (AHI) computed as the number of events per hour of sleep. The apnea duration and the maximum esophageal pressure deflection prior to event termination (DPmax) were measured. The apnea + hypopnea frequency, event duration, nadir in Sa_O2, and DPmax in the control and lidocaine monitoring periods were compared through use of the paired t test.

We determined the phasic EMG-GG, tonic EMG-GG, and DP for the initial occluded breath and the three final occluded breaths of every other obstructive apnea (mixed or obstructive) during NREM sleep in which four or more obstructed inspiratory efforts occurred. When evidence of arousal by standard criteria (9) was noted before or during the last obstructed effort prior to apnea termination, the previous effort was considered the final obstructed effort. The phasic EMG-GG of the effort at apnea termination (usually coincident with arousal) for each of the analyzed events was also determined. The mean ± SEM number of events analyzed was 40.9 ± 6.1 during control and 45.0 ± 5.7 during lidocaine periods (p = NS). The mean phasic EMG-GG, DP, and phasic EMG-GG/DP for the final three occluded breaths under the control and lidocaine conditions were compared through use of the paired t test. The phasic EMG-GG and the values of the EMG-GG/DP ratio for the first effort during apnea and at apnea termination were similarly compared.

The phasic EMG-GG of the three final occluded breaths was plotted against the associated DP for the first (control) and second (lidocaine) 120-min monitoring periods. Using linear regression, we computed the slope of the phasic EMG-GG-versus-DP relationship. Shifts in position of regression lines after lidocaine were determined by computing the EMG-GG at 50 cm H₂O (EMG-GG-50), using the regression-line relationship.

The tonic EMG-GG for the final three breaths of the analyzed apneas in the control and lidocaine periods was compared through a two-way analysis of variance (ANOVA) for repeated measures, with an analysis of factors (control versus lidocaine period) and breath number (1st, 2nd, or last). In all statistical analyses, a value of p < 0.05 was considered statistically significant. Group data are presented as the mean ± SEM unless otherwise indicated.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The subjects were 50.8 ± 7.0 yr of age (mean ± SD) and were 171 ± 43.7% of ideal body weight (10). The amount of NREM sleep recorded during the control and lidocaine periods did not differ (Table 1). The recorded sleep consisted of Stages 1 and 2, and the amounts of each of these stages also did not differ between control and lidocaine monitoring periods. The amount of apnea and hypopnea was high, as expected, but the AHI was similar during control and lidocaine periods. The mean duration of events was longer during lidocaine periods (p < 0.05) and the DPmax was also significantly greater than during control periods (p < 0.01) (Table 1).

During control periods, the phasic EMG-GG increased, as did the DP during the terminal phase of apnea (Figure 1, top panel). During lidocaine periods, the EMG-GG did not always show phasic activity. When phasic activity was noted, the amount was lower at the same DP. In one subject, monitoring continued after the 120-min lidocaine period was completed. The third panel of Figure 1 shows recovery of the phasic EMG-GG activity at 120 min after the nasal lidocaine instillation was terminated.

View larger version (33K): [in this window] [in a new window]   Figure 1.   Portions of polygraph tracings during obstructive apneas of comparable duration in the same subject. Airflow, the moving-time average (MTA) of the genioglossus EMG (MTA EMG-GG), and esophageal pressure are shown. The top panel shows tracings during the initial (control) monitoring period, the middle panel after/during lidocaine administration, and the third panel at 120 min after lidocaine administration was discontinued. Lidocaine decreased the phasic EMG-GG for a given DP. The bottom panel shows recovery of phasic activity after the effects of topical lidocaine were no longer present.

The mean phasic EMG-GG for the last three occluded breaths was lower during the lidocaine than control monitoring periods (Table 2). In contrast, the mean DP was higher during the lidocaine periods. Thus, the mean EMG-GG/DP ratio for the lidocaine periods was reduced to about 23% of its value in the control monitoring periods. The mean tonic EMG-GG for the last three occluded breaths did not differ between the control and lidocaine periods. In addition, there was no significant change in the tonic EMG-GG from the first to the last of the final three occluded breaths during either control (4.9 ± 0.8, 4.9 ± 0.7, 4.8 ± 0.7 au) or lidocaine (4.2 ± 0.4, 4.2 ± 0.4, 4.2 ± 0.4 au) periods.

On the first occluded breath of apnea, both the phasic EMG-GG and the EMG-GG/DP ratio were also lower during the lidocaine than during the control monitoring periods (2.7 ± 0.3 versus 6.4 ± 1.2 au, respectively, p < 0.03) and (0.09 ± 0.02 versus 0.25 ± 0.05 au/cm H₂O, respectively, p < 0.04). The tonic EMG-GG on the first breath of apnea during the lidocaine monitoring period was also lower (4.1 ± 0.64 versus 5.9 ± 1.02 au, p < 0.05). Thus, the effect of lidocaine was not limited to the final three occluded breaths of apnea.

During both control and lidocaine periods, there was considerable augmentation of the EMG-GG at apnea termination. The phasic EMG-GG at apnea termination was slightly lower in the lidocaine periods (32.0 ± 6.9 versus 38.5 ± 8.1 au), although this difference did not quite reach statistical significance (p = 0.06). However, the ratio of the phasic EMG-GG to DP at apnea termination was significantly lower in the lidocaine than in the control periods (0.80 ± 0.21 versus 1.14 ± 0.27 au/cm H₂O, p < 0.05).

The relationship of the phasic EMG-GG to DP during the last three breaths of obstructive apnea during both control and lidocaine periods is shown for one subject in Figure 2. For all subjects, the regression between the phasic EMG-GG and DP was highly significant (p < 0.001). The mean and range of correlation coefficients is shown in Table 3. The slope of the EMG-GG-versus-DP curve was significantly smaller during lidocaine than during control periods for each subject (Figure 3) and for the group as a whole (Table 3). The entire curve was shifted downward, since the EMG-GG at 50 cm H₂O was lower during lidocaine periods (Table 3). Thus, a given level of DP was associated with a lower amount of phasic EMG-GG activity.

View larger version (29K): [in this window] [in a new window]   Figure 2.   The phasic EMG-GG versus DP relationship for one subject. In the control condition, the regression line is (EMG-GG = 0.895 + (0.182) DP; r = 0.76, p < 0.0001). In the lidocaine condition the regression line is (EMG-GG = 0.475 + (0.032) DP; r = 0.61, p < 0.0001). After upper-airway anesthesia (UAA), the phasic EMG deflections for a given DP were reduced.

View larger version (16K): [in this window] [in a new window]   Figure 3.   The slope of the EMG-GG-versus-DP regression line for each patient, and the value of EMG-GG at 50 cm H2O determined from the regression equation. The values for each individual were reduced during lidocaine as compared with control monitoring periods.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Our results indicate that topical UAA greatly diminishes the amount of phasic genioglossus muscle activity for a given DP during obstructive apnea in NREM sleep. This finding is consistent with our hypothesis that stimulation of superficial receptors in the upper airway during OSA augments genioglossus muscle activity. In fact, the majority of the augmentation in the phasic EMG-GG during obstructive apnea appears to be due to upper airway receptor stimulation.

Previous studies of negative pressure pulses applied to the upper airway during wakefulness in humans have shown a reflex activation of the genioglossus muscle (1) that is greatly decreased by topical anesthesia of the upper airway (3). Wheatley and associates also found this reflex to be greatly diminished during NREM sleep as compared with wakefulness in normal human subjects (2). Our findings do not necessarily conflict with this work. The negative pressures in Wheatley and associates' study were applied quickly during inspiration at a time when ventilatory drive was not high (stable breathing). In addition, the mean negative pressure in the pharynx was around 10 cm H₂O. The pressures in our study were applied by a slower onset of inspiratory effort during a time when ventilatory drive was increasing. The pressures in our study were also considerably more negative (usually 20 cm H₂O or more) than in that of Wheatley and associates. Studies with anesthetized (11, 12, 13) and nonanesthetized (14) animals have also shown that wakefulness does not appear necessary for negative upper-airway pressure to augment upper airway muscle activity, although sleep may diminish the magnitude of the augmentation (14).

Our study did not directly address the nature of the receptors influenced by topical lidocaine. Mechanoreceptors in the laryngeal area that specifically respond to flow, transmural airway pressure, and airway distortion (also called "drive" receptors) have been identified (5). With airway occlusion (no flow), the latter two types of receptors would be stimulated during inspiratory efforts. However, studies in animals comparing tracheal (distortion) and supralaryngeal occlusions (distortion and transmural pressure) suggest that upper airway muscle activity (13) or afferent activity from the superior laryngeal nerve (15) is augmented primarily by stimulation of pressure-sensitive receptors as opposed to receptors responding to airway distortion. Thus, the effect of lidocaine on the phasic EMG-GG in our study was most likely due to an effect of lidocaine on pressure-sensitive mechanoreceptors.

During obstructive apnea, only areas below the point of upper airway closure are exposed to negative airway pressure swings. Thus, the mechanoreceptors potentially stimulated should be present in the hypopharynx and laryngeal areas. Horner and coworkers used topical anesthesia selectively applied to different areas of the upper airway to estimate that 32% of the response of the genioglossus muscle to negative upper-airway pressure was due to receptors in the laryngeal area (3). Although we dripped lidocaine into the nose, it rapidly ran into the hypopharynx and laryngeal areas. All of our subjects reported having "lumps in their throat" and most coughed at the end of the lidocaine application (presumably from lidocaine reaching the trachea), implying some impairment of laryngeal sensation and reflex glottic closure. We therefore feel confident that the mechanoreceptors exposed to negative pressure during obstructive apnea were anesthetized.

Augmentation of genioglossus muscle activity during airway occlusion could result from increases in chemical as well as mechanoreceptor stimuli. The relative importance of chemical and mechanoreceptor stimuli is difficult to determine because increases in ventilatory drive are accompanied by increases in mechanoreceptor stimulation (suction pressure). Previous studies have demonstrated a parallel increase in the EMG-GG and the EMG of the diaphragm (16) or intercostal muscles (4) during the terminal phase of obstructive apnea. We found a linear relationship between DP (an index of ventilatory drive and mechanoreceptor stimulation) and the phasic EMG-GG (Figure 2). A similar linear relationship between pharyngeal pressure (below the site of obstruction) and averaged EMG-GG during occluded breaths has been previously reported (17). However, documenting these relationships does not prove that mechanoreceptor stimulation augments genioglossus muscle activity. In contrast, the large reduction in the phasic EMG-GG for a given DP after topical UAA is evidence that stimulation of upper-airway receptors during obstructive apnea accounts for a significant portion of the augmentation in genioglossus muscle activity during airway occlusion.

Additional evidence for the role of upper airway mechanoreceptors during airway occlusion is provided by a study by Issa and coworkers comparing nasal and tracheal occlusions in spontaneously breathing dogs during NREM sleep (14). A progressive augmentation in the EMG-GG during sleep was noted only during nasal occlusions. Although ventilatory drive increased during both types of occlusion, tracheal occlusions would not expose the upper airway to negative pressure swings. Issa and coworkers' study therefore suggests that negative pressure stimulation of upper airway receptors, rather than increased ventilatory drive, was the major factor responsible for the observed augmentation of the EMG-GG during nasal occlusions.

We found a persistent relationship between EMG-GG and DP after UAA. This implies that a portion of the phasic EMG was due to increasing ventilatory drive, that mechanoreceptors at locations other than the upper airway were active, or that lidocaine did not completely abolish activity of the upper airway mechanoreceptors. A study in patients whose upper airways were isolated from negative pressure during inspiration (laryngectomy) suggested that as ventilatory drive increases, genioglossus muscle activity can increase independently of upper airway mechanoreceptor input (18). However, in our study, the phasic EMG-GG was markedly diminished after airway anesthesia, suggesting that during airway obstruction in NREM sleep, augmentation of the phasic EMG-GG was principally dependent on upper airway mechanoreceptor input to the CNS.

In contrast to the lidocaine-induced reduction in the phasic EMG-GG during the last three occluded breaths in our study, the small decrease in the tonic EMG-GG after lidocaine was not statistically significant. We did note a small but significant decrease in the tonic EMG-GG of the first apneic effort during the lidocaine period. In anesthetized dogs, the constant application of negative upper airway pressure augmented the tonic as well as the phasic EMG-GG (12). Issa and coworkers (14) found an augmentation of both the phasic and tonic EMG-GG on the first breath after a nasal occlusion in awake dogs. However, during NREM sleep, only the phasic activity was augmented. Thus, during NREM sleep, intermittent stimulation of upper airway mechanoreceptors during occluded respiratory efforts appears to augment mainly phasic EMG-GG activity. Perhaps this is the reason why we did not find a significant decrease in tonic EMG-GG activity during the last three occluded efforts of apnea after UAA.

The clinical significance of mechanoreceptor augmentation of upper airway muscle activity has been evaluated by studying the effects of topical UAA on airway patency during sleep. DeWeese and coworkers found that topical anesthesia increases upper airway resistance during sleep in normal subjects (19). Studies have also demonstrated an increase in apnea after UAA in normal subjects (20) and snorers (21). In contrast, previous studies in moderate (22) and severe (23) OSA did not find an increase in the AHI after UAA. In the current study, we also did not document an increase in the AHI during lidocaine periods in a group of subjects with severe OSA. One explanation for this finding is that upper airway mechanoreceptors in patients with significant OSA may require greater pressure changes before they contribute substantially to the augmentation of upper airway activity during sleep. Thus, UAA would alter upper airway activity very little in the preapneic period. We did find that UAA increased the mean duration of apnea and mean maximum DP prior to apnea termination, in accord with the findings in a previous study (23).

In an evaluation of our results, some possible confounding effects must be considered. First, changes in head position between the control and lidocaine periods could have caused changes in the EMG-GG. We studied patients only in the supine position and used a special pillow to attempt to standardize head position. Although changes in head position could still have influenced our readings, this should have occurred during both control and anesthesia periods. It seems unlikely that changes in head position would have produced a systematic decrease in the EMG after lidocaine. Second, one could hypothesize that systemic absorption of lidocaine could have altered the control of the genioglossus muscle through CNS effects. In a previous study performed with similar doses of lidocaine, we were unable to document evidence of a central depressant effect with respect to the auditory arousal threshold (23). Basner and coworkers, using doses of lidocaine similar to those employed in the present study, also found no evidence of central effects of lidocaine on auditory stimulus-induced arousal (24). We therefore doubt that central effects of lidocaine produced our results. Additionally, the precedence of the control monitoring periods to the lidocaine periods in our study could conceivably have altered our results. For example, time-of-night effects or deterioration of the integrity of our EMG electrodes could have reduced the EMG-GG amplitude independent of effects of lidocaine. However, these factors are unlikely to have produced our results. In all of our subjects, a large phasic EMG-GG was seen at apnea termination during the lidocaine periods, showing that the electrodes were intact. In one subject who was monitored for a prolonged period after the nasal lidocaine drip was discontinued, recovery of EMG-GG activity was noted (Figure 1). We also restudied two subjects and instilled saline at the end of an initial 120-min monitoring period. The mean EMG-GG/DP ratio of the final occluded breath for these two subjects was 0.27 during the initial 120-min monitoring period and 0.26 during the subsequent (saline) monitoring period. For the foregoing reasons, we do not believe that the order of control and lidocaine monitoring periods in our study produced the observed changes in the phasic EMG-GG.

In summary, our study found that topical lidocaine applied to the upper airway reduced the amount of phasic activity of the EMG-GG for a given level of inspiratory effort (DP) during the terminal portion of obstructive apnea. This suggests that upper airway mechanoreceptor stimulation during obstructed breathing efforts augments phasic genioglossus muscle activity during NREM sleep.