INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Many studies have shown that respiratory-related electromyographic (EMG) activity recorded from the genioglossus (tongue protrudor) muscle in humans and animals (1) increases in response to hypoxia and/or hypercapnia. This finding has led to the suggestion that activation of the genioglossus muscle under hypoxic and hypercapnic conditions is accompanied by tongue protrusion and dilation of the pharyngeal airway, which ultimately increases inspiratory airflow. Consequently, it has been hypothesized that the genioglossus muscle has a primary role in protecting or enhancing upper airway patency in individuals who have obstructive sleep apnea (OSA), since this syndrome is characterized by periodic upper airway obstruction that results in hypoxemia and hypercapnia (7, 8). However, recent studies in rats have shown that the hyoglossus and styloglossus muscles (tongue retractors) are coactivated with the genioglossus muscle in response to hypoxia, hypercapnia, or stimulation of the hypoglossal nerve (5, 9). This finding is consistent with the understanding that the medial branch of the hypoglossal nerve innervates the tongue protrudor, whereas the lateral branch innervates the retractor muscles (10).

Given the foregoing finding, it is possible that the retractor muscles of the tongue are activated in response to hypoxia and/or hypercapnia in humans. Furthermore, it is conceivable that maintenance of airway patency is not simply the result of tongue protrusion, but is more complex and involves coactivation of the protrudor and retractor muscles. However, only one published investigation has so far examined the response of the tongue retractors in humans to hypoxia and hypercapnia, and the results were inconclusive, since the findings from only one subject were published (11). Nonetheless, the results showed that the tongue retractor muscles were not coactivated with the genioglossus muscle during a breathhold maneuver. This finding lead Sauerland and Mitchell (11) to suggest that the retractor muscles of the tongue do not have a significant role in maintaining airway patency in humans during breathing.

Given the limited data published by Sauerland and Mitchell (11), and the recent findings obtained with rats (5), further investigation is required to examine the role of the retractor muscles in protecting airway patency in humans. As a first step toward this goal, the present investigation was designed to examine the response of human tongue protrudor and retractor muscles during a breathhold maneuver and in steady-state hypoxic hypercapnia.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects and Physiologic Measurements

Eight healthy subjects (four males and four females) between the ages of 24 and 36 yr gave their informed written consent to participate in the study, which was approved by the Institutional Review Board of Columbia Presbyterian Medical Center. On one occasion the subjects visited the laboratory to become acquainted with the experimental apparatus and conditions for the study. On a second, subsequent occasion, electromyographic (EMG) activity of the protrudor and retractor muscles of the tongue was recorded. To record EMG activity of the upper airway muscles, we inserted fine-wire electrodes into the genioglossus muscle and the hyoglossus and styloglossus muscles. To insert the fine wires, we placed a 25-gauge needle, containing two Teflon-coated stainless steel recording wires bonded together, into the body of each muscle. The fine-wire electrodes used to record from the genioglossus muscle were inserted 20 mm deep and at right angles to the oral mucosa at a point that was 3 to 4 mm lateral to the frenulum and anterior to the lingular salivary duct. To record retractor muscle activity, we inserted electrodes in the area of interdigitation between the styloglossus and hyoglossus muscles at a point approximately 50 mm posterior to the tip of the tongue and 15 mm below the circumvallate papillae. At this point the electrodes were inserted superficially (depth approximately 2 to 3 mm) in order to avoid the underlying substance of the genioglossus muscle. Illustrations of the anatomic landmarks and the angles and depths of needle insertion have been previously published (11) (Figures 1 and 2).

View larger version (10K): [in this window] [in a new window]   Figure 1.   Representative recording obtained simultaneously from tongue protrudor and retractor muscles of one subject before, during, and after completion of a protrusive (top and bottom left) and retractor isometric maneuver (top and bottom right).

View larger version (23K): [in this window] [in a new window]   Figure 2.   Representative example of raw and integrated EMG recordings obtained from tongue protrudor and retractor muscles of one subject before and during completion of a breathhold maneuver. Note the gradual and simultaneous increase in protrudor and retractor EMG activity.

After the electrodes were inserted, the subjects breathed through a sealed face mask while in a seated position. The mask was attached to a two-way nonrebreathing valve. The expiratory port of the value was attached to a pneumotachometer (Model R3813; Vacumed, Ventura, CA) that was used to measure expiratory airflow on a breath-by-breath basis. Depending on the experimental condition, the inspiratory port of the valve was either open to room air or attached to a spirometer (30-L capacity) (Warren E. Collins Inc., Braintree, MA) that contained a mixture of 7% oxygen (O₂) and 7% carbon dioxide (CO₂), with N₂ as the balance. The fractional concentrations of end-tidal O₂ and CO₂ were sampled at the level of the face mask by using rapidly responding O₂ (Model OM-11; Beckman, Anaheim, CA) and CO₂ (Model CD-3A; Ametek, Pittsburgh, PA) analyzers. Heart rate (HR) and oxygen saturation (Sa_O2) were monitored on a beat-by-beat basis during each protocol through the use of an HR monitor (Model CS-625; Burdick, Milton, WI) and a pulse oximeter (Biox 3700; Ohmeda Corp., Boulder, CO), respectively. The EMG signals recorded from the tongue protrudor and retractor muscles were amplified (Model NL104; Harvard Apparatus, Holliston, MA), filtered (Model NL126; Harvard Apparatus), and integrated (Model NL703; Harvard Apparatus).

Expiratory airflow, integrated airflow (tidal volume, [VT]), raw and integrated EMG recordings, electrocardiograms (ECGs), Sa_O2, and end-tidal CO₂ and O₂ signals were digitized on-line at a sampling frequency of 1,000 Hz and fed into a microcomputer with a commercially available software package (CODAS; Dataq Instruments, Akron OH).

Breathhold Maneuvers

After an accommodation period of 30 min, the subjects completed a minimum of two breathhold maneuvers. Each maneuver was separated from the next by at least a 5-min period of rest. Before each maneuver the subjects breathed room air for 5 min in order to provide baseline measurements. Subsequent to the baseline period and at the end of a normal inspiration, the subjects held their breath for as long as possible. After each maneuver, all physiologic measurements were recorded for an additional 1 min.

Steady-State Hypoxia and Hypercapnia

After the breathhold maneuvers were completed, EMG activity was recorded from the tongue protrudor and retractor muscles during hyperpnea, which was induced by the inspiration of a gas mixture comprising 7% O₂ and 7% CO₂ with a balance of N₂. Prior to their exposure to the gas mixture, the subjects breathed room air for 5 min before being connected to a spirometer that was prefilled and replenished throughout the test with the desired gas mixture. Seven of the eight subjects inspired the gas mixture for 4 min. The remaining subject inspired the mixture for 3 min. After the test period the subjects breathed room air for an additional 1 min. The subjects completed two trials that were separated by at least 20 min of rest.

Isometric Maneuvers

Before completion of the breathhold maneuvers and after completion of the steady-state trials, the subjects completed a series of isometric maneuvers in order to ensure that the protrudor and retractor muscles were properly isolated, and to record a maximal EMG response from each type of muscle. To generate a maximal protrusive EMG response, the subjects were instructed to force their tongue against the mandibular incisor teeth and to maintain this maneuver for a minimum of 3 s. To generate a maximal EMG response from the retractor muscles, the subjects were instructed to complete two maneuvers. First, the subjects were directed to voluntarily retract the tongue while attempting to maintain the intrinsic shape of the tongue (a maneuver used by Sauerland and Mitchell) (11). However, all subjects found it difficult to maintain the intrinsic shape of the tongue during this maneuver, and therefore completed a second maneuver during which the tongue was initially protruded beyond the oral cavity before being retracted while an investigator provided resistance by grasping the tongue. In all cases this latter maneuver was used to obtain the maximal response of the retractor muscles. The tongue protrusion and retraction maneuvers were completed three times, with at least a 1-min rest between each trial, to ensure that each subject was exerting a maximum voluntary effort that was reproducible.

Data Analysis

Six to ten breaths recorded immediately before and after breathholding or the inspiration of the hypoxic, hypercapnic gas mixture were used to measure inspiratory duration (TI), expiratory duration (TE), and VT during baseline and recovery from these maneuvers. The fractional concentrations of CO₂ and O₂, Sa_O2, and EMG activity recorded from the protrudor and retractor muscles were measured for the corresponding breaths. In addition, the measurements described were obtained from six to 10 breaths recorded during each minute in which the subjects were inspiring the hypoxic hypercapnic gas mixture. Furthermore, EMG activity was measured from the tongue protrudor and retractor muscles for the last 10 s of a given breathhold maneuver.

In order to quantify EMG activity, we identified an electrical zero during baseline conditions of a given trial (breathhold or steady-state hypoxic hypercapnia). Tonic EMG activity (defined as the minimum EMG value recorded during expiration), phasic EMG activity (defined as the difference between peak inspiratory and tonic EMG activity), and peak EMG activity (tonic plus phasic EMG activity) were used to quantify the response of the upper airway muscles to hypoxia and hypercapnia.

Each of the previously described EMG variables was reported as a percentage of the maximal response that was recorded during either the protrudor or retractor isometric maneuvers. The onset time of protrudor and retractor phasic EMG activity for the six to 10 breaths recorded during steady-state hypoxic hypercapnia was determined and quantified relative to the onset of inspiration, which was identified by cessation of the expiratory flow signal. A negative value indicated that a phasic burst of EMG activity began prior to inspiration, whereas a positive value indicated that phasic EMG activity occurred after the onset of inspiration. The measurement of onset time was based on visual inspection of the integrated EMG recordings. Two of the authors examined the EMG recordings and compared results in order to ensure reproducibility.

Mean values for each physiologic variable were calculated for each six- to 10-breath epoch. Subsequently, a group mean value was calculated for each experimental condition. Paired t tests were used to determine whether the tonic, phasic, and peak EMG responses recorded from the protrudor and retractor muscles during the breathhold maneuvers were significantly different from each other and from baseline measurements. For each muscle, a one-way analysis of variance (ANOVA) with repeated measures was used to determine whether a significant difference in tonic, phasic, and peak EMG activity existed between each time period (baseline, at 1 to 4 min of the experimental condition, and recovery) during steady-state hypoxic hypercapnia. If a significant difference existed among the time periods, individual comparisons were made using the Student-Newman-Keuls post hoc test. A two-way ANOVA, in conjunction with the Student-Newman-Keuls post hoc test, was used to determine whether significant differences in onset time of phasic EMG activity during steady-state hypoxic hypercapnia existed among the different time periods and between upper airway muscle (protrudor versus retractor) activities. All values in the text and figures are presented as means ± SE, and the level of significance chosen was p < 0.05.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Figure 1 shows EMG activity recorded from the protrudor and retractor muscles of one subject during the completion of an isometric tongue protrusion and retraction maneuver. Note that during completion of the protrusive maneuver, a minimal change in EMG activity was recorded from the tongue retractor muscles as compared with baseline. Similarly, minimal EMG activity was recorded from the tongue protrudor muscle during completion of the retraction maneuver.

Figure 2 shows EMG activity recorded from the tongue protrudor and retractor muscles of one subject during a breathhold maneuver. The figure shows that during the maneuver, a gradual increase in both tonic and phasic EMG activity occurred in the protrudor and retractor muscles. This finding is supported by the average results shown in Figure 3. The average tonic, phasic, or peak EMG activity recorded from the protrudor and retractor muscles during the breathhold maneuver was similar and significantly greater than the corresponding activity recorded under baseline conditions. Furthermore, the average results obtained from the breathhold maneuvers revealed that the onset of protrudor EMG activity was not significantly different from that of retractor EMG activity relative to the beginning of the maneuver (onset of protrudor activity = 45.33 ± 8.71 s, versus onset of retractor activity = 41.38 ± 9.07 s).

View larger version (15K): [in this window] [in a new window]   Figure 3.   Histograms showing the average tonic, phasic, and peak EMG activity recorded from the protrudor and retractor muscles during a breathhold maneuver. Note that no significant differences in tonic, phasic, or peak EMG activity occurred between protrudor and retractor muscles during completion of the maneuvers.

Figure 4 shows a recording of raw and integrated EMG activity obtained from one subject before, during, and after exposure to hypoxic hypercapnia. This individual response was similar to the average results shown in Figure 5, which demonstrate that tonic, phasic, and peak EMG activity recorded from the protrudor and retractor muscles was significantly greater during hypoxic hypercapnia than during baseline and recovery (p  0.003 for all comparisons). Note that one subject terminated the hypoxic hypercapnia protocol after 3 min. Therefore, the average results shown for Min 4 are based on data obtained from seven of the eight subjects.

View larger version (17K): [in this window] [in a new window]   Figure 4.   Respresentative example of raw and integrated recordings of protrudor and retractor muscle activity obtained from one subject before (baseline), during (Min 1 to 4), and after (recovery) inspiring gas mixture consisting of 7% oxygen and 7% carbon dioxide, with the balance consisting of nitrogen.

View larger version (18K): [in this window] [in a new window]   Figure 5.   Line plots showing tonic (top), phasic (middle), and peak (bottom) increases in integrated EMG activity before (baseline), during (Min 1 to 4), and after (recovery) the inspiration of a gas mixture consisting of 7% oxygen, and 7% carbon dioxide, with the balance consisting of nitrogen. *Significantly different from baseline and recovery.

Figure 6 (top panel ) shows the average onset time of phasic protrudor and retractor muscle activity relative to the onset of inspiration. An average for the entire duration (4 min) of steady-state hypoxic hypercapnia is presented, since statistical analysis revealed no difference in onset time for a given muscle between each minute of the maneuver. The results show that the average onset time of phasic protrudor EMG activity occurred before the onset of inspiration. Similarly, in all but one subject, retractor muscle activity occurred before the onset of inspiration. In the one subject in whom the onset of retractor activity occurred in the middle of inspiration, an additional, smaller phasic response was often present earlier in inspiration (Figure 4). However, this smaller response was often obscured by the larger response, and we therefore recorded the onset of phasic EMG activity that had the greatest amplitude. Figure 6 shows the average onset time of retractor muscle activity for all subjects (Figure 6, middle histogram) and subsequent to removal of the one subject who showed a divergent response (Figure 6, far-right histogram). Figure 6 (bottom panel ) shows the average inspiratory and expiratory time of each breath recorded during the 4 min of hypoxic hypercapnia.

View larger version (13K): [in this window] [in a new window]   Figure 6.   Histograms showing the average onset time of phasic protrudor and retractor EMG activity (top) and average inspiratory time and expiratory time (bottom) recorded during inspiration of a gas during mixture consisting of 7% oxygen and 7% carbon dioxide, with the balance consisting of nitrogen for 4 min. Note that the average onset time for retractor muscle activity is displayed for the entire group (middle histogram) and subsequent to the removal of one subject who presented a response divergent from that of the other subjects.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Critique of the Methods

In the present investigation it was necessary to ensure that the retractor muscles were adequately isolated, in order to avoid the underlying substance of the genioglossus muscle. Therefore, subjects completed isometric tongue protrusion and retraction maneuvers before and after completion of the experimental protocols, to ensure that the activities of these muscles were separate from one another. If the isometric maneuvers indicated that the muscles were not properly isolated, the fine-wire electrodes were reinserted or the data collected was not included in the analysis. For five of eight subjects, electrode insertion was successful on the initial attempt.

It was important to confirm that the changes in phasic and tonic EMG activity that we recorded were the result of the experimental perturbation and not a function of changes in posture or other artifacts that have been shown to alter protrudor and retractor muscle activity in previous investigations (11, 12). Hence, we initially secured and isolated the electrodes to ensure that the EMG recordings would be free from motion artifact. Additionally, to control for the effect of posture on tongue muscle activation, we instructed the subjects to maintain their head and body in one position during the completion of an experimental protocol. Subjective observation indicated that the subjects were able to maintain a given position. Additionally and more importantly, the results showed that EMG activity recorded immediately before (baseline) and after (recovery) a given experimental condition were not significantly different. This finding suggests that recorded increases in tonic and phasic activity were probably a function of the experimental perturbation. We also verified that no ECG artifact was present in the EMG signal, by subjectively viewing the EMG signal and by ensuring that the frequencies of ECG and EMG phasic activity were dissimilar. This verification was important, since ECG artifact is often misinterpreted as phasic activity in EMG recordings.

Tongue Protrudor and Retractor Responses to Apnea and Steady-State Hypoxic Hypercapnia

The results obtained in many investigations have suggested that the genioglossus muscle has a primary role in protecting or enhancing airway patency (7, 8). This belief has been based partly on findings obtained with humans showing that the genioglossus muscle is activated in response to hypoxia and hypercapnia during wakefulness and sleep (4, 13), and that tongue protrusion, elicited by genioglossus activation, may be accompanied by dilation of the pharyngeal airway (17). Although this hypothesis is plausible, past studies have largely ignored the role of the styloglossus and hyoglossus muscles in maintaining upper airway patency, despite the understanding that the hypoglossal nerve innervates the genioglossus muscle (medial branch of the nerve) and both the hyoglossus and styloglossus muscles (lateral branch of the nerve), and that coactivation of protrudor and retractor muscles of the tongue is required to perform other behaviors such as swallowing and vocalizing (10, 18, 19).

To our knowledge only one prior investigation (11) examined the response of both human tongue protrudor and retractor muscles to alternations in arterial blood gas concentrations, and the published results were inconclusive, since data from only one of 14 subjects who participated in the investigation were shown, and average findings were not reported. Nonetheless, Sauerland and Mitchell (11) showed that the retractors were not coactivated with the tongue protrudor during a breathhold maneuver, and they concluded on this basis that the tongue retractors do not play a significant role in maintaining airway patency during breathing. This finding was in direct contrast to our finding that the protrudor and retractor muscles of the tongue are coactivated in response to breathholding. It may be argued that the retractors were not activated in the Sauerland and Mitchell study (11) because the threshold of activation for these muscles was greater than that of the tongue protrudor, and the subjects did not prolong their breathholding to a point at which the threshold for retractor activity was attained. However, the results of our investigation suggest coactivation of the protrudor and retractor muscles, since the onset times of increases in activity of these muscles were similar. Nevertheless, further investigation is required to support our findings, since our suggestion is based on the measurement of onset time and not on end-tidal O₂ or CO₂ values, which we were unable to measure during the breathhold maneuvers.

The results obtained from the breathhold maneuvers completed in our study were supported by the findings made during steady-state hypoxic hypercapnia, which showed significant and parallel increases in tonic, phasic, and peak protrudor and retractor activity relative to baseline and recovery measurements of EMG activity. Collectively, our results strongly suggest that coactivation of the tongue protrudor and retractor muscles occurs during both apnea and hyperpnea in humans. Therefore, previous hypotheses that increased phasic genioglossus EMG activity in humans during hypoxia and/or hypercapnia is associated with tongue protrusion (7, 8) may be incorrect, since our results suggest that under these conditions, it is equally plausible that the tongue may retract. The latter suggestion is supported by recent studies with animals and humans. Fuller and colleagues (7) showed that coactivation of rat protrudor and retractor muscles in response to hypoxia and hypercapnia produced a net retractive tongue force. Similarly, Eisele and coworkers (20) showed in humans that stimulation of the whole hypoglossal nerve (prior to its bifurcation into medial and lateral branches) results in tongue retraction. However, further investigation of tongue position during apnea or hyperpnea induced by hypoxia and hypercapnia in humans is required, since tongue location was not monitored in the present investigation.

Onset Time of Human Tongue Protrudor and Retractor Muscles during Hyperpnea

In addition to demonstrating that human tongue protrudor and retractor muscles are coactivated during hyperpnea, we showed that the onset of this parallel activation occurred before or during the early part of inspiration in all but one subject. The onset time of phasic protrudor muscle activity was similar to that obtained in previous investigations (2, 6, 21), whereas to our knowledge, our study represents the first time that onset times have been reported for phasic retractor muscle activation in humans. Although the onset of retractor muscle phasic activity in one subject showed a divergent response compared with that of the remainder of the group (the onset of EMG activity occurred in midinspiration), this is not a unique finding. Hwang and colleagues (22) had previously shown in cats that during exposure to severe hypoxic hypercapnia, activities recorded from single hypoglossal nerve fibers could be characterized by a variety of discharge patterns, including the pattern observed in the present investigation. The parallel and early activation of the tongue protrudor and retractor muscles in our study supports the suggestion that coactivation of these muscles is required for protecting airway patency before the development of inspiratory flow or attainment of peak inspiratory flow rates.

Physiologic Mechanisms Responsible for Activation of the Protrudor and Retractor Muscles

The parallel increase in protrudor and retractor EMG activity during the breathhold maneuvers and steady-state hypoxic hypercapnia was probably elicited by the excitation of central and peripheral chemoreceptors, since a number of studies have shown that hypoxia and hypercapnia are accompanied by an increase in hypoglossal motor neuron activity (23). Similarly, it has been shown that activity of the tongue protrudor in humans and animals (3, 15, 16, 24, 25), and of the tongue retractors in animals (5), is activated by hypoxia and/or hypercapnia. However, it is possible that the increase in protrudor and retractor EMG activity during the breathhold maneuvers in our study was partly mediated by removal of the vagal inhibitory feedback that occurs in response to lung inflation. This suggestion was supported most recently by the finding that lung inflation causes inhibition of both the protrudor and retractor muscles in rats (8). Nonetheless, it is unlikely that removal or attenuation of lung afferent feedback was the single or primary stimulus responsible for the changes in protrudor and retractor EMG activity in our subjects, since we observed a progressive increase in EMG activity throughout the breathhold maneuver, which probably would not occur if removal of phasic volume feedback was the only stimulus responsible for the observed response. The role of central and peripheral chemoreceptors in eliciting coactivation of the protrudor and retractor muscles is further supported by the finding in the present study that protrudor and retractor muscle activity increased during steady-state exposure to hypoxic hypercapnia despite the presence of inhibitory afferent feedback resulting from lung inflation during hyperpnea.

Conclusion

In conclusion, we have shown that human tongue protrudor and retractor muscles are coactivated in response to hypoxia and hypercapnia. This finding suggests that activation of both protrudor and retractor muscles may be required to protect or enhance upper airway patency under hypoxic and hypercapnic conditions.