Response of Genioglossus Muscle to Increasing Chemical Drive in Sleeping Obstructive Apnea Patients

Abstract

Study Objectives:

Subjects with a collapsible upper airway must activate their pharyngeal dilators sufficiently in response to increasing chemical drive if they are to maintain airway patency without arousal from sleep. Little is known about the response of pharyngeal dilators to increasing chemical drive in these subjects. We wished to determine, in obstructive apnea patients, the response of the genioglossus to increasing chemical drive and the contribution of mechanoreceptor feedback to this response.

Design:

Physiological study.

Setting:

University-based sleep laboratory.

Patients:

20 patients with obstructive apnea.

Interventions:

Genioglossus activity was monitored during overnight polysomnography on optimal continuous positive airway pressure (CPAP). Intermittently, inspired gases were altered to produce different levels of ventilatory stimulation. CPAP was then briefly reduced to 1.0 cm H₂O (dial-down), inducing an obstruction.

Measurements and Results:

Without mechanoreceptor feedback (i.e., on CPAP) the increase in genioglossus activity as ventilation increased from 6.1 ± 1.4 to 16.1 ± 4.8 L/min was modest (ΔTonic activity 0.3% ± 0.5%maximum; ΔPhasic activity 1.7% ± 3.4%maximum). Genioglossus activity increased immediately upon dial-down, reflecting mechanoreceptor feedback, but only when ventilation before dial-down exceeded a threshold value. This threshold varied among patients and, once surpassed, genioglossus activity increased briskly with further increases in chemical drive (1.1% ± 0.84%GG_MAX per L/min increase in V_E).

Conclusions:

In sleeping obstructive apnea patients: (1) Mechanoreceptor feedback is responsible for most of the genioglossus response to chemical drive. (2) Mechanoreceptor feedback is effective only above a threshold chemical drive, which varies greatly among patients. These findings account in part for the highly variable relation between pharyngeal mechanical abnormalities and apnea severity.

Citation:

Loewen AHS; Ostrowski M; Laprairie J; Maturino F; Hanly PJ; Younes M. Response of genioglossus muscle to increasing chemical drive in sleeping obstructive apnea patients. SLEEP 2011;34(8):1061-1073.

INTRODUCTION

Subjects with a collapsible upper airway who develop obstructive apneas or hypopneas at sleep onset must activate their pharyngeal dilators sufficiently through reflex mechanisms that are independent of arousal if they are to remain asleep. The reflex mechanisms known to increase pharyngeal dilator activity during sleep are: (1) an increase in chemical drive (i.e., excitability of the respiratory centers related to hypoxia and hypercapnia) acting via central connections between the respiratory centers and the pharyngeal motor nuclei in the brain stem, and (2) an increase in negative pharyngeal pressure acting reflexly via mechanoreceptors in the pharynx and larynx.^1,2 During an obstructive respiratory event, pharyngeal pressure can become more negative only if diaphragm and/or intercostal muscle activity increases. Since, during sleep, such an increase occurs principally in response to increasing chemical drive,³ activation of pharyngeal dilators by either mechanism requires an increase in chemical drive. Accordingly, the response of pharyngeal dilators to increasing chemical drive is of considerable importance in determining whether the subject can re-open the airway without arousal from sleep. Specifically, a high chemical drive threshold for recruiting dilators would make it more likely that arousal occurs before reflex opening, with an arousal-associated ventilatory overshoot.⁴ Even without arousal, a high chemical drive threshold would result in a greater ventilatory overshoot when the airway opens because of the more negative intrathoracic pressure at the time of opening.² A large ventilatory overshoot, with or without arousal, promotes instability and a high apnea-hypopnea index (AHI).^2,5

In many patients with recurrent obstructive apneas, genioglossus (GG) activity increases prior to airway opening,^6–11 thereby indicating this muscle can respond to increasing chemical drive before arousal occurs. However, the level of chemical drive, relative to eupneic drive, at which pharyngeal dilators begin responding, and the gain of the response beyond this threshold are not known. As indicated in the preceding paragraph, these attributes of the response are critical to understanding the mechanism of ventilatory instability in OSA.

The primary objectives of this study were to determine: (1) How much chemical drive must increase before the GG begins responding to further increases in chemical drive (GG recruitment threshold) and the extent to which this threshold varies among individuals with a collapsible upper airway. (2) Whether variability in GG recruitment threshold arises because of differences in central response to increasing drive (mechanism A) or to differences in mechanoreceptor-mediated activation (mechanism B).

To achieve these objectives, we determined the response of GG activity to different levels of chemical stimulation in patients with OSA. The increase in excitability of the respiratory centers related to hypoxia and hypercapnia (chemical drive) was evaluated from the change in respiratory pump muscle output, as reflected in unloaded minute ventilation (i.e., while the patient was on optimal continuous positive airway pressure [CPAP]). The contribution of central mechanisms to GG responses (mechanism A above) was evaluated from the change in GG activity as ventilation increased on CPAP. In addition, CPAP was dialed-down to 1 cm H₂O after chemical drive was increased to different levels on optimal CPAP, thereby inducing an obstructive hypopnea or apnea. The change in GG activity between the last breath on optimal CPAP (with open airway), and the first (obstructed) breath after CPAP dial-down was used to evaluate the contribution of mechanoreceptor feedback at different levels of chemical drive. The results indicate that the GG response to increasing chemical drive in OSA patients is almost exclusively mediated by mechanoreceptor feedback and that, in most patients, a threshold increase in chemical drive must occur before the mechanoreceptor pathway becomes active. The threshold varies greatly among patients. The latter finding provides a partial explanation for the highly variable relation between the extent of abnormality in pharyngeal mechanics and the polysomnographic severity of OSA (e.g., AHI).

METHODS

The methods were similar to those used in previous studies from this laboratory,^5,12 with the addition of genioglossus recordings.

Patients referred to the sleep center for evaluation of OSA were screened with ambulatory monitoring (Remmers Sleep Recorder Model 4.2, Saga Tech Electronic, Calgary, Canada).^13,14 Patients who had a respiratory disturbance index > 15/h, and who did not have any exclusion criteria were invited to participate. Exclusion criteria included significant comorbidities (dialysis-dependent renal failure, congestive heart failure, severe COPD, previous stroke), obesity-hypoventilation syndrome, pregnancy, use of sedatives or antidepressants, or use of antiplatelet or anticoagulant medication or medications that reduce the metabolism of zopiclone. The Conjoint Health Research Ethics Board at the University of Calgary approved the study protocol, and all subjects gave written informed consent to participate. Twenty patients were studied.

Patients underwent attended polysomnography in the sleep laboratory on 2 separate nights, a diagnostic study and the research study. With 3 exceptions (17, 27, and 65 days), the research study was done within 10 days of the diagnostic study (4.5 ± 2.9 days). Patients did not receive treatment for OSA during the interval.

Diagnostic Polysomnography

The diagnosis and severity of OSA was confirmed by overnight, attended polysomnography using standard polysomnography equipment.¹² Registered technologists scored respiratory events using the Chicago criteria.¹⁵ Sleep and arousals were scored using standard criteria.^16,17 AHI, average and minimum O₂ saturation, and number of respiratory events with arousal were calculated for the supine and lateral positions during NREM sleep.

Research Study

The monitoring was identical to that used for diagnostic polysomnography with the addition of GG monitoring and the dial-down set-up that has been used in previous studies.^5,12 Genioglossus electrodes were inserted to measure GG activity (EMG),^18,19 see acknowledgement). Two sterile stainless steel, Teflon coated, 30-gauge wires, threaded through 25-gauge needles, were inserted into the floor of the mouth, one on each side, 3 mm from the midline. The needles were advanced approximately 15 mm below the surface and were subsequently withdrawn. The external ends of the wires were connected to an amplifier (Grass, Quincy, MA) with a common ground. The raw signal was band-pass filtered (30–500 Hz) and displayed on the computer screen. The patient was asked to perform several maneuvers to determine maximum genioglossus activity (GG_MAX). These included tongue protrusion, swallowing and maximally pushing the tongue against the front upper or lower teeth.

CPAP was then applied via nasal mask connected to a special ventilator, described previously,^4,20,21 which allowed reduction of CPAP to 1.0 cm H₂O. Flow was recorded from a pneumotachograph in the hose of the ventilator and mask pressure was recorded from a side port in the mask. Airway CO₂ was monitored at the nares (AEI Technologies, Pittsburgh, PA).

The GG EMG was sampled at 1000 Hz, while all other signals were sampled at 120 Hz. The signals were recorded on a Windaq data acquisition system (DATAQ Instruments, Akron, OH).

Procedure

Patients were studied supine except for one who could not sleep supine. Following sleep onset, CPAP was titrated to correct flow limitation and snoring. If flow limitation appeared during ventilatory stimulation, CPAP was increased to correct this. The following interventions were carried out during stable NREM sleep:

Intervention 1

Three-breath dial-downs of CPAP to different pressures. These maneuvers were done to determine the near-passive mechanical properties of the pharynx.^22–24

Intervention 2

Three-breath dial-downs of CPAP to 1 cm H₂O during air breathing (Figure 1A) and following ≈ 30 sec of breathing different gas mixtures to stimulate breathing to different levels before the dial-down (Figure 1B). Dial-downs were always applied at the beginning of the expiratory phase (Figure 1). We have found in several previous studies using this approach^4,5,12,20 that there is no appreciable further reduction in end expiratory level of the Respitrace tracings beyond the volume reached during the expiratory phase in which the dial-down was initiated. In each patient, hypoxic and normoxic gas mixtures were used. The normoxic mixtures contained different levels of inspired CO₂ in room air. The hypoxic mixtures contained different levels of CO₂ in 14% O₂. A range of CO₂ concentration was used with both types of gas mixtures, with this range being determined in each patient by the highest CO₂ level that could be tolerated for 30 sec without arousal in most trials. Changes in inspired gas mixture were made by use of an apparatus, described in detail previously,⁵ that allowed very rapid changes in inspired O₂ (between 11% and 21%) and CO₂ (between 0% and 10%).

Figure 1.

Tracings illustrating the methods used and an example of the response to increasing chemical drive on continuous positive airway pressure (CPAP) and immediately following induced obstruction. (A) Patient breathing room air. CPAP was reduced to 1.0 cm H₂O (dial-down) inducing a severe hypopnea (arrow in flow tracing). Note that there was no increase in genioglossus activity during the obstructed breath. (B) In the same patient, inspired CO₂ was increased for 30 sec prior to dial-down. Note that genioglossus activity increased little on CPAP despite doubling of ventilation (V_E). However, there was a large increase in genioglossus activity following dial-down. C3/A2: Electroencephalogram. P_ETCO₂: Airway PCO₂. MA-GG: moving average of genioglossus activity, expressed as percent of maximum activity.

Intervention 3

Three-breath dial-downs of CPAP to 1 cm H₂O following 2 min of breathing hypercapnic, normoxic mixtures (1%–3% CO₂ in air). These longer exposures to CO₂ in air prior to the dial-downs were intended to contrast GG responses when ventilatory stimulation is mediated primarily by central chemoreceptors (2-min exposures) with the responses mediated primarily by peripheral chemoreceptors (30-sec exposures; intervention #2 above).

The short and long gas exposures with the different gas mixtures were performed in a certain sequence. For example, a sequence would consist of 3% CO₂ in air, 6% CO₂ in air, 3% CO₂ in 14% O₂, each for 30 sec, followed by 1%, 2%, and 3% CO₂ in air (each for 2 min). The highest (most stimulating) mixture reached by each patient was determined by the patient's tolerance (ability to remain asleep). Once the appropriate sequence was established for the patient, it was repeated throughout the study until “lights on.”

In earlier experiments using this type of ventilatory stimulation we found that many patients aroused easily,^5,12 thereby seriously limiting the range of chemical drive that could be examined, and some patients remained awake for prolonged periods. In the current protocol, 10 mg zopiclone was administered at 01:00. To the extent tolerated, gas challenges were carried out without zopiclone until 01:00. Zopiclone was then administered whether these challenges were tolerated or not and the challenges resumed thereafter. This approach was chosen to avoid having 2 types of studies (one on and one off zopiclone) with no way to assess the effect of this drug on the relevant responses. The independent effect of zopiclone on the outcome measures was assessed by comparing responses before and after zopiclone in patients who tolerated the interventions before and after the drug was administered.

Analysis

The flow signal was corrected for leaks as described previously.⁵ Leak levels were subtracted from the flow signal to obtain patient flow. Corrected flow was integrated to obtain volume changes.

The raw GG signal was rectified, and 2 moving averages were obtained (200 and 500 ms). The 200 ms moving average was used except where the signal was choppy (usually at low levels of activity). The highest moving average value associated with maximum voluntary maneuvers done at the beginning of the study was noted (GG_MAX).

Determination of closing pressure (P_CLOSE) from data collected in Intervention 1

Breaths with flow limitation during the dial-downs were identified by their characteristic flow contour.^25–27 Maximum flow during flow-limited breaths was plotted against the corresponding dial-down pressure and a linear regression was performed.^20,22,24 Regression analysis was performed from data obtained during the second breath of the dial-down to minimize the effect of viscoelastic behavior of the pharynx on the P_CLOSE estimate.^20,22 The intercept on the pressure axis is reported as P_CLOSE. When dial-down to 1 cm H₂O was associated with a hypopnea instead of apnea, P_CLOSE was determined from back extrapolation of the regression equation. In all cases where back extrapolation was used the upper limit of the confidence interval for the intercept (i.e., estimated P_CLOSE) was within 2.0 cm H₂O of the intercept.

Ventilatory and GG responses to increasing chemical drive (Interventions 2 and 3)

Using standard criteria,¹⁶ an experienced technologist (MO) determined whether a cortical arousal occurred during gas administration or during the first dial-down breath that followed. The 3-sec rule was, however, waived so that briefer high-frequency shifts were considered as arousals. Analysis was limited to data obtained without or preceding cortical arousals.

Minute ventilation (V_E) was calculated breath by breath starting 30 sec before the onset of change in inspired gas (Baseline) and ending with the last breath preceding the dial-down (Last Breath). For each gas challenge, the increase in chemical drive was assessed from the difference between V_E in the Last Breath and baseline V_E (Figure 1B), and is expressed as % baseline. Maximum flow during the first dial-down breath was also recorded.

Breath by breath GG activity was assessed over the same intervals as ventilation. For each breath, the lowest GG EMG value preceding the inspiratory phase (tonic activity) and the highest value during the inspiratory phase (peak activity) were measured (Figure 1), and the difference between the 2 values was calculated (phasic activity). For each intervention, we obtained tonic, peak, and phasic GG activities at baseline V_E, at the Last Breath on CPAP following a change in inspired gas, and at the first breath following dial-downs (Figure 1B). All values were expressed as % of GG_MAX.

In each patient there was a range of Last Breath ventilation values accumulated over the entire study. Interventions were stratified into 4–6 bins with the first bin containing all observations made on air breathing (eupneic drive) and the last bin containing the 2 or 3 interventions with the highest ventilation in the Last Breath. Average Last Breath ventilation within each bin was calculated. The corresponding averages of tonic, peak, and phasic GG activities in the Last Breaths and in breath 1 of the dial-downs were also calculated. Thus, for each patient, GG activity was available at 4 to 6 drive levels on CPAP (Last Breath) and during breath 1 of the dial-down. Differences in GG activity between the Last Breath on CPAP and Baseline activity on CPAP were attributed to the central effects of increased chemical drive. Differences between GG activity during the Last Breath and the immediately following obstructed breath (B1 of the dial-downs) were attributed to mechanoreceptor feedback. In this fashion the contributions of central mechanisms and mechanoreceptor feedback to GG activity could be assessed at different levels of chemical drive.

The difference between responses during the 30-sec and 2-min gas exposures was assessed as follows: The most stimulating gas mixture used in the 2-min exposures was identified. Average V_E and GG activity during the Last Breaths of all exposures to this mixture were computed. GG activity so calculated was compared to GG activity at the same V_E obtained with the 30-sec exposure. Likewise, the difference between responses to hypoxic/hypercapnic mixtures and to pure hypercapnic stimulation was assessed by comparing GG activity during the Last Breath on CPAP and the first breath of the dial-downs at the highest V_E common to both types of stimuli.

The effect of zopiclone on ventilatory and GG responses was assessed as follows: Baseline V_E and GG activity at eupneic breathing (on CPAP) before zopiclone was compared with the corresponding baseline values after zopiclone. In patients who tolerated gas challenges before zopiclone, V_E and GG activity during the Last Breaths were calculated for exposures to the most stimulating gas before zopiclone. These values were compared to ventilation and GG activity during exposures to the same gas mixtures after zopiclone.

All values are expressed as mean ± SD, unless otherwise indicated. T test, paired t-test and analysis of variance (ANOVA) were used, as appropriate, for comparing averages. Analysis ToolPack (Microsoft Excel) was used.

RESULTS

We studied 20 patients (13 males) who varied widely in their age (28–60 years), body mass index (25.0–43.8), and apnea-hypopnea index (9–133/h) (Table 1). P_CLOSE ranged from −10.0 to 4.5 cm H₂O (−1.0 ± 4.5 cm H₂O). There was a weak correlation between P_CLOSE and the AHI (Figure 2). Maximum GG activity (GG_MAX) was 0.27 ± 0.10 volts. Tonic GG activity on optimal CPAP ranged from 0.9% to 6.1% GG_MAX. Phasic GG activity on optimal CPAP was not discernible in 11 patients and ranged from 0.8% to 4.1% GG_MAX in the others.

Table 1.

Subject characteristics

Age (years)	41.9 ± 7.2
M/F	13/7
Body mass index	30.7 ± 4.3
Apnea-hypopnea index (/h)*	73 ± 43
Average SpO2 (%)*	92.9 ± 2.0
% time SpO2 < 90%*	8.3 ± 8.4
Optimal CPAP (cm H2O)	11.1 ± 2.4
PCLOSE (cm H2O)	−1.0 ± 4.5
Flow at dial-down (%peak on CPAP)	23 ± 24
GGMAX (Volts)	0.27 ± 0.10
Baseline VE (L/min)	6.1 ± 1.4
Tonic GG activity at Baseline VE (%GGMAX)	2.6 ± 1.4
Phasic GG activity at Baseline VE (%GGMAX)	0.9 ± 1.0

CPAP, continuous positive airway pressure; GG, genioglossus; GG_MAX, maximum genioglossus activity; P_CLOSE, closing pressure; SpO₂, oxyhemoglobin saturation; V_E, minute ventilation on CPAP.

^*Polysomnography data are those observed during NREM sleep (stages 1–4) in the body position used in the dial-down study (19 supine, 1 lateral).

Figure 2.

Relation between critical closing pressure (P_CLOSE) and the apnea-hypopnea index in 20 patients. Open circles: patients with type A response. Shaded circles: patients with type B response. Closed circles: patients with type C response. Open triangles: patients in whom the response type could not be determined. See text or Figure 5 for explanation of response types.

Response of Tonic Genioglossus Activity to Increasing Chemical Drive

The points at which this tonic activity was measured on CPAP and in breath 1 of dial-down are shown in Figure 1 (Tonic, Last Breath and Tonic B1). The top 2 panels of Figure 3 show the individual responses (n = 20) when patients were on optimal CPAP (panel A) and immediately preceding breath 1 of the dial down to (panel B). Reflecting the differences in arousal threshold among patients, the range of chemical drive over which these responses could be documented varied considerably, from as little as 4 to as much as 22 L/min increase in ventilation (abscissa, Figure 3). With the exception of one patient in whom tonic activity increased a modest amount when ventilation exceeded 20 L/min, tonic activity on optimal CPAP remained essentially constant over the entire range of chemical drive (Figure 3A). In the vast majority of patients, tonic activity of dial-down breath 1 also did not increase as chemical drive preceding the dial-down increased (Figure 3B).

Figure 3.

Response of tonic and phasic genioglossus activity in individual patients to increasing chemical drive while on continuous positive airway pressure (left panels) and during the first obstructed breath.

The left panel of Figure 4 shows the average responses of tonic activity in the 20 patients. These average responses were obtained by averaging GG activity and minute ventilation during room air breathing (leftmost point), at the highest V_E reached without arousal (rightmost point) and at the midpoint of the ventilatory range (middle point). On optimal CPAP (solid symbols), tonic activity was very slightly but significantly higher at the highest chemical drive than at baseline (2.9% vs. 2.6%GG_MAX, P < 0.05). At the highest chemical drive reached, tonic activity immediately preceding breath 1 of the dial-down was slightly greater than in the Last Breath before dial-down (3.2% vs. 2.9%GG_MAX, P < 0.05).

Figure 4.

Average responses of tonic and phasic genioglossus activity on CPAP (solid lines) and during the first obstructed breath (dashed lines). +P < 0.05 compared to baseline values on CPAP. *P < 0.05, **P < 0.005, ***P < 0.0005 compared to CPAP at the same chemical drive.

Response of Phasic Genioglossus Activity to Increasing Chemical Drive

Figures 3C and 3D show individual responses in the 20 patients. In 17 patients, phasic activity did not increase or increased only slightly when chemical drive increased while the patient was on optimal CPAP (Figure 3C). The average (±) increase in these patients across the entire chemical drive range was 0.50% ± 0.37%GG_MAX. In the remaining 3 patients, phasic activity on optimal CPAP responded to chemical drive at an appreciable rate. In 2 of these 3 patients, the high gain response began only after ventilation reached a threshold level. By contrast, phasic activity during the first obstructed breath (dial-down B1) increased substantially as chemical drive increased in more than half the patients (Figure 3D).

Figure 4B shows the average responses on optimal CPAP (solid symbols) and during the first obstructed breath (open symbols). The difference between the 2 lines reflects the average contribution of mechanoreceptor feedback associated with the obstruction. The increase in phasic activity on optimal CPAP was marginally significant at the highest level of stimulation (2.6% ± 4.2% vs. 0.9% ± 1.0%GG_MAX, P < 0.05). At baseline chemical drive there was very little increase in phasic activity during the first obstructed breath but the difference was significant (1.3% ± 1.4% vs. 0.9% ± 1.0%GG_MAX, P < 0.05). The increase in phasic activity during the first obstructed breath, relative to the last unobstructed breath, increased progressively as chemical drive increased. At the highest chemical drive (average V_E = 16.2 L/min), the values were 6.0% ± 6.9%GG_MAX during the first obstructed breath and 2.6% ± 4.2% GG_MAX in the immediately preceding unobstructed breath (P < 0.0005).

Impact of Mechanoreceptor Feedback on Genioglossus Activity in Individual Patients

In 3 patients, with highly negative P_CLOSE (−6.9, −7.8, −8.6 cm H₂O), there was only mild hypopnea upon dial-down from air breathing and all signs of obstruction disappeared when dial-down was preceded by very modest increases in V_E (+37% baseline V_E). The increase in GG activity at these levels of V_E was minimal (0.3%, 0.2%, and 0% GG_MAX). Accordingly, these patients could not be used to assess the effect of increasing chemical drive on the GG response to obstruction. These 3 patients had mild OSA (AHI: 11, 11, and 21/h; open triangles, Figure 2).

Figure 5 illustrates the range of immediate GG responses to obstruction in the remaining 17 patients. In 6 patients, the difference between the first obstructed (open symbols) and the last unobstructed (solid symbols) breaths increased progressively throughout the entire range of ventilatory stimulation, thereby indicating that mechanoreceptor feedback is effective with minimal increases in chemical drive (type A response). Two examples of this low threshold response are shown in Figures 5A and 5D. In 6 other patients, there was no increase in GG activity during the first obstructed breath until chemical drive exceeded a finite threshold (type B response). Figures 5B and 5E illustrate 2 examples, one in whom the threshold was only slightly above baseline ventilation (panel 5B) and another where the threshold was much higher (panel E). In the remaining 5 patients, there was no increase in GG activity during the first obstructed breath across the whole range of chemical drive examined (i.e., below the arousal threshold) (type C response). Figures 5C and 5F illustrate 2 such examples.

Figure 5.

Response of phasic genioglossus activity to increasing chemical drive on CPAP (solid lines) and during the first obstructed breath in 6 patients representing the 3 response types. Type A response (panels Aand D): mechanoreceptor effect (difference between the 2 lines) is evident beginning with baseline (lowest) drive. Type B response (panels Band E): mechanoreceptor effect appears only after a threshold increase in chemical drive. Type C response (panels Cand F): No mechanoreceptor effect across the entire range of chemical drive examined.

Table 2 shows some characteristics of the 3 groups of patients. There were no differences in age, gender, body mass index, baseline ventilation on CPAP, or maximum GG activity. All type A patients and all but one of type B patients had a P_CLOSE > 0. These 2 groups developed complete obstruction or severe hypopnea on dial-down from air breathing. There was no significant difference in P_CLOSE between these 2 groups (P = 0.24). Type C patients were a mixed group. In 3 patients P_CLOSE was negative (−2.8, −5.5, −10.0 cm H₂O), and the patients developed only mild hypopnea upon dial-down from air breathing. The average data for one such patient are shown in Figure 5C, and representative tracings are shown in Figure 6. The other 2 patients had positive P_CLOSE and complete obstruction upon dial-down. GG activity in these 2 patients did not increase despite ≈ 3-fold increase in respiratory drive prior to dial-down. One example is shown in panel F, Figure 5. Notwithstanding the severity of their mechanical abnormality, 3 of the 6 type A patients had mild OSA (AHI: 9, 19, and 25/h, open circles, Figure 2). By contrast, all patients with types B and C responses had severe OSA (AHI > 40), including the 3 type C patients with negative P_CLOSE (AHI: 68, 73, and 108/h, solid circles, Figure 2).

Table 2.

Characteristics of 3 types of responders to mechanoreceptor feedback

	Type A	Type B	Type C	P value (ANOVA)
Number of patients	6	6	5
Age (years)	45.0 (7.8)	42.2 (7.5)	42.8 (4.1)	0.76
Gender	4M/2F	4M/2F	3M/2F
Body mass index (kg/m2)	32.1 (6.1)	30.0 (3.1)	29.1 (1.2)	0.49
PCLOSE (cm H2O)	2.0 (1.7)	1.2 (2.1)	−3.1 (4.9)	0.04
Range	0.0 to 3.8	−1.0 to 4.5	−10.0 to 2.5
Dial-down flow on air (% Flow on CPAP)	8 (9)	9 (8)	34 (26)	0.03
Range	0 to 20	0 to 21	0 to 61
Baseline ventilation on CPAP (L/min)	6.5 (2.3)	5.9 (0.8)	6.1 (1.0)	0.80
GGMAX (volts)	0.23 (0.12)	0.20 (0.08)	0.24 (0.07)	0.73
Tonic GG at baseline VE on CPAP (%GGMAX)	3.0 (1.3)	2.9 (0.8)	1.4 (0.7)	0.03
Phasic GG at baseline VE on CPAP (%GGMAX)	1.4 (1.4)	1.0 (1.0)	0.4 (0.6)	0.40
Increase in phasic GG @ twice baseline VE on CPAP (%GGMAX)	1.5 (1.5)	0.5 (0.5)	0.1 (0.1)	0.05
Apnea Hypopnea Index (/h)	61 (48)	109 (19)	79 (28)	0.08
Range	9-108	81-133	42-108

CPAP, continuous positive airway pressure; GG, genioglossus; GG_MAX, maximum genioglossus activity; P_CLOSE, closing pressure; V_E, minute ventilation.

Figure 6.

Tracing from the patient illustrated in Figure 5C. Legend is as in Figure 1. These tracings show failure of genioglossus activity to increase appreciably during the first obstructed breath despite a 3-fold increase in ventilation prior to dial-down (compare last breath in panels A and B). This patient had only mild hypopnea during dial-down from air breathing (A), indicating mild abnormalities in pharyngeal mechanics (P_CLOSE was −10 cm H₂O). His apnea-hypopnea index was 68/h (Figure 2).

The response of GG activity to further increases in respiratory drive beyond the threshold level was, on average, quite brisk (1.1% ± 0.84%GG_MAX per L/min increase in V_E). The range was, however, wide (0.3% to 3.4%GG_MAX per L/min increase in V_E).

Impact of Duration and Type of Chemical Stimulation

Minute ventilation reached at the Last Breath during the 2-min exposures to 3% CO₂ in air averaged 11.5 ± 2.6 L/min. In each patient peak GG activity obtained at the Last Breath and during the first obstructed breath following these 2-min exposures was compared to peak GG activity when the same level of ventilation was reached through shorter (30-sec) exposures. The results are shown in Figure 7A. At similar minute ventilation, peak GG activity was not different during the Last Breath or during the first obstructed breath whether ventilatory stimulation was achieved by long or short exposures inspired gas mixtures. The change in peak GG activity from the Last Breath to the first obstructed breath was also not different (1.9% ± 2.9% vs. 2.4% ± 4.4% GG_MAX, respectively, for 2-min and 30-sec exposures; P > 0.1).

Figure 7.

(A) Comparison of genioglossus responses at the same ventilation when chemical stimulation was produced by 30-sec vs. 2-min exposure to stimulating inspired gases. (B) Comparison of genioglossus responses at the same ventilation when chemical stimulation was produced by hypercapnia alone (solid circles) and by a combination of hypoxia and hypercapnia (open circles). Bars are SEM.

The highest ventilation common to both hypoxic/hypercapnic and pure hypercapnic stimulation was 13.2 ± 3.6 L/min, approximately twice baseline ventilation on air (6.2 ± 1.4 L/min). Peak GG activity was not significantly different between the 2 types of stimuli during the last breath on CPAP (4.5% ± 3.4% vs. 4.6% ± 3.7%GG_MAX, P = 0.47; Figure 7B), with both values representing an increase of 1% GG_MAX relative to air breathing. The increase in GG activity during the first obstructed breath was slightly but not significantly higher during the hypoxic/hypercapnic stimulation (ΔGG 3.5% ± 5.7% vs. 2.6% ± 3.4%GG_MAX, P = 0.10; Figure 7B).

Effect of Zopiclone on Ventilatory and Genioglossus Responses

Seven patients aroused very easily with ventilatory stimulation prior to zopiclone administration, and no useful data could be obtained until after the drug was administered. Four other patients tolerated the inspired gas mixtures on CPAP but aroused immediately upon dial-down. Thus, no dial-down data were available. In the remaining 9 patients, adequate data were available before and after zopiclone was administered. In these 9 patients, we compared minute ventilation and GG activity obtained with the most stimulating gas mixture tolerated before zopiclone with the responses to the same gas mixture after zopiclone. No significant differences were observed between any of these responses before and after zopiclone (Table 3).

Table 3.

Effect of zopiclone on ventilatory and GG responses (n = 9)

	Pre-zopiclone	Post-zopiclone
On room air
Minute ventilation (L/min)	5.6 (1.0)	5.6 (0.9)
Peak GG on CPAP (%GGMAX)	3.8 (2.5)	3.9 (3.0)
ΔGG on DD (%GGMAX)	0.1 (0.3)	0.3 (0.4)
On strongest matched inspired gas
Minute ventilation (L/min)	13.2 (4.3)	13.4 (3.6)
Peak GG on CPAP (%GGMAX)	5.2 (4.4)	5.8 (5.9)
ΔGG on DD (%GGMAX)	5.3 (6.6)	5.0 (9.4)

GG, genioglossus activity; CPAP, continuous positive airway pressure; GG_MAX, maximum voluntary genioglossus activity; DD, dial-down of CPAP; ΔGG on DD, the increase in peak GG activity during the first obstructed breath.

DISCUSSION

The main findings from the present study are: (A) The increase in GG activity in response to increases in chemical drive in sleeping OSA patients is predominantly mediated by mechanoreceptor feedback. (B) There is virtually no immediate increase in GG activity in response to upper airway obstruction at resting (eupneic) levels of chemical drive. (C) Recruitment of GG activity requires the presence of upper airway obstruction in the presence of increased (above eupnea) chemical drive. (D) The increase in chemical drive required before there is a response to upper airway obstruction varies greatly among patients. (E) Genioglossus responses to increased chemical drive are not different whether the increase in drive is produced by brief (30-sec) or long (2-min) exposures to stimulating inspiratory gases and whether the stimulus is pure hypercapnia or a combination of hypoxia and hypercapnia. (F) The responses described in A-E above are minimally altered by 10 mg of zopiclone.

Genioglossus Responses on Optimal CPAP

Although, on average, tonic and phasic GG activity increased significantly with the increase in chemical drive on CPAP (solid lines, Figure 4), the increase was modest, amounting to < 1% GG_MAX, over the entire ventilatory range in 17 of the 20 patients (Figure 3, left panels). In 2 of the 3 responders, the response was not appreciable until ventilation increased to 3- or 4-fold baseline ventilation (Figure 3C). These responses contrast with the steep responses observed in many of the same patients when the same increase in chemical drive was associated with upper airway obstruction (Figure 3D). Even in the 3 patients in whom a substantial response occurred on optimal CPAP, some flow limitation during these high levels of ventilation, resulting in negative pharyngeal pressure, could not be ruled out. Thus, it may be concluded that when GG activity increases rapidly during obstructed breaths in association with increasing chemical drive, the increase is predominantly mediated through mechanoreceptor feedback. This conclusion is supported by the observation that topical anesthesia of the upper airway of OSA patients attenuates the progressive increase in GG activity that occurs in the course of their spontaneous obstructive events.¹¹

It is not clear whether the weak response observed here in OSA patients on optimal CPAP is abnormal. There is only one previous study in which GG response to increased chemical drive was measured on CPAP in normal subjects.¹⁸ The responses observed in that study were considerably larger than what was observed here. Taken at face value, this comparison suggests that GG responses on optimal CPAP are abnormally low in OSA patients. However, caution must be exercised since there were important technical differences that could account for the different results. Most importantly, we titrated CPAP to eliminate all snoring and flow limitation, even at the highest levels of ventilation, whereas in the study of Lo et al., CPAP was set to the level that resulted in the lowest GG activity during wakefulness, with a maximum of 8 cm H₂O.¹⁸ Thus, it is possible that the higher responses in the latter study were due to the presence of flow limitation on CPAP during high levels of ventilation. Another possibility is that the denominator of the reported responses (i.e., GG_MAX) may have been different in the two studies, either because the technique used to obtain maximum activity was different or because OSA patients are capable of generating a greater GG_MAX than normal subjects. In such case, the same absolute increase in GG activity would appear to be lower when expressed as % GG_MAX. Further studies, in which the same methods are used for OSA patients and normal subjects, are needed to address this issue.

Chemical Drive and Mechanoreceptor Feedback

Negative pressure pulses applied to the pharynx during wakefulness result in reflex activation of the GG.^28–32 The response is attenuated during sleep.^28,33 Several studies found no immediate (first breath) increase in GG activity when an obstructive hypopnea or apnea is induced by suddenly reducing CPAP (i.e., dial-down, as done here) at eupneic ventilation.^34,35 This indicates that the negative pressure associated with eupneic drive is insufficient to increase GG activity.² The current study supports this observation since there was minimal increase in GG activity during B1 of dial-downs from air breathing (Figures 4 and 5). Thus, pharyngeal pressure during obstruction must become more negative than its eupneic level in order for this mechanoreceptor feedback to be effective. Once an obstruction develops, the only way by which pharyngeal pressure can become more negative is for chemical drive (and hence pump muscle activity) to increase.² Our study is the first to determine how much chemical drive must increase in order for the negative pharyngeal pressure to reach levels that reflexly activate the GG. Figure 5 shows that in OSA patients the required increase in chemical drive (mechanoreceptor response threshold)varies greatly among patients, from just above eupnea to greater than three-fold eupneic levels. It is important to note that the threshold referred to here is not a mechanical (pharyngeal pressure) threshold but a chemical drive threshold below which the mechanoreceptor response is not active.

The reason why mechanoreceptor response threshold varies so much among patients is not clear. Broadly, this may be due to (A) differences in the negative pressure generated in the pharynx at the same chemical drive and/or (B) differences in the central mechanisms that mediate the mechanoreceptor response. Our study was not designed to address the mechanisms of the differences in this mechanoreceptor response threshold but, rather, to document what the threshold is in different patients in view of the intrinsic importance of this threshold (see Clinical Relevance, below). Distinction between these two possibilities requires measurement of pharyngeal pressure at different chemical drives, which we did not do. Nonetheless, it is not likely that the different thresholds are related exclusively to different levels of negative pharyngeal pressure in the first obstructed breath. All three types of responders (A, B, and C, Figure 5) included patients in whom P_CLOSE was > 1 and, hence, developed complete obstruction during the dial down. Since, by definition, there is no flow during the obstructed breaths, the entire pressure generated by pump muscles in such patients is transmitted to the pharynx during the first, obstructed dial-down breath. For differences in mechanoreceptor response threshold among such patients (with P_CLOSE > 1) to be exclusively due to differences in the magnitude of negative pharyngeal pressure, pump muscle pressure would have to be less at three-times respiratory drive in type C patients than the pressure generated at eupneic drive in type A responders. This, we believe, is unlikely particularly since patients with the different response types did not differ in age, gender, body mass index, or baseline V_E on CPAP.

That mechanoreceptor feedback during sleep becomes effective only above a certain chemical drive is probably not limited to OSA patients. A number of studies have shown that the immediate GG response to external airway obstruction⁸ or to externally applied inspiratory resistors^19,34 in normal subjects is small and not present in all subjects.

Beyond the threshold V_E value, the difference in GG activity between the last breath on CPAP and the first obstructed breath increased progressively as a function of increasing chemical drive (Figures 4 and 5). It is not clear whether this simply reflects the greater negative pressure associated with the increased drive or is in part related to heightened sensitivity of central mechanisms to peripheral mechanoreceptor input. Further studies, in which the response to fixed pressure stimuli is measured at different levels of chemical drive, are needed to address this issue.

Our study has several advantages and limitations that should be addressed. The main advantages are:

• 1) This is the first study to quantify GG responses to measurable increases in chemical drive in patients with OSA during sleep. Most previous studies on GG responses were either on normal subjects,^{8,19,28,29,33,36–39} or OSA patients under conditions where the level of chemical drive was unknown.^6,7,9–11

• 2) The chemical stimulation used included hypoxia and/or hypercapnia that evolved over short periods (30-sec). Unlike the steady increases in chemical drive used in earlier studies,^18,37,38 this pattern, which is largely mediated by peripheral chemoreceptors, is more representative of the pattern of stimulation encountered during spontaneous obstructive events in patients. The current study is the first to demonstrate that GG responses are linked to the level of stimulation reached regardless of the duration or type (hypoxia vs. hypoxia plus hypercapnia) of stimulus.

• 3) Our approach, in which GG responses were measured on optimal CPAP and immediately following the induction of upper airway obstruction, made it possible to determine the relative importance of mechanoreceptor feedback to the increase in GG activity observed during obstructive events in OSA patients.

On the other hand, this study has some shortcomings:

• (1) Since the majority of the data were obtained after zopiclone was administered, it may be argued that the responses do not represent what happens in unmedicated patients. It was reassuring, however, to find that there were no differences in GG responses before and after zopiclone in 9 patients in whom adequate data were available (Table 2).

• (2) Many of the chemical challenges we used were brief, and the ventilatory response had not completely stabilized before the dial-down was performed. Thus, it may be argued that the increase in GG activity during the first obstructed breath was the result of a further increase in chemical drive, between the Last breath on CPAP and the first obstructed breath, rather than being the result of obstruction-related mechanoreceptor feedback. This is unlikely since the increase during the first obstructed breath following brief gas exposures was not different from the increase following long exposures, when ventilation had stabilized long before the dial-down (Figure 7).

Clinical Relevance

Several previous studies have demonstrated that the severity of OSA, as expressed by the AHI, is weakly correlated with the severity of the pharyngeal mechanical abnormality, as expressed by P_CLOSE (or P_CRIT).^20,40–42 This indicates that control mechanisms play a dominant role in determining whether the obstructive event will evolve into stable breathing or will recur, thereby resulting in frequent events and a high AHI.^2,20 The weak correlation between P_CLOSE and AHI found in the current study (Figure 2) is consistent with these previous reports. Thus, some subjects with minimal pharyngeal abnormalities (substantially negative P_CLOSE) had very high AHI, while others with very abnormal mechanics had very low AHI. Our findings expand the current literature by identifying specific control mechanisms that may determine whether the response to abnormal pharyngeal mechanics is stable or not.

Specifically, the findings that mechanoreceptor feedback augments GG activity only above a threshold increase in chemical drive (mechanoreceptor response threshold), that this threshold varies widely among different patients, and that the response of GG activity to increasing chemical drive is modest in the absence of effective mechanoreceptor feedback, have major implications to ventilatory stability in the face of abnormal pharyngeal mechanics. Thus, if the threshold for activating mechanoreceptor feedback is high, only the mildest mechanical abnormality, that which can be overcome by modest activation of the dilators, can be resolved without an excessive increase in chemical drive. In such cases, the small increase in activity that occurs in the absence of effective mechanoreceptor feedback (solid lines, Figure 4) may suffice (e.g., 3 open triangles, Figure 2). In cases where the mechanical abnormality necessitates a more robust increase in dilator activity, no compensation can take place until chemical drive rises at least to the required mechanoreceptor response threshold. A high threshold promotes instability and recurrence of the obstruction, for two important reasons. First, since arousal is directly related to the pressure generated by the respiratory muscles,^43–45 the more chemical drive and effort must increase before the airway opens reflexly, the more likely arousal will occur first. The occurrence of arousal in the setting of obstructive events is quite destabilizing since arousals are associated with a large ventilatory overshoot that washes out the chemical stimulus required for stable compensation, thereby promoting re-obstruction.⁴ Second, if the threshold is high, a substantial overshoot is more likely to occur even in the absence of arousal since intrathoracic pressure at the time of opening will be more negative.²

The importance of the mechanoreceptor response threshold to ventilatory instability was illustrated in this study by three patients who had very mild mechanical abnormalities but whose mechanoreceptor response threshold was greater than the range of chemical drive allowed by the arousal threshold (type C responders). These patients had severe OSA (closed circles, Figure 2). Conversely, three patients with more severe mechanical abnormalities who had a low threshold (type A response) had very low AHI (open circles, Figure 2).

Our data also show that a low mechanoreceptor response threshold does not guarantee a stable response to obstruction or a low AHI. Three of the six patients with low threshold (type A response) had severe OSA (open circles, Figure 2). A plausible explanation is that the mechanoreceptor response threshold is the minimal increase in drive below which there is no dilator response. In order for the airway to open reflexly, chemical drive must rise not only to the threshold level but must further increase until dilator activity is sufficient to overcome the mechanical challenge. This extra increase in drive is determined by the passive mechanical properties of the pharynx, as well as by other factors that affect the GG activation level required to open the airway at the same P_CLOSE. Such factors include the site of obstruction, negative effort dependence, adhesiveness of pharyngeal secretions, and contractility of the dilator muscles.² In addition, the amount of required increase in chemical drive is determined by the slope of the relationship between GG activity and chemical drive above threshold. This slope was found to vary from 0.3% to 3.4% GG_MAX per L/min increase in V_E among patients.

In summary, we found that in OSA patients GG activity responds weakly to increases in chemical drive in the absence of mechanoreceptor feedback, and that mechanoreceptor feedback is not effective until a highly variable (among patients) threshold increase in chemical drive is reached. These findings provide a partial explanation for the poor correlation between the severity of OSA and abnormal pharyngeal mechanics.

DISCLOSURE STATEMENT

This was not an industry supported study. Mrs. Ostrowski has consulted for YRT Ltd. Mr. Laprairie has received remuneration from YRT Ltd., and Respironics, and Dr. Younes is President and CEO of YRT Ltd., and has received honoraria from Respironics and Covidien. He has also served as an expert witness for Respironics. The other authors have indicated no financial conflicts of interest.