INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

THE NEGATIVE EXPIRATORY PRESSURE (NEP) method has been recently introduced to detect intrathoracic expiratory flow limitation (EFL) under different conditions in humans (6, 12). It consists in applying a negative pressure (usual range: 3 to 5 cmH₂O) during tidal expiration, thus widening the pressure difference between the alveoli and the airway opening. In the absence of EFL, there is an increase in expiratory flow, whereas, in the presence of EFL, the expiratory flow during NEP does not increase throughout the entire or part of the tidal expiration compared with the flow of the preceding control expiration. In spontaneously breathing subjects at rest, however, the application of such levels of NEP at the onset of expiration sometimes results in a drop in flow below the flow rate generated during the preceding tidal expiration. This finding, which is believed to reflect partial or total upper airway narrowing (11), may occur in healthy subjects (9, 15), suggesting some degree of upper airway instability.

It has been shown in awake, normal humans that sudden application of NEP at end expiration was associated with reflex activation of genioglossus (4). This phenomenon may be viewed as a compensatory mechanism to maintain upper airway patency.

We reasoned that airway narrowing or collapse during application of NEP at the beginning of expiration could be due to neurophysiological characteristics of the above-mentioned reflex. For instance, its latency could be too long for it to operate after application of NEP in early expiration, or there could be interindividual variations in the sensitivity of the trigger or in the intensity of response. Such reflex elements might be relevant to certain aspects of the assessment of upper airway function in sleep-disordered breathing.

Therefore, the present study was designed to assess and compare the effect of NEP applied at the mouth at the onset of expiration and during the end-expiratory pause on genioglossus muscle reflex activity in normal subjects.

	METHODS

Top Abstract Introduction Methods Results Discussion References

Ten healthy, nonobese subjects without evidence or history of cardiopulmonary disease, obstructive sleep apnea syndrome, or upper airway abnormality were studied (Table 1). Subjects took no medications or alcohol on the day of the study. Informed consent was obtained from each individual.

Flow () was measured with a Hans-Rudolph pneumotachograph with a ±2.6 l/s linearity range (model 4700A, Hans-Rudolph, Kansas City, MO) connected to the mouthpiece and a differential pressure transducer (MP45, ±2 cmH₂O; Validyne, Northridge, CA). Pressure was measured at the mouth (Pmouth) via a noncompliant polyethylene tube (1.7 mm ID), connected to a differential pressure transducer (DP15, ±150 cmH₂O; Validyne). The system used to measure Pmouth had no appreciable shift or alteration in amplitude up to 20 Hz.

A Venturi device capable of rapidly generating a negative pressure (Aeromech Devices, Almonte, ON, Canada) was connected in series with the pneumotachograph. The Venturi device was attached via an electrically operated solenoid valve to a source of compressed air. A pressure regulator between the source of compressed air and the solenoid valve was used to obtain the desired levels of negative pressure at the mouth (range: 3 to 10 cmH₂O). The solenoid valve (Asco electrical valve model 8262G208; Ascolectric, ON, Canada) was driven by a computer (Direc Physiologic Recording System; Raytech Instruments, Vancouver, BC, Canada) and had an opening time of 28 ms. The solenoid valve was activated by computer when expiratory flow reached a threshold level of 20 ml/s with an optional predetermined delay or manually with a remote system during the end-expiratory pause (Fig. 1). The dead space of the equipment assembly was <30 ml, and its pressure (P)- relationship was characterized by the following equation: P = 0.85+0.70², where P is in centimeters of water and is in liters per second. Calibration of the flow and pressure transducers was done before and after each study.

View larger version (21K): [in this window] [in a new window]   Fig. 1.   Experimental setup used to produce sudden negative expiratory pressure (NEP) in a subject breathing through a mouthpiece. Flow (), pressure at the mouth (Pmouth), and raw surface EMG of genioglossus (EMGgg) were simultaneously recorded before and during application of NEP pulses (500 ms), both at onset of tidal expiration and during end-expiratory pause. V, volume; P, pressure.

Surface recordings of the genioglossal electromyogram (EMGgg) were obtained by using a pair of disposable unipolar skin-taped electrodes clipped on the chin and connected to a multichannel recording system (Neuropack Sigma, Nihon Kohden, Tokyo, Japan). The electrode-to-skin impedance was checked before each experiment to keep it always <10 k. The raw EMGgg signal was amplified, band-pass filtered between 100 and 5,000 Hz, full-wave rectified, and electronically integrated by using a time constant of 100 ms (Neuropack Sigma). In two of the subjects, the intraoral EMGgg was also recorded by means of a noninvasive intraoral electrode connected to a mouthpiece and constructed from dental impression material (2). While the impression material was still soft, it was fitted to the subject's lower teeth and pushed onto the floor of the mouth so that the material was closely apposed to the upper surface of the genioglossus. Electrodes were sewn through the impression material and placed parallel and symetrically on each side of the midline. The terminal 5 mm of the electrodes, which were in the contact with the floor of the mouth, were bared. A grounding electrode was placed on the right arm. To compare the individual EMGgg activity obtained at different levels of NEP, we determined the maximal EMGgg activity by performing the following maneuvers: maximal inspiratory effort against occluded airway, sniff, and maximal protrusion of the tongue. In all subjects, the maximal peak integrated EMGgg activity was always achieved during maximal protrusion of the tongue. Once maximum (100%) was defined, the peak of the integrated EMGgg signal could be scaled between 0 (electrical 0) and 100%.

The and Pmouth signals were amplified, low-pass filtered (Demodulator model CD 15, Validyne), and sent as analog signals to Neuropack Sigma. The time course of the raw EMGgg, , and Pmouth signals were continuously displayed on the Neuropack Sigma screen. When required, the signals were recorded and stored in the hard disk of the computer installed on the Neuropack for subsequent analysis. Data analysis was performed by using the managing waveforms software provided with the Neuropack Sigma system (Fig. 1).

The and Pmouth signals were also amplified (AC Bridge Amplifier-ABC module, Raytech Instruments), low-pass filtered at 50 Hz, sent to a 16-bit analog-to-digital converter (Direc Physiologic Recording System, Raytech Instruments) installed on a IBM-compatible computer (486DX, 66 MHz), and sampled at 200 Hz. Both digitized signals were displayed in real time on the computer screen together with the volume (V) signal obtained by numerical integration of the signal. The tracings were continuously monitored both with respect to time and as -V curves (Fig. 1). The recordings were stored on the computer hard disk in Direc format and used for subsequent analysis when necessary. Data analysis was performed by using the Direc data-analysis software (version 3.1, Direc NEP software, Raytech Instruments).

Procedure and data analysis. During the study, the subjects, wearing a noseclip and breathing through a rigid mouthpiece, were sitting upright in a comfortable dentist's chair with the neck fixed in neutral position. After regular quiet breathing had been achieved, NEP values of 3, 5, 7, and 10 cmH₂O were randomly applied 0.2 s after the onset of expiration and maintained for 500 ms. NEP of 10 cmH₂O was also applied at the end-expiratory pause. Each level of NEP was repeated six times. Recordings of Pmouth, , and raw EMGgg (surface or intraoral) were initiated before NEP application and lasted 2 s.

	RESULTS

Top Abstract Introduction Methods Results Discussion References

NEP at onset of expiration. In seven subjects, no detectable reflex surface EMGgg activity was observed after application of all NEP levels, as shown for a representative subject (subject 3) with NEP of 10 cmH₂O (Fig. 2). A consistent reflex response to all levels of NEP was found in only one subject (subject 6), with a mean latency of 57 ± 6 ms with NEP of 10 cmH₂O (Fig. 3). Comparable mean latencies were found with NEP values of 3 (56 ± 9 ms), 5 (56 ± 8 ms) and 7 cmH₂O (59 ± 8 ms) for this subject.

View larger version (23K): [in this window] [in a new window]   Fig. 2.   Records of flow, Pmouth, and surface EMGgg in a representative subject (subject 3) before and during application of NEP (10 cmH2O) at onset of expiration, as shown in volume vs. time graph. No EMGgg reflex activity was elicited by NEP. Also shown for comparison is a record of EMGgg during maximal tongue protrusion of the same subject. Max, maximum.

View larger version (22K): [in this window] [in a new window]   Fig. 3.   Records of flow, Pmouth, and surface EMGgg in subject 6 before and during application of NEP (10 cmH2O) at onset of expiration. Note short latency of activation after pressure change.

View larger version (19K): [in this window] [in a new window]   Fig. 4.   Peak integrated surface EMGgg (EMGi gg) activity in subjects 2, 6, and 9 at all levels of NEP applied at onset of expiration. Values are mean ± SD. * P < 0.001 vs. baseline (NEP = 0 cmH2O).

NEP at end-expiratory pause. When NEP of 10 cmH₂O was applied during the end-expiratory pause, that is, at functional residual capacity before the ensuing inspiration, consistent reflex responses in surface EMGgg activity were found in seven subjects, as exemplified for a representative subject (subject 3) in Fig. 5, and occasionally in three subjects (once in 5 NEP tests). The latency of the reflex of the seven responders amounted to 68 ± 5 ms and was significantly quicker than the reaction time for protruding the tongue (203 ± 5 ms; range 168-228 ms).

View larger version (25K): [in this window] [in a new window]   Fig. 5.   Records of flow, Pmouth, and surface EMGgg in subject 3 before and during application of NEP (10 cmH2O) at end expiration (end-expiratory pause), as shown in volume vs. time graph. Note reflex activation of genioglossus induced by NEP.

NEP at onset of expiration and at end-expiratory pause with intraoral EMGgg electrode. In two of the above subjects (subjects 1 and 3) who did not respond to NEP applied at the onset of expiration, we also repeated the measurements by using intraoral EMGgg electrodes and again found no response to NEP applied at the onset of expiration, as shown in subject 3 (Fig. 6). In contrast, when NEP was applied during the end-expiratory pause, in both subjects there was an increase in EMGgg activity similar to that found with the surface electrodes.

View larger version (23K): [in this window] [in a new window]   Fig. 6.   Records of flow, Pmouth, and raw EMGgg obtained via intraoral electrode in subject of Fig. 2 during application of NEP (10 cmH2O) at onset of expiration. Again no EMGgg reflex activity was observed with NEP. Note mechanical artifact on EMGgg record occurring simultaneously with NEP application and removal.

Relationship between expiratory flow and reflex activation of genioglossus during NEP at onset of expiration. The individual measurements of expiratory flow obtained during application of the different levels of NEP were reported in relation to the genioglossus reflex activation in Table 2. When the values of the mean expiratory flow during the first 150 and 500 ms of NEP were expressed as percent ratio (i.e., 150/500%), two distinct patterns emerged. The former, characterized by a higher 150/500% ratio, was displayed by subjects 2, 3, 7, and 9 and by subject 6 and documented a stable increase in expiratory flow; the latter, which had a lower 150/500% ratio, was observed in subjects 1, 4, 5, 8, and 10 and was determined by a substantial, early decrease in expiratory flow. Thus, in the absence of a reflex response of genioglossus to NEP applied at the beginning of expiration, both patterns in the expiratory flow could be observed, likely reflecting a sustained patency or a transient narrowing of the upper airways in response to NEP, respectively. In Fig. 7, examples were given of -V curves obtained in three representative subjects when NEP was applied at the onset of expiration. Despite the absence of a reflex-mediated activation of genioglossus, subject 3 (left panels) exhibited a steady increase in expiratory flow during NEP in each test, as did subjects 2, 7, and 9. In contrast, subject 4 (middle panels) showed almost at all times an abrupt fall in expiratory flow soon after the application of NEP, which was generally more marked at higher levels of NEP, as did subjects 8 and 10. Subjects 1 and 5 behaved similarly to subject 4, although in a less regular fashion. Finally, subject 6 (right panels), who had an invariable EMGgg reflex activity in response to NEP, usually showed a steady increase in expiratory flow (right); once or twice at each level of NEP, however, she exhibited a sudden drop in expiratory flow (left).

View larger version (49K): [in this window] [in a new window]   Fig. 7.   Tidal flow-volume curves before and during different levels of NEP applied for 500 ms at onset of expiration in 2 subjects (3 and 4) who never exhibited an EMGgg reflex activation in response to NEP in early expiration (left and middle panels) and in subject 6, who invariably had an EMGgg response to NEP (right panels). 3, 3 cmH2O; 5, 5 cmH2O; 7, 7 cmH2O; 10, 10 cmH2O;. See text for explanation.

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

The main finding of this study is the modulation of the reflex activity of genioglossus muscle in response to NEP according to the timing of application of NEP during expiration. In fact, whereas a reflex response of genioglossus was elicited in most subjects by the application of NEP at end expiration as previously described (4), definitely this was not the case at the onset of expiration. In addition, no relationship was found between steady increase or abrupt fall in expiratory flow and the presence or the absence of a reflex activity of genioglossus during sudden application of NEP in early expiration. It follows that, under these dynamic conditions, the maintenance of upper airway patency does not seem to imply the occurrence of a reflex-mediated activation of genioglossus and vice versa.

Studies in awake normal subjects have shown that continuous subatmospheric pressure applied selectively to upper airways during active inspiration increases the phasic activity of the upper airway dilator muscles by reflex mechanisms in a proportional way to stabilize and maintain pharyngeal patency (1, 7). Furthermore, by application of pulses of negative pressure in the upper airway at the onset of inspiration, a reflex activation of genioglossus with a mean latency of 54 ± 11 ms was consistently found in six normal subjects during wakefulness (13).

Recently, by using intraoral bipolar surface electrodes to record EMGgg activity, in 10 awake supine normal subjects application of pulses of 500 ms of negative airway pressure at end expiration has been shown to induce a significant reflex response of genioglossus at levels equal to and higher than 10 cmH₂O with glottis closed and 15 cmH₂O with glottis open (4). The results of our study showing a reflex activity of surface and intraoral EMGgg during the end-expiratory pause in the majority of awake, seated normal subjects in response to NEP of 10 cmH₂O add evidence for reflex pharingeal dilator muscle activation by negative airway pressure at end expiration in normal humans.

This finding seems consistent with phasic activation of upper airway dilator muscles before inspiration in humans acting to stiffen the compliant structures of upper airway against the collapsing forces generated by inspiratory efforts of diaphragm and chest wall muscles (10).

We focused, however, on the genioglossus reflex response after application of NEP in early expiration because sometimes during resting breathing in normal subjects transient reduction and even cessation of flow can occur by application of NEP at the onset of expiration, as used to detect EFL.

In this respect, the substantial lack of reflex EMGgg activity we observed in awake normal subjects suggests that sensory feedback from upper airways in response to subatmospheric pressure might be differently modulated to mediate the reflex activation of upper airway dilator muscles during expiration in humans.

In 24 anesthetized spontaneously breathing rabbits, the effect of upper airway negative pressure pulses (between 5 and 20 cmH₂O) on EMGgg activity has been shown to depend on the time of application within the breathing cycle (14). Indeed, after application of negative pressure pulses, the activation of genioglossus was much lower in the middle expiration than at the onset of inspiration and decreased when negative pressure was applied in late rather than in early inspiration. Thus the time of the application of negative pressure during the respiratory cycle seems to play an important role in determining a reflex response of pharyngeal dilator muscles to upper airway negative pressure.

In one subject (subject 6), a reflex surface EMGgg activity has been invariably found at all levels of NEP applied at onset of expiration.

We do not have a clear explanation for this response, which is opposite to that exhibited by the other subjects. However, in this subject after surface anesthesia of the mouth floor sprayed with 60 mg of 10% lidocaine solution, the reflex surface EMGgg activity after application of NEP of 5 cmH₂O at the onset of expiration was markedly reduced (peak integrated surface EMGgg activity was almost 50% less than before anesthesia). This fact suggests that in some individuals the reflex activation of genioglossus muscle may be mediated, at least in part, by inputs coming from superficial receptors stimulated by NEP and probably traveling by nonvagal afferents (lingual and glossopharyngeal nerves) (5, 8). A similar observation was reported by Horner and colleagues (3), who found an EMGgg reflex activity decreased by 21% in response to negative airway pressure of 25 cmH₂O applied at end expiration in four normal subjects after selective anesthesia of the oropharyngeal mucosa.

Surface records of EMGgg activity are less sensitive and specific than those obtained by invasive technique or by using a special intraoral electrode, which has been recently validated for this purpose (2). In our study, however, the reflex EMGgg activity was always detected by surface electrodes in all subjects during different maneuvers such as inspiratory effort, sniff, and protrusion of the tongue and was consistently observed when NEP was applied at the end expiration and in early expiration in subject 6. In addition, the EMGgg tracings obtained via intraoral electrode in two subjects under the same experimental conditions reflected exactly what was recorded with surface technique.

Contamination of surface EMGgg records by electrical acitivity from other upper airway muscles cannot be obviously excluded. While this criticism does not apply in case of unchanged electromyographic (EMG) activity after stimulus, in the presence of reflex EMG activity induced by NEP, it should be accepted that, because the genioglossus is the biggest and the most important pharyngeal dilator muscle, the greatest amount of the recorded EMG activity would pertain to it.

Thus, although we are well aware of these technical limitations, we are very confident that our surface EMG records were sufficiently reliable to exclude a reflex activity of genioglossus when they were unchanged and to detect an activation pertaining mainly to genioglossus when they displayed a markedly increased activity with NEP.

In conclusion, our results confirm that a reflex response of genioglossus may occur in humans when negative airway pressure is applied during the expiratory phase, showing, however, that this happens much more commonly at the end-expiratory pause than at the onset of expiration. These findings also indicate that when NEP is applied in early expiration the maintenance of upper airway patency does not imply a reflex-mediated activation of genioglossus, suggesting that in these circumstances a partial or total narrowing of upper airways cannot be ascribed to an impaired reflex activity of pharyngeal dilator muscles.