INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

OBSTRUCTIVE SLEEP APNEA (OSA) is more common in men than in women for reasons that are not clearly understood. Although many investigations have focused on identifying upper airway anatomic differences between men and women (2, 17), more recently a difference in respiratory neural control of upper airway dilator muscles has been implicated in this gender difference (27). The respiratory activity of the genioglossus muscle is due partly to a reflex response to the negative airway pressure during inspiration and partly to central activation independent of the reflex (spontaneous activity from respiratory premotoneurons) (6). Either of these elements may differ between the genders and may contribute to the high male prevalence of OSA.

The resting inspiratory activity of the genioglossus muscle has been reported to be higher in women than in men (27). This could, at least in part, be a result of hormonal differences because the genioglossal activity was later reported to be higher in the luteal menstrual phase than in the follicular phase (28). However, more recent studies have failed to replicate the finding of a gender difference in resting genioglossal activity during wakefulness (32) or sleep (26). There is, therefore, uncertainty as to whether a gender difference in resting genioglossal activity actually exists.

After removal of a respiratory stimulus, there is a respiratory afterdischarge (RAD) in inspired minute ventilation (I) that can be observed as a slow decline in ventilation to the resting level (3). The duration of RAD has been proposed to be important in determining ventilatory stability (33), with a short RAD representing an underdamped respiratory controller that is predisposed to unstable patterns. Consistent with this theory is the finding that the duration of RAD is shorter in OSA patients than non-OSA controls (4), suggesting a possible respiratory control abnormality in these patients. Our laboratory recently found RAD, as measured by the rate of decline in ventilation after brief hypoxia, to be the same in men and women and the same between the luteal and follicular menstrual cycle phases (9). In this study, we wished to extend these observations to examine whether there was RAD in the genioglossus muscle and, if so, whether it parallels that seen in respiratory pump muscles or is different between the genders. A slow decline in respiratory phasic hypoglossal nerve activity has been described in cats after carotid sinus or superior laryngeal nerve stimulation (8), and in humans the activities of the diaphragm and genioglossus muscles have been shown to change proportionately during hypoxia (23) and hypercapnia (22). These findings suggest that phrenic and hypoglossal motoneuron control are coupled during periods of physiological stimulation. However, it is not known whether, on abrupt termination of a respiratory stimulus, there is RAD in hypoglossal motorneurons and, if so, whether this is coupled to the RAD of respiratory pump muscles. There are a variety of situations where the activities of the hypoglossal and phrenic motoneuron pools become uncoupled. These include after alcohol ingestion (11), benzodiazepene administration (14) anesthesia, and cyanide brain hypoxia (30). The two motoneuron pools are also seen to behave differently in patients with OSA because, although both muscles increase to a similar degree during an apnea, the genioglossus increases out of proportion to the diaphragm on termination of the apnea (24). We therefore wished to determine whether the genioglossus and diaphragm muscle activities follow each other and ventilation after sudden termination of hypoxia and whether the muscle groups behave similarly between genders. If gender differences in RAD were observed in the electrical behavior of these two muscles, it may have implications for the mechanical behavior of the upper airway after episodes of brief hypoxia such as those that occur in OSA.

In awake normal subjects, the activity of the genioglossus muscle [measured as electromyogram (EMG) of the genioglossus (EMGgg)] has recently been reported to correlate with pharyngeal pressure during application of inspiratory resistive loads (16). It is postulated that, in awake subjects, a negative pressure neural reflex (7, 13) is a major contributing factor to genioglossal activity during tidal and stimulated breathing (16). Whether the negative pressure reflex or the relationship of pharyngeal pressure to upper airway dilator activity during tidal breathing is different between genders has not been investigated previously.

We therefore aimed to compare three aspects of upper airway dilator function in healthy young men and women: 1) resting inspiratory EMGgg activity, 2) the EMGgg RAD after a brief hypoxic stimulus, and 3) the relationship between the EMGgg and pharyngeal airway pressure during normoxia and isocapnic hypoxia.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects. Sixteen men and fourteen women gave informed written consent and participated in the study. All subjects were healthy nonsmokers with normal lung function (spirometry and body plethysmography) and did not take any regular medication, including the oral contraceptive pill. Female subjects had regular menstrual cycles not longer than 35 days and were tested in the luteal menstrual phase (days 20-24). Plasma progesterone levels were measured (ACS:180 Progesterone Assay, Chrion Diagnostics, Chrion Healthcare, Victoria, Australia). One subject was subsequently found to be anovulatory (plasma progesterone <7 nmol/l in the luteal phase), and her results were excluded from analysis. Inadequate data were obtained in 4 men and 1 woman (see Data analysis), leaving 12 men and 12 women for analysis. The study was approved by the Research and Ethics Committee of the Repatriation General Hospital, Daw Park, South Australia.

Equipment and measurements. Subjects wore a nasal mask (Gel mask, Respironics, Murrysville, PA) with a two-way non-rebreathing valve attached (series 2600, Hans Rudolph, Kansas City, MO). The inspiratory side of the breathing valve was connected to a pneumotachograph (model PT36, Erich Jaeger, Würzburg, Germany) and a Gatlin-Shape valve system (series 2440C, Hans Rudolph) for delivery of inspiratory gases. The Gatlin-Shape valve consisted of one output port attached to the pneumotachograph and four inputs, three of which were connected to foil bags (300 liters, Scholle Industries, Adelaide, Australia) containing the following inspiratory gas mixes: 100% O₂, 100% N₂, and 9% O₂ in N₂. The fourth port was open to room air. Only one input port was open at a time, and all changes between ports were conducted during expiration. The inspiratory flow signal (pneumotachograph) was electronically integrated to give inspiratory tidal volume (VT). Inspiratory (TI), expiratory (TE) and total breath (TT) times and I were determined from the flow and volume signals. Mask pressure (P_mask) and end-tidal PCO₂ and PO₂ (PET_CO2 and PET_O2, respectively) were measured continuously from the mask (P_mask: model 78342A, Hewlett Packard, Andover, MI; PET_CO2 and PET_O2: POET II model 602-3, Criticare Systems, Waukesha, WI).

Protocol. Subjects arrived at the laboratory in the morning after a light breakfast without caffeine. They were instrumented (as described in Equipment and measurements) and lay supine with one pillow. Maximal EMGgg activity was determined by having the subjects perform three of each of the following maneuvers: swallows, deep sniffs, and maximal tongue protrusions against the top teeth. The highest activity recorded during these maneuvers was taken to be the maximal activity of the genioglossus. The EMG activity was then expressed as a percentage of maximal activity by scaling the moving-time-averaged EMG between electrical zero and the maximum activity level. This well-established technique (19) allowed the EMGgg signals to be averaged within and between subjects and compared between genders.

Data analysis. To enable RAD to be measured, it was necessary to select trials in which the baseline breath-by-breath ventilation was relatively stable and free from artifact and in which there was a clear hypoxic respiratory neural stimulus (evident in I and both the EMGdi and EMGgg activities). Therefore, individual trials were only included in the analysis if the following criteria were met. 1) Baseline I was stable (the coefficient of variation in resting I in the 30 s before a trial was <25%). 2) I at the end of hypoxia was increased 120% above baseline. 3) The EMGdi and EMGgg both increased 10% above the resting level by the end of hypoxia. 4) Isocapnia was maintained (the PET_CO2 both during and after hypoxia was within 2 Torr of the mean of the 30-s prehypoxia level). 5) There were no sighs in the last 20 s of the hypoxic period or in the 25 s immediately after the hypoxia (sighs were characterized by a change in VT of >75% between 2 adjacent breaths when accompanied by a change in TE of >40% between the same two breaths). 6) There were no swallows in the last 20 s of the hypoxic period or in the 25 s immediately after the hypoxia (swallows were visually determined from the P_epi and EMGgg traces). If swallows or sighs occurred elsewhere within a trial, then the breath during which it occurred was removed, but the trial remained in the analysis. One female and four male subjects had no trials that met these criteria and were excluded from further analysis. To allow breath-by-breath measurements to be averaged within and between subjects, the remaining trials were interpolated at 4-s intervals starting 10 s after the hyperoxic breath to allow for inspiratory circuit (4 s) and peripheral circulatory (6 s) delays (10). The last 20 s of the hypoxic period was also interpolated at 4-s intervals to ascertain the degree of hypoxic stimulation. Resting variables were determined for the 20 s immediately before each trial. EMGgg vs. P_epi comparisons were made for breaths during and after hypoxia to give a range of P_epi and EMGgg activities. The slopes, correlation coefficients, and theoretical EMGgg activity at zero pharyngeal pressure (y-intercept of the P_epi vs. EMGgg activity plot) were compared between genders. The latter should theoretically reflect the inspiratory premotoneuron activity of the muscle, that is, inspiratory phasic activity arising from the respiratory pattern generator uninfluenced by the negative pressure reflex.

Statistical procedures. Anthropometric and resting characteristics were compared between men and women by two-sample Student's t-tests. Student's t-tests were also used to compare the slope, intercept, and correlation coefficients for P_epi-EMGgg activity relationships. All measured variables [I, VT, TT, phasic and tonic EMGgg and EMGdi, EMGgg₂₀₀, PET_CO2, Sa_O2, P_epi, P_epi at 200 ml/s flow, P_mask, peak inspiratory flow (PIF), and R₂₀₀] were compared between men and women for the last 20 s of hypoxia and for 24 s after termination of hypoxia with two-way ANOVA for repeated measures, including the Greenhouse-Geisser correction for multisample asphericity (15). Data are expressed as means ± SE. P < 0.05 was considered statistically significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Trial exclusion. There were 117 hypoxia trials that met the inclusion criteria (57 trials in 12 men, 60 in 12 women). There were approximately equal numbers of trials excluded in men and women for each exclusion criteria except the "I increase" criterion in which 32 trials were excluded in men and only 16 in women.

Resting characteristics. There were no differences in age, body mass index (BMI), or lung function between the 12 men and 12 women in this study (Table 1). Resting I did not differ between the men and women in the luteal menstrual phase; however, the PET_CO2 was significantly lower in the women, consistent with the respiratory stimulant effect of progesterone. The mean plasma progesterone level in the 12 women studied was 36.12 ± 4.16 nmol/l (range 15.0-62.8 nmol/l), confirming recent ovulation in all women. PIF, P_epi, R₂₀₀, and EMGgg (phasic, tonic, and EMGgg₂₀₀) were not different between the genders at rest.

Hypoxic stimulation and afterdischarge. There were no time or gender-by-time interaction effects in PET_CO2 during or after hypoxia, indicating satisfactory maintenance of isocapnia (Fig. 1). The mean Sa_O2 values at rest (98.1 ± 0.2 vs. 97.9 ± 0.3% for women vs. men) and in the last 20 s of hypoxia (86.0 ± 2.0 vs. 87.1 ± 0.9% for women vs. men) were not significantly different between genders. The increase in I and EMG activities during the hypoxic stimulus and the decline after removal of the stimulus are shown in Fig. 2. There were no significant gender or gender-by-time interaction effects in I, P_epi, or phasic or tonic activity of the EMGgg or EMGdi during or after hypoxia. There were also no significant gender or gender-by-time interactions in R₂₀₀, P_mask , PIF, VT, or TT.

View larger version (12K): [in this window] [in a new window]   Fig. 1.   End-tidal PCO2 (PETCO2) control. PETCO2 before (Pre), during (20 to 4 s), and after removal (0-24 s) of brief hypoxia in men (n = 12) and women in the luteal menstrual phase (n = 12). Values are means ± SE. There were no significant time or gender-by-time interaction effects, indicating strict maintenance of isocapnia. *Significant effect of gender on PETCO2 consistent with the resting differences (P = 0.002, men higher than women).

View larger version (26K): [in this window] [in a new window]   Fig. 2.   Ventilatory, airway pressure, and genioglossal and diaphragmatic electromyogram activity during and after removal of a brief hypoxic stimulus. There was a decline in inspired minute ventilation (I), peak epiglottic pressure (Pepi), and peak inspiratory phasic (Phasic) and expiratory tonic (Tonic) genioglossus (GG) and diaphragm (DI) muscle activity after brief isocapnic hypoxia in men (n = 12) and women (n = 12). Values are means ± SE expressed as a percentage of the preintervention level. There were no significant gender or gender-by-time interactions in any variable.

P_epi-EMGgg activity relationship. There were no gender differences in the correlation coefficient, slope, or intercept of the relationship between EMGgg activity and P_epi when group mean values were compared. In some subjects, the relationship between P_epi and EMGgg activity did not reach statistical significance; however, when only the subjects with significant relationships were compared, there were still no differences in correlation coefficient, slope, or intercept between genders. Individual responses are shown in Table 2.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In this study, we compared the electrical activity of an inspiratory pump muscle (diaphragm) and an important upper airway dilator muscle (genioglossus) at rest and during and after brief hypoxic episodes. The hypoxic exposures were deliberately kept brief (45-75 s) and given at infrequent intervals (7 min between episodes) to avoid the suppressant effect on the hypoglossal motoneuron pool, which has been demonstrated previously with either prolonged (20) or more condensed intermittent hypoxic exposures (18). We believe, therefore, that these data provide important information on the intrinsic neural activity and behavior of hypoglossal and phrenic motoneuron pools at rest and in response to a brief respiratory stimulus. The main findings of this study were that the resting respiratory-related activity of the genioglossus muscle was not different between women in the luteal menstrual phase and men and that both the genioglossus and diaphragm muscles declined proportionately during RAD after removal of brief hypoxia. The genioglossus muscle activity was correlated with the negative P_epi during inspiration in most men and women studied, and the mean slope of this relationship was not different between the genders.

Resting genioglossal activity. Previous reports of resting genioglossal activity in healthy men and women are inconsistent (26, 27, 32). Contrary to the results of the present study, Popovic and White (27) reported that the resting level of genioglossal activity was 11% higher in women than in men. The 22 subjects studied by Popovic and White were slightly older (men, 34 yr; women, 32 yr) and had similar BMIs (men 24.4 kg/m²; women, 22.2 kg/m²) compared with those in the present study; however, in their study, the BMI was significantly higher in the men than in the women. It would seem unlikely that the elevated BMI in men would explain the difference in genioglossal activity found because a higher BMI would be expected, if anything, to be associated with a reduced upper airway size and therefore a higher resting level of genioglossal activity (25). Popovic and White found the women to have a smaller mean BMI yet a higher resting genioglossal activity.

Genioglossal and diaphragmatic activity after brief hypoxia. The duration of RAD is thought to be important in stabilizing respiratory patterns (33). If RAD is short, ventilation may become unstable and the subject more susceptible to an undershoot in respiratory drive after any brief respiratory stimulus (such as a brief arousal from sleep). Furthermore, the upper airway is more prone to collapse at times of low respiratory drive (1, 21). Our laboratory has recently demonstrated that the RAD duration is not different between healthy young men and women (9), suggesting it is unlikely that a fundamental gender difference in RAD exists. However, whether RAD occurs in the genioglossus and parallels that of the diaphragm RAD in humans has not been previously studied, nor has the RAD in the two muscles groups been compared in men and women. The phrenic and hypoglossal motoneuron pools have been shown to behave similarly in some circumstances [at rest, during hypoxia (23), and during hypercapnia (22)] but differently in other situations such as after alcohol (11), diazepam (14), cyanide hypoxia, and anesthesia (30). In male cats, the alcohol-induced reduction of hypoglossal output was reduced by pretreatment with a progesterone analog (29), suggesting that progesterone may influence hypoglossal activity and that there may be a gender difference in genioglossal control. In this study, we have found that the decline in genioglossus activity on abrupt termination of a brief hypoxic stimulus parallels the decline in diaphragm muscle activity and ventilation and that the decline was not different between genders. This suggests that the phrenic and hypoglossal motoneuron pools are closely coupled during RAD after hypoxia and are not different between young men and women during wakefulness. Whether this coupling of RAD activity in the genioglossus and diaphragm is the result of similar RAD in respiratory premotoneurons projecting to hypoglossal and phrenic motoneuron pools, or whether the genioglossal activity is coupled via the negative pressure reflex, cannot be determined by this experiment. Caution must be exercised when extrapolating the findings to the mechanical behavior of the upper airway in men and women after a brief respiratory stimulus. Little is known about the electromechanical coupling of upper airway dilator muscles in humans, and no data exist comparing this in men and women. It is possible for example, that the same degree of electrical activation of upper airway dilator muscles in the genders may result in different degrees of stiffening or dilatation of the airway.

Relationship between E_epi and EMGgg activity. Consistent with the finding that there are no gender differences in the EMGgg activity at rest or after increased respiratory drive, the slopes, intercepts, and correlation coefficients of the EMGgg vs. P_epi relationship were also not different between men and women. It is important to note the correlation coefficients in this and the earlier study (16) although significant, were relatively weak (mean R = 0.47), suggesting that other factors importantly influence genioglossal activity. Behavioral influences or postural reflexes may be important. Furthermore, although it seems that the stimulus for negative pressure activation of genioglossal activity is probably stretch receptor activation due to airway distortion, an alternative view is that the muscle activity may affect airway compliance and size and, therefore, the airway pressure. In this way, the most negative pressures would be expected when genioglossal activity is lowest. It is therefore possible, and we believe probable, that the P_epi is both a stimulus for, and a result of, the upper airway EMG activity. This may explain the relatively weak correlation between the variables.

Methodological considerations. Given the negative results of the present study, power calculations were performed to determine the minimum difference in genioglossal activity that could be detected given the number of subjects used and standard deviations measured in this study. It was determined that a gender difference in resting genioglossal activity of 2.7% of maximal activity could be detected with a study power of 80%. This corresponds to a doubling of resting genioglossus activity. In the study of Popovic and White (27), a difference of 10% of maximal genioglossal activity was reported between genders, and this also corresponded to a twofold difference. It would therefore appear unlikely that our results reflect a type II statistical error.

Summary. In summary, we have measured diaphragm and genioglossus muscle activity in a group of young healthy men and women at rest and during and after a brief hypoxic stimulus while awake. We also compared the pharyngeal pressure and genioglossal activity relationship between the genders. We found that there was no difference in the genioglossal activity at rest or during hypoxia. There was also no gender difference in the genioglossal or diaphragmatic RAD after abrupt removal of hypoxia, and the relationship between pharyngeal pressure and genioglossal activity was similar between the two genders. This study is the first to demonstrate that the hypoglossal and phrenic motoneuron pools behave similarly after removal of a brief hypoxic stimulus in healthy men and women. These results do not necessarily exclude gender-specific changes in the output of the hypoglossal or phrenic motorneron pools during sleep or with increased age or body weight.