INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Obstructive sleep apnea (OSA) is a common disorder characterized by the repetitive collapse of the pharyngeal airway during sleep. Its prevalence is 2-4%, and OSA is associated with sequelae including reduced quality of life, increased risk of motor vehicle accidents, hypertension, and possibly an increased frequency of myocardial infarction and stroke. OSA is substantially more common in men than women, with male:female ratios ranging between 2:1 and 10:1, depending on the study design (1, 2). Considerable effort has focused on the mechanism underlying this male predominance, but no clear explanation has emerged. Differences in pharyngeal anatomy and dilator muscle activation/function have been proposed. In theory, a smaller pharyngeal lumen in men could lead to increased susceptibility to OSA, but imaging studies have failed to show such a difference (3). Consistent sex-related differences in pharyngeal dilator muscle activation have not been demonstrated as well. Finally, the simple assessment of pharyngeal resistance in normal men and women during the wake-to-sleep transition revealed no sex-based difference, although airflow resistance was slightly higher in men in deeper non-rapid eye movement (NREM) sleep (6). However, the effect of sleep alone may not be an adequately provocative test to demonstrate sex effects in normal subjects.

One technique for probing upper airway physiology is the use of inspiratory resistive loading (IRL). The effects of externally added IRL on respiratory muscles and ventilation have been previously tested during wakefulness and NREM sleep in healthy subjects (7). During wakefulness, the level of respiratory drive in response to IRL, assessed using a variety of techniques, is increased immediately and serves to maintain tidal volume (VT) and minute ventilation (E). There is little change in upper airway resistance as well. In contrast, during the first few breaths following load application during NREM sleep, there is a substantial increase in pharyngeal resistance (airway collapse) and a decrease in VT and E (7).

Sex differences in the IRL response, however, have been minimally studied. We therefore designed the current protocol to achieve three objectives. First, we explored sex differences in the IRL response at physiologic loads (5 to 25 cm H₂O/L/s). Second, we sought to define the mechanisms underlying potential sex-related differences in load response, that is, differences in upper airway collapse versus central drive. Third, we hoped to define the pharyngeal dilator muscle response to loading during NREM sleep to determine if potential sex-related differences in collapsibility were related to muscle responsiveness. Therefore, we measured the activation of both a tonic and phasic pharyngeal dilator muscle during basal breathing and load applications.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects

Eight healthy women (age = 27.0 ± 2.1 yr, BMI [body mass index] = 22.0 ± 0.9 kg/m²) and eight healthy matched men (age = 27.4 ± 2.1 yr, BMI = 24.3 ± 0.6 kg/m²) were studied. None had any pharyngeal anatomical abnormality on physical examination. Subjects denied any chronic diseases, daytime somnolence, or snoring. All women were studied in the follicular stage of their menstrual cycle, defined as Days 5-11 with Day 1 being the first day of the menses. The study was approved by the Brigham and Women's Human Subjects Review Committee, and the subjects gave written informed consent prior to participation in the study.

Instrumentation and Techniques

Ventilation. Subjects wore a nasal mask (Healthdyne Technologies, Marietta, GA) connected to a two-way valve partitioning inspiration and expiration. Inspiratory flow was determined with a pneumotachometer (Fleish, Inc., Lausanne, Switzerland) and differential pressure transducer (Validyne Corp., Northridge, CA), calibrated with a rotameter. The subjects breathed exclusively nasally as assured by mouth tape and video camera. Tidal volume (VT) was obtained from the integrated flow signal, and minute ventilation (E) was calculated on a breath-by-breath basis using the equation E = VT · (60/Ttot), where Ttot is the breath duration in seconds.

Pressure and resistance. Pressures were monitored in the nasal mask (Validyne Corp.) and in the airway at the level of the choanae and the epiglottis. One nostril was decongested (oxymetazalone HCl) and anesthetized (approximately 0.5 ml of 4% lidocaine HCl), and two pressure-tipped catheters (MPC-500; Millar, Houston, TX) were inserted through this nostril and localized to measure choanal and epiglottic pressures. Prior to insertion, all three pressure signals were calibrated simultaneously in a rigid cylinder using a standard water manometer. These three signals plus flow were demonstrated to be without amplitude or phase lags at up to 2 Hz.

Inspiratory resistive loading (IRL). During all studies, subjects breathed exclusively via the nasal route through tubing that incorporated a one-way valve, a pneumotachograph, and a variable inspiratory resistance device. Expiration was unimpeded via an expiratory pathway, whereas inspiration could be loaded to any desired level by varying the effective caliber of the inspiratory pathway. The total unloaded baseline resistance was approximately 2 cm H₂O/L/s. The variable inspiratory resistance device consisted of a water-filled latex balloon with a wall thickness of 0.15 mm, mounted on a 6.0-mm-outer diameter tube, which was centered within the inspiratory pathway (12.4 mm inner diameter). This balloon could be inflated using a graduated syringe. As the balloon was increasingly distended, it filled more of the tube's caliber and effectively reduced the cross-sectional area available for inspiratory airflow. This resistor has been tested and shown to produce a linear pressure flow relationship (13). Four loads (5, 10, 15, and 25 cm H₂O/L/s) were used during this study with all inspiratory loads being initiated during the previous expiration.

Muscle activation. The genioglossal electromyogram (GG EMG) was measured with a pair of unipolar intramuscular electrodes referenced to a single ground, thus producing a bipolar recording. Two stainless-steel Teflon-coated 30-gauge wire electrodes were inserted 15-20 mm into the body of the genioglossal muscle near its insertion into the mandible and 3 mm lateral to the frenulum on each side, using a 25-gauge needle, which was quickly removed, leaving the wires in place. Tensor palatini electromyogram (TP EMG) was measured in a manner similar to that of the GG muscle, using a pair of referenced unipolar intramuscular electrodes producing a bipolar recording. On each side of the palate, the tip of the pterygoid hamulus was located at the junction of the hard and soft palate. A 25-gauge needle with a 30-gauge stainless-steel Teflon-coated wire was then inserted at a 45° angle along the lateral surface of the medial pterygoid plate, to a depth of approximately 10-15 mm into the palate. The needle was then removed, leaving the electrode in place. These techniques have been used previously in our laboratory (14, 15). To confirm TP electrode placement, the following respiratory maneuvers, which have been shown previously to activate the TP muscle (16), were performed: sucking, blowing, and swallowing.

For both muscles, the raw EMG was amplified, band pass filtered (between 30 and 1,000 Hz), rectified, and electronically integrated on a moving-time-average (MTA) basis with a time constant of 100 ms (CWE, Inc., Ardmore, PA). The EMG was quantified as a percentage of maximal activation. To define maximal muscle EMG activity subjects performed four maneuvers: they were instructed to maximally inspire against an occluded inspiratory airway, maximally protrude their tongue against the maxillary alveolar ridge, swallow, plus repetitive sucking and blowing. Each of these maneuvers was performed several times, and the maximal EMG recorded for each muscle during this calibration was called 100%. Electrical zero was then defined as 0%, and thereafter muscle activity for each individual was quantified as a percentage of their maximal activation.

PSG. Wakefulness/sleep was documented with two-channel electroencephalography (EEG) (C3-A2, C4-O1), two-channel electrooculography (EOG), and submental electromyography (EMG). Subjects were maintained in the lateral decubitus posture throughout the study using pillow and back support. This position was continuously verified by video camera.

Study Protocol

Subjects reported to the sleep laboratory in the evening, having been without food intake for at least 4 h. After obtaining informed consent, they were instrumented with the equipment described above. Prior to allowing each subject to fall asleep, tape was applied to the mouth to ensure nasal breathing. After at least 5 min of stable NREM sleep, data collection was initiated (see Figure 1). First basal breathing was recorded, and then IRL was begun with each set consisting of four loads applied for three breaths each (5, 10, 15, 25 cm H₂O/L/s). Three complete data sets were attempted in each individual. At least 30 s of basal breathing separated each loading series. If the subject awoke during loading, this series was excluded, and repeated after the subject fell back to sleep (for at least 5 min prior to reloading).

View larger version (8K): [in this window] [in a new window]   Figure 1.   This schematic diagram describes the study protocol. After stable non-rapid eyemovement (NREM) sleep was recorded for 5 min, basal breathing was quantified, followed by the application of increasing loads of 5, 10, 15, and 25 cm H2O/L/s for three breaths each, with loads being separated by at least 30 s. This procedure was repeated three times in each subject.

Data Recording and Analysis

All signals (EEG, EOG, submental EMG, airway pressures [mask, choanal, epiglottic], inspiratory flow, GG EMG, and TP EMG) were recorded on a 16-channel Grass model 78 polygraph (Grass Instruments, Quincy, MA). Certain signals (VT, E, airway pressures, inspiratory flow plus GG and TP EMG) were also recorded on computer using signal-processing software (Spike 2; Cambridge Electronic Design, Ltd, Cambridge, UK). Sampling frequency was 125 Hz.

For each breath (baseline and breaths 1, 2, and 3 after each load), the pharyngeal resistance (Rpha, choanae to epiglottis), nasal resistance (Rn, mask to choanae), and supraglottic resistance (Rsup, mask to epiglottis) were determined at a flow of 0.2 L/s and at peak negative epiglottic pressure (which is peak resistance). In addition, VT, inspiratory time (TI), expiratory time (TE), total breath time (Ttot), E, GG EMG (tonic and peak phasic) and TP EMG (tonic only, as inspiratory phasic activation was not encountered) were determined on a breath-by-breath basis. These variables were then averaged from the three data sets for each subject. Because no significant differences were found in any of the measures between the first, second, and third breaths following IRL, these were averaged as well. Thus, each data point (i.e., basal breathing and response to four different IRLs) represents an average of nine breaths. Changes in respiratory variables during IRL application were compared to the preloading baseline data.

All statistical analyses were performed with commercially available software (Excel 97, Microsoft; and SigmaStat + Sigmaplot, SPSS, Chicago, IL). All data are presented as mean ± 1 SE unless otherwise stated. Two-tailed t tests for independent samples were used to compare variables between men and women, and repeated measures ANOVA, with Student's Newman-Keuls post-hoc comparisons to statistically assess the impact of progressively greater inspiratory loads. In each case, p < 0.05 was taken to indicate significance.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Full data sets were obtained in 15 subjects; in one male subject the data for IRL of 15 and 25 cm H₂O/L/s are missing due to technical problems. Respiratory variables for men and women during NREM sleep before and during IRL are summarized in Table 1. As expected, VT and E were greater in men than women under baseline conditions. However, there were no significant sex-related differences in baseline inspiratory timing or breathing rate during sleep. In addition, none of the participants experienced spontaneous apneas or hypopneas during basal breathing.

Men responded to all four levels of resistive loading with greater decreases in VT compared to women (Figure 2). These differences were statistically different between sexes at IRLs of 15 and 25 cm H₂O/L/s. At a resistance of 25 cm H₂O/L/s, VT in women decreased by 26 ± 5%, with a reduction of 44 ± 5% in men (p < 0.05). Changes in breathing cycle durations were not significantly different between the two groups. In both groups, IRL resulted in a prolongation of inspiratory time (TI) and a shortening of expiratory time (Table 1, Figure 3). As a result, TI/Ttot gradually increased with progressively applied IRL in both sexes. Ttot and respiratory rate remained largely unchanged in both groups in response to all levels of IRL (Figure 3). Thus, minute ventilation (E) decreased in a similar manner to VT (Figure 4). In women, E fell by 29.0 ± 1.4% in response to an IRL of 25 cm H₂O/L/s (p < 0.001), whereas E fell by 45.4 ± 1.4% at the same IRL in men (p < 0.001 from BL, p < 0.05 between sexes).

View larger version (18K): [in this window] [in a new window]   Figure 2.   Relative decrease in tidal volume with four inspiratory resistive loading (IRL) levels is shown. Data are presented as mean ± SE. In both groups differences from BL were significant for all IRL levels except IRL of 5 in females (§ p < 0.05 different from BL), and differences between sexes were significant with IRL levels of 15 and 25 cm H2O/L/s (* p < 0.05 male different from female).

View larger version (16K): [in this window] [in a new window]   Figure 3.   Graphic representation of the effect of four levels of IRL on tidal volume and respiratory cycle timing. Group average responses of men are shown in the left panel and of women in the right panel. No sex-related differences were observed in TI, and Ttot remained largely unchanged in both sexes. As shown previously, VT fell more in men than women with increasing IRL.

View larger version (18K): [in this window] [in a new window]   Figure 4.   Relative decrease in minute ventilation with four IRL levels is shown. Data are presented as mean ± SE. §p < 0.05 (different from BL); *p < 0.05 (different between sexes).

None of our subjects responded to any load with complete apnea, but six men and two women responded with snoring and flow limitation. Flow limitation was defined as a plateau or decrease in flow despite an increase in negative epiglottic pressure of at least 1 cm H₂O (17, 18). A representative example for a male and female with an IRL = 10 cm H₂O/L/s is shown in Figure 5. This figure shows prominent flow limitation in the male and none in the female. Indeed, the increase in pharyngeal resistance with loading was greater in men than women for all loads with this difference reaching significance at IRLs of 10, 15 and 25 cm H₂O/L/s (Figure 6, Table 2) when measured at peak resistance. At lower flow rates (0.2 L/s) there were no sex-related differences in resistance observed. Thereafter (within the breath) men tended to flow limit and increase Rpha, and these differences became significant between sexes at peak negative epiglottic pressure (peak resistance). In men, pharyngeal resistance increased from a baseline of 7.6 ± 2.9 to 44.8 ± 13.8 cm H₂O/L/s with an IRL of 25 cm H₂O/L/s (p < 0.05), whereas in women, Rpha increased from 4.9 ± 2.3 to 11.5 ± 6.0 cm H₂O/L/s with this same load (p < 0.05 between sexes, Table 2). As nasal resistances did not change substantially with any of the loads in either sex (all values ranged between 0.5 and 1.5 cm H₂O/L/s), supraglottic resistance tracked the pharyngeal resistance.

View larger version (37K): [in this window] [in a new window]   Figure 5.   One representative breath without and with IRL is shown, for a male and a female. Vertical lines indicate flow = 0.2 L/s, and peak negative epiglottic pressure (peak resistance). In both cases (man and women) with IRL (10 cm H2O/L/s) flow decreased and epiglottic plus choanal pressures became more negative, indicating increased airflow resistance in both subjects. However, the male demonstrated prominent flow limitation that was not observed in the female.

View larger version (15K): [in this window] [in a new window]   Figure 6.   Relative increase in pharyngeal resistance with four IRL levels is shown. In the left panel resistances are measured at flow = 0.2 L/s, whereas in the right panel at peak resistance (peak negative epiglottic pressure). Data are presented as mean ± SE. At peak resistance in men, differences from BL were significant for all IRL levels, whereas in women, a significant difference was observed only at an IRL = 25 cm H2O/L/s. §p < 0.05 (different from BL); *p < 0.05 (different between sexes).

Due to the marked increase in pharyngeal resistance described above in men, the total load to which men were exposed (intrinsic + applied) was higher than women. When the change in ventilation is plotted as a function of total load, the graphs are colinear (Figure 7), indicating that the sex-related difference in ventilatory load response can be completely accounted for by the greater increase in pharyngeal resistance observed during IRL in men. However, the determination of resistance at a single flow rate may be problematic during flow limitation.

View larger version (18K): [in this window] [in a new window]   Figure 7.   Changes in minute ventilation (loaded E  baseline E) as a function of the total load (applied extrinsic plus pharyngeal intrinsic) are shown. Both sexes had a similar response, implying that the mechanism of increased load-induced hypoventilation in men was related to upper airway collapse rather than central drive. Data are presented as means ± SE (on each axis).

GG and TP EMG tended to increase with IRL, but these changes did not reach statistical significance. In men, peak phasic GG EMG increased from 11.1 ± 3.1(% of maximal activation) at baseline to 13.3 ± 4.7% with a load of 25 cm H₂O/L/s (NS), whereas in women, the equivalent increase was from 6.7 ± 1.8 to 7.4 ± 1.9 (% of maximal activation, NS, differences between sexes NS). Similarly, TP activation increased from 4.3 ± 1.2 to 4.8 ± 1.6 (% of maximal activation) in men (baseline versus highest IRL, NS), and remained unchanged in women (4.7 ± 1.6 versus 4.6 ± 1.7% of maximal activation, NS, differences between sexes NS).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The results of this study indicate that healthy men adapt significantly less well than women to resistive loading during sleep. The increase in pharyngeal resistance induced by loading was significantly greater in men than in women. Thus, the male pharyngeal airway is considerably more collapsible than the female one, when exposed to greater intraluminal negative pressure. In addition, by determining the total load (applied plus intrinsic), it becomes clear that the ventilatory response to this total load is similar between the sexes. These data strongly suggest that the mechanism of increased load-induced hypoventilation (relative to baseline, see Table 1) in men is related to greater upper airway collapse rather than reduced central drive.

IRL served in this study as a provocative test of ventilatory control/upper airway mechanics during sleep. The decreased ventilatory response to externally applied loading found in men compared with women could be the result of several mechanisms. These include increased pharyngeal collapsibility (i.e., increased pharyngeal wall compliance or decreased pharyngeal muscle activation) or decreased central ventilatory drive. It has been previously shown that when IRL is applied during sleep, pharyngeal resistance increases (9), due to airway collapse. As a result, the respiratory system is working against both the externally applied resistance plus the internal (pharyngeal) one that subsequently develops. How far ventilation falls therefore is a product of both the total load to which the respiratory system is exposed and the responsiveness of the ventilatory control system to this load. By measuring both internal and external resistance, we were able to calculate the total load to which our subjects were exposed. By plotting the ventilatory response of all subjects versus total load (Figure 7), it becomes clear that the load response of males and females is essentially identical. This implies that the central response to loading is similar between sexes and that the more marked hypoventilation induced in men resulted from greater upper airway collapse. In addition, most previous data suggest that during NREM sleep there is little respiratory system response to brief applications of IRLs in either sex (7, 8, 19). Thus, the greater load-induced hypoventilation in males (relative to baseline) is certainly secondary to prominent airway collapse not encountered in females.

The response of the upper airway to IRL varies substantially with state. During wakefulness, there is little increment in pharyngeal resistance with loading, although a clear pharyngeal muscle response has been difficult to demonstrate (19, 20). During NREM sleep, substantial upper airway collapse has been demonstrated with loading with little to no immediate pharyngeal muscle activation. Over time, genioglossus muscle activity does increase, although this may represent a response to rising PCO₂ rather than a specific response to the loading. Therefore, over the first few breaths after load application during sleep, collapsibility of the pharynx is likely a product of four factors: (1) upper airway dilator muscle activity, (2) the intrinsic characteristics of the airway tissue, (3) the size of the airway lumen prior to load application, and (4) the anatomic structure of the pharynx. As a result, the greater pharyngeal collapsibility in men must be a product of one or several of these variables.

We found no sex differences in muscle activation in response to IRL during sleep. Neither the basal EMG level during NREM sleep nor the response of the muscles to loading differed between men and women. Although one previous study reported greater genioglossal EMG in women than men during wakefulness (21), there are no previous studies addressing sex effects on pharyngeal muscle activation during sleep nor their response to provocation during sleep. Based on the current data, it seems unlikely that the substantial sex differences in pharyngeal collapsibility resulted from differences in muscle activation.

There are currently no studies of actual pharyngeal tissue characteristics in healthy men versus women, so any discussion regarding sex differences in this area would be highly speculative. However, a number of studies suggest that airway size is similar in men and women (22). Rubinstein and coworkers observed a similar reduction of upper airway size during expiration from total lung capacity to residual volume in men and women (23). As a result, at any given lung volume, both groups showed similar supraglottic area during wakefulness. Other studies, using a variety of techniques, have also failed to demonstrate differences in upper airway size in men and women, when measured during wakefulness (3, 24). In addition, airflow resistance prior to loading in this study was similar in men and women. Therefore there is little support for the notion that airway size prior to loading differed between men and women. Finally, the anatomic structure of the pharynx could be affected by gender. In support of this concept is our preliminary observation that the female airway (measured from hard palate to epiglottis) is considerably shorter than that of males (25, 26). As a longer airway, similarly tethered, will be substantially more collapsible than a shorter one, this might, in part, explain the observations of this study. However, considerable further investigation will be required to definitively answer this question.

There are several potential limitations of this study. First, we made no direct measurements of central drive in our subjects. Although the near-identical slopes of the plots of ventilation versus total load imply similar central responses, we recognize that directly measured central drive (P₁₀₀, diaphragmatic EMG activity, etc.) would have been more definitive. Second, our method of comparing EMGs between individuals using the percentage of maximal activation is probably less than ideal due to variability in electrode placement, effort on maximal maneuvers, etc. However, we have used this technique successfully in previous studies and male-female comparisons of basal muscle activation during sleep were not a primary goal of this study (21, 27, 28). Finally, how the observations of this study apply to patients with sleep apnea remains speculative. However, we believe that the observed collapsibility of the pharyngeal airway in men may predispose them to adverse events during sleep.

In conclusion, the most striking observation of this study is the markedly greater pharyngeal collapsibility in response to externally applied load found during NREM sleep in healthy men compared with women. We believe that this is likely a product of either differing anatomic support for the upper airway or differences in pharyngeal tissue characteristics between men and women. It is this difference in airway collapsibility between the sexes that dictates the ventilatory response to loading during sleep as central drive appears quite similar.