Published ahead of print on October 2, 2003, doi:10.1164/rccm.200302-150OC
© 2003 American Thoracic Society Ventilatory Response to Brief Arousal from Non–Rapid Eye Movement Sleep Is Greater in Men Than in WomenAdelaide Institute for Sleep Health, Repatriation General Hospital, Daw Park; Department of Physiology, University of Adelaide, Adelaide; and School of Medicine, Flinders University, Bedford Park, South Australia, Australia Correspondence and requests for reprints should be addressed to Amy Jordan, Ph.D., Sleep Disorders Program at BIDMC, Brigham and Women's Hospital, 75 Francis Street, Boston, MA 02115. E-mail: ajordan{at}rics.bwh.harvard.edu
Sleep apnea syndromes are more common in men than in women. The ventilatory response to arousal from sleep may be an important determinant of respiratory stability/instability and could contribute to this sex difference. We therefore compared changes in ventilation, end-tidal carbon dioxide (CO2), upper airway resistance, heart rate, and finger photoplethysmogram pulse wave amplitude after both spontaneous and tone-induced arousal from non–rapid eye movement sleep in 13 men and 13 women. At sleep onset, ventilation fell and both upper airway resistance and end-tidal CO2 rose, but these changes were not different between sexes. Spontaneous arousal (duration, 6.6 ± 0.2 seconds) resulted in a biphasic ventilatory response consisting of brief hyperventilation (5 seconds) followed by prolonged hypoventilation (30–40 seconds) on resumption of sleep. The biphasic ventilatory response was greater in men than in women and did not appear to be explained by different wake-to-sleep increments in end-tidal CO2 or upper airway resistance between sexes. Peripheral vasoconstriction with arousal was also greater in men than in women. Ventilatory responses were more marked after tone-induced versus spontaneous arousals and when subjects slept supine compared with the left lateral position. These results suggest that male sex and supine position are associated with greater ventilatory instability after arousal from sleep.
Key Words: sleep apnea • arousal • gender • ventilation Obstructive sleep apnea and central sleep apnea associated with congestive heart failure are more common in men than in women (1, 2). The cause for this sex difference in prevalence is not known. Respiratory control instability has been proposed as a mechanism underlying sleep-disordered breathing (3). It is possible that some components of central respiratory control may differ between the sexes and thus help explain the difference in prevalence of these sleep-related breathing problems. Arousal from sleep is believed to have a destabilizing influence on the control of breathing (4–6). In support of this theory, Xie and colleagues (6) found that in patients with central sleep apnea, 37 of 48 episodes of periodic breathing arising from stable sleep were preceded by arousal, hyperventilation, and reduced transcutaneous CO2. The increase in ventilation after arousal from sleep is thought to arise from increased chemoresponsiveness on return to the waking state, a relatively high PaCO2 that develops in sleep before arousal (see Phillipson and Bowes [7] for review), and a sudden reduction in upper airway resistance (5). Given that the awake hypercapnic ventilatory response has been reported to be higher in men than in women (8–10), men may be expected to have a larger ventilatory response on awakening for a given sleep-related increase in PaCO2. However, Khoo and colleagues (5) have suggested that the ventilatory response to arousal is largely determined by upper airway resistance changes, with central drive playing only a small part in the response. During non–rapid eye movement (NREM) sleep, upper airway resistance has been reported to be higher in men than in women in one study (11), but not different between the sexes in another (12). If upper airway resistance were higher in men than in women during sleep, the ventilatory response to arousal may also be elevated in men, secondary to a release of resistance. Horner and coworkers (13) and Trinder and coworkers (14) have suggested that arousal from sleep may elicit an independent stimulus or "waking reflex" with both ventilatory and cardiovascular components. If the waking reflex is similar to a startle reflex, it may be higher in men than in women, as Reyes del Paso and Vila (15) reported a greater increase in breathing amplitude (pneumatic transducer) in response to a brief startling tone (100 dB) in men compared with women during wakefulness. The current study was therefore designed to test the hypothesis that the ventilatory response to arousal from sleep is greater in men than women because of one or more of the above-mentioned factors. Given that a startling tone alters respiration, it is possible that arousal responses are different whether they occur spontaneously or are induced with a tone. For this reason we also compared spontaneous and induced arousal responses. Further aims were to determine whether the ventilatory response to arousal is influenced by sleep stage or posture (supine versus lateral position)-induced changes in upper airway resistance. Preliminary findings of this study have been previously reported in the form of an abstract (16).
Subjects Thirty-three healthy, nonsmoking subjects (15 men, 18 women in the follicular menstrual phase) not taking any medications gave informed written consent and participated in the study. No subject reported regular snoring (more than one night per week) or any auditory, cardiorespiratory, or sleeping problems. The study was approved by the Research and Ethics Committee of the Repatriation General Hospital (Daw Park, South Australia, Australia).
Equipment and Measurements
Protocol
Data Analysis
Anthropometric Characteristics and Sleep Architecture Four female subjects had markedly abnormal sleep (average sleep onset latency, 186.6 ± 112.5 minutes; sleep efficiency, 22.9 ± 7.7%; arousal index, 47.4 ± 8.1 arousals per hour). There were no brief arousals preceded and followed by stable sleep in these subjects, so all their data were excluded. Eleven other subjects (5 men) were found to snore or have flow limitation at some time during the night and arousals after snoring or flow limitation were discarded because the effect of snoring on arousal responses is unknown. This eliminated all otherwise suitable arousals in one female and one male subject. All data were excluded from one male subject who had no suitable arousals without a mask leak, and the genioglossal data from another subject, in whom one wire electrode became dislodged midprotocol, were excluded. Repeated brief arousals from stable sleep, without snoring, flow limitation, or mask leaks were obtained in 13 men and 13 women. The men and women did not differ in terms of age (26.2 ± 1.9 and 24.1 ± 1.8 years, respectively), body mass index (22.6 ± 0.5 and 23.2 ± 0.9 kg/m2, respectively), or respiratory function (FEV1 or FVC); however, the women studied had significantly smaller mean body surface area (1.99 ± 0.04 in men compared with 1.74 ± 0.05 m2 in women, p < 0.01). Sleep architecture did not differ between men and women (see Table E1 in the online supplement) and an average of 63 tones were played to each subject (range, 37–99; not significantly different between sexes).
Resting Data
Ventilatory Response to Spontaneous Arousal in the Left Lateral Position Twenty-five subjects (13 men and 12 women) had brief spontaneous arousals from NREM sleep while in the left lateral position (average, 15.6 ± 1.9 arousals per subject). The average duration of arousal was not different between men and women (6.8 ± 0.3 and 6.3 ± 0.4 seconds, respectively). After arousal,  I rose briefly, before rapidly declining to below the prearousal level (Figure 1)
. This was primarily related to changes in VT; however, significant changes over time were also found in respiratory frequency (Figure 1), TI, and TE. Corresponding to the increase in I was an increase in phasic EMGDI and reductions in PETCO2 and Rua (Figure 1). Both phasic and tonic EMGGG (Figure 1), and tonic EMGDI (data not shown), failed to show statistically significant changes over time after arousal from sleep.
The changes in I were more marked in men than in women (significant sex and sex-by-time interaction effects; Figure 1). Both postarousal hyperventilation (at 4 seconds) and hypoventilation (maximal at about 20 seconds) were significantly greater in men compared with women (p = 0.02 and p = 0.01, respectively). VT and PETCO2 (p = 0.053; Figure 1) also showed more pronounced responses in men than women whereas no other variables differed significantly between the sexes over time. Although men had a greater ventilatory increase after arousal than did women, ventilation did not exceed the level during wakefulness in either men or women (see Figure E1 in the online supplement for absolute changes after arousal). Rua did not decrease to the wakefulness level after arousal and was not elevated above the prearousal sleeping level after return to sleep (see Figure E1 in the online supplement). Ventilatory responses to spontaneous arousal from Stage 2 to Stage 3/4 NREM sleep were not different (see description of findings, and Figure E2, in the online supplement).
Spontaneous versus Tone-induced Arousal Responses
Effect of Body Position on Resting Ventilation and Arousal Responses Twenty-two subjects (11 men and 11 women) had brief spontaneous arousals from stable NREM sleep in both the left lateral and supine positions. The duration of arousal did not differ between positions (6.5 ± 0.3 seconds, left; and 6.3 ± 0.3 seconds, supine) or between men and women (6.6 ± 0.6 and 6.2 ± 0.3 seconds, respectively). The prearousal characteristics for these subjects in both left and supine positions are presented in Table 2 and Table E3 (see the online supplement). Rua was increased in the supine position and TI was prolonged; however, TE was reduced, resulting in respiratory frequency remaining unchanged. PIF and VT were also reduced in the supine position. I fell on assuming the supine position in the whole group; however, this was related to a reduction in men but not in women (significant sex-by-position interaction). Correspondingly, there were slight changes in PETCO2 in men and women when assuming the supine position (Table 2).
After arousal, TI, PIF, and Rua demonstrated significant position-by-time interaction effects with similar trends seen in I (p = 0.055; Figure 2B), VT (p = 0.056), and PETCO2 (p = 0.073) (see Figure E3 in the online supplement for complete results). Sex differences over time were again found for I (p = 0.051), VT, and PETCO2 (p = 0.055) and tonic EMGGG also showed greater responses in men than women when both positions were considered. Although the normalization process used (expressing postarousal data as a percentage of the prearousal level) was required to assess the effect of arousal per se, it does not allow valid comparison of the effect of position on arousal responses in men compared with women, because the change in position from left to supine affected prearousal conditions differently between sexes. Repeated measures analysis of variance was therefore performed on the absolute changes in ventilation from prearousal, to the maximum (4 seconds postarousal) and minimum (20 seconds postarousal) levels in men and women in both left and supine positions (Figure 3)
. This analysis showed that, although resting ventilation was reduced in the supine versus left lateral position in men but not women, and the increment in ventilation on arousal was greater from supine versus lateral sleep (p = 0.04), the effect of posture on the ventilatory increase after arousal was not influenced by sex (p = 0.34; Figure 3). The absolute changes in PETCO2 and Rua after arousal were also greater in the supine versus lateral position (p = 0.04 and p < 0.01, respectively), but this effect of posture was not different between sexes (p = 0.22 and p = 0.29, respectively).
Cardiovascular Responses to Spontaneous and Tone-induced Arousals Spontaneous and tone-induced arousals produced brief tachycardia followed by a small but more prolonged bradycardia, marked finger vasoconstriction lasting 30–40 seconds, and a small decrease in finger pulse arrival time (Figure 4) . Heart rate and pulse arrival time responses were similar between the sexes, whereas females showed a more variable finger pulse wave response to arousal than males. All but one male showed a clearly defined postarousal vasoconstrictor response. In contrast, 5 of the 13 females showed either no response or vasodilatation. As a result, finger photophlethysmogram responses were more marked in men than in women (spontaneous arousal sex effect, p = 0.022; spontaneous arousal sex-by-time effect, p = 0.005; and tone-induced arousal sex-by-time effect, p = 0.045). Heart rate and pulse wave amplitude responses were significantly greater for tone-induced versus spontaneous arousals (heart rate arousal type-by-time effect, p < 0.001; pulse wave amplitude arousal type effect, p = 0.019). Prearousal heart rate was lower and pulse arrival time was longer in men versus women (Table E4; see the online supplement).
The major finding of this study was that the ventilatory response to arousal from NREM sleep was higher in men than in women, and men developed greater subsequent hypoventilation on resumption of sleep. This occurred in both the lateral and supine positions and in Stage 2 and Stage 3/4 NREM sleep. The sex difference persisted whether the arousal was induced by a tone or occurred spontaneously. In addition, the ventilatory response to arousal was greater in the supine position than in the left lateral position, and when induced by a tone rather than occurring spontaneously. Finger vasoconstriction associated with spontaneous and tone-induced arousals was also more pronounced in men compared with women. These sex, type of arousal, and position-related differences were not explained by differences in arousal duration, which was 6–7 seconds for all comparisons.
Influence of Sex on Ventilatory Response to Arousal After the initial hyperpnea and the return to sleep, a reduction in ventilation was observed that was more pronounced in men than in women. During the period of hypoventilation airway resistance was lower than the prearousal level and was not different between sexes. It therefore seems unlikely that the more marked hypoventilation in men was due to elevated upper airway resistance. It is possible that the greater hypocapnia that developed immediately after arousal, when combined with the sleep-related fall in chemoresponsiveness, contributes to a reduction of central respiratory drive on return to sleep. However, we believe this would not entirely explain our findings based on the relatively small sex difference in sleeping hypocapnic ventilatory responsiveness reported by Zhou and coworkers (24). Neural inhibition of respiration after the greater hyperpneic phase in men could further contribute to sex differences in the postarousal ventilatory response.
Tone-induced versus Spontaneous Arousal from Sleep The ventilatory response to tone-induced arousal reported in the current study appeared smaller and followed a different time course than some (4, 5), but not all (25) previous studies conducted in healthy individuals. This may reflect different auditory stimuli (5), the phase of respiration at which tones were presented, or different techniques for measuring respiration (4) in previous studies.
Influence of Position on Arousal Responses The ventilatory response to arousal from sleep was greater in the supine than in the left position. This may be related to prearousal ventilation being lower when supine, or to a larger decrease in resistance after arousal in the supine position (Figure 3). Alternatively, the neural waking reflex may be influenced by posture. The subsequent undershoot in ventilation did not appear to be greater in the supine than in the lateral position (Figure 2). In the supine position, there was a greater postarousal reduction in upper airway resistance than occurred in the lateral position, which gradually returned to the prearousal level over 30 seconds (see Figure E3 in the online supplement). It is possible that the more prolonged reduction in airway resistance reflects greater postarousal upper airway stability in the supine position (compared with prearousal), which protects against more marked hypoventilation. It is important to note that the change in resistance from the lateral to supine position was relatively small in these young healthy subjects. Thus larger increases in resistance in older or snoring subjects may more markedly influence the stable sleeping ventilation and arousal responses than demonstrated in this study. It may appear somewhat surprising that changes in ventilation and PETCO2 on assuming the supine posture were greater in men than in women, and yet men did not have a more marked position-related change in the arousal response when compared with women. One explanation for this result is that the ventilatory response to arousal may be determined more by a behavioral neural response to waking (i.e., a "waking reflex") than the prearousal conditions. In this way, the waking reflex may mask the effect of the small differences in prearousal conditions found in the current study. This is consistent with the finding of Trinder and coworkers (14) that the magnitude of diaphragm activation after arousal from sleep in normal subjects was not different between spontaneous breathing and mechanical ventilation despite PETCO2 and ventilation being markedly different between conditions. Also, Carley and colleagues (25) have reported that steady state hypercapnia does not alter the ventilatory response to tone-induced arousal from sleep, providing additional support for this hypothesis.
Possible Relevance to Sleep Apnea Syndromes The relevance of the findings presented in this study to obstructive sleep apnea is unclear. Patients with obstructive sleep apnea have large increases in resistance during sleep and might be expected to develop obstructive apnea during hypoventilation after a brief arousal from sleep. However, after arousal in the supine position (when prearousal resistance was elevated compared with the left lateral position) a larger initial response was observed but the subsequent hypoventilation was no more marked than in the lateral position. This may reflect the slow return of resistance to the sleeping level. If patients with obstructive sleep apnea are simply further along a continuum of increased upper airway resistance in sleep, they might be expected to show a larger initial ventilatory response without greater undershoot. However, if the time course of change in resistance is shorter in patients with obstructive sleep apnea then an obstructive respiratory event may occur during the period of low respiratory drive. Clearly the role of arousal from sleep in the pathogenesis and high male prevalence of sleep apnea syndromes requires further investigation.
Methodologic Considerations
The particular arousal responses and subjects used for each analysis (influence of sex, sleep stage, and arousal type) varied slightly. Thus there may have been prearousal differences between comparison groups. The sleep-related changes in In this study breath-by-breath measurements were interpolated after arousal, whereas previous studies have averaged data by breath number after arousal (4, 5, 14, 25, 26, 29). Interpolation was performed so that changes in respiratory frequency were taken into consideration. This technique would not account for observed differences between sexes or conditions. However, interpolation tends to regress data toward baseline and will therefore tend to underestimate the magnitude of the initial hyperpnea. To minimize this effect, an interpolation interval close to total respiratory cycle time was selected and interpolation was not performed across the arousal. A related limitation of the data analysis was that an interaction effect over time was required to identify differences between groups because of the biphasic nature of the response. It is possible that differences in some variables may be lost if the difference is restricted to a small portion of the response analyzed. In the case of spontaneous arousals, ventilation increased before EEG changes such that the prearousal baseline level of ventilation was slightly elevated. However, when the data immediately before arousal (–4 seconds) were excluded the average baseline ventilation was only 0.03 L · min–1 lower. This would lower the whole spontaneous arousal response curve down by 0.6%, a change too small to contribute importantly to the results observed. Finally, measurement of PETCO2 on breaths with low tidal volumes can be problematic, as a plateau in mask CO2 at end expiration may not be reached. It is important to note that the magnitude of hypoventilation observed in this study was not sufficient for this to be a problem and the end-expiratory plateau in mask CO2 was still observed.
Conclusions
A.S.J. has no declared conflict of interest; D.J.E. has no declared conflict of interest; P.G.C. has no declared conflict of interest; R.D.M. has no declared conflict of interest.
Supported by the National Health and Medical Research Council of Australia, grant number 138403. This article has an online supplement, which is accessible from this issue's table of contents online at www.atsjournals.org Received in original form February 2, 2003; accepted in final form September 27, 2003
This article has been cited by other articles:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
METHODS
significant sex-by-time interaction effect; 
trend (p = 0.053) to sex-by-time interaction effect.
