INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Episodes of increased upper airway resistance can cause sleep disruption and daytime sleepiness (1). Increased upper airway resistance is indicated by increased pleural pressure swings and a plateau on the inspiratory flow time profile (inspiratory flow limitation) (1). Such events during sleep may not fit the usual definitions of hypopneas, as any reduction in thoracoabdominal movement (4) or thermal flow signal is insufficient. These more subtle nocturnal respiratory events may represent an early stage in the continuum from snoring to the sleep apnea-hypopnea syndrome (SAHS) (1, 5, 6). While heavy snoring is often associated with increased levels of respiratory effort and a flattened flow-time profile, Guilleminault and coworkers (1, 6) found that episodes of increased pleural pressure swings can occur in the absence of snoring.

Increased upper airway resistance or increased pleural pressure swings are not routinely directly measured in many sleep laboratories during overnight sleep studies, as they require the relatively invasive technique of esophageal intubation. The shape of the inspiratory flow contour has been proposed as a useful noninvasive predictor of increased upper airway obstruction (2, 3, 7). However, the clinical significance of these events in terms of causing cortical/autonomic arousal is not yet clear, nor is the precise definition of such events. Preliminary data suggest that inspiratory flow limitation without snoring can occur in healthy asymptomatic subjects during sleep (8, 9), but it is unclear whether these episodes are more frequent and/or more severe in symptomatic individuals than in normal subjects.

The aim of this study was to examine the frequency of increased upper airway resistance events and how often they caused arousal in a consecutive group of symptomatic patients with upper airway resistance syndrome (UARS) compared with a matched asymptomatic control group. We hypothesized that in patients with the upper airway resistance syndrome there would be a higher frequency of airflow limitation leading to more episodes of increased respiratory resistance and more arousals from sleep and thus greater objective sleepiness. We have also compared the number of events detected by invasive (nasal pressure and pleural pressure) and noninvasive (nasal pressure only) techniques and their significance in both groups of subjects.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Study Population

Eight consecutive symptomatic patients (1 female) with UARS and eight normal asymptomatic subjects (1 female) were studied. The normal subjects were matched for age and body mass index (BMI) to be within 10% of the patient group on a one-for-one basis. The normal subjects were screened with an in house questionnaire, had no sleep complaints, were not sleepy (Epworth score < 10), had no witnessed apneas, were reported by their bedpartners not to be habitual snorers, and had no respiratory or cardiovascular disease. All patients had presented to the sleep clinic, were subjectively sleepy, had Epworth scores  10 (10), and had a prior full-night polysomnography (PSG) that confirmed their apnea-hypopnea indexes (AHIs) to be < 15 and excluded periodic limb movement disorder. There were no coexisting causes of daytime sleepiness, and all patients were free from respiratory or cardiovascular disease. The arousal index on the diagnostic night was not a selection criterion.

Measurements

Both patients and control subjects underwent full night polysomnography (Compumedics, Victoria, Australia), using standard techniques (11). In the patient group, the study night took place within 2 mo of the diagnostic PSG, with no intervening treatment or significant weight change. Measurements included central and frontal electroencephalograms (EEG, C3-C4, CZ-PZ, F3-FP1, and F4-FP2), two electrooculograms (EOGs), submental and right and left tibial electromyograms (EMGs), and the electrocardiogram (ECG). Nasal airflow was determined by the AutoSet system (ResMed UK, Oxford, UK), which measured pressure fluctuations in the anterior nares, square rooted, and recorded on-line on the PSG. Oral airflow was measured by thermistor, thoracoabdominal movement by inductance plethysmography, and oxygen saturation by pulse oximetry (3700; Ohmeda, Essex, UK). Pleural pressure (Ppl) was used as a measure of respiratory effort, using a conventional 10-cm latex balloon catheter system (PK Morgan, Gillingham, UK) placed transnasally under topical anesthesia in the esophagus (typically 35 cm from the nares). The balloon catheter was connected to a pressure transducer of the Compumedics system and recorded on-line to the PSG. Digital arterial beat-to-beat blood pressure was measured in all patients and in the four of eight normal subjects who would tolerate it using an infrared plethysmographic volume clamp method (Finapres 2300; Ohmeda) and recorded on-line to the PSG. The position of the hand during plethysmography was verified by infrared video recording to exclude movement artifact.

Within 1 wk of the overnight study after a normal night's sleep at home, objective sleepiness was assessed in both groups by the Multiple Sleep Latency Test (MSLT) and the Maintainance of Wakefulness Test (MWT) (12, 13).

Within 2 wk of the research study night, patients were given an overnight trial of continuous positive airway pressure (CPAP), using the AutoSet device in therapeutic mode (3), with signals recorded online to the PSG. The number of apneas and hypopneas, cortical arousals, and periods of flow limitation (on the CPAP night derived from the nasal pressure trace only) were compared between the study and treatment nights.

Definitions of Nocturnal Events

Respiratory events. Apneas were defined as a cessation of airflow for at least 10 s and hypopneas as a 50% reduction in thoracoabdominal movement for 10 s or more (4). Analysis of the records was performed using only the nasal pressure signal (with the researcher blinded to the pleural pressure trace) to identify periods of inspiratory flow limitation occurring in the absence of hypopneas from plateauing of inspiratory nasal flow. Upper airway resistive events were defined as a plateauing of inspiratory nasal flow with increasingly negative pleural pressure swings for 2 breaths or more compared with the preceding baseline in the absence of hypopnea (Figure 1). Termination of events was marked as a return of pleural pressure to the sleeping baseline and a rounded flow contour.

View larger version (31K): [in this window] [in a new window]   Figure 1.   An example of a resistive event with heavy snoring. The flattened inspiratory flow time profile is associated with increased pleural pressure swings. The event is terminated by cortical arousal, pleural pressure swings return to the sleeping baseline, and the inspiratory flow-time profile is rounded. Boxes on EEG channel represent arousals.

The number of flow limitation events and how frequently they were associated with respiratory arousal were compared with resistive events identified by both nasal and pleural pressure.

Arousals. Cortical arousals were defined as a return of alpha or theta rhythm for at least 1.5 s with an associated transient EMG rise (14). Respiratory arousals were distinguished from spontaneous arousals on the basis of the pleural pressure and nasal pressure signals. Transient increases in blood pressure of at least 5 mm Hg (avoiding the usual respiratory variation in blood pressure) were marked as nonvisible or autonomic arousals.

Analysis

All polysomnograms were scored manually by standard criteria (11). Arousals were scored blind to respiratory events. The total number and frequency of resistive events per hour slept, and the total duration of each as a percentage of the total sleep time (TST), were calculated and compared between the patients with UARS and normal subjects. The level of pleural pressure swings at the termination of events (the arousal threshold), the presence of cortical arousals, and sleep stage distribution were compared between the two groups of subjects. The number of events detected by the nasal pressure signal alone were compared with those detected by both nasal and pleural pressure in both groups of subjects.

Differences between the patient and normal group were determined by using the Wilcoxon rank test for paired data, with Bonferroni correction as appropriate. Significance is taken at p < 0.05. Results are expressed as median (and interquartile range [IQR]) values.

All subjects gave written informed consent to the study, which had the approval of the local Ethics Advisory Committee.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Group Characteristics

There was no significant difference in age between the two groups (control subjects, 44 [IQR, 37-52] yr; patients, 41 [37- 53] yrs) but the BMI was higher in the patient group (28 [25- 30] kg/m²; 31 [26-32] kg/m²; p < 0.05) despite matching each control to within 10% of each patient's BMI. The AHI was significantly higher in the patient group (Table 1). Among the patients, there was no difference between the study and diagnostic night in terms of the AHI (11 [9-14]/h slept diagnostic, 12 [10-16]/h slept study; p = 0.3), or the arousal index (23 [19- 30]/h slept diagnostic, 28 [18-37]/h slept study; p = 0.5).

The total number of cortical arousals per hour slept on the study night was similar for the patients and normal control subjects but the patients had significantly more respiratory arousals than the normal control subjects (Table 1). The TST and the percentage of time spent in each sleep stage was similar between the groups, except that the normal subjects had more REM sleep.

Both the MSLT and MWT were significantly shorter in the patient group (Table 1).

Resistive Events

There was no difference between control subjects and patients with UARS in the number of resistive events (Table 2), but there was a trend for the total duration of resistive events as a percentage of sleep time to be higher in the patients (UARS, 27 [13-45]% TST; normal subjects, 12 [2-32]% TST; p = 0.1). The within-breath pleural pressure swing at the termination of resistive events ("arousal threshold") was significantly higher in the patient group (Ppl at arousal: patients, 15 [9 to 19] cm H₂O; normal subjects, 11 [8 to 12] cm H₂O; p = 0.02]. Furthermore, more of these resistive events were associated with arousal in the patients with UARS than in the normal subjects (Table 2).

Flow Limitation Events

In the patients with UARS, a similar number of episodes was identified when an esophageal pressure trace was available resistive eventsas when the nasal pressure signal alone was visibleflow limitation events (Table 2). However, a higher number of resistive than flattening events was associated with arousal from sleep in patients with UARS (Table 2). Many episodes of flattening (UARS, 2 [1-4]/h slept; normal subjects, 5 [1-9]/h slept) had no discernible decrease in pleural pressure, and these were associated with fewer arousals than those with increased respiratory effort (percentage of flow limitation events with and without change in pleural pressure associated with cortical arousal: patients, 51 [39-61] and 17 [2-44]%, p = 0.025; normal subjects, 30 [20-43] and 15 [12-23]%, p = 0.01).

Beat-to-beat blood pressure was successfully measured for a significant portion of the night only in the patient group. Among the patients, resistive events that did not result in cortical arousal usually (85 [74-92]% of events) terminated with a significant rise in blood pressure (autonomic arousals). There were insufficient data to report from the normal group because of intolerance of the Finapres device.

Sleep Stage Distribution of Events

The frequency of resistive events was similar between NREM and REM sleep in the patient group (NREM, 18 [12-31] events per hour; REM, 20 [15-31] events per hour; p = 0.8), but the duration of events as a percentage of time in the sleep stage was higher in NREM sleep (NREM, 29 [16-64]%; REM, 11 [10-23]%; p = 0.01). In the control group, the frequency of resistive events was significantly higher in REM sleep (NREM, 6 [2-25] events per hour; REM, 23 [9-41] events per hour, p = 0.025) but the duration of resistive events as a percentage of sleep stage was not different between stages (NREM, 4 [1- 35]%; REM, 12 [4-18]%; p = 0.8).

CPAP Titration Night

All patients had a CPAP titration night, using the AutoSet system attached to full PSG. Respiratory events were detected from thoracoabdominal movement (A + H) and the AutoSet nasal pressure signal was recorded on-line to the PSG and manually scored to give flow limitation events. In all patients, CPAP reduced the AHI (study night, 12 [10-16]/h; titration night, 6 [5-13]/h, p = 0.01), the frequency of flow-limited events (study night, 11 [7-23]/h; titration night, 3 [2-6]/h; p = 0.01), and the number of cortical arousals (study night, 28 [18- 27]/h; titration night, 21 [16-23]/h, p = 0.05).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

This study confirms that symptomatic patients with a low AHI may have periods of increased upper airway resistance that are associated with significant sleep disruption (1, 6). In addition, we have shown that periods of increased inspiratory effort with inspiratory flow limitation occur frequently in normal asymptomatic subjects. Pleural pressure swings were greater at arousal in patients with UARS than in control subjects, suggesting tolerance to the sleep-disrupting effects of increased effort.

As far as we are aware, this is the first study to systematically compare the frequency and significance of increased upper airway resistance between patients with UARS/mild SAHS and a matched asymptomatic normal group of subjects. Normal subjects were asymptomatic and had AHIs in the normal range (15). It is possible that both groups of subjects might have had a slightly higher AHI and arousal frequency in this study compared with their normal because of the instrumentation, but no significant changes were shown in sleep quality between baseline polysomnography and the current study in the patient group. Similar arousal indices have been reported previously in normal asymptomatic subjects during less instrumented overnight studies (16). When arousals were categorized into those associated with respiratory events and those occurring spontaneously, normal subjects had significantly fewer respiratory arousals than the patient group.

Preliminary reports have shown that inspiratory flow limitation occurs during sleep in the normal population and is associated with sleep disruption (8, 9). This study shows no increase in the frequency of either flow limitation or resistive events in patients with UARS. The total number of arousals was not different between the two groups, although the patients had a greater number of cortical arousals associated with these resistive events. However, the pleural pressure at the termination of events was more negative in the patient group. The level of pleural pressure is the likely stimulus for arousal from sleep during apneas, hypopneas, and resistive events (17, 18); however, other factors may also impact. Recurrent episodes of respiratory related arousals may blunt the arousal threshold (19), explaining why patients had more negative pressures at arousal than normal subjects. Sleep fragmentation caused by respiratory arousal seems to have different physiological effects than nonrespiratory arousals (20), perhaps explaining why, although the total number of arousals was not significantly different, the arousal thresholds for negative pleural pressure were different. Such recurring sleep fragmentation may also predispose to upper airway collapsibility (21), thus setting up a vicious circle.

Treatment with nasal CPAP in the patient group significantly reduced both the number of resistive events and associated cortical arousals, which confirms previous findings (1, 2, 7). However, the decrease in arousals in the UARS group with CPAP (median, 7/h slept) was only marginally greater than the decrease in hypopneas with CPAP (median, 6/h slept) and smaller than the total decrease in apneas/hypopneas and flow-limited events (median, 14/h slept). Thus, the physiological significance of decreasing the number of flow-limited events with CPAP remains unclear.

The precise definition of inspiratory flow limitation and increased upper airway resistance, and the best techniques to detect such events, are as yet unclear. It was first described as periods of high esophageal pressure swings with associated inspiratory flow limitation, which do not fit the current definitions of hypopneic episodes (1). To detect such events requires invasive techniques, and several studies have explored the use of noninvasive signals (7, 2, 22, 23). The shape of the inspiratory flow-time profile derived from nasal pressure signals has been found to be closely associated with changes in respiratory effort as measured by esophageal pressure (7, 2). In the present study, we have examined the use of the nasal pressure signal alone to detect resistive events and compared this with the "gold standard" definition of both nasal pressure and esophageal pressure signals combined. When using the nasal pressure signal in isolation, a number of events were detected that had a "flattened" inspiratory flow-time profile, but with little or no change in the esophageal pressure swing. In the patient group, events identified by nasal pressure alone were associated with significantly fewer arousals than those identified by using both nasal pressure and esophageal pressure.

Limitations to the study include both methodology and power. We measured resistive events using pleural pressure swings and nasal pressure as a flow indicator. We adopted this technique in preference to using occlusive masks to measure ventilation in an effort to optimize sleep quality. We were careful to exclude all periods from analysis in which the thermistor signal showed evidence of mouth flow. Thus, we believe that the reported episodes of increased resistance in which there was a combination of plateauing of nasal flow with a more negative pleural pressure are reasonable markers of episodes of increased airflow resistance. This study was performed on 16 subjects and thus the power to detect subtle changes between the number of events in the patient and normal groups was limited. However, the main thrust of the investigation was not to quantitate differences in frequency of events between these two groupsany such differences that do exist would clearly vary between the populations studied but to identify whether resistive events and flow-limited events were common in normal subjects. We believe that the data show that such events do occur commonly in normal subjects.

This study has confirmed the presence of episodes of increased upper airway resistance in symptomatic patients with UARS, but has also demonstrated a high level of these events in a matched group of normal asymptomatic subjects. However, patients with UARS generated more negative pleural pressures during events and had more associated arousals. The data suggest that caution needs to be applied to the interpretation of the clinical significance of resistive and flattening events in order to avoid overdiagnosis of the upper airway resistance syndrome.