INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION CONCLUSIONS REFERENCES

Obstructive sleep apnea (OSA) causes daytime sleepiness (1) and impairs general health status (2). These are the main reasons for treating OSA with nasal continuous positive airway pressure (nCPAP), which alleviates both sleepiness and health status (3). Daytime sleepiness is due to recurrent arousal from sleep triggered by repetitive episodes of partial or complete obstruction to the pharyngeal airway (1, 6), and such sleep fragmentation may also contribute to the impaired quality of life of subjects with this condition. Methods for quantifying these recurrent nocturnal events are complex and varied, and include respiratory measures (7), cortical sleep fragmentation indices (8), and monitoring of autonomic activation, which detects very minor events not even detectable on the electroencephalogram (EEG) (9). Although several studies have investigated the relationship between these measures and daytime sleepiness (10) and its response to nCPAP (13), there are no studies of similar relationships with health status.

General health evaluation questionnaires aim to quantify self-perceptions concerning general health and functional and emotional well-being. The use of these questionnaires as outcome measures in clinical trials is increasing. Generic questionnaires enable comparisons between the health effects of different diseases, as well as the effect of therapeutic interventions. The Short Form-36 Health Survey questionnaire (SF-36) has been shown to be sensitive to treatment effects in several medical conditions (14, 15) including OSA (2, 5). OSA is associated with impairment of several aspects (or dimensions [16]) of health status, and it is yet to be determined whether all of these abnormalities are due to OSA itself, or whether some are caused by other common coexisting conditions such as obesity, cardiovascular disease, and poor lung function with daytime ventilatory failure. Studies of the effect of OSA on health status may be confounded by these factors. We have previously used the improvement in sleepiness with nCPAP in OSA as a method of distinguishing confounding causes of sleepiness in OSA patients (13). The present study used a similar approach to distinguish impairment of health status due directly to OSA from that caused by other problems that are unlikely to change with treatment. The OSA symptoms likely to improve with nCPAP can broadly be divided into two categories: symptoms directly related to sleep fragmentation (e.g., sleepiness or cognitive impairment) and those more indirectly related to OSA (e.g., reduced libido, snoring, nocturia).

This study used the SF-36 to quantify health status in a population of subjects referred to a sleep clinic for investigation of possible snoring and OSA. This population was specifically chosen as having a wide range of both baseline health status and change in health status with nCPAP. To obtain data on the response to nCPAP in this group, all subjects, regardless of disease severity, were treated with nCPAP.

The primary aims of the study were: (1) To explore the relationships between pretreatment health status and severity of sleep-disordered breathing (using both respiratory and sleep fragmentation indices). (2) To explore the relationships between the change in health status with nCPAP and severity of sleep-disordered breathing (using both respiratory and sleep fragmentation indices).

The secondary aims of the study were: (1) To compare health status in a sleep clinic population with normal community data, using the SF-36. (2) To examine the effect of nCPAP on health status. (3) To explore the relationships between health status and daytime sleepiness.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION CONCLUSIONS REFERENCES

Subjects

Fifty-one subjects (46 men and five women) were recruited from the Oxford Sleep Clinic, to which they had been referred for investigation of possible snoring and OSA. The study population was selected on the basis of a prior sleep study to represent the full spectrum of upper airway narrowing during sleep, from sleep without snoring, through simple snoring and to mild/moderate and severe OSA (median apnea/ hypopnea index [AHI] 25, 90% central range: 1 to 98). The study was approved by the Central Oxford Research Ethics Committee, and subjects gave consent in accordance with the Committee's requirements.

Protocol

Subjects spent 1 d in the laboratory before the study polysomnogram, for assessment of pretreatment health status (SF-36) and daytime sleepiness. For the purposes of the study, all subjects, regardless of the sleep study findings, were initially given nCPAP on the night after the polysomnogram. At the end of 4 wk of nCPAP treatment, subjects returned for health status assessment.

Health Status Assessment: The SF-36 Health Survey Questionnaire

The SF-36 is a widely used and robust measure of subjective health status, and has undergone considerable reliability and validity testing (17). It is a 36-item, self-completed questionnaire that measures eight multiitem aspects (dimensions) of health. These are: physical functioning (10 items), role limitation due to physical problems (four items), role limitation due to emotional problems (three items), social functioning (two items), mental health (five items), energy/vitality (four items), bodily pain (two items), and general health perceptions (five items). From these dimensions, two summary scores of physical well-being (physical component summary [PCS]) and emotional well-being (mental component summary [MCS]) are derived. Both the dimensions and component summaries are transformed onto a scale from 0 (worst possible health) to 100 (best possible health).

Daytime Sleepiness: Subjective and Objective Measurements

Epworth Sleepiness Scale. The Epworth Sleepiness Scale (ESS) (18) is a questionnaire specific to symptoms of daytime sleepiness. Subjects are asked to score the likelihood of falling asleep in eight different situations with different levels of stimulation, resulting in a final score of 0 to 24. The ESS has been shown to be abnormal in subjects with sleep disorders including sleep apnea syndrome (18), and correlates significantly with an objective measure of sleepiness: the Multiple Sleep Latency Test (19) (r = 0.51, p < 0.01)

The Oxford Sleep Resistance test. The Oxford Sleep Resistance test (OSLER test) (20) is a behavioral test of sleepiness similar in structure to the 40-min electroencephalogram (EEG) based Maintenance of Wakefulness Test (MWT). It separates patients with OSA from normal subjects in a similar way to the traditional MWT (21). To identify sleep onset, the OSLER test uses failure to continue to respond to a light flashing regularly at 3-s intervals.

Subjects received written instructions to have a normal night of sleep before attending the laboratory for the test, and to abstain from beverages containing alcohol or caffeine from midnight before and throughout the day of the test. Compliance with these instructions was confirmed on arrival at the sleep laboratory. During the OSLER test, four sleep-resistance challenges were given to each subject in a darkened room at 2-hourly intervals throughout the day, and sleep was concluded to have occurred when there had been no response for 21 s to the flashing light used in the test. The results of these four tests were averaged to give a single figure.

Sleep Recording

Sleep was recorded with an eight-channel tape recorder (MPA-2/9200; Oxford Medical Instruments, Abingdon, UK). Recording included a frontal EEG (Fp2-A2), a two-channel electrooculogram (EOG), and a submental electromyogram (EMG). Respiration was recorded from oronasal airflow, ribcage and abdomen movements, and arterial pulse oximetry (Model 3700, pulse oximeter; Ohmeda, Boulder, Colorado). Apnea was defined as a decrease in airflow to < 20% of baseline for more than 10 s, and hypopnea was defined as a decrease in airflow to < 50% for the same period. Autonomic activation was monitored via indirect arterial beat-to-beat blood pressure measurements (pulse transit time [PTT] measured between the R-wave of the electrocardiogram (ECG) and the photoplethysmographically detected arrival of the pulse wave at the left ring finger (RM10; Parametric Recorders, London, UK) (22). Movement was detected by subtracting serial, digitized video images gathered every 2 s (Visilab Sleep Monitoring System; Stowood Instruments, Oxford, UK) (13).

Initiating nasal CPAP Therapy

Information to patients. On the night after the sleep study, nCPAP was started in all patients, including patients with a subjective complaint of snoring but in whom no snoring had been detected in a single-night sleep study. The reasons given for treatment were the same for all subjects regardless of the severity of their disease. It was explained that nCPAP would correct snoring but that it was not possible to confidently predict from the sleep study whether the subject would gain any alleviation of daytime symptoms, and that treatment would therefore be given for 4 wk with measurements before and after treatment for comparison.

Nasal CPAP titration. Subjects viewed an educational video, and skilled technical staff members chose the most comfortable nasal mask and headgear for the subjects, ensuring that they were comfortable with the system. The first night of treatment was given in the laboratory, using the Horizon Auto-adjust nCPAP system (7354I Series; Sunrise Medical, Devilbiss Division, West Midlands, UK) which adjusts nCPAP according to the presence of snoring, apneas, and hypopneas (23). During the nCPAP titration night, subjects were monitored with a multichannel system, monitoring a dip rate of > 4% in oxygen saturation, snoring and arousal confirmed by movement (24), and brief increases in heart rate (autonomic nervous system activation) (9). The following morning, the level of nCPAP for home use was chosen as that which controlled all snoring and sleep fragmentation related to upper-airway incompetence (in all subjects this was the 90th centile of overnight nCPAP pressure or higher). For the nonsnoring subjects, in whom the autoadjust nCPAP did not rise above 3 cm H₂O, the minimum nCPAP used was set to 5 cm H₂O to ensure adequate flow of air through the nasal mask and to prevent hypoxemia associated with rebreathing. On the following morning, subjects were reviewed and supplied with an nCPAP machine (7353 Series; Sunrise Medical, Devilbiss Division, West Midlands, UK) that included a compliance counter based on actual machine use (by detecting changes in motor speed with breathing). Subjects were also provided with our usual support information, including a contact telephone number which it was emphasized that they should use for any problems with their treatment.

Follow-up assessment. After 1 mo of nCPAP all subjects returned to the sleep laboratory, where the SF-36 was repeated and nCPAP compliance was noted. Changes in SF-36 dimensions and summary scores were calculated by subtracting pretreatment from posttreatment values, therefore quantifying the change in these health status measures with nCPAP.

Sleep Study Analysis

EEG indices of sleep fragmentation. Sleep was staged manually according to standard criteria (25). Arousals as defined by the American Sleep Disorders Association (ASDA) were scored according to the ASDA criteria (8), and respiratory arousals were scored as ASDA arousals preceded by an apnea or hypopnea. These were expressed as the total number of events divided by total sleep time, to give the ASDA arousal index and respiratory arousal index, respectively.

Autonomic events. The PTT is measured with every heartbeat, and is oversampled and stored at 5 Hz to ensure that no values are missed. In keeping with our previous development work, a decrease in PTT of greater than 15 ms occurring over a 4- to 45-s period was defined as an autonomic arousal event (22). The start and end times for PTT analysis were taken from sleep period time (SPT), which was used as the denominator for PTT to give the autonomic events/h.

Statistical Analysis

The differences in the SF-36 dimensions and summary scores before and after nCPAP are quoted as effect sizes. Because the pre- and posttreatment SF-36 dimensions were closely correlated, the effect sizes were calculated with the method recommended by Kazis and coworkers (26), by dividing the mean change in a variable by the standard deviation of the variable at baseline. The advantage of using effect size is that it provides quantitative information about the magnitude of change, rather than yielding statistical significance alone. As a guide to interpreting these values, Cohen (27) identified an effect size of 0.20 as small, 0.50 as moderate, and 0.80 as large.

Associations between the SF-36 dimensions and sleep fragmentation indices were explored with Spearman's correlation analysis (SAS statistical software package; SAS institute, Cary, NC). In order to adjust for other factors contributing to diminished health status, the change in health status with nCPAP was calculated and correlated with respiratory indices of OSA severity, cortical and autonomic sleep fragmentation indices, and daytime sleepiness.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION CONCLUSIONS REFERENCES

The study population exhibited a wide spectrum of sleep-disordered breathing, ranging from sleep without snoring to mild, moderate, and severe OSA in the diagnostic study, with a consequent wide range of sleep fragmentation indices (Table 1). The SF-36 was repeated on nCPAP in 50 subjects (one subject refused to attend for follow-up assessment).

Relationship between SF-36 dimensions and OSA severity. The relationships between the pretreatment SF-36 dimensions and indices of sleep fragmentation and OSA severity were examined with Spearman's correlation analysis. Those dimensions with the greatest magnitudes of treatment effect were assumed to be the most relevant in assessing treatment response (energy and vitality, and role limitation due to physical problems, as subsequently discussed), and these were analyzed together with the PCS and MCS of the SF-36 (Table 2). Neither the sleep fragmentation indices (ASDA arousals, respiratory arousals, and autonomic events), nor the AHI correlated significantly with the pretreatment energy and vitality dimension or the MCS. The relationships with the PCS and role limitation due to physical problems were a little stronger.

Age, obesity, and lung function were examined as possible confounding factors in these relationships. As might be expected, the PCS correlated with body mass index (BMI) (r = 0.31, p < 0.05) and with FEV₁ (r = 0.31, p < 0.05), but the ASDA arousal index remained independently associated with the PCS even after allowing for these confounding factors, despite the factors' own correlation with OSA severity (residual variance 10%, p < 0.02).

In contrast to the pretreatment values, all of the improvements in the health status dimensions of the SF-36 with treatment correlated significantly with the sleep fragmentation indices and AHI (Table 2), with the closest correlation being between the ASDA arousal index and PCS (Figure 1). In general, these correlations were closer than those of the pretreatment measures, with little difference between the various sleep fragmentation indices. To ensure that these correlations were not artifactually produced by a deterioration in health status in the less severely affected subjects, a paired t test was performed in a subgroup of patients with an AHI of less than 5. Even in this group there was a significant improvement in energy and vitality (p < 0.05), PCS (p < 0.0001), and MCS (p < 0.0001); however, the magnitude of effect was clinically significant only in the energy and vitality dimension (see below). There was no significant deterioration in any of the other dimensions.

View larger version (18K): [in this window] [in a new window]   Figure 1.   Scatter plot showing relationship between improvement in the physical component summary (PCS) of the SF-36 with nCPAP and the ASDA arousal index.

SF-36 before and with nCPAP. Table 3 shows the results for the individual dimensions of the SF-36, PCS, and MCS both before and with nCPAP. The pretreatment values were compared with data from a community population (16) (corrected for age, sex, and social class), also shown in Table 3. Both the pretreatment energy and vitality dimension (p < 0.0001) and pretreatment physical function dimension (p < 0.05) were significantly lower in the study group and returned to normal with nCPAP.

The effect sizes for each dimension were calculated as described earlier and are shown in Table 4. The largest effect sizes were seen in the energy and vitality dimension (0.69) and role limitation due to physical problems (0.50). These effect sizes were smaller than those reported elsewhere (5). This is probably because the study population included subjects who did not have OSA (e.g., nonsnorers) and who were only receiving nCPAP for the purposes of this study. If the subjects with mild or no OSA are excluded (AHI < 10), the effect sizes become greater, with the largest effect sizes seen in the same dimensions: energy and vitality (0.80) and role limitation due to physical problems (0.66). The PCS and MCS also showed moderately large effect sizes (0.69 and 0.80, respectively).

A post hoc analysis was done of effect sizes in the subjects at the mild end of the sleep-disordered breathing spectrum (AHI < 10 and AHI < 5), and the results are also shown in Table 4. In these two groups, clinically significant effects were seen only in the energy and vitality dimensions.

Compliance with nCPAP. Compliance with nCPAP correlated neither with pretreatment health status measured with the SF-36 (PCS: r = 0.11; MCS: r = 0.16; energy and vitality: r = 0.19; role limitation due to physical problems: r = 0.02; p > 0.1 for all Spearman's correlations) nor with the change in SF-36 dimensions with treatment (PCS: r = 0.11; MCS: r = 0.12; energy and vitality: r = 0.20; role limitation due to physical problems: r = 0.15; p > 0.1 for all Spearman's correlations). Even when subjects with less severe disease (AHI < 10) were excluded, compliance correlated neither with pretreatment health status (PCS: r = 0.10; MCS: r = 0.20; energy and vitality: r = 0.25; role limitation due to physical problems: r = 0.03; p > 0.1 for all Spearman's correlations) nor with the change in SF-36 dimensions with treatment (PCS: r = 0.09; MCS: r = 0.14; energy and vitality: r = 0.09; role limitation due to physical problems: r = 0.01; p > 0.1 for all Spearman's correlations).

Relationship between SF-36 dimensions and daytime sleepiness. The relationships between the SF-36 dimensions and subjective and objective daytime sleepiness are shown in Table 5. Subjective sleepiness correlated significantly with all dimensions studied. The OSLER test correlated weakly with both energy and vitality and with PCS.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION CONCLUSIONS REFERENCES

This study investigated the best predictors of improvement in health status (as measured with the SF-36) with nCPAP, and explored the relationship between health status and daytime sleepiness in a diverse group of patients presenting to a sleep clinic. In keeping with the findings of others (2), health status as measured with the SF-36 was impaired in this group, which had a wide range of OSA severity, and improved with nCPAP treatment. The relationship between the SF-36 dimensions and OSA severity as assessed both with the AHI and other sleep fragmentation indices (ASDA arousal index, respiratory arousal index, and autonomic events) was poor. To avoid the confounding effects of coexisting conditions such as obesity, cardiovascular disease, and poor lung function, the SF-36 was repeated during nCPAP and the change with treatment was calculated (representing health status impairment due to OSA) and then correlated with the pretreatment sleep fragmentation measures. The change in health status with nCPAP correlated significantly with all sleep fragmentation indices studied, although the relationship was not particularly close.

The weak relationship between pretreatment SF-36 scores and sleep fragmentation indices is not surprising, since many other factors contribute to quality of life (e.g., age, sex, social status, and coexisting disease). Many of these factors are not related to polysomnographic OSA severity at all. Some symptoms of OSA, such as snoring, nocturnal polyuria, and reduced libido may also impair health status but would not be expected to have close correlations with indices of sleep fragmentation. Even some of the daytime consequences of OSA more plausibly attributable to disturbed sleep have only weak relationships with sleep fragmentation indices (e.g., cognitive dysfunction [28], depression [29]). Furthermore, depending on the circumstances of each individual, these impairments will have diverse effects on overall subjective well-being.

Calculating the change in health status with nCPAP provided a method of quantifying the health status impairment caused by OSA alone. This measure of treatment responsiveness, in the health status dimensions studied, correlated loosely but significantly with all the sleep fragmentation indices. The weakness of these correlations was probably caused by the imperfect nature of methods of quantifying sleep fragmentation, with variable definition and problems of night-to-night variability. Another reason for the weak correlations may be that impairment of health status by OSA is also due to symptoms that are indirectly related to sleep fragmentation (e.g., snoring, reduced libido, nocturia), but which improve with nCPAP treatment. The correction of snoring alone may improve health status even if snoring is not associated with sleep fragmentation (30). Alternatively, the SF-36 may not adequately measure the general health problems of individuals with OSA, although this seems unlikely with the large effect sizes seen in this and other studies (5). However, future work needs to explore this further. Two specific tools for measuring functional status have recently been developed and may be more suited to quantifying health status in OSA. These are the Functional Outcomes of Sleep Questionnaire (31), which is designed to assess the impact of disorders of excessive sleepiness, and a disease-specific measure, the Calgary Sleep Apnea Quality of Life Index (32). These more specific health status measures may show closer relationships than we have seen in this study.

To explore the relationship between sleep fragmentation and health status, we needed to generate a large spectrum nCPAP treatment responsiveness, and this was accomplished by treating subjects who, on the basis of our sleep study, would not normally have been considered for nCPAP treatment. To ensure that this approach had not produced artifactual correlations by making subjects with less severe disease worse (by fragmenting their sleep with nCPAP), we performed a subgroup analysis on subjects at the mild end of the sleep-disordered breathing spectrum (AHI < 5), and found no deterioration in their health status. In fact, there were statistically significant improvements in energy and vitality in this group, probably due either to the correction of snoring-related sleep fragmentation or a placebo effect. The inclusion of these subjects reduced the magnitude of change in the SF-36, but this was not the main aim of the study. Despite this, energy and vitality was abnormal in our study group, and returned to normal with nCPAP treatment, with moderate treatment effect sizes in this dimension and in those relating to physical well-being (0.69 and 0.50, respectively). Because the energy and vitality dimension relates to symptoms of sleepiness, the large effect sizes in this dimension demonstrates face validity of the SF-36 in OSA.

Compliance did not correlate with baseline health status or change in health status with nCPAP, even when subjects with an AHI < 10 were excluded. This finding is in keeping with findings in similar studies of the relationship between compliance and daytime sleepiness, including earlier work in our unit (13). Engleman and colleagues (33) found no correlation between compliance and AHI or multiple sleep latency, and no correlation between nCPAP use and improvement in multiple sleep latency. These findings imply that those subjects who benefit from nCPAP are not always those who use it the most. One reason for this is probably considerable intersubject variation in the extent of nightly nCPAP use required for symptom improvement, in the same way that there is intersubject variation in the amount of normal sleep required per night to prevent daytime symptoms (34).

In the literature, sleepiness is the most widely studied symptom of the sleep apnea syndrome, and the relationship between this symptom and health status in the present study is interesting. We and others have found (35) that subjective sleepiness (ESS) correlates more closely with health status (quantified with the SF-36) than does objectively measured sleepiness. This is probably because the ESS and SF-36 are both subjective measures and both quantify symptoms in the recent past. In contrast, measures of objective sleepiness quantify sleepiness during a single day. The weak correlations between sleepiness and health status emphasize that there is more to the symptom complex of OSA than just falling asleep.

	CONCLUSIONS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION CONCLUSIONS REFERENCES

This study confirmed that patients with OSA have impaired health status as compared with the general population, and that this is corrected with nCPAP treatment. Before treatment there is little correlation between the severity of health status impairment and the severity of OSA. This is probably due to confounding factors such as obesity and coexistent disease. Measuring improvement in health status with nCPAP removes the effects of these confounding factors, as is confirmed by closer relationships between the change in SF-36 dimensions with nCPAP and the pretreatment severity of OSA and sleep fragmentation. However, these relationships remain weak, emphasizing that much of the impairment in health status experienced by OSA patients may be due to factors that are not themselves closely related to OSA severity as currently measured with polysomnographic sleep studies.