Quality of life in patients with obstructive sleep apnea

Abstract

The aim of this study was to investigate the relative contributions of daytime sleepiness, sleep quality, depression, and apnea severity to mental and physical quality of life (QoL) in obstructive sleep apnea (OSA) patients. This was a cross-sectional study. Participants were adults diagnosed with OSA. Medical Outcomes Study–Short Form 36 (SF-36), Epworth Sleepiness Scale (ESS), Medical Outcomes Study–Sleep Scale, and Beck Depression Inventory (BDI) were used. The factors predicting the physical and mental QoL were evaluated using multiple linear regression analysis. Seven hundred ninety three OSA patients participated in the study. The average age was 48.9 years (SD = 11.7 years). The mean apnea–hypopnea index (AHI) was 29.5 hour⁻¹ (SD = 20.6 hour⁻¹). The SF-36 scores were 72.6 (SD = 18.5). The BDI, sleep quality, and age were related to both mental and physical QoL. However, ESS, minimal arterial oxygen saturation, gender, and body mass index were associated with the physical but not mental QoL. The BDI was the strongest predictor of both physical and mental QoL. AHI was related to neither physical nor mental QoL. The potential factors affecting QoL are different between physical and mental dimensions of QoL. Depressive mood was the strongest predictor of both the physical and mental QoL.

Introduction

Obstructive sleep apnea (OSA) is a common sleep disorder.¹ Recurrent episodes of airway obstruction cause intermittent periods of oxygen desaturation and arousals from sleep, which make sleep fragmented.² Untreated OSA is often associated with cardiovascular comorbidity.^3,4 Symptoms reported by OSA patients include fatigue, daytime sleepiness, poor sleep quality, impaired concentration, memory loss, headache, and mood and affect disturbance.⁵ They have a significant impact on the individual's health-related quality of life (QoL).

Assessing QoL is one of the most important outcome measures in the current managed care. The relations between QoL in OSA patients and individual contributing factor such as daytime sleepiness or depression have received considerable attention.⁶ For example, in a large population-based study, subjective sleepiness was found to be significantly related to the diminished QoL.⁷ Poor sleep quality in OSA patients is also associated with their decreased QoL.⁸ Depression is also frequently present in OSA patients⁹ and is considered an important determinant of QoL in OSA⁶ as well as other diseases.¹⁰ However, few studies have assessed the extent to which these factors independently contribute to QoL in OSA patients.⁶ In order to formulate strategies to improve QoL, we need to understand the variables predictive of QoL in this patient population.

In addition, the association of polysomnography (PSG) indices and OSA-related QoL is still of concern. Although treatment of OSA with continuous positive airway pressure (CPAP) or surgery has shown to improve QoL in literature,¹¹ apnea severity usually measured by apnea–hypopnea index (AHI) has failed to show a correlation with QoL.^12,13 This suggests that the AHI may not be the most appropriate measures of apnea severity when studying the clinical impact of OSA on subjective QoL. However, few studies have determined whether other PSG indices such as desaturation profiles are strongly predictive of QoL.¹⁴

We investigated determinants of poor QoL in adults with newly diagnosed, untreated OSA. The aims of this study were threefold: (1) to determine the relative contributions of the potential predictors including daytime sleepiness, poor sleep quality, and depression in OSA patients; (2) to determine whether mental and physical domains of QoL are affected differently by these potential factors; and (3) to determine whether some PSG indices are independently associated with QoL in OSA patients.

Materials and methods

Subjects

This was a cross-sectional study with adult patients who visited sleep laboratories for the evaluation of suspected OSA between 2010 and 2012. They were recruited from a single tertiary hospital in Korea. Criteria for inclusion were as follows: being over 18 years of age, undertaking an overnight PSG, and being diagnosed with OSA (AHI ≥ 5 hour⁻¹). Patients were excluded if they had a periodic limb movement arousal index ≥5 hour⁻¹, they had an active or chronic psychiatric or medical disorder based on the self-reported checklist for medical history, if they were previously diagnosed with and treated for OSA or if they filled out a set of sleep-related questionnaires incompletely. We did not exclude the patients whose Beck Depression Inventory (BDI) scores were over the threshold of depression if they were not previously diagnosed as psychiatric disorder or were taking medication for the treatment of their condition. Hypertensive or diabetic patients without overt cardiovascular complication were not excluded. The study was reviewed and approved by the institutional review board of the Asan Medical Center.

Data collection

Patients completed a battery of sleep-related questionnaires that was routinely administered to all patients undergoing PSG at our sleep laboratory. Basic demographic information, medical comorbidities, and medication information were obtained from the questionnaire and patient's electronic medical record. Body mass index (BMI) was determined on the day of the PSG.

Objective apnea severity was evaluated using AHI, respiratory distress index (RDI), oxygen desaturation index (ODI), and minimal arterial oxygen saturation (MinSaO₂). The AHI was defined as the average number of episodes of apnea and hypopnea per hour.¹⁵ The RDI was defined as the average number of episodes of apnea, hypopnea, and respiratory effort-related arousals per hour. The ODI was defined as the number of times that oxygen desaturation was ≥3% per hour of sleep.¹⁵

QoL was evaluated using the Medical Outcomes Study (MOS) Short-Form Health survey (SF-36).¹⁶ The SF-36 is a multipurpose, self-administered, and non-disease–specific health survey consisting of 36 questions divided into 8 individual domains, which are used to form two distinct high-order summary scales: the physical component summary (PCS) and the mental component summary (MCS). The PCS includes the physical functioning, role physical, bodily pain, and general health scales, and the MCS includes the vitality, social functioning, role emotional, and general mental health scales. All domain scores and summary measures are transformed, resulting in scale scores from 0 (lowest level of functioning) to 100 (highest level of functioning). A higher score indicates a better health-related QoL. The Korean version of the SF-36 was recently validated.¹⁶

Sleep quality was assessed using the MOS-Sleep Scale,¹⁷ which comprises 12 items and measures key sleep structures across 6 domains. In the assessment, subjects are asked to recall the past 4 weeks and to answer the questions based on this. The Sleep Problem Index-2 (SPI-2) was calculated from the MOS-Sleep Scale score, with a higher SPI-2 representing a more severe sleep problem. Daytime sleepiness was assessed using the 8-item Epworth Sleepiness Scale (ESS).¹⁸ Excessive daytime sleepiness was defined as an ESS score ≥11. Depressive symptoms were assessed using the BDI.¹⁹

Statistical analysis

Univariate analysis was performed to identify variables that were related to SF-36 scores. Pearson correlation was used for numeric variables, and Student's t-test was used for nominal variables. Variables with the values of p < 0.1 by the univariate analysis were then included in a multiple hierarchical linear regression analysis. The dependent variables were the PCS and MCS scores of SF-36. The independent variables included demographic variables including age, sex, and BMI; apnea severity-related variables (AHI, RDI, ODI and MinSaO₂); sleep quality (SPI-2); daytime sleepiness (ESS); and the degree of depressive mood (BDI). Log transformation was applied to variables with right-skewed distribution (AHI, RDI, and ODI) or left-skewed distribution (MinSaO₂) to satisfy the normality assumption. Two-tailed statistical significance level was set at p < 0.05. Data were analyzed using SPSS version 21.0 (SPSS Inc., Chicago, Illinois, USA).

Results

Patient characteristics

Of 1103 consecutive patients who underwent overnight PSG for suspected OSA, 937 patients were diagnosed with OSA (AHI ≥ 5 hour⁻¹). Among them, 144 patients were excluded due to periodic limb movement arousal index ≥5 hour⁻¹ (n = 37), neurologic disease (n = 18), psychiatric disease (n = 8), sleep disorder (n = 5), medical problem (n = 43), those previously diagnosed with and treated for OSA (n = 8), and incomplete data (n = 25). The remaining 793 OSA subjects (689 men and 104 women) participated in the study (Table 1). The average age was 48.9 years (SD = 11.7 years). Patient characteristics are shown in Table 1.

Table 1.

Patient characteristics (N = 793).

Male, n (%)	689 (86.9)
Age (years), mean (SD)	48.9 (11.7)
Body mass index (kg m−2), mean (SD)	26.0 (3.4)
AHI, mean (SD)	29.5 (20.6)
AHI ≥ 30, n (%)	324 (40.9)
30 > AHI ≥ 15, n (%)	235 (29.6)
15 > AHI ≥ 5, n (%)	234 (29.5)
RDI, mean (SD)	36.7 (19.3)
ODI, mean (SD)	26.3 (20.1)
MinSaO2, mean (SD)	80.4 (8.1)
PLM index (hour−1), mean (SD)	7.7 (20.4)
PLM arousal index (hour−1), mean (SD)	0.8 (2.9)
SF-36 questionnaire, mean (SD)
Total scores	72.6 (18.5)
Physical component scores	69.1 (19.5)
Mental component scores	70.1 (18.6)
ESS, mean (SD)	9.7 (5.0)
≥11, n (%)	319 (40.2)
SPI-2, mean (SD)	34.0 (16.9)
BDI, mean (SD)	10.1 (7.2)
≥10, n (%)	366 (46.2)

AHI: apnea–hypopnea index, ODI: oxygen desaturation index, PLM: periodic leg movement, RDI: respiratory distress index; MinSaO₂: minimal arterial oxygen saturation; SF-36: Medical Outcomes Study–Short Form-36; ESS: Epworth Sleepiness Scale; SPI-2: Sleep Problem Index-2; BDI: Beck Depression Inventory.

Factors associated with the PCS scores of SF-36

Based on multiple linear regression analysis, all variables with the values of p < 0.05 by the univariate analysis (Tables 2 and 3) were identified as independent factors predicting the PCS scores of SF-36: age (p = 0.003), sex (p < 0.001), BMI (p = 0.003), MinSaO₂ (p = 0.014), SPI-2 (p < 0.001), ESS (p = 0.012), and BDI (p < 0.001; Table 4). Overall, 54.2% of the variance in the physical component scores was explained by the final model. In the first step, demographic factors including age, sex, and BMI accounted for 13.1% of the variance. MinSaO₂ (β = 3.283; p = 0.014) additionally accounted for 1.2% of the variance, and SPI-2 (β = −0.246; p < 0.001) and ESS (β = −0.258; p = 0.012) in the third step additionally accounted for 20.3% of the variance. The BDI (β = −1.393; p < 0.001) in the last step additionally explained 19.6% of the variance in the physical component scores. Based on the standardized coefficient, the strongest predictor of the physical component scores was BDI scores. The second important predictor was the SPI-2 scores.

Table 2.

Correlation between QoL scores and independent variables.

	Physical component scores	Mental component scores
AHI	0.041	0.063
RDI	0.056	0.075a
ODI	0.040	0.071a
MinSaO2	0.070a	0.072a
Age	−0.178b	−0.065
BMI	−0.087a	−0.055
SPI-2	−0.505b	−0.520b
ESS	−0.235b	−0.207b
BDI	−0.679b	−0.751b

PCS: physical component summary; MCS: mental component summary; AHI: apnea–hypopnea index; QoL: quality of life; RDI: respiratory distress index; ODI: oxygen desaturation index; MinSaO₂: minimal arterial oxygen saturation; BMI: body mass index; SPI-2: Sleep Problem Index-2; ESS: Epworth Sleepiness Scale; BDI: Beck Depression Inventory.

^ap < 0.05.

^bp < 0.01.

Table 3.

Gender differences in SF-36 scores of OSA patients.

	Men (n = 689)	Women (n = 104)	p Value
Total scores	74.6 ± 17.3	59.1 ± 20.5	<0.001
Physical component scores	71.5 ± 17.8	52.8 ± 22.1	<0.001
Mental component scores	71.9 ± 17.9	58.6 ± 19.5	<0.001

OSA: obstructive sleep apnea, SF-36: Medical Outcomes Study–Short Form 36.

Table 4.

Hierarchical linear regression analysis of the factors associated with QoL in patients with OSA.^a

		Non-standardized coefficient	SE	Standardized coefficient	p Value	R2 change	p Value
Physical component scores
Step 1	Age	−0.125	0.042	−0.075	0.003	0.131	0.000
	Sex	−8.104	1.501	−0.140	0.000
	BMI	−0.451	0.151	−0.080	0.003
Step 2	MinSaO2	3.283	1.333	0.065	0.014	0.012	0.001
Step 3	SPI-2b	−0.246	0.033	−0.213	0.000	0.203	0.000
	ESS	−0.258	0.102	−0.066	0.012
Step 4	BDI	−1.393	0.076	−0.516	0.000	0.196	0.000
Mental component scores
Step 1	Age	0.076	0.037	0.048	0.042	0.055	0.000
	Sex	−2.569	1.319	−0.046	0.052
Step 2	MinSaO2	1.756	1.094	0.036	0.109	0.003	0.099
Step 3	SPI-2b	−0.224	0.029	−0.203	0.000	0.241	0.000
	ESS	−0.126	0.090	−0.034	0.164
Step 4	BDI	−1.657	0.067	−0.644	0.000	0.307	0.000

BMI: body mass index; SPI-2: Sleep Problem Index-2; ESS: Epworth Sleepiness Scale; BDI: Beck Depression Inventory; MinSaO₂: minimal arterial oxygen saturation; SF-36: Medical Outcomes Study–Short Form 36; MOS: Medical Outcome Study.

^aDependent variable = the scores of SF-36 questionnaire.

^bSPI-2 of MOS-Sleep Scale

Factors associated with the MCS scores of SF-36

Variables with the values of p < 0.05 by the univariate analysis (Tables 2 and 3) were entered in a multiple linear regression analysis. Age and AHI that tended to be weakly related to MCS scores of SF-36 (both with p < 0.1) were also included in a multiple linear regression. All apnea-related PSG variables such as AHI, RDI, ODI, and MinSaO₂ were not identified as one of independent factors predicting the MCS scores. Age (p = 0.042), SPI-2 (p < 0.001), and BDI (p < 0.001) were independently related to the MCS scores (Table 4). Overall, 60.6% of the variance in the mental component scores was explained by the final model. In the first step, age accounted for 5.5% of the variance. SPI-2 in the third step additionally accounted for 24.1% of the variance. The BDI in the last step additionally explained 30.7% of the variance in the mental component scores (Table 4). Based on the standardized coefficient, the strongest predictor of the mental component scores was BDI scores. The second important predictor was the SPI-2 scores.

Discussion

We determined the relative contribution of daytime sleepiness, sleep quality, depression, and apnea severity to QoL in OSA patients. We found that depression, poor sleep quality, and age were related to both mental and physical component scores of SF-36. However, daytime sleepiness, MinSaO₂, gender, and BMI were associated with the PCS but not MCS of SF-36. The BDI was the strongest predictor of both physical and mental component scores. AHI, RDI, and ODI were related to neither MCS nor PCS of SF-36.

Depression is common in OSA patients.⁹ In this study, 46.2% of OSA patients had depressive mood, and depressive symptoms were the strongest predictor of both the physical and mental component scores of SF-36. This finding was consistent with Akashiba et al.,¹⁴ showing that depressive symptoms, the lowest arterial oxygen saturation, and daytime sleepiness were identified as independent factors predicting the SF-36 total score. Among them, depressive mood was the most important determinant of QoL. These three variables accounted for 62.2% of the total variance of SF-36 in their study. In contrast, Ye et al.²⁰ examined anxiety and depression to find determinants of QoL in OSA patients and found that anxiety but not depression was the stronger independent predictor of QoL. In their study, depression lost its statistical significance for QoL in multiple regression analysis.²⁰

Subjective sleep quality is an important determinant of health-related QoL. In the healthy Austrian population, self-assessed sleep quality was found to be closely related to QoL.⁸ Studies about the relation between sleep quality and QoL are relatively rare in the OSA population. Moore et al.²¹ found that PSG indicators of sleep quality and sleep continuity could determine many aspects of QoL in OSA patients. Lee at al.²² found that insomnia was more common (35.5%) in patients with OSA compared to 22.8% of the Korean general population²³ and that the presence of insomnia symptom had negative effects on the total scores of SF-36 only in men but not in women. Silva et al.²⁴ showed in their 5-year follow-up study that a longitudinal increase in the difficulty initiating and maintaining sleep was associated with a decline in mental but not physical components of QoL. In the present study, poor sleep quality was an important factor next to depressive symptoms and was related to both lower physical and mental QoL unlike Silva et al.²⁴

Several studies have shown a significant correlation between daytime sleepiness and QoL. Pichel et al.²⁵ found that patients treated with CPAP for a long period (18 months) showed clinically relevant improvement at posttreatment in both physical (physical functioning and general health perception) and mental (social functioning and vitality) health dimensions of SF-36. Silva et al.²⁴ also found in a longitudinal OSA study after adjusting for age, gender, and BMI that increase in daytime sleepiness was associated with decrease in both the PCS and MCS of SF-36. In contrast, the present study found that daytime sleepiness was related to the PCS but not to MCS of SF-36, after adjustment for age, gender, BMI, sleep quality, and depression. This discrepancy may be caused by different study populations and designs and/or cultural differences in perception of daytime sleepiness.

Although severe OSA determined by RDI was associated with a reduction in vitality of the SF-36 subscales in the Sleep Heart Health Study,⁷ PSG indices such as AHI or RDI have been reported to be not necessarily associated with QoL.^12,13 It suggested that AHI or RDI may not quantify some important aspects of OSA disease burden. The degree of oxygen desaturation, especially MinSaO₂, may be a better predictor of various OSA symptoms, including daytime sleepiness, depression, or QoL than the number and frequency of hypoxic events. Roure et al.²⁶ reported that MinSaO₂ was related to daytime sleepiness in OSA patients, and Mediano et al.²⁷ described that nocturnal hypoxemia measured by MinSaO₂ and mean arterial oxygen saturation can be a major determinant of daytime sleepiness in OSA patients. Similarly, the present study showed that MinSaO₂ was weakly but significantly related to the physical but not mental QoL, while AHI and RDI were not related to the scores of SF-36. This finding was similar to Akashiba et al.,¹⁴ showing that MinSaO₂ was one of the independent predictors for the total scores of SF-36, though the magnitude of its effect was small. In addition, Asghari et al.²⁸ reported that AHI did not show any correlation with QoL, while mean O₂ saturation was weakly correlated with physical domain of World Health Organization QoL instrument. Thus, although the relation between oxygen desaturation during sleep and QoL is still uncertain, it is possible that severe oxygen desaturation plays a role in the development of a reduced physical health dimension in OSA patients.

Demographic features such as age, gender, and obesity have been examined as important OSA disease pathways. The association between obesity and QoL has been investigated in different populations. Larsson et al.²⁹ found in the general population that obesity was negatively more related to physical health than mental health. Zwaan et al.³⁰ found in a cross-sectional controlled study of normal weight and obese individuals that higher BMI negatively predicted the physical but not mental dimension of QoL. In a recent longitudinal study in OSA population, Silva et al.²⁴ also showed that higher BMI was negatively related to the PCS of SF-36. Consistent with previous studies, we found that higher BMI was one of independent predictors for poorer QoL in OSA patients, and its effect was limited to physical—but not mental—component scores of SF-36. In contrast, Ye et al.²⁰ did not find that BMI was a significant predictor of QoL in OSA patients.

The relations between age or gender and QoL in OSA patients varied among the studies. Silva et al.²⁴ found, using SF-36, that older age was related to lower physical but not mental QoL and that gender was not a significant predictor for physical and mental QoL. In the present study, older age was found to be related to lower physical QoL as well as higher mental QoL, and women were more likely to have lower physical QoL but not mental QoL. Ye et al.²⁰ used disease-specific QoL measures (the Calgary sleep apnea QoL index). They found that age has a positive relationship with QoL and that women were more likely to have poorer QoL than men with OSA, especially in emotional functioning.

Certain limitations should be noted when interpreting the results of our present study. First, the design of the study was cross-sectional. This did not therefore address issues of causality. Second, our study population was derived from a single tertiary sleep laboratory. Two-thirds of our series had moderate-to-severe OSA with an AHI ≥ 15. Therefore, the extent to which the results may be generalized is limited. However, our identification of factors contributing to QoL in OSA patients should be generally applicable. Third, the SF-36 may be subject to floor and ceiling effects³¹ that could confound the relation between QoL and OSA severity.

In conclusion, the potential factors affecting QoL were different between physical and mental dimensions of QoL in OSA patients. Depressive mood was the strongest predictor of both the physical and mental QoL. AHI or RDI was not related to QoL in OSA patients.