ORIGINAL ARTICLE
The Clinical Respiratory Journal
Fatigue and serum testosterone in obstructive sleep apnea patients Raluca Mihaela Bercea1,2, Traian Mihaescu1,2, Cristian Cojocaru1,2 and Bjørn Bjorvatn3,4 1 ‘Grigore T. Popa’ University of Medicine and Pharmacy, Iasi, Romania 2 Clinic of Pulmonary Diseases, Iasi, Romania 3 Department of Global Public Health and Primary Care, University of Bergen, Bergen, Norway 4 Norwegian Competence Center for Sleep Disorders, Haukeland University Hospital, Bergen, Norway
Abstract Introduction: Obstructive sleep apnea (OSA)-related fatigue is a common understudied symptom. Fatigue is associated with low serum testosterone level in nonOSA patients. No data are available about this association in OSA patients. Objectives: To investigate in adult obese males affected by OSA, the relationship between fatigue and serum testosterone in order to identify predictors for OSArelated fatigue. Methods: Fifteen OSA patients and 15 control subjects participated. The parameters analyzed were serum testosterone morning concentration, polysomnography parameters, daytime sleepiness (Epworth Sleepiness Scale) and fatigue (Multidimensional Fatigue Inventory). Regression test was applied in order to show predictors of fatigue. Kruskal–Wallis test followed by post-hoc analysis was performed to test for differences between controls and OSA subgroups for testosterone, fatigue components and sleepiness. Results: Mean testosterone level was 3.55 ± 0.7 ng/mL in the OSA group, significantly lower than in controls (4.26 ± 1.1 ng/mL, P = 0.049). An inverse correlation was found between testosterone and fatigue scores (P < 0.01). Furthermore, a statistically significant difference was found between the control group and the severe OSA subgroup for general fatigue, physical fatigue, reduced activity and mental fatigue. However, no significant differences were found between controls and mild OSA. Among all variables, testosterone was the only independent significant predictor of physical fatigue (t = −2.56, P = 0.033, R = 0.978, R2 = 0.958) and reduced activity (t = −4.41, P = 0.002, R = 0.966, R2 = 0.934) in the OSA patients. Conclusions: OSA-related fatigue was strongly associated with serum testosterone, together with OSA severity.
Key words fatigue – obstructive sleep apnea – sleepiness – testosterone
Please cite this paper as: Bercea RM, Mihaescu T, Cojocaru C and Bjorvatn B. Fatigue and serum testosterone in obstructive sleep apnea patients. Clin Respir J 2014; ••: ••–••. DOI:10.1111/crj.12150.
Ethics All subjects gave written informed consent prior to their inclusion in the study, and the study was approved by the local ethical committee in accordance with the standards laid down in the Declaration of Helsinki.
Correspondence Raluca Mihaela Bercea, MD, PhD, ‘Grigore T. Popa’ University of Medicine and Pharmacy, Clinic of Pulmonary Diseases, 30 Dr. I. Cihac Street, 700115 Iasi, Romania. Tel: 0047 46575799 Fax: 0040 332815550 email: [email protected] Received: 27 February 2013 Revision requested: 15 March 2014 Accepted: 08 April 2014 DOI:10.1111/crj.12150 Authorship and contributorship R.M. Bercea and T. Mihaescu designed the study and performed research; R.M. Bercea and C. Cojocaru collected data and performed the statistics; C. Cojocaru, T. Mihaescu and B. Bjorvatn provided input for the tables and figures; R.M. Bercea and B. Bjorvat analyzed data and wrote the paper. All authors have seen and approved the final version of the manuscript.
Conflict of interest The authors (R.M.B., T.M., C.C. and B.B.) have stated explicitly that there are no conflicts of interest in connection with this article.
The Clinical Respiratory Journal (2014) • ISSN 1752-6981 © 2014 John Wiley & Sons Ltd
1
Fatigue and testosterone in sleep apnea patients
Bercea et al.
Introduction
Materials and methods
Obstructive sleep apnea (OSA) is a common disorder, possibly affecting more than 10% of the adult population (1, 2). Typical symptoms and signs of OSA are loud snoring, breathing pauses during sleep, as well as daytime sleepiness and fatigue. However, although fatigue is a common symptom of sleep apnea (3), fatigue in OSA patients remains understudied. Chervin explored different symptoms reported by OSA patients, such as sleepiness, tiredness, fatigue and lack of energy, and showed that the patients more frequently reported problems with fatigue, tiredness and lack of energy than sleepiness (4). When required to select the one most significant symptom, more patients chose lack of energy (about 40%) than any other problem, including sleepiness (about 22%) (4). Moreover, Bailes et al. showed that fatigue-related dysfunction was more accentuated than sleepiness impairments in sleep apnea patients (5). While excessive daytime sleepiness is typically considered the primary clinical symptom of OSA (6), fatigue in sleep apnea has been related more with psychiatric or other medical conditions, such as depression (3, 7). Bardwell et al. showed that depressive symptoms are involved in fatigue development, rather than oxygen saturation levels or respiratory pattern, in OSA patients (8). Fatigue has also been reported as a major symptom associated with low testosterone level in non-OSA patients, among other symptoms such as reduced libido, reduced muscle mass and strength, increased adiposity, osteoporosis and depressed mood (9, 10). Moreover, previous data report that serum total testosterone concentrations are modified after chronic exposure to smoking (11, 12). However, no studies are available about the association between fatigue and low testosterone in OSA patients. Previous data link low serum testosterone in OSA patients to hypoxia, sleep fragmentation and obesity (13, 14), as well as to accentuated depression in OSA (15). What if fatigue in OSA may have other explanations than nocturnal oxygen desaturation, sleep fragmentation or the occurrence of obesity and depressive mood? The aim of the present study was to first investigate different aspects of fatigue in relation to OSA and OSA severity, and second, to investigate the relationship between fatigue and serum testosterone in adult males affected by OSA. Furthermore, the relationship between serum testosterone, polysomnography parameters and excessive daytime sleepiness was analyzed to identify any independent predictor for OSA-related fatigue.
This cross-sectional study was conducted between November 2011 and August 2012 in the Sleep Laboratory of the Clinic of Pulmonary Diseases (Iasi, Romania), in accordance with the declaration of Helsinki and with the approval of the ethical committee of the clinic. All Caucasian subjects, recruited among patients admitted to the centre for the study of sleep disturbances, gave their written informed consent. In total, 140 patients were evaluated, and 15 OSA patients (OSA group) were selected based on the following inclusion criteria: male, 40–60 years old, obese [body mass index (BMI) > 30 kg/m2), non-smokers, diagnosed with OSA based on overnight polysomnography (PSG) evaluation with the number of episodes of apneas and hypopneas per hour of sleep [apnea– hypopnea index(AHI)] above 5 (16). Exclusion criteria included any other acute or chronic disease, chronic medication as well as uncooperative attitude. All patients with depression or anxiety were also excluded after psychiatric evaluation. Fifteen subjects were enrolled as controls matched for age and BMI, not affected by OSA or other medical disorders. PSG parameters [AHI, oxygen desaturation index (ODI), minimum oxygen saturation (minimum SpO2) and total arousal index] measured with conventional sleep recordings (SOMNOlab software version V2.01, Weinmann, Germany) were collected between 22:00– 07:00H. All PSG data were manually validated using standard criteria; mild OSA was considered for AHI between 5 and 15 (n = 5), moderate OSA for AHI between 16 and 30 (n = 5), and severe OSA for AHI above 30 (n = 5) respiratory events per hour of sleep (16). Furthermore, in all subjects, the lung physiological function was assessed by spirometry. The total serum testosterone levels were determined in blood samples, collected in the morning, within 1 h after awakening, using a chemiluminescent method (Monobind, Inc., Lake Forest, CA, USA). Daytime sleepiness was measured with the Epworth Sleepiness Scale (ESS), a self-administered eight-item questionnaire in which a score higher than 10 indicates excessive daytime sleepiness (17). The psychometric evaluation as exclusion tool was performed using Hamilton Depression Rating Scale (HDRS) and Hamilton Anxiety Rating Scale (HARS) under specialist supervision (18, 19). In addition, fatigue was evaluated with Multidimensional Fatigue Inventory (MFI), a self-report instrument which covers different aspects of fatigue (20). It consists of 20 items organized in five scales in order to assess the following conditions: general fatigue,
2
The Clinical Respiratory Journal (2014) • ISSN 1752-6981 © 2014 John Wiley & Sons Ltd
Bercea et al.
Fatigue and testosterone in sleep apnea patients
Table 1. Somatic characteristics, polysomnography parameters and serum testosterone of selected groups
Somatic characteristics Age, years BMI, kg/m2 Polysomnography parameters AHI, events/hour Minimum SpO2, % ODI, events/hour Total arousal index/hour Biochemical tests Morning testosterone (ng/mL)
OSA (n = 15)
Controls (n = 15)
P value*
53 ± 6.6 33.3 ± 2.4
55 ± 5.9 32.4 ± 1.9
0.405 0.260
27.9 ± 20.7 82.2 ± 7.1 26.8 ± 20 36.2 ± 19.2
2.5 ± 1.3 88.6 ± 2.5 2.8 ± 0.8 9.1 ± 2.6
0.0003 0.004 0.0004 0.0001
3.55 ± 0.7
4.26 ± 1.1
0.049
Data are presented as means ± standard deviation. *Student’s t-test with correction for unequal variances (Welch test), where appropriate. Significance level: P < 0.05. OSA, obstructive sleep apnea; BMI, body mass index; AHI, apnea–hypopnea index; SpO2, pulse oximetry saturation; ODI, oxygen desaturation index.
physical fatigue, reduced activity, reduced motivation and mental fatigue.
Statistical analysis The MedCalc version 11.1.1.0., Regged-rG software was used for the statistical analysis. The descriptive characteristics of the group variables were expressed as mean values and standard deviations (SD). T-test and t-test corrected for unequal variances (Welch test) were applied in order to assess the significant differences between the OSA group and controls. For the OSA group, Spearman multiple correlation tests among polysomnographic measurements, serum testosterone level and questionnaire scores (ESS and MFI) were performed. Multiple linear regressions (using the Enter method) were performed to study the relationship among MFI items (as dependent variable) and PSG parameters, serum testosterone and ESS (as independent variable), in order to identify predictor factors for fatigue in the OSA patients. Most of the assessed variables were not normally distributed; therefore, Kruskal–Wallis test followed by post-hoc analysis for multiple comparisons was performed in order to compare the study characteristics between the control group and OSA subgroups (mild OSA, moderate OSA and severe OSA). The statistical significance was set at P < 0.05.
Results Somatic characteristics, polysomnography parameters and serum testosterone are presented in Table 1. The
The Clinical Respiratory Journal (2014) • ISSN 1752-6981 © 2014 John Wiley & Sons Ltd
age and BMI of the selected groups are similar and there was no significant difference recorded (for BMI, OSA: 33.3 ± 2.4 kg/m2 vs Controls: 32.4 ± 1.94 kg/m2, P = 0.260). As expected, the OSA group presented values significantly different from controls for all PSG parameters: AHI (OSA group: 27.9 ± 20.7 events/ hour of sleep; control group: 2.5 ± 1.3 events/hour of sleep, P = 0.0003); minimum SpO2 (OSA group: 82.2 ± 7.1; control group: 88.6 ± 2.5, P = 0.004); ODI (OSA group: 26.8 ± 20.0 events/hour of sleep; control group: 2.8 ± 0.8 events/hour of sleep, P = 0.0004); total arousal index (OSA group: 36.2 ± 19.2 events/ hour of sleep; control group: 9.1 ± 2.6 events/hour of sleep). In addition, ESS was clearly higher in the OSA groups than controls (10.40 ± 2.26 vs 4.73 ± 2.46, P < 0.0001). Total serum testosterone concentrations were significantly lower in OSA group compared with controls (OSA: 3.55 ± 0.7 ng/mL vs Controls: 4.26 ± 1.1 ng/mL, P = 0.049). In our study, we found 13.3% of the OSA patients with testosterone level under lower limit accepted, but most of the OSA patients (86.7%) has testosterone level within normal range. All control subjects have testosterone level within normal ranges. According to selection criteria, Hamilton depression and anxiety scores were in normal ranges for the entire study group. All biochemistry and haematological values, blood pressure, heart rate and pulmonary function tests were in normal ranges for all subjects included (data not shown). Kruskal–Wallis tests followed by post-hoc analysis were performed in order to assess the statistically significant differences between control group (n = 15) 3
Fatigue and testosterone in sleep apnea patients
P = 0.023 Serum testosterone P = 0.793 (ng/mL) P = 0.861 P = 0.009 5 P = 0.008 P = 0.008
4
3
2
1
0 Controls
mild OSA
moderate OSA
severe OSA
Figure 1. Serum testosterone levels in the control group and obstructive sleep apnea (OSA) subgroups: mild OSA, moderate OSA and severe OSA. Statistical analysis: Kruskal–Wallis test, P = 0.026.
and OSA subgroups (mild OSA, n = 5; moderate OSA, n = 5; severe OSA, n = 5) for serum testosterone (Fig. 1), MFI item scores (Fig. 2) and Epworth score (Fig. 3). The serum testosterone was significantly different between controls and OSA subgroups (Controls: 4.3 ± 1.1 ng/mL vs mild OSA: 4.3 ± 2.3 ng/mL vs moderate OSA: 3.5 ± 0.09 ng/mL vs severe OSA: 2.8 ± 0.5 ng/mL; Kruskal–Wallis test, null-hypothesis rejected, P = 0.026). Post-hoc analysis for all pairwise comparisons confirmed significant differences between controls and severe OSA group, but also between mild and severe OSA groups (P < 0.05) for serum testosterone (Fig. 1). General fatigue scores were assessed for difference between controls and OSA subgroups (Controls: 9.2 ± 1.7 vs mild OSA: 8.2 ± 0.4 vs moderate OSA: 10.4 ± 0.5 vs severe OSA: 15 ± 2.3; Kruskal–Wallis test, null-hypothesis rejected, P = 0.0008). Post-hoc analysis for all pairwise comparisons confirmed significant differences between controls and severe OSA group; all OSA subgroups were significantly different from each other for general fatigue (P < 0.05) (Fig. 2A). Figure 2B shows that Physical fatigue scores were significantly different between controls and OSA subgroups (Controls: 9.2 ± 1.5 vs mild OSA: 8.6 ± 0.5 vs moderate OSA: 10.8 ± 0.4 vs severe OSA: 14.8 ± 2.5; Kruskal–Wallis test, null-hypothesis rejected, P = 0.001). Post-hoc analysis for all pairwise comparisons confirmed significant differences between con-
4
Bercea et al.
trols and severe OSA group; all OSA subgroups were significantly different from each other for physical fatigue (P < 0.05) (Fig. 2B). Reduced activity scores were also significantly different (Controls: 9.3 ± 1.4 vs mild OSA: 8.4 ± 0.5 vs moderate OSA: 11.6 ± 0.5 vs severe OSA: 12.4 ± 1.5; Kruskal–Wallis test, null-hypothesis rejected, P = 0.0008). Post-hoc test for all pairwise comparisons of Reduced activity scores confirmed significant differences between controls and the OSA subgroups moderate and severe. Mild OSA was also significantly different from moderate and severe OSA (Fig. 2C). Kruskal–Wallis test for differences in Reduced Motivation score between controls and OSA subgroups was negative (null-hypothesis accepted, P = 0.080). Therefore, post-hoc tests for pairwise comparisons of OSA subgroups and controls were not performed (Controls: 9.9 ± 1.5 vs mild OSA: 9.0 ± 1.0 vs moderate OSA: 11.2 ± 1.0 vs severe OSA: 11.6 ± 2.7) (Fig. 2D). For Mental fatigue scores there was a significant difference (Controls: 9.3 ± 1.5 vs mild OSA: 8.4 ± 0.5 vs moderate OSA: 9.6 ± 1.5 vs severe OSA: 11.8 ± 1.4; Kruskal–Wallis test, null-hypothesis rejected, P = 0.023). Post-hoc test for all pairwise comparisons of Mental fatigue scores confirmed significant differences between controls and severe OSA. Similarly, there were significant differences between mild OSA and severe OSA, and between moderate OSA and severe OSA (Fig. 2E). Figure 3 shows differences in Epworth sleepiness scores between groups (Controls: 4.7 ± 2.4 vs mild OSA: 8.4 ± 1.1 vs moderate OSA: 10.4 ± 1.1 vs severe OSA: 12.4 ± 2.3, Kruskal–Wallis test, null-hypothesis rejected, P = 0.0001). Post-hoc analysis for all pairwise comparisons confirmed significant differences between controls and all OSA subgroups. Furthermore, mild OSA was significantly different from both the other OSA subgroups (Fig. 3). The statistical relationships among study variables in the OSA group were tested by Spearman’s correlation analysis and results are reported in Table 2. Data showed significant relationships between most of MFI items (General fatigue, Physical fatigue, Reduced activity and Mental fatigue) and PSG parameters, serum testosterone value and ESS score. Reduced motivation item was not correlated with any of the parameters. To determine which variable under study was able to predict fatigue in the OSA group, a multiple linear regression analysis was carried out using MFI items as dependent variables, while PSG parameters (AHI, ODI, minimum SpO2 and total arousals index), serum testosterone and ESS were studied as independent variables (Table 3).
The Clinical Respiratory Journal (2014) • ISSN 1752-6981 © 2014 John Wiley & Sons Ltd
Bercea et al.
Fatigue and testosterone in sleep apnea patients
A
D Reduced motivation
P = 0.001 P = 0.008 P = 0.008
General fatigue
16
18
14
16 14 12
P = 0.120 P = 0.008 P = 0.172
12 10
10
8
8
6
6
4
4
2
2
0
0
Controls mild OSA moderate OSA severe OSA
Controls mild OSA moderate OSA severe OSA
B
E
Physical fatigue
P = 0.001 P = 0.008 P = 0.008
20
Mental fatigue
14 12
15
P = 0.199 P = 0.365 P = 0.008
P = 0.015 P = 0.008 P = 0.012 P = 0.587 P = 0.133 P = 0.614
10 8
10
6 4
5
2 0
0 Controls mild OSA moderate OSA severe OSA
C Reduced activity
P = 0.004 P = 0.008 P = 0.520
16 14 12
Controls mild OSA moderate OSA severe OSA
P = 0.006 P = 0.008 P = 0.195
10 8 6 4 2 0 Controls
mild OSA moderate OSA severe OSA
Figure 2. Multidimensional Functional Inventory (MFI) items score in the control group and obstructive sleep apnea (OSA) subgroups: mild OSA, moderate OSA and severe OSA. Statistical analysis: Kruskal–Wallis test for General fatigue, P = 0.0008 (A); for Physical fatigue, P = 0.001 (B); for Reduced activity, P = 0.0008 (C); for Reduced motivation, P = 0.08 (D); for Mental fatigue, P = 0.023 (E).
The dependent variable General fatigue can be predicted from a linear combination of the independent variables: AHI (P = 0.0001), ODI (P = 0.006), total arousals index (P = 0.019) and serum testosterone The Clinical Respiratory Journal (2014) • ISSN 1752-6981 © 2014 John Wiley & Sons Ltd
(P = 0.0002). Minimum SpO2 and Epworth sleepiness score did not significantly add to the ability of the equation to predict General fatigue and were not included in the final equation [normality test 5
Fatigue and testosterone in sleep apnea patients
Bercea et al.
equation to predict reduced motivation and were not included in the final equation. Mental fatigue can be predicted only by minimum SpO2 (P = 0.025); no other independent variable had the ability to predict mental fatigue in OSA group studied.
P = 0.001
Epworth scores
P = 0.012 P = 0.116
16 P = 0.036 14
P = 0.001
12 P = 0.009
Discussion
10 8 6 4 2 0 Controls
mild OSA
moderate OSA
severe OSA
Figure 3. Epworth scores of the control group and obstructive sleep apnea (OSA) subgroups: mild OSA, moderate OSA and severe OSA. Statistical analysis: Kruskal–Wallis test, P = 0.0001.
(Shapiro-Wilk), Passed (P = 0.687); Constant variance test: Passed (P = 0.219); Power of performed test with alpha = 0.050: 1.000]. Physical fatigue and Reduced activity can be predicted only by serum testosterone concentrations (P = 0.033, P = 0.002, respectively); no other independent variable had the ability to predict Physical fatigue or Reduced activity in the group studied. The dependent variable Reduced motivation can be predicted from a linear combination of the independent variables: ODI (P = 0.010), minimum SpO2 (P = 0.016) and ESS (P = 0.017). AHI and serum testosterone did not significantly add to the ability of the
In the present study, we have shown a statistically significant correlation between serum testosterone and fatigue in subjects with obstructive sleep apnea. Results demonstrated that the serum testosterone level was, among other variables in this study (PSG parameters, ESS), one of the predictors of general fatigue and the only independent predictor of physical fatigue and reduced activity in the OSA group. Although the distinction between sleepiness and fatigue has been accepted as useful for the field of insomnia evaluation (21), the role of both daytime sleepiness and fatigue-related symptoms continue to be a debate in OSA (6). More recent studies found complaints of fatigue to be more common than complaints of sleepiness per se (4) and that fatigue is an important symptom of sleep apnea as well (3, 5). The present study is consistent with these previous data and show a statistically significant difference between the control group and the severe OSA subgroup for most of MFI components. However, no significant differences were found between controls and mild OSA. It seems that fatigue is clearly more common in severe OSA. In addition, we reported for the first time in the same patients affected by OSA, a statistically significant correlation between serum testosterone and fatigue components. Moreover, we have shown here that the serum testosterone represented a predictive factor for general
Table 2. Spearman’s correlation coefficients among study variables (OSA group, n = 15) AHI AHI ODI Minimum SpO2 Total arousal index Serum testosterone ESS General fatigue Physical fatigue Reduced activity Reduced motivation Mental fatigue
ODI 0.988*
0.988* −0.744** 0.936* −0.951* 0.847* 0.967* 0.923* 0.784* 0.493 0.728**
−0.737** 0.943* −0.950* 0.882* 0.959* 0.922* 0.764* 0.493 0.720**
Minimum SpO2
Total arousal index
Serum testosterone
ESS
−0.744** −0.737**
0.936* 0.943* −0.832*
−0.951* −0.950* 0.728** −0.911*
0.847* 0.882* −0.580*** 0.842* −0.857*
−0.832* 0.728** −0.580*** −0.727** −0.692** −0.412 −0.030 −0.432
−0.911* 0.842* 0.902* 0.900* 0.690** 0.393 0.752**
−0.857* −0.989* −0.965* −0.824* −0.410 −0.759**
0.841* 0.793* 0.717** 0.413 0.687**
Significance level: *, **, ***: P < 0.001, P < 0.01 and P < 0.05, respectively. AHI, apnea–hypopnea index; SpO2, pulse oximetry saturation; ODI, oxygen desaturation index; ESS, Epworth Sleepiness Scale.
6
The Clinical Respiratory Journal (2014) • ISSN 1752-6981 © 2014 John Wiley & Sons Ltd
Bercea et al.
Fatigue and testosterone in sleep apnea patients
Table 3. Multiple linear regression analysis for Multidimensional Fatigue Inventory items dependent variables in obstructive sleep apnea (OSA) group (n = 15) Independent variables AHI
ODI
Minimum SpO2
Total arousal index
Serum testosterone
−0.072 0.034 −2.121 0.066
−0.070 0.023 −2.925 0.019
−2.368 0.367 −6.445 0.0002
0.050 0.087 0.581 0.577
ESS
General fatigue
β-coefficient 0.208 −0.101 SE of β–coefficient 0.029 0.027 t 7.166 −3.638 P value 0.0001 0.006 R = 0.997 R2 = 0.995 AdjR2 = 0.992
Physical fatigue
β-coefficient 0.046 SE of β–coefficient 0.086 t 0.542 P-value 0.602 R = 0.978 R2 = 0.958 AdjR2 = 0.926
0.006 0.082 0.079 0.938
0.026 0.101 0.261 0.800
0.050 0.071 0.715 0.495
−2.793 1.090 −2.563 0.033
−0.439 0.260 −1.688 0.130
Reduced activity
β-coefficient −0.025 −0.004 SE of β–coefficient 0.071 0.068 t −0.357 −0.058 P value 0.730 0.954 R = 0.966 R2 = 0.934 AdjR2 = 0.884
0.115 0.084 1.363 0.209
0.031 0.059 0.525 0.613
−4.016 0.909 −4.415 0.002
−0.148 0.217 −0.685 0.512
Reduced Motivation
β-coefficient −0.206 SE of β–coefficient 0.115 t −1.785 P value 0.112 R = 0.913 R2 = 0.833 AdjR2 = 0.708
0.367 0.110 3.328 0.010
0.408 0.135 3.011 0.016
0.135 0.095 1.422 0.192
−0.419 1.461 −0.287 0.781
−1.037 0.349 −2.972 0.017
Mental fatigue
β-coefficient 0.182 −0.126 SE of β–coefficient 0.093 0.089 t 1.955 −1.421 P-value 0.086 0.193 R = 0.934 R2 = 0.872 AdjR2 = 0.776
0.300 0.109 2.742 0.025
0.120 0.077 1.566 0.156
−0.252 1.18 −0.214 0.835
−0.109 0.282 −0.387 0.708
Statistical analysis: data are presented as: β-coefficient (regression coefficient), SE of β-coefficient (standard error of β-coefficient), t (t-value), significance level: P < 0.05; R, multiple correlation coefficient; R2, coefficient of determination; R2, AdjR2, R2-adjusted. AHI, apnea–hypopnea index; SpO2, pulse oximetry saturation; ODI, oxygen desaturation index; ESS, Epworth Sleepiness Scale. P values < 0.05 were presented in bold style.
fatigue (in the same manner with OSA severity represented by PSG parameters) and it was the only predictor for physical fatigue and reduced activity in OSA group. Although the leading contribution of a decreased serum testosterone level to the development of fatigue in OSA patients remains to be elucidated, it is notable that low testosterone level is associated with fatigue in non-OSA patients (9, 10). Previous data suggested that fatigue is more related with psychiatric conditions in OSA patients, rather than low oxygen saturations and sleep fragmentations (3, 7, 8). However, all patients in the present study were tested for depression and anxiety, and no one suffered from such conditions or other chronic diseases. As expected, sleepiness was clearly increased in OSA, and sleepiness correlated with all other parameters taken into consideration (PSG parameters, testoster-
The Clinical Respiratory Journal (2014) • ISSN 1752-6981 © 2014 John Wiley & Sons Ltd
one and fatigue scores). However, ESS was a predictor only for the reduced motivation component of fatigue. One major limitation of the present study was the low number of patients included. Still, clear dosedependent differences between the OSA subgroups were seen. Another limitation might be lack of controlling for obesity, which has in previous studies been related to fatigue (22, 23). However, all the subjects (including controls) were obese with the same BMI, over 30 kg/m2 and this reduces obesity as a possible confounding factor. To conclude, the results of our study showed that OSA-related fatigue was strongly influenced by serum testosterone together with OSA severity and sleepiness. Serum testosterone was proven to be the only independent predictor for physical fatigue and reduced activity in OSA patients. This finding suggests that serum testosterone is of clinical importance in OSA patients. 7
Fatigue and testosterone in sleep apnea patients
Acknowledgements The authors wish to thank MASOPHRD (The Managing Authority for Sectoral Operational Programme Human Resources Development) for financial support to R.M.B. in the frame of the POSDRU/107/1.5/S/ 78702 project, entitled ‘Inter-university partnership for increasing the medical doctoral research quality and interdisciplinary through doctoral scholarships – DocMed.net’.
References 1. Hrubos-Strøm H, Randby A, Namtvedt SK, et al. A Norwegian population-based study on the risk and prevalence of obstructive sleep apnea. The Akershus Sleep Apnea Project (ASAP). J Sleep Res. 2011;20: 162–70. 2. Tsara V, Amfilochiou A, Papagrigorakis MJ, Georgopoulos D, Liolios E. Guidelines for diagnosis and treatment of sleep-related breathing disorders in adults and children. Definition and classification of sleep-related breathing disorders in adults: different types and indications for sleep studies (Part 1). Hippokratia. 2009;13: 187–91. 3. Hossain JL, Ahmad P, Reinish LW, Kayumov L, Hossain NK, Shapiro C. Subjective fatigue and subjective sleepiness: two independent consequences of sleep disorders? J Sleep Res. 2005;14: 245–53. 4. Chervin RD. Sleepiness, fatigue, tiredness, and lack of energy in obstructive sleep apnea. Chest. 2000;118: 372–9. 5. Bailes S, Libman E, Baltzan M, Grad R, Kassissia I, Creti L, Rizzo D, Amsel R, Fichten CS. Fatigue: the forgotten symptom of sleep apnea. J Psychosom Res. 2011;70: 346–54. 6. Roehrs T, Carskadon MA, Dement WC, Roth T. Daytime sleepiness and alertness. In: Kryger MH, Roth T, Dement WC, editors. Principles and Practice of Sleep Medicine, 3rd edn. Philadelphia, W. B. Saunders Company, 2000: 43–52. 7. Jackson ML, Stough C, Howard ME, Spong J, Downey LA, Thompson B. The contribution of fatigue and sleepiness to depression in patients attending the sleep laboratory for evaluation of obstructive sleep apnea. Sleep Breath. 2011;15: 439–45. 8. Bardwell WA, Moore P, Ancoli-Israel S, Dimsdale JE. Fatigue in obstructive sleep apnea: driven by depressive symptoms instead of apnea severity? Am J Psychiatry. 2003;160: 350–5. 9. Bassil N. Late-onset hypogonadism. Med Clin North Am. 2011;95: 507–23.
8
Bercea et al.
10. Hyde Z, Flicker L, Almeida OP, Hankey GJ, McCaul KA, Chubb SA, Yeap BB. Low free testosterone predicts frailty in older men: the health in men study. J Clin Endocrinol Metab. 2010;95: 3165–72. 11. Ravnborg TL, Jensen TK, Andersson AM, Toppari J, Skakkebaek NE, Jørgensen N. Prenatal and adult exposures to smoking are associated with adverse effects on reproductive hormones, semen quality, final height and body mass index. Hum Reprod. 2011;26: 1000–11. 12. Svartberg J, Jorde R. Endogenous testosterone levels and smoking in men. The fifth Tromso study. Int J Androl. 2007;30: 137–43. 13. Luboshitzky R, Aviv A, Hefetz A, Herer P, Shen-Orr Z, Lavie L, Lavie P. Decreased pituitary-gonadal secretion in men with obstructive sleep apnea. J Clin Endocrinol Metab. 2003;87: 3394–8. 14. Feldman HA, Longcope C, Derby CA, Johannes CB, Araujo AB, Coviello AD, Bremner WJ, McKinlay JB. Age trends in the level of serum testosterone and other hormones in middle-aged men: longitudinal results from Massachusetts Male Aging Study. J Clin Endocrinol Metab. 2002;87: 589–98. 15. Bercea RM, Patacchioli FR, Ghiciuc CM, Cojocaru E, Mihaescu T. Serum testosterone and depressive symptoms in severe OSA patients. Andrologia. 2013;45: 345–50. 16. AASM American Academy of Sleep Medicine Task Force. Sleep-related breathing disorders in adults: recommendations for syndrome definition and measurement techniques in clinical research. Sleep. 1999;22: 667–89. 17. Johns MW. A new method for measuring daytime sleepiness: the Epworth Sleepiness Scale. Sleep. 1991;14: 540–5. 18. Hamilton M. A rating scale for depression. J Neurol Neurosurg Psychiatry. 1960;23: 56–62. 19. Hamilton M. The assessment of anxiety states by rating. Br J Med Psychol. 1959;32: 50–5. 20. Smets EM, Garssen BJ, Bonke B, De Haes JC. The Multidimensional Fatigue Inventory (MFI) psychometric qualities of an instrument to assess fatigue. J Psychosom Res. 1995;39: 315–25. 21. Buysse DJ, Ancoli-Israel S, Edinger JD, Lichstein KL, Morin CM. Recommendations for a standard research assessment of insomnia. Sleep. 2006;29: 115–73. 22. Vgontzas AN, Bixler EO, Chrousos GP. Obesity-related sleepiness and fatigue. Ann N Y Acad Sci. 2006;1083: 329–44. 23. Grieve FG, Harris DS, Fairbanks SD. Extending the fatigue severity scale to an obese population. Eat Weight Disord. 2000;5: 161–5.
The Clinical Respiratory Journal (2014) • ISSN 1752-6981 © 2014 John Wiley & Sons Ltd
The Clinical Respiratory Journal
Fatigue and serum testosterone in obstructive sleep apnea patients Raluca Mihaela Bercea1,2, Traian Mihaescu1,2, Cristian Cojocaru1,2 and Bjørn Bjorvatn3,4 1 ‘Grigore T. Popa’ University of Medicine and Pharmacy, Iasi, Romania 2 Clinic of Pulmonary Diseases, Iasi, Romania 3 Department of Global Public Health and Primary Care, University of Bergen, Bergen, Norway 4 Norwegian Competence Center for Sleep Disorders, Haukeland University Hospital, Bergen, Norway
Abstract Introduction: Obstructive sleep apnea (OSA)-related fatigue is a common understudied symptom. Fatigue is associated with low serum testosterone level in nonOSA patients. No data are available about this association in OSA patients. Objectives: To investigate in adult obese males affected by OSA, the relationship between fatigue and serum testosterone in order to identify predictors for OSArelated fatigue. Methods: Fifteen OSA patients and 15 control subjects participated. The parameters analyzed were serum testosterone morning concentration, polysomnography parameters, daytime sleepiness (Epworth Sleepiness Scale) and fatigue (Multidimensional Fatigue Inventory). Regression test was applied in order to show predictors of fatigue. Kruskal–Wallis test followed by post-hoc analysis was performed to test for differences between controls and OSA subgroups for testosterone, fatigue components and sleepiness. Results: Mean testosterone level was 3.55 ± 0.7 ng/mL in the OSA group, significantly lower than in controls (4.26 ± 1.1 ng/mL, P = 0.049). An inverse correlation was found between testosterone and fatigue scores (P < 0.01). Furthermore, a statistically significant difference was found between the control group and the severe OSA subgroup for general fatigue, physical fatigue, reduced activity and mental fatigue. However, no significant differences were found between controls and mild OSA. Among all variables, testosterone was the only independent significant predictor of physical fatigue (t = −2.56, P = 0.033, R = 0.978, R2 = 0.958) and reduced activity (t = −4.41, P = 0.002, R = 0.966, R2 = 0.934) in the OSA patients. Conclusions: OSA-related fatigue was strongly associated with serum testosterone, together with OSA severity.
Key words fatigue – obstructive sleep apnea – sleepiness – testosterone
Please cite this paper as: Bercea RM, Mihaescu T, Cojocaru C and Bjorvatn B. Fatigue and serum testosterone in obstructive sleep apnea patients. Clin Respir J 2014; ••: ••–••. DOI:10.1111/crj.12150.
Ethics All subjects gave written informed consent prior to their inclusion in the study, and the study was approved by the local ethical committee in accordance with the standards laid down in the Declaration of Helsinki.
Correspondence Raluca Mihaela Bercea, MD, PhD, ‘Grigore T. Popa’ University of Medicine and Pharmacy, Clinic of Pulmonary Diseases, 30 Dr. I. Cihac Street, 700115 Iasi, Romania. Tel: 0047 46575799 Fax: 0040 332815550 email: [email protected] Received: 27 February 2013 Revision requested: 15 March 2014 Accepted: 08 April 2014 DOI:10.1111/crj.12150 Authorship and contributorship R.M. Bercea and T. Mihaescu designed the study and performed research; R.M. Bercea and C. Cojocaru collected data and performed the statistics; C. Cojocaru, T. Mihaescu and B. Bjorvatn provided input for the tables and figures; R.M. Bercea and B. Bjorvat analyzed data and wrote the paper. All authors have seen and approved the final version of the manuscript.
Conflict of interest The authors (R.M.B., T.M., C.C. and B.B.) have stated explicitly that there are no conflicts of interest in connection with this article.
The Clinical Respiratory Journal (2014) • ISSN 1752-6981 © 2014 John Wiley & Sons Ltd
1
Fatigue and testosterone in sleep apnea patients
Bercea et al.
Introduction
Materials and methods
Obstructive sleep apnea (OSA) is a common disorder, possibly affecting more than 10% of the adult population (1, 2). Typical symptoms and signs of OSA are loud snoring, breathing pauses during sleep, as well as daytime sleepiness and fatigue. However, although fatigue is a common symptom of sleep apnea (3), fatigue in OSA patients remains understudied. Chervin explored different symptoms reported by OSA patients, such as sleepiness, tiredness, fatigue and lack of energy, and showed that the patients more frequently reported problems with fatigue, tiredness and lack of energy than sleepiness (4). When required to select the one most significant symptom, more patients chose lack of energy (about 40%) than any other problem, including sleepiness (about 22%) (4). Moreover, Bailes et al. showed that fatigue-related dysfunction was more accentuated than sleepiness impairments in sleep apnea patients (5). While excessive daytime sleepiness is typically considered the primary clinical symptom of OSA (6), fatigue in sleep apnea has been related more with psychiatric or other medical conditions, such as depression (3, 7). Bardwell et al. showed that depressive symptoms are involved in fatigue development, rather than oxygen saturation levels or respiratory pattern, in OSA patients (8). Fatigue has also been reported as a major symptom associated with low testosterone level in non-OSA patients, among other symptoms such as reduced libido, reduced muscle mass and strength, increased adiposity, osteoporosis and depressed mood (9, 10). Moreover, previous data report that serum total testosterone concentrations are modified after chronic exposure to smoking (11, 12). However, no studies are available about the association between fatigue and low testosterone in OSA patients. Previous data link low serum testosterone in OSA patients to hypoxia, sleep fragmentation and obesity (13, 14), as well as to accentuated depression in OSA (15). What if fatigue in OSA may have other explanations than nocturnal oxygen desaturation, sleep fragmentation or the occurrence of obesity and depressive mood? The aim of the present study was to first investigate different aspects of fatigue in relation to OSA and OSA severity, and second, to investigate the relationship between fatigue and serum testosterone in adult males affected by OSA. Furthermore, the relationship between serum testosterone, polysomnography parameters and excessive daytime sleepiness was analyzed to identify any independent predictor for OSA-related fatigue.
This cross-sectional study was conducted between November 2011 and August 2012 in the Sleep Laboratory of the Clinic of Pulmonary Diseases (Iasi, Romania), in accordance with the declaration of Helsinki and with the approval of the ethical committee of the clinic. All Caucasian subjects, recruited among patients admitted to the centre for the study of sleep disturbances, gave their written informed consent. In total, 140 patients were evaluated, and 15 OSA patients (OSA group) were selected based on the following inclusion criteria: male, 40–60 years old, obese [body mass index (BMI) > 30 kg/m2), non-smokers, diagnosed with OSA based on overnight polysomnography (PSG) evaluation with the number of episodes of apneas and hypopneas per hour of sleep [apnea– hypopnea index(AHI)] above 5 (16). Exclusion criteria included any other acute or chronic disease, chronic medication as well as uncooperative attitude. All patients with depression or anxiety were also excluded after psychiatric evaluation. Fifteen subjects were enrolled as controls matched for age and BMI, not affected by OSA or other medical disorders. PSG parameters [AHI, oxygen desaturation index (ODI), minimum oxygen saturation (minimum SpO2) and total arousal index] measured with conventional sleep recordings (SOMNOlab software version V2.01, Weinmann, Germany) were collected between 22:00– 07:00H. All PSG data were manually validated using standard criteria; mild OSA was considered for AHI between 5 and 15 (n = 5), moderate OSA for AHI between 16 and 30 (n = 5), and severe OSA for AHI above 30 (n = 5) respiratory events per hour of sleep (16). Furthermore, in all subjects, the lung physiological function was assessed by spirometry. The total serum testosterone levels were determined in blood samples, collected in the morning, within 1 h after awakening, using a chemiluminescent method (Monobind, Inc., Lake Forest, CA, USA). Daytime sleepiness was measured with the Epworth Sleepiness Scale (ESS), a self-administered eight-item questionnaire in which a score higher than 10 indicates excessive daytime sleepiness (17). The psychometric evaluation as exclusion tool was performed using Hamilton Depression Rating Scale (HDRS) and Hamilton Anxiety Rating Scale (HARS) under specialist supervision (18, 19). In addition, fatigue was evaluated with Multidimensional Fatigue Inventory (MFI), a self-report instrument which covers different aspects of fatigue (20). It consists of 20 items organized in five scales in order to assess the following conditions: general fatigue,
2
The Clinical Respiratory Journal (2014) • ISSN 1752-6981 © 2014 John Wiley & Sons Ltd
Bercea et al.
Fatigue and testosterone in sleep apnea patients
Table 1. Somatic characteristics, polysomnography parameters and serum testosterone of selected groups
Somatic characteristics Age, years BMI, kg/m2 Polysomnography parameters AHI, events/hour Minimum SpO2, % ODI, events/hour Total arousal index/hour Biochemical tests Morning testosterone (ng/mL)
OSA (n = 15)
Controls (n = 15)
P value*
53 ± 6.6 33.3 ± 2.4
55 ± 5.9 32.4 ± 1.9
0.405 0.260
27.9 ± 20.7 82.2 ± 7.1 26.8 ± 20 36.2 ± 19.2
2.5 ± 1.3 88.6 ± 2.5 2.8 ± 0.8 9.1 ± 2.6
0.0003 0.004 0.0004 0.0001
3.55 ± 0.7
4.26 ± 1.1
0.049
Data are presented as means ± standard deviation. *Student’s t-test with correction for unequal variances (Welch test), where appropriate. Significance level: P < 0.05. OSA, obstructive sleep apnea; BMI, body mass index; AHI, apnea–hypopnea index; SpO2, pulse oximetry saturation; ODI, oxygen desaturation index.
physical fatigue, reduced activity, reduced motivation and mental fatigue.
Statistical analysis The MedCalc version 11.1.1.0., Regged-rG software was used for the statistical analysis. The descriptive characteristics of the group variables were expressed as mean values and standard deviations (SD). T-test and t-test corrected for unequal variances (Welch test) were applied in order to assess the significant differences between the OSA group and controls. For the OSA group, Spearman multiple correlation tests among polysomnographic measurements, serum testosterone level and questionnaire scores (ESS and MFI) were performed. Multiple linear regressions (using the Enter method) were performed to study the relationship among MFI items (as dependent variable) and PSG parameters, serum testosterone and ESS (as independent variable), in order to identify predictor factors for fatigue in the OSA patients. Most of the assessed variables were not normally distributed; therefore, Kruskal–Wallis test followed by post-hoc analysis for multiple comparisons was performed in order to compare the study characteristics between the control group and OSA subgroups (mild OSA, moderate OSA and severe OSA). The statistical significance was set at P < 0.05.
Results Somatic characteristics, polysomnography parameters and serum testosterone are presented in Table 1. The
The Clinical Respiratory Journal (2014) • ISSN 1752-6981 © 2014 John Wiley & Sons Ltd
age and BMI of the selected groups are similar and there was no significant difference recorded (for BMI, OSA: 33.3 ± 2.4 kg/m2 vs Controls: 32.4 ± 1.94 kg/m2, P = 0.260). As expected, the OSA group presented values significantly different from controls for all PSG parameters: AHI (OSA group: 27.9 ± 20.7 events/ hour of sleep; control group: 2.5 ± 1.3 events/hour of sleep, P = 0.0003); minimum SpO2 (OSA group: 82.2 ± 7.1; control group: 88.6 ± 2.5, P = 0.004); ODI (OSA group: 26.8 ± 20.0 events/hour of sleep; control group: 2.8 ± 0.8 events/hour of sleep, P = 0.0004); total arousal index (OSA group: 36.2 ± 19.2 events/ hour of sleep; control group: 9.1 ± 2.6 events/hour of sleep). In addition, ESS was clearly higher in the OSA groups than controls (10.40 ± 2.26 vs 4.73 ± 2.46, P < 0.0001). Total serum testosterone concentrations were significantly lower in OSA group compared with controls (OSA: 3.55 ± 0.7 ng/mL vs Controls: 4.26 ± 1.1 ng/mL, P = 0.049). In our study, we found 13.3% of the OSA patients with testosterone level under lower limit accepted, but most of the OSA patients (86.7%) has testosterone level within normal range. All control subjects have testosterone level within normal ranges. According to selection criteria, Hamilton depression and anxiety scores were in normal ranges for the entire study group. All biochemistry and haematological values, blood pressure, heart rate and pulmonary function tests were in normal ranges for all subjects included (data not shown). Kruskal–Wallis tests followed by post-hoc analysis were performed in order to assess the statistically significant differences between control group (n = 15) 3
Fatigue and testosterone in sleep apnea patients
P = 0.023 Serum testosterone P = 0.793 (ng/mL) P = 0.861 P = 0.009 5 P = 0.008 P = 0.008
4
3
2
1
0 Controls
mild OSA
moderate OSA
severe OSA
Figure 1. Serum testosterone levels in the control group and obstructive sleep apnea (OSA) subgroups: mild OSA, moderate OSA and severe OSA. Statistical analysis: Kruskal–Wallis test, P = 0.026.
and OSA subgroups (mild OSA, n = 5; moderate OSA, n = 5; severe OSA, n = 5) for serum testosterone (Fig. 1), MFI item scores (Fig. 2) and Epworth score (Fig. 3). The serum testosterone was significantly different between controls and OSA subgroups (Controls: 4.3 ± 1.1 ng/mL vs mild OSA: 4.3 ± 2.3 ng/mL vs moderate OSA: 3.5 ± 0.09 ng/mL vs severe OSA: 2.8 ± 0.5 ng/mL; Kruskal–Wallis test, null-hypothesis rejected, P = 0.026). Post-hoc analysis for all pairwise comparisons confirmed significant differences between controls and severe OSA group, but also between mild and severe OSA groups (P < 0.05) for serum testosterone (Fig. 1). General fatigue scores were assessed for difference between controls and OSA subgroups (Controls: 9.2 ± 1.7 vs mild OSA: 8.2 ± 0.4 vs moderate OSA: 10.4 ± 0.5 vs severe OSA: 15 ± 2.3; Kruskal–Wallis test, null-hypothesis rejected, P = 0.0008). Post-hoc analysis for all pairwise comparisons confirmed significant differences between controls and severe OSA group; all OSA subgroups were significantly different from each other for general fatigue (P < 0.05) (Fig. 2A). Figure 2B shows that Physical fatigue scores were significantly different between controls and OSA subgroups (Controls: 9.2 ± 1.5 vs mild OSA: 8.6 ± 0.5 vs moderate OSA: 10.8 ± 0.4 vs severe OSA: 14.8 ± 2.5; Kruskal–Wallis test, null-hypothesis rejected, P = 0.001). Post-hoc analysis for all pairwise comparisons confirmed significant differences between con-
4
Bercea et al.
trols and severe OSA group; all OSA subgroups were significantly different from each other for physical fatigue (P < 0.05) (Fig. 2B). Reduced activity scores were also significantly different (Controls: 9.3 ± 1.4 vs mild OSA: 8.4 ± 0.5 vs moderate OSA: 11.6 ± 0.5 vs severe OSA: 12.4 ± 1.5; Kruskal–Wallis test, null-hypothesis rejected, P = 0.0008). Post-hoc test for all pairwise comparisons of Reduced activity scores confirmed significant differences between controls and the OSA subgroups moderate and severe. Mild OSA was also significantly different from moderate and severe OSA (Fig. 2C). Kruskal–Wallis test for differences in Reduced Motivation score between controls and OSA subgroups was negative (null-hypothesis accepted, P = 0.080). Therefore, post-hoc tests for pairwise comparisons of OSA subgroups and controls were not performed (Controls: 9.9 ± 1.5 vs mild OSA: 9.0 ± 1.0 vs moderate OSA: 11.2 ± 1.0 vs severe OSA: 11.6 ± 2.7) (Fig. 2D). For Mental fatigue scores there was a significant difference (Controls: 9.3 ± 1.5 vs mild OSA: 8.4 ± 0.5 vs moderate OSA: 9.6 ± 1.5 vs severe OSA: 11.8 ± 1.4; Kruskal–Wallis test, null-hypothesis rejected, P = 0.023). Post-hoc test for all pairwise comparisons of Mental fatigue scores confirmed significant differences between controls and severe OSA. Similarly, there were significant differences between mild OSA and severe OSA, and between moderate OSA and severe OSA (Fig. 2E). Figure 3 shows differences in Epworth sleepiness scores between groups (Controls: 4.7 ± 2.4 vs mild OSA: 8.4 ± 1.1 vs moderate OSA: 10.4 ± 1.1 vs severe OSA: 12.4 ± 2.3, Kruskal–Wallis test, null-hypothesis rejected, P = 0.0001). Post-hoc analysis for all pairwise comparisons confirmed significant differences between controls and all OSA subgroups. Furthermore, mild OSA was significantly different from both the other OSA subgroups (Fig. 3). The statistical relationships among study variables in the OSA group were tested by Spearman’s correlation analysis and results are reported in Table 2. Data showed significant relationships between most of MFI items (General fatigue, Physical fatigue, Reduced activity and Mental fatigue) and PSG parameters, serum testosterone value and ESS score. Reduced motivation item was not correlated with any of the parameters. To determine which variable under study was able to predict fatigue in the OSA group, a multiple linear regression analysis was carried out using MFI items as dependent variables, while PSG parameters (AHI, ODI, minimum SpO2 and total arousals index), serum testosterone and ESS were studied as independent variables (Table 3).
The Clinical Respiratory Journal (2014) • ISSN 1752-6981 © 2014 John Wiley & Sons Ltd
Bercea et al.
Fatigue and testosterone in sleep apnea patients
A
D Reduced motivation
P = 0.001 P = 0.008 P = 0.008
General fatigue
16
18
14
16 14 12
P = 0.120 P = 0.008 P = 0.172
12 10
10
8
8
6
6
4
4
2
2
0
0
Controls mild OSA moderate OSA severe OSA
Controls mild OSA moderate OSA severe OSA
B
E
Physical fatigue
P = 0.001 P = 0.008 P = 0.008
20
Mental fatigue
14 12
15
P = 0.199 P = 0.365 P = 0.008
P = 0.015 P = 0.008 P = 0.012 P = 0.587 P = 0.133 P = 0.614
10 8
10
6 4
5
2 0
0 Controls mild OSA moderate OSA severe OSA
C Reduced activity
P = 0.004 P = 0.008 P = 0.520
16 14 12
Controls mild OSA moderate OSA severe OSA
P = 0.006 P = 0.008 P = 0.195
10 8 6 4 2 0 Controls
mild OSA moderate OSA severe OSA
Figure 2. Multidimensional Functional Inventory (MFI) items score in the control group and obstructive sleep apnea (OSA) subgroups: mild OSA, moderate OSA and severe OSA. Statistical analysis: Kruskal–Wallis test for General fatigue, P = 0.0008 (A); for Physical fatigue, P = 0.001 (B); for Reduced activity, P = 0.0008 (C); for Reduced motivation, P = 0.08 (D); for Mental fatigue, P = 0.023 (E).
The dependent variable General fatigue can be predicted from a linear combination of the independent variables: AHI (P = 0.0001), ODI (P = 0.006), total arousals index (P = 0.019) and serum testosterone The Clinical Respiratory Journal (2014) • ISSN 1752-6981 © 2014 John Wiley & Sons Ltd
(P = 0.0002). Minimum SpO2 and Epworth sleepiness score did not significantly add to the ability of the equation to predict General fatigue and were not included in the final equation [normality test 5
Fatigue and testosterone in sleep apnea patients
Bercea et al.
equation to predict reduced motivation and were not included in the final equation. Mental fatigue can be predicted only by minimum SpO2 (P = 0.025); no other independent variable had the ability to predict mental fatigue in OSA group studied.
P = 0.001
Epworth scores
P = 0.012 P = 0.116
16 P = 0.036 14
P = 0.001
12 P = 0.009
Discussion
10 8 6 4 2 0 Controls
mild OSA
moderate OSA
severe OSA
Figure 3. Epworth scores of the control group and obstructive sleep apnea (OSA) subgroups: mild OSA, moderate OSA and severe OSA. Statistical analysis: Kruskal–Wallis test, P = 0.0001.
(Shapiro-Wilk), Passed (P = 0.687); Constant variance test: Passed (P = 0.219); Power of performed test with alpha = 0.050: 1.000]. Physical fatigue and Reduced activity can be predicted only by serum testosterone concentrations (P = 0.033, P = 0.002, respectively); no other independent variable had the ability to predict Physical fatigue or Reduced activity in the group studied. The dependent variable Reduced motivation can be predicted from a linear combination of the independent variables: ODI (P = 0.010), minimum SpO2 (P = 0.016) and ESS (P = 0.017). AHI and serum testosterone did not significantly add to the ability of the
In the present study, we have shown a statistically significant correlation between serum testosterone and fatigue in subjects with obstructive sleep apnea. Results demonstrated that the serum testosterone level was, among other variables in this study (PSG parameters, ESS), one of the predictors of general fatigue and the only independent predictor of physical fatigue and reduced activity in the OSA group. Although the distinction between sleepiness and fatigue has been accepted as useful for the field of insomnia evaluation (21), the role of both daytime sleepiness and fatigue-related symptoms continue to be a debate in OSA (6). More recent studies found complaints of fatigue to be more common than complaints of sleepiness per se (4) and that fatigue is an important symptom of sleep apnea as well (3, 5). The present study is consistent with these previous data and show a statistically significant difference between the control group and the severe OSA subgroup for most of MFI components. However, no significant differences were found between controls and mild OSA. It seems that fatigue is clearly more common in severe OSA. In addition, we reported for the first time in the same patients affected by OSA, a statistically significant correlation between serum testosterone and fatigue components. Moreover, we have shown here that the serum testosterone represented a predictive factor for general
Table 2. Spearman’s correlation coefficients among study variables (OSA group, n = 15) AHI AHI ODI Minimum SpO2 Total arousal index Serum testosterone ESS General fatigue Physical fatigue Reduced activity Reduced motivation Mental fatigue
ODI 0.988*
0.988* −0.744** 0.936* −0.951* 0.847* 0.967* 0.923* 0.784* 0.493 0.728**
−0.737** 0.943* −0.950* 0.882* 0.959* 0.922* 0.764* 0.493 0.720**
Minimum SpO2
Total arousal index
Serum testosterone
ESS
−0.744** −0.737**
0.936* 0.943* −0.832*
−0.951* −0.950* 0.728** −0.911*
0.847* 0.882* −0.580*** 0.842* −0.857*
−0.832* 0.728** −0.580*** −0.727** −0.692** −0.412 −0.030 −0.432
−0.911* 0.842* 0.902* 0.900* 0.690** 0.393 0.752**
−0.857* −0.989* −0.965* −0.824* −0.410 −0.759**
0.841* 0.793* 0.717** 0.413 0.687**
Significance level: *, **, ***: P < 0.001, P < 0.01 and P < 0.05, respectively. AHI, apnea–hypopnea index; SpO2, pulse oximetry saturation; ODI, oxygen desaturation index; ESS, Epworth Sleepiness Scale.
6
The Clinical Respiratory Journal (2014) • ISSN 1752-6981 © 2014 John Wiley & Sons Ltd
Bercea et al.
Fatigue and testosterone in sleep apnea patients
Table 3. Multiple linear regression analysis for Multidimensional Fatigue Inventory items dependent variables in obstructive sleep apnea (OSA) group (n = 15) Independent variables AHI
ODI
Minimum SpO2
Total arousal index
Serum testosterone
−0.072 0.034 −2.121 0.066
−0.070 0.023 −2.925 0.019
−2.368 0.367 −6.445 0.0002
0.050 0.087 0.581 0.577
ESS
General fatigue
β-coefficient 0.208 −0.101 SE of β–coefficient 0.029 0.027 t 7.166 −3.638 P value 0.0001 0.006 R = 0.997 R2 = 0.995 AdjR2 = 0.992
Physical fatigue
β-coefficient 0.046 SE of β–coefficient 0.086 t 0.542 P-value 0.602 R = 0.978 R2 = 0.958 AdjR2 = 0.926
0.006 0.082 0.079 0.938
0.026 0.101 0.261 0.800
0.050 0.071 0.715 0.495
−2.793 1.090 −2.563 0.033
−0.439 0.260 −1.688 0.130
Reduced activity
β-coefficient −0.025 −0.004 SE of β–coefficient 0.071 0.068 t −0.357 −0.058 P value 0.730 0.954 R = 0.966 R2 = 0.934 AdjR2 = 0.884
0.115 0.084 1.363 0.209
0.031 0.059 0.525 0.613
−4.016 0.909 −4.415 0.002
−0.148 0.217 −0.685 0.512
Reduced Motivation
β-coefficient −0.206 SE of β–coefficient 0.115 t −1.785 P value 0.112 R = 0.913 R2 = 0.833 AdjR2 = 0.708
0.367 0.110 3.328 0.010
0.408 0.135 3.011 0.016
0.135 0.095 1.422 0.192
−0.419 1.461 −0.287 0.781
−1.037 0.349 −2.972 0.017
Mental fatigue
β-coefficient 0.182 −0.126 SE of β–coefficient 0.093 0.089 t 1.955 −1.421 P-value 0.086 0.193 R = 0.934 R2 = 0.872 AdjR2 = 0.776
0.300 0.109 2.742 0.025
0.120 0.077 1.566 0.156
−0.252 1.18 −0.214 0.835
−0.109 0.282 −0.387 0.708
Statistical analysis: data are presented as: β-coefficient (regression coefficient), SE of β-coefficient (standard error of β-coefficient), t (t-value), significance level: P < 0.05; R, multiple correlation coefficient; R2, coefficient of determination; R2, AdjR2, R2-adjusted. AHI, apnea–hypopnea index; SpO2, pulse oximetry saturation; ODI, oxygen desaturation index; ESS, Epworth Sleepiness Scale. P values < 0.05 were presented in bold style.
fatigue (in the same manner with OSA severity represented by PSG parameters) and it was the only predictor for physical fatigue and reduced activity in OSA group. Although the leading contribution of a decreased serum testosterone level to the development of fatigue in OSA patients remains to be elucidated, it is notable that low testosterone level is associated with fatigue in non-OSA patients (9, 10). Previous data suggested that fatigue is more related with psychiatric conditions in OSA patients, rather than low oxygen saturations and sleep fragmentations (3, 7, 8). However, all patients in the present study were tested for depression and anxiety, and no one suffered from such conditions or other chronic diseases. As expected, sleepiness was clearly increased in OSA, and sleepiness correlated with all other parameters taken into consideration (PSG parameters, testoster-
The Clinical Respiratory Journal (2014) • ISSN 1752-6981 © 2014 John Wiley & Sons Ltd
one and fatigue scores). However, ESS was a predictor only for the reduced motivation component of fatigue. One major limitation of the present study was the low number of patients included. Still, clear dosedependent differences between the OSA subgroups were seen. Another limitation might be lack of controlling for obesity, which has in previous studies been related to fatigue (22, 23). However, all the subjects (including controls) were obese with the same BMI, over 30 kg/m2 and this reduces obesity as a possible confounding factor. To conclude, the results of our study showed that OSA-related fatigue was strongly influenced by serum testosterone together with OSA severity and sleepiness. Serum testosterone was proven to be the only independent predictor for physical fatigue and reduced activity in OSA patients. This finding suggests that serum testosterone is of clinical importance in OSA patients. 7
Fatigue and testosterone in sleep apnea patients
Acknowledgements The authors wish to thank MASOPHRD (The Managing Authority for Sectoral Operational Programme Human Resources Development) for financial support to R.M.B. in the frame of the POSDRU/107/1.5/S/ 78702 project, entitled ‘Inter-university partnership for increasing the medical doctoral research quality and interdisciplinary through doctoral scholarships – DocMed.net’.
References 1. Hrubos-Strøm H, Randby A, Namtvedt SK, et al. A Norwegian population-based study on the risk and prevalence of obstructive sleep apnea. The Akershus Sleep Apnea Project (ASAP). J Sleep Res. 2011;20: 162–70. 2. Tsara V, Amfilochiou A, Papagrigorakis MJ, Georgopoulos D, Liolios E. Guidelines for diagnosis and treatment of sleep-related breathing disorders in adults and children. Definition and classification of sleep-related breathing disorders in adults: different types and indications for sleep studies (Part 1). Hippokratia. 2009;13: 187–91. 3. Hossain JL, Ahmad P, Reinish LW, Kayumov L, Hossain NK, Shapiro C. Subjective fatigue and subjective sleepiness: two independent consequences of sleep disorders? J Sleep Res. 2005;14: 245–53. 4. Chervin RD. Sleepiness, fatigue, tiredness, and lack of energy in obstructive sleep apnea. Chest. 2000;118: 372–9. 5. Bailes S, Libman E, Baltzan M, Grad R, Kassissia I, Creti L, Rizzo D, Amsel R, Fichten CS. Fatigue: the forgotten symptom of sleep apnea. J Psychosom Res. 2011;70: 346–54. 6. Roehrs T, Carskadon MA, Dement WC, Roth T. Daytime sleepiness and alertness. In: Kryger MH, Roth T, Dement WC, editors. Principles and Practice of Sleep Medicine, 3rd edn. Philadelphia, W. B. Saunders Company, 2000: 43–52. 7. Jackson ML, Stough C, Howard ME, Spong J, Downey LA, Thompson B. The contribution of fatigue and sleepiness to depression in patients attending the sleep laboratory for evaluation of obstructive sleep apnea. Sleep Breath. 2011;15: 439–45. 8. Bardwell WA, Moore P, Ancoli-Israel S, Dimsdale JE. Fatigue in obstructive sleep apnea: driven by depressive symptoms instead of apnea severity? Am J Psychiatry. 2003;160: 350–5. 9. Bassil N. Late-onset hypogonadism. Med Clin North Am. 2011;95: 507–23.
8
Bercea et al.
10. Hyde Z, Flicker L, Almeida OP, Hankey GJ, McCaul KA, Chubb SA, Yeap BB. Low free testosterone predicts frailty in older men: the health in men study. J Clin Endocrinol Metab. 2010;95: 3165–72. 11. Ravnborg TL, Jensen TK, Andersson AM, Toppari J, Skakkebaek NE, Jørgensen N. Prenatal and adult exposures to smoking are associated with adverse effects on reproductive hormones, semen quality, final height and body mass index. Hum Reprod. 2011;26: 1000–11. 12. Svartberg J, Jorde R. Endogenous testosterone levels and smoking in men. The fifth Tromso study. Int J Androl. 2007;30: 137–43. 13. Luboshitzky R, Aviv A, Hefetz A, Herer P, Shen-Orr Z, Lavie L, Lavie P. Decreased pituitary-gonadal secretion in men with obstructive sleep apnea. J Clin Endocrinol Metab. 2003;87: 3394–8. 14. Feldman HA, Longcope C, Derby CA, Johannes CB, Araujo AB, Coviello AD, Bremner WJ, McKinlay JB. Age trends in the level of serum testosterone and other hormones in middle-aged men: longitudinal results from Massachusetts Male Aging Study. J Clin Endocrinol Metab. 2002;87: 589–98. 15. Bercea RM, Patacchioli FR, Ghiciuc CM, Cojocaru E, Mihaescu T. Serum testosterone and depressive symptoms in severe OSA patients. Andrologia. 2013;45: 345–50. 16. AASM American Academy of Sleep Medicine Task Force. Sleep-related breathing disorders in adults: recommendations for syndrome definition and measurement techniques in clinical research. Sleep. 1999;22: 667–89. 17. Johns MW. A new method for measuring daytime sleepiness: the Epworth Sleepiness Scale. Sleep. 1991;14: 540–5. 18. Hamilton M. A rating scale for depression. J Neurol Neurosurg Psychiatry. 1960;23: 56–62. 19. Hamilton M. The assessment of anxiety states by rating. Br J Med Psychol. 1959;32: 50–5. 20. Smets EM, Garssen BJ, Bonke B, De Haes JC. The Multidimensional Fatigue Inventory (MFI) psychometric qualities of an instrument to assess fatigue. J Psychosom Res. 1995;39: 315–25. 21. Buysse DJ, Ancoli-Israel S, Edinger JD, Lichstein KL, Morin CM. Recommendations for a standard research assessment of insomnia. Sleep. 2006;29: 115–73. 22. Vgontzas AN, Bixler EO, Chrousos GP. Obesity-related sleepiness and fatigue. Ann N Y Acad Sci. 2006;1083: 329–44. 23. Grieve FG, Harris DS, Fairbanks SD. Extending the fatigue severity scale to an obese population. Eat Weight Disord. 2000;5: 161–5.
The Clinical Respiratory Journal (2014) • ISSN 1752-6981 © 2014 John Wiley & Sons Ltd
