The effectiveness of the epworth sleepiness scale as an auxiliary resource in the diagnosis of obstructive sleep apnea syndrome

Resumen en: Objective: To analyze the effectiveness of the Epworth Sleepiness Scale (ESS) as an auxiliary resource in the diagnosis of the Obstructive Sleep Apnea Sy...


INTrODUCTION
The obstructive sleep apnea syndrome (OSAS) is defined by the occurrence of several episodes of apnea, with interruption of air flow of 10-second duration at minimum observed during nocturnal sleep and more than 30 times in a range of 7 hours (1) . It is estimated that OSAS affects 1.2% of women and 3.9% of men worldwide, with the ratio of 3:1 for male-female incidence (2) . The ratio equals after menopause and may increase with age (3) .
The apnea-hypopnea (AHI) is used to classify OSAS, being obtained through polysomnography, by dividing the sleep hours into the total number of respiratory events. OSAS is classified as: mild -between 5 and 14.9 events / hour of sleep; moderate -between 15 and 30 events / hour of sleep, and severe -more than 30 events / hour of sleep (3,4) .
The gold standard for diagnosis of OSAS is the assisted polysomnography (PSG), and its essential parameters established by electroencephalogram, electrooculogram and electromyogram (5,6) . Through PSG, are evaluated type, frequency and duration of apnea, snoring characteristics, the degree of oxygen desaturation, cardiac arrhythmias, sleep fragmentation and severity of the condition (7) . The big problem of polysomnography is that it is not always well accepted by the patient, as it is an expensive and difficult to access exam (6) .
In order to contribute to the evaluation of the OSAS condition, an Australian physician, Dr. W. Johns Murray, developed in 1991 the Epworth Sleepiness Scale (ESS) (6) . The ESS consists of a questionnaire that aims to quantify the likelihood of falling asleep in eight everyday situations, with values varying from 0 point (minimum) to 24 points (maximum), where 10 is considered the normality limit (6) . According to the author of the scale, the ESS is not subjective, but like any other method, the questionnaire will depend on the understanding, interpretation and honesty in the patient is responses. It's quick and easy to apply and without costs (6) .
Bertolazi and colleagues performed a validation of the scale to Portuguese and confirmed ESS as a valid and reliable instrument for the assessment of daytime sleepiness (8) .
The present study aimed to analyze the effectiveness of the Epworth Sleepiness Scale as an aid in the diagnosis of the obstructive sleep apnea syndrome.

MeThODs
This is an observational study, comprising a retrospective and a prospective step, of quantitative analysis, with exploratory and diagnostic purposes. This research followed the precepts on research involving human subjects and was approved by ethics commitee of University of Fortaleza (Opinion, 278/07) (9) .
The study included a total of 475 patients (n = 475), who were selected at the Sleep Study Center in Fortaleza (Centro de Estudo do Sono de Fortaleza -CESF). This site was chosen for study because it is a clinic specializing in the study object and has a large influx of patients with sleep disorders. The survey was conducted with patients who sought the CESF for probable diagnosis of a sleep disorder and included males and females of all ages. The study excluded cases of incomplete data, patients with concomitant diseases (heart or thyroid disease, chronic obstructive pulmonary disease and neurological conditions) and patients undergoing uvulopalatopharyngoplasty.
Data collection included the medical records of patients whose admission occurred in the period between October Data was collected from medical records, which included questionnaires designed by CESF professionals, based on models used in other polysomnography laboratories and recommended by the Brazilian Society of Sleep. Questionnaires were filled in previously by the patients on the day of the exam. These questionnaires contained initially general information for registration, such as age, sex, education and marital status.
Subsequently, specific questions concerning the presleep period were applied, which reported the events of the last 6 months related to the patient's daily life, such as sleep quality, memory, concentration, libido, mood, medications, smoking and use of drugs and alcohol. Then, inquiries were made about existing illnesses that could be related to OSAS, such as hypertension, heart disease, diabetes, among others.
The Epworth Sleepiness Scale (Table I), included among the questionnaires used in the institution, assesses the likelihood of patients dozing or falling asleep in some everyday situations.
The study compared the results obtained in the Epworth scale to the polysomnography data. The Epworth Sleepiness Scale scores zero to someone who has no chance of dozing, 1 for small chance of dozing, 2 for moderate chance of dozing and 3 for high chance of dozing. This score is evaluated on 8 everyday situations: sitting and reading; watching TV;  sitting quietly in a public place without activity (waiting room, cinema, theater, meeting); as a passenger on a train, car or bus, walking an hour without stop; laying to rest in the afternoon, when circumstances allow; sitting and talking to someone; sitting quietly after lunch without drinking alcohol and, if you are driving, while stopping for a few minutes in heavy traffic. The outcome was measured by adding the scores obtained in all the listed situations. It is important that the patient responds honestly, trying to differentiate the condition of being only tired. According to Echavarría (10) and Gus (11) , a score greater than 10 is indicative of sleep disorders.
Data was tabulated and analyzed using the software Statistical Package For Social Sciences®, version 15.0 for Windows® (SPSS Inc., Chicago, IL, USA), based on descriptive and inferential statistics, presented in the form of graphs, tables and/or frames. The average was used considering the standard deviation and chi-square test was applied in the intersection of variables. Values of p < 0.05 were considered statistically significant.

resULTs
The study included 475 patients, 61.3% (n = 291) were male and 38.7% (n = 184) were female. Patients' characteristics and mean values for ESS are described in Table II. With regard to fragmented sleep, 48.8% (n = 232) showed this disorder. The alveolar hypoventilation syndrome was present in only 4.4% (n = 21) of cases. While disorders such as restless legs syndrome, insomnia and bruxism were, respectively, present in 5.9% (n = 28), 3.8% (n = 18) and 2.9% (n = 14) of the studied subjects.
Through statistical analysis, it was found that 38.9% (n = 185) of men were diagnosed with OSAS, while this percentage is lower among women, being 15.2% (n = 72). These results show a statistically significant difference (p < 0.000) (Table III).
By correlating the diagnosis of OSAS with age and BMI, it was found that age and BMI > 30 kg/m 2 are predisposing factors to the development of OSAS (p < 0.001) (Figure 1 and 2).
Yet, when values of ESS above 10 are correlated with BMI, a higher prevalence (16.6%) was found for the group with BMI between 25 and 30 kg/m 2 (Figure 3). A significant relationship was found between the ESS score and OSAS (p = 0.001), showing that 25.9% (n = 123) of patients, who had values greater than 10 in the ESS, were diagnosed with OSAS ( Figure 4).
The alveolar hypoventilation syndrome also showed a statistically significant difference when compared to the ESS (p = 0.02). But there was no statistically significant correlation between the ESS score and other variables related to sleep disorders (p > 0.05).
The sensitivity and specificity of ESS were, respectively, 47% and 64%. The positive predictive value of the scale reached 61% and the negative predictive value was 51%. The positive likelihood ratio of ESS was 134% and its accuracy was 55%.

DIsCUssION
This study corroborates the existing literature pointing to age, male gender and obesity as major factors predisposing to the development of OSAS, once it occurs most often among middle-aged men, aged between 30 and 60, with a peak incidence between 40 and 50, and is usually associated with obesity (1,3,(12)(13)(14)(15) .
The mechanisms by which excess weight influences on this syndrome have not been fully elucidated. The suspicious is that there is a decrease in the size of the pharynx and that its collapsibility increases in obese individuals (16) .
In the study of Fuhrman et al (17) and Larsson et al (18) , the male / female ratio was 2.5:1, similar to the findings of this research. Bouloukaki evaluated the prevalence of OSAS in the Greek population and found a five times higher prevalence in men (19) . Recent studies point out that women may be underdiagnosed, for they express less complaining and present distinct symptoms, besides the stigma of OSAS being considered a disease of men (20,21,18) .
Anatomical differences of the upper airways, hormone profile and the central type of fat distribution in men (trunk and neck) may justify higher incidence in males (22) . After menopause, the female becomes as affected as male (3) . Apparently, the progestin protects the pharyngeal collapse and testosterone appears to facilitate such collapse (23) .
Research has suggested that the prevalence of sleep apnea increases with age while the clinical significance of apnea (gravity) decreases with age (24) . This finding contradicts the data of our study regarding the correlation between age and the severity of OSAS, where it was observed a linear growth of gravity as age increases.
The first analysis of the European Sleep Apnea Database (ESADA) showed that 72% of patients referred for polysomnography realization were obese men, with cardiovascular and metabolic diseases (25) .
In this study, among the patients who sought the sleep assistance service, the most prevalent group was aged between 41 and 60 (45.3% of the sample). Studies  indicate that this is the stage of full professional activity, and perhaps this is why the symptoms of OSAS, such as daytime sleepiness, morning headaches and problems with memory and learning, interfere with social and working lives (3,26) .
Sleep disorders can even compromise the public safety, since they directly influence the increase in the number of industrial and traffic accidents. Estimates in the number of accidents and deaths caused by sleepiness or fatigue vary from 2% to 41%, with a high cost in financial terms and in life, itself (14) .
Due to data presented, the importance of applying the ESS is justified, in order to prevent such grievances. We believe that, through simple measures, such as performing adequate clinical history and directed physical examination, along with the implementation of the ESS, we can have a clearer idea on which patients may have sleep disorders.
Recent studies have demonstrated that ESS is a valid and reliable instrument for assessment of daytime sleepiness, thus corroborating this research (8) . According to Murray (27) , author of the scale, the ESS is an objective assessment of sleepiness. However, like any other questionnaire-based method, it depends on the skill of reading, understanding and honest response of the patient (27) .
Studies show that values above 10 in the ESS are indicative of daytime sleepiness, a major symptom of OSAS (28) . Researches have shown that the AHI was significantly associated with excessive daytime sleepiness, measured by ESS (29) .
Some studies show that patients with diagnosis of OSAS, regardless its severity (mild, moderate, or severe), have obtained a high score in the ESS, confirming the findings of this research. However, other studies found no significant correlation between the ESS score and AHI (30) . In the study by Gondim (31) , the ESS score of patients undergoing polysomnography was analyzed and it was found that 70% of patients who had an ESS score above 10 were diagnosed with OSA.
Our research is in agreement with the literature, for confirming that the ESS is not 100% sensitive (32) . A study by Echavarría (10) found that ESS had a sensitivity of 60%, specificity of 82%, positive predictive value of 85% and negative predictive value of 52%. The low sensitivity (50-60%) and low specificity (63-70%) of clinical diagnosis of sleep disorders points out a need of polysomnography for confirmation, with its sensitivity and specificity of approximately 95% (33) . The ESS can be considered as a useful proceeding for research and clinical practice and is now being used to screen for sleep disorders and also as an indicator for polysomnography (6,8) .

CONCLUsION
Data from this study shows that the Epworth Sleepiness Scale fits as an auxiliary resource in the diagnosis of the obstructive sleep apnea syndrome, which can be applied by any health professional, during the clinical case histories. However, clinical signs are not sufficient to diagnose OSAS because of the low clinical sensitivity and specificity of questionnaires used in clinical services. Professionals should refer patients for polysomnography tests when he wants confirmation of suspected sleep disorders.