Prepolysomnography evaluation can predict obstructive sleep apnea and is correlated to its severity

Background Obstructive sleep apnea (OSA) is increasingly identified as a disease with major health consequences. The limited availability of the gold standard diagnostic test, polysomnography (PSG), mandates careful clinical evaluation of suspected patients. This can allow better patient selection for referral for confirmatory diagnostic test. Objective The study aimed at identifying the importance of pre-PSG evaluation in prediction of obstructive sleep apnea and its relation to disease severity. Patients and methods A total of 170 patients were included. Detailed demographic characteristics, anthropometric measures, and comorbid conditions were recorded. Correlation to PSG results was done, and multivariate analysis was used to identify predictors of disease. Results OSA was diagnosed in 58.8% of our studied patients. The patients with OSA and notably the severe subgroup were of older age, predominantly male, and current or ex-smoker. Mean BMI was highest in the patients with severe OSA (41.99±8.92) and same for mean neck and waist circumference, both were significantly higher in patients with severe OSA. In multivariate logistic regression analysis, significant predictive factors for OSA were older age, male sex, being nonemployed, having hypertension, and larger tonsillar size. Conclusion Patient demographics, anthropometric characteristics, and presence of comorbid conditions such as hypertension are strong predictors of having OSA and justify referral for diagnostic sleep study.


Introduction
In the past two to three decades, obstructive sleep apnea (OSA) has been identified as an important cause of public health concern [1].
OSA remained underdiagnosed partly because of the costly nature and limited availability of the gold standard diagnostic test, polysomnography (PSG) [2,3].
Careful clinical evaluation may identify patients at high risk of OSA who strongly need referral to diagnostic PSG.

Study objective
The aim was to analyze the role of demographic characteristics, anthropometric measures, and comorbid conditions as predictors of OSA and their relation to disease severity in our studied population sample.

Study design
This was a cross-sectional study. It was conducted at sleep disordered breathing unit of the Chest Department.
The study was approved by the faculty research ethics committee.
An informed written consent was obtained from all study patients.

Study patients
Adult patients aged more than 18 years attending the outpatient sleep clinic were consecutively recruited, with total inclusion of 170 patients.

Methods
All participants were subjected to the following: (1) History taking.
BMI ¼ weight kg ð Þ=height m 2 À Á  [11]. Classification of patients into OSA and non-OSA and subclassification of OSA severity group was done using the international classification of sleep disorders [12].

Statistical analysis
STATA intercooled, version 14.2, was used for data analysis.
Multivariate regression analysis was used for identification of significant predictors for patients with OSA. Graphs were produced by STATA program (Stata Corp, California, USA). P value was considered significant if it was less than 0.05.

Results
The study participants (n=170) after completing their clinical evaluation were classified according to PSG results as follows: with no OSA (n=70), mild OSA (n=21), moderate OSA (n=11), and severe OSA (n=68). Table 1 shows the sociodemographic characteristic of the studied population with pairwise comparison between non-OSA group and OSA groups of various severities. All factors show significant difference except for residence.
Comparison between non-OSA and OSA with various severity groups showed that mean BMI, NC, and WC were significantly higher in severe OSA in comparison with non-OSA and mild OSA groups. Almost 56% of patients with severe OSA were obese grade III ( Table 2).
More than half of patients with OSA (54%) had FTP grade III. Highest mean Friedman OSAS scoring value was measured in patients with severe OSA, and the difference in Friedman OSAS score was significant between mild and severe case (Table 3).
Regarding comorbidities, diabetes mellitus, hypertension, and coronary artery disease were significantly more common in OSA than non-OSA patients (Fig. 1).
In this study, factors that showed significance in the univariate analysis were subjected to multivariate regression analysis with calculation of adjusted odds ratio with 95% confidence interval. Identification of independent variables for OSA prediction was done by final model.  Age, sex, hypertension, occupation, and tonsillar size score remained predictive of OSA in the final model (Tables 4 and 5).

Discussion
This study was conducted to detect the value of demographic, clinical, and comorbid characteristics  Comorbidities in the studied population.
in prediction of OSA and its severity, which was diagnosed by the PSG.
The study recruited 170 patients.
OSA was diagnosed in 100 (58.9%) of our studied individuals. Other nationally published studies reported that 80% and 87.1 of their recruited patients were diagnosed as patients with OSA, respectively [13,14].
The mean age of patients with OSA was significantly higher than non-OSA patients. Higher age showed 1.07-fold increased risk of having OSA. In agreement with our results, Martins et al. [15] proved that age is an independent risk factor for OSA with the assumption that this can be explained by the age-related reduction in muscle tone that results in decrease diameters of upper airway lumen. Researcher proved that the age effect was obesity independent [16].
Nearly two-thirds of the studied patients were female (65.88%). Male sex was associated with a 3.23-fold higher risk of having OSA and increased risk of severity of the disease in the present study. Consistent reporting of higher OSA prevalence in men and association of OSA with male sex was found in the literature [17].
Explanation for this was made by different adipose tissue distribution, anatomical feature of upper airways, and different muscle function in men. Moreover, leptin and sex hormones exert their own endocrine effects in men [18].
Most of the patients in OSA in this study were not employed (73%), with the risk of OSA being 8.20 higher than those who were employed. This may reflect the sequence of unhealthy sedentary life with subsequent weight gain. However, workers at stressful jobs were proved to be also at risk of OSA in a published research [19].
We found more current smokers in the OSA group (22%) than in the non-OSA group (7.14) and the difference was statistically significant. However, smoking status did not show significance in the final model of regression analysis. Controversy was found in the literature about relation of smoking and OSA. Some researcher found it a nonsignificant risk factor for reporting sleep disordered breathing symptoms [20], whereas others reported that current and past smokers were significantly at higher risk of OSA [21].
Obesity is known as a major risk factor for OSA. The pathogenesis may be mediated by the primary effect of obesity in the form of fat deposition subcutaneously and intraluminal in the upper airway. This changed the upper airway compliance and predispose to its collapse [22]. The secondary effects of obesity on lung functions  also mediate its pathogenic role in OSA [23]. This study used three measurement tools of body morphology: BMI, NC, and WC.
The study results showed that mean BMI was higher, and percentage of obese type II and III was higher in OSA than non-OSA participants. The difference in these parameters showed significance in relation to disease severity. This finding was in in accordance with many studies in the literature that documented the association between higher BMI and increased risk of OSA [24,25].
Our results revealed that higher tonsil size score according to Friedman classification was associated with nearly four-fold increased risk of having OSA in the final model of multivariate analysis. These findings are in accordance with the evidence in the literature that there is good correlation between the Friedman classification and the apnea-hypopnea index [10].
Regarding comorbidities, diabetes mellitus was found to be more prevalent in the OSA group than in non-OSA group and in particular in the severe OSA subgroup, and the difference was statistically significant. There is accumulating evidence in the literature about the association between OSA and diabetes mellitus. Some researchers reached to a conclusion in their review about this patient that sleep duration and quality are linked to the glycemic control in patients with type 2 diabetes [26].
In the final analysis model, hypertension was associated with 6.98-fold increased risk of having OSA. Hypertension and cardiovascular consequences are known to have important implications in the coexistence of obesity and OSA [27]. It is assumed that OSA is an essential differential diagnostic consideration in obese hypertensive patients [28].
In this study, factors identified by the final model of regression analysis as predictors for OSA were age, male sex, hypertension, and tonsillar size. These results are close to the published research work [29], as the researchers in this work concluded that sex, age, hypertension, and obesity are independent risk factors for OSA in their studied Saudi population, which is similar to results from western studies.

Study strength
OSA was confirmed in our study group by PSG, which adds reliability to our study results. Many published studies assessed the risk factors of OSA while utilizing validated simple questionnaires only without performing PSG [20,30].

Study limitation
One of the study limitations was that radiographic measure to directly quantify fat deposition in the upper airway was not done. However, this can be argued that we aimed at utilizing clinical evaluation and noncostly investigation to define their predictive ability. Other limitation was that it is a single-center study.

Conclusion
Detailed clinical evaluation, anthropometric measures, and presence of comorbidity could identify predictive factors for OSA and help in patient selection for referral for PSG diagnostic sleep study, which is relatively expensive and not widely available.

Financial support and sponsorship
Nil.

Conflicts of interest
There are no conflicts of interest.