Abstract
Background
This retrospective study of 3129 individuals aimed to screen obstructive sleep apnea (OSA) with the modified STOP-Bang questionnaire (SBQ) and Neck, Obesity, Snoring, Age, Sex (NoSAS) scores, establish more appropriate cutoff values for these 2 questionnaires in Chinese patients, and investigate whether adding minimum oxygen saturation (LSpO2) as an item could improve the ability of these questionnaires.
Material/Methods
We collected polysomnography data, demographic data, anthropometric measurements, and the SBQ and NoSAS scores from 3129 participants. The ability of the SBQ and NoSAS score with different cutoff values to detect OSA was assessed.
Results
In the SBQ, the area under the receiver operating characteristic curve (AUC) was maximized with the cutoff value of 4 in women, and the AUC values were 0.682 (apnea-hypopnea index [AHI] ≥5/h), 0.680 (AHI ≥15/h), and 0.694 (AHI v30/h). For men, with a cutoff value of 5, the AUCmax values were 0.655 (AHI ≥5/h), 0.632 (AHI ≥15/h), and 0.624(AHI ≥30/h). For women, the optimal NoSAS score cutoff values were 8 (AHI ≥5/h and AHI ≥15/h) and 10 (AHI ≥30/h), and the AUCmax values were 0.687, 0.649, and 0.662. For men, the optimal NoSAS score cutoff value was 12, and the AUCmax values were 0.628, 0.618, and 0.626. When LSpO2 <85% was added as an item to these 2 questionnaires, all AUC values increased.
Conclusions
We recommend using new cutoff values in the SBQ and NoSAS scores in Chinese patients and adding LSpO2 <85% in these 2 questionnaires for predicting OSA.
Keywords: Sleep Apnea, Obstructive; Surveys and Questionnaires; Oxygen Saturation
Introduction
Obstructive sleep apnea (OSA) is characterized by recurrent episodes of partial or total collapse of the upper airway during sleep that is caused by anatomical and non-anatomical factors and affects approximately 1 billion adults aged 30 to 69 years worldwide [1]. Prevalence rates are comparable between Western and Asian populations [2]. Previous studies have estimated the adverse health consequences of OSA, including mood disorders, cognitive dysfunction, cardiovascular disease, stroke, metabolic disease, and an increased all-cause mortality rate [3–5]. Continuous positive airway pressure is the most effective treatment for OSA, and other treatment options include oral appliances, surgical treatment, and lifestyle changes [6]. Adequate therapy has been reported to reduce the risk of major adverse events [7,8].
Full-night polysomnography is considered the criterion standard for OSA diagnosis, as it records various physiological parameters during sleep, including brain waves, blood oxygen levels, heart rate, breathing patterns, eye movements, and leg movements [9]. However, this procedure requires trained personnel, specific equipment, and dedicated space. It is also expensive, lab-intensive, and time-consuming, making it unavailable to most patients suspected of having OSA. A nationwide analysis in the United States revealed that only 20% of patients with OSA have had a diagnosis [10]. Therefore, effective and accessible screening questionnaires for potential OSA patients in clinical practice are urgently needed.
Commonly used screening tools for OSA include the STOP-BANG questionnaire (SBQ), which includes 8 items: snoring, tiredness, observed apnea, blood pressure, body mass index, age, neck circumference (NC), and sex, and the Neck, Obesity, Snoring, Age, Sex (NoSAS) score [11,12]. The SBQ is regarded as the most useful tool, and the NoSAS score is a newer but effective alternative. Both scores have high predictive values for OSA detection. For the SBQ and NoSAS, the sensitivity for predicting OSA was found to be 88% and 80%, and the specificity was found to be 53% and 58%, respectively [13,14]. However, a previous study indicated that the diagnostic accuracy of the SBQ is lower in East Asia than in other regions (0.52 vs >0.80) [15].
The original SBQ and NoSAS scoring systems were developed for Western populations [11,12]. Studies have demonstrated that ethnic differences significantly impact the prevalence of OSA, including anatomical and non-anatomical factors [6]. For example, Asian populations typically have a lower body mass index (BMI) than do Western populations [16], and even with similar a BMI, Asians are known to have higher visceral adiposity [17]. Moreover, oxygen-related parameters obtained from medical examinations and daily life are vital and readily available for assessing or predicting OSA. Therefore, in this retrospective study of 3129 individuals, we aimed to screen OSA with the modified SBQ and NoSAS scores, establish more appropriate cutoff values for these 2 questionnaires in Chinese patients, and investigate whether adding LSpO2 as an item could improve the ability for these questionnaires.
Material and Methods
Patients
A total of 3365 patients underwent full-night polysomnography at the Sleep Medicine Center at Beijing Anzhen Hospital between December 1, 2015, and December 30, 2017. The Ethics Committee of Beijing Anzhen Hospital approved the study protocol (No. 2020067X). The study’s registration number is ChiCTR2000040890. All patients provided informed consent for the anonymous use of their data.
Of these, 3129 patients were included in this retrospective clinical study, as shown in Figure 1. Chinese adults were recruited from multiple regions, with 12.1% from the southern region of China, 28.3% from the northeast region, 33.6% from the northern region (eg, Beijing), 15.9% from the northwest region, and 10.1% from the southern region, covering almost all geographical areas of China. Exclusion criteria were as follows: age under 18 years, unwillingness to attend the visit at the research center, incomplete data, presence of central sleep apnea (more than 50% central apneic events), and sleep duration of less than 180 min during the study.
Figure 1.
Flow diagram of recruitment and exclusion of participants (Microsoft Word, Microsoft Office 365, Microsoft).
General Data
We collected demographic and anthropometric measurements from patients using forms during their first visit, including date of birth, sex, height, weight, BMI, NC, history of alcohol and tobacco consumption, medical history, list of current medications taken, and place of residence.
Polysomnography
All patients underwent polysomnography, using 8 electroencephalogram channels, with recordings lasting at least 180 min (Siesta, Compumedics, Melbourne, Australia). The monitoring parameters included electroencephalogram, electromyography, electro-oculogram, electrocardiogram, blood oxygen saturation, snoring, mouth airflow, nasal airflow, chest movement, and body position. All parameters, sleep staging, and events were scored according to manual version 2.4 published by the American Academy of Sleep Medicine [9]. Patients who experienced an apnea-hypopnea index (AHI) ≥5 events/hour during sleep were considered to have OSA. Moderate OSA was defined as an AHI between 15 and 29 events/hour, and severe OSA was defined as an AHI ≥30 events/hour. The oxygen desaturation index was specified to a value of 4, as the number of 4% desaturations per hour. The evaluation of polysomnography was performed by trained, certified polysomnographic technologists certified by the American Academy of Sleep Medicine.
Questionnaires
Patients completed the following questionnaires prior to undergoing polysomnography: the SBQ, specific items from the Basic Nordic Sleep Questionnaire, and the NoSAS. The SBQ consists of 8 items regarding snoring, tiredness, observed apnea, hypertension, BMI >35 kg/m2, age >50 years, NC >40 cm, and being male, with 1 point awarded for each “yes” response. A score of ≥3 indicates that the patient is at high risk for OSA globally [11]. The NoSAS score, ranging from 0 to 17, includes 5 items, with 4 points for NC >40 cm, 3 points for BMI between 25 and 30 kg/m2 or 5 points for BMI ≥30 kg/m2, 2 points for snoring, 4 points for age >55 years, and 2 points for being male [12].
Statistical Analysis
All statistical analyses were performed using SPSS (Statistical Package for Social Sciences) version 21.0 for Windows (IBM Corp, Armonk, NY, USA). The normality of the data was assessed using the Kolmogorov-Smirnov test. The descriptive data are presented as mean±standard deviation or median with interquartile range (IQR) for numerical variables, while continuous data are expressed as frequencies and percentages. Independent sample t tests, one-way analysis of variance, and chi-square tests were used for intergroup comparisons. The receiver operating characteristic (ROC) curve was used, and the area under the ROC curve (AUC) was calculated to analyze the diagnostic performance of the SBQ and NoSAS scores for OSA. To enhance clinical utility, we systematically refined the scoring criteria for individual variables (eg, age, BMI, and NC) in the SBQ and NoSAS scores. The ability of the SBQ and NoSAS scores to detect OSA was assessed by evaluating sensitivity, specificity, and the Youden index. The logistic regression model as the classifier was used in the ROC analysis. Statistical significance was defined as a P value <0.05.
Results
Baseline Clinical Characteristics and Polysomnography Findings of Patients
Of the 3129 patients, 74.3% were men, mean age was 52.9±13.3 years, BMI was 27.9±4.3 kg/m2, and NC was 40.2±3.7 cm. OSA was diagnosed in 85.9% of patients, with 18.6% classified with mild OSA, 46.3% with moderate OSA, and 35.1% with severe OSA.
The OSA group was older than the non-OSA group, had a higher proportion of men, and had higher BMI, NC, Epworth Sleepiness Scale scores, SBQ scores, and NoSAS scores (all P<0.05). Furthermore, men had a higher NC in both groups (P<0.001). In the OSA group, women were older (P<0.001); however, men had higher AHI and oxygen desaturation index 4 values (P<0.001). The clinical characteristics and polysomnography findings of the patients are presented in Table 1.
Table 1.
Baseline clinical characteristics and polysomnography findings of patients according to the sleep laboratory report.
| Total (n=3129) | Non-OSA Group (n=441) | P value# | OSA Group (n=2688) | P value# | |||
|---|---|---|---|---|---|---|---|
| Women (n=185) | Men (n=256) | Women (n=620) | Men (n=2068) | ||||
| Physiological characteristics | |||||||
| Age; years | 52.9±13.3 | 50.4±14.2 | 49.2±15.3 | .410 | 59.9±12.2 | 51.5±12.5 | <0.001* |
| BMI; kg/m2 | 27.9±4.3 | 24.9±3.9 | 25.8±3.3 | .008* | 28.4±5.1 | 28.2±4.0 | .23 |
| Neck circumference; cm | 40.2±3.7 | 35.0±2.9 | 40.1±2.8 | <0.001* | 37.0±3.0 | 41.6±3.0 | <0.001* |
| Questionnaire scores | |||||||
| ESS score | 9 (6, 9) | 7 (5, 10) | 7 (5, 11) | .428 | 8 (5, 11) | 9 (6, 13) | <0.001* |
| SBQ score | 4 (3, 5) | 3 (2, 4) | 4 (3, 5) | <0.001* | 3 (3, 4) | 5 (4, 5) | <0.001* |
| High risk;% | 88.1 | 64.8 | 81.3% | <0.001* | 75.8 | 95.8 | <0.001* |
| NoSAS score | 11 (7, 13) | 5 (2, 7) | 8 (6, 11) | <0.001* | 9 (6, 9) | 11 (8, 13) | <0.001* |
| High risk;% | 69.0 | 18.4 | 59.7% | <0.001* | 51.5 | 80.0 | <0.001* |
| Polysomnography parameters | |||||||
| AHI; event/h | 26.5±23.0 | 2.2±1.5 | 2.4±1.4 | .134 | 27.6±21.6 | 31.4±22.6 | <0.001* |
| AHI ≥5 event/h; % | 85.5 | / | / | 77.5 | 88.2 | <0.001* | |
| AHI ≥15 event/h; % | 63.7 | / | / | 53.9 | 67.1 | <0.001* | |
| AHI ≥30 event/h; % | 35.0 | / | / | 24.5 | 38.6 | <0.001* | |
| ODI 4; event/h | 25.8±21.9 | 2.5±2.3 | 2.8±3.8 | .351 | 27.5±21.6 | 30.2±21.2 | .005* |
| CT90; % | 8.8±7.5 | 0.6±5.2 | 0.5±5.3 | .159 | 10.3±20.5 | 10.0±17.8 | .75 |
| Mean oxygen saturation; % | 94.2±3.2 | 96.9±1.9 | 96.0±1.9 | <0.001* | 94.0±3.5 | 93.9±3.0 | .44 |
| Min oxygen saturation; % | 81.8±9.3 | 91.5±3.9 | 90.4±3.7 | .003* | 80.9±9.0 | 80.2±9.2 | .12 |
P values present the difference between women and men;
P<0.05.
ODI 4 – oxygen desaturation index 4% desaturation; ESS – Epworth Sleepiness Scale; CT90 – cumulative time percentage with oxygen saturation below 90%.
Diagnostic Performance of SBQ and NoSAS Scores Using Different Total Score Cutoff Values in Female and Male Patients with OSA
The original thresholds were set at 3 for the SBQ and 8 for the NoSAS [11,12]. However, under these thresholds, the model’s sensitivity and specificity for predicting OSA were not optimal (Tables 2, 3). To get potential points for higher AUC values and improve their practical applicability in clinic care, we tested every integer value in age from 45 to 55 years, BMI from 24 kg/m2 to 35 kg/m2, and NC from 37 cm to 43 cm in our predictive models. Then, we listed the top 6 models of AUC values at different AHI levels (Tables 4–7).
Table 2.
Comparing diagnostic performance of STOP-Bang questionnaire using different total scores cutoff values in female and male patients with obstructive sleep apnea.
| AUC (95% CI) | Se, % | Sp, % | Youden index | AUC (95% CI) | Se, % | Sp, % | Youden index | AUC (95% CI) | Se, % | Sp, % | Youden index | |||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| AHI ≥5/h | AHI ≥15/h | AHI ≥30/h | ||||||||||||
| Female | ||||||||||||||
| ≥2 | 0.570 (0.520–0.619) | 96.6 | 17.3 | 0.139 | ≥2 | 0.568 (0.528–0.607) | 98.0 | 15.5 | 0.135 | ≥2 | 0.548 (0.504–0.591) | 99.0 | 10.6 | 0.096 |
| ≥3 | 0.656 (0.607–0.705) | 87.4 | 43.8 | 0.312 | ≥3 | 0.623 (0.585–0.662) | 89.7 | 35.0 | 0.247 | ≥3 | 0.612 (0.571–0.653) | 94.7 | 27.6 | 0.223 |
| ≥4 | 0.682 (0.639–0.725) | 64.5 | 71.9 | 0.364 | ≥4 | 0.680 (0.643–0.717) | 71.4 | 64.5 | 0.359 | ≥4 | 0.694 (0.655–0.733) | 85.6 | 53.3 | 0.389 |
| ≥5 | 0.621 (0.579–0.663) | 33.4 | 90.8 | 0.242 | ≥5 | 0.640 (0.602–0.679) | 42.4 | 85.7 | 0.281 | ≥5 | 0.661 (0.617–0.706) | 56.3 | 76.0 | 0.323 |
| ≥6 | 0.538 (0.493–0.584) | 8.2 | 99.5 | 0.077 | ≥6 | 0.566 (0.526–0.605) | 17.3 | 95.8 | 0.131 | ≥6 | 0.569 (0.522–0.617) | 22.6 | 91.3 | 0.139 |
| Male | ||||||||||||||
| ≥3 | 0.564 (0.523–0.604) | 97.5 | 15.2 | 0.127 | ≥3 | 0.541 (0.516–0.566) | 96.8 | 11.4 | 0.082 | ≥3 | 0.533 (0.509–0.557) | 97.9 | 8.7 | 0.066 |
| ≥4 | 0.624 (0.584–0.664) | 86.9 | 37.9 | 0.248 | ≥4 | 0.602 (0.577–0.627) | 87.8 | 32.6 | 0.204 | ≥4 | 0.583 (0.560–0.606) | 90.7 | 25.9 | 0.166 |
| ≥5 | 0.655 (0.619–0.690) | 64.6 | 66.4 | 0.310 | ≥5 | 0.632 (0.608–0.656) | 67.3 | 59.0 | 0.263 | ≥5 | 0.624 (0.601–0.647) | 73.1 | 51.7 | 0.248 |
| ≥6 | 0.604 (0.570–0.637) | 34.4 | 86.3 | 0.207 | ≥6 | 0.615 (0.592–0.638) | 38.3 | 84.8 | 0.231 | ≥6 | 0.619 (0.595–0.643) | 44.7 | 79.2 | 0.239 |
| ≥7 | 0.544 (0.509–0.579) | 11.9 | 96.9 | 0.088 | ≥7 | 0.553 (0.529–0.577) | 14.4 | 96.5 | 0.106 | ≥7 | 0.566 (0.542–0.591) | 18.4 | 94.8 | 0.132 |
AHI – apnea-hypopnea index; Se – sensitivity; Sp – specificity.
Table 3.
Comparing diagnostic performance of NoSAS score using different total scores cut-off values in female and male patients with obstructive sleep apnea.
| AUC (95% CI) | Se, % | Sp, % | Youden index | AUC (95% CI) | Se, % | Sp, % | Youden index | AUC (95% CI) | Se, % | Sp, % | Youden index | |||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| AHI ≥5/h | AHI ≥15/h | AHI ≥30/h | ||||||||||||
| Female | ||||||||||||||
| ≥6 | 0.657 (0.609–0.706) | 85.0 | 46.5 | 0.315 | ≥6 | 0.634 (0.595–0.672) | 77.4 | 49.3 | 0.267 | ≥9 | 0.614 (0.571–0.656) | 83.2 | 39.5 | 0.227 |
| ≥7 | 0.680 (0.635–0.726) | 73.4 | 62.7 | 0.361 | ≥7 | 0.649 (0.611–0.687) | 73.4 | 56.4 | 0.298 | ≥10 | 0.662 (0.621–0.704) | 74.0 | 58.5 | 0.325 |
| ≥8 | 0.687 (0.643–0.731) | 69.4 | 68.1 | 0.375 | ≥8 | 0.649 (0.611–0.687) | 76.2 | 53.7 | 0.299 | ≥11 | 0.662 (0.621–0.704) | 74.0 | 58.5 | 0.325 |
| ≥9 | 0.646 (0.603–0.690) | 53.5 | 75.7 | 0.292 | ≥9 | 0.554 (0.515–0.594) | 15.5 | 95.3 | 0.108 | ≥12 | 0.535 (0.488–0.582) | 12.5 | 94.5 | 0.070 |
| ≥10 | 0.575 (0.531–0.619) | 19.4 | 95.7 | 0.151 | ≥10 | 0.554 (0.515–0.594) | 15.5 | 95.3 | 0.108 | ≥13 | 0.535 (0.488–0.582) | 12.5 | 94.5 | 0.070 |
| Male | ||||||||||||||
| ≥10 | 0.612 (0.573–0.650) | 76.3 | 46.1 | 0.224 | ≥10 | 0.591 (0.566–0.616) | 84.8 | 33.3 | 0.181 | ≥10 | 0.589 (0.566–0.612) | 82.6 | 35.3 | 0.179 |
| ≥11 | 0.616 (0.577–0.654) | 76.3 | 46.9 | 0.232 | ≥11 | 0.593 (0.568–0.618) | 84.8 | 33.8 | 0.186 | ≥11 | 0.590 (0.567–0.613) | 82.6 | 35.4 | 0.180 |
| ≥12 | 0.628 (0.590–0.665) | 70.9 | 54.7 | 0.256 | ≥12 | 0.618 (0.594–0.642) | 66.2 | 57.5 | 0.237 | ≥12 | 0.626 (0.602–0.649) | 65.7 | 59.4 | 0.251 |
| ≥13 | 0.540 (0.505–0.576) | 11.6 | 96.5 | 0.081 | ≥13 | 0.616 (0.593–0.640) | 64.7 | 58.6 | 0.233 | ≥13 | 0.597 (0.573–0.621) | 53.9 | 65.6 | 0.195 |
| ≥14 | 0.524 (0.488–0.560) | 7.1 | 92.9 | 0.048 | ≥14 | 0.590 (0.566–0.614) | 43.6 | 74.4 | 0.180 | ≥14 | 0.546 (0.522–0.570) | 28.3 | 80.8 | 0.091 |
AHI – apnea-hypopnea index; Se – sensitivity; Sp – specificity.
Table 4.
Comparing diagnostic performance of STOP-Bang questionnaire using a cutoff value of 4 at different body mass index, age and neck circumference level in female patients with obstructive sleep apnea.
| AUC (95% CI) | Se, % | Sp, % | Youden index | |
|---|---|---|---|---|
| AHI ≥5/h | ||||
| >26 kg/m2, >47 y, >40 cm | 0.682 (0.639–0.725) | 64.5 | 71.9 | 0.364 |
| >26kg/m2, >50 y, >40 cm | 0.679 (0.636–0.722) | 62.3 | 73.5 | 0.358 |
| >26 kg/m2, >45 y, >40 cm | 0.678 (0.634–0.722) | 65.3 | 70.3 | 0.356 |
| >26 kg/m2, >47 y, >43 cm | 0.677 (0.633–0.720) | 62.9 | 72.4 | 0.353 |
| >26 kg/m2, >47 y, >37 cm | 0.676 (0.631–0.721) | 70.3 | 64.9 | 0.352 |
| >26 kg/m2, >45 y, >43 cm | 0.674 (0.630–0.718) | 64.0 | 70.8 | 0.348 |
| AHI ≥15/h | ||||
| >28 kg/m2, >53 y, >37 cm | 0.680 (0.643–0.717) | 71.4 | 64.5 | 0.359 |
| >28 kg/m2, >50 y, >37 cm | 0.679 (0.641–0.716) | 72.9 | 62.8 | 0.357 |
| >30 kg/m2, >53 y, >37 cm | 0.678 (0.641–0.716) | 67.4 | 68.2 | 0.356 |
| >26 kg/m2, >55 y, >37 cm | 0.678 (0.640–0.715) | 74.7 | 60.8 | 0.355 |
| >28 kg/m2, >55 y, >37 cm | 0.678 (0.641–0.716) | 68.7 | 67.0 | 0.357 |
| >30 kg/m2, >47 y, >37 cm | 0.676 (0.639–0.713) | 70.2 | 65.0 | 0.352 |
| AHI ≥30/h | ||||
| >26 kg/m2, >55 y, >37 cm | 0.694 (0.655–0.733) | 85.6 | 53.3 | 0.389 |
| >26 kg/m2, >50 y, >40 cm | 0.693 (0.654–0.733) | 82.7 | 55.9 | 0.386 |
| >26 kg/m2, >50 y, >43 cm | 0.693 (0.653–0.732) | 81.3 | 57.3 | 0.386 |
| >26 kg/m2, >45 y, >40 cm | 0.692 (0.653–0.731) | 85.6 | 52.8 | 0.384 |
| >26 kg/m2, >47 y, >40 cm | 0.692 (0.653–0.731) | 84.6 | 53.8 | 0.384 |
| >26 kg/m2, >47 y, >43 cm | 0.691 (0.652–0.739) | 83.2 | 55.1 | 0.383 |
AHI – apnea-hypopnea index; Se – sensitivity; Sp – specificity.
Table 5.
Comparing diagnostic performance of STOP-Bang questionnaire using a cutoff value of 5 at different body mass index, age and neck circumference level in male patients with obstructive sleep apnea.
| AUC (95% CI) | Se, % | Sp, % | Youden index | |
|---|---|---|---|---|
| AHI ≥5/h | ||||
| >26 kg/m2, >45 y, >43 cm | 0.655 (0.619–0.690) | 64.6 | 66.4 | 0.310 |
| >30 kg/m2, >45 y, >37 cm | 0.655 (0.617–0.692) | 74.3 | 56.6 | 0.309 |
| >30 kg/m2, >47 y, >37 cm | 0.650 (0.613–0.687) | 72.9 | 57.0 | 0.299 |
| >28 kg/m2, >45 y, >37 cm | 0.649 (0.611–0.688) | 78.3 | 51.6 | 0.299 |
| >24 kg/m2, >45 y, >43 cm | 0.648 (0.611–0.685) | 71.4 | 58.2 | 0.296 |
| >32 kg/m2, >45 y, >37 cm | 0.648 (0.611–0.685) | 71.7 | 57.8 | 0.275 |
| AHI ≥15/h | ||||
| >28 kg/m2, >55 y, >40 cm | 0.632 (0.608–0.656) | 67.3 | 59.0 | 0.263 |
| >28 kg/m2, >45 y, >43 cm | 0.629 (0.606–0.653) | 62.8 | 63.1 | 0.259 |
| >28 kg/m2, >53 y, >40 cm | 0.628 (0.604–0.652) | 68.7 | 56.9 | 0.256 |
| >28 kg/m2, >47 y, >40 cm | 0.627 (0.603–0.651) | 72.6 | 52.7 | 0.253 |
| >30 kg/m2, >55 y, >40 cm | 0.627 (0.603–0.651) | 61.1 | 64.3 | 0.254 |
| >28 kg/m2, >47 y, >43 cm | 0.626 (0.602–0.649) | 61.3 | 63.8 | 0.251 |
| AHI ≥30/h | ||||
| >28 kg/m2, >55 y, >40 cm | 0.624 (0.601–0.647) | 73.1 | 51.7 | 0.248 |
| >28 kg/m2, >45 y, >43 cm | 0.624 (0.601–0.647) | 68.7 | 56.0 | 0.247 |
| >28 kg/m2, >47 y, >43 cm | 0.622 (0.598–0.645) | 67.3 | 57.1 | 0.244 |
| >30 kg/m2, >47 y, >43 cm | 0.622 (0.598–0.645) | 61.0 | 63.3 | 0.243 |
| >28 kg/m2, >53 y, >40 cm | 0.621 (0.598–0.645) | 74.5 | 49.8 | 0.243 |
| >30 kg/m2, >53 y, >40 cm | 0.621 (0.597–0.644) | 69.0 | 55.2 | 0.242 |
AHI – apnea-hypopnea index; Se – sensitivity; Sp – specificity.
Table 6.
Comparing diagnostic performance of NoSAS scores using different body mass index, age and neck circumference level in female patients with obstructive sleep apnea.
| AUC (95% CI) | Se, % | Sp, % | Youden index | |
|---|---|---|---|---|
| AHI ≥5/h | ||||
| >24 kg/m2, >50 y, >40 cm (score of ≥8) | 0.687 (0.643–0.731) | 69.4 | 68.1 | 0.375 |
| >26 kg/m2, >50 y, >40 cm (score of ≥8) | 0.687 (0.643–0.731) | 68.2 | 69.2 | 0.374 |
| >32 kg/m2, >50 y, >40 cm (score of ≥8) | 0.686 (0.642–0.730) | 66.9 | 70.3 | 0.372 |
| >28 kg/m2, >50 y, >40 cm (score of ≥8) | 0.686 (0.642–0.730) | 67.4 | 69.7 | 0.371 |
| >32 kg/m2, >50 y, >40 cm (score of ≥7) | 0.684 (0.639–0.729) | 72.4 | 64.3 | 0.367 |
| >26 kg/m2, >45 y, >40 cm (score of ≥8) | 0.684 (0.639–0.729) | 73.5 | 63.2 | 0.367 |
| AHI ≥15/h (score of ≥8) | ||||
| >30 kg/m2, >55 y, >37 cm | 0.649 (0.611–0.687) | 76.2 | 53.7 | 0.299 |
| >32 kg/m2, >55 y, >37 cm | 0.649 (0.611–0.687) | 76.2 | 53.7 | 0.299 |
| >32 kg/m2, >50 y, >40 cm | 0.649 (0.611–0.687) | 73.4 | 56.4 | 0.298 |
| >28 kg/m2, >50 y, >40 cm | 0.647 (0.609–0.685) | 73.7 | 55.7 | 0.294 |
| >26 kg/m2, >50 y, >37 cm | 0.647 (0.609–0.685) | 61.4 | 68.0 | 0.294 |
| >26 kg/m2, >50 y, >40 cm | 0.647 (0.609–0.685) | 74.4 | 54.9 | 0.293 |
| AHI ≥30/h (score of ≥10) | ||||
| >24 kg/m2, >45 y, >40 cm | 0.662 (0.621–0.704) | 74.0 | 58.5 | 0.325 |
| >24 kg/m2, >45 y, >43 cm | 0.661 (0.618–0.703) | 71.6 | 60.5 | 0.321 |
| >24 kg/m2, >50 y, >43 cm | 0.661 (0.618–0.704) | 66.8 | 65.3 | 0.321 |
| >24 kg/m2, >50 y, >40 cm | 0.659 (0.617–0.702) | 69.2 | 62.6 | 0.318 |
| >24 kg/m2, >55 y, >37 cm | 0.659 (0.617–0.701) | 75.0 | 56.8 | 0.318 |
| >24 kg/m2, >47 y, >43 cm | 0.658 (0.616–0.701) | 69.7 | 62.0 | 0.317 |
AHI – apnea-hypopnea index; Se – sensitivity; Sp – specificity.
Table 7.
Comparing diagnostic performance of NoSAS scores using different body mass index, age and neck circumference level in male patients with obstructive sleep apnea.
| AUC (95% CI) | Se, % | Sp, % | Youden index | |
|---|---|---|---|---|
| AHI ≥5/h | ||||
| >24 kg/m2, >45 y, >40 cm (score of ≥12) | 0.628 (0.590–0.665) | 70.9 | 54.7 | 0.256 |
| >24 kg/m2, >47 y, >40 cm (score of ≥12) | 0.628 (0.591–0.665) | 69.7 | 55.9 | 0.256 |
| >24 kg/m2, >45 y, >43 cm (score of ≥12) | 0.627 (0.592–0.661) | 52.7 | 72.7 | 0.254 |
| >24 kg/m2, >50 y, >40 cm (score of ≥12) | 0.624 (0.587–0.661) | 67.7 | 65.7 | 0.247 |
| >28 kg/m2, >45 y, >40 cm (score of ≥11) | 0.622 (0.583–0.661) | 80.3 | 44.1 | 0.244 |
| >26 kg/m2, >45 y, >40 cm (score of ≥12) | 0.622 (0.586–0.658) | 60.3 | 64.1 | 0.244 |
| AHI ≥15/h (score of ≥12) | ||||
| >26 kg/m2, >45 y, >40 cm | 0.618 (0.594–0.642) | 66.2 | 57.5 | 0.237 |
| >26 kg/m2, >53 y, >40 cm | 0.618 (0.594–0.642) | 58.7 | 64.9 | 0.236 |
| >26 kg/m2, >47 y, >40 cm | 0.617 (0.593–0.641) | 64.9 | 58.6 | 0.235 |
| >26 kg/m2, >50 y, >40 cm | 0.617 (0.593–0.641) | 62.7 | 60.7 | 0.234 |
| >26 kg/m2, >55 y, >40 cm | 0.617 (0.593–0.640) | 56.5 | 66.8 | 0.233 |
| >26 kg/m2, >45 y, >43 cm | 0.613 (0.590–0.636) | 44.3 | 78.3 | 0.226 |
| AHI ≥30/h (score of ≥12) | ||||
| >26 kg/m2, >53 y, >40 cm | 0.626 (0.602–0.649) | 65.7 | 59.4 | 0.251 |
| >26 kg/m2, >45 y, >40 cm | 0.623 (0.600–0.646) | 72.8 | 51.8 | 0.246 |
| >26 kg/m2, >55 y, >40 cm | 0.622 (0.599–0.646) | 63.2 | 61.3 | 0.245 |
| >26 kg/m2, >50 y, >40 cm | 0.622 (0.599–0.645) | 69.3 | 55.1 | 0.244 |
| >26 kg/m2, >47 y, >40 cm | 0.622 (0.598–0.645) | 71.4 | 52.9 | 0.243 |
| >26 kg/m2, >45 y, >30 cm | 0.617 (0.593–0.640) | 50.5 | 72.8 | 0.233 |
AHI – apnea-hypopnea index; Se – sensitivity; Sp – specificity.
For the SBQ, the AUC was maximized with thresholds of 4 in women, and the AUC values were 0.682 (AHI ≥5/h), 0.680 (AHI ≥15/h), and 0.694 (AHI ≥30/h). For men, with a cutoff value of 5, the maximum AUC values were 0.655 (AHI ≥5/h), 0.632 (AHI ≥15/h), and 0.624 (AHI ≥30/h) (Tables 4, 5). For the NoSAS, as the screening threshold increased from 6 to 14, the AUC was maximized with thresholds of 8 (AHI ≥5/h and AHI ≥15/h) and 10 (AHI ≥30/h) for OSA in women, with AUC values of 0.687, 0.649, and 0.661, respectively (Table 3). In men, when the screening threshold was set at 12, the maximum AUC values were 0.628, 0.618, and 0.626, respectively (Tables 6, 7).
Diagnostic Performance of SBQ and NoSAS with LSpO2 Compared with SBQ without LSpO2 in Female and Male Patients with OSA
Next, we added LSpO2 <85% as an item in the SBQ and NoSAS scores, to predict OSA. For women, the optimal SBQ cutoff value was 5 for predicting OSA, and the maximum AUC values were 0.706 (AHI ≥5/h), 0.717 (AHI ≥15/h), and 0.742 (AHI ≥30/h). For men, the optimal SBQ cutoff value was 6, and the maximum AUC values were 0.678 (AHI ≥5/h), 0.692 (AHI ≥15/h), and 0.697 (AHI ≥30/h). The new cutoff value for the NoSAS scores in women was 12 for predicting OSA and moderate-severe OSA, yielding AUC values of 0.703 and 0.725, respectively. For severe OSA, the cutoff value was 15, resulting in an AUC of 0.719. In men, the optimal cutoff value was 16 for predicting OSA, with maximum AUC values of 0.700, 0.693, and 0.704, respectively (Figure 2).
Figure 2.
Performance of SBQ and NoSAS with LSpO2 compared with SBQ without LSpO2 (Microsoft Excel, Microsoft Office 365, Microsoft). SBQ – STOP-Bang questionnaire; LSpO2 – minimum oxygen saturation; SBQoL – SBQ without LSpO2; SBQwL – SBQ with LSpO2; NoSASoL – NoSAS without LSpO2; NoSASwL – NoSAS with LSpO2.
Discussion
In this study, we found that using new cutoff values improved the performance of the SBQ and NoSAS scores for predicting OSA in a Chinese population, and that specific cutoff values should be used for men and for women. Moreover, adding the lowest oxygen saturation <85% as an item in these 2 scoring systems improved the diagnostic performance of OSA. The standard thresholds of the SBQ and NoSAS systems were found to be 3 and 8 in Western populations [11,12]. In the present study, we found that the cutoff value of SBQ score at 4 for women and 5 for men yielded higher AUC values and greater specificity for screening OSA, compared with the original SBQ cutoff value of 3. For the NoSAS score, the cutoff values rose to 8/10 for women and 12 for men for predicting OSA, providing higher AUC values and specificity than the cutoff value of 8. Compared with the NoSAS, the SBQ had better diagnostic performance for OSA, with higher AUC and Youden index values. Furthermore, the AUC and Youden index for women were significantly higher than those for men at all AHI levels.
Studies have shown that ethnic differences, which can arise from anatomical and non-anatomical factors, are risk factors for OSA prevalence. Anatomical factors include obesity, fat distribution (NC), and upper airway structure. Non-anatomical factors include age and sex. Items in both questionnaires, such as sex, snoring, tiredness, observed apnea, and hypertension are widely recognized as risk factors for OSA and considered as binary variables. In our models, we focused on remaining continuous variables, such as BMI, NC, and age, which can offer optimal predictive value by awarding points that modify the threshold. Therefore, we adjusted BMI cutoffs and NC cutoffs.
Obesity, a vital item in OSA screening questionnaires, is defined more strictly in Asian populations. This indicates that more appropriate cutoff values can be necessary to enhance the accuracy of questionnaires for Asian populations. According to reports by the World Health Organization [16,18], the BMI threshold for obesity in Western populations (BMI ≥30 kg/m2) is significantly higher than that for Asian populations (BMI ≥28 kg/m2). In our study, the mean BMI was 27.9 kg/m2, with 26.1% of patients having a BMI of 30 kg/m2 or higher, and only 5.8% of OSA patients having a BMI of 35 kg/m2 or higher. Consistent with 2 previous studies [19,20], we observed that cutoff values of 30 kg/m2 or higher are inappropriate for the Chinese population. Thus, using an excessively high BMI cutoff value can reduce the effectiveness of the questionnaire by misidentifying potential patients.
Another crucial item is NC, because it is regarded as a promising tool for detecting body fat distribution than BMI. Asian populations are known to have higher NC than Western populations with similar AHI values. In our study, the mean NC was 40.2 cm, with 52.0% of patients having an NC <40 cm, including 45% who were women. Consequently, the BMI and NC cutoff values suggested by Zhang et al [21] differed from our findings, which may be attributed to variations in the research populations. Their patients had a higher mean BMI (28.5 vs 27.9 kg/m2), were younger (41.4 vs 52.9 years), and had a higher AHI (47.7 vs 26.5 events/hour), all of which can affect body fat distribution. Meanwhile, the NC threshold was 35 cm in Belgium, 36 cm in Turkey for women, 36.9 cm in Colombia for women and 41.2 cm for men, and 44 cm in Australia for men and 38 cm for women. The optimal BMI and NC cutoff values for detecting OSA may vary for different ethnicities.
The upper airway comprises the craniofacial structure and soft tissues. The main differences in craniofacial structure between Asian and Western populations include shorter mandibular length and maxillary depth, but longer mandibular ramus and total length [22], resulting in no significant difference in intramandibular volume between the different populations.
In our study, we found that the OSA diagnostic performance of the SBQ and NoSAS scores was better in women than in men, with higher specificity and sensitivity. Similar to previous studies [23–25], our findings demonstrated that OSA was more prevalent and severe in men than in women, with men being younger and having larger NC, while there was no significant difference in BMI between the 2 groups. Moreover, compared with the NoSAS score, the SBQ includes more subjective variables, such as tiredness and observed apnea, which are more common in men [26].
Men are at a higher risk for cardiovascular morbidity, including hypertension, than are women, unless women are postmenopausal, at which point the loss of protective hormonal effects can increase their risk [27]. These differences can contribute to the disparities in OSA predictive performance between the SBQ and NoSAS scores for men and women.
Furthermore, sex hormones significantly influence the occurrence of OSA. Studies [26,28] observed that the prevalence of OSA in women doubles after menopause, regardless of age or BMI, peaking at age 65, which is a decade later than in men. Age is a significant factor in the prevalence of obesity, due to various physiological changes, including a reduced metabolic rate and increased fat accumulation, particularly in central body regions. Fat distribution varies by sex, with men more likely to exhibit central or abdominal obesity, while peripheral fat distribution is more typical in women. However, for menopausal women, hormonal factors combined with age lead to a gradual diminishment of sex-based differences in fat distribution [28]. Additionally, the increased collapsibility of soft tissue and decreased activity of upper airway dilator muscles in postmenopausal women can exacerbate upper airway obstruction during sleep. A study by Kim et al [29] found that women had more collapsible airways with smaller area, diameter, and perimeter than men. The study by Xu et al [22] noted significant differences in mandibular measurements between men and women, with men exhibiting greater mandibular length, while women had greater mandibular width. Consequently, the smaller airways in women can result in greater airway resistance and ventilatory restriction.
In our study, we added LSpO2 <85% as an item in the SBQ and NoSAS scores, to predict OSA to enhance the accuracy. The AUCs, sensitivities, and specificities for men and women improved significantly, with these metrics increasing in parallel with the severity of OSA in both groups. Screening tools, including the SBQ and NoSAS scores, primarily consist of self-reported items (anthropometric indicators and subjective variables) and lack OSA-related variables. While the AHI is a critical standard metric for diagnosing OSA, it does not always reflect the severity of hypoxia experienced and is often difficult to obtain in daily life. LSpO2 is valuable for assessing the severity of OSA, and with the widespread adoption of wearable devices, routine monitoring of LSpO2 has become feasible and accurate. However, further specific studies are needed to elucidate the potential of oxygen-related parameters in OSA screening tools.
This study had certain limitations. The patients enrolled in this study were drawn from only a single sleep center, which can introduce potential sampling bias. The percentage of female patients with OSA in our study was lower than the percentage of male patients with OSA. Moreover, there were wide variations in BMI in this study. Although our patients came from almost all over the country, the composition ratio still mismatches the actual population ratio of the regions. Therefore, additional research should be conducted.
Conclusions
Using new cutoff values, the SBQ and NoSAS scores performed very well in screening for OSA among Chinese patients recruited in this study. Moreover, adding LSpO2 <85% as an item in the SBQ and NoSAS scores for predicting OSA provided excellent diagnostic efficiency, offering a simple and effective screening tool for OSA.
Acknowledgments
We thank all the patients and the staff from the Department of Otolaryngology and Sleep Medical Center, Beijing Anzhen Hospital, China, who participated in this study. We also thank Prof. Zijun Liao (Department of Epidemiology, Capital Institute of Pediatrics, Beijing, China) for her guidance on statistical consult.
Footnotes
Conflict of interest: None declared
Publisher’s note: All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher
Patient Consent: Informed consent was obtained from all individual participants included in the study.
Declaration of Figures’ Authenticity: All figures submitted have been created by the authors, who confirm that the images are original with no duplication and have not been previously published in whole or in part.
Financial support: The Research Foundation of Capital Institute of Pediatrics of China (LCYJ-2025-40), Capital’s Funds for Health Improvement and Research (2022-1-2101), and Beijing Hospitals Authority’s Ascent Plan (DFL20221102) supported this study. The sponsor had no role in the design or conduct of this research
References
- 1.Benjafield AV, Ayas NT, Eastwood PR, et al. Estimation of the global prevalence and burden of obstructive sleep apnoea: A literature-based analysis. Lancet Respir Med. 2019;7(8):687–98. doi: 10.1016/S2213-2600(19)30198-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Lyons MM, Bhatt NY, Pack AI, Magalang UJ. Global burden of sleep-disordered breathing and its implications. Respirology. 2020;25(7):690–702. doi: 10.1111/resp.13838. [DOI] [PubMed] [Google Scholar]
- 3.Bassetti CLA, Randerath W, Vignatelli L, et al. EAN/ERS/ESO/ESRS statement on the impact of sleep disorders on risk and outcome of stroke. Eur J Neurol. 2020;27(7):1117–36. doi: 10.1111/ene.14201. [DOI] [PubMed] [Google Scholar]
- 4.Chami HA, Resnick HE, Quan SF, Gottlieb DJ. Association of incident cardiovascular disease with progression of sleep-disordered breathing. Circulation. 2011;123(12):1280–86. doi: 10.1161/CIRCULATIONAHA.110.974022. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Leng Y, McEvoy CT, Allen IE, Yaffe K. Association of sleep-disordered breathing with cognitive function and risk of cognitive impairment: A systematic review and meta-analysis. JAMA Neurol. 2017;74(10):1237–45. doi: 10.1001/jamaneurol.2017.2180. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Patel SR. Obstructive sleep apnea. Ann Intern Med. 2019;171(11):ITC81-C96. doi: 10.7326/AITC201912030. [DOI] [PubMed] [Google Scholar]
- 7.Gottlieb DJ, Punjabi NM, Mehra R, et al. CPAP versus oxygen in obstructive sleep apnea. N Engl J Med. 2014;370(24):2276–85. doi: 10.1056/NEJMoa1306766. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Lajoie AC, Lafontaine AL, Kimoff RJ, Kaminska M. Obstructive sleep apnea in neurodegenerative disorders: Current evidence in support of benefit from sleep apnea treatment. J Clin Med. 2020;9(2):297. doi: 10.3390/jcm9020297. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Berry RB, Brooks R, Gamaldo CE, et al. The AASM manual for the scoring of sleep and associated events: Rules, terminology and technical specifications. Darien, IL: American Academy of Sleep Medicine; 2017. Version 2.4. [Google Scholar]
- 10.American Academy of Sleep Medicine. Hidden health crisis costing America billions Underdiagnosing and undertreating obstructive sleep apnea draining healthcare system. Mountain View, CA: Frost & Sullivan; 2016. [Google Scholar]
- 11.Chung F, Abdullah HR, Liao P. STOP-Bang questionnaire: A Practical Approach to Screen for Obstructive Sleep Apnea. Chest. 2016;149(3):631–38. doi: 10.1378/chest.15-0903. [DOI] [PubMed] [Google Scholar]
- 12.Marti-Soler H, Hirotsu C, Marques-Vidal P, et al. The NoSAS score for screening of sleep-disordered breathing: A derivation and validation study. Lancet Respir Med. 2016;4(9):742–48. doi: 10.1016/S2213-2600(16)30075-3. [DOI] [PubMed] [Google Scholar]
- 13.Chen H, Zheng Z, Chen R, et al. A meta-analysis of the diagnostic value of NoSAS in patients with sleep apnea syndrome. Sleep Breath. 2022;26(2):519–31. doi: 10.1007/s11325-021-02410-3. [DOI] [PubMed] [Google Scholar]
- 14.Bernhardt L, Brady EM, Freeman SC, et al. Diagnostic accuracy of screening questionnaires for obstructive sleep apnoea in adults in different clinical cohorts: A systematic review and meta-analysis. Sleep Breath. 2022;26(3):1053–78. doi: 10.1007/s11325-021-02450-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Pivetta B, Chen L, Nagappa M, et al. Use and performance of the STOP-Bang questionnaire for obstructive sleep apnea screening across geographic regions: A systematic review and meta-analysis. JAMA Netw Open. 2021;4(3):e211009. doi: 10.1001/jamanetworkopen.2021.1009. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.WHO Expert Consultation. Appropriate body-mass index for Asian populations and its implications for policy and intervention strategies. Lancet. 2004;363(9403):157–63. doi: 10.1016/S0140-6736(03)15268-3. [DOI] [PubMed] [Google Scholar]
- 17.Lim U, Monroe KR, Buchthal S, et al. Propensity for intra-abdominal and hepatic adiposity varies among ethnic groups. Gastroenterology. 2019;156(4):966–975e10. doi: 10.1053/j.gastro.2018.11.021. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Chen K, Shen Z, Gu W, et al. Meinian Investigator Group. Prevalence of obesity and associated complications in China: A cross-sectional, real-world study in 15.8 million adults. Diabetes Obes Metab. 2023;25(11):3390–99. doi: 10.1111/dom.15238. [DOI] [PubMed] [Google Scholar]
- 19.Xia M, Liu S, Ji N, et al. BMI 35 kg/m2 does not fit everyone: A modified STOP-Bang questionnaire for sleep apnea screening in the Chinese population. Sleep Breath. 2018;22(4):1075–82. doi: 10.1007/s11325-017-1610-6. [DOI] [PubMed] [Google Scholar]
- 20.Wang L, Zhao W, Liang C, et al. Accuracy and modification of the STOP-Bang questionnaire for screening patients with obstructive sleep apnea in China. J Sleep Res. 2023;32(3):e13781. doi: 10.1111/jsr.13781. [DOI] [PubMed] [Google Scholar]
- 21.Zhang Z, Yang D, Wang H, Liu X. Effects of age and sex on the performance of the NoSAS score as a screening tool for obstructive sleep apnea: A hospital-based retrospective study in China. Sleep Breath. 2021;25(3):1407–17. doi: 10.1007/s11325-020-02254-3. [DOI] [PubMed] [Google Scholar]
- 22.Xu L, Keenan BT, Wiemken AS, et al. Differences in three-dimensional upper airway anatomy between Asian and European patients with obstructive sleep apnea. Sleep. 2020;43(5):zsz273. doi: 10.1093/sleep/zsz273. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Bauters FA, Loof S, Hertegonne KB, et al. Sex-specific sleep apnea screening questionnaires: Closing the performance gap in women. Sleep Med. 2020;67:91–98. doi: 10.1016/j.sleep.2019.10.023. [DOI] [PubMed] [Google Scholar]
- 24.Zhou X, Zhou B, Li Z, et al. Gender differences of clinical and polysomnographic findings with obstructive sleep apnea syndrome. Sci Rep. 2021;11(1):5938. doi: 10.1038/s41598-021-85558-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Pataka A, Kotoulas S, Kalamaras G, et al. Gender differences in obstructive sleep apnea: The value of sleep questionnaires with a separate analysis of cardiovascular patients. J Clin Med. 2020;9(1):130. doi: 10.3390/jcm9010130. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Bonsignore MR, Saaresranta T, Riha RL. Sex differences in obstructive sleep apnoea. Eur Respir Rev. 2019;28(154):190030. doi: 10.1183/16000617.0030-2019. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Regitz-Zagrosek V, Gebhard C. Gender medicine: Effects of sex and gender on cardiovascular disease manifestation and outcomes. Nat Rev Cardiol. 2023;20(4):236–47. doi: 10.1038/s41569-022-00797-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Schwarz EI, Schiza S. Sex differences in sleep and sleep-disordered breathing. Curr Opin Pulm Med. 2024;30(6):593–99. doi: 10.1097/MCP.0000000000001116. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Kim S, Lee KY, Siddiquee AT, et al. Gender differences in association between expiratory dynamic airway collapse and severity of obstructive sleep apnea. Eur Radiol. 2024;34(6):3730–41. doi: 10.1007/s00330-023-10322-x. [DOI] [PMC free article] [PubMed] [Google Scholar]