Gender Phenotyping of Patients with Obstructive Sleep Apnea Syndrome Using a Network Science Approach

Screening the risk of obstructive sleep apnea by utilizing supervised learning techniques based on anthropometric features and snoring events

Objectives

Obstructive sleep apnea (OSA) is typically diagnosed by polysomnography (PSG). However, PSG is time-consuming and has some clinical limitations. This study thus aimed to establish machine learning models to screen for the risk of having moderate-to-severe and severe OSA based on easily acquired features.

Methods

We collected PSG data on 3529 patients from Taiwan and further derived the number of snoring events. Their baseline characteristics and anthropometric measures were obtained, and correlations among the collected variables were investigated. Next, six common supervised machine learning techniques were utilized, including random forest (RF), extreme gradient boosting (XGBoost), k-nearest neighbor (kNN), support vector machine (SVM), logistic regression (LR), and naïve Bayes (NB). First, data were independently separated into a training and validation dataset (80%) and a test dataset (20%). The approach with the highest accuracy in the training and validation phase was employed to classify the test dataset. Next, feature importance was investigated by calculating the Shapley value of every factor, which represented the impact on OSA risk screening.

Results

The RF produced the highest accuracy (of >70%) in the training and validation phase in screening for both OSA severities. Hence, we employed the RF to classify the test dataset, and results showed a 79.32% accuracy for moderate-to-severe OSA and 74.37% accuracy for severe OSA. Snoring events and the visceral fat level were the most and second most essential features of screening for OSA risk.

Conclusions

The established model can be considered for screening for the risk of having moderate-to-severe or severe OSA.

Sex differences in polysomnographic findings in patients with obstructive sleep apnea

BACKGROUND AND PURPOSE

Sex differences in the clinical findings and the polysomnographic presentation of patients with obstructive sleep apnea (OSA) are compelling current research issues. For example, patients suffering from obstructive sleep apnea are predominantly male. While women are older than men and tend to have a higher body mass index, men typically present with a more severe form of obstructive sleep apnea. Using polysomnography, we investigated a German cohort, subdivided per severity levels of obstructive sleep apnea (apnea-hypopnea index: ≥5 to < 15/h (mild), ≥15 to < 30/h (moderate), and ≥30/h (severe)) to provide a detailed analysis of breathing and sleep parameters, accounting for body position effects and severity of illness. A deeper understanding of sex differences may allow targeted diagnosis and treatment adjustment.

PATIENTS AND METHODS

This retrospective study included a cohort of 1242 German patients (940 male, 302 female) who underwent overnight polysomnography at the private sleep laboratory "Intersom Köln", Center for Sleep Medicine and Sleep Research. In 1125 subjects (878 male, 247 female), obstructive sleep apnea was diagnosed. All patients were examined between January 01, 2018 and December 31, 2020, comparing anthropometric, sleep morphological, and respiratory polysomnographic findings.

RESULTS

Female patients with obstructive sleep apnea were significantly older than male patients (60.9 ± 12.3 vs. 56.9 ± 12.5 years, P < .001), also among OSA subgroups per OSA severity. The body mass index was similar in male and female patients (29.6 ± 5.1 vs. 29.2 ± 7.3 kg/m², P > .05), including the three subgroups. Men were more likely to have severe obstructive sleep apnea (46.9%) than women (35.2%). Women exhibited a higher proportion of slow-wave sleep than men (129.4 ± 52.8 vs. 104.2 ± 53.2 min; P < .001). The apnea-hypopnea index of total sleep time was significantly greater in male than female patients (32.9 ± 21.2 vs. 27.2 ± 20.2 per hour; P < .001). Female patients had a higher apnea-hypopnea index during rapid-eye-movement (REM) sleep (34.0 ± 23.8 vs. 31.8 ± 22.3 per hour; P = .171). A statistically significant difference in the apnea-hypopnea index during REM sleep between sexes was found when the obstructive sleep apnea severity was considered. Women had a lower apnea-hypopnea index in non-rapid eye-movement (NREM) sleep than men (25.7 ± 21.1 vs. 32.7 ± 22.3 per hour; P < .001). The oxygen desaturation index (29.9 ± 20.3 vs. 22.4 ± 19.4%; P < .001) and an oxygen desaturation below 90% (9.4 ± 14.0 vs. 6.8 ± 11.7%; P = .003) was greater in men than in women. In severe obstructive sleep apnea, the oxygen desaturation index was similar between the sexes (45.0 ± 17.8 vs. 41.1 ± 20.9%; P = .077). Male patients showed a higher supine apnea-hypopnea-index than female patients. (45.7 ± 26.7 vs 36.1 ± 22.7 per hour; P < .001).

CONCLUSION

The present noninvasive, retrospective registry study is the first to examine sex differences in OSA in such a large German population in terms of respiratory and sleep parameters, taking into account the effects of body position and severity of the disease. We could confirm and extend observations from previous studies. Female patients were significantly older than the male patients. The apnea-hypopnea index was higher in male than in female patients. Women showed a higher apnea-hypopnea index in REM sleep and a lower one in NREM sleep. Men were desaturated more often and were more affected by supine-dependent obstructive sleep apnea than women. Contrary to the literature, there were no significant differences in body mass index (BMI) between the sexes. With increasing age and BMI, the gender differences become less significant.

Screening for Obstructive Sleep Apnea Risk by Using Machine Learning Approaches and Anthropometric Features

Obstructive sleep apnea (OSA) is a global health concern and is typically diagnosed using in-laboratory polysomnography (PSG). However, PSG is highly time-consuming and labor-intensive. We, therefore, developed machine learning models based on easily accessed anthropometric features to screen for the risk of moderate to severe and severe OSA. We enrolled 3503 patients from Taiwan and determined their PSG parameters and anthropometric features. Subsequently, we compared the mean values among patients with different OSA severity and considered correlations among all participants. We developed models based on the following machine learning approaches: logistic regression, k-nearest neighbors, naïve Bayes, random forest (RF), support vector machine, and XGBoost. Collected data were first independently split into two data sets (training and validation: 80%; testing: 20%). Thereafter, we adopted the model with the highest accuracy in the training and validation stage to predict the testing set. We explored the importance of each feature in the OSA risk screening by calculating the Shapley values of each input variable. The RF model achieved the highest accuracy for moderate to severe (84.74%) and severe (72.61%) OSA. The level of visceral fat was found to be a predominant feature in the risk screening models of OSA with the aforementioned levels of severity. Our machine learning models can be employed to screen for OSA risk in the populations in Taiwan and in those with similar craniofacial structures.

Incorporating neck circumference or neck-to-height ratio into the GOAL questionnaire to better detect and describe obstructive sleep apnea with application to clinical decisions

Objective

Although neck circumference (NC) and neck-to-height ratio (NHR) have been recognized as effective predictors of the clinical diagnosis of adult obstructive sleep apnea (OSA), they have not been included in the widely used GOAL questionnaire. Not coincidentally, the NHR has not been adequately considered in the development and validation of the STOP-Bang questionnaire, No-Apnea score and the NoSAS score. The motivation for the study was (1) to combine the GOAL questionnaire with the NC and NHR, respectively, to evaluate its predictive performance and (2) to compare it with the STOP-Bang questionnaire, the No-Apnea score, the NOSAS score, and the GOAL questionnaire.

Materials and methods

This retrospectively allocated cross-sectional study was conducted from November 2017 to March 2022 in adults who underwent nocturnal polysomnography (PSG) or home sleep apnea testing (HSAT). In this paper, the GOAL questionnaire was combined with the NC and NHR, respectively, using logistic regression. The performance of the six screening tools was assessed by discriminatory ability [area under the curve (AUC) obtained from receiver operating characteristic (ROC) curves] and a 2 × 2 league table [including sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), positive likelihood ratio (LR+), and negative likelihood ratio (LR−)] and compared under AHI ≥5/h, AHI ≥15/h, and AHI ≥30/h conditions.

Results

A total of 288 patients were enrolled in the study. For all severity OSA levels, the sensitivity of GOAL+NC ranged from 70.12 to 70.80%, and specificity ranged from 86.49 to 76.16%. The sensitivity of GOAL+NHR ranged from 73.31 to 81.75%, while specificity ranged from 83.78 to 70.86%. As for area under the curve (AUC) value under ROC curve, when AHI ≥5/h, compared with GOAL (0.806), No-Apnea (0.823), NoSAS (0.817), and GOAL+NC (0.815), GOAL+NHR (0.831) obtained the highest AUC value, but lower than STOP-Bang (0.837).

Conclusion

The predictive power of incorporating NC or NHR into the GOAL questionnaire was significantly better than that of the GOAL itself. Furthermore, GOAL+NHR was superior to GOAL+NC in predicting OSA severity and better than the No-Apnea score and the NoSAS score.

Obstructive Sleep Apnea Syndrome Comorbidity Phenotypes in Primary Health Care Patients in Northern Greece

Background: Obstructive sleep apnea syndrome (OSAS) is a significant public health issue. In the general population, the prevalence varies from 10% to 50%. We aimed to phenotype comorbidities in OSAS patients referred to the primary health care (PHC) system. Methods: We enrolled 1496 patients referred to the PHC system for any respiratory- or sleep-related issue from November 2015 to September 2017. Some patients underwent polysomnography (PSG) evaluation in order to establish OSAS diagnosis. The final study population comprised 136 patients, and the Charlson comorbidity index was assessed. Categorical principal component analysis and TwoStep clustering was used to identify distinct clusters in the study population. Results: The analysis revealed three clusters: the first with moderate OSAS, obesity and a high ESS score without significant comorbidities; the second with severe OSAS, severe obesity with comorbidities and the highest ESS score; and the third with severe OSAS and obesity without comorbidities but with a high ESS score. The clusters differed in age (p < 0.005), apnea–hypopnea index, oxygen desaturation index, arousal index and respiratory and desaturation arousal index (p < 0.001). Conclusions: Predictive comorbidity models may aid the early diagnosis of patients at risk in the context of PHC and pave the way for personalized treatment.

Enhancing Obstructive Sleep Apnea Diagnosis With Screening Through Disease Phenotypes: Algorithm Development and Validation

BACKGROUND

The American Academy of Sleep Medicine guidelines suggest that clinical prediction algorithms can be used in patients with obstructive sleep apnea (OSA) without replacing polysomnography, which is the gold standard.

OBJECTIVE

This study aims to develop a clinical decision support system for OSA diagnosis according to its standard definition (apnea-hypopnea index plus symptoms), identifying individuals with high pretest probability based on risk and diagnostic factors.

METHODS

A total of 47 predictive variables were extracted from a cohort of patients who underwent polysomnography. A total of 14 variables that were univariately significant were then used to compute the distance between patients with OSA, defining a hierarchical clustering structure from which patient phenotypes were derived and described. Affinity from individuals at risk of OSA phenotypes was later computed, and cluster membership was used as an additional predictor in a Bayesian network classifier (model B).

RESULTS

A total of 318 patients at risk were included, of whom 207 (65.1%) individuals were diagnosed with OSA (111, 53.6% with mild; 50, 24.2% with moderate; and 46, 22.2% with severe). On the basis of predictive variables, 3 phenotypes were defined (74/207, 35.7% low; 104/207, 50.2% medium; and 29/207, 14.1% high), with an increasing prevalence of symptoms and comorbidities, the latter describing older and obese patients, and a substantial increase in some comorbidities, suggesting their beneficial use as combined predictors (median apnea-hypopnea indices of 10, 14, and 31, respectively). Cross-validation results demonstrated that the inclusion of OSA phenotypes as an adjusting predictor in a Bayesian classifier improved screening specificity (26%, 95% CI 24-29, to 38%, 95% CI 35-40) while maintaining a high sensitivity (93%, 95% CI 91-95), with model B doubling the diagnostic model effectiveness (diagnostic odds ratio of 8.14).

CONCLUSIONS

Defined OSA phenotypes are a sensitive tool that enhances our understanding of the disease and allows the derivation of a predictive algorithm that can clearly outperform symptom-based guideline recommendations as a rule-out approach for screening.

Evaluation of Obstructive Sleep Apnea Phenotypes Treatment Effectiveness

Background and Objective: Obstructive sleep apnea (OSA) is a heterogeneous chronic sleep associated disorder. A common apnea-hypopnea index (AHI)-focused approach to OSA severity evaluation is not sufficient enough to capture the extent of OSA related risks, it limits our understanding of disease pathogenesis and may contribute to a modest response to conventional treatment. In order to resolve the heterogeneity issue, OSA patients can be divided into more homogenous therapeutically and prognostically significant groups-phenotypes. An improved understanding of OSA phenotype relationship to treatment effectiveness is required. Thus, in this study several clinical OSA phenotypes are identified and compared by their treatment effectiveness. Methods and materials: Retrospective data analysis of 233 adult patients with OSA treated with continuous positive airway pressure (CPAP) was performed. Statistical analysis of data relating to demographic and anthropometric characteristics, symptoms, arterial blood gas test results, polysomnografic and respiratory polygraphic tests and treatment, treatment results was performed. Results: 3 phenotypes have been identified: "Position dependent (supine) OSA" (Positional OSA), "Severe OSA in obese patients" (Severe OSA) and "OSA and periodic limb movements (PLM)" (OSA and PLM). The highest count of responders to treatment with CPAP was in the OSA and PLM phenotype, followed by the Positional OSA phenotype. Treatment with CPAP, despite the highest mean pressure administered was the least effective among Severe OSA phenotype. Conclusions: Different OSA phenotypes vary significantly and lead to differences in response to treatment. Thus, treatment effectiveness depends on OSA phenotypes and treatment techniques other than CPAP may be needed. This emphasizes the importance of a more individualized approach when treating OSA.

Gender differences of clinical and polysomnographic findings with obstructive sleep apnea syndrome

Obstructive sleep apnea syndrome (OSAS) is underdiagnosed in females and gender differences in clinical and polysomnographic findings have not been widely investigated in China. We examined clinical and polysomnographic differences between males and females with OSAS in order to determine the influence of gender on clinical presentation and polysomnographic features. Data were collected from 303 adult patients diagnosed with OSAS (237 males and 66 females) from 2017 to 2019. All the patients completed physical examination, Epworth sleepiness scale, and whole night polysomnography. AVONA, univariate and multivariate logistic regression analyses were conducted to assess gender differences of clinical and polysomnographic findings with OSAS. P < 0.05 was statistically significant. The average age was 48.4 ± 12.6 years for females and 43.4 ± 12.4 years for males. Compared with female patients with OSAS, male patients were taller and heavier, had higher systolic blood pressure in the morning, shorter duration of slow wave sleep, more micro-arousal events, greater AHI, and more complex sleep apnea events. There are obvious gender differences of clinical and polysomnographic characteristics with OSAS. Understanding gender differences will contribute to better clinical recognition of OSAS in females as well as the provision of proper health care and therapeutic practice.