Sleep Tests in the Non-Contact Era of the COVID-19 Pandemic: Home Sleep Tests Versus In-Laboratory Polysomnography

Standardized image-based polysomnography database and deep learning algorithm for sleep-stage classification

STUDY OBJECTIVES

Polysomnography (PSG) scoring is labor-intensive, subjective, and often ambiguous. Recently several deep learning (DL) models for automated sleep scoring have been developed, they are tied to a fixed amount of input channels and resolution. In this study, we constructed a standardized image-based PSG dataset in order to overcome the heterogeneity of raw signal data obtained from various PSG devices and various sleep laboratory environments.

METHODS

All individually exported European data format files containing raw signals were converted into images with an annotation file, which contained the demographics, diagnoses, and sleep statistics. An image-based DL model for automatic sleep staging was developed, compared with a signal-based model, and validated in an external dataset.

RESULTS

We constructed 10253 image-based PSG datasets using a standardized format. Among these, 7745 diagnostic PSG data were used to develop our DL model. The DL model using the image dataset showed similar performance to the signal-based dataset for the same subject. The overall DL accuracy was greater than 80%, even with severe obstructive sleep apnea. Moreover, for the first time, we showed explainable DL in the field of sleep medicine as visualized key inference regions using Eigen-class activation maps. Furthermore, when a DL model for sleep scoring performs external validation, we achieved a relatively good performance.

CONCLUSIONS

Our main contribution demonstrates the availability of a standardized image-based dataset, and highlights that changing the data sampling rate or number of sensors may not require retraining, although performance decreases slightly as the number of sensors decreases.

Surface Active Salivary Metabolites Indicate Oxidative Stress and Inflammation in Obstructive Sleep Apnea

PURPOSE

Obstructive sleep apnea (OSA), a highly prevalent and potentially serious sleep disorder, requires effective screening tools. Saliva is a useful biological fluid with various metabolites that might also influence upper airway patency by affecting surface tension in the upper airway. However, little is known about the composition and role of salivary metabolites in OSA. Therefore, we investigated the metabolomics signature in saliva from the OSA patients and evaluated the associations between identified metabolites and salivary surface tension.

METHODS

We studied 68 subjects who visited sleep clinic due to the symptoms of OSA. All underwent full-night in-lab polysomnography. Patients with apnea-hypopnea index (AHI) < 10 were classified to the control, and those with AHI ≥ 10 were the OSA groups. Saliva samples were collected before and after sleep. The centrifuged saliva samples were analyzed by liquid chromatography with high-resolution mass spectrometry (ultra-performance liquid chromatography-tandem mass spectrometry; UPLC-MS/MS). Differentially expressed salivary metabolites were identified using open source software (XCMS) and Compound Discoverer 2.1. Metabolite set enrichment analysis (MSEA) was performed using MetaboAnalyst 5.0. The surface tension of the saliva samples was determined by the pendant drop method.

RESULTS

Three human-derived metabolites (1-palmitoyl-2-[5-hydroxyl-8-oxo-6-octenoyl]-sn-glycerol-3-phosphatidylcholine [PHOOA-PC], 1-palmitoyl-2-[5-keto-8-oxo-6-octenoyl]-sn-glycerol-3-phosphatidylcholine [KPOO-PC], and 9-nitrooleate) were significantly upregulated in the after-sleep salivary samples from the OSA patients compared to the control group samples. Among the candidate metabolites, only PHOOA-PC was correlated with the AHI. In OSA samples, salivary surface tension decreased after sleep. The differences in surface tension were negatively correlated with PHOOA-PC and 9-nitrooleate concentrations. Furthermore, MSEA revealed that arachidonic acid-related metabolism pathways were upregulated in the after-sleep samples from the OSA group.

CONCLUSIONS

This study revealed that salivary PHOOA-PC was correlated positively with the AHI and negatively with salivary surface tension in the OSA group. Salivary metabolomic analysis may improve our understanding of upper airway dynamics and provide new insights into novel biomarkers and therapeutic targets in OSA.

Obstructive Sleep Apnea and Cardiovascular Diseases: Sad Realities and Untold Truths regarding Care of Patients in 2022

Obstructive sleep apnea (OSA) is one of the most common and serious sleep-related breathing disorders with a high prevalence among patients with cardiovascular (CV) diseases. Despite its widespread presence, OSA remains severely undiagnosed and untreated. CV mortality and morbidity are significantly increased in the presence of OSA as it is associated with an increased risk of resistant hypertension, heart failure, arrhythmias, and coronary artery disease. Evaluation and treatment of OSA should focus on recognizing patients at risk of developing OSA. The use of screening questionnaires should be routine, but a formal polysomnography sleep study is fundamental in establishing and classifying OSA. Recognition of OSA patients will allow for the institution of appropriate therapy that should alleviate OSA-related symptoms with the intent of decreasing adverse CV risk. In this review, we focus on the impact OSA has on CV disease and evaluate contemporary OSA treatments. Our goal is to heighten awareness among CV practitioners.

A Vision-Based System for In-Sleep Upper-Body and Head Pose Classification

Sleep quality is known to have a considerable impact on human health. Recent research shows that head and body pose play a vital role in affecting sleep quality. This paper presents a deep multi-task learning network to perform head and upper-body detection and pose classification during sleep. The proposed system has two major advantages: first, it detects and predicts upper-body pose and head pose simultaneously during sleep, and second, it is a contact-free home security camera-based monitoring system that can work on remote subjects, as it uses images captured by a home security camera. In addition, a synopsis of sleep postures is provided for analysis and diagnosis of sleep patterns. Experimental results show that our multi-task model achieves an average of 92.5% accuracy on challenging datasets, yields the best performance compared to the other methods, and obtains 91.7% accuracy on the real-life overnight sleep data. The proposed system can be applied reliably to extensive public sleep data with various covering conditions and is robust to real-life overnight sleep data.

Impacts of the COVID-19 pandemic on sleep center operations and sleep apnea treatment in Korea: A multicenter survey

ABSTRACT

The coronavirus disease-2019 (COVID-19) pandemic has not only changed the lives of people around the world but also affected all areas of the healthcare system, including sleep medicine. However, no studies in Korea have investigated the status of domestic sleep centers and their challenges during the pandemic.An online survey was performed from December 2020 to January 2021. Hospitals that belonged to sleep-related academic societies and were considered well managed were included in this survey. The questionnaire focused on changes in sleep center operations, infection control policies, and patient treatment since the start of the COVID-19 pandemic. Telemedicine and future directions for sleep medicine services were also investigated.Of the 20 sleep centers that responded, 80% were at university hospitals with more than 500 inpatient beds. During the third wave of the COVID-19 pandemic in Korea (November-December 2020), the routine operating schedule of the sleep study room was reduced in 30% of the sleep centers compared to November-December 2019 (before COVID-19). The number of type 1 polysomnographies performed decreased in 85% of the sleep centers. In contrast, in-lab positive airway pressure (PAP) titrations decreased in 40%, remained unchanged in 35%, and increased in 25%. With respect to prescriptions, 30% of the sleep centers increased the number of prescriptions for auto-titrating continuous PAP. However, 60% of the sleep centers reported no change in the rate of fixed continuous PAP and auto-titrating continuous PAP prescriptions. All sleep centers that participated in this survey agreed that the need for documented infection control regulations will continue after the COVID-19 pandemic. Since the beginning of the pandemic, 30% of the centers have tried telemedicine. However, respondents expressed concern about telemedicine, citing a number of practical issues.Compared to countries where the COVID-19 pandemic was severe, Korea had less impact of COVID-19 on the sleep center operations and sleep apnea treatment. Infection and quality control in the sleep study room are important and inevitable issues, and regulation within each institution is necessary. Further research and discussion are needed regarding telemedicine and home sleep apnea test in Korea.

Urine 5-Eicosatetraenoic Acids as Diagnostic Markers for Obstructive Sleep Apnea

Early detection of obstructive sleep apnea (OSA) is needed to reduce cardiovascular sequelae and mortality. Full-night polysomnography has been used for diagnosing OSA, but it is too expensive and inconvenient for patients to handle. Metabolome-wide analyses were performed to find and validate surrogate markers for OSA. We further investigated the mechanism underlying hypoxic induction of the markers in human cells and mice. Arachidonic acid derivatives 5-HETE and 5-oxoETE were detected in urine samples. The levels (mean ± SD, ng per mg creatinine) of 5-HETE and 5-oxoETE were 56.4 ± 26.2 and 46.9 ± 18.4 in OSA patients, respectively, which were significantly higher than those in controls (22.5 ± 4.6 and 18.7 ± 3.6). Both levels correlated with the apnea-hypopnea index and the lowest oxygen saturation on polysomnography. After the treatment with the continuous positive airway pressure, the metabolite levels were significantly reduced compared with those before the treatment. In human mononuclear cells subjected to intermittent hypoxia, 5-HETE and 5-oxoETE productions were induced by hypoxia-inducible factor 1 and glutathione peroxidase. When mice were exposed to intermittent hypoxia, 5-HETE and 5-oxoETE were excreted more in urine. They were identified and verified as new OSA markers reflecting hypoxic stress. The OSA markers could be used for OSA diagnosis and therapeutic evaluation.