A Minimalist Method Toward Severity Assessment and Progression Monitoring of Obstructive Sleep Apnea on the Edge

Artificial Intelligence-enabled applications on edge devices have the potential to revolutionize disease detection and monitoring in future smart health (sHealth) systems. In this study, we investigated a minimalist approach for the severity classification, severity estimation, and progression monitoring of obstructive sleep apnea (OSA) in a home environment using wearables. We used the recursive feature elimination technique to select the best feature set of 70 features from a total of 200 features extracted from polysomnogram. We used a multi-layer perceptron model to investigate the performance of OSA severity classification with all the ranked features to a subset of features available from either Electroencephalography or Heart Rate Variability (HRV) and time duration of SpO2 level. The results indicate that using only computationally inexpensive features from HRV and SpO2, an area under the curve of 0.91 and an accuracy of 83.97% can be achieved for the severity classification of OSA. For estimation of the apnea-hypopnea index, the accuracy of RMSE = 4.6 and R-squared value = 0.71 have been achieved in the test set using only ranked HRV and SpO2 features. The Wilcoxon-signed-rank test indicates a significant change (p < 0.05) in the selected feature values for a progression in the disease over 2.5 years. The method has the potential for integration with edge computing for deployment on everyday wearables. This may facilitate the preliminary severity estimation, monitoring, and management of OSA patients and reduce associated healthcare costs as well as the prevalence of untreated OSA.

Protocol for detecting unrecognized sleep apnea in patients with atrial fibrillation by a home-monitoring device: the DAN-APNO study

Background Determining the presence of modifiable risk factors for atrial fibrillation (AF), such as sleep apnea is of clinical importance due to the potential impact targeting these risk factors can have on the progression and burden of AF. Using new digital-based technology is a promising solution to the underreporting of sleep apnea highlighted by academical societies in recent years. The aim of this study is to report the prevalence and severity of sleep apnea in patients with AF and, secondarily, assess the accuracy and feasibility of a new home-screening device for sleep apnea (NightOwl™ by Ectosense). Methods DAN-APNO is a cross-sectional study at the Department of Cardiology, Herlev-Gentofte Hospital recruiting patients with AF referred to anticoagulation initiation aged 18 to 90 years without known sleep apnea. At least 150 patients will be recruited and undergo medical history, clinical evaluation, several sleep-apnea questionnaires, and a sleep-recording evaluation for four nights with sleep apnea home-monitoring device NightOwl™. Additionally, the first 20 participants and participants with moderate-severe sleep apnea by screening are referred to cardio-respiratory monitoring (CRM). This clinical evaluation allows the comparison of standard evaluation method and the NightOwl™. Clinical measures include Apnea–Hypopnea Index (AHI), Oxygen Desaturation Index (ODI), pulse rate, as well as questionaries about sleep apnea assessment and the clinical feasibility of the NightOwl™ device. Main outcomes comprise analysis of the prevalence and severity of sleep apnea, and clinical and demographic predictors of moderate and severe sleep apnea. In addition, correlation analyses for accuracy measures between CRM and NightOwl™ will be conducted along with patient ease-of-use and satisfaction questionnaires. Discussion This study is limited by selection bias; only patients with atrial fibrillation from anticoagulation clinic is asked to participate, which could limit the generalizability of our results. However, this study aims to test whether a miniaturized simple home-monitoring device for detecting sleep apnea in patients with AF potentially can evaluate sleep apnea more conveniently and easier. Trial Registration The study is registered the 18-02-2021 at clinicaltrials.gov with registration number: NCT04760002.

Characteristics of Obstructive Sleep Apnea Patients With Hypertension and Factors Associated With Autotitration Acceptance

In order to explore the characteristics and treatment status of obstructive sleep apnea (OSA) patients with hypertension, a retrospective study was conducted on 306 patients admitted from October 2018 to December 2019. According to the apnea hypopnea index (AHI), OSA patients with hypertension were divided into three groups. 69 cases were mild OSA (5 ≤ AHI &lt; 15), 86 cases were moderate (15 ≤ AHI &lt; 30), and 151 cases were severe (AHI ≥ 30). Compared with patients in the mild and moderate groups, the severe group had more male patients, with higher body mass index (BMI) and non-rapid eye movement stage 1 accounted for total sleep time (N1%), and lower non-rapid eye movement stage 2 accounted for total sleep time (N2%), average and minimum blood oxygen. Among all the patients, those who underwent the titration test accounted for 20.6% (63/306). Multivariate analysis showed that sleep efficiency (p &lt; 0.001) and AHI (p &lt; 0.001) were independent factors for patients to accept titration test. OSA patients with hypertension had a low acceptance of titration therapy. These people with higher sleep efficiency and AHI were more likely to receive autotitration.

Automatic analysis of muscular activity in the flexor digitorum superficialis muscles: A fast screening method for rapid eye movement sleep without atonia

Study objectives To identify a fast and reliable method for rapid eye movement (REM) sleep without atonia (RWA) quantification. Methods We analyzed 36 video-polysomnographies (v-PSGs) of isolated REM sleep behavior disorder (iRBD) patients and 35 controls’ v-PSGs. Patients diagnosed with RBD had: i) RWA, quantified with a reference method, i.e. automatic and artifact-corrected 3-s Sleep Innsbruck Barcelona (SINBAR) index in REM sleep periods (RSPs, i.e. manually selected portions of REM sleep); and ii) v-PSG-documented RBD behaviors. We quantified RWA with other (semi)-automated methods requiring less human intervention than the reference one: the indices proposed by the SINBAR group (the 3-s and 30-s phasic flexor digitorum superficialis (FDS), phasic/”any”/tonic mentalis), and the REM atonia, short and long muscle activity indices (in mentalis/submentalis/FDS muscles). They were calculated in whole REM sleep (i.e. REM sleep scored following international guidelines), in RSPs, with and without manual artifact correction. Area under curves (AUC) discriminating iRBD from controls were computed. Using published cut-offs, the indices’ sensitivity and specificity for iRBD identification were calculated. Apnea-hypopnea index in REM sleep (AHIREM) was considered in the analyses. Results RWA indices from FDS muscles alone had the highest AUCs and all of them had 100% sensitivity. Without manual RSP selection and artifact correction, the “30-s phasic FDS” and the “FDS long muscle activity” had the highest specificity (85%) with AHIREM&lt;15/h. RWA indices were less reliable when AHIREM≥15/h. Conclusions If AHIREM&lt;15/h, FDS muscular activity in whole REM sleep and without artifact correction is fast and reliable to rule out RWA.

STOP-BANG: a Mandatory Tool for Targeted Respiratory Therapy in Bariatric Patients

Identification of patients with obstructive sleep apnea syndrome and high respiratory risk, optimization of the screening algorithm for these patients and administration of preventive non-invasive lung ventilation, makes it possible to prevent the development of perioperative complications, reduce duration of hospital stay and reduce mortality in patients undergoing surgery and bariatric surgery specifically.The objective: to evaluate the effectiveness of STOP-BANG questionnaire for preventive targeted respiratory therapy to reduce the risk of complications in bariatric patients. Subjects and Methods. We examined 60 patients with BMI above 30 kg/m2 referred to elective secondary surgery, the age made 44.2 ± 10.1 years, 23 men and 37 women. Before the operation, patients underwent STOP-BANG questionnaire survey, night respiratory monitoring with the calculation of the apnea/hypopnea index (AHI) and/or saturation during sleep. The standard preoperative examination included clinical and biochemical analyzes.Results. Based on results of STOP-BANG survey, a correlation was revealed between the score and AHI as well as the score and average saturation. The higher score the patients had according to the STOP-BANG questionnaire, the higher AHI was (r = 0.4748, p = 0.002), and the lower mean SpO2 was (r = -0.6958, p < 0.001). Using the ROC analysis, we chose the optimal threshold value - 4 points according to STOP-BANG questionnaire, where the sensitivity of the method was 93% for the AHI, the specificity was 56%, and for the average saturation it was 100% and 63%, respectively. Of the total number of bariatric patients included in the study, 30% required preventive ventilation. In the high respiratory risk group, no significant intraoperative incidents and deaths were reported by the surgical and anesthetic teams. All patients were discharged on time (5‒7 days). Based on the results, a screening procedure has been offered for bariatric patients with high respiratory risk associated with obstructive sleep apnea syndrome.Conclusion. The STOP-BANG questionnaire is a reliable screening tool for high respiratory risk in morbid obese patients. Early diagnosis of high respiratory risk and implementation of preventive ventilation reduces the incidence of perioperative respiratory and cardiovascular complications.

Association Between Arousals During Sleep and Hypertension Among Patients With Obstructive Sleep Apnea

Background Sleep fragmentation induced by repetitive arousals is a hallmark of obstructive sleep apnea (OSA). Sleep fragmentation has been linked to hypertension in community‐based studies, but it is unclear if this association is manifest in OSA. We aimed to explore whether frequent arousals from sleep modify the relationship between OSA and prevalent hypertension. Methods and Results A total of 10 102 patients with OSA and 1614 primary snorers were included in the study. Hypertension was defined on either direct blood pressure measures or diagnosis by a physician. Spontaneous, respiratory, and movement arousals were derived by polysomnography. Logistic regression models were used to estimate the associations between arousals and prevalent hypertension in patients with OSA and primary snorers. For every 10‐unit increase of total arousal index, odds of hypertension significantly increased in both the total sample (odds ratio [OR], 1.08; 95% CI, 1.03–1.14; P =0.002) and patients with OSA (OR, 1.10; 95% CI, 1.04–1.16; P <0.001), but not in the primary snoring group. Total arousal index was significantly associated with systolic blood pressure and diastolic blood pressure in the total sample (β=0.05 and β=0.06; P <0.001) and in patients with (β=0.05 and β=0.06; P <0.01), but not in primary snorers. In addition, a greater influence of respiratory events with arousals than respiratory events without arousals on blood pressure in OSA was also noted. Results were independent of confounders, including apnea‐hypopnea index and nocturnal hypoxemia. Conclusions We conclude that repetitive arousals from sleep are independently associated with prevalent hypertension in patients with OSA.

The effect of adenotonsillectomy and rapid maxillary expansion on the upper airway in pediatric obstructive sleep apnea: a randomized crossover-controlled trial

Study Objectives We aimed to determine the effects of adenotonsillectomy (AT) and rapid maxillary expansion (RME) on the apnea-hypopnea index (AHI) and compare volumetric changes in the upper airway (UA) arising from AT and RME. Methods Thirty-nine children who presented with maxillary constriction and grade III/IV tonsillar hypertrophy were randomized into two groups. One group underwent AT as the first treatment, and the other group underwent RME. Polysomnography (PSG) and cone-beam computed tomography (CBCT) were conducted before (T0) and 6 months after the first treatment (T1). In a crossover design, individuals with AHI&gt;1 received the second treatment. Six months later, they underwent PSG and CBCT (T2). The influence of age, sex, tonsil and adenoid hypertrophy, initial AHI severity, initial volume of the UA, first treatment, and maxillary expansion amount was evaluated using linear regression analysis. Intra- and inter-group comparisons for AHI and inter-group comparisons of volumetric changes in each region of the UA were performed using a paired t-test and Wilcoxon test. Results The initial AHI severity and therapeutic sequence in which AT was the first treatment explained for 95.6% of AHI improvement. AT caused significant improvements in the AHI and volumetric increases in the buccopharynx and total UA areas compared to RME. Conclusions The initial AHI severity and AT as the first treatment accounted for most of the AHI improvement. Most reductions in AHI were due to AT, which promoted more volumetric increases in UA areas than RME. RME may have a marginal effect on pediatric obstructive sleep apnea.

A Comparison of Ultrasound Echo Intensity to MRI as a Metric for Tongue Fat Evaluation

Study Objectives Tongue fat is associated with obstructive sleep apnea (OSA). Magnetic resonance imaging (MRI) is the standard for quantifying tongue fat. Ultrasound echo intensity has been shown to correlate to fat content in skeletal muscles but has yet to be studied in the tongue. The objective of this study is to evaluate the relationship between ultrasound echo intensity and tongue fat. Methods Ultrasound coronal cross-sections of ex-vivo cow tongues were recorded at baseline and following three 1 milliliter serial injections of fat into the tongue. In humans, adults with and without OSA had submental ultrasound coronal cross-sections of their posterior tongue. Average echo intensity of the tongues (cow/human) were calculated in ImageJ software. Head and neck MRI were obtained on human subjects to quantify tongue fat volume. Echo intensity was compared to injected fat volume or MRI derived tongue fat percentage. Results Echo intensity in cow tongues showed a positive correlation to injected fat volume (rho = 0.93, p&lt;0.001). In human subjects, echo intensity of the tongue base strongly correlated with MRI-calculated fat percentage for both the posterior tongue (rho = 0.95, p&lt;0.001) and entire tongue (rho = 0.62, p&lt;0.001). Larger tongue fat percentages (rho = 0.38, p=0.001) and higher echo intensity (rho = 0.27, p=0.024) were associated with more severe apnea-hypopnea index, adjusted for age, BMI, sex and race. Conclusions Ultrasound echo intensity is a viable surrogate measure for tongue fat volume and may provide a convenient modality to characterize tongue fat in OSA.

Objective Sleep Profile in LGI1/CASPR2 Autoimmunity

Study Objectives Rapid eye movement (REM) sleep behavior disorder (RBD) and other sleep disturbances are frequent in leucine-rich, glioma inactivated protein 1-IgG (LGI1) and contactin-associated protein 2-IgG (CASPR2) autoimmunity, yet polysomnographic analyses of these disorders remain limited. We aimed to characterize clinical presentations and analyze polysomnographic manifestations, especially quantitative REM sleep without atonia (RSWA) in LGI1/CASPR2-IgG seropositive (LGI/CASPR2+) patients. Methods We retrospectively analyzed clinical and polysomnographic features and quantitative RSWA between LGI1+/CASPR2+ patients and age-sex matched controls. Groups were compared with Wilcoxon rank-sum and chi-square tests. Combined submentalis and anterior tibialis (SM+AT) RSWA was the primary outcome Results Among 11 (LGI1+, n=9; CASPR2+, n=2) patients, Morvan syndrome sleep features were present in 7 (63.6%) LGI1+/CASPR2+ patients, with simultaneous insomnia and DEB in 3 (27.3%), and the most common presenting sleep disturbances were dream enactment behavior (DEB, n=5), insomnia (n=5), and sleep apnea (n=8; median apnea hypopnea index=15/hour). Median Epworth Sleepiness Scale (ESS) was 9 (range 3-24; n=10), with hypersomnia in 4 (36.4%). LGI1+/CASPR2+ patients had increased N1 sleep (p=0.02), decreased REM sleep (p=0.001), and higher levels of SM+AT any RSWA (p &lt; 0.001). Eight of 9 (89%) LGI1+ exceeded RBD RSWA thresholds (DEB, n=5; isolated RSWA, n=3). RSWA was greater in anterior tibialis than submentalis. All 10 LGI1+/CASPR2+ patients treated with immunotherapy benefitted, and 5/10 had improved sleep disturbances. Conclusion LGI1/CASPR2-IgG autoimmunity is associated with prominent dream enactment, insomnia, RSWA, sleep apnea, and shallower sleep. Polysomnography provides objective disease markers in LGI1+/CASPR2+ autoimmunity and immunotherapy may benefit associated sleep disturbances.

The Snoring Index Identifies Risk of Non-Alcoholic Fatty Liver Disease in Patients with Obstructive Sleep Apnea Syndrome

Background: The aim of this observational cohort study was to explore the severity of liver disease in patients with suspected obstructive sleep apnea in Germany. Methods: Patients undergoing polysomnography or home sleep apnea testing (HSAT) as an evaluation for the presence of OSA were screened using vibration-controlled transient elastography (VCTE) and continuous attenuation parameter (CAP) with a Fibroscan ® Mini 430. Clinical and laboratory data were collected following the overnight exam. Results: In total, 78 patients (28 female (35.9%), mean age 54.2 years) with OSA defined by an apnea-hypopnea-index >5 events/hour were included between OCT 2020 and APR 2021. Patients exhibited a high metabolic risk profile with 17% known diabetes mellitus type 2 (T2D), 62% arterial hypertension, 14% hyperlipidemia and 36% BMI > 30 kg/m2. The prevalence of steatosis defined by a CAP > 280 dB/m was 54%. The prevalence of at least significant fibrosis was 16% (E > 9.0 kPa). Interestingly, patients with a snoring index above the median of 278/h showed significantly higher CAP-values (p = 0.0002). In addition, the proportion of oxygen saturations below 90% (t90) correlated with CAP-values (p = 0.02), as well as metabolic risk factors including increased waist circumference (p = 0.005) and body mass index (BMI) (p = 0.035). On the other hand, the apnea-hypopnea-index (AHI) as a marker of OSA severity did not correlate with VCTE, CAP or laboratory parameters. Conclusion: Patients with moderate to severe OSA have a high prevalence of hepatic steatosis. The snoring index is an easy-to-use clinical tool to identify patients at risk for relevant liver disease within the larger group of patients with OSA.