Influence of the device used for obstructive sleep apnea diagnosis on body position: a comparison between polysomnography and portable monitor.

PURPOSE: Different devices have been used for the diagnosis of obstructive sleep apnea (OSA), which differ in the number of sensors used. The numerous sensors used in more complex sleep studies such as in-lab polysomnography may influence body position during sleep. We hypothesized that patients submitted to in-lab polysomnography (PSG) would spend more time in the supine position than patients submitted to an ambulatory Portable Monitor (PM) sleep study. METHODS: Body position during PSG and PM studies was compared among two distinct groups of patients matched for age, body-mass index (BMI), apnea-hypopnea index (AHI), and gender. Predictors of time spent in the supine position were determined using a multiple linear regression model. RESULTS: Of 478 participants who underwent either PSG or PM studies, mean age: 61[43-66] years; males: 43.9%; BMI: 28.4[26.1-31.1]kg/m2; AHI 14[7-27] events/hour). Participants who underwent PSG studies spent more time in the supine position (41[16-68]% than participants who underwent PM studies (34[16-51]%), P = 0.014. Participants with OSA spent more time in the supine position than participants without OSA, both among the PSG and PM groups P < 0.05). Gender, BMI, OSA severity, and sleep study type were independent predictors of time spent in the supine position. CONCLUSION: In-lab PSG may increase time spent in the supine position and overestimate OSA severity compared to a PM sleep study. OSA diagnosis is also associated with increased time spent in the supine position. The potential influence on the sleeping position should be taken into account when choosing among the different sleep study types for OSA diagnosis.

Development and Performance Evaluation of a Low-Cost Portable PM2.5 Monitor for Mobile Deployment.

The concentration of fine particulate matter (PM2.5) is known to vary spatially across a city landscape. Current networks of regulatory air quality monitoring are too sparse to capture these intra-city variations. In this study, we developed a low-cost (60 USD) portable PM2.5 monitor called Smart-P, for use on bicycles, with the goal of mapping street-level variations in PM2.5 concentration. The Smart-P is compact in size (85 × 85 × 42 mm) and light in weight (147 g). Data communication and geolocation are achieved with the cyclist's smartphone with the help of a user-friendly app. Good agreement was observed between the Smart-P monitors and a regulatory-grade monitor (mean bias error: −3.0 to 1.5 µg m−3 for the four monitors tested) in ambient conditions with relative humidity ranging from 38 to 100%. Monitor performance decreased in humidity > 70% condition. The measurement precision, represented as coefficient of variation, was 6 to 9% in stationary mode and 6% in biking mode across the four tested monitors. Street tests in a city with low background PM2.5 concentrations (8 to 9 µg m−3) and in two cities with high background concentrations (41 to 74 µg m−3) showed that the Smart-P was capable of observing local emission hotspots and that its measurement was not sensitive to bicycle speed. The low-cost and user-friendly nature are two features that make the Smart-P a good choice for empowering citizen scientists to participate in local air quality monitoring.

Evaluating the performance of five scoring systems for prescreening obstructive sleep apnea-hypopnea syndrome.

PURPOSE: A comparison of all scoring systems used for screening for obstructive sleep apnea-hypopnea syndrome (OSAHS) is lacking. The aim of this investigation was to evaluate the performance of five scoring systems for screening for OSAHS, as well as to validate the use of the NoSAS and SACS in the Chinese population. METHODS: Data were retrospectively collected from hospital-based, manned, overnight sleep monitoring studies for 105 consecutive outpatients using a portable monitor (PM) device. RESULTS: The 105 participants had an average age of 46 years and were mostly men (75%). STOP-Bang, SACS, and NoSAS scoring exhibited moderate predictive values at different AHI cutoffs (AUC 0.761-0.853, 0.722-0.854, and 0.724-0.771 respectively), followed by the STOP and Berlin questionnaire (AUC 0.680-0.781vs 0.624-0.724). Both STOP-Bang and SACS showed excellent sensitivity (89.5-100% vs 93.4-94.6%) and negative predictive value (68-100% vs 77.3-90.9%), while STOP-Bang, STOP, and SACS showed low negative likelihood ratios (- LR) (0-0.2). CONCLUSIONS: Our study indicated that the STOP-Bang questionnaire and the SACS both show better predictive value than other scoring systems among the five screening tools for OSAHS. Both scoring systems are simple and easy to implement for screening for OSAHS in the community and in hospitals.

Primary Care Physicians Can Comprehensively Manage Patients with Sleep Apnea. A Noninferiority Randomized Controlled Trial.

Rationale: General practitioners play a passive role in obstructive sleep apnea (OSA) management. Simplification of the diagnosis and use of a semiautomatic algorithm for treatment can facilitate the integration of general practitioners, which has cost advantages.Objectives: To determine differences in effectiveness between primary health care area (PHA) and in-laboratory specialized management protocols during 6 months of follow-up.Methods: A multicenter, noninferiority, randomized, controlled trial with two open parallel arms and a cost-effectiveness analysis was performed in six tertiary hospitals in Spain. Sequentially screened patients with an intermediate to high OSA probability were randomized to PHA or in-laboratory management. The PHA arm involved a portable monitor with automatic scoring and semiautomatic therapeutic decision-making. The in-laboratory arm included polysomnography and specialized therapeutic decision-making. Patients in both arms received continuous positive airway pressure treatment or sleep hygiene and dietary treatment alone. The primary outcome measure was the Epworth Sleepiness Scale. Secondary outcomes were health-related quality of life, blood pressure, incidence of cardiovascular events, hospital resource utilization, continuous positive airway pressure adherence, and within-trial costs.Measurements and Main Results: In total, 307 patients were randomized and 303 were included in the intention-to-treat analysis. Based on the Epworth Sleepiness Scale, the PHA protocol was noninferior to the in-laboratory protocol. Secondary outcome variables were similar between the protocols. The cost-effectiveness relationship favored the PHA arm, with a cost difference of €537.8 per patient.Conclusions: PHA management may be an alternative to in-laboratory management for patients with an intermediate to high OSA probability. Given the clear economic advantage of outpatient management, this finding could change established clinical practice.Clinical trial registered with www.clinicaltrials.gov (NCT02141165).

[Integration of Micro Sensors with Mobile Devices for Monitoring Vital Signs of Sleep Apnea Patients].

This paper demonstrates a BSN-based portable monitor integrating micro sensors with mobile devices for monitoring and diagnosing obstructive sleep apnea syndrome (OSAS) at home. The system uses a micro hot-film flow sensor to detect respiratory flow, uses tri-axis micro accelerometer to detect body posture and motion intensity, and uses a micro photoelectric sensor to detect oxygen saturation of blood. The real-time vital signs are detected and transmitted wirelessly to a mobile device. The transmitted data are processed and analyzed in the mobile device and can be further transmitted to a remote medical center and physicians through mobile cellular networks or Internet for comprehensively analyzing, assessing, and diagnosing to the diseases.

Conventional Polysomnography Is Not Necessary for the Management of Most Patients with Suspected Obstructive Sleep Apnea. Noninferiority, Randomized Controlled Trial.

RATIONALE: Home respiratory polygraphy may be a simpler alternative to in-laboratory polysomnography for the management of more symptomatic patients with obstructive sleep apnea, but its effectiveness has not been evaluated across a broad clinical spectrum. OBJECTIVES: To compare the long-term effectiveness (6 mo) of home respiratory polygraphy and polysomnography management protocols in patients with intermediate-to-high sleep apnea suspicion (most patients requiring a sleep study). METHODS: A multicentric, noninferiority, randomized controlled trial with two open parallel arms and a cost-effectiveness analysis was performed in 12 tertiary hospitals in Spain. Sequentially screened patients with sleep apnea suspicion were randomized to respiratory polygraphy or polysomnography protocols. Moreover, both arms received standardized therapeutic decision-making, continuous positive airway pressure (CPAP) treatment or a healthy habit assessment, auto-CPAP titration (for CPAP indication), health-related quality-of-life questionnaires, 24-hour blood pressure monitoring, and polysomnography at the end of follow-up. The main outcome was the Epworth Sleepiness Scale measurement. The noninferiority criterion was -2 points on the Epworth scale. MEASUREMENTS AND MAIN RESULTS: In total, 430 patients were randomized. The respiratory polygraphy protocol was noninferior to the polysomnography protocol based on the Epworth scale. Quality of life, blood pressure, and polysomnography were similar between protocols. Respiratory polygraphy was the most cost-effective protocol, with a lower per-patient cost of 416.7€. CONCLUSIONS: Home respiratory polygraphy management is similarly effective to polysomnography, with a substantially lower cost. Therefore, polysomnography is not necessary for most patients with suspected sleep apnea. This finding could change established clinical practice, with a clear economic benefit. Clinical trial registered with www.clinicaltrials.gov (NCT 01752556).

Detection of sleep-disordered breathing with ambulatory Holter monitoring.

PURPOSE: Obstructive sleep apnea (OSA) syndrome is a common condition that can impact clinical outcomes among patients with cardiovascular disease. Screening all subjects with heart disease via polysomnography (PSG) is costly and resource-limited. We sought to compare a Holter monitor-based algorithm to detect OSA to in-laboratory polysomnography (PSG). METHODS: Prospective cohort study of patients undergoing in-laboratory attended PSG for the evaluation of OSA. A standard 12-lead Holter monitor was attached to patients at the initiation of PSG. Holter-derived respiratory disturbance index (HDRDI) was extracted from the respiratory myogram, based on detecting skeletal muscle "noise" detected on the baseline. Apneic and hypopneic episodes were identified by comparing sudden changes in the myogram to abrupt increases in heart rate. The HDRDI was compared with the PSG-derived apnea-hypopnea index (PDAHI). RESULTS: Thirty patients underwent simultaneous Holter monitoring and overnight diagnostic PSG. An ROC curve for peak HDRDI was 0.79 (95% CI 0.61, 0.97) for OSA, with sensitivity of 94.4% and specificity of 54.5%. A cutoff value of HDRDI < 10 appeared to identify those individuals without clinically significant sleep-disordered breathing. CONCLUSION: Holter-derived respiration detected OSA comparable to PSG. Further study is warranted to determine its utility for screening and diagnosing OSA in appropriately selected patients.

Mandibular position and movements: Suitability for diagnosis of sleep apnoea.

BACKGROUND AND OBJECTIVE: Mandibular movements (MMs) and position during sleep reflect respiratory efforts related to increases in upper airway resistance and micro-arousals. The study objective was to assess whether MM identifies sleep-disordered breathing (SDB) in patients with moderate to high pre-test probability. METHODS: This was a prospective study of 87 consecutive patients referred for an in-laboratory sleep test. Magnetometer-derived MM signals were incorporated into standard polysomnography (PSG). Respiratory events detected with MM analysis were compared with PSG for respiratory disturbance index (RDI) with a blinded scoring. All records were scored manually according to American Academy of Sleep Medicine rules. Primary outcome was to rule-in obstructive sleep apnoea syndrome (OSAS) defined as RDI cut-off value ≥5 or 15/h total sleep time (TST). RESULTS: High concordance emerged between MM and PSG-derived RDI with high temporal coincidence between events (R2 = 0.906; P < 0.001). The mean diagnostic accuracy of MM for OSAS using RDI MM cut-off values of 5.9 and 13.5 was 0.935 (0.86-0.97) and 0.913 (0.84-0.95), with a mean positive likelihood ratio (LLR+) of 3.73 (2.7-20.4) and 8.46 (2.3-31.5), respectively. Receiver operating characteristic (ROC) curves at PSG cut-off values of 5 and 15/h TST had areas under the curve (AUC) of 0.96 (95% CI: 0.89-0.99) and 0.97 (95% CI: 0.91-0.99) (P < 0.001), respectively. MM analysis accurately identified SDB at different levels of severity. CONCLUSION: RDI assessed by MM is highly concordant with PSG, suggesting a role of ambulatory MM recordings to screen for SDB in patients with moderate to high pre-test probability.

Utility of home sleep apnea testing in high-risk veterans.

PURPOSE: Many Veterans Affairs Medical Centers (VAMCs) have implemented home sleep apnea testing (HSAT) in lieu of traditional in-lab testing to establish a timely and cost-sensitive diagnosis of obstructive sleep apnea (OSA). However, concern remains for the sensitivity and specificity of said technology in this population as many veterans are at increased risk for many of the comorbid conditions that can limit the accuracy of HSAT results. Hence, the purpose of this study is to evaluate rate of incongruent outcomes (e.g., negative HSAT results despite high clinical symptomology) as well as differences in study quality metrics and predictors of OSA between veteran sleep patients and general sleep patients being evaluated by a home sleep test. METHODS: A random sample of HSAT outcomes from 1500 veterans and 1500 general sleep clinic patients was retrieved from a repository of anonymized HSAT outcomes from 2009 to 2013. General sleep clinic data were from patients referred for home sleep testing from a variety of clinical practices across North America, whereas VAMC patients were tested using a central dissemination process. All patients were tested for OSA using the Apnea Risk and Evaluation System (ARES), an HSAT that simultaneously records airflow, pulse oximetry, snoring, accelerometry, and EEG. Sample differences and rates of comorbidities, HSAT outcomes, predictors of OSA, and pretest OSA risk information were evaluated between groups. The presence of OSA was defined as an apnea-hypopnea index (AHI; using 4% desaturation criterion) of ≥5 and ≥15 events per hour. Sample differences in predictors of OSA were evaluated using logistic multiple regression. RESULTS: Veterans (91.3% male) were more likely to report comorbidities, especially depression, insomnia, hypertension, diabetes, restless legs syndrome (RLS), and use of sleep and pain medications compared to general sleep clinic patients (57.1% male). Despite differences in the rate of medical comorbidities, no differences were observed between groups with regard to rates of positive studies, study integrity indicators, or predictors of OSA. Veterans, on average, had 30 min less recording time compared to those in the general clinic sample (p < .01). However, these differences did not impact the amount of the record that was deemed valid nor were veterans more likely to have wakefulness after sleep onset. Predictors of OSA for both groups included advancing age, and increased measures of adiposity (neck circumference and BMI). Mean AHI and respiratory disturbance index (RDI) were statistically similar for both groups and were similar for sleep stage and position. CONCLUSIONS: Home sleep apnea testing for the diagnosis of OSA appears to yield similar results for VAMC patients deemed at high risk for OSA as it does with general sleep clinic patients.

Comparison of a novel non-contact biomotion sensor with wrist actigraphy in estimating sleep quality in patients with obstructive sleep apnoea.

Ambulatory monitoring is of major clinical interest in the diagnosis of obstructive sleep apnoea syndrome. We compared a novel non-contact biomotion sensor, which provides an estimate of both sleep time and sleep-disordered breathing, with wrist actigraphy in the assessment of total sleep time in adult humans suspected of obstructive sleep apnoea syndrome. Both systems were simultaneously evaluated against polysomnography in 103 patients undergoing assessment for obstructive sleep apnoea syndrome in a hospital-based sleep laboratory (84 male, aged 55 ± 14 years and apnoea-hypopnoea index 21 ± 23). The biomotion sensor demonstrated similar accuracy to wrist actigraphy for sleep/wake determination (77.3%: biomotion; 76.5%: actigraphy), and the biomotion sensor demonstrated higher specificity (52%: biomotion; 34%: actigraphy) and lower sensitivity (86%: biomotion; 94%: actigraphy). Notably, total sleep time estimation by the biomotion sensor was superior to actigraphy (average overestimate of 10 versus 57 min), especially at a higher apnoea-hypopnoea index. In post hoc analyses, we assessed the improved apnoea-hypopnoea index accuracy gained by combining respiratory measurements from polysomnography for total recording time (equivalent to respiratory polygraphy) with total sleep time derived from actigraphy or the biomotion sensor. Here, the number of misclassifications of obstructive sleep apnoea severity compared with full polysomnography was reduced from 10/103 (for total respiratory recording time alone) to 7/103 and 4/103 (for actigraphy and biomotion sensor total sleep time estimate, respectively). We conclude that the biomotion sensor provides a viable alternative to actigraphy for sleep estimation in the assessment of obstructive sleep apnoea syndrome. As a non-contact device, it is suited to longitudinal assessment of sleep, which could also be combined with polygraphy in ambulatory studies.

Validation of a portable monitor compared with polysomnography for screening of obstructive sleep apnea in polio survivors.

Subjective: Sleep-disordered breathing (SDB) is highly prevalent in polio survivors. Obstructive sleep apnea (OSA) is the most frequent type. Full polysomnography (PSG) is recommended for OSA diagnosis in patients with comorbidities by current practice guidelines, but it is not always accessible. The purpose of this study was to evaluate whether type 3 portable monitor (PM) or type 4 PM might be a viable alternative to PSG for the diagnosis of OSA in postpolio subjects. Methods: A total of 48 community-living polio survivors (39 men and 9 women) with an average age of 54.4 ± 5.3 years referred for the evaluation of OSA and who volunteered to participate were recruited. First, they completed the Epworth Sleepiness Scale (ESS) questionnaire and underwent pulmonary function testing and blood gas tests the day before PSG night. Then, they underwent an overnight in-laboratory PSG with a type 3 PM and type 4 PM recording simultaneously. Results: The AHI from PSG, respiratory event index (REI) from type 3 PM, and ODI3 from type 4 PM was 30.27 ± 22.51/h vs. 25.18 ± 19.11/h vs. 18.28 ± 15.13/h, respectively (P < 0.001). For AHI ≥ 5/h, the sensitivity and specificity of REI were 95.45 and 50%, respectively. For AHI ≥ 15/h, the sensitivity and specificity of REI were 87.88% and 93.33%, respectively. The Bland-Altman analysis of REI on PM vs. AHI on PSG showed a mean difference of -5.09 (95% confidence interval [CI]: -7.10, -3.08; P < 0.001) with limits of agreement ranging from -18.67 to 8.49 events/h. ROC curve analysis for patients with REI ≥ 15/h showed an area under the curve (AUC) of 0.97. For AHI ≥ 5/h, the sensitivity and specificity of ODI3 from type 4 PM were 86.36 and 75%, respectively. For patients with AHI ≥ 15/h, the sensitivity was 66.67%, and the specificity was 100%. Conclusion: Type 3 PM and Type 4 PM could be alternative ways to screen OSA for polio survivors, especially for moderate to severe OSA.

Obstructive sleep apnea event detection using explainable deep learning models for a portable monitor.

Background: Polysomnography (PSG) is the gold standard for detecting obstructive sleep apnea (OSA). However, this technique has many disadvantages when using it outside the hospital or for daily use. Portable monitors (PMs) aim to streamline the OSA detection process through deep learning (DL). Materials and methods: We studied how to detect OSA events and calculate the apnea-hypopnea index (AHI) by using deep learning models that aim to be implemented on PMs. Several deep learning models are presented after being trained on polysomnography data from the National Sleep Research Resource (NSRR) repository. The best hyperparameters for the DL architecture are presented. In addition, emphasis is focused on model explainability techniques, concretely on Gradient-weighted Class Activation Mapping (Grad-CAM). Results: The results for the best DL model are presented and analyzed. The interpretability of the DL model is also analyzed by studying the regions of the signals that are most relevant for the model to make the decision. The model that yields the best result is a one-dimensional convolutional neural network (1D-CNN) with 84.3% accuracy. Conclusion: The use of PMs using machine learning techniques for detecting OSA events still has a long way to go. However, our method for developing explainable DL models demonstrates that PMs appear to be a promising alternative to PSG in the future for the detection of obstructive apnea events and the automatic calculation of AHI.

Efficacy of portable sleep monitoring device in diagnosing central sleep apnea in patients with congestive heart failure.

Introduction: Central sleep apnea (CSA) is a common and serious comorbidity mainly occurring in patients with heart failure (HF), which tends to be underdiagnosed and has not been widely studied. Overnight polysomnography (PSG) is the gold standard for diagnosing CSA; however, the time and expense of the procedure limit its applicability. Portable monitoring (PM) devices are convenient and easy to use; however, they have not been widely studied as to their effectiveness in detecting CSA in patients with HF. In the current study, we examined the diagnostic value of PM as a screening tool to identify instances of CSA among patients with HF. Methods: A total of 22 patients under stable heart failure conditions with an ejection fraction of <50% were enrolled. All patients underwent PM and overnight PSG within a narrow time frame. The measurements of the apnea-hypopnea index (AHI), hypopnea index (HI), central apnea index (CAI), and obstructive apnea index (OAI) obtained from PSG, automatic scoring, and manual scoring of PM were recorded. The results obtained from PSG and those from PM (automatic and manual scoring) were compared to assess the accuracy of PM. Results: Among the patients, CSA in 11 patients was found by PSG. The AHI measurements performed using manual scoring of PM showed a significant correlation with those performed using PSG (r = 0.69; P = 0.01). Nonetheless, mean AHI measurements showed statistically significant differences between PSG and automatic scoring of PM (40.0 vs. 23.7 events/hour, respectively; P < 0.001), as well as between automatic and manual scoring of PM (23.7 vs. 29.5 events/hour; P < 0.001). Central sleep apnea was detected by PM; however, the results were easily misread as obstructive apnea, particularly in automatic scoring. Conclusion: PM devices could be used to identify instances of central sleep apnea among patients with HF. The results from PM were well-correlated with standard PSG results, and manual scoring was preferable to automated scoring.

Pulse wave amplitude drops during sleep are reliable surrogate markers of changes in cortical activity.

BACKGROUND: During sleep, sudden drops in pulse wave amplitude (PWA) measured by pulse oximetry are commonly associated with simultaneous arousals and are thought to result from autonomic vasoconstriction. In the present study, we determine whether PWA drops were associated with changes in cortical activity as determined by EEG spectral analysis. METHODS: A 20% decrease in PWA was chosen as a minimum for a drop. A total of 1085 PWA drops from 10 consecutive sleep recordings were analyzed. EEG spectral analysis was performed over 5 consecutive epochs of 5 seconds: 2 before, 1 during, and 2 after the PWA drop. EEG spectral analysis was performed over delta, theta, alpha, sigma, and beta frequency bands. Within each frequency band, power density was compared across the five 5-sec epochs. Presence or absence of visually scored EEG arousals were adjudicated by an investigator blinded to the PWA signal and considered associated with PWA drop if concomitant. RESULTS: A significant increase in EEG power density in all EEG frequency bands was found during PWA drops (P<0.001) compared to before and after drop. Even in the absence of visually scored arousals, PWA drops were associated with a significant increase in EEG power density (P<0.001) in most frequency bands. CONCLUSIONS: Drops in PWA are associated with a significant increase in EEG power density, suggesting that these events can be used as a surrogate for changes in cortical activity during sleep. This approach may prove of value in scoring respiratory events on limited-channel (type III) portable monitors.

The relationship between personal exposure and ambient PM2.5 and black carbon in Beijing.

This study is part of the "Air Polluion Impacts on Cardiopulmonary disease in Beijing: an integrated study of Exposure Science, Toxicologenomics & Environmental Epidemiology (APIC-ESTEE)" project under the UK-China joint research programme "Atmospheric Pollution and Human Health in a Chinese Megacity (APHH-China)". The aim is to capture the spatio-temporal variability in people's exposure to fine particles (PM2.5) and black carbon (BC) air pollution in Beijing, China. A total of 120 students were recruited for a panel study from ten universities in Haidian District in northwestern Beijing from December 2017 to June 2018. Real-time personal concentrations of PM2.5 and BC were measured over a 24-h period with two research-grade portable personal exposure monitors. Personal microenvironments (MEs) were determined by applying an algorithm to the handheld GPS unit data. On average, the participants spent the most time indoors (79% in Residence and 16% in Workplace), and much less time travelling by Walking, Cycling, Bus and Metro. Similar patterns were observed across participant gender and body-mass index classifications. The participants were exposed to 33.8 ± 27.8 µg m-3 PM2.5 and to 1.9 ± 1.2 µg m-3 BC over the 24-h monitoring period, on average 24.3 µg m-3 (42%) and 0.8 µg m-3 (28%) lower, respectively, than the concurrent fixed-site ambient measurements. Relative differences between personal and ambient BC concentrations showed greater variability across the MEs, highlighting significant contributions from Dining and travelling by Bus, which involve potential combustion of fuels. This study demonstrates the potential value of personal exposure monitoring in investigating air pollution related health effects, and in evaluating the effectiveness of pollution control and intervention measures.

The Evaluation of Autonomic Arousals in Scoring Sleep Respiratory Disturbances with Polysomnography and Portable Monitor Devices: A Proof of Concept Study.

BACKGROUND: Autonomic arousals can be considered as surrogates of electroencephalography (EEG) arousals when calculating respiratory disturbance index (RDI). The main objective of this proof of concept study was to evaluate the use of heart rate acceleration (HRa) arousals associated with sleep respiratory events in a population undergoing full polysomnography (type 1) and in another undergoing portable monitor study (type 3). Our hypothesis is that when compared to other commonly used indexes, RDI based on HRa will capture more events in both types of recording. MATERIALS AND METHODS: A retrospective analysis was performed in two different populations of patients with suspected OSA: a) 72 patients undergoing one night of type 1 recording and b) 79 patients undergoing one night of type 3 recording. Variables for type 1 were 4% oxygen desaturation index (ODI), apnea/hypopnea index (AHI), RDI based on EEG arousals (RDIe), and RDI based on HRa with threshold of 5bpm (RDIa5). For type 3, variables were 4% ODI, AHI, and RDIa5 (it is not possible to calculate RDIe due to the absence of EEG). Calculated data were 1) Mean values for each sleep disturbance index in type 1 and 3 recordings; 2) Frequency of migration from lower to higher OSA severity categories using RDIa5 in comparison to AHI (thresholds: ≥5/h mild, ≥15/h moderate, ≥30/h severe); and 3) Bland-Altman plots to assess agreement between AHI vs RDIe and RDIa5 in type 1 population, and AHI vs RDIa5 in type 3 populations. RESULTS: More respiratory disturbance events were captured with RDIa5 index in both type 1 and type 3 recordings when compared to the other indexes. In type 1 recording, when using RDIa5 37% of patients classified as not having OSA with AHI were now identified as having OSA, and a total of 59% migrated to higher severity categories. In type 3 recording, similar results were obtained, as 37% of patients classified as not having OSA with AHI were now identified as having OSA using RDIa5, and a total of 55% patients migrated to higher severity categories. Mean differences for RDIa5 and AHI in type 1 and 3 populations were similar. CONCLUSION: The use of autonomic arousals such as HRa can help to detect more respiratory disturbance events when compared to other indexes, being a variable that may help to capture borderline mild cases. This becomes especially relevant in type 3 recordings. Future research is needed to determine its validity, optimization, and its clinical significance.

Tracheal Sound Analysis Using a Deep Neural Network to Detect Sleep Apnea.

STUDY OBJECTIVES: Portable devices for home sleep apnea testing are often limited by their inability to discriminate sleep/wake status, possibly resulting in underestimations. Tracheal sound (TS), which can be visualized as a spectrogram, carries information about apnea/hypopnea and sleep/wake status. We hypothesized that image analysis of all-night TS recordings by a deep neural network (DNN) would be capable of detecting breathing events and classifying sleep/wake status. The aim of this study is to develop a DNN-based system for sleep apnea testing and validate it using a large sampling of polysomnography (PSG) data. METHODS: PSG examinations for the evaluation of sleep-disordered breathing (SDB) were performed for 1,852 patients: 1,548 PSG records were used to develop the system, and the remaining 304 records were used for validation. TS spectrogram images were obtained every 60 seconds and labeled with the PSG scoring results (breathing event and sleep/wake status), then introduced to DNN learning. Two different DNNs were trained for breathing status and sleep/wake status, respectively. RESULTS: A DNN with convolutional layers showed the best performance for discriminating breathing status. The same DNN structure was trained for sleep/wake discrimination. In the validation study, the DNN analysis was capable of discriminating the sleep/wake status with reasonable accuracy. The diagnostic sensitivity, specificity, and area under the receiver operating characteristic curves for diagnosis of SDB with apnea-hypopnea index of > 5, 15, and 30 were 0.98, 0.76, and 0.99; 0.97, 0.90, and 0.99; and 0.92, 0.94, and 0.98, respectively. CONCLUSIONS: The developed system using the TS DNN analysis has a good performance for SDB testing. CITATION: Nakano H, Furukawa T, Tanigawa T. Tracheal sound analysis using a deep neural network to detect sleep apnea. J Clin Sleep Med. 2019;15(8): 1125-1133.

Accuracy of Type III Portable Monitors for Diagnosing Obstructive Sleep Apnea.

OBJECTIVES: We evaluated whether the results from portable monitor (PM) devices for the diagnosis of obstructive sleep apnea (OSA), classified into type III and type IV devices by the American Academy of Sleep Medicine, correlated with the results from polysomnography (PSG) testing. METHODS: Sixty-four patients with a sleep-breathing disorder used type III or type IV PM devices at home and were subsequently admitted for testing using PSG. The apnea-hypopnea index (AHI) from each machine was measured, and the AHI component, apnea index (AI), and hypopnea index (HI) were also analyzed. RESULTS: There was a stronger correlation between the AHI values from PSG testing and those from the type III PM devices (r = 0.92, p < 0.001) than for the data from type IV devices (r = 0.69, p < 0.001). However, the correlation of HI values (type III: r = 0.43, p = 0.024; type IV: r = 0.14, p = 0.41) was poorer than that of the AI values (type III: r = 0.95, p < 0.001; type IV: r = 0.68, p < 0.001). Moreover, the type III PM devices tended to evaluate a patient's condition as less severe than did PSG testing when the AHI value was over 30. CONCLUSIONS: Although type III PM devices outperformed type IV devices as substitutes for PSG, the clinical state must be evaluated for patients suspected of having obstructive sleep apnea.