Sleep staging in the ICU with heart rate variability and breathing signals. An exploratory cross-sectional study using deep neural networks.

Introduction: To measure sleep in the intensive care unit (ICU), full polysomnography is impractical, while activity monitoring and subjective assessments are severely confounded. However, sleep is an intensely networked state, and reflected in numerous signals. Here, we explore the feasibility of estimating conventional sleep indices in the ICU with heart rate variability (HRV) and respiration signals using artificial intelligence methods Methods: We used deep learning models to stage sleep with HRV (through electrocardiogram) and respiratory effort (through a wearable belt) signals in critically ill adult patients admitted to surgical and medical ICUs, and in age and sex-matched sleep laboratory patients Results: We studied 102 adult patients in the ICU across multiple days and nights, and 220 patients in a clinical sleep laboratory. We found that sleep stages predicted by HRV- and breathing-based models showed agreement in 60% of the ICU data and in 81% of the sleep laboratory data. In the ICU, deep NREM (N2 + N3) proportion of total sleep duration was reduced (ICU 39%, sleep laboratory 57%, p < 0.01), REM proportion showed heavy-tailed distribution, and the number of wake transitions per hour of sleep (median 3.6) was comparable to sleep laboratory patients with sleep-disordered breathing (median 3.9). Sleep in the ICU was also fragmented, with 38% of sleep occurring during daytime hours. Finally, patients in the ICU showed faster and less variable breathing patterns compared to sleep laboratory patients Conclusion: The cardiovascular and respiratory networks encode sleep state information, which, together with artificial intelligence methods, can be utilized to measure sleep state in the ICU.

High prevalence of sleep-disordered breathing in the intensive care unit - a cross-sectional study.

PURPOSE: Sleep-disordered breathing may be induced by, exacerbate, or complicate recovery from critical illness. Disordered breathing during sleep, which itself is often fragmented, can go unrecognized in the intensive care unit (ICU). The objective of this study was to investigate the prevalence, severity, and risk factors of sleep-disordered breathing in ICU patients using a single respiratory belt and oxygen saturation signals. METHODS: Patients in three ICUs at Massachusetts General Hospital wore a thoracic respiratory effort belt as part of a clinical trial for up to 7 days and nights. Using a previously developed machine learning algorithm, we processed respiratory and oximetry signals to measure the 3% apnea-hypopnea index (AHI) and estimate AH-specific hypoxic burden and periodic breathing. We trained models to predict AHI categories for 12-h segments from risk factors, including admission variables and bio-signals data, available at the start of these segments. RESULTS: Of 129 patients, 68% had an AHI ≥ 5; 40% an AHI > 15, and 19% had an AHI > 30 while critically ill. Median [interquartile range] hypoxic burden was 2.8 [0.5, 9.8] at night and 4.2 [1.0, 13.7] %min/h during the day. Of patients with AHI ≥ 5, 26% had periodic breathing. Performance of predicting AHI-categories from risk factors was poor. CONCLUSIONS: Sleep-disordered breathing and sleep apnea events while in the ICU are common and are associated with substantial burden of hypoxia and periodic breathing. Detection is feasible using limited bio-signals, such as respiratory effort and SpO2 signals, while risk factors were insufficient to predict AHI severity.

Validation Study of Airgo, an Innovative Device to Screen Sleep Respiratory Disorders.

Background: Obstructive sleep apnea affects a consistent percentage of the population, and only a minority of patients have been diagnosed and treated because of a discrepancy between resources available for diagnosis and the epidemiology of a disorder possibly affecting nearly one billion people in the world. Aim: We conducted a study to compare a standard home respiratory monitoring system (Nox T3) with a novel device (Airgo™) consisting of an elastic band and a small recorder, light, comfortable for the patient, and low-cost complete with automatic analysis of the data that produces a screening report indicating the type and severity of sleep respiratory disorder. Patients and Results: We examined 120 patients, reduced to 118 for technical problems. The mean (SD) age of the patients is 55.7 ± 13 years, their BMI is 27.8 ± 4.3 kg/m2, and their AHI is 22 ± 22 events/h. Patients belong to all the different severity rates of OSA, with a percentage of them classified as free of respiratory disorders. The Airgo™ showed excellent agreement with the results of the gold standard, reporting high levels of sensitivity, specificity, positive and negative predicted value, and accuracy. Conclusion: Airgo™ is a reliable tool to screen patients with suspected sleep respiratory disorders, well tolerated by the patient based on totally automatic analysis and reporting system, leading to more efficient use of doctor's and clinician's time and resources and extending the opportunity to diagnose more possible candidates for treatment.

Sleep apnea and respiratory anomaly detection from a wearable band and oxygen saturation.

OBJECTIVE: Sleep-related respiratory abnormalities are typically detected using polysomnography. There is a need in general medicine and critical care for a more convenient method to detect sleep apnea automatically from a simple, easy-to-wear device. The objective was to detect abnormal respiration and estimate the Apnea-Hypopnea Index (AHI) automatically with a wearable respiratory device with and without SpO2 signals using a large (n = 412) dataset serving as ground truth. DESIGN: Simultaneously recorded polysomnography (PSG) and wearable respiratory effort data were used to train and evaluate models in a cross-validation fashion. Time domain and complexity features were extracted, important features were identified, and a random forest model was employed to detect events and predict AHI. Four models were trained: one each using the respiratory features only, a feature from the SpO2 (%)-signal only, and two additional models that use the respiratory features and the SpO2 (%) feature, one allowing a time lag of 30 s between the two signals. RESULTS: Event-based classification resulted in areas under the receiver operating characteristic curves of 0.94, 0.86, and 0.82, and areas under the precision-recall curves of 0.48, 0.32, and 0.51 for the models using respiration and SpO2, respiration-only, and SpO2-only, respectively. Correlation between expert-labelled and predicted AHI was 0.96, 0.78, and 0.93, respectively. CONCLUSIONS: A wearable respiratory effort signal with or without SpO2 signal predicted AHI accurately, and best performance was achieved with using both signals.

A Home Telemedicine System for Continuous Respiratory Monitoring.

This article presents a continuous home telemonitoring system for chronic respiratory patients using 5G connectivity developed in partnership with Vodafone as a part of the 5G Trial in Milan established by the Italian Ministry of Economic Development. The system features a wearable respiratory and activity monitor, an environmental sensor and a pulse oximeter sending the data through a 5G router to a Multi-Edge Computing server, incorporated in the Vodafone 5G infrastructure, where they are stored and accessible for visualization. In particular, activity, respiratory and environmental data are continuously streamed and collected. The solution has been tested on 18 healthy volunteers during non-supervised recordings lasting at least 48 hours. The combination of recognized activities and associated respiratory parameters provided statistically significant variations in breathing patterns between one activity and the other, thus giving more complete information to the clinicians than previously studied telemedicine systems based on spot-checks. In particular, statistically significant differences are found in tidal volume and minute ventilation between horizontal and vertical postures (p < 0.001) and between vertical postures and dynamic activities (p < 0.001); the respiratory rate shows statistically significant differences between horizontal and vertical postures (p < 0.001). Some environmental parameters have different mean values between day and night, such as carbon dioxide (p < 0.001). Trials on patients are needed to further study this telemedicine solution and make it commercially available in the future. The main further technical development suggested is the use of commercial 5G smartphones as routers, in order to make the system usable outside of home settings.

Comparison between the Airgo™ Device and a Metabolic Cart during Rest and Exercise.

The aim of this study is to compare the accuracy of Airgo™, a non-invasive wearable device that records breath, with respect to a gold standard. In 21 healthy subjects (10 males, 11 females), four parameters were recorded for four min at rest and in different positions simultaneously by Airgo™ and SensorMedics 2900 metabolic cart. Then, a cardio-pulmonary exercise test was performed using the Erg 800S cycle ergometer in order to test Airgo™'s accuracy during physical effort. The results reveal that the relative error median percentage of respiratory rate was of 0% for all positions at rest and for different exercise intensities, with interquartile ranges between 3.5 (standing position) and 22.4 (low-intensity exercise) breaths per minute. During exercise, normalized amplitude and ventilation relative error medians highlighted the presence of an error proportional to the volume to be estimated. For increasing intensity levels of exercise, Airgo™'s estimate tended to underestimate the values of the gold standard instrument. In conclusion, the Airgo™ device provides good accuracy and precision in the estimate of respiratory rate (especially at rest), an acceptable estimate of tidal volume and minute ventilation at rest and an underestimation for increasing volumes.