Achieving Better Understanding of Obstructive Sleep Apnea Treatment Effects on Cardiovascular Disease Outcomes through Machine Learning Approaches: A Narrative Review.

Obstructive sleep apnea (OSA) affects almost a billion people worldwide and is associated with a myriad of adverse health outcomes. Among the most prevalent and morbid are cardiovascular diseases (CVDs). Nonetheless, randomized controlled trials (RCTs) of OSA treatment have failed to show improvements in CVD outcomes. A major limitation in our field is the lack of precision in defining OSA and specifically subgroups with the potential to benefit from therapy. Further, this has called into question the validity of using the time-honored apnea-hypopnea index as the ultimate defining criteria for OSA. Recent applications of advanced statistical methods and machine learning have brought to light a variety of OSA endotypes and phenotypes. These methods also provide an opportunity to understand the interaction between OSA and comorbid diseases for better CVD risk stratification. Lastly, machine learning and specifically heterogeneous treatment effects modeling can help uncover subgroups with differential outcomes after treatment initiation. In an era of data sharing and big data, these techniques will be at the forefront of OSA research. Advanced data science methods, such as machine-learning analyses and artificial intelligence, will improve our ability to determine the unique influence of OSA on CVD outcomes and ultimately allow us to better determine precision medicine approaches in OSA patients for CVD risk reduction. In this narrative review, we will highlight how team science via machine learning and artificial intelligence applied to existing clinical data, polysomnography, proteomics, and imaging can do just that.

A foundational transformer leveraging full night, multichannel sleep study data accurately classifies sleep stages.

Study Objectives: To investigate whether a foundational transformer model using 8-hour, multi-channel data from polysomnograms can outperform existing artificial intelligence (AI) methods for sleep stage classification. Methods: We utilized the Sleep Heart Health Study (SHHS) visits 1 and 2 for training and validation and the Multi-Ethnic Study of Atherosclerosis (MESA) for testing of our model. We trained a self-supervised foundational transformer (called PFTSleep) that encodes 8-hour long sleep studies at 125 Hz with 7 signals including brain, movement, cardiac, oxygen, and respiratory channels. These encodings are used as input for training of an additional model to classify sleep stages, without adjusting the weights of the foundational transformer. We compared our results to existing AI methods that did not utilize 8-hour data or the full set of signals but did report evaluation metrics for the SHHS dataset. Results: We trained and validated a model with 8,444 sleep studies with 7 signals including brain, movement, cardiac, oxygen, and respiratory channels and tested on an additional 2,055 studies. In total, we trained and tested 587,944 hours of sleep study signal data. Area under the precision recall curve (AUPRC) scores were 0.82, 0.40, 0.53, 0.75, and 0.82 and area under the receiving operating characteristics curve (AUROC) scores were 0.99, 0.95, 0.96, 0.98, and 0.99 for wake, N1, N2, N3, and REM, respectively, on the SHHS validation set. For MESA, the AUPRC scores were 0.56, 0.16, 0.40, 0.45, and 0.65 and AUROC scores were 0.94, 0.77, 0.87, 0.91, and 0.96, respectively. Our model was compared to the longest context window state-of-the-art model and showed increases in macro evaluation scores, notably sensitivity (3.7% increase) and multi-class REM (3.39% increase) and wake (0.97% increase) F1 scores. Conclusions: Utilizing full night, multi-channel PSG data encodings derived from a foundational transformer improve sleep stage classification over existing methods.

Metabolomic profiles of sleep-disordered breathing are associated with hypertension and diabetes mellitus development.

Sleep-disordered breathing (SDB) is a prevalent disorder characterized by recurrent episodic upper airway obstruction. Using data from the Hispanic Community Health Study/Study of Latinos (HCHS/SOL), we apply principal component analysis (PCA) to seven SDB-related measures. We estimate the associations of the top two SDB PCs with serum levels of 617 metabolites, in both single-metabolite analysis, and a joint penalized regression analysis. The discovery analysis includes 3299 individuals, with validation in a separate dataset of 1522 individuals. Five metabolite associations with SDB PCs are discovered and replicated. SDB PC1, characterized by frequent respiratory events common in older and male adults, is associated with pregnanolone and progesterone-related sulfated metabolites. SDB PC2, characterized by short respiratory event length and self-reported restless sleep, enriched in young adults, is associated with sphingomyelins. Metabolite risk scores (MRSs), representing metabolite signatures associated with the two SDB PCs, are associated with 6-year incident hypertension and diabetes. These MRSs have the potential to serve as biomarkers for SDB, guiding risk stratification and treatment decisions.

Heterogeneous Effects of Continuous Positive Airway Pressure in Non-Sleepy Obstructive Sleep Apnea on Cardiovascular Disease Outcomes: Post Hoc Machine Learning Analysis of the ISAACC Trial (ECSACT Study).

Rationale: Randomized controlled trials of continuous positive airway pressure (CPAP) therapy for cardiovascular disease (CVD) prevention among patients with obstructive sleep apnea (OSA) have been largely neutral. However, given that OSA is a heterogeneous disease, there may be unidentified subgroups demonstrating differential treatment effects. Objectives: We sought to apply a novel data-drive approach to identify nonsleepy OSA subgroups with heterogeneous effects of CPAP on CVD outcomes within the Impact of Sleep Apnea Syndrome in the Evolution of Acute Coronary Syndrome (ISAACC) study. Methods: Participants were randomly partitioned into two datasets. One for training (70%) our machine-learning model and a second (30%) for validation of significant findings. Model-based recursive partitioning was applied to identify subgroups with heterogeneous treatment effects. Survival analysis was conducted to compare treatment (CPAP vs. usual care [UC]) outcomes within subgroups. Results: A total of 1,224 nonsleepy OSA participants were included. Of 55 features entered into our model, only two appeared in the final model (i.e., average OSA event duration and hypercholesterolemia). Among participants at or below the model-derived average event duration threshold (19.5 s), CPAP was protective for a composite of CVD events (training hazard ratio [HR], 0.46; P = 0.002). For those with longer event duration (>19.5 s), an additional split occurred by hypercholesterolemia status. Among participants with longer event duration and hypercholesterolemia, CPAP resulted in more CVD events compared with UC (training HR, 2.24; P = 0.011). The point estimate for this harmful signal was also replicated in the testing dataset (HR, 1.83; P = 0.118). Conclusions: We discovered subgroups of nonsleepy OSA participants within the ISAACC study with heterogeneous effects of CPAP. Among the training dataset, those with longer OSA event duration and hypercholesterolemia had nearly 2.5 times more CVD events with CPAP compared with UC, whereas those with shorter OSA event duration had roughly half the rate of CVD events if randomized to CPAP.