Machine learning derived serum creatinine trajectories in acute kidney injury in critically ill patients with sepsis.

BACKGROUND: Current classification for acute kidney injury (AKI) in critically ill patients with sepsis relies only on its severity-measured by maximum creatinine which overlooks inherent complexities and longitudinal evaluation of this heterogenous syndrome. The role of classification of AKI based on early creatinine trajectories is unclear. METHODS: This retrospective study identified patients with Sepsis-3 who developed AKI within 48-h of intensive care unit admission using Medical Information Mart for Intensive Care-IV database. We used latent class mixed modelling to identify early creatinine trajectory-based classes of AKI in critically ill patients with sepsis. Our primary outcome was development of acute kidney disease (AKD). Secondary outcomes were composite of AKD or all-cause in-hospital mortality by day 7, and AKD or all-cause in-hospital mortality by hospital discharge. We used multivariable regression to assess impact of creatinine trajectory-based classification on outcomes, and eICU database for external validation. RESULTS: Among 4197 patients with AKI in critically ill patients with sepsis, we identified eight creatinine trajectory-based classes with distinct characteristics. Compared to the class with transient AKI, the class that showed severe AKI with mild improvement but persistence had highest adjusted risks for developing AKD (OR 5.16; 95% CI 2.87-9.24) and composite 7-day outcome (HR 4.51; 95% CI 2.69-7.56). The class that demonstrated late mild AKI with persistence and worsening had highest risks for developing composite hospital discharge outcome (HR 2.04; 95% CI 1.41-2.94). These associations were similar on external validation. CONCLUSIONS: These 8 classes of AKI in critically ill patients with sepsis, stratified by early creatinine trajectories, were good predictors for key outcomes in patients with AKI in critically ill patients with sepsis independent of their AKI staging.

Analysis by sex of safety and effectiveness of transvenous phrenic nerve stimulation.

PURPOSE: Little is known about sex differences in the treatment of central sleep apnea (CSA). Our post hoc analysis of the remede System Pivotal Trial aimed to determine sex-specific differences in the safety and effectiveness of treating moderate to severe CSA in adults with transvenous phrenic nerve stimulation (TPNS). METHODS: Men and women enrolled in the remede System Pivotal Trial were included in this post hoc analysis of the effect of TPNS on polysomnographic measures, Epworth Sleepiness Scale, and patient global assessment for quality of life. RESULTS: Women (n = 16) experienced improvement in CSA metrics that were comparable to the benefits experienced by men (n = 135), with central apneas being practically eliminated post TPNS. Women experienced improvement in sleep quality and architecture that was comparable to men post TPNS. While women had lower baseline apnea hypopnea index than men, their quality of life was worse at baseline. Additionally, women reported a 25-percentage point greater improvement in quality of life compared to men after 12 months of TPNS therapy. TPNS was found to be safe in women, with no related serious adverse events through 12 months post-implant, while men had a low rate of 10%. CONCLUSION: Although women had less prevalent and less severe CSA than men, they were more likely to report reduced quality of life. Transvenous phrenic nerve stimulation may be a safe and effective tool in the treatment of moderate to severe CSA in women. Larger studies of women with CSA are needed to confirm our findings. CLINICAL TRIAL REGISTRATION: ClinicalTrials.gov NCT01816776; March 22, 2013.

Novel non-invasive assessment of upper airway inflammation in obstructive sleep apnea using positron emission tomography/magnetic resonance imaging.

PURPOSE: To develop a novel non-invasive technique to quantify upper airway inflammation using positron emission tomography/magnetic resonance imaging (PET/MRI) in patients with obstructive sleep apnea (OSA). METHODS: Patients with treatment naïve moderate-to-severe OSA underwent [18F]-fluoro-2-deoxy-D-glucose (FDG) PET/MRI. Three readers independently performed tracings of the pharyngeal soft tissue on MRI. Standardized uptake values (SUV) were generated from region of interest (ROI) tracings on corresponding PET images. Background SUV was measured from the sternocleidomastoid muscle. SUV and target-to-background (TBR) were compared across readers using intraclass correlation coefficient (ICC) analyses. SUV from individual image slices were compared between each reader using Bland-Altman plots and Pearson correlation coefficients. All tracings were repeated by one reader for assessment of intra-reader reliability. RESULTS: Five participants completed our imaging protocol and analysis. Median age, body mass index, and apnea-hypopnea index were 41 years (IQR 40.5-68.5), 32.7 kg/m2 (IQR 28.1-38.1), and 30.7 event per hour (IQR 19.5-48.1), respectively. The highest metabolic activity regions were consistently localized to palatine or lingual tonsil adjacent mucosa. Twenty-five ICC met criteria for excellent agreement. The remaining three were TBR measurements which met criteria for good agreement. Head-to-head comparisons revealed strong correlation between each reader. CONCLUSIONS: Our novel imaging technique demonstrated reliable quantification of upper airway FDG avidity. This technology has implications for future work exploring local airway inflammation in individuals with OSA and exposure to pollutants. It may also serve as an assessment tool for response to OSA therapies.

Associations of variants In the hexokinase 1 and interleukin 18 receptor regions with oxyhemoglobin saturation during sleep.

Sleep disordered breathing (SDB)-related overnight hypoxemia is associated with cardiometabolic disease and other comorbidities. Understanding the genetic bases for variations in nocturnal hypoxemia may help understand mechanisms influencing oxygenation and SDB-related mortality. We conducted genome-wide association tests across 10 cohorts and 4 populations to identify genetic variants associated with three correlated measures of overnight oxyhemoglobin saturation: average and minimum oxyhemoglobin saturation during sleep and the percent of sleep with oxyhemoglobin saturation under 90%. The discovery sample consisted of 8,326 individuals. Variants with p < 1 × 10(-6) were analyzed in a replication group of 14,410 individuals. We identified 3 significantly associated regions, including 2 regions in multi-ethnic analyses (2q12, 10q22). SNPs in the 2q12 region associated with minimum SpO2 (rs78136548 p = 2.70 × 10(-10)). SNPs at 10q22 were associated with all three traits including average SpO2 (rs72805692 p = 4.58 × 10(-8)). SNPs in both regions were associated in over 20,000 individuals and are supported by prior associations or functional evidence. Four additional significant regions were detected in secondary sex-stratified and combined discovery and replication analyses, including a region overlapping Reelin, a known marker of respiratory complex neurons.These are the first genome-wide significant findings reported for oxyhemoglobin saturation during sleep, a phenotype of high clinical interest. Our replicated associations with HK1 and IL18R1 suggest that variants in inflammatory pathways, such as the biologically-plausible NLRP3 inflammasome, may contribute to nocturnal hypoxemia.

Admixture mapping identifies novel loci for obstructive sleep apnea in Hispanic/Latino Americans.

Obstructive sleep apnea (OSA) is a common disorder associated with increased risk of cardiovascular disease and mortality. Its prevalence and severity vary across ancestral background. Although OSA traits are heritable, few genetic associations have been identified. To identify genetic regions associated with OSA and improve statistical power, we applied admixture mapping on three primary OSA traits [the apnea hypopnea index (AHI), overnight average oxyhemoglobin saturation (SaO2) and percentage time SaO2 < 90%] and a secondary trait (respiratory event duration) in a Hispanic/Latino American population study of 11 575 individuals with significant variation in ancestral background. Linear mixed models were performed using previously inferred African, European and Amerindian local genetic ancestry markers. Global African ancestry was associated with a lower AHI, higher SaO2 and shorter event duration. Admixture mapping analysis of the primary OSA traits identified local African ancestry at the chromosomal region 2q37 as genome-wide significantly associated with AHI (P < 5.7 × 10-5), and European and Amerindian ancestries at 18q21 suggestively associated with both AHI and percentage time SaO2 < 90% (P < 10-3). Follow-up joint ancestry-SNP association analyses identified novel variants in ferrochelatase (FECH), significantly associated with AHI and percentage time SaO2 < 90% after adjusting for multiple tests (P < 8 × 10-6). These signals contributed to the admixture mapping associations and were replicated in independent cohorts. In this first admixture mapping study of OSA, novel associations with variants in the iron/heme metabolism pathway suggest a role for iron in influencing respiratory traits underlying OSA.

Modifying pathways by age and sex for the association between combined sleep disordered breathing and long sleep duration with neurocognitive decline in the Hispanic Community Health Study/Study of Latinos (HCHS/SOL).

INTRODUCTION: We aimed to determine whether obesity or metabolic syndrome (MetS) modify associations between sleep-disordered breathing (SDB), self-reported sleep duration (SD), and phenotypes of combined SDB/SD with 7-year neurocognitive decline (ND) in a community based-cohort of U.S. Hispanic/Latinos (N = 5500) in different age and sex groups. METHODS: The exposures were baseline SDB (respiratory event index ≥ 15), sleepiness (Epworth Sleepiness Scale ≥ 10), SD (< 6 hours, 6-9 hours, ≥ 9 hours). The outcome was 7-year ND. RESULTS: Mean age was 56.0 years, 54.8% were females. Obesity modified the association between SDB/SD and ND in memory (F = 21.49, P < 0.001) and global cognition (F = 9.14, P < 0.001) in the oldest age group. Women without MetS with combined long sleep/SDB exhibited most pronounced decline in global cognition (F = 3.07, P = 0.010). DISCUSSION: The association between combined SDB/long sleep and declines in memory and global cognition was most pronounced in obese older adults. Among women, MetS status modified the association between long sleep/SDB and decline in global cognition.

Self-reported snoring and incident cardiovascular disease events: results from the Jackson Heart Study.

PURPOSE: Evidence suggests that snoring is associated with increased risk for cardiovascular disease (CVD) events such as myocardial infarction and stroke. Limited data exists pertaining to this association among African Americans. We therefore examined the association between self-reported habitual snoring and incident CVD in the Jackson Heart Study (JHS), a population-based cohort study of African Americans. METHODS: Self-reported data on snoring and risk factors for CVD were collected at baseline (2000-2004). Participants were followed prospectively for the development of incident CVD. Habitual snoring was defined as present if the participants reported it as "often" or "almost always" or absent if reported as "sometimes," "never," or "seldom." A CVD event included stroke, myocardial infarction, coronary revascularization procedure, or fatal CHD event. Cox proportional hazards models assessed the independent association between self-reported habitual snoring and incident CVD event adjusting for multiple covariates, including age, sex, hypertension, body mass index, diabetes, hypercholesterolemia, and smoking status. RESULTS: The snorer group consisted of 787 participants (mean age 52.1 years) and the nonsnorer group consisted of 3708 participants (mean age 54.9 years). Frequency of incident CVD events in the snorer group was not significantly different from the nonsnorer group. The fully adjusted hazard ratio for a CVD event in the snorer group was 1.01 (95% confidence interval [0.69, 1.47], p value of 0.96). CONCLUSION: In conclusion, self-reported habitual snoring was not associated with incident CVD among this large African American cohort. Future studies providing objective data on snoring and sleep apnea may provide more information on the snoring-CVD association among African Americans. TRIAL REGISTRATION: Identification Number: NCT00005485.

Multiethnic Meta-Analysis Identifies RAI1 as a Possible Obstructive Sleep Apnea-related Quantitative Trait Locus in Men.

Obstructive sleep apnea (OSA) is a common heritable disorder displaying marked sexual dimorphism in disease prevalence and progression. Previous genetic association studies have identified a few genetic loci associated with OSA and related quantitative traits, but they have only focused on single ethnic groups, and a large proportion of the heritability remains unexplained. The apnea-hypopnea index (AHI) is a commonly used quantitative measure characterizing OSA severity. Because OSA differs by sex, and the pathophysiology of obstructive events differ in rapid eye movement (REM) and non-REM (NREM) sleep, we hypothesized that additional genetic association signals would be identified by analyzing the NREM/REM-specific AHI and by conducting sex-specific analyses in multiethnic samples. We performed genome-wide association tests for up to 19,733 participants of African, Asian, European, and Hispanic/Latino American ancestry in 7 studies. We identified rs12936587 on chromosome 17 as a possible quantitative trait locus for NREM AHI in men (N = 6,737; P = 1.7 × 10-8) but not in women (P = 0.77). The association with NREM AHI was replicated in a physiological research study (N = 67; P = 0.047). This locus overlapping the RAI1 gene and encompassing genes PEMT1, SREBF1, and RASD1 was previously reported to be associated with coronary artery disease, lipid metabolism, and implicated in Potocki-Lupski syndrome and Smith-Magenis syndrome, which are characterized by abnormal sleep phenotypes. We also identified gene-by-sex interactions in suggestive association regions, suggesting that genetic variants for AHI appear to vary by sex, consistent with the clinical observations of strong sexual dimorphism.

Assessment of atherosclerotic plaque activity in patients with sleep apnea using hybrid positron emission tomography/magnetic resonance imaging (PET/MRI): a feasibility study.

PURPOSE: Evidence suggests that the inflammatory state of an atherosclerotic plaque is important in predicting future risk of plaque rupture. This study aims to investigate the feasibility of measuring plaque inflammation in patients with obstructive sleep apnea (OSA) utilizing advanced vascular imaging - hybrid positron-emission tomography/magnetic resonance imaging (PET/MRI) with fluorodeoxyglucose (FDG) tracer-before and after continuous positive airway pressure (CPAP). METHODS: Patients with newly diagnosed moderate to severe OSA underwent baseline PET/MRI for assessment of vascular inflammation of the carotid arteries and thoracic aorta prior to initiation of CPAP. Those adherent to CPAP returned for repeat imaging after 3-6 months of CPAP use. Atherosclerotic plaque activity, as measured by arterial wall FDG uptake, was calculated using target-to-background ratios (TBR) before and after CPAP. RESULTS: Five patients were recruited as part of a focused project. Mean age was 52 years (80% male), and mean apnea-hypopnea index (AHI) was 33. Three patients were objectively adherent with CPAP. In the pre-CPAP phase, all patients had focal FDG uptake in the carotid arteries and aorta. After CPAP, there was an average reduction in TBR of 5.5% (TBRmean) and 6.2% (TBRmax) in carotid and aortic plaque inflammation, similar in magnitude to the reduction observed with statin therapy alone in non-OSA patients (previously reported by others). CONCLUSIONS: We demonstrate the feasibility of using hybrid PET/MRI to assess atherosclerotic plaque inflammation in patients with OSA before and after CPAP. Use of the vascular PET/MRI platform in patients with OSA may provide better insight into the role of OSA and its treatment in reducing atherosclerotic inflammation.

Genetic Associations with Obstructive Sleep Apnea Traits in Hispanic/Latino Americans.

RATIONALE: Obstructive sleep apnea is a common disorder associated with increased risk for cardiovascular disease, diabetes, and premature mortality. Although there is strong clinical and epidemiologic evidence supporting the importance of genetic factors in influencing obstructive sleep apnea, its genetic basis is still largely unknown. Prior genetic studies focused on traits defined using the apnea-hypopnea index, which contains limited information on potentially important genetically determined physiologic factors, such as propensity for hypoxemia and respiratory arousability. OBJECTIVES: To define novel obstructive sleep apnea genetic risk loci for obstructive sleep apnea, we conducted genome-wide association studies of quantitative traits in Hispanic/Latino Americans from three cohorts. METHODS: Genome-wide data from as many as 12,558 participants in the Hispanic Community Health Study/Study of Latinos, Multi-Ethnic Study of Atherosclerosis, and Starr County Health Studies population-based cohorts were metaanalyzed for association with the apnea-hypopnea index, average oxygen saturation during sleep, and average respiratory event duration. MEASUREMENTS AND MAIN RESULTS: Two novel loci were identified at genome-level significance (rs11691765, GPR83, P = 1.90 × 10-8 for the apnea-hypopnea index, and rs35424364; C6ORF183/CCDC162P, P = 4.88 × 10-8 for respiratory event duration) and seven additional loci were identified with suggestive significance (P < 5 × 10-7). Secondary sex-stratified analyses also identified one significant and several suggestive associations. Multiple loci overlapped genes with biologic plausibility. CONCLUSIONS: These are the first genome-level significant findings reported for obstructive sleep apnea-related physiologic traits in any population. These findings identify novel associations in inflammatory, hypoxia signaling, and sleep pathways.

Comparison of Self-Reported Sleep Duration With Actigraphy: Results From the Hispanic Community Health Study/Study of Latinos Sueño Ancillary Study.

Most studies of sleep and health outcomes rely on self-reported sleep duration, although correlation with objective measures is poor. In this study, we defined sociodemographic and sleep characteristics associated with misreporting and assessed whether accounting for these factors better explains variation in objective sleep duration among 2,086 participants in the Hispanic Community Health Study/Study of Latinos who completed more than 5 nights of wrist actigraphy and reported habitual bed/wake times from 2010 to 2013. Using linear regression, we examined self-report as a predictor of actigraphy-assessed sleep duration. Mean amount of time spent asleep was 7.85 (standard deviation, 1.12) hours by self-report and 6.74 (standard deviation, 1.02) hours by actigraphy; correlation between them was 0.43. For each additional hour of self-reported sleep, actigraphy time spent asleep increased by 20 minutes (95% confidence interval: 19, 22). Correlations between self-reported and actigraphy-assessed time spent asleep were lower with male sex, younger age, sleep efficiency <85%, and night-to-night variability in sleep duration ≥1.5 hours. Adding sociodemographic and sleep factors to self-reports increased the proportion of variance explained in actigraphy-assessed sleep slightly (18%-32%). In this large validation study including Hispanics/Latinos, we demonstrated a moderate correlation between self-reported and actigraphy-assessed time spent asleep. The performance of self-reports varied by demographic and sleep measures but not by Hispanic subgroup.

Sleep apnea is independently associated with peripheral arterial disease in the Hispanic Community Health Study/Study of Latinos.

OBJECTIVE: Sleep apnea (SA) has been linked with various forms of cardiovascular disease, but little is known about its association with peripheral artery disease (PAD) measured using the ankle-brachial index. This relationship was evaluated in the Hispanic Community Health Study/Study of Latinos. APPROACH AND RESULTS: We studied 8367 Hispanic Community Health Study/Study of Latinos participants who were 45 to 74 years of age. Sleep symptoms were examined with the self-reported Sleep Health Questionnaire. SA was assessed using an in-home sleep study. Systolic blood pressure was measured in all extremities to compute the ankle-brachial index. PAD was defined as ankle-brachial index <0.90 in either leg. Multivariable logistic regression was used to investigate the association between moderate-to-severe SA, defined as apnea-hypopnea index ≥15, and the presence of PAD. Analyses were adjusted for covariates. The prevalence of PAD was 4.7% (n=390). The mean apnea-hypopnea index was significantly higher among adults with PAD compared with those without (11.1 versus 8.6 events/h; P=0.046). After adjusting for covariates, moderate-to-severe SA was associated with a 70% increase in the odds of PAD (odds ratio, 1.7; 95% confidence interval, 1.1-2.5; P=0.0152). This association was not modified by sex (P=0.8739). However, there was evidence that the association between moderate-to-severe SA and PAD varied by Hispanic/Latino background (P<0.01). Specifically, the odds were stronger in Mexican (adjusted odds ratio, 2.9; 95% confidence interval, 1.3-6.2) and in Puerto Rican Americans (adjusted odds ratio, 2.0; 95% confidence interval, 0.97-4.2) than in other backgrounds. CONCLUSIONS: Moderate-to-severe SA is associated with higher odds of PAD in Hispanic/Latino adults.

Sleep-disordered breathing in Hispanic/Latino individuals of diverse backgrounds. The Hispanic Community Health Study/Study of Latinos.

RATIONALE: Hispanic/Latino populations have a high prevalence of cardiovascular risk factors and may be at risk for sleep-disordered breathing (SDB). An understanding of SDB among these populations is needed given evidence that SDB increases cardiovascular risk. OBJECTIVES: To quantify SDB prevalence in the U.S. Hispanic/Latino population and its association with symptoms, risk factors, diabetes, and hypertension; and to explore variation by sex and Hispanic/Latino background. METHODS: Cross-sectional analysis from the baseline examination of the Hispanic Community Health Study/Study of Latinos. MEASUREMENTS AND MAIN RESULTS: The apnea-hypopnea index (AHI) was derived from standardized sleep tests; diabetes and hypertension were based on measurement and history. The sample of 14,440 individuals had an age-adjusted prevalence of minimal SDB (AHI ≥ 5), moderate SDB (AHI ≥ 15), and severe SDB (AHI ≥ 30) of 25.8, 9.8, and 3.9%, respectively. Only 1.3% of participants reported a sleep apnea diagnosis. Moderate SDB was associated with being male (adjusted odds ratio, 2.7; 95% confidence interval, 2.3-3.1), obese (16.8; 11.6-24.4), and older. SDB was associated with an increased adjusted odds of impaired glucose tolerance (1.7; 1.3-2.1), diabetes (2.3; 1.8-2.9), and hypertension. The association with hypertension varied across background groups with the strongest associations among individuals of Puerto Rican and Central American background. CONCLUSIONS: SDB is prevalent in U.S. Latinos but rarely associated with a clinical diagnosis. Associations with diabetes and hypertension suggest a large burden of disease may be attributed to untreated SDB, supporting the development and evaluation of culturally relevant detection and treatment approaches.

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.

Obstructive sleep apnea alters microRNA levels: Effects of continuous positive airway pressure.

Background: Obstructive sleep apnea (OSA) has been linked to cytokine-mediated chronic inflammatory states. Continuous positive airway pressure (CPAP) is an established therapy for OSA, but its effects on inflammation remain unclear. A recent study from our group identified soluble cytokine receptors altered in OSA patients and modified by CPAP adherence. However, the upstream regulatory pathways responsible for these shifts in proinflammatory cascades with OSA and CPAP therapy remained unknown. Accordingly, this study mapped OSA and CPAP-modulated soluble cytokine receptors to specific microRNAs and then tested the hypothesis that OSA and CPAP adherence shift cytokine-related microRNA expression profiles. Study Design: Plasma samples were collected from patients with OSA (n=50) at baseline and approximately 90 days after CPAP initiation and compared to referent control subjects (n=10). Patients with OSA were further divided into cohorts defined by adherence vs nonadherence to CPAP therapy. The microRNAs that mapped to soluble cytokine receptors of interest were subjected to quantitative polymerase chain reaction. Results: At baseline, increased hsa-miR-15a-5p, hsa-miR-15b-5p, hsa-miR-16-5p, hsa-miR-195-5p, hsa-miR-424-5p, hsa-miR-223-3p, and hsa-miR-223-5p were observed in patients with OSA compared to controls (p<0.05). In CPAP adherent patients (n=22), hsa-miR233-3p and hsa-miR233-5p decreased at follow-up (p<0.05) whereas there was no change in miR levels from baseline in non-adherent CPAP patients (n=28). The miRs hsa-miR233-3p and hsa-miR233-5p mapped to both proinflammatory and innate immunity activation; the inflammasome. Conclusion: A specific set of microRNAs, including hsa-miR233-3p and hsa-miR233-5p, may serve as a marker of inflammatory responses in patients with OSA, and be used to assess attenuation of inflammasome activation by CPAP.

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.

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.