Long-term effect of obstructive sleep apnoea management on blood pressure in patients with resistant hypertension: the SARAH study.

BACKGROUND: There is a close relationship between obstructive sleep apnoea (OSA) and resistant hypertension (RH). However, studies assessing the long-term effect of diagnosing and treating OSA on blood pressure (BP) control in these patients are lacking. METHODS: To address this gap, we recruited 478 RH patients from hypertension units and followed them prospectively after they were screened for OSA through a sleep study. By performing 24-h ambulatory BP monitoring (ABPM) annually, the effect of OSA management was assessed. RESULTS: The patients had a median (interquartile range (IQR)) age of 64.0 (57.2-69.0) years, 67% were males and most were nonsleepy, with a median (IQR) apnoea-hypopnoea index (AHI) of 15.8 (7.9-30.7) events·h-1. The median (IQR) follow-up time was 3.01 (2.93-3.12) years. At baseline, severe OSA was associated with uncontrolled BP, nocturnal hypertension and a nondipper circadian BP pattern. Moreover, these patients had higher BP values during follow-up than did patients in the other groups. However, among patients with moderate and severe OSA, the management of sleep disordered breathing, including the implementation of continuous positive airway pressure treatment, was associated with a reduction in 24-h ABPM parameters, especially night-time BP values, at the 1-year follow-up. These benefits were attenuated over time and only subjects with severe OSA maintained an ABPM night-time reduction at 3 years. Furthermore, clinical variables such as uncontrolled BP, sex and age showed a predictive value for the BP response at 1 year of follow-up. CONCLUSION: A favourable long-term decrease in BP was detected by diagnosing and treating OSA in a cohort of RH patients from hypertension units, but over time this decrease was only partially maintained in severe OSA patients.

Sleep and Circadian Health of Critical Survivors: A 12-Month Follow-Up Study.

OBJECTIVES: To investigate the sleep and circadian health of critical survivors 12 months after hospital discharge and to evaluate a possible effect of the severity of the disease within this context. DESIGN: Observational, prospective study. SETTING: Single-center study. PATIENTS: Two hundred sixty patients admitted to the ICU due to severe acute respiratory syndrome coronavirus 2 infection. INTERVENTIONS: None. MEASUREMENTS AND MAIN RESULTS: The cohort was composed of 260 patients (69.2% males), with a median (quartile 1-quartile 3) age of 61.5 years (52.0-67.0 yr). The median length of ICU stay was 11.0 days (6.00-21.8 d), where 56.2% of the patients required invasive mechanical ventilation (IMV). The Pittsburgh Sleep Quality Index (PSQI) revealed that 43.1% of the cohort presented poor sleep quality 12 months after hospital discharge. Actigraphy data indicated an influence of the disease severity on the fragmentation of the circadian rest-activity rhythm at the 3- and 6-month follow-ups, which was no longer significant in the long term. Still, the length of the ICU stay and the duration of IMV predicted a higher fragmentation of the rhythm at the 12-month follow-up with effect sizes (95% CI) of 0.248 (0.078-0.418) and 0.182 (0.005-0.359), respectively. Relevant associations between the PSQI and the Hospital Anxiety and Depression Scale (rho = 0.55, anxiety; rho = 0.5, depression) as well as between the fragmentation of the rhythm and the diffusing lung capacity for carbon monoxide (rho = -0.35) were observed at this time point. CONCLUSIONS: Our findings reveal a great prevalence of critical survivors presenting poor sleep quality 12 months after hospital discharge. Actigraphy data indicated the persistence of circadian alterations and a possible impact of the disease severity on the fragmentation of the circadian rest-activity rhythm, which was attenuated at the 12-month follow-up. This altogether highlights the relevance of considering the sleep and circadian health of critical survivors in the long term.

Predicting Alzheimer's disease CSF core biomarkers: a multimodal Machine Learning approach.

Introduction: Alzheimer's disease (AD) is a progressive neurodegenerative disorder. Current core cerebrospinal fluid (CSF) AD biomarkers, widely employed for diagnosis, require a lumbar puncture to be performed, making them impractical as screening tools. Considering the role of sleep disturbances in AD, recent research suggests quantitative sleep electroencephalography features as potential non-invasive biomarkers of AD pathology. However, quantitative analysis of comprehensive polysomnography (PSG) signals remains relatively understudied. PSG is a non-invasive test enabling qualitative and quantitative analysis of a wide range of parameters, offering additional insights alongside other biomarkers. Machine Learning (ML) gained interest for its ability to discern intricate patterns within complex datasets, offering promise in AD neuropathology detection. Therefore, this study aims to evaluate the effectiveness of a multimodal ML approach in predicting core AD CSF biomarkers. Methods: Mild-moderate AD patients were prospectively recruited for PSG, followed by testing of CSF and blood samples for biomarkers. PSG signals underwent preprocessing to extract non-linear, time domain and frequency domain statistics quantitative features. Multiple ML algorithms were trained using four subsets of input features: clinical variables (CLINVAR), conventional PSG parameters (SLEEPVAR), quantitative PSG signal features (PSGVAR) and a combination of all subsets (ALL). Cross-validation techniques were employed to evaluate model performance and ensure generalizability. Regression models were developed to determine the most effective variable combinations for explaining variance in the biomarkers. Results: On 49 subjects, Gradient Boosting Regressors achieved the best results in estimating biomarkers levels, using different loss functions for each biomarker: least absolute deviation (LAD) for the Aß42, least squares (LS) for p-tau and Huber for t-tau. The ALL subset demonstrated the lowest training errors for all three biomarkers, albeit with varying test performance. Specifically, the SLEEPVAR subset yielded the best test performance in predicting Aß42, while the ALL subset most accurately predicted p-tau and t-tau due to the lowest test errors. Conclusions: Multimodal ML can help predict the outcome of CSF biomarkers in early AD by utilizing non-invasive and economically feasible variables. The integration of computational models into medical practice offers a promising tool for the screening of patients at risk of AD, potentially guiding clinical decisions.

Metabolipidomic Analysis in Patients with Obstructive Sleep Apnea Discloses a Circulating Metabotype of Non-Dipping Blood Pressure.

A non-dipping blood pressure (BP) pattern, which is frequently present in patients with obstructive sleep apnea (OSA), confers high cardiovascular risk. The mechanisms connecting these two conditions remain unclear. In the present study we performed a comprehensive analysis of the blood metabolipidome that aims to provide new insights into the molecular link between OSA and the dysregulation of circadian BP rhythmicity. This was an observational prospective longitudinal study involving adults with suspected OSA who were subjected to full polysomnography (PSG). Patients with an apnea-hypopnea index ≥ 5 events/h were included. Fasting plasma samples were obtained the morning after PSG. Based on the dipping ratio (DR; ratio of night/day BP values) measured via 24 h ambulatory BP monitoring, two groups were established: dippers (DR ≤ 0.9) and non-dippers (DR > 0.9). Treatment recommendations for OSA followed the clinical guidelines. Untargeted metabolomic and lipidomic analyses were performed in plasma samples via liquid chromatography-tandem mass spectrometry. Non-dipper patients represented 53.7% of the cohort (88/164 patients). A set of 31 metabolic species and 13 lipidic species were differentially detected between OSA patients who present a physiologic nocturnal BP decrease and those with abnormal BP dipping. Among the 44 differentially abundant plasma compounds, 25 were putatively identified, notably glycerophospholipids, glycolipids, sterols, and fatty acid derivates. Multivariate analysis defined a specific metabotype of non-dipping BP, which showed a significant dose-response relationship with PSG parameters of OSA severity, and with BP dipping changes after 6 months of OSA treatment with continuous positive airway pressure (CPAP). Bioinformatic analyses revealed that the identified metabolipidomic profile was found to be implicated in multiple systemic biological pathways, with potential physiopathologic implications for the circadian control of BP among individuals with OSA.