The Effect of Continuous Positive Airway Therapy on the Blood Pressure in Sleepy vs Non-sleepy Patients With Obstructive Sleep Apnea

Overview

Obstructive sleep apnoea (OSA) is one of the most common noncommunicable types of disease, it affects about 1 billion people across the world. Left untreated, it causes apnoeas and hypopnoeas to fragment sleep, with frequent arousal from sleep and intermittent hypoxia associated with increased work of breathing. Frequently, it leads to excessive daytime sleepiness, as measured subjectively by the Epworth Sleepiness Scale, or, objectively, by the multiple sleep latency test (MSLT) or the maintenance of wakefulness test (MWT). OSA can lead to sustained high sympathetic tone at night, which in the long-term may impact on the cardiovascular risk. The investigators hypothesised that any primary airway therapeutic effect on the cardiovascular system, as measured by the blood pressure, in patients with OSA will differ dependent on whether subjects are excessively sleepy, or remain so when treated. Hypothesis 1. Office blood pressure (SBP, DBP) responses to CPAP in patients with OSA who are excessively sleepy (ESS\>10) at baseline vs non-sleepy patients at baseline. 2. 24-hour BP data (SBP, DBP, dipping, nocturnal and daytime) in sleepy patients in response to CPAP vs non-sleepy patients. 3. Adherence to treatment in sleepy patients may be different to non-sleepy patients and the observed effect effects will be adjusted in a secondary analysis according to available adherence data and follow up time.

Introduction Obstructive sleep apnoea (OSA) is one of the most common noncommunicable types of disease, it affects about 1 billion people across the world. Left untreated, it causes apnoeas and hypopnoeas to fragment sleep, with frequent arousal from sleep and intermittent hypoxia associated with increased work of breathing. Frequently, it leads to excessive daytime sleepiness, as measured subjectively by the Epworth Sleepiness Scale, or, objectively, by the multiple sleep latency test (MSLT) or the maintenance of wakefulness test (MWT). OSA can lead to sustained high sympathetic tone at night, which in the long-term may impact on the cardiovascular risk. With optimised primary airway treatment, normal breathing at night resumes and patients experience a restoration of normal sleep architecture, with little or no further sleep fragmentation due to the underlying sleep disordered breathing. Daytime symptoms and quality of life typically improve with primary airway treatment, and long-term cardiovascular risks follows. Data from the ANDANTE collaboration have recently shown that the effect of primary airway patency treatment in patients with obstructive sleep apnoea on the blood pressure control differs according to whether patients with OSA have normal blood pressure, controlled or uncontrolled hypertension. Furthermore, the high prevalence of OSA in the general population, frequently diagnosed randomly (e.g., screening prior to elective surgery) in otherwise asymptomatic patients raises the question whether symptoms, like sleepiness, that are associated with the syndrome predict outcomes on the cardiovascular system. There are clinically relevant implications in who should and who should not be treated with OSA, particularly in the cohort of mild/moderate and asymptomatic patients with OSA, as recently discussed in the Baveno, and the modified Baveno classification. Furthermore, a proportion of patients with OSA who are adherent to primary airway therapy remain symptomatic, as measured by the Epworth Sleepiness Scale, a phenomenon that has been labelled residual excessive daytime sleepiness. The cardiovascular risk associated with patients who are excessively sleepy at baseline, and who remain so, is unclear. The investigators hypothesised that any primary airway therapeutic effect on the cardiovascular system, as measured by the blood pressure, in patients with OSA will differ dependent on whether subjects are excessively sleepy, or remain so when treated. Hypothesis 1. Office blood pressure (SBP, DBP) responses to CPAP in patients with OSA who are excessively sleepy (ESS\>10) at baseline vs non-sleepy patients at baseline. 2. 24-hour BP data (SBP, DBP, dipping, nocturnal and daytime) in sleepy patients in response to CPAP vs non-sleepy patients. 3. Adherence to treatment in sleepy patients may be different to non-sleepy patients and the observed effect effects will be adjusted in a secondary analysis according to available adherence data and follow up time. The investigators expected that the ANDANTE dataset could be used to address the clinically relevant question whether primary airway therapy of sleepy/non-sleepy patients with OSA provides benefits for associated long-term cardiovascular risks, as measured by the blood pressure values and 24h modulation (dipping vs non dipping). These questions form an essential aspect when discussing therapeutic options with patients in the sleep department, particularly in asymptomatic patients with mild/moderate OSA. Methods The purpose of this study was to retrospectively analyse the ANDANTE database using the individual patient data (IPD) from the recently published meta-analysis using the ANDANTE database, details of which have been published elsewhere (Pengo M et al, European Respiratory Journal 2025). The study was approved by the King's College London ethics committee (17/12/2025; MRA-25/26-53856) and, following peer-review, by the scientific committee of the ANDANTE collaboration (Application date: 17/12/2024). The investigators sought to investigate the dataset by stratifying patients based on excessive sleepiness, defined as an Epworth Sleepiness Scale (ESS) score greater than 10 points (a) at baseline, and (b) at follow up. Statistical Analysis Plan Data management This study will retrospectively analyse the ANDANTE database using individual patient data (IPD) from the previously published meta-analysis (Pengo M et al, European Respiratory Journal 2025). As a first step, we will stratify patients based on the Epworth Sleepiness Scale (ESS) both at baseline and at follow-up, considering a threshold of 10 points. From this stratification, we will create multiple study groups from which we will evaluate our primary and secondary outcomes: 1. Baseline excessively sleepy patients (ESS \>10). 2. Baseline non-sleepy patients (ESS ≤10). 1. Patients with residual sleepiness (maintained ESS\>10 at follow-up). 2. Patients with improved sleepiness (ESS reduced to ≤10). Blood pressure responses are described as office blood pressure measurement, 24h, diurnal / nocturnal SBP and DBP, and nocturnal dipping measurements at follow-up. Primary Outcome The effect of baseline sleepiness classification (ESS \> or ≤ 10) on the blood pressure response to CPAP treatment in patients with OSA (groups 1 vs 2). Secondary Outcomes I) The effect of residual vs improving sleepiness on the blood pressure response to CPAP treatment in patients with OSA (groups a vs b). II) Difference in adherence levels (measured by follow-up average daily use) between sleepy and non-sleepy patients at baseline and follow-up (groups 1 vs 2, and a vs b). III) Effect of adherence in treatment responses (groups 1 vs 2, and a vs b). Analytical Approach 1\. Linear mixed models, predicting follow-up blood pressure measurements (dependent variables) by the above-described groups (ESS classifications), considering baseline blood pressure measurements and CPAP adherence as covariate, fixed effects. 1. A set of models will be computed to assess effects of baseline classifications (groups 1 vs 2) and another set to assess differences at follow-up classifications (groups a vs b). 2. These models will account for several random effects in time to follow-up, treatment site, age, gender, body mass index and comorbidities. Statistical Reporting 1. Baseline and follow-up data distributions will be presented as mean (standard deviation) for normally distributed data and median (interquartile range) for non-normally distributed data. 2. Main results will be presented as the difference in follow-up blood pressure measurements between groups, accounting for baseline measurements, adherence and random effects. Adherence effects will be described as its coefficient of interaction with grouping in prediction. Differences will be described as 95% confidence intervals and significance defined as p\<0.05.