Two-phase interventional non-inferiority trial (phase 1: at least 8 weeks of CPAP; phase 2: 2 weeks of CPAP withdrawal) including 32 patients with moderate to severe OSA to compare the physiological consequences of a short-term CPAP withdrawal to the changes in previously untreated OSA. The trial has been designed as a validation of the CPAP-withdrawal model. Baseline in-laboratory sleep studies will be performed prior to CPAP initiation, after 6-8 weeks on CPAP (treatment effect) and following 2 weeks of CPAP therapy withdrawal (withdrawal effect, effect of OSA recurrence).

Background and Aim: Obstructive sleep apnoea (OSA) is a prevalent sleep-related breathing disorder. OSA is associated with daytime sleepiness, impaired quality of life, and adverse cardiovascular outcome. OSA can be effectively treated with nocturnal continuous positive airway pressure (CPAP) that restores normal breathing during sleep by preventing upper airway collapse. However, many patients in whom treatment of OSA is indicated are not able to adhere to CPAP, and the consequences of untreated OSA seem to differ between OSA phenotypes. Therefore, treatment alternatives to CPAP and a better understanding of treatment implications in different phenotypes of OSA are important targets in this research field. Thus, effective and controlled study designs are needed to address these research questions. Conventional randomised controlled CPAP trials are limited by low CPAP adherence resulting in an underestimation of the treatment effect. In addition, the impracticability of withholding an effective treatment from symptomatic patients resulting in inclusion of less symptomatic and patients with less severe OSA has limited the conclusions from previous trials. The CPAP therapy withdrawal model was introduced to allow studying the pathophysiological consequences of OSA and treatment effects of CPAP in an effective and controlled way. Optimally treated patients previously diagnosed with OSA established on CPAP therapy are included and undergo a short-term therapy interruption. A therapy withdrawal results in recurrence of OSA and pathophysiological changes can be attributed to OSA in this controlled model. However, it is uncertain whether the consequences of OSA during a short-term CPAP therapy withdrawal are comparable to long-term consequences of untreated OSA.

The aim of this two-phase interventional trial is to assess the effects of intermediate and long-term CPAP therapy as well as acute CPAP therapy withdrawal resulting in OSA recurrence compared to longstanding untreated OSA on cardiovascular and symptom outcomes in moderate and severe OSA. The objective is to study the comparability of conventional CPAP therapy trials with the CPAP withdrawal model in terms of evaluating relevant cardiovascular and patient-centred outcomes in OSA.

The investigators hypothesise that the treatment effect of CPAP in previously treatment-naïve patients with OSA - if adherent to CPAP - will support the same conclusions on OSA pathophysiology and therapy effects as a short-term CPAP withdrawal in treatment-adherent patients having a recurrence of OSA in response to CPAP withdrawal and thus support the use of this effective CPAP withdrawal model to use as future study design in interventional trials in OSA. It is hypothesised that the CPAP withdrawal is non-inferior to conventional CPAP trials in studying the effect on blood pressure and OSA-symptoms.

Inclusion Criteria:

• Age ≥ 18 years
• Untreated moderate or severe OSA with an apnoea-hypopnoea-index (AHI) of ≥ 20/h and an oxygen desaturation index (>3%-dips) of ≥ 20/h in an in-laboratory sleep study
• Epworth Sleepiness Scale Score ≥7/24 points
• Indication for CPAP therapy
• Written informed consent

Exclusion Criteria:

• Moderate or more severe obstructive or restrictive ventilatory disorder (FEV1 or FVC < 70%)
• Heart failure
• Other types of sleep-related breathing disorders, e.g. sleep-associated hypoventilation, central sleep apnoea, Cheyne Stoke's Breathing
• Hypoxic or hypercapnic respiratory failure (awake paO2 < 9 kPa or paCO2 > 6 kPa)
• Active treatment for OSA (CPAP, mandibular advancement device, hypoglossal nerve stimulation)
• Current professional driver or previous sleepiness-related accidents
• Pregnancy