The software embedded in recent continuous positive pressure devices enables the analysis of the level of leakage that occurs during treatment around the mask or from the mouth. However, there is no clear threshold defining when it would be appropriate to implement means for correcting these leaks. In addition, the calculation methods used to detect/measure leaks and the terminology used vary according to the manufacturers of continuous pressure devices, which also complicates the clinical interpretation of the leakage reports provided by the various devices.

In daily practice, when a patient complains of unintentional leakage, technicians adjust the mask, replace the nasal-mask by an oro-nasal mask or install a chinstrap. Nevertheless, those strategies are not always effective in reducing leakage or patient's complaints. A better understanding of the mechanisms beyond leak appearance for a given patient could lead to individualised leak-treatment solutions. Pressure level, mandibular behaviour, respiratory effort, sleep position or sleep stages may contribute to leak emergence.

A recent exploratory study has proposed a genuine analysis method of determining factors of unintentional-leaks using polysomnographic recordings in OSA patients treated with automatic Continuous Positive Airway Pressure (auto-CPAP). Based on the same methodology, the NOMICS company (Liège, Belgium) has developed a polygraphic device (Brizzy®) recording unintentional leakage during auto-CPAP treatment and providing an etiological analysis of unintentional-leaks via a software named APIOS.

Due to the innovative nature of the analysis, it appears necessary to validate it onto a population of OSA patients treated with auto-CPAP. Also, to our knowledge, there is no descriptive data of the distribution of the determinants of unintentional leaks in such a population. From this perspective, the investigators aim to retrospectively analyse some polygraphic recordings that have been performed with the Brizzy® on OSA patients treated with auto-CPAP.

The primary objective of this study is to identify and describe the determinants of unintentional leaks for a population of patients with OSA and treated with an auto-CPAP device

The secondary objectives are:

1. to evaluate the concordance of the determinants of leaks according to different leak thresholds (0 l/min ; 5l/min ; 10l/min : 20l/min).
2. to describe any technical problems that arise in relation to the analysis of leaks in the study population.

Inclusion Criteria:

• Sleep recordings from patients with OSA treated via auto-CPAP and monitored with Brizzy® device (including measures of non-intentional leaks) between June 1st and December 31th, 2017.

Exclusion Criteria:

• The patient opposes the usage of his/her data
• Minors or adults under any kind of guardianship
• Fixed pressure
• Bilevel positive airway pressure devices
• Adaptive servo-ventilation

The study population corresponds to patients (all meeting eligibility criteria) with obstructive sleep apnea syndrome treated via continuous positive pressure and monitored according to usual practice with the latest Brizzy device.