• Number of respiratory events during non invasive ventilation [ Time Frame: 2 nights ]
  According to consensus group, nocturnal respiratory events are : unintentional leak, obstruction at different levels of the upper airways and asynchrony (ineffective effort, auto-triggering, and double triggering, premature or delayed cycling)
• Sleep quality [ Time Frame: 2 nights ]
  Changes in data of polysomnography (total sleep time (TST) ; total recording time ; sleep latency, sleep efficiency (TST/time in bed) ; time in each stages plus percent of TST in each stage ; stage REM latency ; wake time after sleep onset ; arousal index)
• Ventilation quality [ Time Frame: 2 nights ]
  Subjective changes in quality of ventilator by specific auto questionnaire (quality of night, feeling, dyspnea, quality of ventilation, respiratory disturbance caused by NIV, sleep disturbance caused by NIV). The patient answers to each question using a "likert scale" in terms of intensity. The second questionnaire is the S3-NIV which is a tool for the routine clinical assessment of patients undergoing home NIV.

• Number of respiratory events during non invasive ventilation [ Time Frame: 2 nights ]
  According to consensus group, nocturnal respiratory events are : unintentional leak, obstruction at different levels of the upper airways and asynchrony (ineffective effort, auto-triggering, and double triggering, premature or delayed cycling)
• Sleep quality [ Time Frame: 2 nights ]
  Changes in data of polysomnography
• Ventilation quality [ Time Frame: 2 nights ]
  Subjective changes in quality of ventilator by specific auto questionnaire (quality of night, feeling, dyspnea, quality of ventilation, respiratory disturbance caused by NIV, sleep disturbance caused by NIV). The patient answers to each question using a "likert scale" in terms of intensity. The second questionnaire is the S3-NIV which is a tool for the routine clinical assessment of patients undergoing home NIV.

In COPD patients with chronic hypercapnic respiratory failure, the prognosis is poor and the treatment with non invasive ventilation is actually well established. However the best mode of ventilation is not well known.

In severe COPD patients various disorders of respiratory mechanics result in insufficient ventilation, which can be life-threatening or create NIV discomfort. The main characteristic of these disorders is a cyclical closing of small airways that can limit an expiratory flow and provoked some fluctuations in flow curve. To our knowledge, the management of dynamic hyperinflation seems to play an important role in explaining the effect of the NIV. Few studies have examined the effects of the machine's adjustments on dynamic hyperinflation.

The main objective of this study is to analyze the impact of specific ventilatory modes supposed to reduce the dynamic hyperinflation on the hematosis, by studying transcutaneous pressure of carbon dioxide, in severe hypercapnic COPD patients ventilated by NIV. Two modes of ventilation will be compared. First one is an algorithmic mode developed by the company Löwenstein (AirTrap Control, Trigger Lockout and the Expiratory Pressure Ramp). The second one is a standard algorithmic mode, used in the same ventilator. These two ventilatory modes will be evaluated in each patient, during two consecutive nights in current living conditions at home.

• Chronic Obstructive Pulmonary Disease
• Hypercapnia

• standard mode
  after adjusting pressure and different modes of ventilation to each individual patient. Each one of them will receive NIV using Löwenstein mode during one night at home. Patients will be under transcutaneous PCO2 measurement and polysomnographic surveillance.
  Intervention: Device: Non Invasive Ventilation
• Lowenstein mode
  after adjusting pressure and different modes of ventilation to each individual patient. Each one of them will receive NIV without Löwenstein mode during one night at home. Patients will be under transcutaneous PCO2 measurement and polysomnographic surveillance.
  Intervention: Device: Non Invasive Ventilation

Inclusion Criteria:

• Diagnosis of COPD using Gold 2017 criterion
• Optimal medical treatment of COPD, ie. inhaled steroids, long-acting β2-agonist, long-acting anticholinergic, according to GOLD 2017 guidelines.
• At least one acute exacerbation of COPD with hypercapnic respiratory failure
• Introduction of NIV distant to acute exacerbation of COPD with persisting hypercapnia (> 7kPa) on arterial blood gas at rest.
• NIV equipment from the service provider "Don Du Souffle (DDS)".
• Patient able to give their consent

Exclusion Criteria:

• Contraindication to NIV (consensus conference of SFAR-SPLF-SRLF)
• Obstructive sleep apnea with apnea/hypopnea index > to 20/h
• BMI > 35 kg/m2
• Low compliance with NIV treatment (<3h per day)
• Acute exacerbation COPD during inclusion or pH <7.30 in routine blood gas arterial
• Severe heart failure (New York Heart Association stage IV/IV)
• Unstable angina
• Severe arrhythmias
• Severe renal insufficiency stage > 4
• Other conditions than COPD resulting in hypercapnia
• Pregnancy
• Patient without health insurance
• Patient excluded by another study