This cross-over study will compare the asynchrony index between standard manual ventilator settings, optimized manual ventilator settings, and automated ventilator setting in intensive care patients ventilated in non-invasive ventilation with a high asynchrony index. The hypothesis is that both manual optimized ventilator settings and automated ventilator settings are associated with a lower patient-ventilator asynchrony index as compared to manual standard ventilator settings.

A randomized cross-over design method will be used. Patient requiring NIV with an asynchrony index over 35% will be included. An esophageal catheter with a balloon will be inserted to monitor esophageal pressure. Patients will be ventilated during 3 periods of 30 min, with 10 minutes of washout in between. Recordings of airway pressure, airway flow, and esophageal pressure will be analyzed by two investigators blinded of the trigger settings.

The primary outcome will be the asynchrony index. The secondary outcome will be the ineffective inspiratory effort index, autotrigering index, double triggering index, inspiratory trigger delay, cycling delay, total time spent in asynchrony, patient comfort, and blood gas results.

Non-invasive ventilation (NIV) is used in 35% of patient admitted in intensive care unit (ICU) with a failure rate of 10 to 70% depending on the indication and clinician experience. Patient-ventilator asynchrony is a frequent cause of NIV failure. Therefore, optimizing patient-ventilator synchronization is important for its comfort, tolerance, and efficacy. An optimal patient-ventilation is achieved when the mechanical breath provided by the ventilator match the patient inspiratory effort. The ratio between the number of asynchronies divided by the number of patient inspiratory effort define the asynchrony index (AI). AI over 10% is considered as severe and occurs in 30 to 43% of patients ventilated in NIV. Patient ventilator asynchronies occurs because ventilator settings of inspiratory and expiratory triggers remain constant in patient with variable respiratory drive, and unintentionnals leaks that are difficult to control in NIV. Thus using an automatic adjustment of inspiratory and expiratory triggers setting according to patient effort and unintentional leaks may decrease the number of patient-ventilator asynchronies. This cross-over study will compare the asynchrony index between standard manual ventilator settings, optimized manual ventilator settings, and automated ventilator setting in intensive care patients ventilated in non-invasive ventilation with a high asynchrony index. The hypothesis is that both manual optimized ventilator settings and automated ventilator settings are associated with a lower patient-ventilator asynchrony index as compared to manual standard ventilator settings.

A randomized cross-over design method will be used. Patient requiring NIV with an asynchrony index over 30% will be included. An esophageal catheter with a balloon will be inserted to monitor esophageal pressure. Patients will be ventilated during 3 periods of 30 min, with 10 minutes of washout in between. Recordings of airway pressure, airway flow, and esophageal pressure will be analyzed by two investigators blinded of the trigger settings.

The primary outcome will be the asynchrony index. The secondary outcome will be the ineffective inspiratory effort index, autotrigering index, double triggering index, inspiratory trigger delay, cycling delay, total time spent in asynchrony, patient comfort, and blood gas results.

The sample size was calculated from the total asynchrony index (primary outcome). Patients with an asynchrony index over 30% in using manual standard ventilator settings will be included. Considering an asynchrony index of 30 ± 15 % in manual standard ventilator settings with a clinically significant objective to reduce the asynchrony index to 15% in manual optimized ventilator settings and automated ventilator settings, a sample size of 30 patients is required with a risk at 0.05 and a power at 80%. Therefore, 35 patients are planned.

Inclusion Criteria:

• Patient aged over 18 years old
• Covered by social insurance
• Consent for study signed by patient or next-of-kin
• NIV session indicated for at least 2 hours
• Asynchrony index ≥ 30% with standard manual settings

Exclusion Criteria:

• Patient requiring continuous NIV
• Contra-indication to esophageal catheter insertion: gastric ulcer, esophageal varices, pharyngeal or laryngeal tumor.
• Patient with withholding decision about intubation
• Moribund patient
• Patient included in another interventional study in the last 30 days
• Patient that does not speak French
• Pregnant women