Non invasive ventilation (NIV) is generally delivered by pneumatically triggered and cycled-off Pressure Support (PSP) through a facial mask. Compared to PSP, Neurally Adjusted Ventilatory Assist (NAVA), which is the only ventilatory mode using a non-pneumatic signal, i.e., diaphragm electrical activity (EAdi), to trigger and drive ventilator assistance, improves patient-ventilator interaction. A specific setting to generate neurally controlled Pressure Support (PSN) was recently proposed for delivering NIV by helmet. The investigators here compare PSN with PSP and NAVA during NIV by facial mask, with respect to arterial blood gases (ABGs), patient comfort, and patient-ventilator interaction and synchrony.

Three 30-minute trials of NIV were randomly delivered to 14 patients immediately after extubation to prevent post-extubation respiratory failure: 1) PSP, with an inspiratory support ≥8 cmH2O; 2) NAVA, adjusting the NAVA level to achieve a comparable peak EAdi (EAdipeak) as during PSP; 3) PSN, setting the NAVA level at 15 cmH2O/mcV with an upper airway pressure (Paw) limit such to obtain the same overall Paw applied during PSP. We assessed EAdipeak, ABGs, peak inspiratory flow (PIF), time to reach PIF (PIFtime), pressure-time product of the first 300 (PTP300-index) and 500 (PTP500-index) milliseconds after initiation of patient effort, patient comfort, inspiratory trigger delay (DelayTR-insp), and the rate of asynchrony, as assessed by the Asynchrony Index (AI%).

Non Invasive Ventilation (NIV) is increasingly used for treatment of Acute Respiratory Failure (ARF) and is commonly applied through an oral-nasal mask by means of pneumatically triggered and cycled-off Pressure Support (PSP). Although better tolerated than invasive mechanical ventilation, NIV is characterized by drawbacks such as poor patient-ventilator interaction and discomfort, which are major determinants of NIV failure.

In particular, the pneumatic signals, i.e., flow, volume and airway pressure (Paw), are leak sensitive and frequently cause patient-ventilator asynchrony. The only mode that does not use pneumatic signals to trigger and drive the ventilator is Neurally Adjusted Ventilator Assist (NAVA). In fact, with NAVA the ventilator assistance is under control of diaphragm electrical activity (EAdi), as assessed through an esophageal catheter. As opposed to PSP, NAVA has been repeatedly shown to improve patient-ventilator interaction and reduce asynchronies, both during invasive ventilation and NIV. However, NAVA is characterized by a lower rate of pressurization than PSP.

Very recently, a specific NIV setting to generate a neurally, i.e., EAdi, controlled Pressure Support (PSN) has been proposed and applied during both invasive ventilation and NIV delivered via helmet. PSN consists in increasing the user-controlled gain factor (NAVA level) at maximum, while restraining inspiratory Paw by adjusting on the ventilator the upper pressure limit.

During NIV delivered through helmets, PSN has been shown, compared to both PSP and NAVA, to result in better pressurization and triggering performance, which improve patient's comfort while reducing EAdi, without affecting respiratory rate and arterial blood gases (ABGs). Due to the different characteristics of helmets and masks, it is unclear whether these advantages could be extended to NIV delivered by mask. The investigators therefore designed this physiological study aimed at comparing PSN with PSP and NAVA, with respect to breathing pattern, respiratory drive, ABGs, pressurization and triggering performance, patient's comfort and patient-ventilator synchrony.

After patient enrollment, a nasal-gastric feeding tube able to detect EAdi (EAdi catheter, Maquet Critical Care, Solna, Sweden) was placed and the correct positioning ascertained. The study was performed using a Servo-I ventilator (Maquet Critical Care, Solna, Sweden) equipped with a NIV software for air-leaks. The oral-nasal mask was individually selected for each patients based on the anthropometric characteristics and in order to minimize air leaks and optimize patient tolerance, among three different models, FreeMotion RT041 Non Vented Full Face Mask (Fisher and Paykel, Auckland, New Zealand), Ultra Mirage FFM-NV (ResMed, San Diego, CA, USA) and PerforMax Face Mask (Philips Respironics, Murrysville, PA, USA).

All patients underwent three 30-minute trials in random order: 1) PSP, setting the inspiratory pressure support ≥8 cmH2O to obtain a tidal volume of 6-8 mL•kg-1 of body weight, the fastest rate of pressurization (0.0 sec) and I/E cycling at 35% of peak inspiratory flow (PIF); 2) NAVA, adjusting the NAVA level in order to achieve a comparable peak EAdi (EAdipeak) as during PSP with a safety Paw upper limit of 30 cmH2O; 3) PSN, setting the NAVA level at its maximum (i.e; 15 cmH2O/mcV), and an upper Paw limit such to obtain the same overall Paw applied during PSP. During both NAVA and PSN the trigger sensitivity was set at 0.5 mcV while, the default cycling-off is 70% EAdipeak, as fixed by the manufacturer. PEEP was set by the attending physicians in a range between 5 and 10 cmH2O, and left unmodified throughout the whole study period. The inspiratory oxygen fraction (FiO2) was regulated to obtain peripheral oxygen saturation (SpO2) between 94% and 96%, before starting the protocol, and kept unmodified throughout the study period.

The three modes of ventilation were applied according to a computer-generated random sequence using sealed, opaque numbered envelops. The envelopes were kept in the head of nurses' office in both institutions. The envelope was opened by the nurse in charge of the patient, and the sequence of modes to be applied communicated to the investigators.

Predefined criteria for protocol interruption were: 1) need for emergency re-intubation; 2) SpO2 <90%, 3) acute respiratory acidosis, as defined by PaCO2 >50 mmHg and pH <7.30; 3) inability to expectorate secretions; 4) hemodynamic instability (i.e.; need for continuous infusion of dopamine or dobutamine >5 µg∙kg-1∙min-1, norepinephrine >0.1 µg∙kg-1∙min-1 or epinephrine or vasopressin at any dosage to maintain mean arterial blood pressure >60 mmHg); 5) life-threatening arrhythmias or electrocardiographic signs of ischemia; or 6) loss of 2 or more points of GCS.

The investigators considered eligible:

• any patient ≥18 years admitted in ICU undergoing invasive mechanical ventilation for at least 48 hours with a Glasgow Coma Scale (GCS) of 11 (i.e.; spontaneous eyes opening, obeys to command, no verbal response because of the endotracheal tube in place),
• no infusion of midazolam and propofol in the previous 24 and 4 hours, respectively
• ready for extubation with indication, prior to extubation, to receive prophylactic NIV to prevent post-extubation respiratory failure.

Patients were excluded in case of

• need for analgesic or sedative drugs,
• recent cervical spine injury,
• obstructive sleep apnoea syndrome,
• pregnancy,
• contraindications to placement of a nasal-gastric feeding tube,
• inclusion in other research protocol
• lack of consent.