Study of the diaphragmatic function in patients with respiratory failure treated by standard oxygenotherapy, high-flow nasal oxygenotherapy, and noninvasive ventilation

Three different modes of non-invasive ventilation will be applied to the included patients: High-flow nasal oxygenotherapy (HFNO): HFNO will be performed using the Optiflow'Trademark' system (Fisher & Paykel Healthcare, Auckland, NZ) with gas flow = 50L/min and inspired FiO2 to target SpO2 between 93 and 97%. Because HFNO is our respiratory failure first-line treatment, baseline PaO2/FiO2 ratio and the first diaphragmatic ultrasound examination will be recorded with this respiratory support mode. Standard oxygenotherapy (SO): SO will be then delivered using a reservoir facemask with O2 flow between 9 and 15 L/min ( FiO2 approximatively 55 to 80%) to get equal SpO2 as baseline. To avoid potential recruitement effects of the NIV, and because patients will move from one mode to another, without washout period, SO diaphragmatic thickness fraction (DTF) measurements will be performed before BIPAP (second measurement). Noninvasive ventilation (NIV): NIV will be delivered as Bilevel positive airway pressure (BiPAP). BIPAP will be applied with BIPAP Vision or Respironics V60 ventilators (Philips Healthcare, Andover, MA, USA) according to availability, with an oronasal mask. Standard BIPAP settings will be: pressure support (PS) = 4 cmH2O, PEEP = 4 cmH2O (BIPAP 4/4), and inspired FiO2 equal to previously set FIO2 with HNFO. Recordings will be required during two extra BIPAP periods if DTF is above 10%, to assess the specific effects of PS and PEEP level (fourth measurement: PS =8 cmH2O and PEEP = 4 cmH2O [BIPAP 8/4]; fifth measurement: PS = 4 cmH2O and PEEP = 8 cmH2O [BIPAP 4/8]). The inspired FiO2 will be kept constant during the three BIPAP measurement periods. If DTF is inferior to 10% with BIPAP 4/4, we will consider increasing PS or PEEP futile, and will not record the 4th and 5th measures. Ultrasound examination: The diaphragmatic ultrasound examinations will be performed after at least 5 minutes of stable and constant breathing with each technique. Every measurement will last 5 to 10 minutes, giving a total of 10 to 15 minute in each ventilation mode. Ultrasound examinations will be performed by a single operator, experienced in diaphragmatic ultrasound examination. Available echographs were Vivid 6 (GE Healthcare, Little Chalfont, UK) or CX50 POC (Philips Healthcare, Andover, MA, USA), with high frequency probes. Measurements were performed in the semi-recumbent position. The right side was the first choice for ultrasound examination because of generally easier visualization compared to the left hemidiaphragm. However, the latter was chosen in case of poor visualization of the right hemidiaphragm (chest drains, surgical wound, pneumothorax, dressings), to limit the high risk of multiple protocol exclusions in our thoracic surgery patients. The diaphragm was located at its zone of apposition to the thorax, by placing the probe in the 9th intercostal space in the midaxillary line with the ultrasound beam in the cephalocaudal direction. The diaphragm will be identified between the two hyperechogenic parallel lines of the pleura and the peritoneum. Measurements were then performed in 2D mode according to the leading-edge method in order to minimize the ultrasound overestimation of the peritoneal membrane thickness. The measurements will include diaphragmatic end-inspiratory thickness (eIT) and end-expiratory thickness (eET). eET was recorded just before the thickening start and eIT at maximal thickening. The diaphragmatic thickening fraction (DTF) was calculated as follows: DTF = (eIT - eET)/eET.