Abstract
This overview addresses the pathophysiology of the acute respiratory distress syndrome (ARDS; conventional vs. COVID), the use of oxygen high flow (HFN) vs. noninvasive ventilation (NIV; conventional vs. helmet) and a multi-modal approach to avoid endotracheal intubation ("intubation"): low normal temperature, cooperative sedation, normalized systemic and microcirculation, anti-inflammation, reduced lung water, upright position, lowered intra-abdominal pressure. Increased ventilatory muscle activity ("respiratory drive") is observed in early ARDS, at variance with ventilatory fatigue observed in decompensated chronic obstructive pulmonary disease (COPD). This increased drive leads to impending then overt ventilatory failure. Therefore, muscle relaxation presents little rationale and should be replaced by lowering the excessive respiratory drive, increased work of breathing, continued or increased labored breathing, self-induced lung injury (SILI), i.e. preserving spontaneous breathing. As CMV is a lifesaver in the setting of failure but does not heal the lung, side-effects of intubation, controlled mechanical ventilation (CMV), paralysis and deep sedation are to be avoided. Additionally, critical care resources shortage requires practice changes. Therefore, NIV should be routine when addressing immune-compromised patients. The SARS-CoV2 pandemics extended this approach to most patients, which are immune-compromised: elderly, obese, diabetic, etc. The early COVID is a pulmonary vascular endothelial inflammatory disease requiring lower positive-end-expiratory pressure than the typical pulmonary alveolar epithelial inflammatory diffuse ARDS. This leads one to reassess a) the technique of NIV b) the sedation regimen facilitating continuous and extended NIV to avoid intubation. Autonomic, circulatory, respiratory, ventilatory physiology is hierarchized under HFN/NIV and cooperative sedation (dexmedetomidine, clonidine). A prospective randomized pilot trial, then a larger trial are required to ascertain our working hypotheses.