The hypoxemia of acute respiratory distress syndrome (ARDS) depends chiefly upon shunt and ventilation-perfusion (VA/Q) inequality produced by fluid located in the interstitial space, alveolar collapse and flooding. Variables other tham inspired oxygen fraction and the underlying physiological abnormality can influence arterial oxygen partial pressure (PaO2). Changes in cardiac output, hemoglobin concentration, oxygen consumption and alcalosis can cause changes in PaO2 through their influence on mixed venous PO2. Gas exchange (GE) in ARDS may be studied using the inert gas elimination technique (MIGET) which enables to define the distribution of ventilation and perfusion without necessarily altering the FIO2 differentiating shunt from lung units with low VA/Q ratios and dead space from lung units with high VA/Q ratios. Different ventilatory strategies that increase mean airway pressure (positive end-expiratory pressure, high tidal volumes, inverse inspiratory-expiratory ratio, etc) improve PaO2 through increasing lung volume by recruiting new open alveoli and spreading the intra-alveolar fluid over a large surface area. Also prone-position ventilation would result in a marked improvement in GE enhancing dorsal lung ventilation by the effects on the gravitional distribution of pleural pressure and the reduction in the positive pleural pressure that develops in dorsal regions in ARDS. Inhaled nitric oxide (NO) has been shown to increase PaO2 in ARDS patients by inducing vasodilation predominantly in ventilated areas redistributing pulmonary blood flow away from nonventilated toward ventilated areas of the lung thus resulting in a shunt reduction. On the same way inhaled prostaglandins (PGI2 or PGE1) causes selective pulmonary vasodilation improving pulmonary GE. Intravenous almitrine, a selective pulmonary vasoconstrictor, has been shown to increase PaO2 by increasing hypoxic pulmonary vasoconstriction. A synergistic effect was found between inhaled NO and almitrine. In spite of the improval of GE shown by these different techniques on ARDS, no effect was demonstrated on mortality or duration of mechanical ventilation.