IRV-induced increases in MAP are clearly associated with shunt reduction, but we find no studies that show shunt reduction without increased end-expiratory alveolar pressure. On the other hand, various studies in humans with ARDS and hyaline membrane disease and animal models of acute lung injury indicate that shunt reduction does not occur with IRV if there is no increase in end-expiratory alveolar pressure (21), that shunt reduction is the same with IRV as with conventional ventilation with PEEP when there are comparable levels of end-expiratory volume or alveolar pressure (16, 32), and that shunt reduction is greater when MAP is raised with PEEP than with IRV (27). Improved ventilation-perfusion matching with IRV is theoretically unlikely and, given the high FIO2 used in ARDS, improvements in oxygenation from more even ventilation would not be great. Deadspace reduction by extended inspiratory phase ventilation may allow only minor improvements in gas exchange (14-16, 21, 25, 26, 34). Thus, there is little evidence to indicate that oxygenation can be maintained or improved with IRV while volotrauma risk is reduced. Some have suggested that IRV may promote gradual shunt reduction over hours or days, and that slower inspiratory airflow may reduce injurious parenchymal shear forces. However, these potentially salutary effects of IRV are unproven. On the other hand, there are potential deleterious effects of IRV, including increased risk of volotrauma and the requirements for heavy sedation and neuromuscular blockage. IRV remains of unproven value in the management of ARDS.