OBJECTIVES To assess the feasibility and safety of a conservative approach to oxygen therapy in mechanically ventilated ICU patients. DESIGN Pilot prospective before-and-after study. SETTING A 22-bed multidisciplinary ICU of a tertiary care hospital in Australia. PATIENTS A total of 105 adult (18 years old or older) patients required mechanical ventilation for more than 48 hours: 51 patients during the "conventional" before period and 54 after a change to "conservative" oxygen therapy. INTERVENTIONS Implementation of a conservative approach to oxygen therapy (target SpO2 of 90-92%). MEASUREMENTS AND MAIN RESULTS We collected 3,169 datasets on 799 mechanical ventilation days. During conservative oxygen therapy the median time-weighted average SpO2 on mechanical ventilation was 95.5% (interquartile range, 94.0-97.3) versus 98.4% (97.3-99.1) (p < 0.001) during conventional therapy. The median PaO2 was 83 torr (71-94) versus 107 torr (94-131) (p < 0.001) with a change to a median FIO2 of 0.27 (0.24-0.30) versus 0.40 (0.35-0.44) (p < 0.001). Conservative oxygen therapy decreased the median total amount of oxygen delivered during mechanical ventilation by about two thirds (15,580 L [8,263-29,351 L] vs 5,122 L [1,837-10,499 L]; p < 0.001). The evolution of the PaO2/FIO2 ratio was similar during the two periods, and there were no difference in any other biochemical or clinical outcomes. CONCLUSIONS Conservative oxygen therapy in mechanically ventilated ICU patients was feasible and free of adverse biochemical, physiological, or clinical outcomes while allowing a marked decrease in excess oxygen exposure. Our study supports the safety and feasibility of future pilot randomized controlled trials of conventional compared with conservative oxygen therapy.