Within 4 h of intubation, the decision was made to initiate venovenous ECMO. The patient was taken to the operating room for placement of a 20F cannula in her right internal jugular vein and a 23F cannula in her right femoral vein. We drained blood from the patient’s femoral vein and returned blood through her internal jugular vein in order to avoid recirculation. Following the initiation of ECMO, tidal volumes were decreased and Fio2 was lowered in order to prevent barotrauma and oxygen toxicity. The patient’s initial ventilator settings were pressure-regulated volume control, a rate of 12 breaths/min, a tidal volume of 350 mL (5 mL/kg), a PEEP of 12 mm Hg, and an Fio2 of 0.5. Due to peak inspiratory pressures of > 30 mm Hg, her tidal volume was further reduced to 325 mL, and PEEP was reduced to 10 mm Hg. The initial ECMO settings were a flow of 3.5 L/min and a sweep of 2 L/min. Because of the risk of continued alveolar hemorrhage, anticoagulation therapy was avoided for the first 48 h by using a centripetal pump, which does not require systemic heparinization. She was subsequently changed to a traditional roller pump, and she received full anticoagulation. During her ECMO run, flows were maintained between 2 and 3 L/min, and the sweep was gradually increased to 4.5 L/min. Beginning with hospital day 9, we began attempting breathing trials without ECMO. Over the next few days, no major changes were made with the ventilator, and the chest radiograph findings gradually improved. On hospital day 16 (ECMO day 12), the patient was able to remain without ECMO for our goal of 4 h with oxygen saturations in the mid-90% range. At the conclusion of this trial, an arterial blood gas analysis revealed a pH of 7.27, a Pco2 of 50 mm Hg, and a Po2 of 80 mm Hg. The patient was subsequently decannulated at the bedside.