Objective: Extracorporeal membrane oxygenation (ECMO) is increasingly used as rescue therapy for patients with severe acute respiratory distress syndrome (ARDS). We report our experience with this therapy and evaluate the effect of different cannulation strategies. Methods: Thirty patients received an ECMO for severe ARDS due to pneumonia (n=8), lung graft failure (n=5), primary lung disease (n=5), trauma (n=2), sepsis (n=2), near-drowning (n=1), or before or after major surgery (n=7). Initially, veno-venous ECMO was used in 19, veno-arterial ECMO in 8, and veno-veno/arterial ECMO in 3 patients. Eight patients were upgraded from veno-venous (n=5) or veno-arterial (n=3) to veno-veno/arterial ECMO due to poor oxygenation and hemodynamic instability, respectively. Results: Sixteen patients died within 30 days after ECMO implantation (53%), 13 patients survived to hospital discharge (43%). Causes of death were sepsis and multiorgan failure. There was a trend (p=0.057) for lower 30-day mortality (27%) in patients who were on veno-veno/arterial ECMO as compared to patients on the veno-venous and veno-arterial mode. Multivariate analysis did not identify a risk factor or predictor for outcome on ECMO. Complications were bleeding at the cannulation site in eight patients (27%), limb hypoperfusion and hyperperfusion in one patient each (7%), and wound infection in one patient (3.5%). Conclusions: Considering a predicted mortality rate of 80% in ARDS, ECMO therapy provided a survival benefit in our patients. However, it remained unclear which patients profit most from this therapy. There is a trend for improved survival when the ECMO is run in a veno-veno/arterial mode.