Extracorporeal membrane oxygenation (ECMO) has been successfully used to support patients with cardiac arrest failing to respond to conventional cardiopulmonary resuscitation (CPR). Preimplant factors being indicative for success are unknown up to now. The study describes single center experience with special focus on differences between survivors and nonsurvivors. Between 2002 and 2009, 103 patients were supported within the scope of CPR by means of ECMO. Besides primary diagnosis, duration, and outcome, pH, lactate, mean arterial pressure, aspartate aminotransferase, bilirubin, catecholamine dosage, and oxygenation ratio before ECMO, after 2 h, 1 day, and at explantation were analyzed. One hundred three patients (51.2 ± 16 years, 35 women, 68 men) were analyzed. Primary cardiac failure led to CPR in 54%. Duration of support was 4.8 ± 0.6 days. Twenty-nine (28.1%) patients survived to hospital discharge. On ECMO support, pH, lactate, and mean arterial pressure improved significantly. Catecholamine dosage was significantly reduced after ECMO implantation. Demographic data and primary diagnosis revealed no significant influence on outcome. pH, lactate, creatinine, and bilirubin differed significantly between survivors and nonsurvivors in the course of ECMO support. ECMO support during CPR reliably improves the circulatory and respiratory situation. Considering observed survival critical patient selection is mandatory. Although there are several significant differences between surviving patients and patients with fatal outcome, patient selection turns out to be difficult as clinically relevant factors show only limited predictive value. Future research should focus on better defining a population that may be best of all suited for the use of ECMO support in CPR.