There are at least 300,000 cardiac arrests annually in the United States. Cardiopulmonary resuscitation (CPR) effectively restores hemodynamic stability, return of spontaneous circulation (ROSC), in 40% to 60% of arrests. Prolonged survival is significantly lower because of underlying illness and the postresuscitation syndrome, specifically central nervous system injury and left ventricular stunning after resuscitation. Prognostic variables have been shown to predict survival in multivariate analyses, but no models are sufficiently accurate to predict futility. End-tidal carbon dioxide has prognostic value and can measure the efficacy of CPR. Cardiac arrest outcomes will be most improved with public education and earlier initiation of resuscitative efforts, both Basic Life Support and Advanced Cardiac Life Support, notably defibrillation. Active compression-decompression and interposed abdominal compressions improved ROSC in prospective randomized trials; abdominal compressions have also been shown to increase survival to hospital discharge. Despite 30 years of research, CPR is now performed much as it was initially. Further research into the mechanisms of cardiac arrest, development of predictive models, and improved means to improve cardiac output and survival are needed.