Dr. Gordon Ewy and his group at the University of Arizona are quietly marshalling evidence that challenges the current practice of cardio-pulmonary resuscitation (CPR). One of their latest publications is found here in Circulation. The investigators reviewed resuscitation records of patients experiencing out of hospital cardiac arrest in greater Tucson. Standard automated external defibrillator (AED) protocols were used. This resulted in frequent interruption of chest compressions such that compressions were performed only 43% of the time during resuscitation efforts. Initial defibrillation attempts did not restore perfusing rhythm is any patients, and survival was no better than their historical control.
What is the clinical importance of these observations? Over time following the onset of ventricular fibrillation (VF) myocardial ATP rapidly depletes. The result is that rapidly over time the relative effectiveness of electrical defibrillation diminishes in comparison to chest compressions as the initial modality of treatment, culminating in unresponsiveness to defibrillation unless reperfusion via adequate chest compression (and consequent repletion of myocardial ATP) is accomplished first. 
This has lead to the concept of the three phases of CPR: the electrical, the hemodynamic and the metabolic phases. These phases correspond to time periods of roughly 0-5 minutes, 5-10 minutes, and longer, respectively. During the electrical phase initial defibrillation is the most important. That’s in keeping with popular teaching. However, during the hemodynamic phase compression becomes the most important initial intervention, because by that time there is little hope that defibrillation will restore spontaneous circulation unless there is pre-treatment with chest compression to replete myocardial energy stores. The hemodynamic phase is the one often encountered by responders to out of hospital cardiac arrest. (The metabolic phase, during which measures to decrease brain metabolism such as therapeutic hypothermia are of importance, will not be discussed here).
The Arizona group has also demonstrated that current practices of rescue breathing may compromise myocardial perfusion due to delays and interruptions in chest compression . This and related evidence, coupled with survey data indicating reluctance of bystanders to perform mouth to mouth rescue breathing has lead these researchers to teach continuous compression CPR to the public and institute local fire department protocol changes in Tucson.
Although the American Heart Association (AHA) and International Liaison Committee on Resuscitation (ILCOR) strive to be evidence based and tend to await high level evidence before recommending changes in emergency cardiac care, expect the following or some semblance thereof to show up eventually in the recommendations: 1) The lay public will be taught compression only CPR; 2) the initial modality for emergency personnel responding to out of hospital cardiac arrest, unless compressions are already in progress, will be a prolonged and continuous series of compressions before defibrillation is attempted.
1) For witnessed VT or VF in the health care setting immediate defibrillation remains the initial modality (remember the electrical phase!).
2) This new thinking does NOT apply to pediatric codes or other arrests of suspected respiratory origin. Rescue breathing remains a higher priority in those situations.