A reappraisal of the priorities of cardiopulmonary resuscitation by Pepe and colleagues [1] calls attention to the evidence that defibrillation may not be the optimal initial intervention. Initial precordial compression after more than perhaps 3 minutes of untreated cardiac arrest greatly improves the likelihood of successful conversion of ventricular fibrillation with restoration of spontaneous circulation [2, 3]. Since it is often very difficult to gauge this time interval, and whether it exceeds 3 min, a number of both preclinical investigators [4–6] and clinical investigators [7, 8] have sought an electrocardiographic predictor of the likelihood that an electrical shock will restore circulation. It is to this extent that we applaud the authors' call for caution lest the availability and promotion of automated external defibrillators diminishes the preparedness of the rescuer to proceed with more conventional basic life support. The authors appropriately point to the time dependency of basic life support interventions.

Pepe and colleagues then extend their discussion to advanced cardiac life support, and especially pharmacological interventions. In the context of the time dependency of interventions, they are not prepared to discard the possibility that high-dose epinephrine will improve outcome. Indeed, they favor the use of cocktails, entertaining the possibility that epinephrine may be administered conjointly with antioxidants and anti-arrhythmic drugs. Although we agree with Pepe and colleagues in implicating the myocardial energy/supply relationship as an important issue, we wish to point out that beta-adrenergic agonists, and to a lesser extent the alpha1 actions, greatly increase myocardial energy consumption and thereby intensify the severity of myocardial injury [9, 10]. Although the alpha-adrenergic effects increase coronary perfusion pressure and transiently increase myocardial blood flow, the downside is major. The inotropic and chronotropic effects produce greater global myocardial ischemia, greater post-resuscitation ventricular ectopy and recurrent ventricular tachycardia, and recurrent ventricular fibrillation. Even more importantly, the adrenergic inotropic and chronotropic actions result in greater severity of post-resuscitation myocardial dysfunction.

Although epinephrine has been used as a resuscitative drug for more than a century, and although there is evidence that epinephrine may facilitate initial resuscitation, there is no proof of ultimate clinical benefit in terms of survival. To the contrary, we suspect that vasopressor agents with no inotropic and chronotropic actions are likely to come to the fore. This has prompted interest in more selective vasopressor agents, including α-methylnorepinephrine [10] and nonadrenergic vasopressin [11, 12]. Moreover, as yet unpublished preclinical studies from our group now suggest a place for beta-adrenergic blocking agents.

Looking to the future, we also see an opportunity for much improvement in sequencing interventions. In support of the conclusions reached by Pepe and colleagues, additional precision in sequencing cardiopulmonary resuscitation interventions has a high likelihood of improving outcomes. Perhaps one measurement, not cited by the authors, is end-tidal carbon dioxide. Both experimentally and clinically, end-tidal carbon dioxide has been a noninvasive monitor of blood flow generated by precordial compression [13, 14]. It serves as quality control of precordial compression and allows chest compression to remain uninterrupted. No longer will advanced cardiac life support rescuers need electrocardiographic confirmation of the return of spontaneous circulation because it is overtly signaled by an overshoot in the end-tidal carbon dioxide.