We thank the correspondents for their valuable comments about our concerns that an increasing cumulative epinephrine dose during resuscitation is independently associated with unfavorable neurologic outcome after ventricular fibrillation cardiac arrest.

Although caution must be exercised in generalizing our results, they are important because they are the first to call into question the limit of subsequent epinephrine doses (1). Resuscitation researchers are well aware of the limitations in their study designs and the many other difficulties involved in conducting clinical studies. However, no better methods to overcome these handicaps, other than using an internationally recognized protocol (1), have been suggested.

Emergency physicians on the Vienna ambulance service administer a median epinephrine dose of 1.2 mg (interquartile range, 0.6 to 2.1 mg) every 5 minutes. This practice differs from those reported in informal surveys across the United States, which revealed that most emergency physicians follow an escalating approach—1, 3, and 5 mg given 3 minutes apart (1). Our retrospective cohort study showed that bystander-administered basic life support significantly reduced the no-flow time and therefore may indirectly improve outcome. Our study may have allowed increasing age to be a prognostic factor for the quality of neurologic outcome.

The study sample and methods chosen were based on our study objective. The calculation of the no-flow duration is an estimate; by nature, it cannot be quantitated accurately. Imprecision regarding the exact time of recognition of collapse and the time of emergency medical activation was minimized by using highly accurate data.

For those who conduct prospective or retrospective clinical research on resuscitation, it is always surprising how many more questions will be raised by such an undertaking. Much more surprising, however, is the number of therapies used in daily clinical routine that have not been sufficiently studied. Therefore, every effort in resuscitation research is justified and already recognizes the final denominator. We are not going to throw in the towel! We will continue to try to convince researchers that additional trials and reanalysis of existing data are needed. For example, Gueugniaud and colleagues (2) recently showed that long-term survival after cardiac arrest outside the hospital was no better with repeated high doses of epinephrine than with repeated standard doses.