We compared the incidence of ventricular fibrillation after release of a 20 min-proximal left anterior descending coronary artery ligation in dogs anesthetized with either pentobarbital (N = 26) or morphine sulfate plus alpha-chloralose (N = 26). Results were analyzed using the logistic risk regression model. In each group, outcome correlated with the amount of myocardium perfused by the ligated vessel (myocardium "at risk") as measured by postmortem coronary injection of monastral dyes. At any given myocardium at risk, animals anesthetized with alpha-chloralose were less likely to fibrillate after release than those anesthetized with pentobarbital (p less than 0.01). This difference in outcome can be described by the myocardium at risk at which half the animals are expected to fibrillate (MAR50). MAR50 was 20.9 g for the pentobarbital group and 30.7 g for the alpha-chloralose group. Myocardium at risk should be taken into account in intervention studies using this coronary reperfusion arrhythmia model. We chose the logistic risk regression model to correct for this variable because it allowed a good fit of the data and offered a concise standard format for expressing the results. Outcome in reperfusion studies may be influenced by the choice of anesthetic agent.