Dr. Horton raises two important questions about the care of patients with uncomplicated myocardial infarction: length of stay and whether a stress perfusion imaging study is necessary before hospital discharge. Although it is true that our specific recommendations have yet to be tested in large randomized trials, we can already estimate the potential outcomes and costs associated with this strategy on the basis of information available from large myocardial infarction registries. For example, one can assess the cost-effectiveness of inpatient stays beyond hospital day 5 (for example, staying 7 versus 5 days) in patients with uncomplicated myocardial infarction. The benefit of longer hospitalization can be measured in terms of the patient's higher likelihood of surviving a cardiac arrest if this event occurs in the hospital. The rate of cardiac arrest in patients with uncomplicated myocardial infarction between days 5 and 7 is 0.18%; 52% of these resuscitated patients survive to hospital discharge (Unpublished data from GUSTO). If we assume that every patient who has an arrest out of the hospital dies, then a larger in-hospital strategy would save approximately one life for every 1000 patients treated, at an added cost of $1 080 000 ($540/regular hospital day x 2 days x 1000 patients). If the average life expectancy for post-myocardial infarction patients is 11.3 years [1], then the cost per added life year for a longer-stay strategy would be $95 575, an economically unattractive figure [2]. If the life expectancy for survivors of cardiac arrest is only half that seen for the average postinfarction patient, then a longer in-hospital stay strategy would cost $2 160 000 per life saved (or $191 150 per added life-year).

With regard to which predischarge testing strategy to use in patients with uncomplicated myocardial infarction, our guidelines first recommend assessment of ventricular function [3, 4]. In patients with preserved ventricular function, we recommend some form of noninvasive stress testing before discharge. We believe that exercise stress testing can help set cardiac rehabilitation thresholds and identify a few additional higher-risk patients. However, because the overall 30-day and 1-year mortality rates in this patient population are less than 1% and 3%, respectively, it remains to be proven whether any additional testing strategy (such as one using perfusion imaging) can significantly improve on this already excellent prognosis.