During the recruitment phase of the MIAMI trial (December 1982 to March 1984), data on 26,439 patients eligible for inclusion were entered. Of these, 5,778 patients were included. Current treatment with either beta blockers or calcium-channel blockers (51%) was the most predominant reason for exclusion. The randomized and excluded patients differed. The randomized patients were younger and more often men. The mean age of the patients was 59 years in both the placebo and the metoprolol groups. The 2 groups were evenly balanced with regard to basic demographic variables. The median delay between onset of symptoms and randomization was 7 hours, and 25% of the patients were included within 4 hours. Previous clinical history and pharmacologic treatment given before admission were well balanced in the groups. Mean heart rate for the 2 groups before randomization was 83 beats/min and systolic blood pressure was 141 mm Hg. Approximately 15% of randomized patients presented with pulmonary rales. Before randomization 20% of the patients had normal electrocardiograms; 70% could be classified as having electrocardiographic signs of acute myocardial infarction; and 10% presented with other electrocardiographic abnormalities. Electrocardiographic signs at entry suggested a predominance of anterior wall infarctions. The randomized patients were not representative of eligible patients and the treatment groups were well balanced at entry.