The endomyocardial biopsy (EMB) in heart transplant recipients has been considered the "gold standard" for diagnosis of graft rejection (REJ). The purpose of this retrospective study is to develop long-term strategies (frequency and postoperative duration of EMB) for REJ monitoring. Between 1985 and 1992, 346 patients (mean age 44.5 years, female patients = 14%) received 382 heart grafts. For graft surveillance EMBs were performed according to a fixed schedule depending on postoperative day and the results of previous biopsies. In the first year the average number (no.) of EMBs/patient was 20 with 19% positive for REJ in the first quarter, dropping to 7% REJ/EMB by the end of the first year. The percentage of REJ/EMB declined annually from 4.7% to 4.5%, 2.2% and less than 1% after the fifth year. Individual biopsy results in the first 3 postoperative months had little predictive value. Patients with fewer than two REJ (group 1), vs patients with two or more REJ in the first 6 postoperative months (group 2), were significantly less likely to reject in the second half of the first year (group 1: 0.29 +/- 0.6 REJ/patient; group 2:0.83 +/- 1.3 REJ/patient; P < 0.001) and third postoperative year (group 1:0.12 +/- 0.33 REJ/patients; group 2:0.46 +/- 0.93 REJ/patient; P < 0.05). In conclusion, routine EMBs in the first 3 postoperative months have only limited predictive value, however the number of routine EMBs can be drastically reduced later depending on the intermediate postoperative REJ pattern.