BACKGROUND: The determinants of early and repeated episodes of acute rejection after cardiac transplantation remain elusive. METHODS AND RESULTS: To gain insight into this phenomenon, a multivariate analysis for repeated events was applied to 229 patients receiving 249 transplanted hearts between 1981 and July 1, 1991 (595 rejection episodes). The mean frequency of rejection per patient after initial cardiac transplantation was 1.2 at 3 months, 1.8 at 1 year, and 2.8 at 5 years. The pattern of rejection was characterized by an early period of higher risk (greatest during the first month) followed by a low constant risk that continued throughout the period of follow-up (maximum, 9.5 years). By multivariate analysis, risk factors were identified for the likelihood of subsequent rejection after a previous rejection episode (or time of transplantation). Triple-drug immunosuppression plus induction therapy yielded a higher risk of early subsequent rejection compared with other baseline immunotherapy protocols, but it also provided the greatest freedom (95%) from rejection-related death during the first year. Risk factors in the constant phase of hazard included younger age at transplant, female donor and/or recipient, longer donor ischemic time, greater HLA donor-recipient mismatch, and an increased number of previous rejection episodes. CONCLUSIONS: Immunologic and other patient-specific characteristics as well as rejection history predict the likelihood of future rejection events. The value of any antirejection protocol must be evaluated both in terms of rejection episodes and rejection-related deaths. Future analyses may identify specific high- and low-risk patient subsets for rejection, which may provide a more rational basis for altering the amount of chronic immunosuppressive therapy.