Factors that predict hospital length of stay after heart transplantation were identified from retrospective data of 65 patients (82% male, mean age, 43.3 years). Multiple regression analysis with a stepwise procedure was used to generate three predictive models for length of stay: (1) a model to be used before operation, (2) a model that combines preoperative and donor information, and (3) a model that takes preoperative, donor, and postoperative factors (complete model) into consideration. Hospital length of stay ranged from 15 to 45 days after heart transplantation (median length of stay, 22.5 days; mean length of stay, 24.4 +/- 6.4 days). In the preoperative model, diagnosis, duration of cardiac symptoms, severity of heart failure, and pulmonary vascular resistance were significantly related to length of stay and together accounted for 36% of the variance in length of stay. When donor information (for example, size and ischemic time) was added to preoperative information, the resultant model failed to account for appreciably more of the variance in length of stay. A model that considered preoperative, donor, and postoperative factors accounted for 71% of the variance in length of stay. Significant variables in the model were the month in which the patient had transplantation in the program, duration of cardiac symptoms before transplantation, preoperative severity of heart failure, pulmonary vascular resistance, and postoperative incidence of severe acute rejection. Patient age, sex, and postoperative infections were not related to length of stay. In conclusion, there are cardiopulmonary and immunologic factors that can predict length of stay. The model also suggests that a program's experience with heart transplantation affects length of stay.