Prolactin has been shown to play a role in the modulation of the immune system. In addition, recent laboratory evidence suggests that circulating prolactin might fluctuate as a function of immunologic events. To evaluate the role of prolactin as a predictor of rejection, 22 heart and heart-lung transplant patients were studied from July 1985 to May 1986. Daily plasma prolactin levels were measured by radioimmunoassay. Endomyocardial biopsy and histologic evaluation of the allografts were used to define rejection episodes. Thirteen of 14 documented rejection episodes were preceded by a specific elevation of circulating prolactin. A significant elevation of prolactin occurred 6 and 8 days before 14 rejection episodes that were diagnosed by biopsy but not before 109 biopsies were taken that had negative results. The normal values for circulating prolactin are defined by a preoperative mean in a range of 0 to 15 ng/ml (95% confidence limit). The sensitivity of prolactin to predict rejection is 79%, the specificity is 92%, and the positive predictive value is 61%. A large data variability gives a low accuracy to the test as a predictor of allograft rejection. To be useful clinically, the accuracy needs to be improved through a better understanding of prolactin physiology and its relation to the immune system.