Cyclosporin-A levels were determined by radioimmunoassay in plasma or whole blood, using split samples collected from patients receiving this agent as the only form of immunosuppression following allogeneic bone marrow transplantation. In the plasma assay the temperature at which centrifugation took place was critical since the mean levels were approximately 30% higher with separation at 37 degrees C in comparison to 20 degrees C or lower. Furthermore, the level in whole blood samples was 2.4 times higher than that from the matching serum. In addition, anticoagulated blood that had been frozen and then thawed was technically more difficult to pipette and resulted in a recovery of only 83% of the cyclosporin when compared with assay using fresh blood. In contrast, consistent measurements were obtained either when whole blood was stored at 4 degrees C and then well mixed and diluted in buffer immediately prior to use or when such buffered samples were frozen and thawed immediately before analysis. The latter modifications render the whole blood assay a practical and reliable means for monitoring cyclosporin-A concentrations and may avoid excessive and the potentially nephrotoxic levels achieved when plasma levels are held in ranges previously considered therapeutic.