The aim of the present investigation was to define whether multisite subcutaneous (s.c.) administration in unanesthetized, unrestrained rabbits of human recombinant interferon-alpha 2 (rec. IFN-alpha 2) either in saline, human albumin (ALB) solution (4, 7 and 10% final concentrations), or in a solution containing 75 U of hyaluronidase, modified the pharmacokinetic parameters calculated from the IFN plasma levels. Plasma disappearance rates of rec. IFN-alpha 2 were measured in rabbits after intravenous (i.v.) administration and the kinetic was adequately represented by a three-pools mammillary model. This model was the basis for evaluating the absorption and distribution of rec. IFN-alpha 2 after s.c. administration. The increase of ALB concentration (from 4 to 10%) caused a significant reduction of the plasma IFN Cmax while both the mean residence time and the release time of IFN increased linearly with the ALB concentration. The data support the postulation that s.c. administration of albumin acts as an interstitial fluid expander and may favour absorption of IFN via lymphatics rather than blood capillaries. Improvement of therapeutic index of IFN by using this route remains to be shown in clinical trials.