The pharmacokinetics and pharmacodynamics of recombinant human interferon-beta (rHuIFN-beta 1a) were assessed following administration to 12 healthy male volunteers. Each subject received, in a double-blind, balanced, random-order, crossover sequence, single doses of 6 MIU of rHuIFN-beta 1a (Rebif) i.v., i.m., and s.c. or matching placebo on four occasions separated by washout periods of 1 week. Blood samples were collected at preset times for the measurement of serum IFN-beta levels and of intracellular 2'-5'-oligoadenylate synthetase levels. Blood pressure, sitting heart rate, respiratory rate, oral body temperature, and tolerance were monitored regularly. All administrations of rHuIFN-beta 1a were well tolerated, although about half of the subjects had a flu-like syndrome, as expected. After i.v. bolus injection, the pharmacokinetics of rHuIFN-beta 1a were well described by a classic two-compartment model. Mean total clearance of rHuIFN-beta 1a was about 100 L.h-1. The distribution half-life was 5 min, and the terminal half-life was approximately 5 h. After i.m. or s.c. injection, serum IFN-beta profiles were rather flat, and about one sixth of the administered dose was available systemically. Extent and duration of clinical and biologic effects were independent of the route of administration and of the IFN-beta serum levels. Biologic pharmacodynamic effects persisted even when IFN-beta serum levels had returned to baseline and were still significantly elevated 3 days after a single dose. Because of the independence of the extent and duration of clinical and biologic pharmacodynamic effects from the route of administration and from the IFN-beta serum levels, the s.c route of administration is preferred in indications in which primarily an immunomodulatory action is sought. Predominantly antiviral and antiproliferative activity is enhanced by the i.v. route to provide adequate drug levels at the site of pathology, although its application is limited on practical grounds.