Measurements of erythropoietin in fetal blood obtained by cordocentesis or in umbilical cordblood and maternal blood have shown a lack of correlation indicating an independent regulation of EPO concentration in fetal and maternal compartments. There is a good correlation between amniotic fluid EPO concentration in fetal blood levels. Fetal EPO concentration therefore might serve as an indicator of chronic fetal hypoxia with fetal EPO production being responsive to tissue hypoxia early on in pregnancy. The lack of human placental permeability for EPO was further investigated using a dual in vitro perfusion system of an isolated cotyledon in freshly delivered term placentae. With recirculation of both circuits trace amounts of EPO (0.04% of the amount added to the maternal compartment) were transferred to the fetal side during 4-5 hours of perfusion. This transfer is comparable to the rate determined in the same experiments for albumine, and the biological significance of this very slow transfer is questionable. A very low rate of diffusion across the human placenta has also been shown for dextran, horseradish peroxidase and heparin using an in vitro perfusion system. The only exception among macromolecules are immunoglobulines G, which towards the end of pregnancy are transferred by an Fc-receptor mediated transcellular mechanism from the mother to the fetus. It is concluded, that there is no easy exchange of EPO across the human placenta between maternal and fetal compartments. Changes in EPO concentration in the fetal compartment therefore could serve as indicator of fetal hypoxia. A therapeutic application of EPO in the mother for the treatment of chronic anemia would not have any effect on fetal tissues.(ABSTRACT TRUNCATED AT 250 WORDS)