Sources of amniotic fluid erythropoietin during normoxia and hypoxia in fetal sheep

Objective: Erythropoietin is present in human amniotic fluid and has been suggested as a marker of fetal hypoxia. The objectives of the present study were to determine whether erythropoietin is present in ovine amniotic fluid, fetal urine, and/or lung liquid and whether concentrations in these compartments change in parallel with endogenous fetal plasma erythropoietin concentration when the latter is increased experimentally. Study design: In late gestation chronically catheterized fetal sheep, samples of amniotic fluid and plasma, urine and plasma, lung liquid, amniotic fluid, and plasma were collected before and up to 7 days after induction of 4 types of fetal hypoxia: (1) acute anemic hypoxia that was induced by a single fetal hemorrhage, (2) progressive anemic hypoxia that was induced by daily exchange transfusion, (3) acute hypoxic hypoxia that was induced by the reduction of maternal inspired oxygen content, or (4) chronic placental insufficiency that was induced by daily umbilicoplacental embolization for 4 days. Erythropoietin concentrations were determined by radioimmunoassay. Statistical testing included analysis of variance and least squares regression. Results: Under basal, nonhypoxic conditions, amniotic fluid erythropoietin concentration averaged 33.2% ± 1.6% (SE) of fetal plasma erythropoietin concentration, and basal fetal urine and lung liquid erythropoietin concentrations ranged from low (<10% of plasma concentration) to nondetectable. Unlike the strong correlation in humans, basal amniotic fluid and plasma erythropoietin concentrations were correlated only weakly (r = 0.259; r² = 6.7%; P = .0027; n = 132). Amniotic fluid erythropoietin concentration approximately doubled after 12 hours of severe hypoxic hypoxia or after 24 hours of embolization-induced severe hypoxia but was unchanged after 12 hours of mild-moderate hypoxic hypoxia or 24 hours of anemic hypoxia. Concomitant fetal plasma erythropoietin concentrations increased to 28.1 ± 5.3, 12.5 ± 2.7, 10.8 ± 4.6, and 10.0 ± 1.3 times basal values, respectively. During progressive fetal anemia, urinary erythropoietin concentration increased almost 10-fold (P = .0023) but remained a small fraction (3.7% ± 0.4%) of plasma concentration; at 12 hours of hypoxic hypoxia, lung liquid erythropoietin concentration did not vary with the severity of the hypoxia and remained low relative to plasma concentration (4.2% ± 2.1%). Conclusion: Erythropoietin is present in ovine amniotic fluid, urine, and lung liquid. With only 3 potential sources, the fetal membranes appear to be the primary source of amniotic fluid erythropoietin in the nonhypoxic ovine fetus because basal urine and lung liquid erythropoietin concentrations are much lower than amniotic fluid concentrations. Although unchanged during mild-to-moderate fetal hypoxia, amniotic fluid erythropoietin concentration increases modestly during severe fetal hypoxia. In sheep, fetal urinary erythropoietin may contribute to this rise in amniotic fluid erythropoietin concentration during severe hypoxia, because fetal urinary and plasma concentrations increase in parallel during anemia.