It has been shown previously that erythropoietin expression in uitro by hepatoma cells increases in response to hypoxia. To verify whether hypoxia of the tumor might result in hepatic release of erythropoietin in uiuo, serum erythropoietin concentrations were measured immunoenzymatically in 12 patients (5 women, 7 men) who underwent transarterial chemoembolization for hepatocellular carcinoma. Peripheral blood samples were collected at baseline, and after 6 hours and 1,2,3, and 7 days after the procedure. In a second set of experiments, performed in three male patients also undergoing chemoembolization for hepatocellular carcinoma, paired blood samples were collected after catheterization of the hepatic veins and of the right antecubital vein. None of the patients had erythrocytosis. In comparison with a baseline mean value i-SEM of 100.6 ? 12.6 pg/L, serum erythropoietin concentrations were the following; +6 hours, 55.4 -+ 18.0 (P < .001); +1 day, 102.4 ? 24.7 (P = NS), +2 days, 183.0 ? 31.1 (P < .05); +3 days, 155.0 2: 26.0 (P < .05); +7 days, 153.3 -C 27.4 (P < .05) (matched Student's t-test). The ratio of hepatic veidantecubital vein serum erythropoietin concentrations increased from 0.85 at baseline to 1.30 at +2 days, paralleling the increase of aspartate transaminase (r = .914, P < .005).
After chemoembolization, no correlation was found between serum erythropoietin and alpha-1-fetoprotein concentrations. The concentration of the latter, stable initially, decreased 7 days after the procedure. These findings demonstrate that in patients with hepatocellular carcinoma hepatic release of erythropoietin can oc- cur not only for unregulated paraneoplastic production by tumor cells but also as a predictable response to local hypoxia in the liver.