A 41-year-old hemodialyzed woman developed ascites and was found to have secondary iron overload. The dose ofadministered iron was approximately 11–12 g, and her serum ferritin level was 15,000 ng/ml (15,000 μg/l). There were no signs of congestive heart failure, fluid overload, or liver cirrhosis. A program of weekly phlebotomy combined with recombinant human erythropoietin (rhEPO) therapy was tried to eliminate the iron congestion. After 9 months of this therapy, about 5 g of iron had been removed. The ascites completely disappeared, and her serum ferritin level fell to 5,800 ng/ml (5,800 μg/l). This suggests that such combined therapy would be useful when iron overload must be corrected rapidly. Before therapy, the sterile ascitic fluid showed exudative characteristics with 3.7 g/dl (37 g/l) of total protein. The serum-ascites albumin difference was 0.6 g/dl (6 g/l), and the fluid contained 1,400 inflammatory cells/mm<sup>3</sup> (1.4 × 10<sup>9</sup>/l). Notably, the serum-ascites albumin difference increased in parallel with iron elimination. These findings suggested that iron deposition may have played a role in changing the permeability of the peritoneum, or in impairing lymphatic drainage, both of which are presumed to be pathogenetic factors of nephrogenic ascites.