Background/Aims: Patients on chronic hemodialysis (HD) are often infected with hepatitis C virus (HCV), a common cause of chronic liver disease. In some cases, however, decreases in the serum HCV load after HD have been documented. To better understand this phenomenon, we investigated the effects of various types of dialysis membrane on virus load in the circulation in vivo and in vitro. Methods: HCV RNA levels in patients’ serum, filtrate and dialyzer membranes were analyzed semiquantitatively by reverse transcription-polymerase chain reaction (RT-PCR) before and after HD treatment with two to four different types of dialysis membrane. HCV RNA was also determined from each fraction in in vitro dialysis and ultrafiltration. Results: In HD patients treated with a polysulfone (PS) membrane and a hemophan membrane, the HCV RNA titer reproducibly decreased by a factor of 10<sup>–1</sup>–10<sup>–2</sup>. In contrast, a cuprophan (CU) membrane had no detectable effect on HCV viremia, and HD with the AN69 membrane reduced HCV RNA levels in only a subset of the patients. In addition, a PS membrane, but not a CU membrane, reduced the level of circulating HCV in an in vitro assay. In both in vivo and in vitro experiments, HCV RNA was recovered from the PS membrane itself, but not from the ultrafiltrate. Conclusions: Membrane-dependent adsorption of HCV occurs during HD, causing a transient reduction in HCV in the circulation of patients.