Hemofiltration, or blood ultrafiltration, is a pressure-driven process used to remove waste solutes and excess water from blood in the treatment of kidney dysfunction. This convective process provides better removal of middle molecular weight solutes (1,000-20,000 MW) than hemodialysis, substantially reducing the incidence of hypotension. The ultrafiltration is performed using a cross-flow configuration in which blood flow is parallel to the filtering membrane and perpendicular to the filtrate flow. This configuration minimizes the accumulation of retained species at the membrane surface, significantly increasing filtration rates. Clinical hemofiltration is performed in two distinct modes: post-dilution, in which the fluid removed during filtration is replaced after blood leaves the filter, and pre-dilution, in which the replacement fluid is added to the blood before it enters the filter. Although post-dilution is used more commonly, the high hematocrit at the device outlet limits the maximum attainable flux and can contribute to cell damage. The lower inlet hematocrit and plasma protein concentration in pre-dilution leads to a marked increase in flux, potentially increasing the removal of waste solutes. There is, however, a concomitant increase in the amount of diluting fluid required and in the amount of fluid that must be removed during filtration. Existing models for hemofiltration are unable to explain the complicated dependence of the filtrate flux on shear rate, hematocrit, and plasma protein concentration. These models cannot be used to accurately compare the performance of different devices or of the two clinical modalities.(ABSTRACT TRUNCATED AT 250 WORDS)