Patients who are critically ill with acute renal failure and sepsis have extremely high mortality rates. While it seems reasonable that eliminating the inflammatory mediators (such as cytokines, chemokines, tumor necrosis factor-alpha, etc.) by continuous renal replacement therapy (CRRT) would be effective, studies show that only insubstantial numbers of these mediators are removed in comparison with endogenous clearance. Mass removal seems only to be effective when highly permeable membranes (sieving coefficient of approximately 1.0) are used, there is a filtrate volume greater than 2 liters/hour, and when the half-life of the substance to be eliminated is greater than 60 minutes. Removal of cytokines by membrane adsorption is another possibility. However, because the membrane surfaces are saturated after a few hours, frequent filter changes are necessary for them to generate effective adsorption of these mediators. Despite filter changes, only a brief and transient drop in the TNF plasma level has been observed. Controlled clinical trials are needed to determine whether or not CRRT actually has a beneficial effect on the systemic inflammatory response syndrome (SIRS).