Several studies have attempted to compare different doses and modes of therapy in continuous renal replacement therapies in critically ill patients. It is commonly asserted in the literature that convective therapies can achieve higher clearance of middle molecules than achieved by dialysis alone. However, regardless of the actual prescription, most therapies will actually contain a mixture of both diffusive and convective clearance. Molecular transport in purely convective prescriptions may be hindered by clotting and protein interactions with the dialyser. We measured middle molecule clearance using a tracer molecule, Ficoll, in citrated bovine blood. Using a 2 x 2 factorial design, we examined the impact of prescription [postdilution continuous venovenous hemofiltration (CVVH) vs. continuous venovenous hemodialysis (CVVHD)] and membrane area (0.4 m2 vs. 2.0 m2) on blood-side and dialysate-side middle-molecule clearance. In large dialysers, convective and diffusive prescriptions resulted in nearly identical middle molecule clearance from 10 to 100 kDa molecular weight. In the smaller dialyser, middle molecule clearance was higher when a diffusive therapy (CVVHD) was prescribed versus a convective therapy (postdilution CVVH). We hypothesized that high ultrafiltration rates in the smaller dialyzer resulted in a concentration polarization at the membrane that formed a prefilter, limiting middle-molecule clearance. This effect has implications for design and analysis of clinical trials of continuous renal replacement therapy (CRRT).