Bioartificial livers (BALs) are bioreactors containing liver cells that provide extracorporeal liver support to liver-failure patients. Theoretically, the plasma perfusion flow rate through a BAL is an important determinant of its functionality. Low flow rates can limit functionality due to limited substrate availability, and high flow rates can induce cell damage. This hypothesis was tested by perfusing the AMC-BAL loaded with the liver cell line HepaRG at four different medium flow rates (0.3, 1.5, 5, and 10 mL/min). Hepatic functions ammonia elimination, urea production, lactate consumption, and 6β-hydroxylation of testosterone showed 2–20-fold higher rates at 5 mL/min compared to 0.3 mL/min, while cell damage remained stable. However, at 10 mL/min cell damage was twofold higher, and maximal hepatic functionality was not changed, except for an increase in lactate elimination. On the other hand, only a low flow rate of 0.3 mL/min allowed for an accurate measurement of the ammonia and lactate mass balance across the bioreactor, which is useful for monitoring the BAL's condition during treatment. These results show that (1) the functionality of a BAL highly depends on the perfusion rate; (2) there is a universal optimal flow rate based on various function and cell damage parameters (5 mL/min for HepaRG-BAL); and (3) in the current set-up the mass balance of substrate, metabolite, or cell damage markers between in-and out-flow of the bioreactor can only be determined at a suboptimal, low, perfusion rate (0.3 mL/min for HepaRG-BAL). Biotechnol. Bioeng. 2012; 109: 3182–3188. © 2012 Wiley Periodicals, Inc.