The influence of four variables, specifically PEG molar mass (400, 1,000, and 8,000 g/mol), concentrations of PEG and phosphate salts (15, 20, and 25% for both), and agitation intensity (110, 150, and 200 rpm), on clavulanic acid (CA) extraction by extractive fermentation with PEG/phosphate salts aqueous two-phase system was investigated in shaken flasks using a 2(4-1)-fractional factorial design. After selection of the two most significant variables (agitation intensity and PEG molar mass), an optimization study conducted according to a 2(2)-central composite design revealed that 25% PEG 8,000 g/mol and phosphate salts at 240 rpm (run 6) were the best conditions for the extractive fermentation, leading to the best results in terms of partition coefficient (k = 8.2), yield of CA in the PEG-rich phase (eta(T) = 93%) and productivity (P = 5.3 mg/Lh). As a first attempt to make a scale-up of these results, the effectiveness of the extractive fermentation was then checked in a bench-scale bioreactor under conditions as close as possible to the optimum ones determined in flasks. The highest CA concentration obtained in the PEG-rich phase (691 mg/L) was 30% higher than in flasks, thus demonstrating the potential of such a new process, integrating the production and extraction steps, as a promising, low-cost tool to obtain high yields of this and similar products. (C) 2010 American Institute of Chemical Engineers Biotechnol. Prog., 27: 95-103, 2011