In vitro studies of serum protein binding were conducted to examine the influence of hemodialysis therapy on the stereoselectivity for ketoprofen (KP) in patients with renal dysfunction. In normal volunteers, only small differences in binding to undiluted serum were observed between the enantiomers of racemic KP (10 micrograms/mL). As compared with healthy volunteers, significantly impaired binding to undiluted serum was observed in uremic patients before they received dialysis. This reduced binding was not accompanied by alterations in stereoselectivity. At the end of dialysis, a similar extent of defective binding to undiluted serum was still observed, whereas interestingly, its stereoselectivity was significantly enhanced. These findings indicated that preferential inhibition for S-enantiomer serum binding was caused by changes that occurred during hemodialysis. Among various physiological parameters, only enantioselective inhibition of KP serum binding was correlated with heparin-induced increases in free fatty acids (FFA) levels and with unbound levels of uremic toxins with indole ring, indoxyl sulfate and indole acetate. To elucidate the effects of these accumulated endogenous substances in serum of patients, ultrafiltration was also used to study the binding of KP enantiomers to isolated human serum albumin under various conditions. After dialysis, mean albumin level and pH value in patients were slightly but significantly altered, but in vitro binding data indicated that these factors could clearly be ruled out as an explanation for the origin of the induced stereoselectivity. In contrast to palmitate, oleate and laurate markedly induced the stereoselective inhibition, which indicated that these fatty acids preferentially inhibited the binding of S-KP via a competitive and allosteric mechanism. Among the four uremic toxins, only indoxyl sulfate and indole acetate predominantly displaced the S-enantiomer rather than its antipode through competitive interaction. In addition, serum unbound levels of indole uremic toxins were correlated with elevated FFA levels. FFA displaced those indole uremic toxins in a manner similar to KP, which suggests that FFA also indirectly inhibits the S-enantiomer binding by increasing the unbound levels of indole uremic toxins. Thus it is likely that combined direct and cascade effects of transiently increased FFA and unbound indole uremic toxins contributed to the preferential elevation of the pharmacologically active S-KP fraction in uremic patients at the end of hemodialysis.