1 Single oral doses of 100 mg racemic ketoprofen were given to 15 patients (age range:51-79 years) with rheumatoid arthritis and a range of creatinine clearances (CLCR) from 26 to 159 ml min-1. 2 The fractions unbound of (R)-and (S)-ketoprofen in plasma were determined for each subject after in vitro addition of rac-ketoprofen (enantiomer range: 1.00-6.00 ,ug ml-1) to pre-dose plasma. 3 An index of the antiplatelet effect of ketoprofen in vitro was measured as inhibition of platelet thromboxane B2 (TXB2) generation during the controlled clotting of whole blood (pre-dose) spiked with rac-ketoprofen. 4 In vivo studies revealed significant associations (P < 0.05) between the reciprocal of AUC for both unbound and total (bound plus unbound) (S)-ketoprofen and CLCR. Corresponding relationships were also observed for the (R)-enantiomer of ketoprofen. In addition, the half-life of each enantiomer was negatively correlated with CLCR.There was a positive relationship between the 24 h urinary recovery of combined non-conjugated and conjugated (R)-ketoprofen and CLCR while that for the (S)-stereoisomer failed to reach statistical significance (P > 0.05). 5 There was no difference between AUC for (R)-and (S)-ketoprofen for either unbound or total drug. 6 The mean ± s.d. percentage unbound of (S)-ketoprofen in plasma (0.801 ± 0.194%) exceeded (P < 0.05) the corresponding value for its optical antipode (0.724 + 0.149%). The percentage unbound of the (S)-enantiomer was higher at 6.00 ,ug ml-l than that at enantiomer concentrations of 3.50 ,ug ml-1 and below, where it was invariant. The percentage unbound of (R)-ketoprofen was independent of plasma concentration up to 6.00 ,ug ml-l. There were no correlations between the percentage unbound of each enantiomer and either serum albumin concentration or CLCR. 7 The relationship between the serum concentration of unbound (S)-ketoprofen and the percentage inhibition of platelet TXB2 generation was described by a sigmoidal Emax equation for each patient. There was no correlation between the unbound concentration of (S)-ketoprofen in serum required to inhibit platelet TXB2 generation by 50% (EC50) and CLCR. The mean ± s.d. EC50 was 0.216 ± 0.143 ng ml-1.8 These data indicate that diminished renal function is associated with an increased exposure to unbound (S)-ketoprofen, presumably due to regeneration of parent aglycone arising from the hydrolysis of accumulated acyl-glucuronide conjugates. The apparent sensitivity of platelet cyclo-oxygenase to the inhibitory effect of (S)-ketoprofen was not influenced by renal function.