1. Four healthy male subjects received racemic ibuprofen (200, 400, 800 and 1200 mg), orally, on four occasions, 2 weeks apart, according to a four-way Latin-square design, in order to investigate the influence of increasing dose of ibuprofen on the magnitude and duration of its antiplatelet effect as well as on the relationship between such effect and drug concentration. 2. The antiplatelet effect of ibuprofen was assessed by measuring the inhibition of platelet thromboxane B2 (TXB2) generation during the controlled clotting of whole blood. The plasma unbound concentration of S(+)-ibuprofen, the enantiomer shown in an in vitro study to be responsible for the inhibitory effect of platelet TXB2 generation, was measured using an enantioselective method. 3. The maximum percentage inhibition of TXB2 generation increased significantly with dose from a mean +/- s.d. of 93.4 +/- 1.2% after the 200 mg dose to 98.8 +/- 0.3% after the 1200 mg dose, and there was an increase with dose in the duration of inhibition of TXB2 generation. The effect of ibuprofen on platelet TXB2 generation was transient and mirrored the time-course of unbound S(+)-ibuprofen in plasma; on all but one of the 16 occasions, serum TXB2 concentrations returned to at least within 10% of the pretreatment concentrations within 24 h of ibuprofen administration. 4. For each subject, the relationship between the percentage inhibition of TXB2 generation and the unbound concentration of S(+)-ibuprofen in plasma was modelled according to a sigmoidal Emax equation. The mean plasma unbound concentration of S(+)-ibuprofen required to inhibit platelet TXB2 generation by 50% (EC50) was 9.8 +/- 1.0 micrograms l-1.(ABSTRACT TRUNCATED AT 250 WORDS)