The capacity of platelets to generate thromboxane A2, reflected by measurement of serum thromboxane B2 (TxB2), greatly exceeds the systemic production of thromboxane in vivo. Thus, it is possible that substantial but incomplete inhibition of thromboxane formation ex vivo would still allow marked augmentation of thromboxane production in vivo. To address this hypothesis, we administered aspirin 120 mg, a selective inhibitor of thromboxane synthase (TxSl), 3-(1H-imidazol-1-yl- methyl)-2-methyl-1H-indole-1-propanoic acid (UK-38, 485) 200 mg, and a combination of both drugs to 12 healthy volunteers and measured the effects on serum TxB2 and urinary 2,3-dinor-thromboxane B2 (Tx-M), an index of endogenous thromboxane biosynthesis. Although serum TxB2 was maximally inhibited by 94 +/- 1% after aspirin and 96 +/- 2% after the TxSl, maximal depression of Tx-M was only 28 +/- 8% and 37 +/- 9%, respectively. Combination of aspirin with the TxSl resulted in a small but significant increase in inhibition of thromboxane generation ex vivo (98 +/- 1% v 94 +/- 1%; P less than 0.05), but a disproportionately greater fall in thromboxane synthesis in vivo (58 +/- 7%; P less than 0.01). Consistent with further inhibition of platelet thromboxane synthesis, addition of the TxSl abolished the transient decline in prostacyclin formation after aspirin alone. Administration of a lower dose of aspirin (20 mg) to 6 healthy subjects caused a small reduction in Tx-M (12 +/- 4%; P less than 0.05) and inhibited serum TxB2 by 48 +/- 2%. The relationship between inhibition of platelet capacity to form thromboxane ex vivo (serum TxB2) and synthesis in vivo (Tx-M) departed markedly from the line of identity. When total blockade of the capacity of platelets to generate thromboxane is approached, minor decrements in capacity result in a disproportionate depression of actual thromboxane biosynthesis. These results imply that pharmacologic inhibition of serum TxB2 must be virtually complete before thromboxane- dependent platelet activation is influenced in vivo.