Urinary levels of 2,3-dinorthromboxane B2 (Tx-M) and 2,3-dinor-6-ketoprostaglandin F1 alpha (PGI-M), measured by gas chromatography-negative ion-chemical ionization mass spectrometry, accurately reflect in vivo biosynthesis of thromboxane A2 (TxA2) and prostacyclin (PGI2), respectively. Although the basal excretion of Tx-M and PGI-M is adequately documented, no systemic data on the excretion during controlled exercise have been presented. We studied the effect of maximal tolerated exercise (2 h of bicycle ergometry) on the excretion of Tx-M and PGI-M in healthy humans. In addition, their urinary levels of epinephrine (E) and norepinephrine (NE) were analyzed. To address the impact of locally formed adenosine, all subjects were reinvestigated after administration of theophylline. Exercise did not affect the excretion of Tx-M (47 +/- 21 vs. 34 +/- 9 pg/mg creatinine) but increased the excretion of PGI-M (74 +/- 14 vs. 267 +/- 70 pg/mg creatinine; P less than 0.02). Theophylline augmented urinary NE and E but did not significantly change the PGI-M response to exercise. We suggest that the normal cardiovascular eicosanoid response to exercise is a platelet-independent increase in vascular PGI2 formation.