1. Urinary levels of the 2,3-dinor metabolites of thromboxane A2 (2,3-dinor-thromboxane A2, Tx-M) and prostacyclin (2,3-dinor-6-keto-prostaglandin F1 alpha, PGI-M) are frequently analysed as indices of platelet and endothelial activity and interaction in vivo. Despite this, little is known about the possible diurnal variations in urinary Tx-M and PGI-M in healthy human subjects, and how the urinary levels of Tx-M and PGI-M in single samples reflect their respective 24 h excretion rates. We addressed this by determining Tx-M, PGI-M and creatinine in consecutive portions of urine collected during 24 h in 15 healthy non-smoking subjects. 2. The total 24 h excretion of Tx-M and PGI-M did not differ between men (223 +/- 31 and 132 +/- 27 ng, respectively) and women (215 +/- 44 and 127 +/- 29 ng, respectively). Neither the excretion of Tx-M nor that of PGI-M displayed any significant diurnal variation. 3. The excretion of Tx-M during a 3 h period and the Tx-M/creatinine ratio in a urine sample accurately reflected the 24 h excretion of Tx-M (correlation coefficient ranges 0.74-0.94 and 0.74-0.86, respectively). The excretion of PGI-M during a 3 h period and the PGI-M/creatinine ratio in a urinary sample were accurate measures of 24 h excretion of PGI-M (correlation coefficient ranges 0.76-0.94 and 0.72-0.83, respectively). Urinary Tx-M and PGI-M expressed as simple concentrations were poor indices of their respective 24 h excretion. 4. We conclude that time-related excretions of Tx-M and PGI-M may be the best indices ex vivo of the cardiovascular formation of thromboxane A2 and prostacyclin, but that urinary creatinine-related concentrations of Tx-M and PGI-M in a urine sample are accurate measures as well.