The urinary excretion of stable metabolites of thromboxane A2, such as 11-dehydro-thromboxane B2, reflects platelet activity in vivo. Efficient sample purification is required before analysis of thromboxane metabolites, due to the presence of large amounts of interfering material in urine. Analysis by gas chromatography-mass spectrometry after extensive sample work-up procedures provides the most reliable data, but detection by enzyme immunoassay may be reliable if sample cleanup is adequate. We describe an improved immunoassay procedure for 11-dehydro-thromboxane B2, which is based on a simple one-step solid phase extraction, by using Bond-Elut Certify II columns, followed by enzyme immunoassay by using commercially available reagents. 11-Dehydro-thromboxane B2 exists in two forms, with different chemical and immunological characteristics, which are in pH-dependent equilibrium. We kept 11-dehydrothromboxane B2 in its open ring form throughout the assay, by incubating and handling samples at pH 8.6. The extraction step achieved a recovery of 83% (95% confidence interval 74-92%), the sensitivity of the enzyme immunoassay was doubled, and the reproducibility of the assay improved under these conditions. Intra- and interassay coefficients of variation were 3 and 13.8%, respectively. A single 500-mg dose of aspirin reduced the excretion of 11-dehydro-thromboxane B2 by 77+/-14%, suggesting good specificity. Comparison with gas chromatography-mass spectrometry in 28 urine samples showed excellent agreement between the two methods (r2 = 0.94; p<0.0001), and a regression line with a slope close to 1.0. The presently modified enzyme immunoassay for 11-dehydro-thromboxane B2 is suitable for clinical studies evaluating platelet function in vivo and has the advantage of being simpler and less expensive to use than gas chromatography-mass spectrometry.