Individualised antiplatelet therapy and platelet function testing have attracted considerable clinical interest, but several aspects of test performance have not been thoroughly evaluated. We investigated repeatability and concordance of light transmission aggregometry (LTA) induced with arachidonic acid (AA) 1.0 mM, PFA-100 induced with collagen/epinephrine, multiple electrode aggregometry (MEA) induced with AA 0.5 or 0.75 mM and VerifyNow Aspirin. Patients with stable coronary artery disease (n=43) and healthy individuals (n=21) were included. All tests were performed in duplicate at baseline in healthy individuals and in duplicate for four days in all study participants during aspirin treatment. Serum and urinary thromboxane metabolites were measured several times to evaluate cyclooxygenase-1 inhibition by aspirin. MEA was most sensitive for aspirin as treatment induced a 12-fold difference in AA-induced platelet aggregation. Coefficients of variation for duplicate measurements at baseline (0.4-12%), during aspirin treatment (3-46%) and for day-to-day variability (3-37%) differed markedly between tests and were lowest for VerifyNow. The prevalence of aspirin low-responsiveness also differed between tests (0-9%) and the agreement was low: kappa<or=0.21 for all tests compared with AA-induced LTA (reference test), which correlated best with VerifyNow (r=0.43, p<0.001). Urinary thromboxane metabolites did not correlate with any platelet function test, whereas serum thromboxane correlated with VerifyNow Aspirin (r=0.41, p=0.001). Overall, repeatability was moderate and the correlation between tests was low. VerifyNow Aspirin proved most reproducible, and this was the only assay showing a significant positive correlation with serum thromboxane. This study demonstrated that conclusions based on platelet function testing strongly depend on the assay used.