Recent advances in recombinant technology have led to the development of prothrombin time (PT) reagents containing recombinant tissue factor which has been lipidated to allow expression of procoagulant activity. In this study we have compared International Normalised Ratios (INRs) determined using two such reagents and conventional thromboplastins in widespread use in the UK. Lyophilised plasma samples from eight different warfarinised patients were distributed to 33 laboratories in the UK. Each participant determined prothrombin times on 20 local fresh normal plasmas (used to derive mean normal PT and calculate INR) and the eight lyophilised samples, using manual technique and the following thromboplastins; Recombiplastin (Ortho Diagnostics Ltd); Innovin (Baxter Diagnostics Ltd); the conventional thromboplastin in local use. For eight plasmas the mean INRs determined with different reagents were as follows: Innovin (33 laboratories) - 3.4; Manchester Reagent (MR = 8 laboratories) - 3.4; Recombiplastin (33 laboratories) - 3.7; Instrumentation Laboratory (IL = 13 laboratories) - 4.4. Mean INR results with Recombiplastin were on average 7% greater than those obtained with Innovin, 8% greater than results with MR and 18% less than INRs with IL thromboplastin. There was no significant difference between results obtained with Innovin and MR. In contrast INRs obtained with IL were markedly (mean 28%) greater than results obtained with Innovin. This study employed lyophilised plasma and it is possible that some of the relationships described are influenced by this. However, the lyophilisation process employed did not influence the relationship between INRs of warfarinised plasmas obtained by the four main reagents described, indicating that the results are relevant to routine clinical practice. In conclusion, our data show some important differences are present between INRs determined using Recombiplastin, Innovin and two conventional thromboplastins.