Prothrombin times were measured in 120 orally anticoagulated patients and 122 patients not orally anticoagulated using two different commercial thromboplastin reagents, a rabbit brain thromboplastin (CRB-Thromboplastin, Roche) and a human placenta thromboplastin (Thromborel S, Behring). A good correlation (r = 0.93, slope = 1.089, intercept = 0.014) was found between the International Normalized Ratio (INR) values of the anticoagulated plasmas obtained with both thromboplastins. Defining five degrees of intensity of anticoagulation (between INR less than or equal to 2.0 to INR greater than or equal to 5.1) the intensity of treatment was judged concordantly with both thromboplastins in 67% of the patients and as differing by one degree in 32%. In only one out of 120 patients was a larger discrepancy of INR values found. - A good correlation (r = 0.93) was also found between the prothrombin times expressed as Quick percentage values as assessed with both thromboplastins in 122 patients not orally anticoagulated. In 82 cases a "normal" prothrombin time (Quick greater than or equal to 70%) and in 33 cases a "subnormal/diminished" percentage value (Quick less than 70%) was obtained with both thromboplastins. In only 7 out of 122 plasma samples was the Quick percentage normal with one and subnormal with the other reagent. - The good conformity of the INR values measured with two thromboplastins of differing species but similar sensitivity (International Sensitivity Index [ISI] = 1.14 for CRB-Thromboplastin, and = 1.10 for Thromborel S) supports the recommendation by the World Health Organization that the intensity of oral anticoagulation be expressed in terms of INR. Whether INRs obtained with thromboplastins of largely differing ISIs show sufficient conformity will need to be assessed by further prospective studies.