Validity of international normalized ratio in expressing prothrombin times in anticoagulated and nonanticoagulated patients

We investigated the effect of four thromboplastins with different International Sensitivity Index (ISI) values on prothrombin times (PTs), expressed as International Normalized Ratios (INRs), of 30 normal plasmas, 30 patients on stable oral anticoagulant (OA) therapy, and 30 patients with prolonged PTs for reasons other than OA therapy. Normal ranges became significantly longer when three thromboplastins of rabbit brain origin were used-the lower the ISI, the wider the normal range. The human placenta-derived thromboplastin had a range similar to the values obtained with the most insensitive rabbit reagent. Significantly different INR values were noted when normal plasmas and plasmas of patients with prolonged PTs not due to OA therapy were tested. This was not the case when specimens from patients on OAs were tested. When PT values from 212 consecutively selected patients were tested with the least sensitive rabbit reagent (ISI 2.477) and the most sensitive human placenta reagent (ISI 1.06), significantly more ''abnormal'' PTs were identified with the high-sensitivity thromboplastin. The data indicate that special attention has to be paid to ''normal'' ranges when thromboplastins with low ISI replace less sensitive reagents, that INR should not be used to express PT in patients not on stable OA therapy, and that with the routine use of highly sensitive reagents (low ISI) more abnormal patients are identified.