Failure of the International Normalized Ratio to generate consistent results within a local medical community.

Use of the International Normalized Ratio (INR) has been recommended as a means of standardizing prothrombin time (PT) results for management of oral anticoagulant therapy. During the evaluation of a new lot of thromboplastin reagent, however, INR values were obtained that were inconsistent with results obtained with the prior lot of reagent from the same manufacturer. A local normalized ratio (LNR) was substituted for the INR for the new reagent, based on a calculated local sensitivity index (LSI). Validation of the LSI was performed at the three hospitals in the medical community by testing an identical panel of aliquots from 64 plasmas obtained from patients who were chronically receiving oral anticoagulants. Each test result was classified according to the INR value as low, low therapeutic, high therapeutic, or high. Local normalized ratio values obtained at one hospital in the community were in reasonable agreement with INR determinations in the other two hospital laboratories. Classification mismatches occurred using the INR, however, in 84 of 280 (30%) of the paired samples. Thus, the inability to generate consistent INR values within a local medical community raises serious concern about the reliability of the INR concept in its current form.