Problems with the international normalized ratio.

In a recent letter to Blood, Hirsh called attention to the inadequacy of monitoring warfarin dosage by the prothrombin time assay (PT) as commonly performed in the United States.’ The actual PT value may vary widely from laboratory to laboratory because the result is dependent on the “responsiveness” of the thromboplastin used for the test, and because the thromboplastin reagents distributed commercially by manufacturers in North America display a wide range of “responsi~eness.”~~~ The need to reduce this variability led to the concept of the international normalized ratio (INR), a means of expressing the prothrombin ratio in a form that permits comparison with an international reference ~ t a n d a r d . ~ Yet, even if every laboratory in the United States were to adopt the INR overnight, more work would be needed to achieve practical success. Hirsh mentioned “anecdotal reports of inaccurate calibration by manufacturers.” Such reports are more than anecdotal: at least one recent example has been documented and reported in detail5 Hirsh also cited as major problems the wide range of international sensitivity index (ISI) values of thromboplastins distributed by North American manufacturers and the observation that the system does not work well with certain automatic clot detection systems. In our experience, even in a single laboratory using one reagent, a change of instruments led to INR mismatches.’ This suggests that the existing equation for calculating INR is inadequate for dealing with the real-life variations in materials and methods used in American laboratories. Perhaps the problem can be solved by additional correction factors for these variables in the INR equation itself. Meanwhile, we agree with Hirsh’s recommendations: (1) the range of variation in thromboplastin “responsiveness’’ should be more narrowly restricted by the manufacturers, and “responsiveness” should more closely resemble that of the international reference standard; and (2) an accurate IS1 should be determined by the manufacturer for each anticipated reagent/ instrument combination, including all of the common automated clot detection systems. Until full confidence is established (or restored) in the INR system, it would also be helpful to provide a means of verifying the manufacturer’s IS1 value for each lot of thromboplastin, either through a central regulatory agency or by individual laboratories using a reference thromboplastin.