INRs and point of care testing

The development and adoption of instruments for determining the international normalised ratio (INR) at the point of care for monitoring treatment with coumarin anticoagulants has added another variable to the question of the precision of INR measurements. In this issue Poller et al (p 30) show discrepancies between results when two such instruments are used and the best estimate of the true INR.¹ These observations, however, should be seen in the context of the already recognised limitations of the INR system and the potential for using point of care testing to deliver safe and effective anticoagulant therapy efficiently.

Coumarin anticoagulants remain the preferred method of prophylaxis and treatment for thrombosis in a range of clinical situations. Around 1% of the populations of many countries receive such treatment, a proportion boosted by the increasing use of oral anticoagulation as thromboprophylaxis in atrial fibrillation.² Monitoring of the anticoagulant effect of coumarins, however, is essential, and the prothrombin time is used for this purpose. Its standardisation was achieved with the INR system in the early 1980s and guidelines for monitoring treatment using the INR have been adopted in a range of clinical indications.³ Anticoagulant control, however, lacks precision and in most surveys only 50-60% of INR results of patients are in the predetermined therapeutic range.⁴ This is due, in part at least, to dietary variations, interactions with medications and supplements, and incomplete compliance. In addition there is residual imprecision in the INR, even when experienced laboratories employ automated coagulometers. Finally, in the United Kingdom, most coagulation laboratories take part in the National Quality Assessment Scheme. Samples for INR measurement are distributed regularly in this scheme. On average around 5-10% of laboratories report INR results which are more than 15% outwith the consensus median value in any particular survey (Kitchen S, personal communication, 2003).

Point of care instruments facilitate the management of anticoagulants in primary care, and furthermore they introduce the possibility of patients' self management of oral anticoagulant treatment. The finding of a discrepancy between INRs determined by point of care systems and laboratory measurements is important but not surprising. While it should be the stimulus for better standardisation of point of care testing monitors and improved quality assurance procedures, clinical studies suggest already that point of care testing systems can be effectively introduced into clinical practice. For example, in a trial comparing performance of a point of care system in a number of primary care sites to laboratory results, primary care INR estimations were at least as reliable as laboratory estimations.⁵ The degree of discrepancy between point of care test results in the study was similar to that found between laboratory INR test results. Another study compared three point of care systems previously validated in laboratory conditions for use in a non-laboratory setting.⁶ The results showed no significant disagreement between the systems. Again the discrepancies found were comparable to those between laboratory estimations. Such data show that point of care testing systems can be used for warfarin monitoring in primary care.

A difference exists between statistical and clinical significance when considering the INR. Clinical significance is determined by an impact on clinical decision making and for this purpose instruments for point of care testing have been found to be satisfactory.⁷ However, the size of the discrepancies in measuring INR reported by Poller et al are such that dosing decisions could be affected. Therefore it is reassuring that further evidence is beginning to accumulate to suggest that point of care testing systems can be used to help deliver safe anticoagulant therapy. For example, in a recent study of self management by patients, employing a point of care testing instrument of a type studied by Poller et al, 336 tests done over a six month period found no greater incidence of bleeding or thrombosis than the comparison group that received routine care, using a central laboratory for measuring the INR.⁸ Much larger studies will be needed to exclude important differences in clinical outcomes with confidence. However, the evidence so far indicates that point of care testing systems have great promise in both primary care and for self management by individual patients. There are clear quality of life benefits for patients from easier access to management of oral anticoagulant treatment.

Increasingly responsibility for oral anticoagulant management is being devolved to primary care practitioners. This may lead to an increase of service uptake in general practice and increased use of point of care testing. To minimise clinical problems arising from implementation of this technology collaboration between manufacturers, pathology laboratories, and general practice and adherence to a recognised external quality assurance scheme are essential.⁹ Wherever possible steps should be taken to improve the accuracy and precision of instruments, including the measures proposed by Poller et al. Moreover, all practitioners involved in the delivery of oral anticoagulant treatment should remain conscious of the limitations in the accuracy of the current systems for determining the degree of anticoagulation within individual patients.