Intended for healthcare professionals


Monitoring oral anticoagulation in primary care

BMJ 1996; 312 doi: (Published 08 June 1996) Cite this as: BMJ 1996;312:1431
  1. D A Fitzmaurice,
  2. F D R Hobbs,
  3. J A Murray
  1. Professor Lecturer Department of General Practice, Medical School, University of Birmingham, Birmingham B15 2TT
  2. Consultant haematologist Selly Oak Hospital, Birmingham B29 6JD

    Quality assurance and computerised decision support will help to minimise inconsistencies

    Monitoring treatment with anticoagulants is traditionally done in hospital because it requires laboratory blood tests and expert interpretation of the test results. The growing pressure on general practices to undertake monitoring of warfarin treatment has therefore led to concerns over the lack of resources and expertise in primary care.1 The use of computerised decision support systems in general practices may help with dosing.2 Other problems remain, however, including questions about the reliability of the international normalised ratio, the validity of current reference ranges, and where the test should be performed.

    The purpose of oral anticoagulant treatment is to decrease thrombotic episodes, particularly stroke, by interfering with the haemostatic pathways.3 The intensity of anticoagulant effect required for any particular indication remains controversial.4 There have been problems in standardising test results between different centres at local and international levels. Data showing wide international variation in mean warfarin doses—Hong Kong under 2 mg, North America over 8 mg, and British figures between these extremes5—led to international collaboration to improve testing and reporting for oral anticoagulation. Variation in doses was explained by the different sensitivities of thromboplastins used for testing, with the thromboplastin used in Hong Kong being more sensitive than the one in North America, resulting in lower warfarin doses.6

    The international normalised ratio was established as an attempt to standardise these discrepancies, based on comparison between the responsiveness of a thromboplastin to that of reference thromboplastins from the World Health Organisation.7 This relative responsiveness, called the international sensitivity index (ISI), is calculated from the slope of the regression line of the calibration comparing prothrombin times obtained with the two thromboplastins.7 The international normalised ratio is then calculated as INR = prothrombin ratioISI, where the prothrombin ratio is calculated by testing normal plasma against the patient's plasma in parallel with the local reagent and the reference thromboplastin. Unfortunately, there remains considerable debate about the most appropriate reference standard; several secondary standards are now available from animal origins.7

    Using the international sensitivity index has certainly led to more consistent testing and reporting.8 However, unacceptable differences between tests performed on the same sample remain9 and are particularly evident at the extremes of and outside therapeutic ranges. Differences in test results are attributable to many factors other than the thromboplastin, including operator errors, manual timing of clot formation, and using different clot “thicknesses,” but the principal factor is the use of different thromboplastins, resulting in inaccurate estimation of prothrombin time and errors in determining the international sensitivity index.6 Variation in international normalised ratios within laboratories is minimised by adequate quality assurance procedures, but results from different centres are not directly comparable, which may have important clinical implications. One way of resolving such discrepancy would be to ensure standardisation of thromboplastin reagents, particularly with regard to their derivation and international sensitivity index values, combined with standardisation of equipment for measuring prothrombin time.

    The British Committee for Standards in Haematology produced the currently recommended reference ranges for oral anticoagulation,3 and similar guidelines exist in the United States.10 However, audit of British hospital anticoagulation clinics has shown that these guidelines are not being followed consistently.11

    Unfortunately, current guidelines are largely based on non-randomised and uncontrolled studies.4 The only recommendations supported by more than one randomised controlled study are those for the prophylaxis and treatment of deep vein thrombosis, treatment of pulmonary embolism, and prophylaxis for tissue heart valves.10 Evidence is emerging to support anticoagulation in non-rheumatic atrial fibrillation.12 Although warfarin's role after myocardial infarction is less clear, myocardial infarction has recently been approved as an indication by the United States Food and Drug Administration.13 However, appropriate levels of anticoagulation remain uncertain. There is a trend for lower levels of anticoagulation, in an attempt to reduce bleeding complications while maintaining the benefits of treatment.7 Lower intensities of anticoagulation may reduce the risks while maintaining clinical benefits, especially in patients with proximal deep vein thrombosis.14 The evidence for other indications is not as strong, and long term trials will be needed, especially for atrial fibrillation, to determine the optimum desired therapeutic range.15 16 Evidence based policies on anticoagulation need to consider not only the intended range for international normalised ratio but also the actual ranges achieved, the method of determining the ratio, the thromboplastin used and its international sensitivity, and the mechanism of measuring prothrombin time.

    The advent of near patient testing now allows anticoagulation to be monitored in primary care settings.17 Since there are problems with inter-laboratory comparisons, it is especially important that quality assurance is maintained by training and adherence to traditional laboratory techniques.2 Local pathology departments should probably retain responsibility for training and quality assurance, provision of reagents, and standardisation of equipment, aided by a national external quality assurance scheme. Such a centralised arrangement would minimise the problems of inconsistency between centres. Using computerised decision support systems for dosing and patient recall would also help the safe transfer of anticoagulation monitoring from hospital to primary care.2 Better standardisation of the international normalised ratio system and clearer indications of who would most benefit from oral anticoagulation remain to be resolved.


    1. 1.
    2. 2.
    3. 3.
    4. 4.
    5. 5.
    6. 6.
    7. 7.
    8. 8.
    9. 9.
    10. 10.
    11. 11.
    12. 12.
    13. 13.
    14. 14.
    15. 15.
    16. 16.
    17. 17.