WO1995012127A1

WO1995012127A1 - Calibration of prothrombin time test systems

Info

Publication number: WO1995012127A1
Application number: PCT/GB1994/002377
Authority: WO
Inventors: Leon Poller
Original assignee: Leon Poller
Priority date: 1993-10-29
Filing date: 1994-10-28
Publication date: 1995-05-04
Also published as: AU8000294A; GB9322316D0

Abstract

A method of calibrating local PT test systems in hospitals and clinics involves use of a set of plasma calibrants comprising a series of lyophilized plasma samples each having a PT value assigned with respect to a selected IRP for thromboplastin. The samples may be artificially depleted in coagulation factors and tested to ensure that they provide a range of PT values over the therapeutic range. The plasma calibrants may dispense with the need for the demanding schedule of conventional fresh plasma ISI calibrations using blood samples from normal subjects and from patients stabilised on oral anticoagulant treatment. The invention concerns manufacture and use of the plasma calibrants.

Description

CALIBRATION OF PROTHROMBIN TIME TEST SYSTEMS

Prothrombin time (PT) tests are used, in particular, for monitoring oral anticoagulant treatment. The present invention provides a method of calibrating local PT test systems in hospitals and clinics to conform to the internationally recommended system of measurement. The invention also relates to a set of plasma calibrants for use in a calibration method and to the manufacture of such calibrants. The use of the sets of plasma calibrants could be an essential ingredient in the safety of oral anticoagulant therapy and of its inter-hospital and clinic standardization in years to come.

Oral anticoagulant treatment with warfarin and related drugs reduces coumarin (warfarin) dependent clotting factors II, VII, IX and X. The success and safety of oral anticoagulant treatment is dependent on regular and appropriate laboratory monitoring by the PT test which measures depression of three of these (clotting factors II, VII and X).

In addition to its application in monitoring oral anticoagulant treatment, the PT test is also a useful indication of liver disease and other disturbances of the extrinsic blood clotting system.

The PT test consists of the recalcification of decalcified plasma in the presence of a powerful tissue extract procoagulant (thromboplastin). The PT is measured by adding a thromboplastin reagent (which is an extract of mammalian tissue rich in tissue factor, a similar preparation of human tissue or recombinant human tissue factor with phospholipids) to citrated plasma and recording the time for clotting to occur after recalcification. The latter may be detected in the manual technique (by eye) or automatically by use of a coagulometer. The tissue extract thromboplastin reagent is an important variable in the test system, although other effects including coagulometers also contribute to variability in results of PT tests at different centres. The need for the use of the Internationally recommended system of PT standardisation arises from the differences in PT techniques and PT results in different laboratories. Owing to these variations between results, the amount of anticoagulant drug prescribed to achieve the therapeutic prolongation of the PT test differs greatly from centre to centre. In view of the variability of PT results between centres, and the need to standardize the PT to provide a uniform basis for clinical dosage with warfarin and related drugs, there have been a number of attempts over the last 30 years to provide a reliable scheme to calibrate PT tests at individual centres and thereby give a common scale for clinical dosage. The first attempt at standardization was the Manchester Regional Thromboplastin Scheme in the UK in the early 1960s with the provision of standardized routine thromboplastin Manchester Comparative Reagent to be used in all hospitals over a large geographical area (Poller 1964). This was succeeded in the UK by the establishment of an official national thromboplastin reagent, which was an independently monitored batch of Manchester Comparative Reagent (MCR) British Comparative Thromboplastin (BCT), first used as a reference material and ultimately as a nationwide routine reagent for the UK (Poller 1970). This national scheme was a unique one but regrettably had to be withdrawn in 1986 with the discontinuation of BCT because of possible risk of virus contamination of human brain material.

In the interim a variety of other attempts were made to provide international standardization of the PT test by evolving reference preparations to calibrate local and commercial thromboplastins. An early attempt was to provide lyophilized (freeze-dried) "standard" reference plasmas calibrated with designated "prothrombin ratios", i.e. prothrombin ratio = PT of patient ^■-^■ PT of normal. This initiative began in the United States in the late 1960s and early 1970s (Miale and La Fond 1967, 1969, and Miale and Kent 1972). The "plasma standards" consisted of a set of two commercial control plasmas which were artificially depleted of warfarin dependent clotting factors and a "normal" plasma. However, as the plasmas were only calibrated in terms of one commercial reagent they proved unsuccessful for standardization of other thromboplastins. Thus, the proposed scheme was not accepte .

Attention was then concentrated on the alternative development of reference thromboplastins and their evaluation was described in a series of reports in the late 1970s and early 1980s. An International Reference Preparation (IRP) for thromboplastin had been adopted by WHO (WHO Technical Series 1977). However, the method of standardization originally used with it and termed an "International Standardized Ratio" (Biggs and Denson 1967) proved to be unreliable (Kirkwood 1983).

Some years later after international collaborative studies (Loeliger et al 1981, Hermans et al 1983, Kirkwood 1983), WHO introduced the present approved international scheme for PT standardization. This is based on calibration of local thromboplastins against IRP of human, rabbit or bovine origin to match the species of the local reagent with an approved manual technique and statistical method (WHO Technical Series 1983). The calibration scheme is based on an International Sensitivity Index (ISI) for thromboplastins. The ISI of the first primary IRP was arbitrarily allocated the figure of 1.0 on the scale. The ISI of other thromboplastins are then related to this, on a common scale. The ISI is inversely proportional to the responsiveness to the warfarin-dependent clotting factors, ie. a lower ISI equates to a more responsive thromboplastin.

From the ISI, INR is derived as follows once the ISI of a thromboplastin is established:

PT of patient divided by the normal mean PT (the geometric mean PT (sec) of a minimum of 20 healthy adults) gives the local prothrombin ratio (PR) which is raised to the power of the ISI of the local (in-house) thromboplastin reagent to give the INR

ie. INR = PR^IS1 wherein

PR = PT(patient) PT (normal mean)

According to WHO guidelines, the ISI of a local routine thromboplastin reagent is derived by testing it in parallel with the IRP on individual fresh plasma samples from 60 long-term stabilized patients on oral anticoagulant treatment and fresh plasmas from 20 normal subjects using the approved manual PT techniques as described by Ingram and Hills (1976) i.e. a hand tilt tube method with the observers visual determination of the end point as the moment of the formation of the fibrin web or clot. The result is recorded as a PT in seconds. The ISI of the local reagent is subsequently calculated using the orthogonal regression equation (described later).

The main problem, which has limited the application of the INR system, has been that the recommended WHO procedure (as detailed later) is dependent on the measurement of the PT using this recommended manual PT technique with the IRP. However, since the development of the WHO scheme, circumstances have changed owing to the prevalent use of coagulometers for automatic PT end- point detection. The manual PT technique has therefore been largely discarded and with it the possibility of local ISI calibration in the WHO recommended way.

Local ISI calibrations have, however, become even more important with the development of new types of coagulometers which are far removed from the manual technique. Furthermore different coagulometers, even those of the same manufacture, have been shown to markedly alter the ISI of thromboplastins unpredictably. Thus erroneous ISI values for "in-house" and manufacturers' production batches of thromboplastins hr -3 been obtained (Bussey et al 1992). By the same token the need for local calibration has therefore become even more imperative.

The manufacturers of thromboplastin have also experienced difficulty in fulfilling their obligations under the WHO scheme to calibrate their reagents with the true ISI value as they likewise no longer perform the manual technique routinely. The use of a "system ISI" i.e. for a thromboplastin/coagulometer combination for all instruments of the same brand with the same thromboplastin has been proposed but has not been shown to be as reliable as the local ISI calibrations with the plasma calibrants of the present invention.

The use of lyophilised plasma samples for monitoring local technique of performance of the PT test was initiated by national external quality assessment studies initiated by the applicant in the late 1960's and continued under the applicant's direction for the following 21 years. Lyophilized plasmas were used also for the WHO International External Quality Assessment Scheme in blood coagulation also initiated and organized by the applicant over a similar period. In separate developments, PT proficiency surveys using lyophilized plasmas were conducted in the USA under the College of American Pathologists and from the Netherlands.

The effects of coagulometers on the manually determined INR of lyophilized plasmas issued in different surveys in the UK National External Quality Assessment Scheme has also been reported by the present applicant.

In a subsequent study of the two most widely used coagulometers in the UK in 1990 we showed that both artificially depleted lyophilized plasmas and those lyophilized from warfarin treated patients could be used to assess coagulometer effects on the PT and INR when a single type of thromboplastin (Manchester Reagent) was used (Thomson, Taberner and Poller 1990). The results with the coagulometers at centres using them were compared with INR obtained centrally using the manual technique with the same reagent (Manchester Reagent). It was observed that there were marked coagulometer effects and considerable differences between instruments of the same type in effects on INR.

In a further study from our centre (Clarke et al 1992) the pooled results of three consecutive national surveys of Manchester Reagent users were used to try to correct for coagulometer effects on INR. A local system calibration based on the pooled data from 26 lyophilized plasma samples in the three surveys was performed to obtain an orthogonal regression slope for each local instrument against reference PT values with Manchester Reagent obtained with the manual technique at our centre. Local normal values were obtained to calculate the local system INR using the stated manual ISI of Manchester Reagent. Better correction of INR was obtained with the local calibration than using the consensus ISI from all users of the same instrument demonstrating the importance of local system ISI determination.

In a poster presentation at a meeting in 1993 published in the abstracts of the meeting Houbayan and Goguel (1993) described the correction of INR of a test sample in French national proficiency surveys using two abnormal lyophilized plasmas with established INR. The interlaboratory dispersion of INR of the test sample was reduced by the calibration using two lyophilized warfarin (AVK) plasmas but not with two artificially depleted plasmas. In this case the use of artificially depleted plasmas was not recommended.

The present invention recognises and seeks to alleviate at least some of the aforementioned difficulties by the provision of lyophilized plasma calibrants with certified PT values in terms of the IRP, suitably using the recommended manual PT technique. This procedure facilitates the implementation of ISI calibration and allows a reliable local thromboplastin/coagulometer system ISI and hence the correct INR to be determined. The scheme proposed combines the benefits of thromboplastin ISI calibration with the additional safeguards of lyophilized plasma controls certified in terms of IRP.

The present invention aims to provide a method of calibrating local PT systems so as to provide reliable local system ISI to derive INR accurately for monitoring oral anticoagulant treatment. Accordingly, in one broad aspect the present invention provides a method of calibrating prothrombin time (PT) test systems which involves use of a set of lyophilised plasma samples each having a PT value assigned in terms of the relevant WHO IRP for thromboplastin. The PT value of each plasma sample with respect to an IRP is recorded in seconds and compared with the results in seconds of the PT for the respective samples using the local PT system.

Suitably the samples comprise normal human blood plasma depleted in coagulation factors prior to lyophilisation. It is convenient to provide lyophilized (freeze-dried) plasmas as fresh plasmas are very unstable in clotting factor content as some clotting factors rapidly reduce in storage at ambient and even low temperatures and other clotting factors are artificially increased in potency on storage. Fresh plasmas, if not snap-deep frozen and distributed refrigerated, are generally not suitable for transporting between laboratories or anything but immediate testing within a few hours of collection. Deep frozen plasmas are relatively unstable compared with lyophilised plasmas, unless maintained at very low temperatures (-40°C to - 80°C) and are therefore unsuitable, other than on a small scale.

Prior to the present invention, we appreciated the possibility of using lyophilized plasma samples of known INR to calibrate local coagulometer systems, however the method of reliably doing so across the range of thromboplastin/ coagulometer combinations had not been developed.

The present invention is such a procedure. It relates to a plurality of plasma calibrants which may be artificially depleted to provide sufficient volumes, e.g. test samples for large scale use in local laboratories in different countries to provide an adequate spread and distribution of INR across the therapeutic range. This is not usually possible with warfarin treated patient's samples or their "pools". The invention may also limit the possible risks of infection from HIV and hepatitis transmittable viruses in patients' samples.

In preferred embodiments the method of certification, of assigning the PT value involves multiple replicates using the manual technique by a plurality of operators to obtain a reliable mean value and the benefits of multiple certification not usually available in local laboratories.

The need to certify these plasmas in terms of the different types of thromboplastin IRP (rabbit, bovine and human) was not appreciated previously but has been demonstrated to be important in recent international study with these calibrants.

In another aspect the present invention provides a control set of calibrants for use in calibrating prothrombin time test systems, the set comprising a plurality of lyophilised plasma samples, each respective sample having a PT value assigned with respect to a selected IRP for thromboplastin. Suitably the set comprises plasma samples having a range of PT values to give, as outlined later, a broad spread and even scatter across the desired therapeutic interval employed in clinical practice for anticoagulant treatment. Preferably the set comprises plasma samples selectively depleted in coagulation factors so as to provide this range of PT values. For this purpose in preferred embodiments the calibration set includes plasma samples having various PT values within the established therapeutic interval of from 2.0 to 4.5 INR. Preferably the samples have PT values evenly distributed over the stated therapeutic interval. Suitably between about 70%, and about 85% of the calibrants in the set have an INR of from 2.5 to 4.5. Preferably between about 40% and 60%, most preferably between about 45% and 55% of the samples have an INR of from 2.5 to 4.0. In accordance with particularly preferred embodiments the calibration set includes samples having various PT values to provide an INR distribution as follows:

INR distribution Possible Number Most preferred of samples n u m b e r o f samples

1. 5 • - 2.0 1 - 3 2 2 2..00 - - - 22.. 55 2 - 4 4

2. 5 - - 3.0 2 - 4 4

3.0 - - 3.5 2 - 4 3

3. 5 - - 4.0 2 - 4 3

4.0 - - 4. 5 1 - 3 3 4 4..55 - - - 55..00 1 - 2 1

For these purposes a sample is considered within the relevant INR range if it has an INR greater than that at the lower end of the specified range.

Preferably the set of calibrants includes at least about ten plasma samples, suitably at least about fifteen samples and most preferably about twenty samples.

In this regard it has been found that particularly accurate and reliable results are obtainable by using about twenty plasma samples within the above preferred INR distribution. The optimum number of plasma calibrants has been derived on the basis of the results of an international collaborative survey involving 95 test systems in 37 laboratories and has been found to be 20 on the basis of multicentre clinical trial.

The present invention provides for first time a valid system for local ISI calibration across the range of available thromboplastin/coagulometer calibrants (system).

In the same study it has been shown that artificially depleted plasmas can be used as dependably as lyophilized warfarin plasmas in local systems ISI calibration thus refuting previous misconceptions (e.g.

Houbayan and Goguel 1993).

In another broad aspect the present invention provides a method of preparing a set of calibrants for use in calibrating prothrombin time test systems, which method comprises providing a plurality of lyophilised plasma samples and assigning a PT value to each respective sample with respect to a selected IRP. Preferably the method also comprises selectively depleting the plasma samples of coagulation factors prior to lyophilisation. In particularly preferred embodiments the method involves selection of a number of samples having a spread of PT values within the therapeutic range of 2.0 to 4.5 INR, such as described above. The PT values with respect to the selected IRP are assigned by the manual technique.

In yet another broad aspect the present invention provides a method of calibrating local PT test time systems utilising a control calibration set as described above, which method comprises determining the PT time of plasma samples using the local system, comparing the local PT times with the PT values assigned with respect to a selected IRP and determining the local ISI for the test system.

In preferred embodiments the method also involves determining local PT times for normal samples and the calculation of the local ISI for the test systems is modified to take into account the PT time for normal samples.

The utility of inclusion of the local normal value in the local system ISI calibration using the lyophilized plasmas had not been previously appreciated but is now appreciated to be of importance when comparing different thromboplastin coagulometer interactions in local systems. The inclusion of the mean normal prothrombin time (MNPT) (geometric mean of 20 normals) is a new step in this type of calibration as 20 individual results are included in the fresh plasma calibrations in the WHO procedure. The best balance of normal and abnormal samples has also been derived from survey results. The formula developed is a new one with the MNPT being entered twice with the 20 abnormal samples.

Embodiments of the present invention have various aims and objects.

The provision of control calibrant plasmas in accordance with the present invention may allow world¬ wide implementation of the WHO recommended system of prothrombin time standardization at the vast majority of centres now using automated or semi-automated end-point detection methods (coagulometers) in the PT test. It circumvents the requirement for the local performance of the PT test with the thromboplastin IRP using the obsolete or obsolescent (depending on geographical location) manual PT technique. This was previously mandatory for the calibration of local thromboplastin reagents on the agreed common scale, to provide an International Sensitivity Index (ISI). Local ISI determination is now more than ever necessary because the use of automated end-point detection methods (coagulometers), and other non-traditional PT techniques, eg. chromogenic substrates, has a profound and unpredictable effect on the ISI of thromboplastins.

The value of a single centre ISI calibration performed according to WHO procedure on 60 fresh plasmas from coumarin treated patients and 20 normal subjects may be suspect because of between laboratory variation in technique. As a result multicentre collaborative calibrants involving several expert centres have been performed for important thromboplastin reagents e.g. the IRP and some manufacturers' commercial products. The plasma calibrant sets offers some of the advantages of a multicentre calibration because the certified PT values of each of the calibrant plasmas are based on the mean value of a plurality of independent observers (at least 8).

Previously logistics of supply of fresh plasmas for calibration made local ISI calibration a considerably time consuming, elaborate, difficult and expensive procedure. The present control plasma calibrants substitute for these fresh plasmas and provide a simplified, rapid local ISI calibration procedure for each laboratory's PT system (thromboplastin/coagulometer combination). They thus allow conformity to the ISI scale of calibration of thromboplastins and provision of accurate INR for individual patients. Use of the sets of calibrants of the present invention dispenses with the heavy demands on fresh plasmas for conventional thromboplastin calibration, ie. minimum of 60 fresh plasmas from long-term stabilized patients on oral anticoagulation. Such plasmas may not be readily available in some laboratories, particularly for manufacturers who have a responsibility to assign an ISI value to their products.

The provision of the present calibrants also dispenses with the need for local supplies of thromboplastin IRP at individual laboratories. Batches of primary IRP of human, rabbit and bovine tissue origin are held by WHO but are for the use of designated national control laboratories only. A batch of the human IRP is still available in limited amounts for developing countries producing their own national standards. Batches of secondary IRP are held by the Bureau Commun de Reference of the European Community in Brussels. These are strictly limited in batch size from 8,000 to 10,000 ampoules per batch. 20 to 30 vials are required for one single local calibration of a PT system. Clearly, these batches of IRP would be totally insufficient for all laboratories to test their local systems on one occasion only, particularly as some laboratories have a multiplicity of coagulometer PT systems for which ISI should be calibrated.

This was appreciated when the ISI system was established and the need for development of national reference thromboplastins to support the IRP was stressed (WHO 1983). There are currently no such national reference thromboplastins available. Use of the calibration plasmas of the present invention may resolve this problem.

Embodiments of the present invention may provide the following benefits to patients -

Greater between-centre standardization of PT measurement and oral coagulant dosage by provision of correct ISI and INR. Increased patient security from under or overdosage of oral anticoagulants.

Benefits of local application of INR equivalents from clinical trials at other centres resulting from the correct implementation of the INR/ISI system. Increased precision and accuracy of PT measurement from the implementation of the ISI/INR system via the calibrants.

Greater within-centre standardization of the above in laboratories with more than one coagulometer system or in the substantial minority where technologists still practise the manual PT technique.

Embodiments of the present invention may also provide the following benefits to hospitals -

Considerable reduction in costs of conventional calibrations i.e. costs of purchase of IRP, collection and testing of the 60 patients' blood samples. Provision of inter- and intra-laboratory quality control by comparison of ISI values from results with the calibrant sets between operators (technologists) and between PT systems. Removal of source of transmission of infection from fresh donor samples (HIV, Hepatitis, etc.). It is not the usual practice to screen the 60 fresh donor plasmas used in conventional calibrations for the above as the labour and costs would be prohibitive and prevent testing of fresh donations. Facilitation of provision of lower dose anticoagulation thus increasing patient safety via the introduction of INR control as a result of the correct implementation of the INR/ISI system.

Embodiments of the present invention may also provide the following research benefits -

Safer and more effective therapeutic ranges from the application of ranges in INR advocated by expert groups, eg. Hirsh et al 1992.

Availability of the calibrants will provide a common INR basis for multicentre clinical trials of oral anticoagulation.

Embodiments of the present invention will now be described further, with reference to the accompanying examples:

EXAMPLES

Five to twenty human plasmas each artificially depleted in coagulation factors by established chemical procedures are prepared. (This procedure is discussed in more detail below). They are tested prior to freeze- drying to ensure that they provide an adequate span over the therapeutic range for oral anticoagulation with the IRP (2.0 to 4.5 INR) and are not deficient in other non- coumarin dependent clotting factors which affect the PT e.g. fibrinogen and factor V. The plasmas are then carefully selected after freeze-drying so that the set give not only a wide span but a good spread of values over the therapeutic range. It has been shown that clustering of PT values within this scale leads to less accurate and less precise ISI calibrations.

The lyophilized plasmas are reconstituted with the requisite volume of distilled water and each tested in quadruplicate in the PT procedure in the local test system as opposed to single tests in conventional fresh plasma calibration. Mean results of the 5 to 20 plasmas are recorded and plotted individually on a log/log scale against the results of the certified manual technique PT values of the same plasmas tested with the relevant IRP. The results of the mean normal prothrombin time (MNP T- geometric mean of 20 normal subject PT) are entered twice into the calibration calculation in place of 20 individual PT results in the WHO recommended procedure. Our studies have shown that this proportion of normal to abnormal PT gives the most reliable ISI. From the orthogonal regression analysis the ISI is calculated according to the conventional formula for ISI calibration. From the resulting slope of the orthogonal regression calibration line (b), the ISI of the local system is calculated, ie. local ISI = b x ISI of the IRP. The local ISI for each local system can then be applied to results obtained from testing patients' samples and the correct INR derived.

The effectiveness and safety of oral anticoagulant treatment can therefore be achieved more reliably and therapeutic ranges in INR for various clinical thrombotic disorders established elsewhere by clinical trials can be applied at the local level.

A procedure according to the recommended method of applying the new development can now be described by way of example.

Preparation of plasma calibrants, their certification and testing

A minimum of 5 and up to 20 aliquots of normal blood plasma are taken from donations from subjects who are negative on HIV, HbC and HBsAg testing and free from all other pathogens. The donations may be single or pooled but the original constituent samples should have a normal PT, Activated Partial Thromboplastin Time test, factor V assay and fibrinogen result prior to depletion. These samples are depleted of the vitamin K-dependent clotting factors to predetermined levels to give a set of plasmas each with a different INR level, representative of the various levels of treatment in the broad therapeutic range of 2.0 to 4.5 INR with the IRP of known ISI.

The plasmas are then lyophilized in accordance with established procedures. The procedure incorporates a primary drying cycle of the plasma cooled to -60°C initially. This lasts one day and is followed by a secondary drying period of two days. The lyophilization ensures the stability of the clotting factors of the plasmas. After completion of freeze-drying the plasmas are then tested using the manual technique with IRP for thromboplastin calibrated in terms of this IRP to assess their INR. Long-term stability testing and accelerated stability testing at 4°C, 26°C, 37°C and 42°C are then conducted to ensure adequate long-term stability. Tests are performed daily on vials of plasma at the highest temperature until the PT prolongs by one second. The high temperature accelerated tests provide a guide to any possible future lack of stability at lower temperatures of storage and are needed because lyophilized plasma samples are normally stable for at least one year when stored at 4°C and at lower temperatures.

The certified values for the individual plasmas in the set are derived from the manual testing with the relevant IRP (human plain, rabbit plain or bovine combined) by at least eight experienced laboratory staff. Each plasma is tested in quadruplicate and the mean results for all eight investigators recorded in PT seconds. The overall mean PT result is the certified value. The sets as constituted are then distributed to hospital laboratories with instruction sheets for the performance of the calibration calculation. The certified values are then used with the normal mean values and plotted on double logarithmic graph paper with the reference preparation on the y-axis and the test results on the X-axis to derive the calibration slope of individual local thromboplastin/coagulometer systems and hence ISI. Local system ISI = slope X ISI of IRP.

Local Calibration Procedure

The set of 5 to 20 certified plasma calibrants is stored at -20°C between preparation and use and is stable under these conditions for several years. 0.5ml of sterile distilled water is used to reconstitute each vial and it is mixed gently with the content in situ. The reconstituted plasma is left at room temperature for at least 10 minutes. Each plasma is then tested in at least four replicate tests using the local coagulometer/ thromboplastin system and the mean PT value for each plasma is determined. If a discrepancy between replicates is more than 5% an aberrant result sh ld be excluded.

Recent national and international surveys with these sets of calibrants have indicated that in the great majority of laboratories the use of the above sets of lyophilized calibrants provides adequate precision of ISI calibration according to WHO criteria. These should be for a maximum of 5% CV of the slope with an optimum of 3% or less.

The local normal PT value is also included in duplicate in the calibration and in the calculation of the orthogonal regression slope as this may influence the slope of the line and should not be ignored. The local normal mean is derived with the customary local PT systems from the geometric mean of at least 20 healthy adults at each centre. An example calculation is as follows:

Calculation of Local System I.S.I.

As in current techniques, the local system ISI may be determined using the orthogonal regression equation m, as follows, for an example procedure of the present invention employing 20 plasmas. The log time of the coagulometer PT results for the 20 plasmas are summated, and the result (Σ x) recorded, where x = Log PT). Square this result and divide by n=20. (n= the number of plasmas).

Record the result (Σ x )² n

Repeat this procedure for the log times of the manual results, recording the values ∑y and (∑y )² where y is n log PT, and n=20.

The sums of the squares ∑x² and ∑y₂ are calculated, and the sum of the product for each measurement and each manual result (∑xy.

A value m is calculated according to the formula: ⁽∑x⁾² - (∑y_i² - ∑^χ + ∑y² m = n n

2 (∑xy - ∑x∑v j n

The instrument slope (b) = m + /m²+l

The ISI of the instrument is the product of b calculated above and the manual ISI.

We have found that improved results are obtainable by modifying the calculation of m to take into account the PT results of local normal plasma samples. The mean normal PT (geometric mean of the, say, 20 normals) is introduced as a sample into calculation of m. Best results are obtained by using the MNPT in duplicate, so that n in the above formula is taken as the total number of plasma calibrant samples plus two. A worked example is shown in Table. 1.

EXAMPLE OF METHOD FOR CALCULATION OF ISI - TABLE 1

IRP - CRM 149R ROUTINE REAGENT

No. Time(y) lny (lny)² (lnx)(lny) (lnx)² lnx Time(x)

1 32.30 1.50920 2.27769 1.93676 1.64686 1.28330 19.20

2 41.40 1.61700 2.61469 2.18021 1.81793 1.34830 22.30

3 56.63 1.75305 3.07317 2.56888 2.14735 1.46538 29.20

4 29.45 1.46909 2.15821 1.87523 1.62935 1.27646 18.90

5 37.75 1.57692 2.48667 2.09152 1.75917 1.32634 21.20

6 30.93 1.49038 2.22123 1.88153 1.59378 1.26245 18.30

7 29.63 1.47173 2.16599 1.82576 1.53896 1.24055 17.40

8 37.70 1.57634 2.48485 2.08100 1.74279 1.32015 20.90

9 56.90 1.75511 3.08042 2.52896 2.07622 1.44091 27.60

10 37.90 1.57864 2.49210 2.11608 1.79679 1.34044 21.90

11 42.65 1.62992 2.65664 2.23171 1.87475 1.36922 23.40

12 36.93 1.56738 2.45668 2.09160 1.78077 1.33445 21.60

13 42.40 1.62737 2.64832 2.21910 1.85944 1.36361 23.10

14 38.78 1.58861 2.52367 2.15419 1.83881 1.35603 22.70

15 38.00 1.57978 2.49572 2.07229 1.72070 1.31175 20.50

16 57.25 1.75778 3.08977 2.53831 2.08527 1.44404 27.80

17 46.38 1.66633 2.77666 2.34659 1.98314 1.40824 25.60

18 43.83 1.64177 2.69541 2.30360 1.96875 1.40312 25.30

19 58.55 1.76753 3.12415 2.57148 2.11657 1.45484 28.50

20 50.48 1.70312 2.90062 2.46204 2.08977 1.44560 27.90

21 16.98 1.22994 1.51275 1.35084 1.20626 1.09830 12.54

22 16.98 1.22994 1.51275 1.35084 1.20626 1.09830 12.54 n=22 ∑y=34.7869 ∑y²=55.4482 ∑xy=46.7785 ∑x²=39.4797 ∑x= 29.3918 CALCULATE ISI USING THE ORTHOGONAL REGRESSION EQUATION m:

(∑x)² - (∑V)² _ ∑x² + ∑y² n n m =

∑x∑y

2(∑xy - n

(29.3918)² (34.7869)^:

- 39.4797 + 55.4482

22 22 m =

( 29.3918 ) ( 34.786 )

2( 46.7785

22 0.2297915 m =

0.6070272 m = 0.3785523 slope = m + /m² + 1 slope = 1.4478052 system ISI = 1.4478052 x IRP ISI system ISI = 1.4478052 x 1.348 system ISI = 1.95164 By way of illustration, Table 2 shows means and standard deviations of slope values (b) obtained in calibrations using twenty plasma calibrants for thromboplastin reagents OBT (rabbit brain) and Recombinastin (human) using reference preparations CRM149R and BCT441 respectively. The table shows the mean results obtained from thirty-seven centres using 20 calibrants. A statistical analysis was performed to determine the effect of reducing the number of calibrants on the slope calculation. In accordance with the paired- slope calculation the p value is significant, (less then 5%) when using eighteen or less samples. On the basis of these p values it is concluded that at least about eighteen, and especially about twenty samples, is preferable for the calibration set for these thromboplastin reagents.

TABLE 2

Calibration line slopes and standard deviations. Log PT obtained for OBT using CRM 149R as reference preparation and Log PT for RecombiPlas Tin obtained using BCT441

OBT

No. of Slope Calibrants Mean SD P-values

20 1.30718 0.04783

19 1.30994 0.04762 0.21906

18 1.31326 0.04951 0.03401

17 1.31494 0.04903 0.00648

16 1.31919 0.05273 0.00830

15 1.32101 0.05462 0.00439

14 1.33501 0.05625 0.00004

13 1.34221 0.06006 0.00015

12 1.34413 0.06299 0.00017

11 1.34330 0.06328 0.00029

10 1.34729 0.06845 0.00015

9 1.34485 0.06533 0.00029

8 1.35319 0.07113 0.00002

7 1.35294 0.07203 0.00001

6 1.37325 0.08400 0.00000

5 1.32365 0.07025 0.02510

RecombiPlastin

No. of Slope Calibrants Mean SD P-values

20 0.86931 0.03695

19 0.86902 0.03786 0.54909

18 0.86860 0.03977 0.33777

17 0.86862 0.04075 0.48171

16 0.88139 0.04100 0.00008

15 0.88164 0.04259 0.00002

14 0.87986 0.04535 0.00048

13 0.88544 0.04825 0.00033

12 0.87141 0.04753 0.57796

11 0.87152 0.05008 0.48219

10 0.87315 0.05218 0.27903

9 0.87788 0.04595 0.02367

8 0.87668 0.04742 0.04329

7 0.88232 0.04075 0.00077

6 0.87886 0.04967 0.00681

5 0.86462 0.05807 0.19296 References

Biggs R, Denson KWE. 1967. Standardization of the one- stage prothrombin time for the control of oral anticoagulation ther .y. Brit Med J. 1:84-87.

Bussey HI, Force RW, Bianco TM, Leonard A. 1992. Reliance on prothrombin ratios causes significant errors in anticoagulation. Arch Int Med. 152:278-81.

Clarke K, Taberner DA, Thomson JM, Morris JA, Poller L. 1992. Assessment of value of calibrated lyophilized plasmas to determine International Sensitivity Index for coagulometer. Journal of Clinical Pathology 45:58-60.

Hermans J, van den Besselaar AMHP, Loeliger A, van der Velda EA. 1983. A collaborative study of reference materials for thromboplastins. Thromb Haemostas 50:712-7.

Hirsh J, Dalen JE, Deykin D, Poller L. Oral anticoagulants: Mechanism of action, clinical effectiveness and optimal therapeutic range. Chest. 1992 102(4):312S-326S

Ingram GIC, Hills M. 1979. Reference method for the one stage prothrombin time test on human blood. Thrombos Haemostas. 36:237-8.

Kirkwood TBL. 1983. Calibration of reference thromboplastin and standardization of the prothrombin time ratio. Thromb Haemostas. 49:238-244.

Loeliger EA, van den Beselaar AMHP, Hermans J, van der Velda EA. 1981. Certification of three reference materials for thromboplastin. Commission of the European Communities, Brussels.

Miale JB and La Fond DJ. 1967. Amer J Clin Path 47:40-45.

Miale JB and La Fond DJ. 1969. Thrombos et Diathes Haemorrh Suppl 35:125-129.

Miale JB and Kent JW. 1972. Amer J Clin Path. 57:80

Poller L. 1964. The standardization of anticoagulant treatment: Manchester Regional thromboplastin scheme. British Med Journal. 2:565-566

Poller L. 1970. ACP Broadsheet 71: British Comparative Thromboplastin: The use of the national thromboplastin for the uniformity of laboratory control of oral anticoagulants and expression of results.

Poller L, Thomson JM, Taberner DA, Clarke DK. 1994 a. The correction of coagulometer effects on international normalized ratios: a multicentre evaluation. British Journal of Haematology 86:112-117.

WHO Expert Committee on Biological Standardization. 1977. 28th Report WHO Technical Report Series. 610, Geneva.

WHO Expert Committee on Biological Standardization. 1983. 33rd Report WHO Technical Report Series 687:81-105, Geneva.

Thomson JM, Taberner DA, Poller L. Automation and the prothrombin time. A UK field study of two widely used coagulometers. J. Clin Pathol. 1990. 43:679-684.

Clarke K, Taberner DA, Thomson JM, Morris JA and Poller L. The assessment of the value of calibrated lyophilized plasmas to determine ISI. J. Clin Pathol. 1992. 45:58-60

Houbayan LL, Goguel AF. Procedure of reference calibrated plasmas for prothrombin time standardization. Thrombos Haemostas 1993. 69:663.

Claims

1. A method of preparing a set of calibrants for use in calibrating local prothrombin time test systems, the method comprising providing a plurality of lyophilised plasma samples and determining a PT value for each sample with respect to a selected IRP for thromboplastin.

2. A method according to Claim 1 which includes lyophilising the samples.

3. A method according to Claim 1 or 2 which comprises selectively depleting at least some of the plasma samples of coagulation factors prior to lyophilisation.

4. A method according to any preceding claim which includes determining INR values of the plasma samples and selecting at least some plasma samples having INR values within the therapeutic range 2.0 - 4.5.

5. A method according to Claim 4 which comprises providing samples having the following INR distribution:

INR distribution No. of samples

1.5 - 2.0 1 - 3 2.0 - 2.5 2 - 4

2.5 - 3.0 2 - 4

3.0 - 3.5 2 - 4

3.5 - 4.0 2 - 4

4.0 - 4.5 1 - 3 4.5 - 5.0 1 - 2

6. A method according to Claim 5 which comprises providing samples having the following INR distribution:

INR distribution No. of samples 1.5 - 2.0 2

2.0 - 2.5 4

2.5 - 3.0 4

3.0 - 3.5 3

3.5 - 4.0 3 4.0 - 4.5 3

4.5 - 5.0 1

7. A method according to any preceding claim which comprises performing a plurality of PT tests on the plasma samples using the manual technique with a selected IRP, determining the mean results of the PT tests and assigning the mean results as the PT values for the respective samples.

8. A set of calibrants for use in calibrating prothrombin time test systems, the set comprising a plurality of lyophilised plasma samples, each respective sample having a PT value assigned with respect to a selected IRP for thromboplastin.

9. A set according to Claim 8 comprising plasma samples having a range of PT values.

10. A set according to Claim 8 or 9 comprising plasma samples selectively depleted in coagulation factors to provide a range of PT values.

11. A set according to any one of claims 8 to 10 including at least some plasma samples having INR values within the range of 2.0 - 4.5 INR.

12. A set according to Claim 11 wherein at least 70% of the plasma samples have INR values within the range of 2.0 - 4.5 INR.

13. A set according to Claim 12 wherein no more than 70% of the plasma samples have INR values within the range of 2.5 - 4.0 INR.

14. A set according to Claim 11 wherein no more than 45% of the plasma samples have INR values within the range of

2.5 - 3.5 INR.

15. A set according to any one of claims 8 to 14 including at least fifteen plasma samples.

16. A set according to Claim 15 including between fifteen and twenty five plasma samples.

17. A set according to Claim 16 including twenty samples.

18. A method of calibrating local prothrombin time test systems utilising a set of calibrants as claimed in any one of claims 8 to 17 which method comprises determining the local PT time of the plasma samples using the local system, comparing the local PT time with the PT value assigned with respect to a selected IRP and determining the local ISI for the local system.

19. A method according to Claim 18 wherein the geometric mean PT time of a plurality of local normal samples is determined and the local system ISI is determined using orthogonal regression analysis of local PT times of the calibrants, PT values assigned to the calibrant with respect to a selected IRP, and the geometric mean PT time of local normal samples.

20. A method according to Claim 19 wherein the geometric mean PT time of local normal samples is included in duplicate in the orthogonal regression analysis.