Professional Documents
Culture Documents
Comparison of Quick and Owren Prothrombin Time With Regard To The Harmonisation of The International Normalised Ratio (INR) System PDF
Comparison of Quick and Owren Prothrombin Time With Regard To The Harmonisation of The International Normalised Ratio (INR) System PDF
Juha Horsti*
Laboratory, District Hospital of Valkeakoski, University
Hospital of Tampere, 37600 Valkeakoski Finland
Prothrombin time (PT) is tested mostly to monitor patients on oral anticoagulant treatment. The International Normalised Ratio (INR) was introduced to improve and harmonise PT results and therapeutic range
globally for patient care and the scientific literature.
We studied the Quick PT in 179 patients and the
Owren PT in 137 patients on oral anticoagulant therapy using two different reagents for the two methods
of measuring PT. We assessed the clinical significance
of the INR results obtained by each method using the
two reagents and compared the Quick and Owren
methods.
We conclude that with the Quick method individual
INR results differed from each other too much clinically, while using the Owren method individual INR results were clinically acceptable. Our opinion is that we
should develop the INR system using the Owren PT
method rather than the Quick to improve patient care.
Clin Chem Lab Med 2002; (40):399403
Key words: Prothrombin time (PT); International Normalised Ratio (INR); Oral anticoagulant therapy; Coagulation.
Abbreviations: EDTA, ethylenediaminetetraacetic acid;
INR, International Normalised Ratio; ISI, International
Sensitivity Index; MNPT, mean normal PT; PT, Prothrombin time.
Introduction
The prothrombin time (PT) is a primary coagulation
test used in laboratories to detect preoperative bleeding tendency and to monitor oral anticoagulant therapy
at regular intervals in view of its narrow therapeutic
range.
Two major methods are used globally to measure PT.
Quick and co-workers (1, 2) developed the first method
for PT, and Owren (3) subsequently the Quick
method to overcome its drawbacks. The Owren
reagent (the combined thromboplastin reagent) is the
one used almost exclusively in the Nordic countries,
Benelux and Japan. Because of its different reagent
composition, the Quick technique is sensitive to fibrinogen and coagulation factors II, V, VII and X, while
*E-mail of the corresponding author: :juha.horsti@tays.fi
the Owren technique is affected by deficiencies in factors II, VII and X. Dicoumarol drug is a vitamin K antagonist and antithrombotic treatment reduces the synthesis of coagulation factors in the liver and lengthens
coagulation time in PT.
The aim of using the International Normalised Ratio
(INR) is the worldwide harmonisation of the PT results
and therapeutic ranges both in clinical practice and in
research. The units formerly used were inadequate for
international communication (46).
PT estimation carries its own analytical difficulties
depending, for example, on sampling (79) and on
sample and reagent stability (1012), reagent composition (13, 14), instrumentation (15, 16) and measuring
principles (17), the source of thromboplastins (13) and
the calibration (6, 1820). Thus, there is a need for continuous research to improve the uniformity of INR results.
In the Owren method of PT assessment the sample is
diluted before estimation and the sample volume in the
reaction mixture is 5%. In the Quick method there is no
sample dilution and the sample volume in the reaction
mixture is many times greater (33%), which means
sensitivity to preanalytical factors such as sample citrate concentration (1517, 21, 22).
The Owren method has the advantage of the small
sample volume and the fact that EDTA plasma can also
be used for PT measurement (2325).
The literature contains few publications comparing
the Quick and Owren methods and only few individual
patient results on INR obtained using different
reagents. Most of the studies have used only the mean
values of results, which is not a good method for PT
studies because individual changes may be marked
while the mean is acceptable. The aim of the present
work was to measure PT using two different sensitive
Quick and Owren reagents (International Sensitivity Index (ISI) near 1.0), using the same calibration, and to
compare the influence of different reagents on patient
INR results within the two methods. We sought to establish which of the methods is better for INR result
harmonisation.
400
and as INR using the same calibration. Measurement commenced within 4 hours of blood collection.
The citrate coagulation tube (Greiner Labortechnik GmbH,
Kremsmnster, Austria, Vacuette, cat. no. 454322, 9NC) contained 0.2 ml 0.109 mol/ (3.2%) citrate solution and a blood
volume of 1.8 ml. The sample needle (Becton Dickinson,
Plymouth, UK, Precision Glide, cat. no.360213) was 0.8 38
mm. Sample tubes were centrifuged at 1560 g for 10 min at
20 C to separate plasma.
Instruments and reagents
We measured Quick PT with the ACL 7000 (Instrumentation
Laboratory, IL, Milano, Italy), a fully automatic microcentrifugal analyser. The coagulation reagents were PT-Fibrinogen
Recombinant, rabbit tissue thromboplastin from Instrumentation Laboratory, Lexington, USA and Dade Innovin, human tissue thromboplastin from Dade Behring Marburg GmbH, Marburg, Germany. We used 50 l sample and 100 l of the
reagent.
We measured Owren PT with an ACL analyser with the following coagulation reagents: Nycotest PT, rabbit brain thromboplastin from Axis-Shield as, Oslo, Norway and SPA 20, tissue thromboplastin from Diagnostica Stago, Asnieres,
France, and with a diluent (Nycotest PT, dilution liquid, AxisShield as, Oslo, Norway and SPA buffer, Diagnostica Stago,
Asnieres, France). We used 10 l sample, 50 l diluent and
140 l of the reagent. We used the ACL 7000 internal standard
(IL Test Reference, cat. no. 9756900).
For Quick and Owren reagents ISI calibration was carried on
with ISI calibrators (Etaloquick cat. no. 00496, Diagnostica
Stago, France). The ACL was calibrated with the calibration
plasma (IL cat. no. 08467300): PT-Fibrinogen Recombinant,
n=4 at every point: 100%, 9.38 s, CV=0.48; 50%, 15.7 s,
CV=0.53; 25% 27.7 s, CV=0.44, R2=1.000; Dade Innovin, n=4 at
every point: 100%, 8.18 s, CV=0.55; 50%, 12.9 s, CV=1.29; 25%
22.7 s, CV=1.46, R2=1.000; Nycotest PT, n=4 at every point:
100%, 21.0 s, CV=0.00; 50%, 30.53 s, CV=0.38; 25% 45.8 s,
CV=0.44, R2=0.997; SPA 20, n=4 at every point: 100%, 20.4 s,
CV=0.00; 50%, 28.53 s, CV=0.40; 25%, 40.8 s, CV=0.49,
R2=0.995.
The day-to-day CVs (NKP 160, cat. no. GHI 160, normal
plasma, and OKP 165, cat. no. GHI 165, abnormal plasma from
the Global Hemostasis Institute, Linkping, Sweden) (n=8)
were as follows:
Figure 1
Results
Quick PT results were between 0.9 INR (7.05 s) and 8.1
INR (59.0 s), Owren PT results between 0.9 INR (18.2 s)
and 4.8 INR (97.2 s). The mean of patient results using
PT-Fibrinogen Recombinant reagent was 2.55 INR, using Dade Innovin reagent 2.49 INR and using Nycotest
PT 2.32 INR, SPA 20 2.31 INR.
The regression equations for Quick PT using PT-Fibrinogen Recombinant (x) and Dade Innovin reagent (y)
were:
y = 1.23x 0.63 INR, R2 = 0.98 (Figure 1)
y = 0.92x 3.00 s, R2 = 0.98
The regression equations for Owren PT using Nycotest
PT (x) and SPA 20 reagent (y) were:
y = 0.98x + 0.04 INR, R2 = 1.00 (Figure 2)
y = 0.86x + 3.05 s, R2=1.00
According to the Bland and Altman plot (26) seven
Quick PT INR results were out of range out of the 179
patient results (mean difference 0.05 INR, 2SD limits
Figure 2
Figure 3
Figure 4
0.53 to +0.63 INR). Differences in INR results between the two Quick PT reagents are shown in Figure 3. INR results did not differ with respect to clinical
401
402
Discussion
The objective of harmonising PT results and therapeutic
ranges using the INR units is commendable and worth
pursuing. For most indications the recommended INR
therapeutic range is 2.0 to 3.0 (6). Since the range is narrow there is a need for high quality of PT results. The
overall coefficient of variation of INR is 1113% under
favourable conditions if ISI is around 1 (27).
A variety of factors affect the final PT result in INR
units, and we should minimise the sources of error of
every factor in order to optimise results. The INR formula is a key factor in analysing the sources of error:
INR = (samples/normals )ISI. For evaluating variation in
INR results we took the logarithm from the formula:
Log10 INR = ISI Log10 (samples/normals). Low ISI
means low CV for INR according to the formula, but the
PT ratio is nevertheless not independent of ISI (6).
Using sensitive reagents improves the quality of INR
results and ISI as the power function plays a minimal
role. The role of ISI is enhanced with higher INR results
because the ratio differs from 1 and the power function
(ISI) has more meaning in calculating INR. The mean
normal PT (MNPT = normals) plays an important role in
the formula and markedly affects variation in INR results. The MNPT can be determined from the fresh
plasma of healthy persons or from lyophilised plasma
obtained commercially (13). As they are biological
materials from different sources, differences arise
between preparations, and this affects the INR results.
Sensitive reagents were selected for this study to
eliminate sources of error and compare Quick and
Owren methods. The calibrators for ISI and MNPT calculation were the same for Quick and Owren estimations.
Recombinant Quick PT reagents were chosen for the
study because they ensure excellent lot-to-lot uniformity and better performance than thromboplastins derived from natural sources. The correlation between
Quick PT reagents is good (R2=0.98), but the correlation
equation (y=1.23x0.63 INR) reveals considerable divergence from the ideal correlation line (Figure 1). Although the mean difference was only 0.05 INR, the
Bland and Altman plot (Figure 3) showed increasing
differences towards higher INR results, and this is
caused by ISI. The magnitude of the 2SD (0.53 to +0.63
INR) and deviations of INR results are in our opinion
too wide clinically, considering that the therapeutic
range is only 23 INR. These results show that although the means of INR results may be acceptable,
the individual results may be clinically unacceptable
with the Quick method.
Figure 2 shows good correlation (R2=1.00) for patient
INR results between the two Owren reagents, and the
equation (y=0.98x +0.04) is very close to the ideal line.
Acknowledgements
I thank the staff of the Laboratory at Valkeakoski District Hospital for their co-operation.
References
1. Quick AJ, Stanley-Brown M, Bancroft FW. A study of the
coagulation defect in hemophilia and in jaundice. Am J
Med Sci 1935; 190:50111.
2. Quick AJ. The prothrombin time in haemophilia and in obstructive jaundice. J Biol Chem 1935; 109:734.
3. Owren PA. Thrombotest. A new method for controlling anticoagulant therapy. Lancet 1959; 2:7548.
4. International Committee for Standardisation in Haematology. International Committee on Thrombosis and
Haemostastis. ICSH/ICTH recommendations for reporting
prothrombin time in oral anticoagulant control. Thromb
Haemost 1985; 53:1556.
5. WHO Expert Committee on Biological Standardisation.
Thirty-third Report. Technical Report Series 687. WHO
Geneva 1983:81105.
6. Hirsh J, Dalen JE, Anderson DR, Poller L, Bussey H, Ansell
J, et al. Oral anticoagulants mechanism of action, clinical
effectiveness, and optimal therapeutic range. Chest 1998;
114:44569.
7. Gottfried EL. Prothrombin time and activated partial
thromboplastin time performed on the first tube. Am J Clin
Pathol 1997; 107:6813.
8. McGlasson D, More LE, Best H, Norris W, Doe R, Ray H.
Drawing specimens for coagulation testing: is a second
tube necessary? Clin Lab Sci 1999; 12:1379.
9. Adcock DM, Kressin DC, Marlar RA. Are discard tubes necessary in coagulation studies? Lab Med 1997; 28:5303.
10. Heil W, Grunewald R, Amend M, Heins M. Influence of time
and temperature on coagulation analytes in stored
plasma. Clin Chem Lab Med 1998; 36:45962.
11. Leeming DR, Craig S, Stevenson KJ, Taberner DA. The de-
403
21. Duncan EM, Casey CR, Duncan BM, Lloyd JV. Effect of concentration of trisodium citrate anticoagulant on the calculation of the International Normalised Ratio and the International Sensitivity Index of thromboplastin. Thromb
Haemost 1994; 72:848.
22. Horsti J. Preanalytical aspects of routine coagulation measurements. Scand J Clin Lab Invest 2001; 61:1678.
23. Horsti J. Measurement of prothrombin time in EDTA
plasma with combined thromboplastin reagent. Clin Chem
2000; 46:18446.
24. Horsti J. Use of EDTA samples for prothrombin time measurement in patients receiving oral anticoagulants.
Haematologica 2001; 86:8515.
25. Horsti J. EDTA samples are stable for prothrombin time
measurement by combined thromboplastin reagent. Clin
Chem 2001; 47:17313.
26. Bland JM, Altman DG. Statistical methods for assessing
agreement between two methods of clinical measurement. Lancet 1986; 8:30710.
27. Loeliger EA, van den Besselaar AM, Lewis SM. Reliability
and clinical impact of normalisation of the prothrombin
times in oral anticoagulant control. Thromb Haemost
1985; 53:14854.
28. Adcock DM, Duff S. Enhanced standardisation of the International Normalised Ratio through the use of plasma calibrants: a concise review. Blood Coagul Fibrinolysis 2000;
11:58390.
Received 8 October 2001, revised 11 February 2002,
accepted 20 February 2002
Corresponding author: Juha Horsti, District Hospital of
Valkeakoski, University Hospital of Tampere, Laboratory,
37600 Valkeakoski, Finland
Phone: +358 3 586 7251, Fax: +358 3 586 7435
E-mail:juha.horsti@tays.fi