HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Extract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in Web of Science Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Web of Science (11) Citing Articles via Google Scholar Google Scholar Articles by Horsti, J. Search for Related Content PubMed PubMed Citation Articles by Horsti, J. Related Collections Hemostasis and Thrombosis

Agreement of Owren and Quick Prothrombin Times: Effects of Citrate and Calcium Concentrations and International Sensitivity Index Correction

¹ Valkeakoski District Hospital, Tampere University Hospital Laboratory, 37600 Valkeakoski, Finland; fax 358-3-586-7435, e-mail juha.horsti{at}tays.fi

The prothrombin time (PT) is commonly measured by either the "Quick method", which is based on the technique described by Quick and co-workers in 1935 (1)(2), or by the Owren method (3) (combined thromboplastin reagent). The latter is the predominant approach used in the Nordic countries, Benelux, and Japan. These two methods are rarely mentioned or noted in the scientific literature (4)(5), and more attention is paid to the thromboplastin in the reagent and to its origin (6)(7).

The Owren method is sensitive to coagulation factors II, VII, and X. Fibrinogen (factor I) and factor V are in the reagent. The Quick method is affected by deficiencies of fibrinogen and factors II, V, VII, and X. Both methods are suitable for the control of anticoagulant treatment. The most important technical difference between the two is the sample volume in the reaction mixture: 5% in the Owren method, and 33% in the Quick method. This has the effect of making the coagulation reaction in the Quick method more sensitive to many preanalytical variables, e.g., citrate concentration in the sample (8)(9). The small sample volume in the reaction mixture with the Owren method makes it possible to use even EDTA plasma (10).

The International Normalized Ratio (INR) is recommended to harmonize PT results for oral anticoagulant therapy (11). The recommendation conforms with those made by WHO using reference thromboplastins (12). Regardless of the reagent, instrument, or method used, the results should be the same when the INR system is used.

Because these two major PT methods have not previously been compared, the aim of this study was to compare the PT results obtained with two Owren reagents as INR units in samples containing 0.129 mol/L (3.8%) citrate solution and to compare the PT results and clinical significance with the same calibration between one Owren reagent and one Quick reagent as INR units in samples containing 0.129 mol/L (3.8%) and 0.109 mol/L (3.2%) citrate solution. Because calcium (factor IV) is in a key position in coagulation reactions as a cofactor in several steps, we sought to ascertain the ionized calcium (Ca²⁺) concentration in the reaction mixtures in both methods and to study the effect of sample citrate concentration on the Ca²⁺ concentration in the reaction mixtures by both methods.

All procedures were approved by our institution’s responsible committee in accordance with the Helsinki Declaration of 1975. We studied patient samples chosen without conscious bias from among hospital and health center patients; 145 patients were sampled with a citrate coagulation tube (Vacutainer cat. no. 367702, 9NC; Becton Dickinson) containing 0.35 mL of 0.129 mol/L (3.8%) citrate solution (blood volume, 3.15 mL), and 157 patients were sampled with a citrate coagulation tube (Vacuette cat. no. 454327, 9NC; Greiner Labortechnik GmbH) containing 0.35 mL of 0.109 mol/L (3.2%) citrate solution (blood volume, 3.15 mL). The sample needle (Precision Glide, cat. no. 360213; Becton Dickinson) was 0.8 x 38 mm. Sample tubes were centrifuged at 1560g for 10 min at 20 °C to separate the plasma.

PT was measured in samples collected in both kinds of sample tubes and reported in INR units. Measurement commenced within 2 h of blood collection. In 30 of the same patients, Ca²⁺ concentrations were measured in the reaction mixtures for both sample types (0.129 and 0.109 mol/L citrate) and for both methods.

The ACL 1000 and ACL 7000 [Instrumentation Laboratory (IL)] are automatic microcentrifugal analyzers. We measured PT on the ACL 1000 analyzer using Nycotest PT coagulation reagent [cat. no. 1055510; International Sensitivity Index (ISI), 1.006; rabbit brain thromboplastin] and citrate-barbiturate buffer as diluent (cat. no. 1002834; both from Nycomed Pharma), with a 10-µL of sample, 50 µL of diluent, 140 µL of reagent, and IL Test Reference (cat. no. 97569-00).

We measured PT on the ACL 7000 analyzer using Owren’s PT coagulation reagent (cat. no. GHI 131-10; ISI; 1.22; rabbit brain thromboplastin), 25 mmol/L CaCl₂ (cat. no. GHI 155) as diluent, Owren buffer [cat. no. GHI 150; all from the Global Hemostasis Institute (GHI)], with a 10-µL sample, 50 µL of diluent, and 140 µL of reagent. We also measured PT on the ACL 7000 analyzer using PT-Fibrinogen Recombinant coagulation reagent (cat. no. 20005000; ISI, 1.12; rabbit tissue thromboplastin) and sample diluent (cat. no. 09756800; both from IL), with a 50-µL sample and 100 µL of reagent. All three methods (reagents) were calibrated with the same calibrators: calibration plasma (cat. no. 08467300; IL) and ISI calibrators (Etaloquick, cat. no. 00496; Diagnostica Stago).

ACL calibration is needed for INR units. The ACL 1000 calibration data with Nycotest PT reagents were as follows (n = 6 at every point): 100%, 19.8 s, CV = 0.00%; 50%, 26.4 s, CV = 0.04%; 25%, 40.87 s, CV = 0.75% (R² = 1.00). The ACL 7000 PT calibration data with the Owren’s PT method were as follows (n = 4 at every point): 100%, 20.0 s, CV = 0.00%; 50%, 27.6 s, CV = 0.00%; 25%, 38.80 s, CV = 0.00% (R² = 0.99). The ACL 7000 PT calibration data with the PT-Fibrinogen Recombinant method were as follows (n = 4 at every point): 100%, 11.4 s, CV = 0.73%; 50%, 20.5 s, CV = 0.53%; 25%, 34.2 s, CV = 1.7% (R² =1.00).

Within the same run (n = 6), the CVs were as follows: for Nycotest PT, CV = 0.5% (mean, 21.1 s) and 0.4% (mean, 38.8 s); for Owren’s PT, CV = 0.4% (mean, 21.2 s) and 0.9% (mean, 35.3 s); for PT-Fibrinogen Recombinant, CV = 0.8% (mean, 12.0 s) and 1.0% (mean, 23.3 s).

The day-to-day CVs were as follows: for Nycotest PT, CV = 1.5% (mean, 20.0 s; n = 14); for Owren’s PT, CV = 1.7% (mean, 20.3 s; n = 14); for PT-Fibrinogen Recombinant, CV = 2.0% (mean, 11.7 s; n = 10).

Ionized calcium was measured with an IL 1640 blood gas analyzer, using IL Test contrIL PLUS (cat. no. 98321-30, lot. no. N0208083). The CV was 0.62% at a mean calcium concentration of 1.47 mmol/L (n = 9), 0.71% at 1.09 mmol/L (n = 10), and 1.4% at 0.62 mmol/L (n = 6).

For Ca²⁺, we used IL Test Cal 1 (cat. no. 98318-00), Test Cal 2 (cat. no. 98342-00), and Test BGE Flush (cat. no. 98320-00). For Ca²⁺ measurements, a reaction mixture was prepared with volumes two times those for the ACL (for Owren’s PT, 20 µL + 100 µL + 280 µL; for Quick PT, 100 µL + 200 µL); after coagulation the solution was centrifuged at 12 700g for 10 min, and Ca²⁺ was measured in the supernatant by an IL 1640. The regression equations for calculating the patient data were as follows:

Owren ISI x 1.20 for 0.129 mol/L citrate.

Owren ISI x 1.14 for 0.109 mol/L citrate (Fig. 1B ).

View larger version (20K): [in this window] [in a new window]   Figure 1. Differences between patient results obtained by the Owren and Quick methods as INR in 0.109 mol/L (3.2%) citrate before (A) and after (B) ISI correction.

These regression models were applied to convert the results obtained by the Quick and Owren methods to ISI, making its routine use easy.

The regression equations for PT estimation from citrate plasma (0.129 mol/L) were as follows for results obtained as INR: for two Owren reagents, Nycotest PT (x) vs Owren’s PT (y), y = 1.08x - 0.15 (R² = 0.99); for the Owren (x; Nycotest PT) and Quick (y) methods, y = 0.98x^1.20 (R² = 0.95).

For patient results obtained by the Owren (Nycotest PT) and Quick methods using 0.129 and 0.109 mol/L citrate samples, for results <2.0 INR, the differences between the methods were not clinically significant with both sample types (Table 1 ). For results >2.0 INR, however, the means of the differences were clinically high and thus significant with both sample types. The Bland and Altman method (13) was not suitable for analyzing results in the range >2 INR before correction because the distribution was not gaussian (Table 1 and Fig. 1A ). In the therapeutic range 2–4 INR, the difference in patient results was >0.5 INR for 31 of 83 samples (37%) in 0.129 mol/L citrate and 28 of 103 (27%) samples in 0.109 mol/L citrate.

When the Owren ISI was corrected according to the correlation equation, INR = (sample_s/normal_s)^1.21, between the methods from 1.006 to 1.21 (1.21 = 1.006 x 1.20), eight results were outside 2 SD in the range 2–4 and >4 INR (mean difference, 0.07 INR; limits, -0.55 to 0.69 INR). According to Bland and Altman (13), 7 results of 145 could be outside the limits. With a probability of 94%, the clinical significance of the results is the same.

The regression equations for PT estimation from citrate plasma (0.109 mol/L) were as follows: for the Owren (x; Nycotest PT) and Quick (y) methods, with the results obtained as INR, y = 1.007x^1.14 (R² = 0.97). The difference plot (13) before correction (mean difference, -0.33 INR; limits, -0.99 INR to 0.33 INR) is shown in Fig. 1A .

When the ISI of the Owren method was corrected according to the correlation equation, INR = (sample_s/normal_s)^1.15, between the methods from 1.006 to 1.15 (1.15 = 1.006 x 1.14), nine results were outside 2 SD in the range 2–4 and >4 INR (mean difference, -0.01 INR; limits, -0.49 INR to 0.47 INR). According to Bland and Altman (13), 8 results of 157 could be beyond the limits (Fig. 1B ). With a probability of 94%, the clinical significance of the results is the same.

The mean Ca²⁺ concentrations in the reaction mixtures for the Owren and Quick methods using both citrate samples (0.129 or 0.109 mol/L) are given in Table 1 .

In the comparison of one Owren and the Quick method, we obtained fairly good nonlinear correlation using both citrate samples (R² = 0.95 for 0.129 mol/L; R² = 0.97 for 0.109 mol/L). The correlation equations were the power functions (y = 0.98x^1.20 and y = 1.007x^1.14). The difference in exponents is 0.06, and this change depends on the change in sample citrate concentration (0.02 mol/L). Because the citrate concentration in the sample affects the Ca²⁺ concentration in reaction mixtures with the Quick method but not the Owren method (Table 1 ), we may conclude that the correlation equation formula derives from the Ca²⁺ concentration. The correlation equation shows that by changing ISI, we can improve the comparability of results between the Owren and Quick methods. The correlation between the methods is somewhat better when 0.109 mol/L citrate samples are used rather than 0.129 mol/L citrate samples. The correlation improves as the citrate concentration decreases.

Clinically, there was no significant difference between the Owren and Quick methods when INR results were <2 INR. In the important therapeutic range 2–4 INR and >4 INR, Quick results were markedly higher (mean differences, 0.44 and 0.39 for 2–4 INR, and 1.48 and 1.26 for >4 INR), and the differences were clinically significant for both citrate concentrations (Table 1 ). The patients whose samples were analyzed by Owren PT methods received too much medication, and this should be reduced. Oral anticoagulant therapy recommendations are based on Quick PT (5)(11). In published scientific reports, the situation was the same in that the results were not equal.

Using ISI correction (Owren ISI x 1.20 for 0.129 mol/L; Owren ISI x 1.14 for 0.109 mol/L), we can unify the results clinically with a probability of 94%.

In a previous study (10), the correlation equation for INR units was: y = 1.20x - 0.162 (R² = 1.00) for EDTA (x) and citrate (y) samples using the Owren method. The equation for results obtained by the Quick method (y) in INR units vs results obtained for EDTA samples (x) by the Owren method was: y = (1.20x - 0.162)^{Owren ISI x 1.20 or Owren ISI x 1.14}.

The two different Owren reagents show good correlation (R² = 0.99), and the results involve no clinically significant difference.

The physiological concentration of Ca²⁺ in plasma is 1.3 mmol/L. In 0.129 mol/L citrate samples, the mean Ca²⁺ concentration in the reaction mixture was 0.21 mmol/L (14%) lower than the concentration in 0.109 mol/L citrate samples, using the Quick method. The difference in Ca²⁺ concentrations is the same with other Quick reagents because the relationship between the volumes (50 µL + 100 µL) in the reaction mixture is the same with every reagent. The influence of variable Ca²⁺ concentrations in the reaction mixture can exert different effects on the results obtained with different Quick reagents and thromboplastins.

In addition, in the Quick method, the Ca²⁺ concentration varies widely (1.07–1.79 mmol/L in 0.129 mol/L citrate samples), and the CV is 13%. In the 0.129 mol/L reaction mixtures, in 5 of 30 samples, the Ca²⁺ concentration was <1.3 mmol/L. When 0.109 mol/L citrate samples were used in the Quick method, the lowest Ca²⁺ concentration in the reaction mixture was 1.3 mmol. The Ca²⁺ concentrations in the coagulation mixtures are clearly different between two sample citrate concentrations with the Quick method.

When a sample is diluted (1 + 9) with citrate solution, the dilution occurs in plasma, not in whole blood, and depending on the hematocrit, the plasma volume will differ from sample to sample. The situation is similar if too little blood is drawn into the collection tube and the tube is underfilled. The Ca²⁺ concentration in patient plasma can sometimes be low. The citrate concentration is markedly variable in plasma, as well as in the coagulation reaction mixture in the Quick method.

Because the Ca²⁺ concentration is lower in 0.129 mol/L citrate samples, the times for PT coagulation and activated partial thromboplastin time measurements are longer than the corresponding times for patient samples containing 0.109 mol/L citrate for patient groups designated as "no anticoagulant therapy" and "anticoagulant therapy" when PT (ISI, 1.0; Quick method) and activated partial thromboplastin time reagents are used (9). Duncan et al. (8) showed that the concentration of citrate in the sample affects the results and ISI calibration (Quick methods). The citrate concentration exerts an 10% effect on the ISI of reagents. The above-mentioned results of two studies agree well with the PT and calcium results in our study.

With Owren’s PT, the mean difference in the Ca²⁺ concentration in the reaction mixture was 0.01 mmol/L (0.4%) and the CVs were 4.9% for 0.129 mol/L citrate and 3.0% for 0.109 mol/L. The sample citrate concentration has no effect on the Owren method, and this method is also less sensitive to other preanalytical variables by reason of the small sample volume (5%) in the reaction mixture. Because the Owren and Quick methods depend differently on sample citrate concentrations (Ca²⁺ in the reaction mixture), the correlation is not so good.

We subscribe to the demand for unification of recommendations of a sample citrate concentration of 0.109 mol/L to improve results for PT and activated partial thromboplastin times (8)(9). For results obtained by the Owren method, ISI correction should be used to standardize patient INR results and anticoagulant therapy to be the same as the results obtained by the Quick method because the recommendations for care are made according to the Quick method. The patients whose samples were analyzed by the Owren method were receiving too much medication and were in greater danger of bleeding.

Acknowledgments

I thank the staff of the Laboratory at the Valkeakoski District Hospital for their assistance.

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;190:501-511.[Web of Science]
2. Quick AJ. The prothrombin time in haemophilia and in obstructive jaundice. J Biol Chem 1935;109:73-74.
3. Owren PA. Thrombotest. A new method for controlling anticoagulant therapy. Lancet 1959;2:754-758.
4. van den Besselaar AM. International standardisation of laboratory control of oral anticoagulant therapy: a survey of thromboplastin reagents used for prothrombin time testing. J Heart Valve Dis 1993;2:42-52.[Medline] [Order article via Infotrieve]
5. 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;114:445-469.
6. Kirkwood TB. Calibration of reference thromboplastins and standardisation of the prothrombin time ratio. Thromb Haemost 1983;49:238-244.[Web of Science][Medline] [Order article via Infotrieve]
7. Hermans J, van den Besselaar AM, Loeliger EA, van der Velde EA. A collaborative calibration study of reference materials for thromboplastins. Thromb Haemost 1983;50:712-717.[Medline] [Order article via Infotrieve]
8. Duncan EM, Casey CR, Duncan BM, Lloyd JV. Effect of concentration of trisodium citrate anticoagulant on the calculation of the International Normalized Ratio and the International Sensitivity Index of thromboplastin. Thromb Haemost 1994;72:84-88.[Web of Science][Medline] [Order article via Infotrieve]
9. Adcock DM, Kressin DC, Marlar RA. Effect of 3.2% vs 3.8% sodium citrate concentration on routine coagulation testing. Am J Clin Pathol 1997;107:105-110.[Web of Science][Medline] [Order article via Infotrieve]
10. Horsti J. Measurement of prothrombin time in EDTA plasma with combined thromboplastin reagent. Clin Chem 2000;46:1844-1846.[Free Full Text]
11. . 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;53:155-156.[Web of Science][Medline] [Order article via Infotrieve]
12. . WHO Expert Committee on Biological Standardisation. Thirty-third Report 1983:81-105 WHO Technical Report Series 687. Geneva. .
13. Bland JM, Altmann DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986;8:307-310.

The following articles in journals at HighWire Press have cited this article:

				J. Horsti, H. Uppa, and J. A Vilpo Poor Agreement among Prothrombin Time International Normalized Ratio Methods: Comparison of Seven Commercial Reagents Clin. Chem., March 1, 2005; 51(3): 553 - 560. [Abstract] [Full Text] [PDF]

				J. Horsti EDTA Samples Are Stable for Prothrombin Time Measurement by Combined Thromboplastin Reagent Clin. Chem., September 1, 2001; 47(9): 1731 - 1732. [Full Text] [PDF]

This Article Extract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in Web of Science Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Web of Science (11) Citing Articles via Google Scholar Google Scholar Articles by Horsti, J. Search for Related Content PubMed PubMed Citation Articles by Horsti, J. Related Collections Hemostasis and Thrombosis