European concerted action on anticoagulation. Use of plasma samples to derive international sensitivity index for whole-blood prothrombin time monitors.
We also tested citrated plasmas recalcified with a calcium chloride concentration that we found suitable for PT testing of plasmas on the POCT systems (L. Poller, manuscript in preparation). The aim of the present study was to determine whether the ISIs of these monitors could be derived reliably with citrated plasmas. If successful, ISI calibration of whole-blood POCT monitors could then be simplified.
To avoid invidious comparison with the several different types of POCT monitors on the market that have not been subjected to the same analysis, results with the three POCT systems have been coded.
Materials and Methods
The CoaguChek and TAS systems, consisting of the monitors and specific lots of test strips/cards, were provided for the study by Roche Diagnostics (Mannheim, Germany) and Bayer (Leverkusen, Germany), respectively.
CoaguChek. The CoaguChek system consisted of the meter, test strips incorporating thromboplastin, and a code chip containing calibration information relevant to the specific lot of test strips, such as mean normal prothrombin time (MNPT), ISI, lot number, and expiration date. For this system, the test strip, which contains iron oxide particles and thromboplastin, is inserted into the monitor and warmed to 37[degrees]C. An unmeasured drop of capillary blood (noncitrated) is applied to the strip. The blood is then transported to the reaction area by capillary forces, and the coagulation process is triggered by contact with the thromboplastin. There are two magnets below the test strip, a permanent magnet and an electromagnet. The former causes the iron particles on the test strip to align horizontally, and the latter forces these particles into the vertical plane within a set frequency, giving rise to a regular pulsation pattern. A photodetector above the test strip registers the changes in reflected light caused by this pulsation pattern. As soon as a clot begins to form, movement of the iron particles slows until it finally stops. This produces a decrease in reflection, which is interpreted as the onset of coagulation. An algorithm programmed into the specific code chip converts this result into PT and INR.
Two CoaguChek meters and single lots of two different types of test strips (CoaguChek-S test strip lot no. 828594 and CoaguChek Mini test strip lot no. 079) were provided to each of three centers. CoaguChek-S and Mini test strips contain human recombinant and rabbit plain thromboplastin, respectively.
TAS. The TAS has recently been redesignated the RapidPointCoag, but in this study the established name of TAS has been retained. The TAS is designed primarily for near-patient testing, but it can be used for home PT testing. The TAS works on essentially the same principles as the CoaguChek. The test carriers are referred to as "test cards". A magnetic strip on the back of each test card contains all relevant calibration information, type of test, lot number, expiration date, and other test parameters, such as MNPT and ISI, specific to the individual lot of cards. This information is entered into the TAS by passing the test card through the magnetic card reader at the beginning of each test.
The test cards (lot no. 305099801) used in this study are designated PT-NC (prothrombin time noncitrated) and contain human placental plain thromboplastin.
Until the 1990s, thromboplastins with a high ISI (>2.0) were widely used in North America, and many laboratories were accustomed to the resulting short PT. PTs displayed on some POCT monitors were therefore in terms of a typical North American thromboplastin, which are not real measurements in seconds.
The CoaguChek Mini displays PTs that are considered representative of the result, in seconds, that would have been obtained with a traditional high-ISI thromboplastin. To provide results in real PT in this study, a master code chip containing a correction formula was provided by Roche Diagnostics for the CoaguChek Mini system rather than the conventional code chip.
The TAS PT-NC also does not give real PT results. During the development of the PT-NC test cards, the objective of the manufacturer was to have results as close as possible to laboratory results. The PTs on the TAS PT-NC system are therefore expressed as an assumed laboratory PT result with the actual TAS PT being longer than the displayed result on the monitor. For the TAS PT-NC, a specific correction formula was supplied by the manufacturer for this study for calculation of real PTs from those reported on the monitor display screen.
THREE-CENTER WHO ISI CALIBRATION
A WHO ISI calibration as modified for POCT monitors by Tripodi et al. (3) was conducted at three centers: Leiden, Manchester, and Milan. Each center collected blood samples from 20 healthy blood donors and 60 long-term patients stabilized on oral anticoagulants within the therapeutic interval (INR, 1.5-4.5). Blood samples were collected over a 5- to 10-day period at each center. All volunteers gave informed consent.
COLLECTION AND TESTING OF NONCITRATED WHOLE BLOOD ON THE MONITORS
Noncitrated whole blood was collected by venipuncture using a needle and plastic syringe. The blood was applied with minimum delay to the test strips/cards in the monitors, and the PTs were recorded from the monitor display screens. Single tests were performed on all whole-blood samples.
BLOOD COLLECTION FOR CITRATED PLASMA SAMPLES
An additional sample of venous blood from the same donor was collected at the same venipuncture, either with a Vacutainer system containing 0.105 mol/L citrate or with a Monovette system containing 0.106 mol/L citrate (the difference in citrate concentration between these systems is negligible). Citrated plasma was obtained by centrifugation and stored at room temperature before testing. All tests were completed within 6 h of venipuncture.
TESTING OF CITRATED PLASMA ON THE MONITORS
Citrated plasma (0.1 mL) was transferred to a plastic tube. Calcium chloride (0.1 mL; 16.3 mmol/L) was added and mixed to give a clotting time comparable to the whole-blood clotting time and the traditional manual PT method.
Recalcified plasma was applied within 15 s  to the test strip/cards, and the PTs were recorded from the monitor display screens. Single tests were performed on the monitoring systems with all plasma samples. The 16.3 mmol/L calcium chloride solution was prepared individually at each of the three centers.
Manual PTs for the citrated plasmas were obtained with the European Concerted Action on Anticoagulation (ECAA) rabbit reference plain thromboplastin ([ISI.sub.ref] = 1.67) according to the recommended procedure (5).
Single tests were performed with the patients' plasmas and duplicate tests with the control plasmas.
Calibration data. The ISIs for whole blood and citrated plasma with the three monitoring systems were derived according to the conventional WHO orthogonal regression procedure (1, 2). The logs of individual PTs for healthy blood donors and coumarin-treated patients, respectively, obtained with the ECAA rabbit reference reagent were plotted as recommended on the vertical axis against the log of the PT obtained from the monitor on the horizontal axis (1, 2). The slope of the calibration line was calculated using orthogonal regression analysis. The precision of the calibration slope (b) was determined by its CV: CV(b) = 100 = SE(b)/b, where SE (b) is the standard error of b. The assumption that the mean log PT for the healthy blood donors lies on the calibration line derived from the PTs for patient samples only was assessed by the displacement test of Tomenson (6). This determines the significance of the vertical distance (displacement) of the mean of the log PT for the healthy blood donors from the patient-only line. The estimated slopes of the calibration lines based on the PTs for both healthy blood donors and patients and the corresponding 95% confidence intervals are given to allow a comparison between whole-blood and citrated plasma calibrations for each of the three monitoring systems. The ISI was derived as: ISI = b x [ISI.sub.ref], where [ISI.sub.ref] = 1.67 is the ISI of the ECAA rabbit reference thromboplastin.
MNPT. The MNPTs (geometric mean of the PTs obtained for the 20 samples from healthy blood donors) for whole blood and citrated plasma were compared using paired t-tests on the log PT obtained for the healthy blood donors, for the three centers, and for the three monitoring systems.
Assessment of INR. The prothrombin ratio (PR), which corresponds theoretically to an INR of 3.0 (this being a good representative degree of anticoagulation) with the mean whole-blood ISI of the three centers, was calculated. The INRs derived from this PR and the mean citrated plasma ISI were then obtained, and the absolute percentage of deviation from the whole-blood INR of 3.0 was determined. As in previous reports (7-9), the absolute percentage of INR deviation was deemed clinically relevant if >10%.
[FIGURE 1 OMITTED]
Results for the three monitoring systems are coded as systems A, B, and C. The ISI values and the calibration statistics for whole-blood samples and plasmas on the three monitoring systems at each of the three centers are given in Table 1. The mean results for these three centers are given in Table 2.
The CV of the calibration line slopes was 1.9-3.9% excluding system C results at Milan, which used a different lot number of test cards/strips. For systems A and C at all three centers and system B at two centers, the CV of the slope was smaller for plasmas than for whole-blood samples. For system A at all three centers, there was a small difference between the CVs of the slopes for the whole-blood samples and plasmas. There were larger differences between the whole-blood and citrated plasma CVs at one of the three centers with system B and at all three centers with system C.
There was evidence (P <0.01) that the mean log PT for healthy individuals did not lie on the patient-only calibration line (i.e., significant displacement) in 2 of the 18 calibrations. This is greater than the number of significant displacements that would be expected by chance in 18 tests at the 1% level. The two calibrations giving significant displacement were with plasmas on systems A and B (see Table 1). Fig. 1 shows both the lines for patients and healthy blood donors combined and the patient-only line for these calibrations.
[FIGURE 2 OMITTED]
The clinical significance of the displacement was assessed according to WHO guidelines (2). The PR corresponding to a theoretic Tomenson-corrected (6) INR of 2.0 was calculated. The INR determined with this PR and the ISI derived from the slope of the line for samples from both patients and healthy individuals was compared with the INR of 2.0 (see Table 1). This was repeated for a theoretic INR of 4.5. The absolute percentage of INR deviation exceeded the WHO 10% limit (2) for system B only.
Again there was little difference in mean ISI (see Table 2) for whole-blood and plasma samples at the three centers on systems A and B. However, for system C, the difference between the mean ISIs for whole blood and citrated plasma was much greater. Fig. 2, which gives the calibration slopes and the 95% confidence intervals, allows a comparison of slopes for whole blood and citrated plasma. Although there were between-center differences in the slopes for both systems A and B, there was little within-center difference between the slopes obtained for whole blood and plasma at the individual centers. With system C, the within-center difference between the calibration slopes for whole blood and plasma was much greater at two of the three centers.
The theoretic citrated plasma INR derived from the mean plasma ISI and the PR corresponding to an INR of 3.0 obtained from the mean whole-blood ISI are shown in Table 2. The absolute percentage of deviation of the plasma INR from the whole-blood INR of 3.0 was considerably less than the 10% limit for systems A and B. For system B, there was no difference between INRs because the mean ISIs for whole blood and citrated plasma were the same. With system C, the two centers with the same lot of test cards/strips gave ~12% lower INRs with plasma than with whole blood.
The MNPTs with the three monitoring systems based on both whole-blood and plasma samples are given in Table 3. At all three centers with systems A and B, MNPTs were significantly longer (P <0.0001) with plasmas than with whole blood. With system C, the pattern in the difference was not consistent at the three centers. At Leiden, the MNPT was significantly shorter (P <0.0001) with the plasmas than with the whole blood. At Manchester, the plasma MNPT on system C was shorter than the MNPT for whole blood, but the difference was not significant (P = 0.06). The plasma MNPT was significantly longer (P = 0.03) than the MNPT for whole blood at Milan.
This study demonstrates that provided an appropriate calcium chloride concentration is used to recalcify the citrated plasma, plasma PT results can be used for accurate ISI calibration of two of the three whole-blood POCT monitoring systems considered in this report. Results showed little difference between plasma and whole-blood ISIs. With the third system, the difference in ISI between plasma and whole-blood calibrations produced a clinically relevant difference in INR. A theoretic PR of 2.6 gave a 12% INR difference between whole blood and citrated plasma (3.00 and 2.63, respectively). This would need to be investigated further to determine whether this problem can be resolved.
The precision of the resulting calibrations using the WHO procedure modified by Tripodi et al. (3) based on 20 healthy controls and 60 long-term coumarin-treated patients was satisfactory. The CV of the calibration slope was less than the recommended upper limit of 3% set for conventional manual PT testing (2) at two centers on two monitoring systems with both whole-blood and citrated plasma samples. At the third center, the slope CVs were close to 3%. With the third monitoring system, slope CVs were <3% at one center and <5% at all three centers. There was no suggestion, however, that substitution of plasma samples made the precision of the calibration poorer with the PT monitors; indeed, the mean CV of the slope was greater with whole blood for all three monitoring systems.
In the present study, a secondary reference thromboplastin (ECAA rabbit reference plain thromboplastin) was used in the ISI calibrations. Although WHO guidelines recommend like-species calibrations (2), only one of the three POCT systems was calibrated with like-species thromboplastin. For this reason and because of the important differences between results at the three centers, a larger study is required with relevant WHO International Reference Preparations for accurate ISI determinations. This is currently being undertaken in an ECAA multicenter calibration of the CoaguChek Mini and TAS systems against the human and rabbit thromboplastin International Reference Preparations.
The members of the ECAA Steering Group are as follows: J. Jespersen (Chairperson), L. Poller (Co-ordinator), A.M.H.P. van den Besselaar, F.J.M. van der Meer, C. Shiach, A. Tripodi, and F.E. Preston (Consultant). The study was supported by EC Grant SMT4-CT98-2269 and an additional grant from the Manchester Thrombosis Research Foundation. We express our gratitude to Roche Diagnostics and Bayer for the loan of the CoaguChek and TAS instruments, respectively, and test strips/cards.
(1.) WHO Expert Committee on Biological Standardization. 33rd report. World Health Organ Tech Rep Ser 1983;687:81-105.
(2.) WHO Expert Committee on Biological Standardization. Guidelines for thromboplastins and plasma used to control oral anticoagulation therapy. Annex 3. World Health Organ Tech Rep Ser 1999; 889:64-93.
(3.) Tripodi A, Arbini AA, Chantarangkul V, Bettega D, Mannucci PM. Are capillary whole blood coagulation monitors suitable for the control of oral anticoagulant treatment by the International Normalized Ratio? Thromb Haemost 1993;70:921-4.
(4.) Tripodi A, Chantarangkul V, Clerici M, Negri B, Mannucci PM. Determination of the ISI of a new near-patient testing device to monitor oral anticoagulant therapy. Thromb Haemost 1997;78:855-8.
(5.) Poller L, Barrowcliffe TW, van den Besselaar AMHP, Jespersen J, Houghton D. European Concerted Action on Anticoagulation (ECAA)-the multicentre calibration of rabbit and human ECAA reference thromboplastins. Steering Committee. Thromb Haemost 1996;76:977-82.
(6.) Tomenson JA. A statistician's independent evaluation. In: van den Besselaar AMHP, Lewis SM, Gralnick HR, eds. Thromboplastin calibration and oral anticoagulant control. Boston: Martinus Nijhoff, 1984:87-108.
(7.) Poller L, Barrowcliffe TW, van den Besselaar AMHP, Jespersen J, Tripodi A, Houghton D, for the ECAA. Minimum lyophilized plasma requirement for ISI calibration. Am J Clin Pathol 1998;109:196-204.
(8.) Poller L, van den Besselaar AMHP, Jespersen J, Tripodi A, Houghton D, for the ECAA. The effect of sample size on fresh plasma thromboplastin ISI determination. Br J Haematol 1999; 105:655-63.
(9.) Poller L. Screening INR deviation of local prothrombin time systems. J Clin Pathol 1998;51:356-9.
 Nonstandard abbreviations: POCT, point-of-care testing; PT, prothrombin time; INR, International Normalized Ratio; ISI, International Sensitivity Index; TAS, Thrombolytic Assessment System; PT-NC, prothrombin time, noncitrated; MNPT, mean normal PT; ECAA, European Concerted Action on Anticoagulation; and PR, prothrombin ratio.
 To determine the maximum time allowable between recalcification and testing of plasma samples on the POCT monitoring systems, recalcified plasma was added to two of the three POCT systems after 5, 30, 60, and 90 s with the order of testing on the two instruments being reversed on subsequent samples. Five replicate tests were performed for each time interval. Results obtained on the CoaguChek system showed no change over the time intervals, but results obtained on the TAS system suggested some shortening of PT after 30 s. Detailed study on the TAS performed at intervals of 5 s for up to 30 s indicated a progressive shortening of PT after 15 s. It was therefore decided to perform all tests with recalcified plasmas on all three POCT systems within 15 s.
LEON POLLER,  * MICHELLE KEOWN,  NIKHIL CHAUHAN,  ANTON M.H.P. VAN DEN BESSELAAR,  JOYCE MEEUWISSE-BRAUN,  ARMANDO TRIPODI,  MARIAGRAZIA CLERICI,  AND JORGEN JESPERSEN 
 European Concerted Action on Anticoagulation (ECAA) Central Facility, School of Biological Sciences, The University of Manchester, Manchester M13 9PT, United Kingdom.
 Haemostasis and Thrombosis Research Center, Leiden University Medical Center, 2333ZA Leiden, The Netherlands.
 A Bianchi Bonomi, Haemophilia and Thrombosis Centre, IRCCS Maggiore Hospital, University of Milan, 20122 Milan, Italy.
 Department for Thrombosis Research, University of Southern Denmark, and Department of Clinical Chemistry, Ribe County Hospital, DK-6700 Esbjeig, Denmark.
* Author for correspondence. Fax 44-161-275-5316; e-mail firstname.lastname@example.org. Received July 23, 2001; accepted November 1, 2001.
Table 1. ISI and calibration statistics for individual centers. (a) Center Test Slope CV(b), ISI (b) % System A Leiden Whole blood 0.97 2.1 1.62 Plasma 0.93 1.9 1.56 Manchester Whole blood 0.90 3.2 1.50 Plasma 0.95 3.2 1.58 Milan Whole blood 0.85 2.9 1.42 Plasma 0.79 2.9 1.32 (b) (4%, 8%) System B Leiden Whole blood 0.72 2.5 1.20 Plasma 0.69 1.9 1.15 (b) (28%, 37%) Manchester Whole blood 0.65 2.8 1.08 Plasma 0.65 2.8 1.08 Milan Whole blood 0.58 2.9 0.97 Plasma 0.60 2.8 1.00 System C (c) Leiden Whole blood 0.74 2.8 1.24 Plasma 0.63 2.1 1.06 Manchester Whole blood 0.63 3.9 1.06 Plasma 0.58 3.2 0.96 Milan3 Whole blood 0.72 4.7 1.20 Plasma 0.61 3.2 1.01 (a) Estimated slope (b), CV (%) of b [CV(b)], and ISI for each monitor test system with noncitrated whole-blood and plasma samples. (b) Significant displacement at the 1% level using the Tomenson (6) test (the absolute percentages of INR deviation from theoretic INRs of 2.0 and 4.5, respectively, are given in parentheses). (c) Different lot of test cards/strips used with system C monitor at Milan. Table 2. Mean calibration statistics. (a) Monitoring Test Mean Mean Mean INR %diff system slope CV(b), % ISI A Whole blood 0.91 2.8 1.51 3.00 -1.7 Plasma 0.89 2.7 1.49 2.95 B Whole blood 0.65 2.7 1.08 3.00 0.0 Plasma 0.65 2.6 1.08 3.00 C (b) Whole blood 0.69 3.4 1.15 3.00 -12.3 Plasma 0.61 2.7 1.01 2.63 (a) Mean slopes, mean CV of slopes, and mean ISI obtained for each monitor test system with samples of noncitrated whole blood and citrated plasma. Theoretic plasma INR determined from the mean plasma ISI and the PR corresponding to a whole-blood INR of 3.0 are given with the percentage of deviation (%diff) of plasma INR from the whole-blood INR of 3.0. (b) Mean of results for two centers only. Table 3. MNPT and SD (in parentheses) for 20 healthy blood donors. (a) Mean (SD) MNPT Monitoring Test Leiden Manchester system System A Whole blood 17.7 (1.34) 17.3 (1.05) Plasma 19.3 (1.68) 19.9 (1.11) System B Whole blood 12.5 (0.57) 12.1 (0.57) Plasma 13.6 (0.72) 13.1 (0.97) System C (b) Whole blood 15.3 (1.10) 14.3 (1.15) (c) Plasma 13.6 (1.10) 13.7 (1.07) ECAA rabbit reference Plasma 16.6 (1.25) 17.5 (0.73) thromboplastin Mean (SD) MNPT Monitoring Test Milan Overall system MNPT (SD) System A Whole blood 18.5 (1.02) 17.8 (1.23) Plasma 20.9 (1.53) 20.0 (1.58) System B Whole blood 12.2 (0.69) 12.3 (0.62) Plasma 13.6 (0.76) 13.5 (0.83) System C (b) Whole blood 12.9 (0.82) 14.9 (1.21) Plasma 13.4 (0.68) 13.7 (1.07) ECAA rabbit reference Plasma 18.8 (1.34) 17.7 (1.45) thromboplastin (a) Samples were collected locally and tested as both whole-blood and plasma samples on each monitoring test system. Plasma samples were also tested with ECAA rabbit reference thromboplastin using the manual PT method. The overall MNPT is the geometric mean of PT results for all centers. (b) Overall MNPT (SD) for system C excludes Milan results because a different lot of test strips was used. (c) MNPT (SD) calculated from PT results for 19 healthy blood donors.
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||Hemostasis and Thrombosis|
|Author:||Poller, Leon; Keown, Michelle; Chauhan, Nikhil; van den Besselaar, Anton M.H.P.; Meeuwisse-Braun, Jo|
|Date:||Feb 1, 2002|
|Previous Article:||Methods for measurement of LDL-cholesterol: a critical assessment of direct measurement by homogeneous assays versus calculation.|
|Next Article:||Serum paraoxonase activity: a new additional test for the improved evaluation of chronic liver damage.|