Journal of Thrombosis and Thrombolysis 18(2), 103–107, 2004. C 2004 Kluwer Academic Publishers, Manufactured in The Netherlands.


Analytical Performance of the New Coagulation Monitoring System INRatioTM for the Determination of INR Compared with the Coagulation Monitor Coaguchek
S and an Established Laboratory Method R

Uwe Taborski MD,1 Siegmund L. Braun MD,2 Heinz Voller MD3 ¨ 1 Kerckhoff-Klinik

Bad Nauheim Germany; Herzzentrum Munchen Germany; ¨ 3 Klinik am See Rudersdorf Germany ¨ 2 Deutsches

Abstract. An evaluation of the INRatioTM Prothrombin Time Monitoring system for determination of INR was done in two centers with a total of 5 healthy subjects and 77 subjects on oral anticoagulation. The INRatioTM R and the Coaguchek
S were compared with an established laboratory method. The correlation coefficient of the comparison with the laboratory was r = 0.954 for R INRatioTM and r = 0.937 for Coaguchek
S. The mean relative deviation from the lab method calculated according to Hill was 6.87% for INRatioTM , which is rated R “very good”, and 9.72% for Coaguchek
S (“good”). The imprecision in the normal range (INR = 1.1) showed a coefficient of variation (CV) of 7.8% and a standard deviation (SD) of 0.09. In the therapeutic range (INR 3.9) the CV was 5.4%, the SD 0.21 and above therapeutic range (INR 5.3), the CV was 8.4% (SD 0.44), rated satisfactory. The concordances of the compared methods with the routine were 81% for INRatioTM and 79% R for Coaguchek
S, which is considered state-of-the-art. Most of the patients’ perceptions of the INRatioTM were very positive. In the hands of professionals the INRatioTM demonstrated very good accuracy and precision and an excellent technical reliability. Further studies using INRatioTM for self testing by patients are warranted. Key Words. oral anticoagulation, prothrombin time self monitoring, method comparison, INR determination, point-of-care testing

Introduction More and more patients are on oral anticoagulation therapy worldwide [1–4] and point of care and self testing of INR are becoming more common in many countries. It is desirable that accurate, robust and easy to use monitoring systems are available to determine INR-values from capillary blood samples, both for professional and home use. The new INRatioTM monitor from Hemosense Inc. fulfills all prerequisites as described in the latest guidelines of the German working group on self management of anticoagulation (ASA) [5]. The primary objective of this trial is to evaluate the INRatioTM system for accuracy and precision.

These will be demonstrated by comparative testing with the STA Compact Analyzer with Innovin
R Thromboplastin. The Roche CoaguChe
R S was also compared to the same reference method.

Material and Methods The INRatioTM -Monitoring-System (Producer: Hemosense Inc., Milpitas CA, USA) [6] consists of a small meter and disposable test strips. The meter measures the change in impedance of the bloodreagent-mixture during the process of coagulation and determines the prothrombin time using a programmed algorithm. The INR is calculated using the standard formula, INR = (PT/Mean normal PT)ISI . A drop of about 15 µl of capillary blood is applied to the sample application area on top of the test strip. In addition to the patient test channel the test strip incorporates two quality control channels which contain reagents designed to clot at a predetermined time regardless of the INR of the sample. These on-board controls have been designed to fail if the test strips have been exposed to conditions of high temperature or humidity, verifying the integrity of the test strip. These on-board controls are a new feature of those kinds of systems. After about two minutes, the result is shown on the display, and stored with the time and date in the memory of the meter. Reference testing of the citrated plasma samples was performed with the STA Compact (Manufacturer: Stago; Paris, France) coagulation analyzer using the thromboplastin reagent Innovin
R (DadeBehring, Liederbach, Germany). The other self testing device used in the comparison was the Coaguchek
R S (Roche Diagnostics GmbH,

Address for correspondence: Dr. Uwe Taborski, M.D., Deutsche ¨ Humanplasma mbH, Walzmuhle, ¨ Gesellschaft fur Rheinuferstraße 9 67061 Ludwigshafen, Germany. Tel.: +49 621 595 00 95; Fax: +49 621 595 00 98; E-mail: [email protected] 103

104

Taborski, Braun and V¨oller

Mannheim, Germany) [7]. The Coaguchek
R S system comprises a meter with digital display and disposable test strips. The meter incorporates a detector that recognizes iron oxide particles which move in an oscillating magnetic field, then cease to move after clotting has occurred. To start the reaction, approximately 10 µl of capillary or venous blood is placed on the application field of the test strip. After 1–2 minutes, the INR result is displayed. Quality control is performed using liquid controls on separate strips. 63 anticoagulated subjects were tested in center 1 (Bad Nauheim), 14 anticoagulated subjects and 5 healthy subjects were tested in center 2 (Munich). For all subjects, citrated blood was collected into Sarstedt monovettes
R containing a 3.2% citrate concentration by venous puncture just before the capillary testing. Nine parts of blood were mixed with one part of citrate. The blood was centrifuged within 60 minutes after venous puncture at 4.000 g for 20 minutes, the plasma was separated and shock frozen in portions; the INR determinations of the reference samples were all analyzed in center 1. The capillary measurements were performed as follows: Two INRatioTM instruments (#1 and #2) were used in each site. After capillary punctures of alternate fingers there was the first drop used on instrument 1 and the second on instrument 2. For each subsequent subject, the first drop was alternated between

instrument 2 and instrument 1 to level out possible sampling differences. The same procedure with another finger was used with Coaguchek
R S. The coefficient of variation of the dual measurement was calculated. The statistics for method comparison were determined following the PassingBablok equation [9]. The precision of the INRatioTM system was determined using capillary blood from a healthy blood donor and two patients undergoing oral anticoagulation therapy. Samples were taken multiple fingersticks and the INR values determined for each. All patients were familiar with the self testing of INR values and used BiotrackTM 512, Coaguchek
R or Coaguchek
R S at home. They were asked about their opinion comparing their old meter with the new device. At the beginning of the study all subjects were informed by the investigating physician and signed a consent form. The study was approved by the ethic committee of the Landesaerztekammer Hessen in Frankfurt am Main.

Results 1. Accuracy The following diagram (Fig. 1) shows the results from both centers comparing the INRatioTM and Coaguchek
R S with the reference method of the lab.

Fig. 1. Comparison between the INR tested by capillary methods versus the laboratory method.

Analytical Performance of the New Coagulation Monitoring System INRatioT M

Following the evaluation of the method comparison according to Passing-Bablok [9] the regression equation for INRatioTM was y = 0.98x + 0.08 and for Coaguchek
R S y = 0.87x + 0.174 (n = 82). The correlation coefficient was r = 0.954 (INRatioTM ) and r = 0.937 (Coaguchek
R S). The mean relative deviation (MRD) according to Hill [8] was 6.87% for INRatioTM and 9.72% for Coaguchek
R S.

2. Precision The coefficient of variation found in the normal range (INR = 1.1) was 7.8% with a standard deviation of 0.09 (n = 9). In the high therapeutic range (INR = 3.9) we found a CV of 5.4% with a standard deviation of 0.21 (n = 8). In the high range (INR = 5.3) we found a CV of 8.4% with a standard deviation of 0.44 (n = 9). 3. Concordance Figure 2 depicts the concordance between the INRatio and the CoaguChek versus the laboratory method(STA). Both, the number of patients and the percentage, are shown. The nine-field comparisons for the hypo-, normo- and hyper therapeutic ranges put reference method values against INR-values determined with INRatioTM and Coaguchek
R S (mean of dual values). Based on 62 pair of values from patients from center 1 under oral anticoagulation therapy the concordance for INRatioTM was 81%. For Coaguchek
R S we determined 79%. 4. Patient’s observations 62 patients of center 1 were asked their perception of the new device. The statements in Table 1 were the patients’ spontaneous comments when asked to compare the INRatioTM device to the system actually used by the patient for home monitoring (Coaguchek
R , Coaguchek
R S or BiotrackTM 512). The respondents could give more than one argument. The answers were grouped into specific attributes:

Fig. 2. Nine-field-concordance tables.

105

Table 1. Results of the Patient’s Observations on INRatioTM (n = 58) Cumulative statements

Positive/negative Number of observation statements

Small instrument In-built quality control in each strip Good readability from display Low instrument weight Beautiful design of the instrument Several individual singular positive statements Larger drop of blood required Longer time to wait for result PT% display missing Additional singular negative statements

Positive Positive

51 48

Positive Positive Positive

18 5 2 5

Negative Negative Negative

33 12 3 3

5. General observations During the trial period there was no malfunction or problem with the investigated systems in both centers and no case of an erroneous result generated by any of the meters.

Discussion 1. Method comparison and accuracy Tritscher and Zerback [10,11] found in method comparisons of Coaguchek
R between routinely used lab method HepatoQuick
R and CRM149R, the WHO reference thromboplastin, in a six-center-trial coefficients of variation between r = 0.888 and 0.894 with slopes between 1.07 and 1.38 and absolute differences of −0.5 to −1.1 INR. In the measuring range from 1.2 to 8.0 INR they saw a scattering of ±1.2 INR and ±0.6 INR below 2.0 [11]. The INRatioTM system in this trial demonstrated slightly better

106

Taborski, Braun and V¨oller

coefficients of variation. The investigators also found better results compared to earlier publications with Coaguchek
R S. The slope in this study was near to the ideal value 1.0. The scattering was greater at higher INR values, which is in concordance with earlier publications. As it is known that erroneous results may be caused by pressing and squeezing of the finger, where a capillary sample might initiate coagulation due to the increased release of tissue factor before the sample is applied on to the test strip [11], the skilled evaluators in this study tried to avoid excess squeezing of the finger in collecting the sample.

2. Precision Tritschler and Zerback reported for Coaguchek
R coefficients of variation for imprecision of patient samples and control solutions between 3.5 and 8.7% within the therapeutic range. Earlier authors Jennings, Luddington and Baglin report when evaluating BiotrackTM 512 CVs of 7.5% for an INR at 0.79 and 4.3% in the higher therapeutic range [12]. The results of this trial are in good agreement with the published experiences [13]. 3. Concordance For the coagulation monitor CoaguChek
R Tritscher and Zerback reported a concordance of 81–88% compared with Hepato-Quick
R as reference method performed by professional users and 72% in the hands of patients [10, 11] and conclude for CoaguChek
R S similar results [7], finding as many CoaguChek
R values in the hyper-therapeutic range as in the hypo-therapeutic range compared with normotherapeutic values of the reference method, which indicates appropriate handling by the patient as user [7,11]. INRatioTM showed a concordance of 81% with the reference method, which is in line with published data for CoaguChek
R . For CoaguChek
R S a 79% concordance was found. 4. Patients’ observations For physicians to ask patients opinions is always a little cumbersome and may be biased [14], as they tend to answer more positive than in other environments. It is difficult to quantify answers, and probably more appropriate to weight patient answers against a wellknown standard. Tritschler and Zerback [7] report a patient survey of comparing CoaguChek
R with the new CoaguChek
R S, in which all 58 respondents saw positive improvements in any detail they were questioned about. The results of the survey of this study reflect the spontaneous comments of the subjects in comparison to the system the patients had been using and were pleased with. This could bias the opinion in one or another way. A positive remark can be interpreted as a response to a feature the subject sees

as an obvious improvement. The degree of supporting agreement can be correlated based on the number of times the comment was made. The patient survey confirms that the design goal has been accomplished, to build more safety in the performance of the system by the in-built quality control function. There is also no doubt that the majority of respondents realized that a larger drop of blood is required to fill the threefold channel of the test strip. It is not possible to determine by the answers how patients judged these both features in comparison to each other. A few respondents agreed that the size of the drop of blood required is uncritical if a pain-reduce lancet device is creating a reasonable and sufficiently sized droplet. The INRatioTM monitoring system for the measurement of INR values is innovative and reliable , and fulfills the precision and accuracy requirements of a coagulation device while being very easy to use. The on-board controls are a further step towards more safety and quality assurance for both professional and self testing of INR in oral anticoagulation patients. The INRatioTM device seems suitable for use by a trained patient, and this should be evaluated in further studies.

Acknowledgment The authors thank Hemosense Inc. for supporting this study.

References 1. Horstkotte D, Piper C, Wiemer M. Optimal frequency of patient monitoring and intensity of oral anticoagulation therapy in valvular heart disease. J Thromb Thrombolysis 1998;5(Suppl 1):19–24. 2. Ansell JE, Fish L, Nozzolillo E, Ostrovsky D, Patel N, Peterson AM. Long term patient self-management of oral anticoagulation. Arch Intern Med 1995;155:2185– 2189. 3. Cromheecke M, Colly L, deMol B, et al. Oral anticoagulation self-management and management by specialist anticoagulation clinic: A randomized cross-over comparison. Lancet 2000;356:97–102. 4. K¨orfer R, K¨ortke H. ESCAT. In: Krian A, et al. (eds.), Advancing the Technology of Bileaflet Mechanical Heart Valves. Darmstadt: Steinkopff Verlag;1998:12–22. 5. Bernardo A, V¨oller H. ASA Leitlinien “Gerinnungsmanagement”. DMW 2001;126:346–351. 6. Taborski U, Voeller H, Koertke H, Blunt J. Self management of oral anticoagulation with the INRatio system: LabMed 2004;35:17–21. 7. Hill J. Determination of INR accuracy. Methods of analysis. Hemostasis 1996;26(Suppl 3):422. 8. Passing H, Bablok W. A new biometrical procedure for testing the equality of measurements from different analytical methods. Part 1. J. Clin Chem Clin Biochem 1983;21: 709.

Analytical Performance of the New Coagulation Monitoring System INRatioT M

9. van den Besselaar AMHP, Breddin K, Lutze G, et al. Multicenter evaluation of a new capillary blood PT monitoring system. Blood Coag Fibrinol 1995;6:726–732. 10. Tritschler W, Zerback R. A system for monitoring of Oral Anticoagulant Therapy—Coaguchek S. Roche publication 3064131: 2001. Publicly available on request.

107

11. Tritschler W, Zerback R. Multicenter Evaluation and Patient Handling Study Roche publication 1678523:1999. Publicly available on request. 12. Jennings I, Luddington RJ, Baglin T. Evaluation of ciba cornings biotrack 512. J Clin Pathol 1991;44:950– 953.