- Email: [email protected]
Evaluation of the i-STAT point-of-care capillary whole blood prothrombin time and international normalized ratio: Comparison to the Tcoag MDAII coagulation analyzer in the central laboratory
Clinica Chimica Acta 413 (2012) 955–959
Contents lists available at SciVerse ScienceDirect
Clinica Chimica Acta journal homepage: www.elsevier.com/locate/clinchim
Evaluation of the i-STAT point-of-care capillary whole blood prothrombin time and international normalized ratio: Comparison to the Tcoag MDAII coagulation analyzer in the central laboratory J.A. Peña, K.B. Lewandrowski 1, E.L. Lewandrowski, K. Gregory, J.M. Baron, E.M. Van Cott ⁎, 1 Department of Pathology, Massachusetts General Hospital, Boston, MA 02114, United States
a r t i c l e
i n f o
Article history: Received 5 August 2011 Received in revised form 27 December 2011 Accepted 30 January 2012 Available online 4 February 2012 Keywords: Point-of-care testing POCT Prothrombin time International normalized ratio INR Capillary blood
a b s t r a c t Background: Point-of-care devices for performing a prothrombin time/international normalized ratio (PT/ INR) using capillary blood samples are being increasingly used to monitor patients receiving anticoagulation therapy. However, the performance of some devices has been shown to be suboptimal and there are only limited published data comparing specific devices to various central laboratory coagulation analyzers. We report an evaluation of the iSTAT PT/INR with a comparison to the Tcoag MDA II analyzer. Methods: We obtained simultaneous capillary/venous samples on 20 healthy volunteers for a normal range study and on 50 anticoagulated patients for a clinical evaluation. Testing was performed by phlebotomists. We also obtained 68 near simultaneous capillary/venous test results for assessment of performance by non-laboratory personnel. The criteria for determining clinical equivalence of the iSTAT to the MDA II were (1) same clinical category (subtherapeutic INR b 2, therapeutic INR 2–3, and supratherapeutic INR > 3) or (2) paired values within ≤ 0.4 INR. Results: Forty nine of 50 patient sample pairs collected by phlebotomists showed acceptable clinical agreement. Sixty one (61) of 68 patient sample pairs collected by nurses showed acceptable agreement. In all discordant cases the differences were minor and would have resulted in either no or minimal change in therapy. Conclusions: The iSTAT PT/INR compares well to the MDA II when performed by phlebotomists or nurses. © 2012 Elsevier B.V. All rights reserved.
1. Introduction The laboratory evaluation of coagulation status is important for clinical decision making in a number of situations. The prothrombin time (PT) and the activated partial thromboplastin time (PTT) are commonly used screening tests for coagulation disorders. Monitoring of the PT and INR (International Normalized Ratio) is also utilized to monitor patients receiving anticoagulation therapy, particularly with oral vitamin K antagonists (VKA). The current approach to VKA anticoagulation monitoring requires regularly scheduled measurement of the PT/INR. Point-of-care testing offers several potential advantages over conventional centralized laboratory testing. These include improved patient satisfaction due to the use of a less invasive specimen collection technique (fingerstick instead of venipuncture), no need to train personnel to perform phlebotomy, and the availability of rapid results permitting immediate adjustments to therapy. Finally, when patients and/or their caregivers are properly trained, some POCT devices can be used for self monitoring at home [1]. Despite ⁎ Corresponding author at: Massachusetts General Hospital, 55 Fruit Street, Gray 235, Boston, MA 02114, United States. Tel.: + 1 617 726 9468; fax: + 1 617 726 3256. 1 These authors contributed equally. 0009-8981/$ – see front matter © 2012 Elsevier B.V. All rights reserved. doi:10.1016/j.cca.2012.01.035
these advantages, there have historically been questions concerning the quality of the test results obtained from point-of-care devices for PT/INR testing [2,3]. There is relatively little published data comparing the i-STAT PT/INR test to different coagulation analyzers in the central laboratory. Several small studies have demonstrated acceptable performance when compared to specific central laboratory analyzers albeit with some discrepancies and bias at high INR values [2,4–6]. The present study is the first to compare PT/INR values obtained by i-STAT versus an MDAII analyzer. The purpose of this study was to compare the i-STAT PT/INR to the Tcoag MDAII using fingerstick capillary samples and simultaneous venous samples, respectively. This data is important for organizations considering implementation of POCT and is often difficult to obtain because of the requirement for obtaining split samples from anticoagulated patient volunteers. Our results add additional experience to the limited currently available data concerning the i-STAT PT/INR. 2. Methods The study was conducted in a large, metropolitan multi-specialty general hospital (Massachusetts General Hospital (MGH), Boston, MA). We compared the i-STAT® PT/INR test (Abbott Point-of-Care, Princeton, NJ) using a cartridge with an International Sensitivity
956
J.A. Peña et al. / Clinica Chimica Acta 413 (2012) 955–959
Table 1 Within day (intra-run) precision testing of two i-STAT PT/INR devices.
Normal plasma control
PT
INR
Abnormal plasma control
PT
INR
Mean (s) Range (min, max) SD CV (%) Mean Range (min, max) SD CV (%) Mean (s) Range (min, max) SD CV (%) Mean Range (min, max) SD CV (%)
Device 1
Device 2
13.5 (13, 14.1) 0.37 2.8 1.13 (1.1, 1.2) 0.05 4.3 34.8 (22.8, 36.1) 0.76 2.2 3.07 (3, 3.2) 0.07 2.2
13.4 (12.8, 13.7) 0.27 2 1.1 (1.1, 1.1) 0 0 34.6 (33.8, 35.4) 0.55 1.6 3.1 (3, 3.1) 0.05 1.7
PT: prothrombin time, INR: international normalized ratio.
Index (ISI) of 1.05 to the MDAII coagulation analyzer (Tcoag, Parsippany, NJ) in the central laboratory. Within day "intra-run" imprecision testing for the i-STAT PT/INR test was determined by analyzing reconstituted lyophilized control plasma samples provided by the manufacturer 10 times on two i-STAT devices (i-STAT® PT Control Level 1, which has a normal PT/INR, and Level 2, which has an elevated PT/INR). Day to day “inter-run” imprecision was determined by testing aliquots of control plasma over ten consecutive days including a comparison of results on 3 different i-STAT instruments. For comparison of the iSTAT to the central laboratory, blood was collected into plastic tubes containing 3.2% citrate, and centrifuged at 1500 ×g for 10 min to obtain platelet-poor plasma. PT/INR values were obtained on fresh plasma using an MDAII analyzer and Platelin-L reagent (Tcoag). For i-STAT testing of volunteers and patients a capillary whole blood specimen was collected by fingerstick. To establish the normal range of the i-STAT PT/INR test, blood was acquired from 20 healthy volunteers. The PT/INR was determined for each of the volunteers using both the i-STAT PT/INR instrument (by fingerstick) and the core laboratory automated MDAII instrument using a simultaneously collected plasma sample. The Shapiro–Wilk test was used to test for normal distribution of the data. To assess correlation of the i-STAT PT/INR to the core laboratory MDAII in adult patients receiving Coumadin, we collected simultaneous
capillary fingerstick whole blood samples and citrated plasma samples on 50 volunteer patients from our outpatient Anticoagulation Management Service (AMS) at MGH. The patients consisted of 36 men and 14 women with ages ranging from 27 to 92 y (median age = 66.5 y). The POCT testing was performed by trained phlebotomists in our outpatient specimen collection center according to the manufacturer's instructions. All patients in this group had a target INR of 2.0–3.0. Patients were asked if they were taking other anticoagulation medications including enoxaparin (Lovenox®), dalteparin (Fragmin®), or heparin and if they ever had been diagnosed with a lupus anticoagulant which can interfere with PT/INR determinations. Patients with any of these confounding variables were excluded from the study. The patient test results were assigned to clinical categories based on the INR: subtherapeutic if the INR was b2, therapeutic if the INR was 2.0–3.0 and supra-therapeutic if the INR was >3. Paired INR values were considered in clinical agreement or concordant if they fell within the same clinical category or ≤0.4 INR of each other. This method of determining INR agreement has been in use for a number of years [7,8]. Recently, a new method has been proposed by Shermock et al., therefore in addition, results were also analyzed with this method [9]. With this method, a pair is considered in agreement if both INR results are within one of the following ranges: b1.9, 1.9–3.3, 3.4–5.5, 5.6–9.0, or >9.0. To assess the performance of the i-STAT PT/INR when performed by non-laboratory personnel, we compared 68 i-STAT PT/INR results performed by nurses to central laboratory MDAII results performed on the same day in patients presenting for an interventional procedure (e.g. arrhythmia ablation or pacemaker placement) to the Electrophysiology Laboratory (EPL) at the MGH. To further examine elevated INR values, an additional 25 consecutive pairs of results with INR > 3 were collected over the subsequent 4 month period in the Electrophysiology Laboratory. 3. Results Within day (intra-run) imprecision for the i-STAT PT/INR is shown in Table 1. For the PT the coefficient of variation (CV) ranged from 1.6 to 2.8% and for the INR from 0 to 4.3%. Day to day (inter-run) imprecision is shown in Table 2. For the PT the overall CV ranged from 3.9 to 4.4% and for the INR 4.9 to 5.2%. In comparison, the MDAII analyzer intra-run PT imprecision was 0.39–0.75% and the inter-run imprecision was 1.2–2.2%. To establish the normal range for the i-STAT PT and INR, venous blood specimens from healthy volunteers were analyzed on both
Table 2 Day to day (inter-run) precision testing of three i-STAT PT/INR devices.
Normal plasma control
PT
INR
Abnormal plasma control
PT
INR
PT: prothrombin time, INR: international normalized ratio.
Mean (s) Range (min, max) SD CV (%) Inter-device CV Mean Range (min, max) SD CV (%) Inter-device CV Mean (s) Range (min, max) SD CV (%) Inter-device CV Mean Range (min, max) SD CV (%) Inter-device CV
Device 1
Device 2
Device 3
13.08 (12.3, 14) 0.56 4.28
13.43 (12.7, 14.3) 0.48 3.57 3.9 1.13 (1.1, 1.2) 0.05 0.04 5.2 30.2 (28.6, 32.6) 1.3 4.30 4.4 2.65 (2.5, 2.9) 0.13 4.91 4.9
12.99 (12.5, 13.8) 0.45 3.46
1.1 (1, 1.2) 0.07 6.36 29.98 (29, 32.4) 1.12 3.74 2.62 (2.5, 2.8) 0.1 3.82
1.09 (1, 1.2) 0.06 0.06 30 (28.5, 33.8) 1.65 5.50 2.63 (2.5, 3) 0.16 6.08
J.A. Peña et al. / Clinica Chimica Acta 413 (2012) 955–959 Table 3 Determination of in-house normal ranges for the i-STAT and MDAII instruments using healthy volunteer plasma samples.
PT
INR
Mean (s) Median Range (min, max) SD Mean Median Range (min, max) SD
i-STAT
MDAII
12.78 12.60 (11.3, 15.2) 1.09 1.07 1.06 (0.9, 1.3) 0.11
12.84 12.85 (11.4, 14.2) 0.69 1.06 1.06 (0.9, 1.2) 0.07
957
Table 5 Comparison of the uncorrected and corrected INR from the i-STAT analyzer against the MDAII analyzer INR.
INR
Mean Median Range (min, max) SD
i-STAT Uncorrected
i-STAT Corrected
MDAII Analyzer
2.79 2.8 (1.4, 5.6) 0.84
2.61 2.6 (1.3, 5.3) 0.79
2.60 2.5 (1.3, 4.2) 0.76
INR: international normalized ratio. Corrected: Using locally-determined, devicespecific mean normal PT rather than the manufacturer-supplied mean normal PT to calculate the INR.
PT: prothrombin time, INR: international normalized ratio.
the i-STAT and the MDAII (Table 3). For both methods, the PT was found to be consistent with an underlying normal distribution based upon both the Shapiro–Wilk test and visual inspection of a histogram of the data. Therefore, we calculated normal ranges based upon the mean +/− 2SD for the PT. The i-STAT PT/INR normal range was 10.6–15.0 s for the PT, corresponding to an INR of 0.88–1.26. In comparison, the MDA normal range based upon the volunteers was 11.5 to 14.2 s, corresponding to an INR of 0.92–1.19. Of note, the mean INR for both the i-STAT and core laboratory MDAII was almost the same (1.06 vs 1.07), and all twenty paired results agreed within 0.2 INR. The performance of the i-STAT PT/INR using capillary samples from 50 anticoagulated patients was compared to the MDAII using simultaneous citrated venous blood specimens as shown in Table 4. The i-STAT calculates the INR from the whole blood PT as well as mean normal i-STAT PT (12.0 s) and the cartridge ISI of 1.05. Using the formula: ISI
INR ¼ ½Patient i−STAT PTðsÞ=Mean normal i−STAT PTðsÞ
The mean normal PT of 12.0 s is supplied by the manufacturer and is the same for all i-STAT devices. We determined a corrected i-STAT INR using the mean normal i-STAT PT (12.78 s) that we established locally, as described above, on healthy volunteers. The mean corrected INR values for the i-STAT device is closer to the MDAII INR than the uncorrected MDAII INR (Table 5). A Bland–Altman graph showed that i-STAT INR tended to slightly overestimate the core laboratory MDAII INR (Fig. 1). When using the corrected i-STAT INR, the positive bias is somewhat mitigated. Correlation of the paired INR values was determined. Results showed y = 0.81x + 0.35 (R = 0.90, R 2 = 0.80). Using the corrected i-STAT INR did not significantly change the correlation. Forty-nine (49) of 50 uncorrected pairs (98%) showed acceptable clinical agreement or concordance (Table 6). The one pair that did not meet these criteria had an i-STAT INR of 3.2 and a core laboratory INR of 2.6. Using a “corrected” i-STAT INR the value would be 3.0 thus bringing the discordant value into agreement.
Table 4 Comparison of PT/INR values from patients taking warfarin and monitored by the Massachusetts General Hospital Anticoagulation Management Service.
PT
INR
Mean (s) Median Range (min, max) SD Mean Median Range (min, max) SD
i-STAT
MDAII
31.81 31.8 (16.4, 62.4) 9.2 2.79 2.8 (1.4, 5.6) 0.84
27.25 26.6 (15, 41.1) 6.72 2.60 2.5 (1.3, 4.2) 0.76
PT: prothrombin time, INR: international normalized ratio.
Looking at the two criteria for agreement individually, which is more stringent, 45 of 50 uncorrected pairs (90%) were within the same clinical category and 40 of 50 uncorrected pairs (80%) were within 0.4 INR of each other. Using the Shermock method, which uses slightly different definitions for clinical categories, 44 of 50 pairs (88%) were in the same clinical category. Three of the 6 discrepant pairs were within 0.4 INR of each other, 1 was within 0.5 INR, and the remaining two pairs were 5.6 iSTAT vs 4.2 laboratory and 4.3 vs 3.1. Thus 94% of pairs were in the same clinical category as defined by Shermock or within 0.4 INR of each other. As our policy requires that a laboratory INR be performed for patients with iSTAT INR > 4, the 2 remaining discrepant pairs would qualify for a reflexed laboratory INR. To assess the performance of the i-STAT PT/INR when performed by non-laboratory personnel at the point-of-care we compared capillary fingerstick samples from 68 patients in our electrophysiology laboratory to either a simultaneous or same day plasma value performed in our central laboratory. Correlation of the i-STAT INR to the MDAII showed i-STAT = 0.77 MDAII + 0.26, R 2 = 0.80, R = 0.90. Sixty-one of 68 pairs (89.7%) were in the same clinical category or within 0.4 INR of each other (looking at the acceptance criteria separately, 83.8% of pairs were in the same clinical category and 79.4% were within 0.4 INR of each other). Review of the medical records for the 7 discrepant pairs revealed that 5 of the discrepancies were due to warfarin or vitamin K administration shortly after one INR was performed but hours before the other. Thus, there were only 2 true discrepant pairs (61/63 = 96.8% agreement). For one discrepant pair, the laboratory INR was 2.2 and the i-STAT INR was 1.6; for the other pair the laboratory INR was 2.3 and the i-STAT INR was 1.7. Using the Shermock method, 79.4% (54/68) of pairs were in the same clinical category. Six of the discrepancies were within 0.4 INR (in fact, all 6 were within 0.1–0.3 INR) and 5 were due to warfarin or vitamin K as described above, leaving 3 discrepant pairs. Two of the pairs are the same as described above; for the third pair, the laboratory INR was 1.9 and the iSTAT INR was 1.4. To look more closely at elevated INR values, we continued to collect paired specimen results for INR> 3 for an additional 4 month time period, as performed by non-laboratory personnel at the point-of-care. Twenty-five paired results were obtained, 60% of which were INR 3–3.9, 32% were 4–4.9, and 8% were ≥5. Agreement among pairs was 92.0% (looking at the two agreement criteria separately, 80.0% were in the same clinical category and 72.0% were within 0.4 INR of each other). Using the Shermock method, 68.0% were in the same clinical category, but 5 of 8 of the discrepant pairs were within 0.4 INR of each other, leaving 3 of 25 discrepant pairs (88.0% agreement). Shermock et al. reported that 2 point-of-care INR devices, Hemochron Junior and Hemochron Signature Elite, appeared incapable of reporting 7 specific INR values, for unclear reasons (namely, INR 2.1, 2.7, 3.1, 3.5, 3.8, 4.1, and 4.4.) [9]. Therefore, we reviewed all INR results generated by the i-STAT in this range. The i-STAT generated every possible INR value from 1.0 to 4.5, and thus does not appear to have the same problem as the Hemochron devices.
958
J.A. Peña et al. / Clinica Chimica Acta 413 (2012) 955–959
1 uncorrected bias corrected bias Median uncorrected bias
Difference
Median corrected bias
0 0
1
2
3
4
5
-1
INR Fig. 1. Bland Altman Plot comparing the i-STAT international normalized ratio (INR) to the MDAII using both uncorrected and corrected INR values. The difference (Y-axis) indicates the i-STAT INR minus the MDAII central laboratory INR. The INR (X-axis) indicates the mean of the i-STAT and the MDAII INR. Shading indicates the typical therapeutic range for warfarin. Uncorrected INR values are INR values generated by the i-STAT (which uses the manufacturer's assigned mean normal PT). Corrected INR values use the mean normal PT that we established locally on the i-STAT device used in this study. Corrected INR values more closely matched the central laboratory INR values than did the uncorrected INR values. The horizontal line of zero bias is shown but is visually indistinguishable from the median corrected bias line at − 0.002.
4. Discussion Obtaining validation data for point-of-care PT/INR tests that use capillary fingerstick whole blood specimens is challenging because the studies require split sample testing from anticoagulated and normal subjects. Validation also requires comparison to the specific coagulation analyzer used in the central laboratory because the correlation data cannot be generalized across instruments from different manufacturers. There is a paucity of published data comparing the i-STAT PT/INR to various different instruments including no prior data using the MDAII. Most studies have reported data using only a small number of subjects such that the cumulative experience with any one instrument is limited. The utility of any POCT device, including those used for determining the PT/INR, depends on the ability of the device to measure the analyte with clinically acceptable accuracy and precision. In this study, we have shown that the i-STAT PT/INR, can be reliably utilized at the point-ofcare by both phlebotomists and by nursing staff in a clinical setting. Specifically the i-STAT PT/INR exhibits only a modestly higher imprecision than the central laboratory MDAII analyzer. All 50 split sample specimens performed by phlebotomists were in the same clinical category or within 0.4 INR as the laboratory result using a “corrected” PT/INR. In the case of tests performed by nurses 61 of 63 (96.8%) values were in the same clinical category or within 0.4 INR. There were 2 discrepant Table 6 Concordance of clinical categories based on INR for patients followed by the AMS. Therapeutic
Suprab
INR from i-STAT uncorrected Suba 7 Therapeutic 3 b Supra 0
0 24 3
0 0 13
INR from i-STAT corrected Suba Therapeutic Suprab
3 23 1
0 0 13
INR from MDAII
Suba
9 1 0
INR: international normalized ratio. AMS: Anticoagulation Management Service. a Sub-therapeutic INR (b2.0). b Supra-therapeutic INR (> 3.0).
cases that could potentially have resulted in a minor adjustment to therapy. Of note not all specimens performed by the nurses were truly simultaneous. A number of these were simply performed within the same day. For this reason the data on nursing-performed testing might show better correlation if simultaneous paired specimens could be obtained. Nevertheless, these results are considerably better than those reported by Karon et al. [2] (76% agreement and 24% discordant: nurses performed the i-STAT POCT) and by Donaldson et al. [4] (46% agreement and 54% discordant: a pharmacist performed the i-STAT POCT), using other central laboratory coagulation analyzers. There is relatively little published data comparing the i-STAT PT/ INR to central laboratory instruments. In the study by Donaldson et al. [4] the i-STAT and CoaguChek XS Plus PT/INR were compared to the Stago coagulation analyzer (model of instrument not specified) on 50 patients. Both POC devices exhibited a positive bias in INR values when compared to the Stago and the bias increased at higher INR values. This necessitated reflexing all POCT values greater than 3.5 to a central laboratory for confirmation. In the subsequent study of the i-STAT by Karon et al. of 50 patients [2] comparing the iSTAT to the MDA-180 analyzer the authors concluded that the overall agreement between the POC and laboratory values was “very good”. They also noted that newer POCT PT/INR devices appear to perform better than older devices. They used an older version of the MDAII called the MDA-180, and a different reagent than the present study (Innovin reagent from Siemens). The results presented in our study are consistent with the findings reported by Karon et al. and expand on the limited available database for the i-STAT PT/INR. Specifically the i-STAT PT/INR performs with acceptable clinical reliability in the hands of phlebotomists and non-laboratory nursing personnel. While the i-STAT PT/INR appears to have a positive bias at higher INR values there is good concordance to the MDAII. Eighty eight percent of i-STAT values were in the same clinical category as the MDAII and 98% were within the same category or within 0.4 INR of each other. This concordance improves when the corrected i-STAT INR is used (90% within the same clinical category or 100% if the criterion of within 0.4 INR is also used). The recurring finding that POCT devices, including the i-STAT system, impart a bias compared to laboratory instrumentation might, in
J.A. Peña et al. / Clinica Chimica Acta 413 (2012) 955–959
part, be due to a calculation issue. For example, with the i-STAT, the INR is calculated from a specimen's PT, and the divisor in the equation (the mean normal i-STAT PT) is set to be 12.0 s as provided by the manufacturer. However, this mean normal PT of 12.0 s is not what we have established within our institution on the i-STAT analyzer. Our mean normal i-STAT PT is 12.78 s. Using this device-specific mean normal, the bias is partially mitigated. Previous studies have shown that POCT devices for anticoagulation monitoring can be of great utility, especially for patients on long-term anticoagulation [5,10]. These devices have also been shown to be of use in the acute setting as well, particularly for determining thrombolytic therapy eligibility in anticoagulated patients who have suffered a cerebrovascular accident [6,10,11]. Although automated laboratory instruments have been shown to have greater precision compared to POCT devices [12], the i-STAT PT/INR device has been shown to have sufficiently acceptable precision and reproducibility for use in clinical decision making for patients receiving oral VKA therapies [5,6]. In the acute setting, the i-STAT PT/INR has also been shown to be useful in patients presenting with acute stroke symptoms who are eligible for thrombolytic therapy [6,11]. The selection of a particular POCT device for implementation in either outpatient or inpatient use requires careful assessment of individual devices and assessment of how compatible the device is with each institutions existing anticoagulation instrument in the central laboratory. Some POCT devices for PT have been shown to have good correlation with laboratory instruments, but at elevated INRs, this concordance is less reliable [13] which may necessitate reflexing specimens with high INR values to the central laboratory. i-STAT imprecision was performed by laboratory personnel, which might be seen as a limitation of our study, since the precision seen might not be generalizable to real clinical practice settings. This potential limitation is at least partially overcome in our study because subsequent INR testing was performed by both phlebotomy and nursing staff. In summary, we show that the i-STAT PT/INR provides rapid and reliable PT/INR testing to aid in clinical decision making. While the INR determined by this POCT device generally leads to a slightly higher INR, the clinical category in which a patient is placed has 88% concordance with the MDAII analyzer and 98% concordance when the criterion of “same clinical category or within 0.4 INR” is used. Correction of i-STAT INR values using a locally determined mean normal i-STAT PT improves the concordance between the two analyzers. Our data expands on the
959
currently available studies in the literature and should assist clinical services who are considering implementation of the i-STAT POCT PT/INR.
Acknowledgments Abbott Point-of-Care supplied some of the i-STAT cartridges used in this study.
References [1] Newall F, Monagle P, Johnston L. Home INR monitoring of oral anticoagulant therapy in children using the CoaguChek S point-of-care monitor and a robust education program. Thromb Res 2006;118:587–93. [2] Karon BS, McBane RD, Chaudhry R, Beyer LK, Santrach PJ. Accuracy of capillary whole blood international normalized ratio on the CoaguChek S, CoaguChek XS, and i-STAT 1 point-of-care analyzers. Am J Clin Pathol 2008;130:88–92. [3] Khoschnewis S, Hannes FM, Tschopp M, Wuillemin WA. INR comparison between the CoaguChek Pro PT(N) and a standard laboratory method. Thromb Res 2004;113:327–32. [4] Donaldson M, Sullivan J, Norbeck A. Comparison of International Normalized Ratios provided by two point-of-care devices and laboratory-based venipuncture in a pharmacist-managed anticoagulation clinic. Am J Health Syst Pharm 2010;67: 1616–22. [5] Boehlen F, Reber G, de Moerloose P. Agreement of a new whole-blood PT/INR test using capillary samples with plasma INR determinations. Thromb Res 2005;115: 131–4. [6] Drescher MJ, Spence A, Rockwell D, Staff I, Smally AJ. Point-of-care testing for coagulation studies in a stroke protocol: a time-saving innovation. Am J Emerg Med 2011;29:82–5. [7] Anderson DR, Harrison L, Hirsh J. Evaluation of a portable prothrombin time monitor for home use by patients who require long-term oral anticoagulant therapy. Arch Intern Med 1993;153:1441–7. [8] van den Besselaar AM. A comparison of INRs determined with a whole blood prothrombin time device and two international reference preparations for thromboplastin. Thromb Haemost 2000;84:410–2. [9] Shermock KM, Streiff MB, Pinto BL, Kraus P, Pronovost PJ. Novel analysis of clinically relevant diagnostic errors in point-of-care devices. J Thromb Haemost 2011;9:1769–75. [10] Green TL, Mansoor A, Newcommon N, Stephenson C, Stewart E, Hill MD. Reliability of point-of-care testing of INR in acute stroke. Can J Neurol Sci 2009;35:348–51. [11] Rizos T, Herweh C, Jenetzky E, et al. Point-of-care international normalized ratio testing accelerates thrombolysis in patients with acute ischemic stroke using oral anticoagulants. Stroke 2009;40:3547–51. [12] van den Besselaar AM. Accuracy, precision, and quality control for point-of-care testing of oral anticoagulation. J Thromb Thrombolysis 2001;12:35–40. [13] Ryan F, O'Shea S, Byrne S. The reliability of point-of-care prothrombin time testing. A comparison of CoaguChek S and XS INR measurements with hospital laboratory monitoring. Int J Lab Hematol 2010;32:e26–33.
Contents lists available at SciVerse ScienceDirect
Clinica Chimica Acta journal homepage: www.elsevier.com/locate/clinchim
Evaluation of the i-STAT point-of-care capillary whole blood prothrombin time and international normalized ratio: Comparison to the Tcoag MDAII coagulation analyzer in the central laboratory J.A. Peña, K.B. Lewandrowski 1, E.L. Lewandrowski, K. Gregory, J.M. Baron, E.M. Van Cott ⁎, 1 Department of Pathology, Massachusetts General Hospital, Boston, MA 02114, United States
a r t i c l e
i n f o
Article history: Received 5 August 2011 Received in revised form 27 December 2011 Accepted 30 January 2012 Available online 4 February 2012 Keywords: Point-of-care testing POCT Prothrombin time International normalized ratio INR Capillary blood
a b s t r a c t Background: Point-of-care devices for performing a prothrombin time/international normalized ratio (PT/ INR) using capillary blood samples are being increasingly used to monitor patients receiving anticoagulation therapy. However, the performance of some devices has been shown to be suboptimal and there are only limited published data comparing specific devices to various central laboratory coagulation analyzers. We report an evaluation of the iSTAT PT/INR with a comparison to the Tcoag MDA II analyzer. Methods: We obtained simultaneous capillary/venous samples on 20 healthy volunteers for a normal range study and on 50 anticoagulated patients for a clinical evaluation. Testing was performed by phlebotomists. We also obtained 68 near simultaneous capillary/venous test results for assessment of performance by non-laboratory personnel. The criteria for determining clinical equivalence of the iSTAT to the MDA II were (1) same clinical category (subtherapeutic INR b 2, therapeutic INR 2–3, and supratherapeutic INR > 3) or (2) paired values within ≤ 0.4 INR. Results: Forty nine of 50 patient sample pairs collected by phlebotomists showed acceptable clinical agreement. Sixty one (61) of 68 patient sample pairs collected by nurses showed acceptable agreement. In all discordant cases the differences were minor and would have resulted in either no or minimal change in therapy. Conclusions: The iSTAT PT/INR compares well to the MDA II when performed by phlebotomists or nurses. © 2012 Elsevier B.V. All rights reserved.
1. Introduction The laboratory evaluation of coagulation status is important for clinical decision making in a number of situations. The prothrombin time (PT) and the activated partial thromboplastin time (PTT) are commonly used screening tests for coagulation disorders. Monitoring of the PT and INR (International Normalized Ratio) is also utilized to monitor patients receiving anticoagulation therapy, particularly with oral vitamin K antagonists (VKA). The current approach to VKA anticoagulation monitoring requires regularly scheduled measurement of the PT/INR. Point-of-care testing offers several potential advantages over conventional centralized laboratory testing. These include improved patient satisfaction due to the use of a less invasive specimen collection technique (fingerstick instead of venipuncture), no need to train personnel to perform phlebotomy, and the availability of rapid results permitting immediate adjustments to therapy. Finally, when patients and/or their caregivers are properly trained, some POCT devices can be used for self monitoring at home [1]. Despite ⁎ Corresponding author at: Massachusetts General Hospital, 55 Fruit Street, Gray 235, Boston, MA 02114, United States. Tel.: + 1 617 726 9468; fax: + 1 617 726 3256. 1 These authors contributed equally. 0009-8981/$ – see front matter © 2012 Elsevier B.V. All rights reserved. doi:10.1016/j.cca.2012.01.035
these advantages, there have historically been questions concerning the quality of the test results obtained from point-of-care devices for PT/INR testing [2,3]. There is relatively little published data comparing the i-STAT PT/INR test to different coagulation analyzers in the central laboratory. Several small studies have demonstrated acceptable performance when compared to specific central laboratory analyzers albeit with some discrepancies and bias at high INR values [2,4–6]. The present study is the first to compare PT/INR values obtained by i-STAT versus an MDAII analyzer. The purpose of this study was to compare the i-STAT PT/INR to the Tcoag MDAII using fingerstick capillary samples and simultaneous venous samples, respectively. This data is important for organizations considering implementation of POCT and is often difficult to obtain because of the requirement for obtaining split samples from anticoagulated patient volunteers. Our results add additional experience to the limited currently available data concerning the i-STAT PT/INR. 2. Methods The study was conducted in a large, metropolitan multi-specialty general hospital (Massachusetts General Hospital (MGH), Boston, MA). We compared the i-STAT® PT/INR test (Abbott Point-of-Care, Princeton, NJ) using a cartridge with an International Sensitivity
956
J.A. Peña et al. / Clinica Chimica Acta 413 (2012) 955–959
Table 1 Within day (intra-run) precision testing of two i-STAT PT/INR devices.
Normal plasma control
PT
INR
Abnormal plasma control
PT
INR
Mean (s) Range (min, max) SD CV (%) Mean Range (min, max) SD CV (%) Mean (s) Range (min, max) SD CV (%) Mean Range (min, max) SD CV (%)
Device 1
Device 2
13.5 (13, 14.1) 0.37 2.8 1.13 (1.1, 1.2) 0.05 4.3 34.8 (22.8, 36.1) 0.76 2.2 3.07 (3, 3.2) 0.07 2.2
13.4 (12.8, 13.7) 0.27 2 1.1 (1.1, 1.1) 0 0 34.6 (33.8, 35.4) 0.55 1.6 3.1 (3, 3.1) 0.05 1.7
PT: prothrombin time, INR: international normalized ratio.
Index (ISI) of 1.05 to the MDAII coagulation analyzer (Tcoag, Parsippany, NJ) in the central laboratory. Within day "intra-run" imprecision testing for the i-STAT PT/INR test was determined by analyzing reconstituted lyophilized control plasma samples provided by the manufacturer 10 times on two i-STAT devices (i-STAT® PT Control Level 1, which has a normal PT/INR, and Level 2, which has an elevated PT/INR). Day to day “inter-run” imprecision was determined by testing aliquots of control plasma over ten consecutive days including a comparison of results on 3 different i-STAT instruments. For comparison of the iSTAT to the central laboratory, blood was collected into plastic tubes containing 3.2% citrate, and centrifuged at 1500 ×g for 10 min to obtain platelet-poor plasma. PT/INR values were obtained on fresh plasma using an MDAII analyzer and Platelin-L reagent (Tcoag). For i-STAT testing of volunteers and patients a capillary whole blood specimen was collected by fingerstick. To establish the normal range of the i-STAT PT/INR test, blood was acquired from 20 healthy volunteers. The PT/INR was determined for each of the volunteers using both the i-STAT PT/INR instrument (by fingerstick) and the core laboratory automated MDAII instrument using a simultaneously collected plasma sample. The Shapiro–Wilk test was used to test for normal distribution of the data. To assess correlation of the i-STAT PT/INR to the core laboratory MDAII in adult patients receiving Coumadin, we collected simultaneous
capillary fingerstick whole blood samples and citrated plasma samples on 50 volunteer patients from our outpatient Anticoagulation Management Service (AMS) at MGH. The patients consisted of 36 men and 14 women with ages ranging from 27 to 92 y (median age = 66.5 y). The POCT testing was performed by trained phlebotomists in our outpatient specimen collection center according to the manufacturer's instructions. All patients in this group had a target INR of 2.0–3.0. Patients were asked if they were taking other anticoagulation medications including enoxaparin (Lovenox®), dalteparin (Fragmin®), or heparin and if they ever had been diagnosed with a lupus anticoagulant which can interfere with PT/INR determinations. Patients with any of these confounding variables were excluded from the study. The patient test results were assigned to clinical categories based on the INR: subtherapeutic if the INR was b2, therapeutic if the INR was 2.0–3.0 and supra-therapeutic if the INR was >3. Paired INR values were considered in clinical agreement or concordant if they fell within the same clinical category or ≤0.4 INR of each other. This method of determining INR agreement has been in use for a number of years [7,8]. Recently, a new method has been proposed by Shermock et al., therefore in addition, results were also analyzed with this method [9]. With this method, a pair is considered in agreement if both INR results are within one of the following ranges: b1.9, 1.9–3.3, 3.4–5.5, 5.6–9.0, or >9.0. To assess the performance of the i-STAT PT/INR when performed by non-laboratory personnel, we compared 68 i-STAT PT/INR results performed by nurses to central laboratory MDAII results performed on the same day in patients presenting for an interventional procedure (e.g. arrhythmia ablation or pacemaker placement) to the Electrophysiology Laboratory (EPL) at the MGH. To further examine elevated INR values, an additional 25 consecutive pairs of results with INR > 3 were collected over the subsequent 4 month period in the Electrophysiology Laboratory. 3. Results Within day (intra-run) imprecision for the i-STAT PT/INR is shown in Table 1. For the PT the coefficient of variation (CV) ranged from 1.6 to 2.8% and for the INR from 0 to 4.3%. Day to day (inter-run) imprecision is shown in Table 2. For the PT the overall CV ranged from 3.9 to 4.4% and for the INR 4.9 to 5.2%. In comparison, the MDAII analyzer intra-run PT imprecision was 0.39–0.75% and the inter-run imprecision was 1.2–2.2%. To establish the normal range for the i-STAT PT and INR, venous blood specimens from healthy volunteers were analyzed on both
Table 2 Day to day (inter-run) precision testing of three i-STAT PT/INR devices.
Normal plasma control
PT
INR
Abnormal plasma control
PT
INR
PT: prothrombin time, INR: international normalized ratio.
Mean (s) Range (min, max) SD CV (%) Inter-device CV Mean Range (min, max) SD CV (%) Inter-device CV Mean (s) Range (min, max) SD CV (%) Inter-device CV Mean Range (min, max) SD CV (%) Inter-device CV
Device 1
Device 2
Device 3
13.08 (12.3, 14) 0.56 4.28
13.43 (12.7, 14.3) 0.48 3.57 3.9 1.13 (1.1, 1.2) 0.05 0.04 5.2 30.2 (28.6, 32.6) 1.3 4.30 4.4 2.65 (2.5, 2.9) 0.13 4.91 4.9
12.99 (12.5, 13.8) 0.45 3.46
1.1 (1, 1.2) 0.07 6.36 29.98 (29, 32.4) 1.12 3.74 2.62 (2.5, 2.8) 0.1 3.82
1.09 (1, 1.2) 0.06 0.06 30 (28.5, 33.8) 1.65 5.50 2.63 (2.5, 3) 0.16 6.08
J.A. Peña et al. / Clinica Chimica Acta 413 (2012) 955–959 Table 3 Determination of in-house normal ranges for the i-STAT and MDAII instruments using healthy volunteer plasma samples.
PT
INR
Mean (s) Median Range (min, max) SD Mean Median Range (min, max) SD
i-STAT
MDAII
12.78 12.60 (11.3, 15.2) 1.09 1.07 1.06 (0.9, 1.3) 0.11
12.84 12.85 (11.4, 14.2) 0.69 1.06 1.06 (0.9, 1.2) 0.07
957
Table 5 Comparison of the uncorrected and corrected INR from the i-STAT analyzer against the MDAII analyzer INR.
INR
Mean Median Range (min, max) SD
i-STAT Uncorrected
i-STAT Corrected
MDAII Analyzer
2.79 2.8 (1.4, 5.6) 0.84
2.61 2.6 (1.3, 5.3) 0.79
2.60 2.5 (1.3, 4.2) 0.76
INR: international normalized ratio. Corrected: Using locally-determined, devicespecific mean normal PT rather than the manufacturer-supplied mean normal PT to calculate the INR.
PT: prothrombin time, INR: international normalized ratio.
the i-STAT and the MDAII (Table 3). For both methods, the PT was found to be consistent with an underlying normal distribution based upon both the Shapiro–Wilk test and visual inspection of a histogram of the data. Therefore, we calculated normal ranges based upon the mean +/− 2SD for the PT. The i-STAT PT/INR normal range was 10.6–15.0 s for the PT, corresponding to an INR of 0.88–1.26. In comparison, the MDA normal range based upon the volunteers was 11.5 to 14.2 s, corresponding to an INR of 0.92–1.19. Of note, the mean INR for both the i-STAT and core laboratory MDAII was almost the same (1.06 vs 1.07), and all twenty paired results agreed within 0.2 INR. The performance of the i-STAT PT/INR using capillary samples from 50 anticoagulated patients was compared to the MDAII using simultaneous citrated venous blood specimens as shown in Table 4. The i-STAT calculates the INR from the whole blood PT as well as mean normal i-STAT PT (12.0 s) and the cartridge ISI of 1.05. Using the formula: ISI
INR ¼ ½Patient i−STAT PTðsÞ=Mean normal i−STAT PTðsÞ
The mean normal PT of 12.0 s is supplied by the manufacturer and is the same for all i-STAT devices. We determined a corrected i-STAT INR using the mean normal i-STAT PT (12.78 s) that we established locally, as described above, on healthy volunteers. The mean corrected INR values for the i-STAT device is closer to the MDAII INR than the uncorrected MDAII INR (Table 5). A Bland–Altman graph showed that i-STAT INR tended to slightly overestimate the core laboratory MDAII INR (Fig. 1). When using the corrected i-STAT INR, the positive bias is somewhat mitigated. Correlation of the paired INR values was determined. Results showed y = 0.81x + 0.35 (R = 0.90, R 2 = 0.80). Using the corrected i-STAT INR did not significantly change the correlation. Forty-nine (49) of 50 uncorrected pairs (98%) showed acceptable clinical agreement or concordance (Table 6). The one pair that did not meet these criteria had an i-STAT INR of 3.2 and a core laboratory INR of 2.6. Using a “corrected” i-STAT INR the value would be 3.0 thus bringing the discordant value into agreement.
Table 4 Comparison of PT/INR values from patients taking warfarin and monitored by the Massachusetts General Hospital Anticoagulation Management Service.
PT
INR
Mean (s) Median Range (min, max) SD Mean Median Range (min, max) SD
i-STAT
MDAII
31.81 31.8 (16.4, 62.4) 9.2 2.79 2.8 (1.4, 5.6) 0.84
27.25 26.6 (15, 41.1) 6.72 2.60 2.5 (1.3, 4.2) 0.76
PT: prothrombin time, INR: international normalized ratio.
Looking at the two criteria for agreement individually, which is more stringent, 45 of 50 uncorrected pairs (90%) were within the same clinical category and 40 of 50 uncorrected pairs (80%) were within 0.4 INR of each other. Using the Shermock method, which uses slightly different definitions for clinical categories, 44 of 50 pairs (88%) were in the same clinical category. Three of the 6 discrepant pairs were within 0.4 INR of each other, 1 was within 0.5 INR, and the remaining two pairs were 5.6 iSTAT vs 4.2 laboratory and 4.3 vs 3.1. Thus 94% of pairs were in the same clinical category as defined by Shermock or within 0.4 INR of each other. As our policy requires that a laboratory INR be performed for patients with iSTAT INR > 4, the 2 remaining discrepant pairs would qualify for a reflexed laboratory INR. To assess the performance of the i-STAT PT/INR when performed by non-laboratory personnel at the point-of-care we compared capillary fingerstick samples from 68 patients in our electrophysiology laboratory to either a simultaneous or same day plasma value performed in our central laboratory. Correlation of the i-STAT INR to the MDAII showed i-STAT = 0.77 MDAII + 0.26, R 2 = 0.80, R = 0.90. Sixty-one of 68 pairs (89.7%) were in the same clinical category or within 0.4 INR of each other (looking at the acceptance criteria separately, 83.8% of pairs were in the same clinical category and 79.4% were within 0.4 INR of each other). Review of the medical records for the 7 discrepant pairs revealed that 5 of the discrepancies were due to warfarin or vitamin K administration shortly after one INR was performed but hours before the other. Thus, there were only 2 true discrepant pairs (61/63 = 96.8% agreement). For one discrepant pair, the laboratory INR was 2.2 and the i-STAT INR was 1.6; for the other pair the laboratory INR was 2.3 and the i-STAT INR was 1.7. Using the Shermock method, 79.4% (54/68) of pairs were in the same clinical category. Six of the discrepancies were within 0.4 INR (in fact, all 6 were within 0.1–0.3 INR) and 5 were due to warfarin or vitamin K as described above, leaving 3 discrepant pairs. Two of the pairs are the same as described above; for the third pair, the laboratory INR was 1.9 and the iSTAT INR was 1.4. To look more closely at elevated INR values, we continued to collect paired specimen results for INR> 3 for an additional 4 month time period, as performed by non-laboratory personnel at the point-of-care. Twenty-five paired results were obtained, 60% of which were INR 3–3.9, 32% were 4–4.9, and 8% were ≥5. Agreement among pairs was 92.0% (looking at the two agreement criteria separately, 80.0% were in the same clinical category and 72.0% were within 0.4 INR of each other). Using the Shermock method, 68.0% were in the same clinical category, but 5 of 8 of the discrepant pairs were within 0.4 INR of each other, leaving 3 of 25 discrepant pairs (88.0% agreement). Shermock et al. reported that 2 point-of-care INR devices, Hemochron Junior and Hemochron Signature Elite, appeared incapable of reporting 7 specific INR values, for unclear reasons (namely, INR 2.1, 2.7, 3.1, 3.5, 3.8, 4.1, and 4.4.) [9]. Therefore, we reviewed all INR results generated by the i-STAT in this range. The i-STAT generated every possible INR value from 1.0 to 4.5, and thus does not appear to have the same problem as the Hemochron devices.
958
J.A. Peña et al. / Clinica Chimica Acta 413 (2012) 955–959
1 uncorrected bias corrected bias Median uncorrected bias
Difference
Median corrected bias
0 0
1
2
3
4
5
-1
INR Fig. 1. Bland Altman Plot comparing the i-STAT international normalized ratio (INR) to the MDAII using both uncorrected and corrected INR values. The difference (Y-axis) indicates the i-STAT INR minus the MDAII central laboratory INR. The INR (X-axis) indicates the mean of the i-STAT and the MDAII INR. Shading indicates the typical therapeutic range for warfarin. Uncorrected INR values are INR values generated by the i-STAT (which uses the manufacturer's assigned mean normal PT). Corrected INR values use the mean normal PT that we established locally on the i-STAT device used in this study. Corrected INR values more closely matched the central laboratory INR values than did the uncorrected INR values. The horizontal line of zero bias is shown but is visually indistinguishable from the median corrected bias line at − 0.002.
4. Discussion Obtaining validation data for point-of-care PT/INR tests that use capillary fingerstick whole blood specimens is challenging because the studies require split sample testing from anticoagulated and normal subjects. Validation also requires comparison to the specific coagulation analyzer used in the central laboratory because the correlation data cannot be generalized across instruments from different manufacturers. There is a paucity of published data comparing the i-STAT PT/INR to various different instruments including no prior data using the MDAII. Most studies have reported data using only a small number of subjects such that the cumulative experience with any one instrument is limited. The utility of any POCT device, including those used for determining the PT/INR, depends on the ability of the device to measure the analyte with clinically acceptable accuracy and precision. In this study, we have shown that the i-STAT PT/INR, can be reliably utilized at the point-ofcare by both phlebotomists and by nursing staff in a clinical setting. Specifically the i-STAT PT/INR exhibits only a modestly higher imprecision than the central laboratory MDAII analyzer. All 50 split sample specimens performed by phlebotomists were in the same clinical category or within 0.4 INR as the laboratory result using a “corrected” PT/INR. In the case of tests performed by nurses 61 of 63 (96.8%) values were in the same clinical category or within 0.4 INR. There were 2 discrepant Table 6 Concordance of clinical categories based on INR for patients followed by the AMS. Therapeutic
Suprab
INR from i-STAT uncorrected Suba 7 Therapeutic 3 b Supra 0
0 24 3
0 0 13
INR from i-STAT corrected Suba Therapeutic Suprab
3 23 1
0 0 13
INR from MDAII
Suba
9 1 0
INR: international normalized ratio. AMS: Anticoagulation Management Service. a Sub-therapeutic INR (b2.0). b Supra-therapeutic INR (> 3.0).
cases that could potentially have resulted in a minor adjustment to therapy. Of note not all specimens performed by the nurses were truly simultaneous. A number of these were simply performed within the same day. For this reason the data on nursing-performed testing might show better correlation if simultaneous paired specimens could be obtained. Nevertheless, these results are considerably better than those reported by Karon et al. [2] (76% agreement and 24% discordant: nurses performed the i-STAT POCT) and by Donaldson et al. [4] (46% agreement and 54% discordant: a pharmacist performed the i-STAT POCT), using other central laboratory coagulation analyzers. There is relatively little published data comparing the i-STAT PT/ INR to central laboratory instruments. In the study by Donaldson et al. [4] the i-STAT and CoaguChek XS Plus PT/INR were compared to the Stago coagulation analyzer (model of instrument not specified) on 50 patients. Both POC devices exhibited a positive bias in INR values when compared to the Stago and the bias increased at higher INR values. This necessitated reflexing all POCT values greater than 3.5 to a central laboratory for confirmation. In the subsequent study of the i-STAT by Karon et al. of 50 patients [2] comparing the iSTAT to the MDA-180 analyzer the authors concluded that the overall agreement between the POC and laboratory values was “very good”. They also noted that newer POCT PT/INR devices appear to perform better than older devices. They used an older version of the MDAII called the MDA-180, and a different reagent than the present study (Innovin reagent from Siemens). The results presented in our study are consistent with the findings reported by Karon et al. and expand on the limited available database for the i-STAT PT/INR. Specifically the i-STAT PT/INR performs with acceptable clinical reliability in the hands of phlebotomists and non-laboratory nursing personnel. While the i-STAT PT/INR appears to have a positive bias at higher INR values there is good concordance to the MDAII. Eighty eight percent of i-STAT values were in the same clinical category as the MDAII and 98% were within the same category or within 0.4 INR of each other. This concordance improves when the corrected i-STAT INR is used (90% within the same clinical category or 100% if the criterion of within 0.4 INR is also used). The recurring finding that POCT devices, including the i-STAT system, impart a bias compared to laboratory instrumentation might, in
J.A. Peña et al. / Clinica Chimica Acta 413 (2012) 955–959
part, be due to a calculation issue. For example, with the i-STAT, the INR is calculated from a specimen's PT, and the divisor in the equation (the mean normal i-STAT PT) is set to be 12.0 s as provided by the manufacturer. However, this mean normal PT of 12.0 s is not what we have established within our institution on the i-STAT analyzer. Our mean normal i-STAT PT is 12.78 s. Using this device-specific mean normal, the bias is partially mitigated. Previous studies have shown that POCT devices for anticoagulation monitoring can be of great utility, especially for patients on long-term anticoagulation [5,10]. These devices have also been shown to be of use in the acute setting as well, particularly for determining thrombolytic therapy eligibility in anticoagulated patients who have suffered a cerebrovascular accident [6,10,11]. Although automated laboratory instruments have been shown to have greater precision compared to POCT devices [12], the i-STAT PT/INR device has been shown to have sufficiently acceptable precision and reproducibility for use in clinical decision making for patients receiving oral VKA therapies [5,6]. In the acute setting, the i-STAT PT/INR has also been shown to be useful in patients presenting with acute stroke symptoms who are eligible for thrombolytic therapy [6,11]. The selection of a particular POCT device for implementation in either outpatient or inpatient use requires careful assessment of individual devices and assessment of how compatible the device is with each institutions existing anticoagulation instrument in the central laboratory. Some POCT devices for PT have been shown to have good correlation with laboratory instruments, but at elevated INRs, this concordance is less reliable [13] which may necessitate reflexing specimens with high INR values to the central laboratory. i-STAT imprecision was performed by laboratory personnel, which might be seen as a limitation of our study, since the precision seen might not be generalizable to real clinical practice settings. This potential limitation is at least partially overcome in our study because subsequent INR testing was performed by both phlebotomy and nursing staff. In summary, we show that the i-STAT PT/INR provides rapid and reliable PT/INR testing to aid in clinical decision making. While the INR determined by this POCT device generally leads to a slightly higher INR, the clinical category in which a patient is placed has 88% concordance with the MDAII analyzer and 98% concordance when the criterion of “same clinical category or within 0.4 INR” is used. Correction of i-STAT INR values using a locally determined mean normal i-STAT PT improves the concordance between the two analyzers. Our data expands on the
959
currently available studies in the literature and should assist clinical services who are considering implementation of the i-STAT POCT PT/INR.
Acknowledgments Abbott Point-of-Care supplied some of the i-STAT cartridges used in this study.
References [1] Newall F, Monagle P, Johnston L. Home INR monitoring of oral anticoagulant therapy in children using the CoaguChek S point-of-care monitor and a robust education program. Thromb Res 2006;118:587–93. [2] Karon BS, McBane RD, Chaudhry R, Beyer LK, Santrach PJ. Accuracy of capillary whole blood international normalized ratio on the CoaguChek S, CoaguChek XS, and i-STAT 1 point-of-care analyzers. Am J Clin Pathol 2008;130:88–92. [3] Khoschnewis S, Hannes FM, Tschopp M, Wuillemin WA. INR comparison between the CoaguChek Pro PT(N) and a standard laboratory method. Thromb Res 2004;113:327–32. [4] Donaldson M, Sullivan J, Norbeck A. Comparison of International Normalized Ratios provided by two point-of-care devices and laboratory-based venipuncture in a pharmacist-managed anticoagulation clinic. Am J Health Syst Pharm 2010;67: 1616–22. [5] Boehlen F, Reber G, de Moerloose P. Agreement of a new whole-blood PT/INR test using capillary samples with plasma INR determinations. Thromb Res 2005;115: 131–4. [6] Drescher MJ, Spence A, Rockwell D, Staff I, Smally AJ. Point-of-care testing for coagulation studies in a stroke protocol: a time-saving innovation. Am J Emerg Med 2011;29:82–5. [7] Anderson DR, Harrison L, Hirsh J. Evaluation of a portable prothrombin time monitor for home use by patients who require long-term oral anticoagulant therapy. Arch Intern Med 1993;153:1441–7. [8] van den Besselaar AM. A comparison of INRs determined with a whole blood prothrombin time device and two international reference preparations for thromboplastin. Thromb Haemost 2000;84:410–2. [9] Shermock KM, Streiff MB, Pinto BL, Kraus P, Pronovost PJ. Novel analysis of clinically relevant diagnostic errors in point-of-care devices. J Thromb Haemost 2011;9:1769–75. [10] Green TL, Mansoor A, Newcommon N, Stephenson C, Stewart E, Hill MD. Reliability of point-of-care testing of INR in acute stroke. Can J Neurol Sci 2009;35:348–51. [11] Rizos T, Herweh C, Jenetzky E, et al. Point-of-care international normalized ratio testing accelerates thrombolysis in patients with acute ischemic stroke using oral anticoagulants. Stroke 2009;40:3547–51. [12] van den Besselaar AM. Accuracy, precision, and quality control for point-of-care testing of oral anticoagulation. J Thromb Thrombolysis 2001;12:35–40. [13] Ryan F, O'Shea S, Byrne S. The reliability of point-of-care prothrombin time testing. A comparison of CoaguChek S and XS INR measurements with hospital laboratory monitoring. Int J Lab Hematol 2010;32:e26–33.