Available online at www.sciencedirect.com

www.elsevier.com/locate/yabio

Analytical Biochemistry 375 (2008) 144–146

ANALYTICAL

BIOCHEMISTRY

Notes & Tips

Evaluation of cardiac assays on a benchtop chemiluminescentenzyme immunoassay analyzer, PATHFAST

Takashi Kurihara 1,*, Atsushi Yanagida, Hiroyuki Yokoi, Atsushi Koyata, Takeshi Matsuya,Junichi Ogawa, Yoshikazu Okamura, Daishi Miyamoto

Research and Development Department, Mitsubishi Kagaku Iatron, 1144 Ohwadashinden, Yachiyo, Chiba 276-0046, Japan

Received 31 August 2007Available online 31 December 2007

Abstract

Devices for cardiac immunoassay are widely available at point of care settings. The lateral flow method is the most popular solutionfor ease of use. But its major shortcomings, poor assay precision, and low analytical sensitivity may lead to false negative results. There-fore, confirmatory testing by a routine lab analyzer sometimes is necessary. In the current study, we evaluated the cardiac assays troponinI, MB isoenzyme of creatine kinase, myoglobin, and N-terminal pro brain natriuretic peptide on a chemiluminescent enzyme immuno-assay analyzer, PATHFAST. All of the assays demonstrated correlation coefficients higher than 0.97 against the predicate devices alongwith good total precision (coefficients of variation <10%), and the troponin I assay showed analytical sensitivity in conformance with theguidelines of European Society of Cardiology/American College of Cardiology and National Academy of Clinical Biochemistry. Weconcluded that the system was rapid and easy to use without compromising analytical performance.� 2007 Elsevier Inc. All rights reserved.

During recent years, devices for cardiac immunoassay ofsubstances, such as cardiac troponin I (cTnI)2/cardiac tro-ponin T (cTnT), myoglobin, MB isoenzyme of creatinekinase (CK–MB), and N-terminal pro brain natriureticpeptide (NT–proBNP)/brain natriuretic peptide (BNP),have come to be used widely in point-of-care (POC) set-tings where the test results are needed in a rapid and timelymanner for clinical decision making (e.g., emergency

0003-2697/$ - see front matter � 2007 Elsevier Inc. All rights reserved.

doi:10.1016/j.ab.2007.12.030

* Corresponding author. Fax: +81 3 6722 4081.E-mail address: kurihara.takashi@mb.medience.co.jp (T. Kurihara).

1 Current address: PATHFAST Group, POCT Business Department,Mitsubishi Chemical Medience Corporation, 2-8 Shibaura, 4 Minato,Tokyo 108-8559, Japan.

2 Abbreviations used: cTnI, cardiac troponin I; cTnT, cardiac troponinT; CK–MB, MB isoenzyme of creatine kinase; NT–proBNP, N-terminalpro brain natriuretic peptide; BNP, brain natriuretic peptide; POC, point-of-care; CV, coefficient of variation; CLEIA, chemiluminescent enzymeimmunoassay; B/F, bound/free; QC, control; LOB, limit of blank; LOD,limit of detection; LOQ, limit of quantification; cTnC, cardiac troponin C;sTnI, skeletal muscle troponin I; CK–MM, MM isoenzyme of creatinekinase; CK–BB, BB isoenzyme of creatine kinase; ANP, atrial natriureticpeptide; CNP, C-type natriuretic peptide.

rooms, intensive and coronary care units) [1]. The lateralflow method is the most popular solution for ease of use.But its major shortcomings, poor assay precision and lowanalytical sensitivity [2], may lead to false negative results.Therefore, confirmatory testing by a routine lab analyzersometimes is necessary.

cTnI is the most specific cardiac marker for myocardialinfarction. The European Society of Cardiology/AmericanCollege of Cardiology guideline published in 2000 and theNational Academy of Clinical Biochemistry guideline in2007 stated that acceptable imprecision (i.e., coefficientof variation [CV]) at the 99th percentile of a reference con-trol group for each assay should be equal to or less than10% [3,4]; however, only a few products meet theserequirements [5,6].

Recently, we developed a benchtop batch analyzer,PATHFAST, and its single-use reagent cartridges for cTnI,myoglobin, CK–MB, and NT–proBNP intended for rapid(in as short a time as 17 min) and quantitative measure-ment in POC settings (Fig. 1). The system adopted highlysensitive chemiluminescent enzyme immunoassay (CLEIA)

Fig. 1. The PATHFAST system (upper left panel) and a single-use reagentcartridge (lower panel). Reaction and B/F units are magnified (upper rightpanel). Reagent cartridges in the cartridge rack are set on the heat blockon the stage after the samples were dispensed in sample wells. Disposabletips are placed at the designated positions. The B/F separation unitconsists of magnets and a heat block to capture magnetic particles on theinside walls of tips (Magtration) during B/F separation at a constanttemperature. The start button is assembled separately from the liquidcrystal display touch panel, which does not always need to be operated fora patient sample assay, so that unskilled personnel can operate the systemwithout difficulty.

Notes & Tips / Anal. Biochem. 375 (2008) 144–146 145

using CDP–Star/Sapphire-II (Applied Biosystems) aschemiluminescent substrate and Magtration [7] by whichefficient bound/free (B/F) separation can be performed ina disposable pipette tip with a small volume of washingbuffers. Here we evaluate analytical performances ofPATHFAST cardiac assays.

The PATHFAST analyzer and its single-use cartridgesfor cTnI, myoglobin, CK–MB, and NT–proBNP weremanufactured by ourselves for evaluation. The auto-mated clinical analyzer Stratus CS and its cartridgesfor cTnI, myoglobin, and CK–MB were purchased fromDade Behring, and the Elecsys analyzer and its NT–proBNP reagent were purchased from Roche Diagnos-tics. Controls (QCs) for each analyte were prepared frompooled heparinized human plasma by the addition ofcardiac troponin ITC complex (HyTest), myoglobinnumber 8M50 (HyTest), CK–MB (BioCheck), and syn-thetic NT–proBNP (Roche Diagnostics). The controlswere kept frozen at –80 �C and thawed prior to use.All plasma and whole blood samples were obtained oninformed consent.

The evaluation study consists of intrarun precision, totalprecision, limit of blank (LOB), limit of detection (LOD),limit of quantification (LOQ), linearity, recovery, cross-reactivity with related substances, interferences, prozone,method comparison with predicate devices, correlationbetween plasma and whole blood, and normal referencerange.

Intrarun and total precisions were evaluated in accor-dance with EP5-A2 guidelines [8], and LOB, LOD, andLOQ were evaluated in accordance with EP17-A guidelines[9], published by the National Committee for Clinical Labo-ratory Standards. Linearity was tested using three differentplasma samples with an elevated analyte concentration thatwere serially diluted with sample diluents. Recovery wasevaluated by using a 1:1 mixture of plasma samples with highand low analyte concentrations. Cross-reactivities wereassessed against cTnT, cardiac troponin C (cTnC), skeletalmuscle troponin I (sTnI), ternary cardiac troponin ITC com-plex, free cTnI, dephosphorylated cTnI, and phosphorylatedcTnI for cTnI assay; against MM isoenzyme of creatinekinase (CK–MM) and BB isoenzyme of creatine kinase(CK–BB) for CK–MB assay; and against atrial natriureticpeptide (ANP), BNP, and C-type natriuretic peptide(CNP) for NT–proBNP assay. Interference was estimatedfor conjugated and free bilirubin, hemoglobin, lipemia, tri-glyceride, and rheumatoid factor. Prozone was evaluatedby using serial dilution of samples containing cardiac tropo-nin ITC complex, myoglobin number 8M50, CK–MB, andsynthetic NT-proBNP. Method comparison was performedagainst Stratus CS for cTnI, myoglobin, and CK–MB,whereas NT–proBNP was tested against Elecsys using fro-zen heparinized plasma samples. Correlations betweenplasma and whole blood were evaluated using fresh samples,and a normal reference study was performed using plasmasamples from apparently healthy individuals. Statisticalanalysis for method comparison and correlations betweenplasma and whole blood were performed by Passing–Bablokregression analysis [10,11].

The intrarun precisions were 3.7, 2.8, and 3.2% CV withQCs of 0.09, 2.22, and 9.57 ng/ml, respectively, for cTnI;3.0, 2.8, and 1.4% CV at 23.0, 69.1, and 468 ng/ml, respec-tively, for myoglobin; 5.4, 3.9, and 3.2% CV at 4.57, 61.2,and 359 ng/ml, respectively, for CK–MB; and 4.4, 4.1, and4.7% CV at 239, 2388, and 12,058 pg/ml, respectively, forNT–proBNP. Using the same QCs, the total respective pre-cisions were 3.9, 3.1, and 3.7% CV for cTnI; 5.0, 4.6, and5.4% CV for myoglobin; 8.3, 6.4, and 6.8% CV for CK–MB; and 5.0, 4.6, and 5.4% CV for NT–proBNP. The respec-tive LOB/LOD/LOQ values were 0.001/0.003/0.007 ng/mlfor cTnI, 0.37/0.65/1.30 ng/ml for myoglobin, 0.008/0.100/0.330 ng/ml for CK–MB, and 3.3/5.1/11.8 pg/ml for NT–proBNP. The cTnI assay showed linearity up to 50 ng/mlwith recoveries of 87.0 to 99.6%. Linearities of myoglobin,CK–MB, and NT–proBNP were up to 1000 ng/ml, 500 ng/ml, and 25,000 pg/ml with recoveries of 86.7 to 107.7%,89.4 to 102.7%, and 92 to 110.7%, respectively. Recoverieswere 93.3 to 104.3% for cTnI, 100.7 to 109.0% for myoglo-bin, 93.1 to 96.6% for CK–MB, and 99.4 to 100.5% forNT–proBNP. The cTnI assay was less than 0.1% cross-reac-tive with cTnT, cTnC, and sTnI, and it was nearly equallycross-reactive with ternary cardiac troponin ITC complex(100%), free cTnI (98%), dephosphorylated cTnI (86%),and phosphorylated cTnI (108%). The CK–MB assayshowed 0.08% cross-reactivity with CK–BB and less than

146 Notes & Tips / Anal. Biochem. 375 (2008) 144–146

0.01% cross-reactivity with CK–MM. The NT–proBNPassay showed less than 1% cross-reactivity with BNP,ANP, and CNP. The influences of interfering substanceswere investigated for conjugated bilirubin (up to 60 mg/dl),free bilirubin (up to 60 mg/dl), hemoglobin (up to1000 mg/dl), lipemia (up to 3000 FTU), triglyceride (up to1000 mg/dl), and rheumatoid factor (up to 500 IU). Recov-eries for all of the interfering substances on the assays werenot affected (100 ± 10% of the theoretical values) at the high-est concentration tested with the exception of free bilirubinon the myoglobin assay (unaffected up to 36 mg/ml). Pro-zone was investigated, and samples up to 20,000 ng/ml cTnI,30,000 ng/ml myoglobin, 40,000 ng/ml CK–MB, and300,000 pg/ml NT–proBNP were reported as over the upperlimit of the measurement ranges. A comparison with DadeStratus CS (x) found the following correlations: y = 0.371x

– 0.007, r = 0.975 (n = 80) for cTnI; y = 0.679x + 0.81,r = 0.992 (n = 126) for myoglobin; and y = 1.72x – 0.468,r = 0.997 (n = 87) for CK–MB. NT–proBNP assay wascompared with Roche Elecsys proBNP (x), and the correla-tion was found to be y = 1.01x + 2.6, r = 0.994 (n = 795).The correlations between plasma (x) and whole blood (y)were as follows: y = 1.01x + 0.014, r = 0.999 (n = 91) forcTnI; y = 1.00x – 2.82, r = 0.992 (n = 88) for myoglobin;y = 0.98x + 0.15, r = 0.998 (n = 95) for CK–MB; andy = 1.04x – 0.79, r = 0.994 (n = 25) for NT–proBNP. Thevalue for the 99th percentile for cTnI in 363 apparentlyhealthy individuals was less than 0.02 ng/ml. The calculated95% reference intervals (2.5th–97.5th) were 8.95 to 48.80 ng/ml for myoglobin, 0.34 to 4.99 ng/ml for CK–MB, and 0 to112 pg/ml for NT–proBNP.

PATHFAST cardiac assays for cTnI, myoglobin, CK–MB, and NT–proBNP showed analytical performancescomparable to those for routine lab analyzers. In particu-lar, the cTnI assay exhibited superiority in several aspects.First, LOQ of the cTnI assay (0.007 ng/ml) was lower thanthe 99th percentile of the reference control group (0.02 ng/ml). This fulfills the requirement of the guidelines’ recom-mendation that acceptable imprecision (CV) at the 99thpercentile of a reference group should be less than 10%[3,4]. Second, whole blood sample is applicable becausePATHFAST uses chemiluminescent substrate containinglevamisol to inhibit endogenous alkaline phosphataseactivity in white blood cells, enabling the highly sensitiveCLEIA assay with alkaline phosphatase as a reporter mol-ecule. Third, the cTnI assay was standardized to standardreference material 2921 (National Institute of Standardsand Technology). The low slope against Stratus CS(0.371) probably is due to the difference in standardizationprocedure given that it was confirmed by the fact that thediluted standard reference material 2921 with normalhuman plasma consistently showed good recovery byPATHFAST and not by Stratus CS.

In summary, we have evaluated cardiac assay on abenchtop chemiluminescent immunoanalyzer, PATH-FAST. Several features make the system easy to use,including the use of cartridge-type reagents, no need for

quantitative pipettes for sample aliquots, freedom fromkey operations (if necessary) for sample measurement, min-imal preanalytical tasks (no washing buffer, substrate, ordrainage), and the use of whole blood samples. Yet it israpid, sensitive, and quantitative by adopting CLEIA withMagtration. The analytical performance was verified bydemonstrating that the PATHFAST troponin I assay wasin conformance with the guidelines [3,4]. The assay systemwill be particularly useful even where rapid and sensitiveassays are to be conducted by unskilled personnel.

Acknowledgment

We thank Professor Masafumi Yohda (Tokyo Univer-sity of Agriculture and Technology) for extensive reviewof the manuscript.

References

[1] G.J. Kost, N.K. Tran, Point-of-care testing and cardiac biomarkers:The standard of care and vision for chest pain centers, Cardiology 23(2005) 467–490.

[2] M. Plebani, M. Zaninotto, Cardiac markers: Centralized or decen-tralized testing? Clin. Chem. Lab. Med. 37 (1999) 1113–1117.

[3] Joint European Society of Cardiology/American College of Cardiol-ogy Committee, Myocardial infarction redefined: A consensus doc-ument of the Joint European Society of Cardiology/American Collegeof Cardiology Committee for the redefinition of myocardial infarc-tion, J. Am. Coll. Cardiol. 36 (2000) 959–969.

[4] National Academy of Clinical Biochemistry, National Academy ofClinical Biochemistry laboratory medicine practice guidelines: Clin-ical characteristics and utilization of biochemical markers in acutecoronary syndromes, Clin. Chem. 53 (2007) 552–574.

[5] M. Panteghini, F. Pagani, K.T.J. Yeo, F.S. Apple, R.H. Christenson,F. Dati, J. Mair, J. Ravkilde, A.H.B. Wu, Evaluation of imprecisionfor cardiac troponin assays at low-range concentrations, Clin. Chem.50 (2004) 327–332.

[6] F.S. Apple, M.M. Murakami, Serum and plasma cardiac troponin I99th percentile reference values for 32nd-generation assay, Clin.Chem. 53 (2007) 1558–1560.

[7] K. Obata, O. Segawa, M. Yakabe, Y. Ishida, T. Kuroita, K. Ikeda, B.Kawakami, Y. Kawamura, M. Yohda, T. Matsunaga, H. Tajima,Development of a novel method of operating magnetic particles,Magtration technology, and its use for automating nucleic acidpurification, J. Biosci. Bioeng. 91 (2001) 500–503.

[8] National Committee for Clinical Laboratory Standards, Evaluationof Precision Performance of Quantitative Measurement Methods:Approved Guideline–Second Edition, NCCLS document EP5-A2,2004. (NCCLS, 940 West Valley Road, Suite 1400, Wayne, PA 19087,USA).

[9] National Committee for Clinical Laboratory Standards, Protocols forDetermination of Limits of Detection and Limits of Quantitation:Approved Guideline, NCCLS document EP17-A, 2004. (NCCLS, 940West Valley Road, Suite 1400, Wayne, PA 19087, USA).

[10] H. Passing, W. Bablok, A new biometrical procedure for testing theequality of measurements from two different analytical methods:Application of linear regression procedures for method comparisonstudies in clinical chemistry (part I), J. Clin. Chem. Clin. Biochem. 21(1983) 709–720.

[11] H. Passing, W. Bablok, Comparison of several regression proceduresfor method comparison studies and determination of sample sizes:Application of linear regression procedures for method comparisonstudies in clinical chemistry (part II), J. Clin. Chem. Clin. Biochem. 22(1984) 431–445.