Evaluation of analytical performance of the Pathfast cardiac troponin I.
ABSTRACT Cardiac troponins are considered the cornerstone for risk stratification and diagnosis of patients with acute coronary syndrome (ACS). Following Clinical Laboratory Standards Institute (CLSI) guidelines, we assessed the analytical performances of the Pathfast (Mitsubishi, Japan) cTnI method.
We evaluated different sample types. Control materials and lithium heparin plasma pools were used to determine: limit of blank (LoB), limit of detection (LoD), imprecision and linearity. The effects of potential endogenous interfering substances and the possibility of falsely increased cardiac troponin I (cTnI) concentrations attributable to the presence of heterophilic antibodies (HA), rheumatoid factor (RF) and human anti-mouse antibodies (HAMA) in high concentrations were evaluated. The 99th percentile limit of the cTnI value distribution was determined from 320 Caucasian reference individuals.
No significant differences were found when cTnI concentrations of 40 lithium-heparin plasma samples were compared with the matched values of K(2)-EDTA plasma, whole blood and serum samples. The LoB and the LoD of the cTnI method were 0.0048 and 0.0066 microg/L, respectively. cTnI mean values from 0.66 to 6.0 microg/L showed a total CV% from 6.0 to 6.4. cTnI at a concentration of 0.02 microg/L was associated with a total CV of 9.6%. The method gave a linear response for cTnI concentrations within the measurement range. In six of 12 samples containing HA, a positive interference was demonstrated. The 99th percentile limit of the cTnI distribution in the reference population was 0.013 microg/L.
The data indicate that the cTnI Pathfast method may be suitable for helping clinicians in the management of patients with ACS.
- Clinical biochemistry 08/2013; 46(12):961-2. · 2.02 Impact Factor
- European Heart Journal 06/2012; 33(18):2252-2257. · 14.10 Impact Factor
Clin Chem Lab Med 2009;47(7):829–833 ? 2009 by Walter de Gruyter • Berlin • New York. DOI 10.1515/CCLM.2009.182
Article in press - uncorrected proof
Evaluation of analytical performance of the Pathfast?
cardiac troponin I
Francesca Di Serio1,*, Marco Caputo2, Martina
Zaninotto3, Cosimo Ottomano4and Mario
1Department of Clinical Pathology I, University
Hospital, Bari, Italy
2Clinical Chemistry Laboratory, Bussolengo
Hospital, Verona, Italy
3Department of Laboratory Medicine, University
Hospital, Padua, Italy
4Clinical Laboratory, Bergamo Hospital, Bergamo,
Background: Cardiac troponins are considered the
cornestore for risk stratification and diagnosis of
patients whit acute coronary syndrome (ACS). Follow-
ing Clinical Laboratory Standards Institute (CLSI)
guidelines, we assessed the analytical performances
of the Pathfast?(Mitsubishi, Japan) cTnI method.
Methods: We evaluated different sample types. Con-
trol materials and lithium heparin plasma pools were
used to determine: limit of blank (LoB), limit of detec-
tion (LoD), imprecision and linearity. The effects of
potential endogenous interfering substances and the
possibility of falsely increased cardiac troponin I
(cTnI) concentrations attributable to the presence of
heterophilic antibodies (HA), rheumatoid factor (RF)
and human anti-mouse antibodies (HAMA) in high
concentrations were evaluated. The 99th percentile
limit of the cTnI value distribution was determined
from 320 Caucasian reference individuals.
Results: No significant differences were found when
cTnI concentrations of 40 lithium-heparin plasma
samples were compared with the matched values of
K2-EDTA plasma, whole blood and serum samples.
The LoB and the LoD of the cTnI method were 0.0048
and 0.0066 mg/L, respectively. cTnI mean values from
0.66 to 6.0 mg/L showed a total CV% from 6.0 to 6.4.
cTnI at a concentration of 0.02 mg/L was associated
with a total CV of 9.6%. The method gave a linear
response for cTnI concentrations within the measure-
ment range. In six of 12 samples containing HA, a
positive interference was demonstrated. The 99th per-
centile limit of the cTnI distribution in the reference
population was 0.013 mg/L.
Conclusions: The data indicate that the cTnI Pathfast
method may be suitable for helping clinicians in the
management of patients with ACS.
Clin Chem Lab Med 2009;47:829–33.
*Corresponding author: Francesca Di Serio, Department of
Clinical Pathology I, University-Hospital of Bari, Piazza
Giulio Cesare N. 11, 70124 Bari, Italy
Phone: q39-080-5592629/124, Fax: q39-080-5592124,
Received December 16, 2008; accepted March 31, 2009;
previously published online May 20, 2009
Keywords: acute myocardial infarction (AMI); analyt-
ical performance; troponin.
Coronary heart disease is the leading cause of mor-
tality in developed countries (1). Timely evaluation of
patients with chest pain in the emergency/cardiology
department is mandatory in order to provide early
treatment to reduce morbidity and mortality in those
with acute coronary syndrome (ACS). In patients with
no ST elevation recorded by ECG, the diagnosis of
acute myocardial infarction (AMI) depends almost
entirely on serial measurements of biochemical mark-
ers of necrosis. In 2000, the European Society of
Cardiology (ESC) and the American College of Cardi-
ologists (ACC) recognized the pivotal role of biomar-
kers, making their increase in the bloodstream the
‘‘cornerstone’’ for diagnosis of AMI (2). In particular,
cardiac troponins have emerged as a powerful tool for
diagnosis of AMI and for risk assessment of cardiac
events, including death and recurrent ischemia (3). In
2007, the joint ESC/ACCF/American Heart Association
(AHA)/World Heart Federation (WHF) Task Force for
the redefinition of myocardial infarction (4) confirmed
the role of troponin. Thus, in a laboratory medicine
context, a high-quality analytical method is manda-
tory (5, 6). In accordance with Clinical Laboratory
Standards Institute (CLSI) guidelines, we evaluated
the analytical performance of cardiac troponin I (cTnI)
on the Pathfast immunoassay analyzer (Mitsubishi
Kagaku Iatron Inc., Tokyo, Japan). We also defined
the 99th percentile limit of the cTnI value distribution
in our reference population.
Materials and methods
The Pathfast system is a bench-top enzyme immunoassay
analyzer, processing up to six samples at a time and provid-
ing simultaneous quantitative determination of creatine
kinase-MB mass, myoglobin, cTnI, and NT-proBNP using
whole blood or plasma (sodium or lithium-heparin) samples
The chemiluminescence enzyme immunoassay (CLEIA)
principle has been enhanced by employing an efficient
bound/free separation method whereby magnetic particles
are washed and separated at the inner wall of a pipette tip
according to a patented Magtration?technology (Precision
System Science/PSS USA Inc., Livermore, CA, USA). The
term ‘‘Magtration’’ is derived by abbreviating ‘‘Magnetic Fil-
tration’’: in order to remove excessive reagents or residual
materials, the technology performs bound/free separation in
pipette tips using magnetic particles. Briefly, the sample
(25 mL) containing the target analyte is incubated with
830 Di Serio et al.: Analytical performance of the PATHFAST?cTnI method
Article in press - uncorrected proof
magnetic particles coated with the corresponding capture
antibodies and alkaline phosphatase (ALP) conjugated anti-
bodies. The resulting immunocomplexes are separated,
washed and transferred by Magtration technology to a lumi-
nescence measurement well containing a CDP-Star?reagent
(1,2-dioxetane compound Phototope?ALP) chemilumines-
cent substrate. The generated luminescence is measured
and the analyte concentration calculated using a standard
curve relating these two parameters. The cTnI method is a
one-step, two site immunoassay using a mouse monoclonal
antibody (MoAb A) immobilized on magnetic particles (cap-
ture: G00811) recognizing cTnI (residues 41–49), and two
monoclonal antibodies (MoAb B and C) bound to ALP (tag:
G00821 and G00831) recognizing the cTnI mid (residues
71–116) and C-terminal portion (residues 163–210), respec-
tively. The human cardiac troponin complex (CTI), supplied
by Hytest Ltd. (Turku, Finland), was used as calibrationmate-
rial in the Pathfast assay. Upon receiving a new lot of re-
agents, the specifications of the lot are recorded in the
instrument through the master lot entry sheet, included in
each package; the calibration adjustment is made using two
calibrators supplied with the kit. The working range of the
master curve is 0.02–50 mg/L and calibration stability lasts
28 days. All reagents needed for the assay are incorporated
into a single cartridge. Analysis time is 15 min.
Lithium-heparin plasma samples (polyethylene terephtha-
late/PET tubes, 68 IU lithium-heparin in 4 mL blood; code:
367374, Becton Dickinson, Franklin Lakes, NJ, USA) were
used as reference for the Pathfast cTnI analytical evaluation
throughout the entire study. In order to evaluate the poten-
tial impact of anticoagulants on cTnI concentrations, EDTA
plasma, whole blood (K2-EDTA: 1 mg/mL blood, in plastic
tubes; code: 368861, Becton Dickinson) and serum (PET
tubes without a gel barrier; code: 368815; Becton Dickinson)
samples were used. Forty paired samples with detectable
cTnI concentrations were obtained from AMI patients admit-
ted to the Cardiology Department of the University Hospital
of Bari. cTnI was measured in all sample types within 1 h
following collection. cTnI results in whole blood were
obtained after software correction for individual hematocrit
values, according to the following formula: corrected
Estimation of limit of blank (LoB) and limit
of detection (LoD), imprecision studies, linearity
The LoB and the LoD were calculated according to CLSI EP
17-A (8). The LoB, which is defined as the highest measure-
ment result that is likely to be observed (with a stated prob-
abilitys5%) for a blank sample, was estimated from 60
repeated measurements of the cTnI Pathfast zero calibrator;
the analytical signal was expressed in arbitrary units. A sam-
ple with a cTnI concentration of 25.3 mg/L was diluted until
an expected concentration equal to zero was obtained. The
parameters of the linear regression (arbitrary units on the
y-axis and expected cTnI concentrations on the x-axis) were
used to transform the mean analytical signal of LoB into cTnI
concentrations. To determine the LoD, which is defined as
the lowest amount of analyte in a sample that can be detect-
ed with a stated probability (usually, 95%), a pooled estimate
of standard deviations (SDs) was obtained from 12 repeated
measurements of five pools with low cTnI concentrations
(cTnI concentrations from LoB to ;3=LoB). The imprecision
of the cTnI Pathfast method was determined in accordance
with CLSI EP5-A guidelines (9): a lithium-heparin plasma
pool and BioRad Liquicheck?Cardiac Markers control LT
(level 2, 3) (BioRad Laboratories, Headquarters,Hercules,CA,
USA) were analyzed daily, in duplicate, for 20 consecutive
days. Two aliquots of test material, for each concentration,
were tested within each run (80 measurement per sample).
To define the method imprecision in the low concentration
range, a plasma pool containing a known troponin concen-
tration (0.11 mg/L) was serially diluted with a lithium-heparin
plasma sample showing undetectable concentrations ofcTnI.
Eleven pools containing troponin concentrations ranging from
0.1 to 0.01 mg/L were aliquoted and stored frozen at –208C
until analysis. For each of the 11 pools, a new aliquot was
thawed, centrifuged and analysed in duplicate every day
for 10 consecutive days; two different reagent lots and two
different calibration curves were used. The linearity of
the method was determined according to CLSI guidelines
EP6-P (10): a lithium-heparin plasma pool (pool A; TnI
concentrations42 mg/L) was diluted with a lithium-heparin
plasma pool showing undetectable concentrations of cTnI
(pool E), at a ratio of 3/1, 2/2 and 1/3. Using this protocol,
three different samples (pool B, C and D) with theoretical
cTnI concentrations of 31.5, 21.0, 10.5 mg/L, respectively,
were obtained. Each of the five pools were tested in quad-
ruplicate in a single run.
Effect of potential interfering substances
Lithium-heparin plasma pools (P) containing high concentra-
tions of potentially interfering substance (total bilirubin,
771 mmol/L; triglycerides, 8.8 mmol/L; hemoglobin, 8.2 g/L)
and TnI concentrations lower than the detection limit were
prepared. These were added in fixed ratios (0q10, 1q9,
2q8, 3q7, 4q6, 5q5, 6q4, 7q3, 8q2, 9q1, 10q0) to two
lithium-heparin plasma pools containing elevated TnI con-
centrations (S1: 1.13 mg/L; S2: 2.72 mg/L) and physiological
concentrations of each of the potential interfering sub-
stances. In addition, diluting both the S1 and S2 pools (in
the same fixed ratios mentioned above) with a lithium-hep-
arin plasma pool with undetectable concentrations of TnI
allowed us to produce a set of control dilutions (control-1,
C1; control-2, C2).
To evaluate the potential interference due to heterophilic
antibodies (HA), human anti-mouse antibodies (HAMA) and
rheumatoid factor (RF), 12 lithium heparin plasma samples
known to cause false positive cTnI results when tested with
the Dimension RxL (Siemens Healthcare Diagnostics, Deer-
field, IL, USA), and seven commercially available (Scantibo-
dies Inc, Santee, CA, USA) samples containing high HAMA
(concentration range: 100–782 ng/mL) and RF titers (concen-
tration range: 162–1080 UI/mL) were tested using the Path-
fast method. All these samples were treated with heterophile
blocking tubes (Scantibodies Inc).
The Pathfast cTnI method was compared to the Stratus CS
cTnI method (Siemens Healthcare Diagnostics) in accor-
dance with CLSI guidelines (11). A total of 115 paired lithium-
concentrations measured with the Stratus CS and Pathfast.
In the comparison study, only samples with cTnI concentra-
tions )0.015 mg/L (LoD for Stratus CS) and lower than
50 mg/L (linearity limit for Pathfast) were used.
Di Serio et al.: Analytical performance of the PATHFAST?cTnI method831
Article in press - uncorrected proof
Upper reference limit
In order to define the cTnI upper reference limit, lithium-hep-
arin plasma samples were collected from 320 healthy Cau-
casian subjects: (160 women and 160 men aged 26–89 years;
mean"SDs56.7"15.6 years). Exclusion criteria included
subjects with a suspected history of ACS, biochemical
abnormalities in serum creatinine, glucose, urea, alanine
aminotransferase (ALT), aspartata aminotransferase (AST),
C-reactive protein, creatine kinase-MB, leukocyte blood
count, or hemoglobin and hyperpyrexia. According to the
ESC/ACCF/AHA/WHF (2, 4), the upper reference limit was cal-
culated at the 99th percentile limit of the cTnI value distri-
bution in the reference group; the 95% confidence interval
(CI) for the percentile was determined assuming a non-
Gaussian distribution of the data.
Statistical comparison among different sample types was
assessed with the Wilcoxon rank sum test for paired sam-
ples; Deming regression and Bland-Altman analysis (12)
were also performed.
For 60 measured values of the sample blank, a Gaussian
p-0.01) and the 95th percentile of the LoB was calculated
as w60(95/100)q0.5x ordered observations following logarith-
mic transformation of the original data set. The LoD was
defined by the relation LoDsLoBqcb=SDs. SDs is the
pooled estimate of the SD of pools with low cTnI concentra-
tions, and cbsz1–b/(1–1/(4=f)) where the z1–bis the standard
normal deviation (SDs1.645), and f is the number of degrees
of freedom for estimation of SDs (13). ANOVA was used to
assess the total imprecision (CV%) for the cTnI method at
different concentrations of analyte. Dilution linearity was
determined by least-squares linear regression analysis with
the lack-of-fit-test (10). Bland-Altman analysis was used to
evaluate the effect of potentially interfering endogenoussub-
stances. We considered the significance of the bias on the
basis of the 95% CI. Taking into account the analytical impre-
cision of the method, a difference )10% among results
obtained in pools with and without added substances was
considered to be clinically significant. Correlation between
Pathfast and Stratus CS cTnI methods were assessed using
Passing and Bablok regression (14) and Bland-Altman
Statistical analysis was performed using Analyse-it Soft-
ware for Microsoft Excel version 2.11 (Analyse-it Software,
Leeds, UK) and MedCalc version 9.5 (Mariakerke, Belgium).
Results and discussion
cTnI concentrations in EDTA plasma (range: 0.004–
29.8 mg/L; median, 0.56 mg/L), whole blood (range:
0.006–28.3 mg/L; median, 0.58 mg/L), and serum
(range: 0.003–28.8 mg/L; median, 0.48 mg/L) were
compared against the cTnI concentrations measured
in lithium-heparin plasma samples (range: 0.004–
27.4 mg/L; median, 0.52 mg/L).
No significant differences were found between cTnI
concentrations measured in 40 lithium-heparin plas-
ma samples and concentrations measured in matched
K2-EDTA plasma samples (2-tailed ps0.76), whole
blood (2-tailed ps0.56) and serum samples (2-tailed
Regression analysis (Deming) showed the following
K2-EDTA plasma vs. Li-heparin: slopes0.96 ("0.025);
intercepts–0.10 ("0.19) mg/L;
K2-EDTA whole blood vs. Li-heparin: slopes0.96
("0.025); intercepts–0.09 ("0.19) mg/L;
Serum vs. Li-heparin: slopes1.04 ("0.00); inter-
cepts–0.07 ("0.07) mg/L.
No significant bias was detected by Bland and Altman
K2-EDTA plasma vs. Li-heparin: average absolute bias
s–0.2 mg/L (95% CI, from –0.6 to 0.09 mg/L); average
percentage biass0.1% (95% CI, from –3.8% to 4%).
K2-EDTA whole blood vs. Li-heparin: average absolute
biass–0.2 mg/L (95% CI, from –0.6 to 0.07 mg/L); aver-
age percentage biass6% (95% CI, from –0.3% to
Serum vs. Li-heparin: average absolute biass0.09
mg/L (95% CI, from –0.05 to 0.2 mg/L); average per-
centage biass0.3% (95% CI, from –2.4% to 3.2%).
The analytical signal corresponding to 60 measure-
ments of Pathfast zero calibrator ranged from 257 to
590 arbitrary units, and the 95th percentile of value
distribution was 344 arbitrary units. Following loga-
formed using the sample with a cTnI concentration of
25.3 mg/L (ys0.9093xq4.6476; r2s0.94) revealed the
Pathfast LoB to be 0.0048 mg/L. To evaluate the LoD,
12 measurements for each of the five pools (cTnI con-
centration from 0.0061 to 0.0156 mg/L) were per-
formed; thisprovided 55
(fs5=(12-1)). The pooled within-sample SD of the
measurements, calculated as the weighted average of
the variances (squared SDs), was 0.0011 mg/L. The
LoD was calculated as LoDsLoBqz1–b/(1–1/(4=f))=
Imprecision studies gave the following: control mate-
rial level 2: mean, 0.66 mg/L; within run CV, 5.8%; total
CV, 6.3%; control material level 3: mean, 3.1 mg/L;
within-run CV, 5.1%; total CV, 6.0%; lithium-heparin
plasma pool: mean, 6.0 mg/L; within-run CV, 4.5%;
total CV, 6.4%. Figure 1A shows the imprecision pro-
file at low cTnI concentrations. A cTnI concentration
of 0.02 mg/L was associated with a total imprecision
-10% (CVs9.6%). Linearity data showed that the
linearity hypothesis was accepted (Fs2.33; ps0.11),
and a linear response obtained (r2s0.995). No signif-
icant interferences were observed with icteric (biliru-
bin up to 704 mmol/L) or hemolyzed (hemoglobin up
to 7.2 g/L) samples. Triglyceride concentrations at
7.6 mmol/L appeared to interfere with Pathfast cTnI
measurements in one of the two samples tested wS2,
0.04–0.15)x. However, given that this difference is well
within the imprecision of the assay, it could be con-
sidered not clinically relevant.
Six of 12 samples showing elevated cTnI concen-
trations due to HA when tested on Dimension RxL
0.09 mg/L: (95%CI,
832Di Serio et al.: Analytical performance of the PATHFAST?cTnI method
Article in press - uncorrected proof
Pathfast vs. Stratus CS method, ns115 (B).
(A) Plot of cTnI values (x-axis) vs. total imprecision as coefficient of variation (CV%; y axis): cTnI concentration (0.02 mg/L)
associated with a 9.6% CV (thick lines), 99th percentile reference limit of cTnI (0.013 mg/L) associated with a 28.7% CV (dashed
lines) and imprecision 95% confidence intervals (error bars) at each cTnI concentration level, are shown. (B) Results are
absolute (left panel), percentage bias (right panel) and 1.96 SD.
Imprecision profile of Pathfast cTnI assay at low concentrations (A) and Bland-Altman difference plots for the
system were obtained from the same patients at dif-
ferent times. The cTnI values measured in these
patients are consistently higher than the decision
level for AMI (0.15 mg/L). However, there was no typ-
ical rise and fall in cTnI values and clinical findings
were not consistent with AMI. Treatment with hete-
rophilic blocking tubes provided definitive evidence
that all values were false positive; after treatment all
values were -0.15 mg/L. When tested on the Pathfast
system, these samples showed cTnI concentrations
ranging from 0.022 to 0.032 mg/L. These values were
higher than the 99th percentile upper reference limit
obtained in this study, suggesting interference by HA,
as also shown by pre-treatment. cTnI concentrations
after treatment with heterophilic blocking tubes were
lower than detection limit of the assay. The other
samples that showed spuriously high concentrations
of troponin when measured with the RxL Dimension,
including those with high RF and HAMA concentra-
tions, gave troponin I results below the detection limit
of the method when tested on Pathfast.
Comparison (ns115) between Pathfast (y) and Stra-
tus CS (x) cTnI gave the following correlation:
ys0.336(SE, 0.009)x–0.005 (SE, 0.12) mg/L; rs0.98
(ns115). Figure 1B shows the bias plots. The mean
absolute bias was y2.7 mg/L (95% CI, –3.8 to –1.4)
and mean percentage bias was –112% (95% CI, –118.9
to –104.9). The intercept was not significant and
methods showed a constant bias. The 99th (95% CI)
percentile limit of the cTnI values distribution in the
reference population was 0.013 (0.0068–0.019) mg/L:
measurable cTnI concentrations (i.e., )0.0066 mg/L,
assay LoD) were found in only six subjects (2%).
Figure 1A shows the imprecision (28.7% CV) obtained
with a sample with a cTnI value of 0.013 mg/L. Several
assays with relatively high precision for cTnI meas-
urement exist. However, none are able to experimen-
tally achieve a 10% CV at the 99th percentile limit of
the reference population (15); this is true also for the
cTnI Pathfast method.
In conclusion, this study indicates that the cTnI
Pathfast method meets the quality specifications rec-
ommended by NACB and the IFCC Committee for the
Standardization of Cardiac Damage (5, 6). Therefore,
it may be suitable for use in the clinical laboratory to
help clinicians in the management of ACS patients.
Caution should be taken when comparing the pub-
lished findings related to analytical sensitivity of the
Di Serio et al.: Analytical performance of the PATHFAST?cTnI method 833
Article in press - uncorrected proof
cardiac troponin assays, with the findings of this
study. Many LoD values are experimentally calculated
as LoB (e.g., as the concentration corresponding to a
signal of 3 SD above the mean of N replicates for a
sample in which cTnI is absent), so that they can be
markedly lower if more statistically rigorous proce-
dures, like those applied in this study, were used.
1. American Heart Association. Heart and stroke statistical
update. Dallas, TX: American Heart Association; 2001.
2. Alpert JS, Thygesen K, Antman E, Bassand JP. Myocardial
infarction redefined – a consensus document of the Joint
European Society of Cardiology/American College of Car-
diology Committee for the redefinition of myocardial
infarction. J Am Coll Cardiol 2000;36:959–69.
3. Braunwald E, Antman EM, Beasley JW, Califf RM, Cheitlin
MD, Hochman JS, et al. ACC/AHA 2002 guideline update
for the management of patients with unstable angina and
non-ST-segment elevation myocardial infarction: execu-
tive summary and recommendations. A report of the
American College of Cardiology/American Heart Associ-
ation Task Force on Practice Guidelines (Committee on
the Management of Patients with Unstable Angina). Cir-
4. Thygesen K, Alpert JS, White HD, Joint ESC/ACCF/AHA/
WHF Task Force for the Redefinition of Myocardial Infarc-
tion. Universal definition of myocardial infarction. Eur
Heart J 2007;28:2525–38.
5. Panteghini M, Gerhardt W, Apple FS, Dati F, Ravkilde J,
Wu AH. Quality specifications for cardiac troponin assays.
International Federation of Clinical Chemistry and Labor-
atory Medicine (IFCC). IFCC Scientific Division Committee
on Standardization of Markers of Cardiac Damage. Clin
Chem Lab Med 2001;39:174–8.
6. Apple FS, Jesse RL, Newby LK, Wu AH, Christenson RH.
National Academy of Clinical Biochemistry; IFCC Com-
mittee for Standardization of Markers of Cardiac Dam-
age. National Academy of Clinical Biochemistry and
IFCC Committee for Standardization of Markers of Car-
diac Damage Laboratory Medicine Practice Guidelines:
analytical issues for biochemical markers of acute cor-
onary syndromes. Circulation 2007;115:e352–5.
7. Kurihara T, Yanagida A, Yokoi H, Koyata A, Matsuya T,
Ogawa J, et al. Evaluation of cardiac assays on a bench-
top chemiluminescent enzyme immunoassay analyzer,
PATHFAST. Anal Biochem 2008;375:144–6.
8. National Committee for Clinical Laboratory Standards.
Protocols for determination of limits of detection and
limits of quantitation; approved guideline. NCCLS guide-
line EP17-A. Wayne, PA: NCCLS, October 2004.
9. National Committee for Clinical Laboratory Standards.
Evaluation of precision performance of clinicalchemistry
devices; approved guideline. NCCLS guideline EP5-A.
Wayne, PA: NCCLS; February 1999.
10. National Committee for Clinical Laboratory Standards.
Evaluation of the linearity of quantitative measurement
procedures: a statistical approach; proposed guideline.
NCCLS guideline EP6-A. Wayne, PA: NCCLS; September
11. National Committee for Clinical Laboratory Standards.
Method comparison and bias estimation using patient
samples; approved guideline. NCCLS guideline EP9-A.
Wayne, PA: NCCLS; December 1995.
12. Bland JM, Altman DG. Statistical methods for assessing
agreement between two methods of clinical measure-
ment. Lancet 1986;307–10.
13. Linnet K, Kondratovich
approach for determining the limit of detection. Clin
14. Passing H, Bablok W. A new biometrical procedure for
testing the equality of measurements from two different
analytical methods. J Clin Chem Biochem 1983;21:
15. Panteghini M, Pagani F, Yeo KT, Apple FS, Christenson
RH, Dati F, et al. Evaluation of the imprecision at low
range concentrations of the assays for cardiac troponin
determination. Clin Chem 2004;50:327–32.