Clin Chem Lab Med 2020; 58(11): e263–e265

Letter to the Editor

Federica Braga* and Mauro Panteghini

Derivation of performance specifications for

uncertainty of serum C-reactive protein
measurement according to the Milan model 3
(state of the art)
https://doi.org/10.1515/cclm-2020-0532 elevation of CRP is not diagnostic of any one specific disease
Received April 17, 2020; accepted June 1, 2020; published online June as it occurs in many diseases involving tissue damage or
29, 2020 inflammation. Besides not having a central and well-defined
role in the decision making of a specific disease, CRP is a
Keywords: analytical performance specifications; C-reactive biologically challenging analyte. We previously highlighted
protein; measurement uncertainty; state of the art. the difficulties in deriving reliable CRP BV data [6, 7]. The
meta-analysed BV data made recently available on the EFLM
database (https://biologicalvariation.eu/) are 33.5% for
To the Editor,
intraindividual (CVI) and 87.7% for inter-individual vari-
The definition and fulfillment of suitable analytical per- ability (CVG), respectively. As Franzini pointed out, the CV
formance specifications (APS) for the measurement un- describing the distribution of a set of positive distributed
certainty (MU) is essential to ensure that laboratory values, as laboratory results are, cannot exceed 33.3% [8].
measurements are clinically usable [1, 2]. There is now a Higher CVI and CVG values probably represent incorrect es-
global consensus that the derivation of APS for different timates due either to a transitory illness of some subjects in
measurands should be carried out by using the models the evaluated sample group or, more generally, to non-
established in 2014 by the Strategic Conference of the Eu- normal distribution of the CRP data in apparently healthy
ropean Federation of Clinical Chemistry and Laboratory individuals.
Medicine (EFLM) held in Milan, Italy [3]. In particular, Considering that neither of the first two EFLM models
model 1, based on the effect of analytical performance on seems to be suitable for CRP, the model three constitutes
clinical outcome, applies for measurands that have a the best option to derive APS for this measurand. Accord-
central role in diagnosis and monitoring of a specific dis- ing to the definition of the “state of the art” by the EFLM
ease; model 2, based on the components of biological Strategic Conference, to derive APS for MU by using this
variation (BV) of the measurand, should be used for model it is necessary to identify the highest quality of
measurands under strict metabolic control; and model 3, performance that is currently achievable [3]. With this aim,
based on the state of the art of the measurement (defined as we compared experimentally estimated standard MU of
the highest level of analytical performance technically four widely used commercial measuring systems for CRP
achievable) should be used for measurands that cannot be measurement available in our laboratory: Alinity c and
included in the first two models [4]. Architect c16000 (Abbott Diagnostics), AU680 (Beckman
C-reactive protein (CRP) is the most sensitive of the acute Coulter) and Cobas c501 (Roche Diagnostics), all ‘Con-
phase proteins and its concentrations in serum increase formité Européenne’/CE marked. Table 1 summarizes the
rapidly during inflammatory processes [5]. However, an characteristics of the evaluated measuring systems.
According to recently published documents [1, 9], MU
*Corresponding author: Federica Braga, ‘Luigi Sacco’ University associated with laboratory results should combine two un-
Hospital, Department of Biomedical and Clinical Sciences, University certainty components: the MU of calibrator value (ucal) and
of Milan Medical School, Laboratorio Patologia Clinica via G.B. Grassi the MU accounting for random sources (uRw). The former
74 Milano, Milan, Italy, E-mail: [email protected].
should in turn combine the MU of the higher-order reference
https://orcid.org/0000-0003-3562-7180
Mauro Panteghini: Research Centre for Metrological Traceability in
selected by the in vitro diagnostics (IVD) manufacturer for
Laboratory Medicine (CIRME), University of Milan, Milan, Italy implementing traceability with the MU deriving from the

Open Access. © 2020 Federica Braga and Mauro Panteghini, published by De Gruyter. This work is licensed under the Creative Commons
Attribution 4.0 International License.
e264 Braga and Panteghini: Performance specifications for uncertainty of serum C-reactive protein

Table : Characteristics of measuring systems for C-reactive protein measurement included in the study.

Manufacturer Platform Commercial name and assay Manufacturer’s Calibrator name Stated Certified value of
principle instruction traceability reference material ±
for use version standard uncertainty

Abbott Alinity c CRP Vario Wide Range Immu- G/R, CRP Vario Wide ERM-DA . ± . mg/L
Diagnostics noturbidimetric method (art. May  Range Calibrator kit /IFCC
P)
Abbott Architect CRP Vario Wide Range Immu- /R, CRP Calibrator Set ERM-DA . ± . mg/L
Diagnostics c noturbidimetric method (art. August  /IFCC
K)
Beckman AU CRP Latex Immunoturbidimetry BLOSRx., CRP Latex Calibrator ERM-DA . ± . mg/L
Coulter (art. OSR) July  Normal Set /IFCC
Roche Cobas c CRPL Particle Enhanced V ., C.f.a.s. Proteins CRM  . ± . mg/L
Diagnostics Immunoturbidimetric Gen.  December 
(art. )

process for assignment of calibrator values. uRw gives in- reproducibility from 6-month consecutive measurement
formation about the stability of the measuring system over data of a serum pool with a selected CRP concentration near
time and its variability when employed by an individual to 10 mg/L, randomly analysed daily during the ordinary
laboratory. The recommended time span for this evaluation laboratory activity. This material had the characteristics
is 6 months, as it allows covering all the significant sources previously recommended for correctly deriving the MU of
of uRw [9]. We previously described in detail how correctly measuring systems due to random effects [1, 10]. In daily
estimating uRw through the internal quality control (IQC) practice, the system alignment to the manufacturer’s speci-
information [10]. In this study, for each measuring system, fications was checked by measuring, three times per day, i. e.
we collected the information about the metrological trace- every 8 h, the control materials offered by each manufacturer
ability and MU of commercial calibrators at CRP concentra- as part of their CE-marked measuring system and by veri-
tion of 10 mg/L, the 99th percentile limit of reference value fying that results were within the range declared by the
distribution, with the exception of Roche C.f.a.s. calibrator manufacturer. Finally, the standard MU associated to clin-
for which only one level is available at 79.9 mg/L concen- ical samples were calculated as √(ucal2 + uRw2) and the sys-
tration. Then, we calculated the ucal by combining the tem with the best analytical performance identified (Table 2).
standard MU of employed reference material, available on Although it is theoretically possible that there is some
the corresponding certificate of analysis, with the standard option on the in vitro diagnostics market performing better,
MU of commercial calibrator as provided by the manufac- our data show that the Architect c16000 performance in
turer. The uRw was experimentally estimated as intermediate terms of MU on clinical samples (3.76%) may represent the

Table : Estimated combined standard measurement uncertainty of evaluated measuring systems for C-reactive protein (CRP) measurement.

Measuring Intermediate reproducibility data Combined standard Combined standard uncertainty

system uncertainty of calibrator on clinical samplesb, %
value (ucal)a, %
Number of measurements Mean CRP, Standard uncertainty
in the -month period mg/L accounting for random
sources (uRw), %

Alinity c  . . . .

Architect  . . . .
c
AU  . . . .
Cobas c  . . . .
a
Calculated by combining the standard MU of the employed reference material with the standard MU of commercial calibrator as declared by
manufacturers. bEstimated as √(ucal + uRw).
 Braga and Panteghini: Performance specifications for uncertainty of serum C-reactive protein e265

state of the art of the CRP measurement to be employed for (Clinical Pathology Unit, ASST Fatebenefratelli-Sacco) for
the definition of APS for MU according to the model three of collection and processing of IQC data.
the EFLM Strategic Conference. If we consider this APS as Research funding: None declared.
desirable, we can also modulate the quality level to mini- Author contributions: All authors have accepted
mum goal (3.76% + ½ 3.76% = 5.64%), as previously responsibility for the entire content of this manuscript
described [10]. From data in Table 2, all the evaluated and approved its submission.
systems fulfilled this minimum quality goal for CRP MU. It Competing interests: Authors state no conflict of interest.
is however expected that involved manufacturers work for Ethical approval: The local Institutional Review Board
improving the quality of assay performance to move to- deemed the study exempt from review.
wards the desirable quality goal in the next future. In this
context, it is noteworthy the worsening of the analytical
performance of CRP measurement between successive References
generations of platforms when Abbott Diagnostics prod-
1. Braga F, Panteghini M. The utility of measurement uncertainty in
ucts are considered. We recommend that manufacturers,
medical laboratories [published online ahead of print, 2020 Mar
before releasing any kind of technological upgrade, should 3]. Clin Chem Lab Med 2020;58:1381–3.
pay major attention to assuring an equivalent, if not better, 2. Braga F, Panteghini M. Defining permissible limits for the
analytical quality of provided measurements. combined uncertainty budget in the implementation of
In this brief communication, we gave an example of metrological traceability. Clin Biochem 2018;57:7–11.
3. Sandberg S, Fraser CG, Horvath AR, Jansen R, Jones G,
the concepts previously mentioned for deriving APS for MU
Oosterhuis W, et al. Defining analytical performance
using the Milan model 3 [10]. The major components of the specifications: consensus Statement from the 1st
process are as follows: 1) assess combined standard MU for Strategic Conference of the European Federation of Clinical
the selected measurand, based on ucal and random effects Chemistry and Laboratory Medicine. Clin Chem Lab Med 2015;
for commonly used measuring systems; 2) for random ef- 53:833–5.
fects assessment, use material(s) other than those used to 4. Ceriotti F, Fernandez-Calle P, Klee GG, Nordin G, Sandberg S,
Streichert T, et al. Criteria for assigning laboratory measurands
control the assay alignment and randomly spaced
to models for analytical performance specifications defined in
throughout the day; 3) identify the combined standard MU the 1st EFLM Strategic Conference. Clin Chem Lab Med 2017;55:
from the best performing system as the ‘desirable’ APS; and 189–94.
4) establish the ‘minimum’ APS as being 50% greater than 5. Mazlam MZ, Hodgson HJ. Why measure C reactive protein? Gut
the desirable one. One potential limitation of our study was 1994;35:5–7.
6. Braga F, Panteghini M. Biologic variability of C-reactive protein:
that only one level for pool material was employed, based
is the available information reliable? Clin Chim Acta 2012;413:
on the CRP threshold having the major clinical impact (i. e., 1179–83.
the upper reference limit). The need of more than one level 7. Braga F, Ferraro S, Szöke D, Lanzoni M, Panteghini M. Estimate of
that spans the reportable range for the test for the optimal intraindividual variability of C-reactive protein: a challenging
estimate of MU has been previously underlined [1, 2]. issue. Clin Chim Acta 2013;419:85–6.
Another study limitation is the use of only one version of 8. Franzini C. Need for correct estimates of biological variation:
the example of C-reactive protein. Clin Chem Lab Med 1998;36:
each method, including the determined best system which
131–2.
is used to set the APS. This could not allow optimal capture 9. ISO/TS 20914:2019. Medical laboratories – Practical guidance for
of the variation in performance that would be seen in the estimation of measurement uncertainty, 1st ed. Geneva,
various laboratories. Switzerland: ISO, 2019.
10. Braga F, Pasqualetti S, Aloisio E, Panteghini M. The internal
quality control in the traceability era. Clin Chem Lab Med 2020,
Acknowledgments: We are indebted to Simona Borille, Article Number: 20200371. Published online: 28 Apr 2020.
Assunta Carnevale, Concetta Sasso and Sara Zampetti https://doi.org/10.1515/cclm-2020-0371.