Different top-down approaches to estimate

Clinical Biochemistry 57 (2018) 56–61

Contents lists available at ScienceDirect

Clinical Biochemistry journal homepage: www.elsevier.com/locate/clinbiochem

Different top-down approaches to estimate measurement uncertainty of whole blood tacrolimus mass concentration values

T

⁎

Raül Rigo-Bonnina, , Aurora Blanco-Fonta, Francesca Canaliasb a b

Laboratori Clínic, IDIBELL, Hospital Universitari de Bellvitge, L'Hospitalet de Llobregat, Barcelona, Spain Laboratori de Referència d'Enzimologia Clínica, Departament de Bioquímica i Biologia Molecular, Universitat Autònoma de Barcelona, Bellaterra, Spain

A R T I C LE I N FO

A B S T R A C T

Keywords: Single laboratory validation Proficiency testing Uncertainty Tacrolimus Top-down UHPLC-MS/MS

Background: Values of mass concentration of tacrolimus in whole blood are commonly used by the clinicians for monitoring the status of a transplant patient and for checking whether the administered dose of tacrolimus is effective. So, clinical laboratories must provide results as accurately as possible. Measurement uncertainty can allow ensuring reliability of these results. The aim of this study was to estimate measurement uncertainty of whole blood mass concentration tacrolimus values obtained by UHPLC-MS/MS using two top-down approaches: the single laboratory validation approach and the proficiency testing approach. Material and methods: For the single laboratory validation approach, we estimated the uncertainties associated to the intermediate imprecision (using long-term internal quality control data) and the bias (utilizing a certified reference material). Next, we combined them together with the uncertainties related to the calibrators-assigned values to obtain a combined uncertainty for, finally, to calculate the expanded uncertainty. For the proficiency testing approach, the uncertainty was estimated in a similar way that the single laboratory validation approach but considering data from internal and external quality control schemes to estimate the uncertainty related to the bias. Results: The estimated expanded uncertainty for single laboratory validation, proficiency testing using internal and external quality control schemes were 11.8%, 13.2%, and 13.0%, respectively. Conclusions: After performing the two top-down approaches, we observed that their uncertainty results were quite similar. This fact would confirm that either two approaches could be used to estimate the measurement uncertainty of whole blood mass concentration tacrolimus values in clinical laboratories.

1. Introduction Tacrolimus (TAC) is an immunosuppressive drug that is widely administered to recipients of solid organ transplants [1–3]. Therapeutic drug monitoring of TAC is an useful tool for minimizing drug toxicity while maximizing prevention of graft loss and organ rejection [1–3]. Thereby, accurate results of mass concentration of tacrolimus in whole blood (cTAC) should be provided by clinical laboratories to ensure these measurement results are fit for their clinical purpose and do not compromise patient care. The measurement uncertainty can provide a quantitative indication of the quality and accuracy of these results. Measurement uncertainty is a non-negative parameter characterizing the dispersion of the quantity values being attributed to the measurand based on the information used [4]. In other words, the uncertainty is numerical

information that complements a measured value, indicating the magnitude of the doubt about this value and providing a quantitative indication of the quality of a measured value. Also, clinical laboratories looking forward the accreditation under the ISO 15189 standard, shall determine measurement uncertainty for each measurement procedure, and define and regularly check their performance requirements concerning uncertainty [5]. According to the EUROLAB guideline [6], there are different approaches for the estimation of uncertainty: the modelling approach (so called bottom-up approach by the GUM, EURACHEM and CLSI guidelines [7–9]), the single laboratory validation approach, the interlaboratory approach and the proficiency testing approach. The last three approaches refer to the so-called topdown approach. Briefly, the modelling approach is based on statistical and metrological procedures where all conceivable sources of uncertainty are

Abbreviations: TAC, tacrolimus; cTAC, mass concentration of TAC in whole blood; IQCS, internal quality control scheme; EQCS, external quality control scheme; UHPLC-MS/MS, ultrahigh-performance liquid chromatography-tandem mass spectrometry; CF, correction factor; δ, bias; IPTS, Immunosuppresant Proficiency Testing Scheme; HPLC-MS/MS, high-performance liquid chromatography-tandem mass spectrometry; CV, coefficient of variation ⁎ Corresponding author at: Laboratori Clínic, IDIBELL, Hospital Universitari de Bellvitge, Feixa Llarga s/n, 08907, L'Hospitalet de Llobregat, Barcelona, Spain. E-mail address: [email protected] (R. Rigo-Bonnin). https://doi.org/10.1016/j.clinbiochem.2018.05.005 Received 20 February 2018; Received in revised form 19 April 2018; Accepted 7 May 2018 Available online 08 May 2018 0009-9120/ © 2018 The Canadian Society of Clinical Chemists. Published by Elsevier Inc. All rights reserved.


R. Rigo-Bonnin et al.

For the specification of the measurand, we described the pharmacological quantity according to the IUPAC and IFCC recommendations [12]. Estimation of the precision under intermediate conditions includes the same measurement system, different operators, solutions and samples preparation, lot-to-lot reagents (except for calibrators materials), different calibrations, possible variabilities on the measurement system parameters (mainly cone voltage, collision energy, capillary voltage, collision gas flow, source temperature, desolvation temperature, desolvation flow for the tandem-mass spectrometer; and injection volume, mobile phase flow, column chamber and autosampler temperature for the liquid-chromatograph), and data processing (TAC and internal standard —ascomycin— peak areas integration, peak fitting and peak areas variability, as well as the retention times peaks variability). To estimate it, the Liquicheck™ Whole Blood Immunosuppressant quality control data was used. Specifically, 210 internal quality control values were collected over 6 months from August 2017 to January 2018. The uncertainty associated to the measurement system intermediate imprecision (up) was calculated as the relative standard of values using pooled relative standard deviations of inter-month data as:

systematically evaluated, and demands a clear description of what is being measured, including the relationship between the quantity and the parameters upon which it depends. Usually this is achieved using a flowchart of steps in the measurement procedure, and by providing the formula for calculating the result. For the single laboratory validation approach, the major sources of variability can often be assessed by measurement procedure validation study. Estimates of trueness (using a reference material) and intermediate precision (e.g. using internal quality control data) can be obtained by organizing experimental work inside the laboratory. Combined with experimental investigation of important individual effects, this approach provides essentially all of the data required for uncertainty estimation. For the interlaboratory approach, the major sources of variability can often be assessed by interlaboratory studies which provide estimates of repeatability, reproducibility and, sometimes, trueness of the measurement system (estimated as a bias respecting to a known reference value). Finally, for the proficiency testing approach, the uncertainty can be estimated in a similar way that the single laboratory validation approach but considering data from interlaboratory proficiency testing schemes to estimate the trueness. Although the modelling is considered the best approach to estimate the uncertainty, in clinical laboratories real situation it is labor intensive, time base consuming, and typically too complex and cumbersome to implement. Also, few clinical laboratories validate their measurement procedures from interlaboratory studies data. Taking into account these reasons, the aim of this study was to estimate the uncertainty of cTAC values using the single laboratory validation approach, as well as two proficiency testing approaches based on internal and external quality control schemes data, to know the differences of the uncertainty results obtained between them, and to intent to decide and recommend which of these approaches could be more adequate or suitable in the clinical laboratories.

n

up =

∑ i=1

νi ·si2 νi

where n is the number of values for the month i; si, the relative standard deviation obtained in the month i; and νi, the degree of freedom of si. Different sources of uncertainty cannot adequately covered by the imprecision data, such as the uncertainty associated with the assigned values of calibrators and with the bias related to the recovery of extracted samples, the matrix effect, the carry-over and the selectivity. From all of these, only the first one contribute significantly to the overall measurement uncertainty. After performing different compatibility studies, the other sources of uncertainty mentioned could be considered negligible (data not showed). Taking into account all these reasons, only the uncertainty associated with the assigned values of calibrators was considered. According to the manufacturer of ClinCal® Calibrators, their cTAC assigned values (traceable to certified reference material ERM®-DA110a) and respective associated expanded uncertainties (k = 2) were: (1.35 ± 0.04) μg/L, (2.86 ± 0.04) μg/L, (5.66 ± 0.12) μg/L, (11.7 ± 0.2) μg/L, (23.2 ± 0.5) μg/L, and (45.1 ± 0.8) μg/L. Certified reference material ERM®-DA110a was used to evaluate the bias. Their cTAC assigned value with the associated expanded uncertainty (k = 2.3) was (7.82 ± 0.25) μg/L. To perform the evaluation, different aliquots of this material were processed, under intermediate conditions in 15 nonconsecutive days distribute between September 2017 and December 2018, according to the CLSI guideline [13]. The relative bias (δr) and its uncertainty (uδ) were calculated as follows:

2. Material and methods 2.1. Materials A commercially available lyophilized calibrator set ClinCal® Whole Blood Calibrators, for Immunosuppressants (lot: 405) was purchased from Recipe (Darmstadt, Munich, Germany). The Liquicheck™ Whole Blood Immunosuppressant Control Level 2 (lot: 26252) was supplied by Bio-Rad Laboratories (Hercules, Ca, USA). Calibrator and quality control materials were reconstituted with LC-MS grade water and stored as recommended by manufacturers. Certified reference material ERM®-DA110a (LGC Standards; Teddington, Middlesex, UK); the UNITY™ Interlaboratory Program (Bio-Rad Laboratories) (IQCS); and the Immunosuppresant Proficiency Testing Schemes (IPTS) CICTAC−Cyclosporin and Tacrolimus (LGC standards) (EQCS) were used to perform trueness studies. 2.2. Measurement procedure

x − μ⎞ δr = ⎜⎛ ⎟ ·100 ⎝ μ ⎠

Mass concentration of tacrolimus in whole blood was measured using a procedure previously described [10] based on an ultra-highperformance liquid chromatography tandem mass spectrometry (UHPLC-MS/MS) method. The measurement system used was Acquity® UPLC®-TQD® (Waters, Milford, MA, USA).

uδ =

2 δr2 + ubp + uμ2

where x is the mean obtained after processing the certified reference material; μ, the certified reference material assigned value; ubp, the relative uncertainty related to the intermediate imprecision obtained in the bias evaluation (corresponding to relative standard deviation divided by root square of the number of certified reference materials aliquots processed); and uμ, the relative uncertainty associated with the assigned value of the certified reference material. A correction factor (CF) should be applied and included as source of uncertainty if:

2.3. Measurement uncertainty estimation 2.3.1. Single laboratory validation approach Steps followed to estimate the uncertainty were: 1) specification of the measurand, 2) estimation of the uncertainty associated to the intermediate precision of the measurement system, 3) identification of any sources of uncertainty which are not adequately covered by the imprecision data, 4) estimation of the uncertainty related to correction factor for bias of the measurement system, 5) estimation of combined uncertainty, and 6) estimation of expanded uncertainty [6,8,9,11].

|δr | > 2·uδ Since that bias must be eliminated to greatest possible extent, CF 57



was applied and included as a source of uncertainty. Correction factor and their relative uncertainty (uCF) were calculated as:

CF =

Table 1 Internal quality control scheme (UNITY) data utilized to estimate the relative bias of the measurement system, used to measure mass concentration of tacrolimus in whole blood, and their uncertainty.

μ ·100 x

Month

‾x (μg/L)

nlab

μ (μg/L)

δIQCS (%)

sIQCS (%)

m

August 2017 September 2017 October 2017 November 2017 December 2017 January 2018

8.93 8.99 9.00 8.88 9.10 8.98

32 36 38 37 31 36

9.29 9.39 9.31 9.33 9.21 9.15

−3.9 −4.3 −3.4 −4.8 −1.2 −1.9

9.8 8.2 8.2 7.7 8.8 8.2

40 41 46 46 53 53

uCF = uδ Once the individual contribution of uncertainty sources to the overall uncertainty was quantified, we combined them to give a relative combined standard uncertainty (uc) according to the following equation:

uc =

2 2 u p2 + ucal + uCF

‾x, mean value obtained in our laboratory; nlab; number of quality control data obtained in our laboratory; μ, reference value (the conventional value calculated as the mean of the means of all laboratories participating in the internal quality control scheme (UNITY program) with independence of the measurement procedure used); δIQCS (%), relative bias; sIQCS (%) relative standard deviation (coefficient of variation) of all laboratories participating in the UNITY; m, number of laboratories participating in the UNITY; IQCS, internal quality control scheme.

Relative expanded uncertainty (U) was obtained by multiplying the relative combined standard uncertainty by an appropriate coverage factor (k):

U = k·uc According to the GUM [7], the coverage factor was obtained from t distribution table with a level of confidence of 95%, according to the effective degree of freedom (νeff) based on the Welch-Satterthwaite formula:

⎛ uμ − IQCS = 1.25·⎜ ⎝

uc4 (y )

νeff =

n

∑i = 1

uc =

where CVmax is the maximum permissible imprecision. This requirement was obtained from Segel et al. [15] which is based on the Consensus Statement from the 1st Strategic Conference of the European Federation of Clinical Chemistry and Laboratory Medicine (EFLM) [16] based on components of biological variation data of the measurand (CVmax = 6.0% and Umax = 14.3%).

uδ − IQCS =

⎞ / m⎟ ⎠

2 u p2 + ucal + uδ2− IQCS

δEQCS =

n 1 ⎛⎜ x i − μi ⎞⎟ ⎤·100 ·∑ ⎡ ⎥ n i=1 ⎢ ⎣ ⎝ μi ⎠ ⎦

uδ − EQCS =

2 δEQCS + uμ2− EQCS

where n is the number of participations considered to estimate the bias (n = 10 in our case); xi is the measured value obtained in our laboratory for the participation i; μi, the reference value (the conventional value assigned by the proficiency testing manufacturer) for the participation i; and uμ−EQCS, the pooled relative uncertainty associated with the reference value facilitated by manufacture's proficiency testing (Table 2). The relative combined and expanded uncertainties were calculated as described above.

2.3.2. Proficiency testing approach The uncertainty using the proficiency testing approach was estimated in a similar way than the single laboratory validation approach but calculating the bias of the measurement system considering data from two types of interlaboratory proficiency testing schemes: an interlaboratory internal quality control scheme (IQCS) and an interlaboratory external quality control scheme (EQCS). For IQCS, we used the Liquicheck™ Whole Blood Immunosuppressant internal quality control data included in the UNITY™ Interlaboratory Program. This control material shows a measured mean value near to the certified reference material. The relative bias (δIQCS) and its uncertainty (uδ−IQCS) were calculated as [17,18]: n

mi − 1

and the expanded uncertainty was calculated as described above. For EQCS, we used data from 10 external quality control participations during the survey period 2017. The relative bias (δEQCS) and its uncertainty (uδ−EQCS) were calculated as [17,18]:

Umax = 2.39·CVmax

1 ·∑ n i=1

i=1

(mi − 1)·s IQCSi2

where sIQCSi is the relative standard deviation of all laboratories participating in the month i, mi is the number of laboratories participating in the month i, and m is the mean number of participating laboratories (Table 1). The relative combined uncertainty was calculated as:

us4 (xi ) νi

where us(xi) is each individual relative standard uncertainty, and νi is the degree of freedom of us(xi). To know if the relative expanded uncertainty obtained was acceptable, it was compared with the maximum permissible relative expanded uncertainty (Umax) proposed by Haeckel et al. [14]. The uncertainty requirement, considering a 95% probability was obtained as:

δIQCS =

n

∑

3. Results 3.1. Measurement uncertainty using the single laboratory validation approach The measurand was defined as the mass concentration (μg/L) of the tacrolimus in human whole blood measured according to a laboratory-developed measurement procedure using an Acquity® UPLC®-TQD® measurement system. Furthermore, the quantity can be described as: Blood—Tacrolimus; mass concentration(ERM®-DA110a; UHPLCMS/MS) The mean of means of cTAC value and the relative uncertainty associated to the measurement system intermediate imprecision were 8.98 μg/L and 5.02%, respectively. Taking into account each uncertainty calibrator data facilitated from the manufacturer, we combined them to obtain a relative uncertainty related to the assigned values of calibrators of 2.55%.

⎡ ⎛ x i − μi ⎞ ⎤·100 ⎢⎜ μ ⎟⎥ i ⎠⎦ ⎣⎝

2 δIQCS + uμ2− IQCS

where n is the number of months considered to estimate the bias (n = 6 in our case); x i is the mean obtained in our laboratory for the month i; μi, the reference value (the conventional value calculated as the mean of the means of all laboratories participating in the program with independence of the measurement procedure used) for the month i; and uμ−IQCS, the relative uncertainty associated with the reference value calculated as [19]: 58

Mass concentration of tacrolimus in whole blood values are commonly used by the clinicians for monitoring the status of a transplant patient and for checking whether the administered dose of TAC is effective. Given that to regulate the appropriate level of immunosuppression is done to a large extent based on laboratory measured values, the clinical laboratories must provide cTAC results as accurate as possible. The measurement uncertainty can provide a quantitative indication of the quality and accuracy of these results. For a given cTAC value, measurement uncertainty represents the interval associated with a defined probability in which the true measured value should lie. In addition, this interval should fall within limits which guarantee fitness for the clinical purpose of this pharmacologically quantity. Measurement uncertainty requirements for defining fitness59

Experimental Proficiency testing manufacturer A B

Calibration procedure Precision Veracity

Commercial calibrators assigned values Intermediate imprecision Bias

B

Experimental Proficiency testing manufacturer Calibrators manufacturer A B

– 3.34 5.02 – – –

– –

– 0.65

– 3.41

6.58 – – – – 2.55

–

– 3.59 – 1.56 – – – −3.23 5.02 – – –

Experimental Experimental Calibrators manufacturer A A B

– −1.41 – 5.02 – – – – 2.55

– 0.51 –

– 1.39 –

– 2.04 –

6.68

13.0

13.2

14.3 11.8 5.99 – – – – 2.55 Calibrators manufacturer B

Proficiency testing using external quality control scheme (IPTS) (EQCS)

4. Discussion

Proficiency testing using internal quality control scheme (UNITY™) (IQCS)

Relative bias and its relative uncertainty were −3.23% and 3.59%, respectively, using data from IQCS, and 3.34% and 3.41%, respectively, using data from EQCS (Table 3). Taking into account all sources of uncertainty, relative combined and expanded uncertainties were 6.68% and 13.2% (νeff = 125; k = 1.979), respectively, for the IQCS, and 6.58% and 13.0% (νeff = 131; k = 1.978), respectively, for EQCS (Table 3). The relative expanded uncertainties obtained were lower than the relative expanded uncertainty requirement (14.3%).

Commercial calibrators assigned values Intermediate imprecision Bias Commercial calibrators assigned values Intermediate imprecision Bias

3.2. Measurement uncertainty using the proficiency testing approach

Calibration procedure Precision Veracity Calibration procedure Precision Veracity

After processing the certified reference material under intermediate conditions, mean cTAC value, CV and ubp obtained were 7.71 μg/L, 3.96% and 0.51%, respectively. Relative bias and its uncertainty were −1.41% and 2.04%, respectively (Table 3). Absolute value of the relative bias was lower than two-fold their uncertainty. Despite the bias could be considered negligible, a correction factor of 1.0143 was applied for each cTAC patient value and, in consequence, we considered the bias associated to the correction factor as a source of uncertainty. Relative uncertainty related to the correction factor was 2.04%. All sources of uncertainty (up, ucal and uCF) were combined to obtain a relative combined uncertainty of 5.99%. Furthermore, taking into account that the calculated effective degree of freedom was 254, the coverage factor used was 1.969 and the relative expanded uncertainty obtained using the single laboratory validation approach was 11.8% (Table 3). The relative expanded uncertainty obtained was lower than the maximum permissible relative expanded uncertainty (14.3%).

Single laboratory validation

x, value obtained in our laboratory; μ, reference value (the conventional value assigned by manufacture's proficiency testing); uμ-EQCS (%), relative uncertainty associated with the reference value facilitated by manufacture's proficiency testing; δEQCS (%), relative bias; m, number of laboratories participating in the external quality control scheme.

–

U (%) uc (%) uδ (%)

190 190 194 184 193 193 193 193 193 192

uμ (%)

−2.7 5.7 4.9 2.6 4.5 4.4 3.8 −1.5 6.0 5.6

ubp (%)

0.68 0.67 0.80 0.78 0.57 0.59 0.51 0.60 0.60 0.70

δr (%)

7.40 10.5 7.48 7.70 8.80 6.80 9.73 6.70 6.70 7.10

up (%)

7.20 11.1 7.85 7.90 9.20 7.10 10.1 6.60 7.10 7.50

ucal (%)

399-A 399-C 400-B 401-C 402-A 402-C 403-A 403-B 404-A 404-B

Origin of data

m

Distribution type

δEQCS (%)

Category

uμ-EQCS (%)

Uncertainty source

μ (μg/L)

Approach

x (μg/L)

Table 3 Uncertainty budget for the measurement of tacrolimus mass concentration in whole blood using the single laboratory validation and the proficiency testing approaches.

Survey

Umax (%)

Table 2 External quality control scheme data utilized to estimate the relative bias of the measurement system, used to measure mass concentration of tacrolimus in whole blood, and their uncertainty.

ucal (%), relative uncertainty related to the assigned values of calibrators; up (%), relative uncertainty associated to the measurement system intermediate imprecision; δr, relative bias related to the measurement system bias; ubp (%), relative uncertainty associated to the intermediate imprecision obtained in the bias evaluation; uμ (%), relative uncertainty associated with the assigned value of the reference material; uδ (%), relative uncertainty associated with the bias; uc (%), relative combined uncertainty; U (%), relative expanded uncertainty; Umax (%), maximum permissible measurement uncertainty value. IQCS, internal quality control scheme; EQCS, external quality control scheme.





for-purpose limits may be based on clinical outcome studies, biological variation or state-of-the-art [16], being those based on biological variation, despite their limitations [20–23], generally accepted [16,24] and used. On the other hand, as International Association of Therapeutic Drug Monitoring and Clinical Toxicology Immunosuppressive Drug Scientific Committee (IATDMCT) recommends, precision and trueness, must be assessed by comparison with meaningful clinical and physiological target intervals. The drug concentration varies with the state of the patient, so target ranges immediately after transplantation may differ from those for long-term transplant patients. Because data on longitudinal intraindividual biological variation of cTAC are still scarce, it should be feasible to estimate the measurement uncertainty from clinical therapeutic drug monitoring requirements, for example, to measure a trough concentration interval for a patient [15]. Different guidelines and studies based on different approaches have been proposed [6–9,11,17–18,25–29] in order to estimate the measurement uncertainty. All of them are based on the bottom-up approach or in the different models of top-down approaches. Taking into account that the modelling approach is difficult to implement in clinical laboratories, and considering the importance of uncertainty is increasing in all fields of health sciences [5,30–33] we decided to estimate the measurement uncertainty of cTAC values using two simpler top-down approaches. Both differ on how the trueness was estimated: using a certified reference material or a proficiency testing results. Our study showed that all top-down approaches used gave quite similar measurement uncertainty results (11.8%, 13.2% and 13.0%) (Table 3), all lower than the uncertainty requirement (14.3%). This result confirms that either two approaches could be useful to estimate the measurement uncertainty of cTAC values in clinical laboratories. However, several considerations must be taken into account. To simplify, the estimation of measurement uncertainty used only one cTAC value within therapeutic interval. In a real situation, considering the heterocedasticity of the measurement systems based on UHPLC-MS/ MS resulting in a greater uncertainty at values below the therapeutic range and a less uncertainty at values above this interval, several cTAC values covering the measuring interval should be studied. In addition, the control material used could be non-commutable with human samples affecting to the uncertainty associated to the intermediate precision. According to the measurement uncertainty results obtained and taking into account other consideration related to the lack of proficiency testing's traceable reference values, we would recommend the following order approaches: 1) the single laboratory validation approach, 2) the proficiency testing approach using IQCS, and 3) the proficiency testing approach using EQCS, whenever possible according to the each laboratory media. So, novelty part of this study was to use data from IQCS beyond to the already known and recommended use of data from EQCS to estimate measurement uncertainty.

[2] A.J. McShane, D.R. Bunch, S. Wang, Therapeutic drug monitoring of immunosuppressants by liquid chromatography-mass spectrometry, Clin. Chim. Acta 454 (2016) 1–5. [3] K. Freudenberger, U. Hilbig, G. Gauglitz, Recent advances in therapeutic drug monitoring of immunosuppressive drugs, TrAC 79 (2016) 257–268. [4] Joint Committee for Guides in Metrology, International Vocabulary of Metrology, Basic and General Concepts and Associated Terms (VIM), 3er ed. JCGM 200:2012. http://www.bipm.org/en/publications/guides/vim.html, 2012 (accessed 20 January 2018). [5] International Organization for Standardization, Medical Laboratories, Requirements for Quality and Competence, ISO 15189, ISO, Geneva, Switzerland, 2012. [6] European Federation of National Associations of Measurement, Testing and Analytical Laboratories, Measurement Uncertainty Revisited: Alternative Approaches to Uncertainty Evaluation, EUROLAB Technical report No. 1/2007 http://www.eurolab.org/documents/1-2007.pdf, (2007) , Accessed date: 20 January 2018. [7] Joint Committee for Guides in Metrology, Evaluation of Measurement Data. Guide to the Expression of Uncertainty in Measurement (GUM), Vol. 100 JCGM, 2008, http://www.bipm.org/en/publications/guides/#gum , Accessed date: 20 January 2018. [8] EURACHEM/CITAC, Guide CG4: Quantifying Uncertainty in Analytical Measurement, 3er ed., Eurachem, 2012, https://www.eurachem.org/index.php/ publications/guides/quam , Accessed date: 20 January 2018. [9] Clinical and Laboratory Standards Institute, Expression of Measurement Uncertainty in Laboratory Medicine, approved guideline, CLSI EP29-A CLSI, Wayne, PA, 2012. [10] R. Rigo-Bonnin, A. Arbiol Roca, J.M. González de Aledo-Castillo, P. Alía, Simultaneous measurement of cyclosporine A, everolimus, sirolimus and tacrolimus concentrations in human blood by UPLC–MS/MS, Chromatographia 78 (2015) 1459–1474. [11] European Federation of National Associations of Measurement, Testing and Analytical Laboratories, Guide to the Evaluation of Measurement Uncertainty for Quantitative Test Results, EUROLAB Technrical report No. 1/2006 http://www. eurolab.org/documents/EL_11_01_06_387%20Technical%20report%20-%20Guide_ Measurement_uncertainty.pdf, (2006) , Accessed date: 20 January 2018. [12] International Union of Pure and Applied Chemistry, International Federation of Clinical Chemistry, Properties and units in the laboratory sciences, part X, properties and units in general clinical chemistry, Pure Appl. Chem. 72 (2000) 747−972. [13] Clinical and Laboratory Standards Institute, Evaluation of Precision of Quantitative Analysis Measurement Procedures, approved guideline—Third Edition, CLSI EP05A3 CLSI, Wayne, PA, 2014. [14] R. Haeckel, W. Wosniok, E. Gurr, B. Peil, Permissible limits for uncertainty of measurement in laboratory medicine, Clin. Chem. Lab. Med. 53 (2015) 1161–1171. [15] C. Seger, M. Shipkova, U. Christians, E.M. Billaud, P. Wang, D.W. Holt, et al., Assuring the proper analytical performance of measurement procedures for immunosuppressive drug concentrations in clinical practice: recommendations of the International Association of Therapeutic Drug Monitoring and Clinical Toxicology Immunosuppressive Drug Scientific Committee, Ther. Drug Monit. 38 (2016) 170–189. [16] S. Sandberg, C.G. Fraser, A.R. Horvath, R. Jansen, G. Jones, W. Oosterhuis, et al., Defining analytical performance specifications: consensus statement from the 1st strategic conference of the European Federation of Clinical Chemistry and Laboratory Medicine, Clin. Chem. Lab. Med. 53 (2015) 833–835. [17] Nordic Innovation Center, Handbook for Calculation of Measurement Uncertainty in Environmental Laboratories, NORDTEST Technical report 537—Edition 3.1 http://www.nordtest.info/index.php/technical-reports/item/handbook-forcalculation-of-measurement-uncertainty-in-environmental-laboratories-nt-tr-537edition-3.html, (2012) , Accessed date: 20 January 2018. [18] A. Padoan, G. Antonelli, A. Aita, L. Sciacovelli, M. Plebani, An approach for estimating measurement uncertainty in medical laboratories using data from long-term quality control and external quality assessment schemes, Clin. Chem. Lab. Med. 55 (2017) 1696–1701. [19] International Organization for Standardization, Statistical Methods for Use in Proficiency Testing by Interlaboratory Comparisons, ISO 13528 ISO, Geneva, Switzerland, 2005. [20] G. Cooper, N. De Jonge, S. Ehrmeyer, J. Yundt-Pacheco, R. Jansen, C. Ricós, et al., Collective opinion paper on findings of the 2010 convocation of experts on laboratory quality, Clin. Chem. Lab. Med. 49 (2011) 793–802. [21] X. Fuentes-Arderiu, A. Padró-Miquel, R. Rigo-Bonnin, Disadvantages of using biological variation data for reference change values, Clin. Chem. Lab. Med. 50 (2012) 961. [22] A. Carobene, Reliability of biological variation data available in an online database: need for improvement, Clin. Chem. Lab. Med. 53 (2015) 871–877. [23] W.A. Bartlett, F. Braga, A. Carobene, A. Coskun, R. Prusa, P. Fernández-Calle, et al., A checklist for critical appraisal of studies of biological variation, Clin. Chem. Lab. Med. 53 (2015) 879–885. [24] R. Haeckel, W. Wosniok, E. Gurr, B. Peil, Supplements to a recent proposal for permissible uncertainty of measurements in laboratory medicine, Lab. Medizin. 40 (2016) 141–145. [25] R. Zhou, Y. Qin, H. Yin, Y. Yang, Q. Wang, Measurement uncertainty of γ-glutamyltransferase (GGT) in human serum by four approaches using different quality assessment data, Clin. Chem. Lab. Med. 56 (2018) 242–248. [26] M. Cui, L. Xu, H. Wang, S. Ju, S. Xu, R. Jing, Combining Nordtest method and bootstrap resampling for measurement uncertainty estimation of hematology

5. Conclusions We showed a practical and detailed example to estimate the measurement uncertainty for the cTAC using two top-down approaches. After performing the single laboratory validation and proficiency testing approaches, we observed that their results were quite similar. This fact could confirm that both approaches would be suitable for estimating uncertainty of cTAC values in clinical laboratories. Finally, we hope that this study help and motivate clinical laboratories to perform measurement uncertainty studies for this or other pharmacological quantities. References [1] Y. Zhang, R. Zhang, Recent advances in analytical methods for the therapeutic drug monitoring of immunosuppressive drugs, Drug Test Anal. 10 (2018) 81–94.

60



[30] B. Magnusson, H. Ossowicki, O. Rienitz, E. Theodorsson, Routine internal- and external-quality control data in clinical laboratories for estimating measurement and diagnostic uncertainty using GUM principles, Scand. J. Clin. Lab. Invest. 72 (2012) 212–220. [31] W.P. Oosterhuis, H. Bayat, D. Armbruster, A. Coskun, K.P. Freeman, A. Kallner, et al., The use of error and uncertainty methods in the medical laboratory, Clin. Chem. Lab. Med. 56 (2018) 209–219. [32] A. Kallner, Estimation of uncertainty in measurements in the clinical laboratory, Clin. Chem. Lab. Med. 51 (2013) 2249–2251. [33] Comite Europeen de Normalisation, Health Informatics, Expression of Results of Measurements in Health Sciences, ENV 12435 CEN, Brussels, 2006.

analytes in a medical laboratory, Clin. Biochem. 50 (2017) 1067–1072. [27] G. Matar, B. Poggi, R. Meley, C. Bon, L. Chardon, K. Chikh, et al., Uncertainty in measurement for 43 biochemistry, immunoassay, and hemostasis routine analytes evaluated by a method using only external quality assessment data, Clin. Chem. Lab. Med. 53 (2015) 1725–1736. [28] L. Rami, F. Canalias, An approach to establish the uncertainty budget of catalytic activity concentration measurements in a reference laboratory, Clin. Chem. Lab. Med. 53 (2015) 743–751. [29] J.H. Lee, J.H. Choi, J.S. Youn, Y.J. Cha, W. Song, A.J. Park, Comparison between bottom-up and top-down approaches in the estimation of measurement uncertainty, Clin. Chem. Lab. Med. 53 (2015) 1025–1032.

61