IFCC international conventional reference procedure for the measurement of free thyroxine in serum

Transcription

1 Clin Chem Lab Med 2011;49(8): by Walter de Gruyter Berlin Boston. DOI /CCLM IFCC international conventional reference procedure for the measurement of free thyroxine in serum International Federation of Clinical Chemistry and Laboratory Medicine (IFCC) Working Group for Standardization of Thyroid Function Tests (WG-STFT) 1) Sofie K. Van Houcke 1, Katleen Van Uytfanghe 1, Eri Shimizu 2, Wataru Tani 2, Masao Umemoto 2 and Linda M. Thienpont 1, * 1 Laboratory for Analytical Chemistry, Faculty of Pharmaceutical Sciences, Ghent University, Ghent, Belgium 2 Laboratory of Reference Material Institute for Clinical Chemistry Standards (ReCCS), Kawasaki, Kanagawa, Japan Abstract The IFCC Working Group for Standardization of Thyroid Function Tests proposes a candidate international conventional reference procedure (RMP) for measurement of the amount-of-substance concentration of free thyroxine in plasma/serum at physiological ph 7.40 and temperature (37.08C). The unit for reporting measurement results is, by convention, pmol/l. The RMP is based on equilibrium dialysis isotope dilution-liquid chromatography/tandem mass spectrometry (ED-ID-LC/tandem MS). The rationale for proposing a conventional RMP is that, because of the physical separation step, it is unknown whether the measurement truly reflects the concentration of free thyroxine (FT4) in serum. Therefore, the ED part of the RMP has to strictly adhere to the following conditions: use of a dialysis buffer with a biochemical composition resembling the ionic environment of serum/plasma as closely as possible; buffering of the sample to a ph of 7.40 (at 37.08C) before dialysis, however, without additional dilution; dialysis in a device with a dialysand/ dialysate compartment of identical volume and separated by a membrane of regenerated cellulose and adequate cut-off; thermostatic control of the temperature during dialysis at 37.08C"0.508C. The convention does not apply to the ID- LC/tandem MS part, provided it is eligible to be nominated 1) The current membership of the WG-STFT is as follows: L. Thienpont (BE, Chair); S. Faye (UK); G. Beastall (UK); J. Faix (US); T. Ieiri (JP); R. Janzen (US); G. Miller (US); D. Montague (UK); J. Nelson (US); F. Quinn (US); C. Ronin (FR); H.A. Ross (NL); M. Rottmann (DE); J. Thijssen (NL); B. Toussaint (BE). *Corresponding author: Prof. Linda M. Thienpont, Laboratory for Analytical Chemistry, Ghent University, Harelbekestraat 72, 9000 Ghent, Belgium Phone: q , Fax: q , linda.thienpont@ugent.be Received March 14, 2011; accepted March 28, 2011; previously published online June 21, 2011 for review by the Joint Committee for Traceability in Laboratory Medicine. Here, we describe the ED procedure, inclusive its validation and transferability, in greater detail. We recommend a protocol for successful calibration, measurement and monitoring of the accuracy/trueness and precision of the candidate conventional RMP. For details on our ID-LC/tandem MS procedures, we refer to the Supplement. Keywords: equilibrium dialysis; isotope dilution-liquid chromatography-mass spectrometry; ph 7.40; physiological conditions; temperature 37.08C. Introduction The IFCC Working Group for Standardization of Thyroid Function Tests (WG-STFT) is mandated among others with establishing metrological traceability of measurement procedures for free thyroxine (FT4) in plasma/serum. Meanwhile, the rationale for this mandate has been documented in a report on FT4 concentrations measured in a panel of sera by several immunoassays and routinely applied mass spectrometric methods (1). The most discrepant results differed by 40%. In 2007 the WG proposed to define the measurand, which is the quantity intended to be measured by FT4 measurement procedures, Plasma/Serum Thyroxine (free); substance concentration (2). According to this IUPAC/IFCC format, the component is thyroxine that is not bound to proteins, the kind-of-quantity to measure the amount-of-substance concentration and the system plasma or serum at physiological conditions (ph 7.40, temperature 37.08C). The preferred unit for expression of measurement results is pmol/l. The WG proposed to establish metrological traceability of serum FT4 measurements to a candidate international conventional reference measurement procedure (RMP) based on equilibrium dialysis (ED) combined with determination of the T4 concentration of the dialysate with a trueness-based isotope dilution-liquid chromatography/tandem mass spectrometry (ID-LC/tandem MS) RMP (3). The ED part of the RMP is intended to generate a protein-depleted measurement system by physical separation of protein-bound and FT4. The reasons to select ED and not ultrafiltration are explained in (3). The rationale for proposing a conventional RMP was that, because of the physical separation step, even if performed under theoretically optimal conditions to respect the free-bound equilibri- 2011/0156

2 1276 Van Houcke et al.: Free thyroxine conventional reference procedure um, it is unknown whether the T4 measured in the dialysate truly reflects the FT4 concentration in serum (3, 4). Therefore, the WG preferred to adopt the convention that ED should be performed with strict adherence to a predefined procedure. This implies that the measurand is operationally defined as thyroxine in the dialysate from ED of serum prepared under defined conditions. The convention does not apply to the ID-LC/tandem MS part of the RMP, provided it is calibrated with the certified T4 primary calibrator IRMM 468 listed by the Joint Committee for Traceability in Laboratory Medicine (JCTLM) (5). In addition, it is stateof-the-art to start from a JCTLM listed ID-LC/tandem MS method for total T4, but to optimize it towards measurement of picomolar concentrations typical for serum FT4. The WG proposed the ED ID-LC/tandem MS procedure originally described by Van Uytfanghe et al. as preliminary candidate international conventional RMP, however, with the requirement to document its transferability (3, 4). The ED conditions were based on the recommendations of the C45-A consensus guideline, i.e., i) the dialysis buffer has a biochemical composition that mimics the ionic environment of the serum as closely as possible, ii) serum is buffered at ph 7.40 (at 37.08C), iii) the sample is not diluted before dialysis and the dilution factor does not exceed the inherent 1:2 dilution of dialysis, iv) the temperature during dialysis is kept constant at 37.08C within a "0.58C interval (as controlled by a calibrated thermometer) (6). Here, we describe the final version of the candidate international conventional RMP for FT4 in serum as implemented now in 2 laboratories wghent University and Reference Material Institute for Clinical Chemistry Standards (ReCCS)x.We mainly focus on the methodological aspects which are part of the convention, whereas the variants in the ID-MS design are detailed in the online data supplement. Note that to distinguish between conventional and variable analytical conditions, we used for the former the imperative form. We also present the validation and method transfer data. Equilibrium dialysis Reagents (4) Use high-quality reagents and water (water purity: resistivity Mohm* cm; max concentration of calcium, potassium, magnesium and sodium: 0.10 mg/g each, max residue on evaporation: 0.20 wm/vx). Prepare 2 HEPES buffers wnote: for the rationale of the composition, see (4)x: (i) M HEPES buffer (further referred to as concentrated HEPES buffer) comprising mol/l N-(2-hydroxyethyl)piperazine-N9-(2-ethylsulfonic acid) and 0.22 mol/l sodium hydroxide. Adapt the ph to 7.50"0.03 at room temperature (RT; min 208C, max 258C) using 8.18 mol/l sodium hydroxide or 37% hydrochloric acid. Make the buffer in a polypropylene tube and store no longer than 5 days at 48C. Before use, check the ph (at RT) and adapt if necessary; (ii) mm HEPES buffer (further referred to as dialysis buffer): weigh, in the described order, mmol/l N-(2-hydroxyethyl)piperazine-N9-(2-ethylsulfonic acid), 22.5 mmol/l sodium hydroxide, 91.6 mmol/l sodium chloride, 1.65 mmol/l potassium phosphate, 2.68 mmol/l potassium chloride, 1.12 mmol/l magnesium sulfate heptahydrate, 5.00 mmol/l urea. Dissolve in approximately 800 ml water. Weigh 1.90 mmol/l calcium chloride dihydrate, dissolve separately and add to the buffer. Add 8.00 mmol/l sodium azide. Dilute further until 1 L of water is added. Adapt the ph to 7.50"0.03 (at RT) using 8.18 mol/l sodium hydroxide or 37% hydrochloric acid. The deviation of the weight per substance should not exceed the "2% limit. Store 3 50-mL portions of the dialysis buffer in polypropylene tubes no longer than 3 days at 48C. Even under this limitation, do not use if any precipitation is observed. Check the ph (at RT) of each portion before applying and adapt if necessary. Use the rest of the buffer for washing the membranes. Dialysis equipment and membranes Dialyse the serum in a multi-cell equilibrium dialyser. Incubate in a water bath with constant agitation by a rotating apparatus. Maintain, during dialysis, the temperature of the water at 37.08C"0.58C, by constantly monitoring with an immersion thermostatic system. Use dialysis cells made of Teflon and consisting of two identical parts between which a flat membrane is fitted. In our dialysis equipment (7), the volume of each half-cell is 1.36 ml, the optimal working volume 1.0 ml with a minimum of 0.7 ml (Supplement Figure 1). The membrane surface area is 4.5 cm 2. The wall of each half-cell contains three holes, two close to each other (one for filling the cell, one for air displacement) and one approximately 1558 away (for emptying the cells). Always fill the same compartments of the dialysis cells with serum or buffer. Clean the cells by soaking them max 5 min in detergent we.g., 2% (v/v) R.B.S. 35x, followed by thorough cleaning with deionized water and finally soaking in water during 15 min. Dry the cells overnight at 378C. Use membranes of regenerated cellulose with a cut-off value of 5 kda. Wash them 3 times with water (same quality as used for the buffers) and twice with dialysis buffer. Equilibrate each time for 15 min at 378C. Membranes can be washed max one day in advance when stored at 48C indialysis buffer. Conditions for equilibrium dialysis (6) Before ED, adapt the ph of the sera to 7.40"0.03 (measured at 378C), by adding 100 ml of the concentrated HEPES buffer to 1 ml sample. Add 1 ml of ph adapted serum to one side of the dialysis cell, add an equal volume of dialysis buffer to the other side. The latter is necessary to restrict the dilution to its minimum. Incubate the cells for 4 h at 37.0"0.58C under continuous agitation in the temperature controlled water bath (see above). If necessary, working volumes of the dialysand and dialysate buffer down to 0.7 ml can be used, however, this is not recommended.

3 Van Houcke et al.: Free thyroxine conventional reference procedure 1277 Collect the dialysates in weighed 1.5 ml glass vials containing a carrier (see below) and a weighed amount of internal standard (we use w 13 C 9 x-t4 and w 13 C 6 x-t4, see the online data Supplement). Use disposable Teflon liners (with an outer diameter of 2.5 mm) for emptying the cells. Blow air ()1.5 ml, with a pipette) through the cell to obtain max recovery of dialysate. While emptying the first set of cells, keep the remaining cells at 37.08C. Empty the cells one by one, vortex each vial immediately to mix the dialysate with the internal standard. Sample treatment after dialysis and measurement of FT4 in dialysate Equilibrate the dialysate/internal standard mixture for 1 h prior to sample preparation. If necessary, it can be stored overnight at 48C. Because variants are permissible in the sample pretreatment and the ID-LC/tandem MS procedure, we refer to the conditions described in the online data Supplement. Calibration Primary calibrator and internal standard Use as primary calibrator the T4 certified reference material IRMM 468 (8) listed by the JCTLM (5). It has a purity of 98.6%"0.70% (expanded uncertainty, ks2). Use an internal standard that is a T4 analog labeled with stable isotopes wnote: in the JCTLM listed total T4 RMPs ID is done with the following internal standards: w 13 C 2 x-t4, w 13 C 6 x-t4, w 13 C 9 x-t4 or w 2 H 5 x-t4 (5)x. Protection against adsorption Use monoiodotyrosine (MIT) as a carrier to prevent the low amounts of T4 from adsorption to vials etc. Prepare stock and working solutions of MIT in methanol. Add nmol of MIT to the vials used for sampling of the calibrators and those used for collection of the dialysates. Before evaporation of the methanol under N 2, equilibrate the vials containing MIT working solutions by constant rolling for 20 min. The working solution of the internal standard also contains 5000 times as much MIT as there is isotopically labeled T4-analog. Preparation of solutions of standard, internal standard and calibrators Prepare three individual standard stock solutions and dilute each to a working solution. Control all volumetric steps gravimetrically. Weigh the primary calibrator in an amount appropriate to keep the uncertainty below 0.05%, e.g., we weigh approximately 30 mg (38.6=10 6 pmol) with an analytical balance with a readability down to 10 5 g, and dissolve it in 12 ml of methanol (HPLC-grade) containing 100 ml of 25% ammonium hydroxide. After total dissolution, store these stock solutions at 208C. Make consecutive dilutions to reach working solutions of appropriate concentrations. Verify the trueness of the subsequent dilutions. For example, we make two consecutive intermediate solutions in methanol. We first pipet 1 2 ml pure methanol, then add an appropriate amount of respectively the stock solution or higher concentrated intermediate solution. Finally, we add methanol until the required total volume is reached and store the solutions at 208C. To avoid adsorption phenomena wnote: we observed them during optimization of our procedure; see the online data Supplement for the experiments done in this regardx, we make the next dilution in T4-depleted serum using the same pipetting sequence as described above for the solutions in methanol. We assess beforehand the T4 concentration of the depleted serum with a total T4 RMP (5) and account for it in the calculation of the concentration of the prepared dilution. We dilute this solution with dialysis buffer in two consecutive steps (1q7 and 1q7.75) to the final working solution. We always add an appropriate volume of intermediate solution to the dialysis buffer. We divide the final working solutions in 1-mL single-use portions and store at 208C. Note that for dilutions at this level, buffer can be used without adverse effects on the trueness of the final standard solution due to adsorption of T4. Table 1 gives a schematic overview of the complete dilution protocol in our procedure. We express concentrations in amount of substance/g. The final concentration of our T4 working solutions is pmol/g. Make internal standard solutions by use of a similar dilution protocol, however, all dilutions can be made in methanol. The working solutions contain 5000 times as much MIT as there is internal standard. Prepare the measurement calibrators by weighing the internal standard solution in a MIT coated vial (0.173 nmol MIT), followed by the T4 solution, both in an appropriate Table 1 Schematic overview of the dilution protocol used to prepare the T4 working solutions. Solution Approximate Solvent Protocol concentration Stock solution 4.06 mmol/g Methanol 38.6=10 6 pmol standard materialq12 ml methanol Intermediate solution mmol/g Methanol 2 ml of methanol, 300 ml of stock, 9.7 ml of methanol Intermediate solution nmol/g Methanol 2 ml methanol, 380 ml intermediate 1, 9.62 ml methanol Intermediate solution pmol/g T4 depleted serum 2 ml T4 depleted serum, 375 ml intermediate 2, ml T4 depleted serum Intermediate solution pmol/g Dialysis buffer 7 ml of dialysis buffer, 1 ml intermediate 3 Working solution 1.29 pmol/g Dialysis buffer 7.75 ml of dialysis buffer, 1 ml intermediate 4

4 1278 Van Houcke et al.: Free thyroxine conventional reference procedure amount to achieve the required isotope ratio. It is recommended to work at an isotope ratio that warrants minimum contribution to the uncertainty of MS measurement, e.g., the 1/1 or bracketing (0.7/1, 1/1 and 1.3/1) ratio (9, 10). Unknown samples should be analysed at an identical isotope ratio or one encompassed by the bracketing calibrators, which requires their preliminary quantification, e.g., with a routine immunoassay. Since the T4 standard solutions are made in dialysis buffer, the measurement calibrators need to be purified with the same procedure as the dialysates. Measurement protocol Spuriously high FT4 concentrations can be obtained due to occasional protein leakage during ED (11). Therefore, process and measure each sample at least in duplicate, preferably, in triplicate on independent occasions. Monitoring of the accuracy/trueness and precision We recommend to monitor the precision, accuracy and trueness by a 2-step internal quality control (IQC) procedure: i) monitor the whole FT4 measurement procedure, inclusive ED, by use of lyophilized or pooled sera (target: between 10 and 20 pmol/l) further referred to as ED controls ; ii) exclusively monitor the precision and accuracy of the ID/MS measurements by use of sera certified for their total T4 concentration we.g., 95.4 nmol/l"1.0 nmol/l and nmol/l" 1.1 nmol/l (12)x and gravimetrically diluted with dialysis buffer to concentrations of approximately 1.29 pmol/g (further referred to as MS controls ). Dilution and sampling is done according to a strict protocol. For the dilution, pipet in the described sequence, nmol MIT, 10 ml of dialysis buffer and an amount of serum corresponding to 12.9 pmol T4. After equilibration, store the solution in single-use portions at 208C. For sampling of the MS controls, weigh first the internal standard (approx pmol) in a MIT coated vial (0.173 nmol), add 0.9 ml of water and weigh subsequently the sample (a volume corresponding to pmol T4). We recommend to perform each day of analysis both IQC procedures, each with analysis of samples in duplicate. Calculation of results convention In principle the FT4 measurement result obtained by the candidate conventional RMP stands for the T4 concentration in serum water (units: pmol/kg). However, taking into consideration that in the routine laboratory FT4 measurements are expressed as concentrations in serum (pmol/l), the WG- STFT preferred to avoid proliferation of units and adopted the convention to report the ED ID-LC/tandem MS measurements also in pmol/l. Validation Complementary to the initial validation of the candidate conventional RMP by Van Uytfanghe et al. (4), additional validation experiments were performed by the laboratory of Ghent University. They were intended to assess that ED was performed with compliance to the predefined requirements and with sufficient robustness against certain variables. A detailed description of the experimental set up, the samples used, etc. can be found in the online data Supplement. Verification of ph and stability during dialysis Requirement: ph: 7.40"0.03. The ph was verified at different levels: the ph-adapted serum before dialysis (sdialysand) wph"sd: 7.38"0.02 at 36.9"0.1 (SD) 8C; ns9x, the dialysis buffer w7.34"0.01 at 36.7"0.3 (SD) 8C; ns3x, the serum after dialysis w7.40"0.01 at 37.1"0.2 (SD) 8C; ns9x and the dialysate w7.40"0.02 at 37.0"0.1 (SD) 8C; ns9x. The experiments confirmed that the required ph was reached and that the buffer capacity was sufficient to maintain the ph of the dialysand and dialysate constant during the entire dialysis process. Stability of temperature during dialysis Requirement: 37.0"0.58C. Verification of adequate temperature control during dialysis was done using a water resistant temperature logger. Its accuracy was confirmed against a calibrated thermometer ("0.58C). The temperature within one dialysis series was typically 37.18C"0.0028C (SD), the average median temperature over several dialysis series was 37.18C"0.0118C (SD), which confirmed compliance with the required temperature conditions. Dialysis time necessary to reach equilibrium With regard to determining the dialysis time necessary to reach equilibrium, samples submitted to 4, 5, 6 and 7 h of incubation were compared (ns5 each). A p-values0.198 in single-factor ANOVA (as0.05) confirmed that no significant difference could be observed between dialysis during 4 h vs. 5, 6 and 7. Influence of membrane cut-off and brand The candidate RMP should perform ED with a membrane of regenerated cellulose (3), however, we investigated the robustness of the dialysis procedure against changes of the membrane cut-off and brand. Membranes with a cut-off of 5 kda, 10 kda, 6 8 kda and kda and from 2 different manufacturers were tested (see online data Supplement). This was done by dialysis of aliquots taken from a pool of serum samples (17.4 pmol/l FT4) with each of the membranes (within one experiment, ns5 for each membrane), followed by FT4 quantification of the dialysates. No significant difference between the FT4 results was found wps0.422 in single-factor ANOVA (as0.05)x, confirming robustness of the

5 Van Houcke et al.: Free thyroxine conventional reference procedure 1279 dialysis procedure against the impact of changes in cut-off and brand of the membrane of regenerated cellulose. Generation of non-esterified fatty acids during dialysis Since the presence of non-esterified fatty acids is known to alter the FT4 concentration (13), we investigated whether their concentration increased during dialysis. We measured in 20 sera the concentration of non-esterified fatty acids before and after dialysis (see online data Supplement). A paired t-test (2-sided, as0.05) confirmed that there is no significant generation of non-esterified fatty acids in serum submitted to a 4-h dialysis process (ps0.895). Verification of the trueness of the T4 standard solutions Note beforehand: from our total T4 RMP, we can confirm the trueness of the T4 standard solutions. The total T4 working solution corresponds with intermediate three of the described FT4 procedure (see Table 1), however, is prepared in methanol. We used this solution as anchor point for trueness. To verify whether adsorption phenomena upon further dilution with methanol adversely impact the trueness, we compared the total T4 working solution with a solution in methanol at a T4 concentration of approximately 1.29 pmol/g (a calibrator concentration appropriate for measurement of pmolar FT4 concentrations) and found a difference of 7.9% (ns6, p-value ). This observation made additional experiments necessary that finally lead to the protocol described in Table 1 (the online data Supplement summarizes the experiments done). Trueness verification was additionally done from measurement of the MS controls. The results had a mean bias of q0.03%, which was within the 1.5% specification limit whalf of the bias specification for routine measurement procedures (6)x. Precision and trueness data Since the initial publication of the candidate conventional RMP (4), it has been used in different projects we.g., (1, 14)x. Based on the results obtained on these occasions, the withinrun, between-run and total CVs could be calculated using ANOVA. For the procedure, inclusive ED, a within-run CV of 2.8%, a between-run CV of 2.4% and a total CV of 3.7% was obtained (target: 10 pmol/l, ns61 duplicates); for the MS measurement procedure (exclusive ED) the CVs were, respectively: 1.7%; 1.0% and 2.0% (target: 1.29 pmol/g, ns66 duplicates). The contribution of ED to the aforementioned CVs was estimated to be 2.3%, 2.2% and 3.2%, respectively. Monitoring of the trueness of the measurement procedure with and without inclusion of ED, documented a mean deviation of the target of 0.2% (ns61 duplicates) and q0.03% (ns66 duplicates). Uncertainty As described in the online data Supplement, the combined uncertainty was estimated first by grouping the different sources of uncertainty (group 1: all sources attributed to the calibration procedure; group 2: the total imprecision of the measurement procedure, group 3: unspecific interferences) and combining them by error propagation. For calculation of the expanded uncertainty a coverage factor (k) of 2 was used, which is justified because the imprecision component in the combined uncertainty is estimated from a considerable number of replicates (ns61). For the recommended measurement protocol of ns3 (singlicate measurement of a sample on 3 independent occasions), the laboratory of Ghent University estimated the expanded uncertainty at 7.6%. Note that it is constant over the measurement range ( pmol/l). Transferability of the method The WG considered the documentation of the transferability of the candidate conventional RMP as an essential part of the validation process. This required implementation of the procedure in a second laboratory, sufficient familiarization of the latter with the procedure and organization of an intercomparison study. The laboratory of Ghent University Figure 1 Intercomparison between the laboratories of Ghent University and ReCCS. (A) Scatter plot and ordinary least squares regression parameters (xslaboratory of Ghent University; yslaboratory of ReCCS; dashed line: ysx); (B) % difference plot (relative to the mean), full circles: laboratory of Ghent University, open circles: laboratory of ReCCS.

6 1280 Van Houcke et al.: Free thyroxine conventional reference procedure worked in this context together with the laboratory of ReCCS. After training of a staff member of the latter laboratory at Ghent University, the procedure was implemented in the laboratory of ReCCS. Afterwards, all additional information on optimization steps done by the laboratory of Ghent University was exchanged. When both laboratories were ready to perform an intercomparison, 15 samples (concentration range pmol/l) were measured in parallel, each sample in four replicates. The mean total CVs were, respectively 3.1% (range: 1.3% 4.5%) (laboratory of Ghent University) and 4.2% (range 1.0% 8.9%) (laboratory of ReCCS). Figure 1 represents the final intercomparison study in a scatter plot with regression analysis results, and in a % difference plot. Both plots show that the agreement between the results of the laboratories of Ghent University and ReCCS is reasonable (% difference within "4.2%). Acknowledgments We gratefully acknowledge the financial support of the European Commission through project G6RD-CT (for the period ). The IFCC Chair of the WG-STFT gratefully acknowledges the scientific support by the following in-vitro diagnostic companies (in alphabetical order): Abbott Diagnostics (USA), Beckman Coulter Inc. (USA and Europe), biomérieux (France), DiaSorin S.p.A (Italy), Olympus Life Science Europa GmbH (Ireland), Ortho- Clinical Diagnostics (UK), Roche Diagnostics GmbH (Germany), Siemens Healthcare Diagnostics (USA), TOSOH Corp. (Japan). Their continuous interest in the project of the WG was a real stimulus to optimize the FT4 measurement procedure to a level that it became prone to be recommended by the IFCC SD as international conventional reference measurement procedure. She is also grateful for the intercomparisons done with two other ED ID-LC/tandem MS measurement procedures running at ARUP Laboratories (USA, represented by A. Rockwood and B. Yue) and Mayo Foundation (USA, represented by L. Dodge). The exchange of practical information and scientific discussions were a furtherance to improve the FT4 measurement procedure of Ghent University. Last but not least, she particularly acknowledges the invaluable advice regarding ED received from one of the members of the WG, H.A. Ross and his staff member J. Laurant. Author contributions All authors confirm they have contributed to the intellectual content of this paper by significant contributions to the conception and design, acquisition, analysis and interpretation of data, drafting or revising the article for intellectual content, and final approval. Conflict of interest statement Authors conflict of interest disclosure: The authors stated that there are no conflicts of interest regarding the publication of this article. Certain commercial equipment, instruments, materials, and companies are identified in this report to adequately specify the experimental procedure. Such identification does not imply recommendation or endorsement by the IFCC, nor does it imply that the equipment, instruments, materials, and companies identified are the best available for the purpose. Research funding: None declared. Employment or leadership: None declared. Honorarium: None declared. References 1. Thienpont LM, Van Uytfanghe K, Beastall G, Faix JD, Ieiri T, Miller WG, et al. IFCC Working Group on Standardization of Thyroid Function Tests. Report of the IFCC Working Group for Standardization of Thyroid Function Tests; part 2: free thyroxine and free triiodothyronine. Clin Chem 2010;56: Thienpont LM, Beastall G, Christofides ND, Faix JD, Ieiri T, Miller WG, et al. International Federation of Clinical Chemistry and Laboratory Medicine (IFCC) IFCC Scientific Division Working Group for Standardization of Thyroid Function Tests (WG-STFT). Measurement of free thyroxine in laboratory medicine proposal of measurand definition. Clin Chem Lab Med 2007;45: International Federation of Clinical Chemistry and Laboratory Medicine IFCC, IFCC Scientific Division Working Group for Standardization of Thyroid Function Tests WG-STFT. Thienpont LM, Beastall G, Christofides ND, Faix JD, Ieiri T, Jarrige V, et al. Proposal of a candidate international conventional reference measurement procedure for free thyroxine in serum. Clin Chem Lab Med 2007;45: Van Uytfanghe K, Stöckl D, Ross HA, Thienpont LM. Use of frozen sera for FT4 standardization: investigation by equilibrium dialysis combined with isotope dilution-mass spectrometry and immunoassay. Clin Chem 2006;52: Bureau International des Poids et Mesures Joint Committee for Traceability in Laboratory Medicine. JCTLM database: laboratory medicine and in vitro diagnostics. Available at: Accessed: 11 March Clinical Laboratory and Standards Institute (CLSI). Measurement of Free Thyroid Hormones; Approved Guideline. CLSI document C45-A wisbn x. CLSI, Wayne, PA, Harvard Apparatus. Available at: uk/webapp/wcs/stores/servlet/haisku3_10001_11555_64325_-1_ HAUK_ProductDetail_N_37263_37267_ Accessed: 11 March Toussaint B, Klein CL, Wiergowski M. The certification of the mass fraction of thyroxine in a CRM intended for calibration: Certified Reference Material IRMM-468. Report EUR EN. Luxembourg: Office for Official Publications of the European Communities, Thienpont LM, Van Nieuwenhove B, Stöckl D, De Leenheer AP. Calibration for isotope dilution mass spectrometry description of an alternative to the bracketing procedure. J Mass Spectrom 1996;31: Siekmann L, Breuer H. Determination of cortisol in human plasma by isotope dilution-mass spectrometry. Definitive methods in clinical chemistry. J Clin Chem Clin Biochem 1982;20: Fritz KS, Wilcox RB, Nelson JC. Quantifying spurious free T4 results attributable to thyroxine-binding proteins in serum dialysates and ultrafiltrates. Clin Chem 2007;53: Thienpont LM, Van Uytfanghe K, Marriot J, Stokes P, Siekmann L, Kessler A, et al. Metrologic traceability of total thy-

7 Van Houcke et al.: Free thyroxine conventional reference procedure 1281 roxine measurements in human serum: efforts to establish a network of reference measurement laboratories. Clin Chem 2005;51: Lim CF, Bai Y, Topliss DJ, Barlow JW, Stockigt JR. Drug and fatty acid effects on serum thyroid hormone binding. J Clin Endocrinol Metab 1988;67: Anckaert E, Poppe K, Van Uytfanghe K, Schiettecatte J, Foulon W, Thienpont LM. FT4 immunoassays may display a pattern during pregnancy similar to the equilibrium dialysis ID-LC/ tandem MS candidate reference measurement procedure in spite of susceptibility towards binding protein alterations. Clin Chim Acta 2010;411: