Comparability of four clinical laboratory measurement methods for GGT and commutability of candidate reference materials

Abstract Background This study was conducted to evaluate the progress in the standardization of the gamma‐glutamyl transferase (GGT) to achieve metrological traceability of routine in vitro diagnosis (IVD) medical devices. Methods We collected 25 single fresh frozen serum samples for GGT analysis. Candidate reference materials (RMs), calibrators, internal quality controls (IQC), and external quality assessment (EQA) materials from the National Center for Clinical Laboratory (NCCL), Beijing Center for Clinical Laboratory (BCCL), and College of American Pathologists (CAP) were randomly added to these serum samples. A total of 42 samples were examined using IFCC reference method and four different IVD medical devices to perform the comparability and commutability study. Results The four IVD medical devices achieved trueness assessment within the measurement range. Linear analysis showed the agreement of Siemens ADVIA 2400, Hitachi 7600‐020/BioSino, Beckman AU 5800, and Roche Cobas 501 with the reference method. These assay pairs were comparable at the medical decision levels. The GGT in‐house candidate RMs, and Beckmann and Roche calibrators were all within the limits of the 95% prediction intervals, the commutability of BioSino calibrators was indeterminate, and some internal and external quality controls were not commutable for comparisons of certain IVD medical devices vs the reference method. Conclusions By comparing with the reference method, we found that performance of GGT conventional measurement systems to be traceable to the higher order references was improved. The commutable materials for calibration and trueness controls of routine methods were significant to promote the standardization of GGT analysis.


| INTRODUC TI ON
Gamma-glutamyl transferase (GGT) belongs to a group of peptidases that catalyze the transfer of amino acids from one peptide to another. GGT is present in all cells of the body except the muscles, and GGT in the serum originates mainly in the hepatobiliary system. GGT is increased significantly in the presence of intrahepatic or posterior biliary obstruction. It is more sensitive than alkaline phosphatase for detecting obstructive jaundice, cholangitis, and cholecystitis, as it increases earlier and persists for longer. Additionally, recent studies showed that serum GGT plays an important role in the monitoring and prognosis of metabolic syndrome and cardiovascular and cerebrovascular diseases [1][2][3][4] and is a marker of oxidative stress. 5,6 The detection of serum GGT is one of the most commonly used which shows accurate measurement results, but the operation is tedious, time-consuming, costly, and cannot be automated, making it unsuitable for routine detection of clinical patients samples, and (b) conventional measurement systems with simple operation, low-cost, automation, and wide use. However, the results of conventional measurement systems are often inconsistent; Xia Cet al 8 showed deviation was as high as 14.2% compared to IFCC reference procedure and coefficients of variation between systems as high to 11.6%. Additionally, some systems are not consistent in the level of medical decision-making. 9 Reference methods are rarely used as a basis for comparison in these studies. The purpose of this study was to compare the reference method for GGT recommended by the IFCC with the four conventional measurement systems to evaluate the accuracy and reliability of the measurement results according to

EP09-A3.
Commutable substances can be used to detect and monitor differences between the results of different laboratories in the EQA Plan or Performance Verification (PT). This requires an understanding of the commutability of the substances for EQA or PT.
To this end, we assessed the commutability of CAP samples, EQA substances from NCCL and BCCL, and enzymatic in-house materials developed by the reference laboratory of Chaoyang Hospital to obtain experimental data for the PT program. The conventional measurement system consists of calibrators, quality control products, test reagents, and instrument. There is a metrological traceability from the routine result to the manufacturer's calibrator or even to a reference material and thus SI units. Because of the importance of the commutability of calibrators, 10-12 studies conducted nationally and internationally have demonstrated that applying commutable calibrators can reduce differences in the test results between systems. 8,9 Therefore, it is necessary to understand the commutability of calibrators used in mainstream methods in China. IQC products are used to monitor the inter-day stability of the laboratory's internal IVD medical device and evaluate the precision of the method.
However, when differences are observed between IQC products and human blood samples, the coefficient of variation does not reflect the actual detection performance of the instrument, 13 which may affect the detections of the patients. Therefore, in this study, 25 fresh serum samples from patients were used to assess the commutability of the above reference materials among the four commonly IVD medical devices to provide necessary information for relevant decision-making departments.

| Reagents and instruments
According to the IFCC reference measurement for GGT, the rea-

| Ethics statement
The study involved the use of leftover patient serum samples. The leftover patient samples were all de-identified during the collection.
The use of patient samples in the present study has been reviewed by the Ethics Committee of Beijing Chaoyang Hospital. Detailed patient information was not needed, and the data were analyzed anonymously; therefore, participants did not provide written informed consent.
The calibrators and quality controls (except from CAP) are provided in the form of a freeze-dried powder and must be re-dissolved.
Before use, reagent grade laboratory water was added to a bottle based on specification requirement. Each sample was incubated for 30 minutes at 20 ~ 25°C. The bottles were mixed gently until the samples were completely dissolved.

| In-house candidate reference materials
The in-house candidate RMs for GGT were from the reference laboratory in Beijing Chaoyang hospital (3 levels). These materials were from patient leftover sera which were pooled, mixed thoroughly, sterile-filtered to 0.20 µm and aliquoted 0.8 mL sera into 1-mL cryovials and stored under −80°C until use. Three levels of frozen serum RMs were assigned values according to the IFCC reference method for GGT measurement by candidate reference laboratories. One new aliquot of each level was tested four replicates a day for three successive days. Results were expressed as "target value ± uncertainty". The homogeneity and stability of the candidate RMs were evaluated according to ISO Guide 35. 14 Ten vials of each concentration were analyzed in triplicate to determine their homogeneity, which were analyzed using one-way ANOVA. For the stability assessment, linear regression analysis was used. The candidate RMs for GGT were homogeneous and stable for at least 3 days as stored at 4°C and 12 months as stored at −80°C. The in-house prepared materials as the candidate reference materials met the characterization of the secondary reference materials. We will further apply for the secondary reference materials certification to the National Standard Substance Committee.

| Serum samples
Twenty-five low, medium, and high single fresh frozen serum samples for GGT without hemolysis, lipemia, and icteric were tested and found to be negative for HIV1 + 2 antibodies. These samples were obtained from 2 to 8°C refrigerator in the Department of Laboratory Medicine at the Beijing Chao-Yang Hospital within 48 hours. The concentrations were 8-868 U/L. Each sample had a volume of at least 5 mL after mixing evenly, and the samples were divided into five parts and stored at −80°C; the commutability may not be influenced when they were thawed and measured. 15 The single serum samples were homogeneous and stable for at least 3 days as stored at 4°C and showed stability after 3 months of storage at −20°C. 16 19 The primary reference method was performed strictly according to the IFCC publications. 20,21 N-Glycylglycine was weighed 2.73 g (206.3 mmol/L) and dissolved in 100 mL water to prepare the reaction solution by the adjustment of PH. L-γ-Glutamyl-3-carboxy-4-nitroanilide, monoammonium salt, and monohydrate were weighed 0.229 g (33.00 mmol/L) and dissolved in 20 mL water to prepare the start reagent solution.
A 2.000 mL reaction solution was added into the cuvette and equilibrated to 37°C, and then, a 0.250 mL sample as a quality control was added (equilibrated close to 37°C). After mixing thoroughly and incubating the sample for 3 minutes, the temperature of the solution in the cuvette reached 37°C. Finally, 0.500 mL start reagent solution was added, mixed thoroughly, and incubated for 60 seconds, after which absorbance was monitored for 3 minutes. Dynamic modeling software was used to analyze the results. The IFCC External Quality Assessment Scheme for Reference Laboratories in Laboratory Medicine (RELA 2014) was used, and samples A and B were used as quality controls, which were required to be within the uncertainty range of the laboratory mean of the RELA 2014 (limit of equivalence). Each sample of prepared materials was tested twice by the IFCC reference method for GGT.

| Routine measure systems
Forty-two samples were tested three times on four types of IVD

| Comparison and commutability study
According to EP9A-3, 22 25 fresh serum samples were simultaneously evaluated by the reference method and routine measurement systems. All samples were measured twice by the IFCC reference The commutability experiment for GGT was designed according to the CLSI EP-14A-3 guideline. 24 Prepared materials (calibrators, quality control materials, homemade candidate RMs) together with 25 patient samples were measured for GGT using the four routine IVD medical devices and IFCC reference method. The measurement results were analyzed in a pairwise manner, and Deming regression and 95% prediction intervals were calculated between each routine IVD medical device and IFCC reference method. The commutability of these prepared materials between the reference method and four different analyzed systems was estimated. Prepared materials with measurement results inside the prediction intervals were considered as commutable.

| Comparability of IVD medical devices for GGT
The 25 single fresh frozen serum samples were analyzed on 4 IVD medical devices and IFCC reference method for the comparison experiments. The linear relationships between X (IFCC reference method) and Y (routine methods) are shown in Figure 1 and    Table 3. Compared to the moderate allowable deviation (11.1%) derived from biological variability, 23 four measurement procedures were acceptable at the levels of 150 U/L medical decision. The 95%

Bias
confidence limits of Hitachi/BioSino, Beckman, and Roche were exceeded the 11.1% bias acceptability criteria at the 60 U/L medical decision points.   Figure 2 and Table 4).

| D ISCUSS I ON
Obtaining accurate and reliable examination results by clinical laboratories is the goal of laboratory medicine. To achieve accurate test results, testing methods must be standardized to ensure the metrological traceability of the test results. Currently, biochemical reagent manufacturers in China are increasingly considering metrological traceability. We obtained comparable information between the GGT assays in this study, which was significant for the standardization of enzyme.
The CLSI EP09-A3 is an internationally recognized guideline for comparing different methods. 22 In this study, Deming regression was used as a method to estimate slope and intercept parameters from a measurement procedure comparison experiment with allowance for both measurement procedures to have imprecision. The measurement error for each measurement procedure is taken into account in the estimation procedure using Deming regression analysis. EP14-A3 is an international guideline for evaluating matrix effects or commutability and requires at least 20 representative native patient samples as standards for comparison. 23 We used 25 native patient samples at low, medium, and high activities, and the data were processed using the commutability assess- The results also showed the Beckmann and Roche systems had reasonable performance goal in the trueness and comparability assessment (Tables 1 and 3).
The GGT in-house candidate reference materials (3 levels) developed by the human serum pools prepared from patient samples are commutable, which can be further used to certificate national secondary standard substances in the future. The reference materials are useful for external quality assessment and validation of the routine testing systems, as well as for improving the accuracy and consistency of GGT results by different manufacturers.
A major limitation of the study is that all measurements of each method were performed in a single measurement system under specific measuring conditions. The routine methods were performed in four different clinical laboratories, and the reference methods were performed by the reference laboratory. These measuring conditions were assumed to be representative. Therefore, the future measurements should be further performed with the same methods to generalize the research conclusions.

AUTH O R CO NTR I B UTI O N S
All the authors have accepted responsibility for the entire content of this submitted article and approved submission.