Analytical evaluation of a fully automated immunoassay for faecal calprotectin in a paediatric setting

Introduction Faecal calprotectin (FC) is a routinely used marker for identifying and monitoring children with inflammatory bowel disease (IBD). This non-invasive test is useful for screening children with gastrointestinal symptoms to avoid unnecessary invasive procedures. In this study, we validated for the first time the performance of a fully automated particle-enhanced turbidimetric immunoassay (PETIA) on the VITROS® 5600 analyzer for measurement of FC in symptomatic children and adolescents. Materials and methods For performance validation of the PETIA (fCAL® turbo, Bühlmann Laboratories, Switzerland) on the VITROS® 5600 analyzer (Ortho Clinical Diagnostics, USA) limit of quantitation (LoQ), linearity, precision data and calibration curve stability were defined. Additionally, 95 faecal samples were measured using the PETIA, an enzyme-linked immunosorbent assay (ELISA; fCAL®, Bühlmann Laboratories, Switzerland) and a semi-quantitative lateral flow assay (Quantum Blue Reader®, Bühlmann Laboratories, Switzerland) for agreement evaluation. Results The LoQ for calprotectin using PETIA on the VITROS® 5600 analyzer was 21 µg/g. The linearity range was 20 - 2100 µg/g and the precision study showed a total coefficient of variation (CV) between 2.3% and 8.9%. The calibration curve was stable for 4 weeks. Using the clinical samples quantifiable by PETIA, ELISA and the semi-quantitative lateral flow assay, Passing-Bablok regression analysis and Bland-Altman plots showed good agreement. Conclusions Due to good performance characteristics and agreement with established methods, the fully automated PETIA on the routine chemistry analyzer VITROS® 5600 is a new analytical option for the rapid determination of FC.


Introduction
Faecal calprotectin (FC) is a routinely used marker for identifying and monitoring children with inflammatory bowel disease (IBD). This non-invasive test is very useful for screening children with gastrointestinal symptoms and differentiating organic from functional intestinal disorders, to avoid unnecessary invasive procedures. The gold standard for diagnosis of IBD, endoscopic evaluation with ileocolonoscopy, can be very traumatizing, especially for children. Calprotectin, a 36 kDa protein, is predominantly located in the cytosol of neutrophil granulocytes (1). Besides, it is found in small parts in the cytosol of monocytes and macrophages, supporting the theory of having bactericidal and fungicidal properties (2). It is a marker of increased granulocyte turnover and consequently it is involved in the regulation of inflammatory reactions. Literature reported elevated calprotectin concentrations in plasma, synovial fluid, urine and faeces in inflammatory state (3). As it is upregulated in inflammatory gastrointestinal (GI) diseases, FC is used as diagnostic biomarker in the detection and observation of chronic inflammatory bowel diseases (IBD), for more than 10 years (2)(3)(4)(5).
In paediatrics, FC is commonly used in the detection of newly active IBD, in monitoring of therapy success as well as in the prediction of relapse in IBD Noebauer B. et al.
Fully automated PETIA for faecal calprotectin patients (6)(7)(8)(9). With implementation of FC as biomarker in discriminating between functional and organic BD, the number of endoscopies has successfully been reduced in paediatrics (10). In addition, it is a very resistant protein and therefore a stable parameter for laboratory testing, as collected faeces can be sent to a laboratory and stored up to 7 days at room temperature prior to analysis (11,12). In healthy adult patients, FC concentrations < 50 µg/g faeces is considered to be normal, concentrations between 50 and 100 µg/g are treated as borderline while concentrations > 100 µg/g need more precise intervention and deeper clinical workup (13). Importantly, FC concentrations have been shown to vary with age, the highest concentrations reported in new-borns (14)(15)(16)(17).
The initial procedure for measuring FC was to perform an enzyme-linked immunosorbent assay (ELI-SA). Disadvantage of this method is the long turnaround time, as a batch contains around 35-40 samples and, in reality, analysis is not started until a batch is full. For point-of-care (POC) diagnostics of single sample testing in short time (minutes), quantitative and semi-quantitative methods were established, but with the downside of a loss in accuracy compared to ELISA tests. For that reason, companies developed and launched new fully automated immunochemical methods with short testturnaround times on existing routine chemistry analysers. Quite a number of laboratories use the VITROS ® 5600 analyzer, due to its small sample size needed, which is highly advantageous in paediatric settings. Therefore, we aimed to validate, for the first time, the recently launched Bühlmann fCAL ® turbo particle-enhanced turbidimetric immunoassay (PETIA) for FC measurement on the VITROS ® 5600 analyzer.

Methods
Next to performance validation (which included limit of quantitation (LoQ), calibration curve stability, linearity and precision study) of the new PETIA (fCAL ® turbo, Bühlmann Laboratories, Switzerland) on the VITROS ® 5600 Integrated System (Ortho Clinical Diagnostics, New Jersey, USA), faecal calprotectin concentration measurements were compared to concentration values analysed by an ELI-SA test (fCAL ® , Bühlmann Laboratories, Switzerland) and a semi-quantitative lateral flow assay (Quantum Blue Reader ® , Bühlmann Laboratories, Switzerland). Therefore, FC concentrations in frozen faecal extracts were measured in duplicates by means of three different quantification methods. Diagnoses of ulcerative colitis (UC) and Crohn's disease (CD) were confirmed by ileocolonoscopy and histology.

Sample extraction and preparation
Stool samples were extracted using the CALEX ® Cap device (Bühlmann Laboratories, Switzerland) according to the manufacturer's instructions for sample preparation. Briefly, faecal samples were vortexed for 30 seconds and left for 10 minutes prior to analysis with the semi-quantitative lateral flow assay. After the initial determination with the semi-quantitative lateral flow assay for clinical purpose, faecal extracts were frozen at − 20°C. The anonymous frozen extracts, analysed for routine work, were then used for the validation of the PE-TIA on the VITROS ® 5600 Integrated system. Therefore, thawed faecal extracts were analysed in du-

Enzyme-linked Immunosorbent Assay
For performing the Bühlmann fCAL ® ELISA, manufacturer's instructions were followed as briefly summarized. After centrifugation of the prepared extract, for 10 minutes at 3000xg, the supernatant was diluted 1:50 with incubation buffer. A volume of 100 µL of the diluted supernatant was then transferred to an antibody-precoated ELISA plate and incubated for 30 minutes at room temperature. After a washing step, 100 µL of enzyme label was added prior to an additional 30 minutes incubation step. Additionally, 100 µL of substrate was pipetted to the 100 µL of enzyme label and incubated for another 15 minutes, protected from light. The optical density values were read at 450 nm on a spectrophotometer.

Validation of PETIA on the VITROS® 5600 Integrated System
Limit of quantitation Limit of quantitation was defined as the lowest actual amount of the analyte that could be reliably detected with a total error below 20%. To assess the LoQ, a sample with an assigned value of 84 µg/g was diluted (1:4, 1:5 and 1:6, respectively) with extraction buffer, supplied from the manufacturer, for confirmation. All levels ranged from 14 to 21 μg/g. Using the known dilution factor and the assigned sample value, the theoretical concentration of the diluted samples could be calculated. The diluted samples were then aliquoted, frozen at − 20°C and measured in 12 replicates during 3 days, which resulted in 36 measurements for each concentration level in total. Biases from theoretical values were assessed and total error results were defined by the root mean square (RMS) method.

Linearity
For defining the linearity range, a high FC sample was diluted with distilled water in different concentrations, which resulted in eleven levels in a range from 100% -0%. Concentrations above the highest calibrator were diluted 1:10 with distilled

Statistical analysis
A total number of 95 stool samples for PETIA and ELISA and 53 stool samples for PETIA and the semiquantitative assay were included for the agreement assessment between the different methods.
Concentrations above the highest calibrator were diluted 1:10 with distilled water. Using the mean concentrations calculated from the duplicate measurements of each sample, Passing-Bablok regression analysis and Bland-Altman plots were performed from the PETIA with the mean concentration measured by the ELISA and the semi-quantitative lateral flow assay using MedCalc statistical software version 11.6 (MedCalc, Ostend, Belgium).

Results
For the LoQ-analysis, samples in the concentration range of 14 to 21 µg/g were measured in 12 replicates during 3 days. Limit of quantitation was defined as the lowest concentration possible that showed a total error below 20%. The PETIA showed a LoQ of 21 μg/g calprotectin, on the VIT-ROS ® 5600 analyzer, with a total error of 18.98%.
For analysing the linearity of the PETIA, expected values were compared to the observed ones and recovery calculated. Results of linearity testing are listed in Figure 1 and confirmed linearity in the range from 20 -2100 µg/g, the LoQ and the highest calibrator standard used.
Calibration curve stability testing revealed that the calibration curve was stable for a period of 4 weeks. The testing covered the range of expected results, using four different calprotectin concentration levels: 50, 80, 250 and 1000 µg/g. The samples measured in week 4 showed a recovery from baseline between 98% and 106%. Analytical method agreement (Passing-Bablok regression analysis) and possible systematic bias (Bland-Altman plot) were investigated for the fully automated PETIA and the ELISA (Figure 2 and 3), and PETIA and the semi-quantitative lateral flow analyzer (Figure 4 and 5). The Passing-Bablok linear regression analysis of the PETIA on VITROS ® 5600 and ELISA (N = 95) showed   Four distinct sample concentrations and two controls were used for the assessment of different coefficients of variations (CVs). Table 1. Precision data analysis for the fully automated particle-enhanced turbidimetric immunoassay (PETIA) on VITROS ® 5600 good agreement (R = 0.99), with an intercept (95% confidence interval, CI) of 1.53 (-0.69 to 3.14) and a slope (95% CI) of 1.01 (0.99 to 1.02) (Figure 2). The FC concentrations of the faecal samples ranged from 0 to 10,692 µg/g by PETIA and from 0 to 10,237 µg/g by ELISA. In the Bland-Altman plots mean absolute and relative biases with corresponding limits of agreement were -3.8 (-327.7 to 320.1) ( Figure 3A) and 7.2% (-104.8 to 119.2) ( Figure 3B), respectively. The Passing-Bablok regression analysis for the PE-TIA and the semi-quantitative lateral flow assay (N = 53) illustrated acceptable agreement as well (R = 0.93), with an intercept (95% CI) of 0.35 (-10.03 to 7.38) and slope (95% CI) of 0.99 (0.91 to 1.08) (Figure 4). The FC concentrations of faecal samples, used for method validation, ranged from 0 µg/g to 852 µg/g for the PETIA and from 31 µg/g to 820

Discussion
For rapid analysis of FC as a screening-parameter in IBD, an accurate assay with an easy, fast and good performance is appreciated. For that reason, we validated the performance of the fully automated PETIA on the VITROS ® 5600 integrated system, which is often used in paediatric laboratories because of its requires small sample volume (19).
Overall, the immunoassay performed well in terms of LoQ, linearity, precision and calibration curve stability. These results are in line with a recent study of Nilson et al. who evaluated the fully automated PETIA fCAL ® turbo for the analysers Mindray BS-380 and Cobas c501 (20). The study showed an equal linearity range for all three methods and similar LoQs for the PETIA on the VITROS ® 5600 and the Cobas c501 with 20 µg/g (BS-380: 11 µg/g). The precision was found with a total coefficient of variation between 3% and 8% in the 50 -100 µg/g range for both analysers, respectively, and is in agreement with our results of the PETIA on the VITROS ® 5600. However, the BS-380 demonstrated stability for a period of 10 weeks, whereas the calibration curve for the fully automated PETIA on VITROS ® 5600 was stable for 4 weeks.
For method validation, we additionally compared the fully automated PETIA with the most com- and with a POC-method (Bühlmann Quantum Blue Reader ® ). The fully automated PETIA showed good agreement with the ELISA, as well as with the semi-quantitative assay. As already pointed out, the standard ELISA-based methods are generally used as reference methods for FC measurement, albeit they are labour intensive, time-consuming and have long test-turnaround times. As a result, many laboratories use the point-of-care desktop device with its lateral flow technology. Advantage is the possibility of single reliable results, obtained in less than 30 minutes, with simplicity of sample preparation and analysis. However, it is a semiquantitative assay with a limited measurement range (21). In agreement with recent literature, the fully automated FC analysis methods turn out to be the future in routine laboratory work, as they unite short turnaround times, single sample investigation, easy as well as fast handling, a high measuring range and cost-effectiveness (12,20,22).
Two recent studies showed that using different extraction devices and procedures resulted in variability between distinct assays (23,24). Hence, we compared all three methods using only one extraction device, as all three assays are compatible with the same sample preparation for analysis. In this study, the Calex ® Cap device was used for pro-tein extraction, as it can be used with any stool consistency. However, De Sloovere et al. demonstrated that the stool consistency does not affect FC results notably, when commercial extraction devices are used (22). Therefore, those devices are eminently suited in routine laboratory practice as their performance is acceptable and comparable, they are very easy to handle, they reduce handson preparation time and weighting of stool samples can be excluded.
This is the first study investigating a fully automated PETIA for FC measurement on the routine analyzer VITROS ® 5600. This procedure demonstrated an excellent overall performance and met our minimum demands in terms of quality standards. With its fine performance characteristics, its strong correlation with the reference method ELISA and its convenient handling, it is a new analytical option for the rapid determination of FC. Hence, it is the perfect tool for screening paediatric patients with gastrointestinal symptoms and monitoring children with IBD.

Acknowledgments
Calprotectin test kits were provided free of charge by the manufacturer. The author would like to thank Bühlmann Laboratories AG for this support.

Potential conflict of interest
None declared.