Evaluation of the Atellica COAG 360 coagulation analyzer in a central laboratory of a maximum care hospital

Fully‐automated coagulation analyzers are key components of a high‐throughput central laboratory. The novel Atellica COAG 360 (Siemens Healthineers) is a high‐volume coagulation analyzer approved for hemostasis diagnostics. The aim of the study was to evaluate the analytical performance of this coagulation analyzer in a central laboratory.

the extrinsic and intrinsic coagulation system, respectively.
Besides global coagulation tests, determination of specific coagulation parameters, such as single coagulation factor activities, von Willebrand (vWF) activity, or protein C, are essential for diagnosis and management of hemostatic disorders. Increasing centralization and consolidation of hemostasis diagnostics together with a growing global hemostasis market require fully-automated coagulation analyzers for high-throughput coagulation testing. 3,4 Nowadays, high-volume coagulation testing can be performed by fully-automated coagulation analyzers, which allow the fast, accurate, and reliable measurement of coagulation parameters thereby maintaining high quality. [4][5][6][7][8] However, pre-analytical issues are still challenging and have to be considered in laboratory medicine and in particular in hemostasis testing. 4  The aim of the present study was to evaluate the analytical performance and throughput capability of the Atellica COAG 360 analyzer connected to a total laboratory automation in a central laboratory on 24/7 duty.

| Study design and sample collection
The study was conducted at the Institute for Clinical Chemistry and Pathobiochemistry in the Department for Diagnostic Laboratory Medicine at the University Hospital in Tübingen. Measurement of coagulation parameters was performed using citrate-containing plasma patient samples (Sarstedt) from clinical routine (n = 74-104).
All samples were centrifuged for 10 minutes at 2500 g and anonymized using individualized identification numbers for each sample tube to ensure anonymity of patients. Plasma supernatants were transferred into at least two identical aliquots and were immediately measured. Determination of sample throughput rate was performed using freshly collected citrate-containing plasma samples from healthy volunteers (n = 45). Written informed consent was obtained from healthy volunteers prior to blood sample collection. The study was conducted in accordance with the Declaration of Helsinki from 1964 and its later amendments and approved by the local Ethics Committee of the medical faculty of Tübingen (protocol number: 113/2014BO1).

| Reagents, calibrators, and controls used on coagulation analyzers
Reagents, calibrators, and controls were used on the Atellica COAG 360 according to the instructions of the manufacturer and standard operating procedures (see Table S1). The same reagents were used on the Sysmex CS-5100 (Siemens Healthineers) and the STart Max (Diagnostica Stago SAS) coagulation analyzers. Reference ranges (5th-95th percentiles) and onboard reagents stabilities are provided according to the manufacturer for the Atellica COAG 360. A genderspecific reference range is provided for protein S (free) and a 90th percentile for D-dimer.
In total, the analyzer has a maximum load capacity of 150 samples (30 racks à five samples). Furthermore, one rack is defined as STAT sample rack with five priority positions. According to the manufacturer, the Atellica COAG 360 is able to perform 210 single tests of PT/aPTT, 350 simultaneous tests of PT/aPTT, and 310 simultaneous tests of PT/aPTT/AT/DD per hour. For assessment of sample quality, the analyzer performs a primary-tube volume check and a hemolysis, icterus, and lipemia ("HIL") check. Therefore, a 4-channel photometer with simultaneous multiwavelengths scanning (365, 415, 470, and 645 nm) is used. For each HIL-index, nine levels are defined and assay-specific thresholds are provided. Furthermore, the Atellica COAG 360 provides automated reflex, redilution, multidilution analysis and repeat testing using laboratory-specific rules.
Cooled chambers and anti-evaporation caps can be used for most of the reagents. For integration into existing laboratory infrastructures, the Atellica COAG 360 can be connected, according to the manufacturer, to various laboratory automation and online data management systems.

| Assessment of linearity and of intra-and interassay precision and accuracy
Linearity of coagulation measurements was evaluated using samples from clinical routine with high levels of fibrinogen, antithrombin, D-dimer and FXIII. Samples were manually diluted (1:2, 1:4, 1:8, 1:16; and 1:32), and measurement results were correlated with theoretical assigned concentrations. A curve was determined by linear regression analysis, and linearity was assumed as acceptable when R 2 > 0.95.
Commercially available control samples (see Table S1) in the normal and abnormal range of respective parameters were used for determination of intra-and inter-assay precision and accuracy. Intra-assay precision was calculated for all indicated parameters by repeated measurements of the respective control samples 10 times in a single batch and reported as mean ± SD and the calculated coefficient of variation (CV%). Inter-assay precision was determined by measuring control samples twice a day over 20 days and reported as mean ± SD and the calculated CV%. Intra-assay and inter-assay accuracy was calculated as mean percentage difference between measurement results and known target values of commercially available control samples.

| Pre-analytical assessment of sample quality
Centrifuged samples from clinical routine were selected according to optical properties and defined as hemolytic, icteric, or turbid (lipemic) by visual inspection. Triglyceride and total bilirubin concentrations were determined on an ADIVA Centaur XPT clinical chemistry analyzer using enzymatic methods, and plasma hemoglobin concentrations were determined on a Dimension EXL 200 system using a spectrophotometric method (both Siemens Healthineers).
INR, aPTT, fibrinogen, antithrombin, and D-dimer measurements were performed using the Atellica COAG 360 and the Sysmex CS-5100 coagulation analyzers, and results of both analyzers were compared. Furthermore, the STart Max, a coagulation analyzer based on mechanical clot detection was included in the study. The same reagents and calibrators, as provided in Table S1, were used for measurements of INR, aPTT and fibrinogen in the same hemolytic, icteric, and lipemic samples on the STart Max. Results were compared with previously obtained measurement results of the Atellica COAG 360.

| Sample throughput rate and STAT analysis
Plasma samples from healthy volunteers were used to determine sample throughput rate of the Atellica COAG 360 and the Sysmex CS-5100. In total, 45 citrated plasma samples were collected and subsequently centrifuged for 10 minutes at 2500 g.

| Statistical analysis
Results of coagulation measurements are presented as mean values with standard deviation (SD) for normally distributed data. Passing-Bablok regression and Bland-Altman analyses were performed for comparison of coagulation measurement results. 10,11 Statistical analyses were performed and figures were created using GraphPad Prism 7.03 (GraphPad Software).
TA B L E 1 Intra-and inter-assay precision and accuracy of coagulation parameters determined by the Atellica COAG 360 Parameter Intra-assay (n = 10) Inter-assay (n = 20)

| Comparison of coagulation analyzers: Atellica COAG 360 vs Sysmex CS-5100
Plasma samples from clinical routine covering the entire clinically relevant concentration ranges were used for comparison analyses between the Atellica COAG 360 and the Sysmex CS-5100. Results were reported for INR, aPTT, antithrombin, fibrinogen, D-dimer, and FXIII (see Table 2). Passing-Bablok analyses showed good correlation between measurement results of the Atellica COAG 360 and the Sysmex CS-5100 for the indicated parameters (r > 0.88; P < .0001 for all parameters, see Figure 1). Bland-Altman plots revealed high agreement between analyzers for most parameters (see Figure 2).

| Detection of interferences and investigation of possible effects on coagulation measurements
Interferences on coagulation measurements were assessed using icteric (total bilirubin concentrations between 7.  Table 3A and 3B).

| Performance of the Atellica COAG 360 in a high-throughput central laboratory
The

| D ISCUSS I ON
In the present study, the analytical performance of the Siemens Atellica COAG 360, a fully-automated high-throughput coagulation analyzer, was evaluated. Intra-and inter-assay accuracy and precision were determined and compared with the widely used Sysmex CS-5100 automated coagulation analyzer. CVs were optimal (<5%) and accuracy was acceptable (<10%) for most of the routine coagulation parameters. Only few parameters showed slightly elevated CVs in the normal and/or abnormal range. Similar results were obtained for the Sysmex CS-5100 coagulation analyzer. These values are consistent with previously published precision data for the Sysmex CS-5100. 5 Other automated coagulation analyzers, such as the ACL TOP and the Sysmex CA-7000, showed comparable analytical performance. 12,13 Overall, correlation of coagulation measurement results between analyzers was optimal and mean differences were acceptable antithrombin, and D-dimer using the Sysmex CS-5100. 5 However, plasma throughput rates are hardly comparable due to the lack of standardization for sample throughput studies and due to the laboratory automation setting in our study. The throughput rates provided by the manufacturer are usually much higher as they are performed using samples, which are directly loaded to the analyzer.
To our knowledge, this is the first study demonstrating the sample throughput rate of a coagulation analyzer connected to a track-line system of a total laboratory automation. In general, the connection of a coagulation analyzer to a total laboratory automation may a rational and necessary process for central laboratories. Total laboratory automation can improve the efficient sample processing and quality of coagulation measurements. Pre-analytical concerns regarding the transport of coagulation samples via a track-line and automated centrifugation processes should be considered but are assumed not to be relevant. In a study by Da Rin and Lippi, the loading of samples to coagulation analyzers via a track-line did not significantly affect hemostasis testing compared with directly loaded samples. 15 Furthermore, unifying different methodologies on one coagulation analyzer is a main advantage in the field of hemostasis diagnostics. So far, hemostasis systems for routine and specific coagulation tests were operated separately or were used alternately for routine and specific coagulation tests. Together with markedly increased onboard reagent stabilities, a significantly improved workflow and reduced turnaround times in the laboratory are feasible.
The Atellica COAG 360 was launched as the first analyzer combining five different assay technologies on one platform. We investigated clotting (optical detection), immunologic, and chromogenic assays. Additionally, the Atellica COAG 360 is, according to the manufacturer, intended to use optomechanical clot detection and aggregation testing. However, these technologies have not been released yet and were therefore not evaluated in this study.
From a personal view, this analyzer provides a user-friendly interface and facilitates routine maintenance. The daily maintenance required less than 10 minutes. Compared to the Sysmex CS-5100, the main advantage of the Atellica COAG 360 is the continuous access and loading of reagents and consumables at any time together with markedly increased onboard reagents stabilities (see Table S1).
This leads to a simplified handling and an improved workflow.
Taken together, the results of the evaluation demonstrate that the novel Atellica COAG 360 features optimal analytical performance as a high-throughput coagulation analyzer in a central laboratory with total automation. It combines routine and specific high-quality coagulation testing with an effective and efficient workflow.

ACK N OWLED G EM ENTS
We gratefully thank Isolde Riedlinger, Susanne Faix, and Janina Roche for excellent technical assistance.

CO N FLI C T O F I NTE R E S T
The authors state that there is no conflict of interest.