A comparative evaluation of the CN‐6000 haemostasis analyser using coagulation, amidolytic, immuno‐turbidometric and light transmission aggregometry assays

Abstract Background The CN‐6000 (Sysmex Corp.) is a new haemostasis analyser with blood coagulation, amidolytic, immuno‐turbidometric and light transmission aggregometry (LTA) capabilities. Transmitted light is monitored at multiple wavelengths (340, 405, 575, 660, 800 nm), from an LED light source. Aims To evaluate the performance of the CN‐6000 against a predicate device. Methods The CN‐6000 was evaluated against the CS‐5100 (Sysmex) for 14 different tests, using 880 samples from normal subjects, anticoagulated patients, critically ill patients, plasmas with high or low fibrinogen content or abnormal levels of interfering substances. Between‐day assay imprecision was assessed using commercial QC materials (n = 10 replicates on each of 5 days). Results Acceptable levels of imprecision were obtained for all assays. Agreement between the two analysers was excellent for all assays. Throughput was 35% higher using the CN‐6000 (337 vs 250 tests per hour for PT, aPTT and fibrinogen). The CN‐6000 also demonstrated improved clot detection in plasmas with high levels of interfering substances as demonstrated by a 29% reduction in “vote‐outs” due to low light transmission (24 vs 34). Conclusions The CN‐6000 demonstrated excellent comparability with the predicate instrument and acceptable levels of imprecision in all assays. Improvements in throughput and clot detection in the presence of interfering substances were also shown.


| ME THODS
Citrated plasma was obtained from residual, anonymized samples (collected into 0.109 mol/L sodium citrate (Vacutainer; Becton Dickinson) after all routine testing had been completed, in compliance with local ethical committee rules and the Human Tissue Act.
Samples were frozen at −80°C and thawed at 37°C immediately prior to testing. Normal citrated plasma collected locally from apparently normal healthy volunteers and commercially sourced plasmas (CRYOcheck™ Normal Donor Set; Precision BioLogic Inc) were also tested. Informed consent was obtained from normal donors (approved by the UCL Research Ethics committee: Project ID Number: 7029/001). In total, 880 samples from normal subjects, patients receiving warfarin, low molecular weight heparin (LMWH) or direct oral anticoagulants (rivaroxaban, apixaban, edoxaban and dabigatran), as well as patients with critical illness, lupus anticoagulant, haemophilia A or von Willebrand disease (vWD), and samples with low or high fibrinogen levels, high levels of haemolysis, icterus or lipaemia were tested.
Due to limited clinical sample availability, citrated whole blood from several normal donors was spiked with unfractionated heparin (5000 IU/mL Multiparin; Wockhardt UK Ltd) before plasma preparation and dilution with autologous unspiked plasma to yield a range of concentrations. As dabigatran is seldom used in the UK, external quality control samples, quality control (QC) samples and normal plasma were used to prepare plasmas with a range of dabigatran concentrations.
For the comparability study, samples were tested on the CS-5100 and CN-6000 within the same 2-hour period. The same reagents and redilution limits were used on both instruments. FVIII assays were performed in multidilution analysis mode (MDA) using low normal or high dilution ranges to ensure that the clotting times were on the linear part of the calibration curve, and the mean of the Cholesterol and triglycerides were measured using Chol2 and triglyceride reagents using a Cobas C111 analyser (Roche Diagnostics).
Plasma haemoglobin was measured using the Hemocue system (Radiometer).
Sample and reagent carryover were investigated as previously described 6 Briefly, sample carryover was assessed by performing APTT testing on aliquots of normal plasma (A1-A3) and plasma containing 1 IU/mL unfractionated heparin (B1-B3) in the sequence A1, A2, A3, B1, B2, B3 ten times, then repeating the sequence in reverse.
Reagent carryover was investigated by performing APTT testing on five aliquots of normal plasma before performing Clauss fibrinogen assays on the same five aliquots. This sequence was repeated ten times.
Statistical analysis was performed with GraphPad Prism version 8.0.0 for Windows (GraphPad Software). Deming regression analysis and Bland Altman analysis were performed to assess comparability.
The between assay imprecision results are shown in Table 2. In all cases, the %CV values for both instruments were lower than the manufacturer's specifications for within-run imprecision for a normal control plasma for the CS-5100.
The performance of the HIL checking function was assessed by A range of plasmas were used to assess comparability for FVIII, FXIII, vWF:Ac, vWF:RCo, DOACs, D-dimer and antithrombin (Table 3). Only samples giving results within the reportable range for both instruments were used for the analysis. Consequently, although large numbers of samples (180-300) were tested, some tests produced many results outside the reportable range (eg DOAC and. VWF:RCo) and regression analysis was performed on a relatively small number of results. Despite this, r 2 values of >0.9 were reported for all tests except VWF:RCo (Table 3).

| D ISCUSS I ON
The purpose of this study was to compare the performance of the CN-6000 haemostasis analyser with the established CS-5100 system for assays using clotting, amidolytic, immunoturbidimetric and light transmission aggregometry (VWF ristocetin cofactor assays) methodologies. Intra-assay imprecision for both analysers was within the manufacturer's specification for the CS-5100 and similar to previous studies of the CS-5100. [3][4][5] Agreement between the two analysers was excellent for the 14 assays studied in a wide range of samples. clinically important discrepancies in some coagulation tests. [8][9][10] No sample or reagent carryover was detected, and the throughput of the CN-6000 was 35% higher than the CS-5100 in sample representative of a normal hospital laboratory workload. This is the result of an improved high-pressure rinsing system, which reduced the time for pipette rinsing.
We did not evaluate LTA on fresh platelets in this study, as quantitative comparative evaluation of LTA is extremely difficult. There are no standards, no quality control materials and fresh platelets are labile. For this reason, VWF:RCo with freeze-dried platelets was selected to evaluate the LTA measurement feature on the CN-6000.
However, we acknowledge that formalin-fixed platelets do not behave in the same way as fresh platelets, and this is a limitation of this study.
In conclusion, the CN-6000 demonstrated excellent comparability with the predicate instrument and acceptable levels of imprecision in all assays, with improved throughput and HIL detection.