First report of the point-of-care TEG: A technical validation study of the TEG-6S system.

Abstract Thrombelastography (TEG) measured by the TEG5000 Hemostasis Analyzer is an established but the labor-intensive method for assessing global hemostasis. The first true point-of-care TEG, the TEG6s system, uses resonance-frequency viscoelasticity measurements and a disposable multi-channel microfluidic cartridge to assess hemostasis and response to antiplatelet therapy. TEG assays (n = 5,100) were performed on the blood of healthy volunteers (n = 157) and patients undergoing coronary revascularization at three hospitals (n = 300). The results from the TEG6s were compared with the conventional TEG5000 in accordance with Clinical and Laboratory Standards Institute (CLSI) and FDA recommendations. Precision testing was conducted using blood from healthy donors, all assays were run for 5 consecutive days in duplicate using multiple operators, lots, and instruments. Reference ranges were comparable between the TEG systems. Deming regression analysis demonstrated a strong correlation between the two systems for the standard hemostasis tests (R r = 0.932, MA r = 0.972, LY30 r = 0.938). Method comparison analysis showed an acceptable agreement between PlateletMapping (PM) assays for measuring arachidonic acid (indicator of aspirin response)- and adenosine diphosphate (indicator of P2Y12 inhibitor response)-induced platelet aggregation (total agreement = 90%, and 72%, respectively). TEG6s precision testing yielded low variability (CV 0–13%) in all measures. The new point-of-care TEG6s is associated with greater ease of use than the TEG5000 and provides precise results. The results correlated between methods for all variables. TEG6s is a promising device for near-patient hemostasis monitoring and future trials of personalized therapy designed to reduce bleeding and thrombosis.


Introduction
Hemostasis is the result of complex interactions involving diverse biochemical reactions, feedback systems, and cellular elements. Depending on the disease state and balance between antithrombotic and prothrombotic factors, the hemostasis process results in either cessation of bleedinga physiologic response, or thrombosis/ hemorrhage-pathologic responses. Disease, injury, or administration of pharmacological agents may alter hemostasis, requiring laboratory monitoring in a wide range of clinical environments [1]. In this scenario, a comprehensive evaluation of the patient's global hemostatic properties and response to anticoagulant and antiplatelet agents offers a platform for personalized therapy to improve both efficacy and safety.
Since its first documented clinical use in 1960's, thrombelastography, which measures whole blood viscoelastic properties, has been used in clinical practice to quantify coagulability, fibrinolysis, and the effects of anticoagulant and antithrombotic therapy [2]. The Thrombelastograph® (TEG® 5000, Haemonetics Inc., Braintree MA) analyzer has been used in the settings of liver transplantation, cardiovascular surgery and other surgical interventions, percutaneous coronary intervention (PCI), trauma, and coagulation disorders [3][4][5][6]. It has also been used in the fields of obstetrics, pediatrics, veterinary medicine, and basic research. The TEG5000 with PlateletMapping® (PM) assay has been utilized to personalize antiplatelet treatment in patients undergoing percutaneous coronary intervention (PCI) and to determine the timing of coronary artery bypass grafting surgery (CABG) in patients on dual antiplatelet therapy [7][8][9][10].
The TEG5000 has been marketed since 2000, but ease-of-use issues, labor, and time intensive operation have limited more widespread use in practice and in large-scale clinical trials. A novel point-of-care TEG instrument, the TEG6s system (Haemonetics, Braintree, MA under license from Coramed Technologies, Niles, IL) has been developed to mitigate these limitations. In this first report on the TEG6s we sought to: (1) determine equivalency by correlating TEG6s results with those of the TEG5000 System in patients undergoing PCI or open heart surgery; (2) characterize the precision of the TEG6s; and (3) establish reference ranges for TEG6s assays in healthy volunteers.

Study design
The study was a multicenter investigation conducted in healthy volunteers (n = 157) to determine TEG6s reference ranges and in patients undergoing cardiovascular surgery or PCI (n = 300) for method comparison and precision testing. The study was approved by the institutional review board at each of the study hospitals (Sinai Hospital, Baltimore, MD; University of Pittsburgh Medical Center, Pittsburgh, PA; and Mayo Clinic, Rochester, MN). The study design is shown in Figure 1 and was developed in accordance with Clinical and Laboratory Standards Institute (CLSI) Guidelines (EP09-A2, C28-A3) and United States Food and Drug administration (FDA) recommendations and was conducted during February 2013 to November 2014. Written informed consent was obtained prior to blood collection. Healthy subjects and patients >18 years of age were included. Healthy subjects were free of medication use for ≥14 days. Patients with a prior known coagulopathy were excluded.
Additional precision testing was conducted by Coramed Technologies in accordance with CLSI EP05-A3 Evaluation of Precision of Quantitative Measurement Procedures and FDA recommendations. Testing was conducted to establish precision characteristics for each assay parameter by lot, operator, instrument, and repeated measurement variations. Each sample was run in duplicate for 5 non-consecutive days. Each day, two operators ran 6 TEG 6s instruments in parallel and in duplicate, with two instruments running one reagent lot. A total of three reagent lots were evaluated each day (Supplement Table I). Samples from three types of donors were used: hypocoagulable (coagulation level close to the lower limit of reference range for maximum thrombin-induced platelet-fibrin clot strength ([MA] and upper limit of reference range for clotting time [R]), normal and hypercoagulable (coagulation level close to the upper limit of reference range for MA and lower limit of reference range for R) For the adenosine diphosphate (ADP) and arachidonic acid (AA)-based PM assays, samples were characterized as normal (donor with little or no platelet inhibition) or abnormal (donor with platelet inhibition levels above cut-offs for inhibition.

Blood collection
Blood samples from healthy subjects were used to develop reference ranges whereas blood samples from patients were used to correlate results of four channels tested in two TEG5000 instruments (two channels per TEG) and four channels in a single TEG6s unit (method comparison). Tests were performed in duplicate to determine the repeatability of assays using standard assay materials in both systems. Blood samples were obtained by direct venipuncture, indwelling venous sheath, or by central venous line, and transferred to Vacutainer blood collecting tubes (Becton-Dickinson, Franklin Lakes, NJ) containing 3.2% trisodium citrate (for standard hemostasis assay) or 4 ml sodium heparin 74 USP (for PM assay) after discarding the first 2-3 ml of free flowing blood. The Vacutainer tube was filled to capacity and gently inverted 5 times to ensure complete mixing of the anticoagulant. TEG6s and TEG5000 testing occurred within 2 hours of blood drawing.
In patients undergoing cardiac surgery, blood was drawn in~1/ 3 of patients in the pre-operative period, in~1/3 patients during the operation, and in~1/3 patients in the intensive care unit immediately following the operation. In patients treated with protamine, blood samples were drawn 30 minutes after protamine administration. In patients undergoing PCI, half of the blood samples were drawn before PCI and half were drawn within 2 hours following PCI (Supplement Figure 1).

Thrombelastography parameters
Thrombelastography measures the viscoelastic properties of a clot through all phases of hemostasis -from the enzymatic phase through the fibrinolytic phase and displays results numerically and as a TEG tracing. The standard TEG parameters measured are (1) clotting time (R), recorded in minutes from the start of the sample run to the point of clot formation corresponding to an amplitude of 2 mm, a quantitative representation of the initiation phase of enzymatic clotting; (2) clot kinetics (K), recorded in minutes, a measure of the time to reach 20 mm of clot strength from 2 mm; (3) clot strengthening rate (α, angle), the angle recorded in degrees, formed by the tangent to the TEG tracing measured from R, reflects the velocity of clot strength generation; (4) maximum clot strength (MA, maximum amplitude) recorded in millimeters, reflects the maximal platelet-fibrin clot strength, and (5) clot breakdown (LY 30), recorded as the % reduction in amplitude at 30 min after MA has been reached, reflective of clot lysis (Supplement Figure 2).

TEG5000 method
In the two-channel TEG5000, the time-varying physical properties of the clot are measured with a mechanism similar to a low-strain concentric cylinder oscillatory rheometer (Supplement Figure 3). Native or anticoagulated whole blood samples are combined with reconstituted dry reagents, then manually pipetted into a small cylindrical sample cup and placed into the instrument for analysis. Two samples can be run concurrently.
The sample cup holding the blood oscillates through an angle of 4°45' in 10-second cycles. A stationary pin suspended in the blood from a thin tungsten wire is monitored for motion. As fibrin and/or fibrin-platelet bonding begins to link the cup and pin together, the pin begins to move in phase with the cup. The strength and rate of formation of the clot, balanced by the torsional stiffness of the tungsten wire, determines the degree of pin rotation. As the clot retracts, rotation diminishes. A non-contact rotation sensor allows a computer to monitor the pin motion, producing graphical and numeric representation of clot strength.
The Functional Fibrinogen (FF) reagent is a blood modifier intended to measure the contribution of functional fibrinogen to MA. The FF reagent is composed of tissue factor (TF) to activate the clotting process, along with a platelet inhibitor (PI) to exclude the contribution of platelets to the MA (MA PI ) thus measuring only the functional fibrinogen contribution to clot strength (MA FF ). MA FF is measured in millimeters and is transformed to functional fibrinogen level (FLEV) in milligrams per deciliter [11].
The PM assays on heparinized blood included four reagenttreated samples: (1) kaolin, (2) reptilase + factor XIII (ActF, Activator F), (3) ADP + Activator F, and 4) AA) + Activator F. Blood sample volume was 360 ul per channel. One channel was  required for each reagent, for a total of four channels (2 TEG instruments). The standard parameters are displayed for the kaolin sample, where MA K represents maximal clot strength. MA ADP and MA AA are displayed, which reflect clot strength induced by stimulation of platelet ADP receptors and thromboxane (Tx)A 2 receptors, respectively. Corresponding calculated % aggregation and % inhibition are displayed. For calculations of percentage of clot strength reduction, percent aggregation is where MA P can be either MA ADP or MA AA . Percent inhibition is calculated as 100-percent aggregation.

TEG6s method
The TEG6s system is a fully automated diagnostic instrument (Figure 4), that uses a four-channel microfluidics cartridge. An unmetered whole blood sample (~0.4 ml) is pipetted into the entry port in the cartridge (Figure 1). The sample is metered under instrument control into four separate analysis channels. Dried reagents resident within each channel are reconstituted by movement under microfluidic valve and bellow action. After reconstitution, approximately 20 μl of prepared sample is delivered to individual test cells at the terminus of each microchannel, where clotting is monitored. Excess sample is directed into a waste chamber. The test cell is a short vertical tube open at both top and bottom ends, in which blood is supported by surface tension. The slightly convex meniscus that naturally forms at the bottom opening of the tube is positioned at the focus of an optical detection system. When the blood is excited with a multi-frequency signal by a piezoelectric actuator, the resulting meniscus motion can be optically monitored with a photodetector. A Fast Fourier Transform is performed on these frequency components to identify the resonant frequency at which the sample has had the greatest amplitude due to a coagulation event. As clotting occurs, resonant frequencies increase. These frequencies were converted to TEG equivalent units using a mapping function derived in previous in vitro studies (personal communication from Coramed). The standard hemostasis assay cartridge (Citrated Multi-Channel, CMC) uses a citrated blood sample that is mixed with dried reagents within each of the 4 channels, each with calcium chloride (to reverse the sodium citrate) and: (1) kaolin, (2) kaolin + tissue factor activated (RapidTEG), (3) kaolin + heparinase, and (4) kaolin + abciximab (functional fibrinogen) (Table III) The PM cartridge used to monitor response to oral antiplatelet agents employs a heparinized blood sample that is mixed with dried reagents within each of the 4 channels of the cartridge: (1) kaolin + heparinase, (2) Activator F + abxicimab, (3) ADP + Activator F, and (4) AA + Activator F. The same parameter results are displayed as in the TEG5000.

In-vitro studies
In order to cover the full analytical measurement range, samples from the healthy subjects (up to 11 subjects per site,~10% of all samples) were treated with either: (1) abciximab (ReoPro®, Eli Lilly and Company), at concentrations of 4 and 20 μg/ml to inhibit glycoprotein (GP)IIb/ IIIa receptor, (2) heparin, at a concentration of 1.5 USP/ml, to prolong the R, or (3) alteplase (Activase, Genentech USA, South San Francisco, CA) at concentrations of 100-200 ng/ml, to increase the % reduction in MA at 30 minutes (LY30). The results of these in-vitro treated samples were incorporated with results of the patient samples for method comparison. The assay cut-off values for platelet aggregation detection were identified using the reference ranges data from healthy individuals (and determined by the 10% probability quintile for both the ADP-and AA-induced aggregation). In addition, based on the TEG5000 and the TEG 6s cut-off values we measured the positive (PPA), negative (NPA), and total percent agreements (TPA) between the systems for ADP-and AA-induced aggregation in patient samples.

Statistical analysis
Statistical analyses were performed with R Studio (RStudio, Inc.). Descriptive statistics were reported for subjects' age, gender, race, medications, and primary diagnosis for surgery.
Reference ranges were estimated using the CLSI C28-A3c Guideline (www.clsi.org) on three reference sample groups, and constructed using the Shapiro-Wilk test to determine normality of the variable. If normality was satisfactory (p > 0.10), then the reference interval was constructed as mean ± 2 SD. If the Shapiro-Wilk test indicated non-normality (p < 0.10), the 95% reference limits were used as estimates of the reference interval. Method comparisons were carried out using Weighted Deming regressions and Bland Altman analysis. In the Deming Regression analysis, outputs of the two TEG methods are plotted against each other allowing measurement error (imprecision) in outputs from both methods. The Bland Altman analysis method calculates the mean difference between the two methods of measurement, called the bias. The limits of agreement (LOA) refer to lower and upper 95% CI of the mean difference.

Demographics
Healthy volunteers (n = 157) were representative of a broad population with respect to age, gender, and race (Supplement Table II). Among cardiac patients, 264 patients underwent open heart surgery and 36 underwent PCI (at Sinai Hospital only). Fifty-seven percent of patients were on aspirin therapy and 7% and 12% of patients were on P2Y 12 inhibitor therapy and anticoagulant therapy prior to procedure, respectively (Supplement Table III).

Reference ranges
Since Shapiro-Wilk test indicated non-normality (p < 0.10), the 95% reference limits were used as estimates of the reference interval. Reference ranges for all TEG6s tests are shown in Table I and are similar to the established TEG5000 values except LY30 in Kaolin channel, and K and LY30 in Rapid TEG assay [12,13].

TEG5000 vs. TEG 6s method comparison
Deming regression analysis demonstrated a good correlation between the two systems for the standard hemostasis tests (R r = 0.932, K r = 0.627, angle r = 0.627, MA r = 0.972, LY30 r = 0.938) and a low coefficient of variation for the TEG6s system (Figure 2A-E).The MA, R, K, and Alpha parameters had a minimal bias and good agreement, indicating that the measures are comparable between the 2 assays. In contrast, the LY30 had a substantial bias of 0.57 and wide limits of agreement ( Figure 3A-E).
PM assay agreement for ischemia risk between the systems were compared by using the TEG5000 cutoff of MA ADP ≥47 mm, as previously defined [8]. Using a non-linear regression model between the TEG5000 and the TEG 6s MA ADP results, MA ADP >56.1 mm defined the ischemia risk cutoff for the TEG 6s. Qualitative agreement between systems for ischemia risk identification was evaluated based on the cutoffs, with a sensitivity and specificity of 83% and 83%, respectively, for all samples collected (data not provided).

Method performance
Precision results are presented for each reagent and parameter for the citrated multichannel (CMC) and the PM assays as mean, SD, and coefficients of variability (CV) in Tables III and IV. Imprecision was very low within reagent lot, operator, and instrument (≤2.1%) for all tests. Within day, within run, and total precision ranged between 0.8% and 10% for tests except K for CK sample type, which was 13.0%. Imprecision was very low in the PM assay within reagent lot, operator, and instrument (≤4.1%) for all tests. Within run and total precision ranged between 0.3% and 12.9% for tests except MA AA , which was 45.8%. The latter was due to a low value for the mean (3.7 ± 1.69). For comparison, previously published values for within run precision on the TEG5000 for MA ADP and MA AA were 9% and 6%, and total precision were 8% and 7%, respectively. 15

Discussion
The current study provides the first description of a point-of-care TEG. Our study demonstrated that the TEG6s provided data that are comparable to the established TEG5000 with respect to parameters reflecting global hemostasis and platelet function, and that the precision of the TEG6s was high.
Currently, available coagulation tests are mostly plasma based and ignore the role that platelets play and their interaction with coagulation factors in hemostasis. Therefore, an effective measurement of hemostasis requires a whole blood assay to measure the net effect of interactions between cellular and soluble factors. The TEG system with its associated assays reports the net effect of plasmatic and platelet factors that contribute to hemostasis. The TEG has been Activator F + AA MA AA 51-71 †Standard hemostasis assay cartridge ‡PlateletMapping assay cartridge R = clotting time, K = clot kinetics, MA = maximum clot strength or maximum amplitude, α = clot strengthening time, LY30 = clot breakdown, MA PI = maximum amplitude in the presence of platelet inhibitor, ACT = activated clotting time, FLEV = functional fibrinogen level, ActF = activator F, ADP = adenosine diphosphate. Table II. Platelet aggregation cut-offs and agreement. most widely used to manage blood product usage in patients requiring liver transplantation, in patients undergoing open heart surgery, in trauma patients, and in non-surgical diseases [3][4][5][6][7]. The TEG5000 is not a true point-of-care device and requires the skillset of a trained technician, manual sample pipetting and mixing, frequent quality control activities, calibration, careful cleaning due to exposed blood, and leveling. These barriers have limited the widespread use of the system in multi-center clinical trials and in routine clinical practice. The new TEG6s system has been developed to overcome the limitations of the TEG5000. The TEG6s is an automated thrombelastograph that allows for continuous measurement of clot viscoelasticity in the presence of coagulation factors and inhibitors during clot initiation, formation, retraction and subsequent fibrinolysis. In contrast to the TEG5000, the TEG6s system is a stand-alone, self-contained unit that has the same physical footprint size as the TEG5000. The technical issues related to the torsion wire have been eliminated, and therefore there is no requirement to level the instrument or guard against vibratory surfaces. No computer and software installation is required, and operator training is minimal. The automated sample handling within the cartridge-based system reduces the time needed to set up and start a test from approximately 10-15 minutes on the TEG5000 to less than one minute on the TEG6s. No pipetting, adding, reconstitution or mixing of reagents is required, reducing exposure to open containers of blood, reagents, or used sample cups and pins. Overall operation is now streamlined due to simplified quality control procedures resulting in part from a self-test performed before running each sample, along with reduced preventive maintenance requirements.
In TEG6s, 4 assays are run simultaneously in a single cartridge, instead of using two TEG5000 instruments. The use of automated cartridges in the TEG6s system removes sample preparation variability, derived from operator and instrument factors, as demonstrated in the performance data. Agreement of results was demonstrated across multiple operators. The variability due to instrument and user is <1% in most instances. The TEG6s method generates reproducible results, has a short turnaround time, and can be run by non-trained personnel. The novel microfluidics cartridge-based device offers a simple, rapid, and accurate system for evaluating hemostasis status, enabling thrombelastography to be conveniently used in near patient settings and in the management of critical care.
The current validation study was performed with blood samples collected from both healthy volunteers and patients with cardiovascular disease. The results including precision testing  Note: (*) Higher CV is due to the small value of the mean, since small deviations in the mean cause exaggerated increase in CV. †N = 120 ADP = adenosine diphosphate, AA = arachidonic acid, MA = maximum amplitude, KHK = kaolin + heparinase, SD = standard deviation, %CV = percent coefficient of variance data, reference ranges and method comparison analysis data for PM assays were comparable to the conventional TEG5000. Hypercoagulability as indicated by high thrombin induced-platelet fibrin clot strength (>72 mm), and high ADP-induced-platelet fibrin clot strength measured by TEG have been shown to be associated with both short-and long-term clinical events in patients undergoing PCI [9]. It is well established that high on-treatment platelet reactivity to adenosine diphosphate (HPR) during clopidogrel therapy measured by various platelet function assays is an independent risk factor for ischemic event occurrences in post-percutaneous coronary intervention (PCI) patients [14]. In 4 addition, recent data suggest that low on-treatment platelet reactivity to ADP (LPR) is associated with a higher risk of bleeding and a therapeutic window concept has been proposed for P2Y 12 inhibitor therapy [14,15]. However, recent prospective randomized trials evaluating personalized antiplatelet therapy based on PFT in PCI populations did not demonstrate a clinical benefit, thus questioning whether treatment modification based on the results of PFT can actually influence outcomes [13]. These trials were associated with limitations including enrollment of low risk patients, and use of high-dose clopidogrel, a suboptimal strategy to overcome HPR. Most importantly, all these trials used VerifyNow P2Y 12 assay which may not be effective in identifying high risk patients [14].
We have demonstrated that clopidogrel-treated patients undergoing first time on-pump coronary artery bypass grafting had the same perioperative bleeding as clopidogrel-naive patients when their surgery was timed on the basis of a preoperative assessment of platelet reactivity by TEG [10]. In the 2012 Society of Thoracic Surgeons guidelines there is a class IIa recommendation for PFT in clopidogrel-treated patients to shorten the wait time for operation [16]. In addition, 2014 ESC/EACTS Guidelines on myocardial Revascularization states that platelet function testing should be used to guide antiplatelet therapy interruption rather than arbitrary use of a specified period of delay in patients undergoing CABG surgery (Class IIa recommendation) [17].
Inclusion of a new point-of-care TEG6s system with PM assay in future personalized antiplatelet therapy trials may assist in identifying high risk patients for modified therapy to reduce the risk of both ischemic and bleeding event occurrences. In addition, the ease of use of the TEG6s system may facilitate future investigations of precision medicine. Finally, since clinical outcomes were not included in the current study, future studies will be required to establish the TEG 6s as a valid tool to detect etiologies of bleeding and thrombosis.
In conclusion, the new point-of-care TEG6s is associated with greater ease of use as compared to the TEG5000 and has high precision. Results correlated well between the two systems and there was acceptable agreement for all tests. The TEG6s is a promising new point-of-care device for measuring hemostasis in general practice and in future trials of personalized therapy designed to reduce bleeding and thrombotic risk.