- Email: [email protected]
Evaluation of TEG® and RoTEM® inter-changeability in trauma patients
G Model
JINJ-5208; No. of Pages 6 Injury, Int. J. Care Injured xxx (2012) xxx–xxx
Contents lists available at SciVerse ScienceDirect
Injury journal homepage: www.elsevier.com/locate/injury
Evaluation of TEG1 and RoTEM1 inter-changeability in trauma patients Jostein S. Hagemo a,b,*, Paal A. Næss c, Pa¨r Johansson d, Nis A. Windeløv e, Mitchell Jay Cohen f, Jo Røislien a,g, Karim Brohi h, Hans Erik Heier i,j, Morten Hestnes k, Christine Gaarder c a
Department of Research, Norwegian Air Ambulance, Drøbak, Norway Department of Anaesthesiology, Oslo University Hospital, Oslo, Norway c Department of Traumatology, Oslo University Hospital, Oslo, Norway d Section for Transfusion Medicine, Regional Blood Bank, Rigshopitalet, University of Copenhagen, Copenhagen, Denmark e Department of Anaesthesia, Centre of Head and Orthopaedics, Rigshospitalet, University of Copenhagen, Copenhagen, Denmark f Department of Surgery, San Francisco General Hospital, University of California San Francisco, San Francisco, CA, USA g Department of Biostatistics, Institute of Basic Medical Sciences, University of Oslo, Oslo, Norway h The Royal London Hospital, Whitechapel, London, UK i Department of Immunology and Transfusion Medicine, Oslo University Hospital, Oslo, Norway j Institute of Clinical Medicine, Faculty of Medicine, University of Oslo, Norway k The Oslo University Hospital Trauma Registry, Department of Research and Development, Division of Emergencies and Critical Care, Oslo University Hospital, Norway b
A R T I C L E I N F O
A B S T R A C T
Article history: Accepted 17 November 2012
Background: Massive haemorrhage is a leading cause of preventable deaths in trauma. Traumatic coagulopathy is frequently present early after trauma, and is associated with increased mortality. A number of recent trials suggest that viscoelastic haemostatic assays (VHA), such as thromboelastography and thromboelastometry, are useful tools in guiding transfusion. Treatment algorithms exist for the use of VHAs but are not validated in traumatic haemorrhage. In this study we examined the interchangeability of two commonly used VHAs, TEG1 and RoTEM1. Methods: A total of 184 trauma patients over the age of 18, requiring full trauma team activation, were included at three different hospitals in three different countries (Copenhagen, Denmark, San Francisco, CA, USA and Oslo, Norway). Blood samples were drawn immediately upon arrival, and TEG1 and RoTEM1 analyzed simultaneously. Correlations were calculated using. Spearman’s rank correlation coefficient. Agreement was evaluated by Bland–Altman plots and calculation of limits of agreement. Results: The mean ISS in the total population was 17, and the mortality was 16.5%. Mean base excess was 2.8 (SD: 4.2). The correlation coefficient for corresponding values for the two devices was 0.24 for the Rtime vs CT in all centres combined. For the K-time vs CFT the correlation was 0.48, for the a-angleTEG vs a-angleRoTEM 0.44, and for MA vs MCF 0.76. Limits of agreement exceeded the preset clinically acceptable deviation of 10% for all variables in all centres except for MA/MCF in one centre (Copenhagen). Generally, correlation coefficients were lower and agreement poorer in the one centre (Oslo) where measurements were performed bedside by clinicians. Conclusion: Inter-changeability between TEG1 and RoTEM1 is limited in the trauma setting. Agreement seems poorer when clinicians operate the devices. Development and validation of separate treatment algorithms for the two devices is required. ß 2012 Published by Elsevier Ltd.
Keywords: Haemorrhage Trauma Coagulopathy TEG RoTEM Inter-changeability
Introduction Massive haemorrhage accounts for up to 40% of trauma-related deaths in patients reaching hospital, and is considered to be the leading cause of preventable deaths.1,2 Traumatic coagulopathy is
* Corresponding author at: Department of Anaesthesiology, Oslo University Hospital, Ulleval, Kirkeveien 166, 0407 Oslo, Norway. Mobile: +47 975 45 667. E-mail address: [email protected] (J.S. Hagemo).
detectable in 25–34% of patients on admission, and is highly predictive of poor outcome.3,4 A number of recent studies suggest that viscoelastic haemostatic assays (VHAs), such as thromboelastography and thromboelastometry, are useful tools in guiding transfusions and pharmacological coagulation support in trauma patients.5–9 The principle behind VHAs is based on the tendency of blood to increase its viscosity and elasticity through the process of coagulation. The clot that eventually forms subsequently dissolves during fibrinolysis. The clot changes over time are visualized as an
0020–1383/$ – see front matter ß 2012 Published by Elsevier Ltd. http://dx.doi.org/10.1016/j.injury.2012.11.016
Please cite this article in press as: Hagemo JS, et al. Evaluation of TEG1 and RoTEM1 inter-changeability in trauma patients. Injury (2012), http://dx.doi.org/10.1016/j.injury.2012.11.016
G Model
JINJ-5208; No. of Pages 6 2
J.S. Hagemo et al. / Injury, Int. J. Care Injured xxx (2012) xxx–xxx
evolving trace where the time to initiation of clot formation, the rate of clot formation and the maximum strength of the clot are among the most commonly utilized variables.5 VHA has advantages over conventional plasma based coagulations tests in some aspects. The overall coagulation picture expressed by VHA may be more clinically relevant than conventional assays, as it reflects the coagulation process in whole blood, rather than in fragments of the coagulation system.5,10 In addition, VHAs may be performed bedside, as point-of-care (PoC) measurements, reducing the time delay associated with conventional laboratory assays. Two frequently used VHA devices are TEG1 and RoTEM1. Both are marketed as PoC devices that give reliable and comparable results when operated in near patient environment by clinicians. However, they differ markedly in terms of function and user interface, and it is not clear whether this affects their performance in a clinical setting.11 Only a few studies have previously compared the interchangeability of TEG1 and RoTEM1.12–14 They conclude that the assays to a certain extent may be inter-changeable. However, these studies have exclusively been performed within cardiac surgery and liver transplantation procedures, and not in a trauma population. There are indications that the mechanisms behind coagulopathy in trauma differ significantly from coagulopathy in other massive bleeding scenarios.15,16 This may have implications for VHA, and consequently the inter-changeability of TEG1 and RoTEM1. To the best our knowledge, no comparison between the two methods has been performed in a trauma patient population. The aim of the current study was to compare the results of the initial TEG1 and RoTEM1 analyses in a cohort of trauma patients, and to assess the inter-changeability of the two devices. Methods Study design and patient selection We performed a multi-centre observational study in three major trauma centres; San Francisco General Hospital, CA, USA; Rigshospitalet, Copenhagen, Denmark; Oslo University Hospital, Ulleva˚l, Norway. In San Francisco General Hospital and Rigshospitalet, the samples were transported to a laboratory facility and analyzed by laboratory-trained personnel. In Oslo University Hospital the tests were run as PoC measurements by a group of trained clinicians. Patients more than 18 years old and who required full trauma team activation were eligible. Patients were excluded if the time from injury to admission was more than 2 h, if they had received more than 2000 ml of fluids pre-hospitally, if they were pregnant, had known liver failure or bleeding disorders, or were on anticoagulant medications other than acetyl salicylic acid. Data describing demographics, injury severity, admission physiology and outcome were retrieved from the institutional trauma registries. Ethics approval was obtained in accordance with local regulations for each centre. Sampling and VHA methodology In accordance with previously published studies, we chose the TEG1 5000 Hemostasis Analyzer System (Haemonetics Corp., MA, USA) with kaolin as an activator, and RoTEM1 (Tem International GmbH, Munich, Germany) with tissue factor as the activating agent (ExTEM1). Blood was collected from patients upon arrival in the trauma room in a citrated tube, by puncture of the femoral or radial artery. For TEG1, 1000 ml of blood was pipetted and blended in the kaolin containing test tube. After 1 min 340 ml was extracted with a manually operated pipette, and deposited in the designated
plastic cup with 20 ml of CaCl2. The cup was then elevated into test position, and the measurements initiated within 20 s. For RoTEM1, the automated pipette was used to extract 20 ml of CaCl2 (StarTEM1 reagent). An air cushion was applied in the tip before extraction of 20 ml of the ExTEM1 reagent, and finally mixing with 300 ml of blood in the cup and measurement initiated within 30 s. The TEG1 and RoTEM1 assays were run simultaneously in all three centres, and with the temperature set at 37.0 8C. We compared four widely used variables from the VHA trace: the reaction time from initiation of the assay to the first detectable coagulation, denoted R-time for TEG1 and clotting time (CT) for RoTEM1; the time from start of coagulation to clot amplitude of 20 mm, called the K-time in TEG1 and clot formation time (CFT) in RoTEM1; the angle by which the clot strength increases, called the a-angle for both devices; lastly, the maximum amplitude (MA) for TEG1, which corresponds to the maximum clot firmness (MCF) for the RoTEM1 device. Statistical analysis Data are presented as mean and standard deviation (SD) or number (%) unless stated otherwise. For comparison of baseline data between the three centres, one-way ANOVA was used for continuous variables, and Chi Square-test for categorical variables. Injury severity score (ISS) was regarded as a continuous variable. Correlation between TEG1 and RoTEM1 measurements was calculated using Spearman non-parametric correlation. For evaluation of agreement between TEG1 and ROTEM1 measurements, we applied Bland–Altman difference-mean plot, i.e. a plot of the methods differences of measurements (D) against the corresponding average (A), and estimation of corresponding limits of agreement (LoA), calculated as 1.96 SD from the mean difference of measurements.17 As TEG1 and ROTEM1 are technically different measurements of the same concept, a significant linear relationship in the Bland– Altman plot was expected. A log transformation is often recommended17 but was unsuccessful in compensating for this relationship in our data. Instead, a generalized version of LoA using linear regression, as suggested in a revised version of the Bland– Altman methodology,18 was used. In the Bland–Altman plots where there was a significant non-zero linear association between D and A, the linear relationship D = aA + b was estimated using univariate linear regression, the estimated linear association extracted from each of the actual differences D, and subsequently calculating the SD and corresponding LoA on these transformed data D* = D (aA + b). A predefined set of clinically acceptable LoA was established based on clinical experience and previous publications.12 These limits were based on the assumption that a 10% deviation in corresponding variables is acceptable for mean values. Clinically acceptable LoA by these conditions were 20.8 s for R/CT, 10.9 s for K/CFT, 6.78 for the a-angle and 6.2 mm for the MA/MCF. A p-value <0.05 was considered statistically significant. All calculations were made using SPSS 19.0 (SPSS Inc., Chicago, IL, USA). Results A total of 184 patients were included in the study. Patient characteristics are shown in Table 1. No statistically significant differences between centres were found except for ISS, which was significantly higher in the San Francisco population compared to both Oslo and Copenhagen (p = 0.031). Scatter plots of the corresponding TEG1 and RoTEM1 values for the four predefined variables are shown in Fig. 1, and corresponding
Please cite this article in press as: Hagemo JS, et al. Evaluation of TEG1 and RoTEM1 inter-changeability in trauma patients. Injury (2012), http://dx.doi.org/10.1016/j.injury.2012.11.016
G Model
JINJ-5208; No. of Pages 6 J.S. Hagemo et al. / Injury, Int. J. Care Injured xxx (2012) xxx–xxx
3
Table 1 Descriptive statistics of trauma patients included in the study. Location Copenhagen
Oslo Age (years) Sex (% male)* MOI (% Blunt)* ISS§ Base excess R-time K-time a-angleTEG MA CT CFT a-angleRoTEM MCF Mortality (%)*
47.4 (20.0) 60.4 90.6 17 (5, 29) 3.3 (5.0) 427.7 (190.0) 138.9 (73.2) 59.4 (12.3) 62.0 (9.0) 68.6 (17.1) 104.1 (39.9) 69.5 (6.9) 61.1 (10.0) 15.1
n = 53 n = 53 n = 53 n = 50 n = 46 n = 66 n = 66 n = 66 n = 66 n = 66 n = 65 n = 66 n = 66 n = 53
48.4 (20.0) 57.0 91.1 17 (8, 26) 2.3 (3.5) 343.1 (69.4) 111.0 (32.3) 65.1 (6.3) 62.8 (6.1) 63.0 (13.0) 106.0 (30.6) 69.3 (5.2) 61.2 (6.1) 19.0
San Francisco n = 79 n = 79 n = 79 n = 78 n = 74 n = 79 n = 79 n = 79 n = 79 n = 79 n = 79 n = 79 n = 79 n = 79
44.5 (18.2) 71.1 71.8 26 (16, 36) 3.7 (4.2) 237.6 (83.8) 93.4 (71.3) 69.3 (9.0) 64.0 (7.0) 59.4 (12.2) 82.0 (30.4) 74.2 (5.6) 63.5 (6.1) 13.2
Total n = 39 n = 38 n = 39 n = 35 n = 13 n = 38 n = 39 n = 39 n = 39 n = 38 n = 39 n = 39 n = 39 n = 38
47.2 (19.5) 61.2 86.5 17 (7.5, 26.5) 2.8 (4.2) 351.7 (145.7) 117.3 (60.9) 63.9 (10.0) 62.8 (7.5) 64.3 (14.8) 100.2 (35.3) 70.4 (6.2) 61.7 (7.7) 16.5
n = 171 n = 170 n = 171 n = 163 n = 133 n = 183 n = 184 n = 184 n = 184 n = 183 n = 183 n = 184 n = 184 n = 170
Values given as mean (SD), except * given as number (%), and § given as median (inter quartile range). MOI: mechanism of injury; ISS: injury severity score. Mortality is given at 30 days. The TEG parameter R-time (reaction time) corresponds to RoTEM CT (clotting time) and is measured in seconds. K-time (kinetic time) corresponds to CFT (clot formation time), also measured in seconds. a-angleTEG corresponds to a-angleRoTEM and is measured in degrees (8). MA (maximum amplitude) corresponds to MCF (maximum clot formation) and is measured in millimetres. Base excess is measured in mequiv./L.
Fig. 1. Scatter plots showing the relation between the corresponding TEG1 and RoTEM1 measurements in each centre. (a) TEG reaction time (R) vs RoTEM clotting time (CT) in seconds, (b) TEG kinetic-time (K) vs RoTEM clot formation time (CFT) in seconds, (c) TEG a-angle vs RoTEM a-angle in degrees and (d) TEG maximum amplitude (MA) vs RoTEM maximum clot formation (MCF) in millimetres.
Please cite this article in press as: Hagemo JS, et al. Evaluation of TEG1 and RoTEM1 inter-changeability in trauma patients. Injury (2012), http://dx.doi.org/10.1016/j.injury.2012.11.016
G Model
JINJ-5208; No. of Pages 6 J.S. Hagemo et al. / Injury, Int. J. Care Injured xxx (2012) xxx–xxx
4
Table 2 Spearman correlation coefficients for corresponding TEG1 and RoTEM1 values, in each of three centres and total. Oslo (n = 65)
R-time vs CT K-time vs CFT a-angleTEG vs RoTEM MA vs MCF
Copenhagen (n = 79)
San Francisco (n = 38)
All centres (n = 182)
Correlation
p-Value
Correlation
p-Value
Correlation
p-Value
Correlation
p-Value
0.15 0.21 0.17 0.67
0.231 0.093 0.185 0.001
0.11 0.56 0.45 0.84
0.33 <0.001 <0.001 <0.001
0.06 0.57 0.62 0.86
0.72 <0.001 <0.001 <0.001
0.24 0.48 0.44 0.76
0.001 <0.001 <0.001 <0.001
The TEG parameters: R-time (reaction time), K-time (kinetic time), a-angleTEG and MA (maximum amplitude) corresponds to the RoTEM parameters: CT (clotting time), CFT (clot formation time), a-angleRoTEM, and MCF (maximum clot formation), respectively.
correlation coefficients in Table 2. No statistically significant correlation was found for R/CT-time for any centre individually. Only the correlation between MA and MCF was statistically significant within all centres. When data from all centres were combined, a statistically significant correlation was found for all four TEG1/RoTEM1 variables, ranging from 0.21 for R/CT-time to 0.76 for MA/MCF. The Bland–Altman plots for TEG1 and RoTEM1 measurements are shown in Fig. 2. There was a significant linear relationship between difference and average for paired measurements of R/CT, K/CFT and a-angle, but not for MA/MCF. LoAs are given in Table 3. Neither of the LoA, evaluated within each centre or overall, were within the predefined clinically acceptable range for R/CT, K/CFT or a-angleTEG/a-angleRoTEM. For MA/MCF, only measurements from
Copenhagen were within the predefined clinically acceptably limits. Discussion The main finding of this study is that the agreement between kaolin activated TEG1 assay and tissue factor activated RoTEM1 assay is limited in a clinical setting of trauma patients. We found that correlation was highly variable in the different stages of the clotting process, as well as between the different centres. The best correlation was found for MA/MCF with a correlation coefficient of 0.84 and 0.86 in Copenhagen and San Francisco, respectively. In the Oslo centre, where clinicians ran the analyses as PoC, only the MA/ MCF measurements correlated significantly, and with a markedly
Fig. 2. Bland–Altman difference-mean plot for corresponding TEG1 and RoTEM1 values. (a) TEG R-time vs RoTEM CT, (b) TEG K-time vs RoTEM CFT, (c) TEG a-angle vs RoTEM a-angle and (d) TEG MA vs RoTEM MCF. Regression lines for mean bias (middle line), and limits of agreement are indicated.
Please cite this article in press as: Hagemo JS, et al. Evaluation of TEG1 and RoTEM1 inter-changeability in trauma patients. Injury (2012), http://dx.doi.org/10.1016/j.injury.2012.11.016
G Model
JINJ-5208; No. of Pages 6 J.S. Hagemo et al. / Injury, Int. J. Care Injured xxx (2012) xxx–xxx
5
Table 3 Mean bias and limits of agreement for corresponding TEG1 and RoTEM1 measurements. Copenhagen
Oslo D a R-time vs CT K-time vs CFT a-angleTEG vs RoTEM MA vs MCF
1.94 0.82 0.90 0
LoA b 122.9 64.2 67.9 0.9
D a
67.6 111.5 21.8 13.0
1.78 0 0.24 0
San Francisco LoA
b 82.0 5.0 20.2 1.6
D a
49.4 48.4 10.7 5.7
1.9 1.13 0.65 0
Total LoA
b 107.4 87.7 51.1 0.5
D a
48.2 95.7 14.8 8.0
1.89 0.72 0.67 0
LoA b 106.5 61.2 51.6 1.13
57.1 92.9 17.1 9.4
The mean bias (D) is expressed as aA + b, where a is the linear regression derived coefficient, A is the average of two corresponding measurements and b is a constant. LoA: limits of agreement. R-time: reaction time, K-time: kinetic time, MA: maximum amplitude, CT: clotting time, CFT: clot formation time, a-angleRoTEM: MCF (maximum clot formation).
weaker correlation coefficient of 0.67. Bland–Altman analyses showed poor agreement between R/CT, K/CFT and the a-angles in all centres, with wider LoA than what was considered clinically acceptable. MA/MCF had the best agreement over all and in each centre individually, but only the measurements in Copenhagen showed acceptable LoA. In accordance with the weaker correlations, the LoA in Oslo was also broader than in the other two centres. A clinical comparison of TEG1 and RoTEM1 assays was examined by Venema et al. in a group of 46 patients scheduled for cardiac surgery.12 Their findings indicate that the kaolin activated TEG1 assay and the tissue factor activated RoTEM1 assay (ExTEM1) were inter-changeable for the a-angle and MA/MCF, but not for R/CT and K/CFT. The lack of interchangeability of R/CT and K/CFT is in accordance with our findings. In the same study, an evaluation of repeatability of measurements from two healthy volunteers was also performed. Their analyses showed variability for the R/CT and K/CFT variables of more than 10%, and consequently considered inappropriate for clinical purposes. The inter-changeability of the a-angle and MA/MCF reported in the study by Venema et al. is, however, partly in contrast to our findings. This may be due to at least two factors. First, our study population of trauma patients is expected to contain subjects with a significant degree of coagulopathy. The activating agent in our TEG1 measurements, kaolin, activates coagulation by directly binding to and activating factor XII, subsequently activating factors XI, IX and X. The last step however, depends on factor VIII as a cofactor.19 In trauma patients, studies have shown increased levels of activated protein C due to hypoperfusion and tissue injury.20 Activated protein C is a potent inhibitor of factor VIII.21 Hence, the rate of clot formation may be decreased when activated by kaolin compared to tissue factor in this group of patients, and subsequently increasing the discrepancy from tissue factor activated measurements. A direct comparison of TEG1 and RoTEM1 was performed in a porcine model by Tomori et al.22 In their study, blood was collected in three different stages of coagulopathy in combination with trauma. No activating agent was added to the blood sample. They found correlation coefficients higher than in our study, although the correlation for the a-angle was not statistically significant. Analyses of agreement were not performed in their study. Their findings support the fact that the activating agent may have implications for the correlation of the measurements. In a recent paper by Solomon et al.14, TEG and RoTEM assays with platelet inhibition by abciximab and cytochalasin D were evaluated. In their study both reagents were used on both machines, and the samples were subjected to hemostatically active medications, as well as different degrees dilution. Their most notable findings are that the MCF and MA differ depending on both reagent and device used. Moreover, the effect of high dose heparin affected the MA in the TEG assay when abciximab was used,
whereas the cytochalasin D in the RoTEM assay remained unchanged. The authors conclude that target thresholds for fibrinogen substitution should be reviewed for the each device separately. One previous study has shown that kaolin activated TEG1 and tissue factor activated RoTEM1 show better agreement than several alternative assays. We therefore found it reasonable to use kaolin and tissue factor for TEG1 and RoTEM1 respectively in our study.12 Regardless of whether kaolin or tissue factor is used, the addition of an activating agent may be a crucial part of the test procedure in TEG1 and RoTEM1. Only small errors of pipetting may affect the clot initiation time and clot formation rate. The discrepancies found for one of the centres in our study might be due to staff less experienced with laboratory techniques. This finding is nevertheless of clinical importance, as it indicates that agreement between the two VHA devices is subject to variations in operator skills. Other possible explanations for the wide LoA in our study could be the time factor from sampling to initiation of the analyses. The time from sampling to analyses differed between centres. In San Francisco and Oslo, the measurements were performed almost immediately after sampling, whereas in Copenhagen they were performed after 1 h. Previous studies examining the significance of time delay indicate that R-time may be reduced, and the a-angle increased by approximately 10% over a 60-min period after collection in a citrated tube. The MA value, on the other hand, was not affected.23 We find it unlikely that this would inflict a systematic bias to our results, since changes of comparable magnitude also occur when tissue factor was used as an activator, and corresponding TEG1 and RoTEM1 analyses in our study were performed simultaneously. We defined LoA constituting less than 10% of mean values as being clinically acceptable. This limit might be too strict. According to previous studies on method comparisons some authors argue that a lesser extent of agreement is tolerable.24 LoA for MA/MCF in the total population indicate that a deviation of 15.1% of the mean values would have to be accepted. Correspondingly, the acceptable deviation for R/CT would have to be 27.5%, for K/ CFT 85.4% and for a-angleTEG/a-angleRoTEM 25.4%. This study has some limitations. The difference between the numbers of patients from each centre affected the level of significance for the correlations. A more homogenous group of laboratory technicians might have given better agreement between measurements overall, but an evaluation of the practical use of VHA as PoC devices was considered to be one important aspect of this study. Transfusion algorithms based on TEG1 and RoTEM1 do exist, but there are no studies validating their use with regard to patient outcome.25 This study was not powered to determine whether the overall performance in detecting coagulation perturbations or predicting outcome differed between the two devices. The discrepancies found in our data, however, warrant further research to validate the two devices in the trauma setting.
Please cite this article in press as: Hagemo JS, et al. Evaluation of TEG1 and RoTEM1 inter-changeability in trauma patients. Injury (2012), http://dx.doi.org/10.1016/j.injury.2012.11.016
G Model
JINJ-5208; No. of Pages 6 J.S. Hagemo et al. / Injury, Int. J. Care Injured xxx (2012) xxx–xxx
6
TEG1 and RoTEM1 have become widely used in the trauma setting over the past few years. Their proposed advantages in being PoC devices have been a driving force. This study indicate that the inter-changeability between kaolin activated TEG1 and tissue factor activated RoTEM1 measurements is limited in the trauma setting. Agreement is worse when the devices are used as PoC devices by clinicians in our study. Development and validation of separate trauma treatment algorithms for TEG1 and RoTEM1 is required. Conflict of interest statement The research group have received support from both Haemonetics and TEM international for this study in the form of reagents for the analyses, and leasing of devices at reduced prices. KB has also performed consultancy work for Haemonetics. Acknowledgements All of the author institutes are affiliated members of the International Trauma Research Network (www.INTRN.org) and as such this work represents a combined output resulting from this international partnership. The authors would like to thank The Oslo University Hospital Trauma Registry for contributing with descriptive data in the Oslo population of the study. References 1. Sauaia A, Moore FA, Moore EE, Moser KS, Brennan R, Read RA, et al. Epidemiology of trauma deaths: a reassessment. Journal of Trauma 1995;38(February (2)):185–93. 2. Gruen RL, Jurkovich GJ, McIntyre LK, Foy HM, Maier RV. Patterns of errors contributing to trauma mortality: lessons learned from 2,594 deaths. Annals of Surgery 2006;244(September (3)):371–80. 3. Brohi K, Singh J, Heron M, Coats T. Acute traumatic coagulopathy. Journal of Trauma 2003;54(June (6)):1127–30. 4. MacLeod JBA, Lynn M, McKenney MG, Cohn SM, Murtha M. Early coagulopathy predicts mortality in trauma. Journal of Trauma 2003;55(July (1)):39–44. 5. Fries D, Innerhofer P, Schobersberger W. Time for changing coagulation management in trauma-related massive bleeding. Current Opinion in Anesthesiology 2009;22(April (2)):267–74. 6. Johansson PI, Stissing T, Bochsen L, Ostrowski SR. Thrombelastography and tromboelastometry in assessing coagulopathy in trauma. Scandinavian Journal of Trauma Resuscitation and Emergency Medicine 2009;17(Januar (1)):45. 7. Scho¨chl H, Nienaber U, Hofer G, Voelckel W, Jambor C, Scharbert G. Goaldirected coagulation management of major trauma patients using thromboelastometry (ROTEM)-guided administration of fibrinogen concentrate and prothrombin complex concentrate. Critical Care 2010;14(Januar (2)):R55.
8. Davenport R, Manson J, De’ath H, Platton S, Coates A, Allard S, et al. Functional definition and characterization of acute traumatic coagulopathy. Critical Care Medicine 2011;July. 9. Kashuk JL, Moore EE, Sawyer M, Le T, Johnson J, Biffl WL, et al. Postinjury coagulopathy management: goal directed resuscitation via POC thrombelastography. Annals of Surgery 2010;251(April (4)):604–14. 10. Johansson PI, Ostrowski SR, Secher NH. Management of major blood loss: an update. Acta Anaesthesiologica Scandinavica 2010;July. 11. Jackson GNB, Ashpole KJ, Yentis SM. The TEG1 vs the ROTEM1 thromboelastography/thromboelastometry systems. Anaesthesia 2009;64(Februar (2)): 212–5. 12. Venema LF, Post WJ, Hendriks HGD, Huet RCG, de Wolf JTW, De Vries AJ. An assessment of clinical interchangeability of TEG and RoTEM thromboelastographic variables in cardiac surgical patients. Anesthesia and Analgesia 2010;111(August (2)):339–44. 13. Coakley M, Reddy K, Mackie I, Mallett S. Transfusion triggers in orthotopic liver transplantation: a comparison of the thromboelastometry analyzer, the thromboelastogram, and conventional coagulation tests. Journal of Cardiothoracic and Vascular Anesthesia 2006;20(August (4)):548–53. 14. Solomon C, Sorensen B, Hochleitner G, Kashuk J, Ranucci M, Schochl H. Comparison of whole blood fibrin-based clot tests in thrombelastography and thromboelastometry. Anesthesia and Analgesia 2012;114(April (4)):721–30. 15. Cohen MJ, Call M, Nelson M, Calfee CS, Esmon CT, Brohi K, et al. Critical role of activated protein c in early coagulopathy and later organ failure infection and death in trauma patients. Annals of Surgery 2011;December. 16. Brohi K, Cohen MJ, Ganter MT, Matthay MA, Mackersie RC, Pittet JF. Acute traumatic coagulopathy: initiated by hypoperfusion: modulated through the protein C pathway? Annals of Surgery 2007;245(May (5)):812–8. 17. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986;1(Februar (8476)):307–10. 18. Bland JM, Altman DG. Measuring agreement in method comparison studies. Statistical Methods in Medical Research 1999;September (8):135–60. 19. Thorsen K, Ringdal KG, Strand K, Soreide E, Hagemo J, Soreide K. Clinical and cellular effects of hypothermia, acidosis and coagulopathy in major injury. British Journal of Surgery 2011;98(July (7)):894–907. 20. Cohen MJ, Call M, Nelson M, Calfee CS, Esmon CT, Brohi K, et al. Critical role of activated protein c in early coagulopathy and later organ failure. Infection and Death in Trauma Patients Annals of Surgery 2012;255(February (2)):379–85. 21. Esmon CT. The protein C pathway. Chest 2003;124(September (3 Suppl.)):26S– 32S. 22. Tomori T, Hupalo D, Teranishi K, Teranishi K, Michaud S, Hammett M, et al. Evaluation of coagulation stages of hemorrhaged swine: comparison of thromboelastography and rotational elastometry. Blood Coagulation and Fibrinolysis 2010;21(January (1)):20–7. 23. Johansson PI, Bochsen L, Andersen S, Viuff D. Investigation of the effect of kaolin and tissue-factor-activated citrated whole blood, on clot-forming variables, as evaluated by thromboelastography. Transfusion 2008;48(November (11)): 2377–83. 24. Critchley LA, Critchley JA. A meta-analysis of studies using bias and precision statistics to compare cardiac output measurement techniques. Journal of Clinical Monitoring and Computing 1999;15(February (2)):85–91. 25. Afshari A, Wikkelso A, Brok J, Moller AM, Wetterslev J. Thromboelastography (TEG) or thromboelastometry (ROTEM) to monitor haemotherapy versus usual care in patients with massive transfusion. Cochrane Database of Systematic Reviews 2011;(3):CD007871.
Please cite this article in press as: Hagemo JS, et al. Evaluation of TEG1 and RoTEM1 inter-changeability in trauma patients. Injury (2012), http://dx.doi.org/10.1016/j.injury.2012.11.016
JINJ-5208; No. of Pages 6 Injury, Int. J. Care Injured xxx (2012) xxx–xxx
Contents lists available at SciVerse ScienceDirect
Injury journal homepage: www.elsevier.com/locate/injury
Evaluation of TEG1 and RoTEM1 inter-changeability in trauma patients Jostein S. Hagemo a,b,*, Paal A. Næss c, Pa¨r Johansson d, Nis A. Windeløv e, Mitchell Jay Cohen f, Jo Røislien a,g, Karim Brohi h, Hans Erik Heier i,j, Morten Hestnes k, Christine Gaarder c a
Department of Research, Norwegian Air Ambulance, Drøbak, Norway Department of Anaesthesiology, Oslo University Hospital, Oslo, Norway c Department of Traumatology, Oslo University Hospital, Oslo, Norway d Section for Transfusion Medicine, Regional Blood Bank, Rigshopitalet, University of Copenhagen, Copenhagen, Denmark e Department of Anaesthesia, Centre of Head and Orthopaedics, Rigshospitalet, University of Copenhagen, Copenhagen, Denmark f Department of Surgery, San Francisco General Hospital, University of California San Francisco, San Francisco, CA, USA g Department of Biostatistics, Institute of Basic Medical Sciences, University of Oslo, Oslo, Norway h The Royal London Hospital, Whitechapel, London, UK i Department of Immunology and Transfusion Medicine, Oslo University Hospital, Oslo, Norway j Institute of Clinical Medicine, Faculty of Medicine, University of Oslo, Norway k The Oslo University Hospital Trauma Registry, Department of Research and Development, Division of Emergencies and Critical Care, Oslo University Hospital, Norway b
A R T I C L E I N F O
A B S T R A C T
Article history: Accepted 17 November 2012
Background: Massive haemorrhage is a leading cause of preventable deaths in trauma. Traumatic coagulopathy is frequently present early after trauma, and is associated with increased mortality. A number of recent trials suggest that viscoelastic haemostatic assays (VHA), such as thromboelastography and thromboelastometry, are useful tools in guiding transfusion. Treatment algorithms exist for the use of VHAs but are not validated in traumatic haemorrhage. In this study we examined the interchangeability of two commonly used VHAs, TEG1 and RoTEM1. Methods: A total of 184 trauma patients over the age of 18, requiring full trauma team activation, were included at three different hospitals in three different countries (Copenhagen, Denmark, San Francisco, CA, USA and Oslo, Norway). Blood samples were drawn immediately upon arrival, and TEG1 and RoTEM1 analyzed simultaneously. Correlations were calculated using. Spearman’s rank correlation coefficient. Agreement was evaluated by Bland–Altman plots and calculation of limits of agreement. Results: The mean ISS in the total population was 17, and the mortality was 16.5%. Mean base excess was 2.8 (SD: 4.2). The correlation coefficient for corresponding values for the two devices was 0.24 for the Rtime vs CT in all centres combined. For the K-time vs CFT the correlation was 0.48, for the a-angleTEG vs a-angleRoTEM 0.44, and for MA vs MCF 0.76. Limits of agreement exceeded the preset clinically acceptable deviation of 10% for all variables in all centres except for MA/MCF in one centre (Copenhagen). Generally, correlation coefficients were lower and agreement poorer in the one centre (Oslo) where measurements were performed bedside by clinicians. Conclusion: Inter-changeability between TEG1 and RoTEM1 is limited in the trauma setting. Agreement seems poorer when clinicians operate the devices. Development and validation of separate treatment algorithms for the two devices is required. ß 2012 Published by Elsevier Ltd.
Keywords: Haemorrhage Trauma Coagulopathy TEG RoTEM Inter-changeability
Introduction Massive haemorrhage accounts for up to 40% of trauma-related deaths in patients reaching hospital, and is considered to be the leading cause of preventable deaths.1,2 Traumatic coagulopathy is
* Corresponding author at: Department of Anaesthesiology, Oslo University Hospital, Ulleval, Kirkeveien 166, 0407 Oslo, Norway. Mobile: +47 975 45 667. E-mail address: [email protected] (J.S. Hagemo).
detectable in 25–34% of patients on admission, and is highly predictive of poor outcome.3,4 A number of recent studies suggest that viscoelastic haemostatic assays (VHAs), such as thromboelastography and thromboelastometry, are useful tools in guiding transfusions and pharmacological coagulation support in trauma patients.5–9 The principle behind VHAs is based on the tendency of blood to increase its viscosity and elasticity through the process of coagulation. The clot that eventually forms subsequently dissolves during fibrinolysis. The clot changes over time are visualized as an
0020–1383/$ – see front matter ß 2012 Published by Elsevier Ltd. http://dx.doi.org/10.1016/j.injury.2012.11.016
Please cite this article in press as: Hagemo JS, et al. Evaluation of TEG1 and RoTEM1 inter-changeability in trauma patients. Injury (2012), http://dx.doi.org/10.1016/j.injury.2012.11.016
G Model
JINJ-5208; No. of Pages 6 2
J.S. Hagemo et al. / Injury, Int. J. Care Injured xxx (2012) xxx–xxx
evolving trace where the time to initiation of clot formation, the rate of clot formation and the maximum strength of the clot are among the most commonly utilized variables.5 VHA has advantages over conventional plasma based coagulations tests in some aspects. The overall coagulation picture expressed by VHA may be more clinically relevant than conventional assays, as it reflects the coagulation process in whole blood, rather than in fragments of the coagulation system.5,10 In addition, VHAs may be performed bedside, as point-of-care (PoC) measurements, reducing the time delay associated with conventional laboratory assays. Two frequently used VHA devices are TEG1 and RoTEM1. Both are marketed as PoC devices that give reliable and comparable results when operated in near patient environment by clinicians. However, they differ markedly in terms of function and user interface, and it is not clear whether this affects their performance in a clinical setting.11 Only a few studies have previously compared the interchangeability of TEG1 and RoTEM1.12–14 They conclude that the assays to a certain extent may be inter-changeable. However, these studies have exclusively been performed within cardiac surgery and liver transplantation procedures, and not in a trauma population. There are indications that the mechanisms behind coagulopathy in trauma differ significantly from coagulopathy in other massive bleeding scenarios.15,16 This may have implications for VHA, and consequently the inter-changeability of TEG1 and RoTEM1. To the best our knowledge, no comparison between the two methods has been performed in a trauma patient population. The aim of the current study was to compare the results of the initial TEG1 and RoTEM1 analyses in a cohort of trauma patients, and to assess the inter-changeability of the two devices. Methods Study design and patient selection We performed a multi-centre observational study in three major trauma centres; San Francisco General Hospital, CA, USA; Rigshospitalet, Copenhagen, Denmark; Oslo University Hospital, Ulleva˚l, Norway. In San Francisco General Hospital and Rigshospitalet, the samples were transported to a laboratory facility and analyzed by laboratory-trained personnel. In Oslo University Hospital the tests were run as PoC measurements by a group of trained clinicians. Patients more than 18 years old and who required full trauma team activation were eligible. Patients were excluded if the time from injury to admission was more than 2 h, if they had received more than 2000 ml of fluids pre-hospitally, if they were pregnant, had known liver failure or bleeding disorders, or were on anticoagulant medications other than acetyl salicylic acid. Data describing demographics, injury severity, admission physiology and outcome were retrieved from the institutional trauma registries. Ethics approval was obtained in accordance with local regulations for each centre. Sampling and VHA methodology In accordance with previously published studies, we chose the TEG1 5000 Hemostasis Analyzer System (Haemonetics Corp., MA, USA) with kaolin as an activator, and RoTEM1 (Tem International GmbH, Munich, Germany) with tissue factor as the activating agent (ExTEM1). Blood was collected from patients upon arrival in the trauma room in a citrated tube, by puncture of the femoral or radial artery. For TEG1, 1000 ml of blood was pipetted and blended in the kaolin containing test tube. After 1 min 340 ml was extracted with a manually operated pipette, and deposited in the designated
plastic cup with 20 ml of CaCl2. The cup was then elevated into test position, and the measurements initiated within 20 s. For RoTEM1, the automated pipette was used to extract 20 ml of CaCl2 (StarTEM1 reagent). An air cushion was applied in the tip before extraction of 20 ml of the ExTEM1 reagent, and finally mixing with 300 ml of blood in the cup and measurement initiated within 30 s. The TEG1 and RoTEM1 assays were run simultaneously in all three centres, and with the temperature set at 37.0 8C. We compared four widely used variables from the VHA trace: the reaction time from initiation of the assay to the first detectable coagulation, denoted R-time for TEG1 and clotting time (CT) for RoTEM1; the time from start of coagulation to clot amplitude of 20 mm, called the K-time in TEG1 and clot formation time (CFT) in RoTEM1; the angle by which the clot strength increases, called the a-angle for both devices; lastly, the maximum amplitude (MA) for TEG1, which corresponds to the maximum clot firmness (MCF) for the RoTEM1 device. Statistical analysis Data are presented as mean and standard deviation (SD) or number (%) unless stated otherwise. For comparison of baseline data between the three centres, one-way ANOVA was used for continuous variables, and Chi Square-test for categorical variables. Injury severity score (ISS) was regarded as a continuous variable. Correlation between TEG1 and RoTEM1 measurements was calculated using Spearman non-parametric correlation. For evaluation of agreement between TEG1 and ROTEM1 measurements, we applied Bland–Altman difference-mean plot, i.e. a plot of the methods differences of measurements (D) against the corresponding average (A), and estimation of corresponding limits of agreement (LoA), calculated as 1.96 SD from the mean difference of measurements.17 As TEG1 and ROTEM1 are technically different measurements of the same concept, a significant linear relationship in the Bland– Altman plot was expected. A log transformation is often recommended17 but was unsuccessful in compensating for this relationship in our data. Instead, a generalized version of LoA using linear regression, as suggested in a revised version of the Bland– Altman methodology,18 was used. In the Bland–Altman plots where there was a significant non-zero linear association between D and A, the linear relationship D = aA + b was estimated using univariate linear regression, the estimated linear association extracted from each of the actual differences D, and subsequently calculating the SD and corresponding LoA on these transformed data D* = D (aA + b). A predefined set of clinically acceptable LoA was established based on clinical experience and previous publications.12 These limits were based on the assumption that a 10% deviation in corresponding variables is acceptable for mean values. Clinically acceptable LoA by these conditions were 20.8 s for R/CT, 10.9 s for K/CFT, 6.78 for the a-angle and 6.2 mm for the MA/MCF. A p-value <0.05 was considered statistically significant. All calculations were made using SPSS 19.0 (SPSS Inc., Chicago, IL, USA). Results A total of 184 patients were included in the study. Patient characteristics are shown in Table 1. No statistically significant differences between centres were found except for ISS, which was significantly higher in the San Francisco population compared to both Oslo and Copenhagen (p = 0.031). Scatter plots of the corresponding TEG1 and RoTEM1 values for the four predefined variables are shown in Fig. 1, and corresponding
Please cite this article in press as: Hagemo JS, et al. Evaluation of TEG1 and RoTEM1 inter-changeability in trauma patients. Injury (2012), http://dx.doi.org/10.1016/j.injury.2012.11.016
G Model
JINJ-5208; No. of Pages 6 J.S. Hagemo et al. / Injury, Int. J. Care Injured xxx (2012) xxx–xxx
3
Table 1 Descriptive statistics of trauma patients included in the study. Location Copenhagen
Oslo Age (years) Sex (% male)* MOI (% Blunt)* ISS§ Base excess R-time K-time a-angleTEG MA CT CFT a-angleRoTEM MCF Mortality (%)*
47.4 (20.0) 60.4 90.6 17 (5, 29) 3.3 (5.0) 427.7 (190.0) 138.9 (73.2) 59.4 (12.3) 62.0 (9.0) 68.6 (17.1) 104.1 (39.9) 69.5 (6.9) 61.1 (10.0) 15.1
n = 53 n = 53 n = 53 n = 50 n = 46 n = 66 n = 66 n = 66 n = 66 n = 66 n = 65 n = 66 n = 66 n = 53
48.4 (20.0) 57.0 91.1 17 (8, 26) 2.3 (3.5) 343.1 (69.4) 111.0 (32.3) 65.1 (6.3) 62.8 (6.1) 63.0 (13.0) 106.0 (30.6) 69.3 (5.2) 61.2 (6.1) 19.0
San Francisco n = 79 n = 79 n = 79 n = 78 n = 74 n = 79 n = 79 n = 79 n = 79 n = 79 n = 79 n = 79 n = 79 n = 79
44.5 (18.2) 71.1 71.8 26 (16, 36) 3.7 (4.2) 237.6 (83.8) 93.4 (71.3) 69.3 (9.0) 64.0 (7.0) 59.4 (12.2) 82.0 (30.4) 74.2 (5.6) 63.5 (6.1) 13.2
Total n = 39 n = 38 n = 39 n = 35 n = 13 n = 38 n = 39 n = 39 n = 39 n = 38 n = 39 n = 39 n = 39 n = 38
47.2 (19.5) 61.2 86.5 17 (7.5, 26.5) 2.8 (4.2) 351.7 (145.7) 117.3 (60.9) 63.9 (10.0) 62.8 (7.5) 64.3 (14.8) 100.2 (35.3) 70.4 (6.2) 61.7 (7.7) 16.5
n = 171 n = 170 n = 171 n = 163 n = 133 n = 183 n = 184 n = 184 n = 184 n = 183 n = 183 n = 184 n = 184 n = 170
Values given as mean (SD), except * given as number (%), and § given as median (inter quartile range). MOI: mechanism of injury; ISS: injury severity score. Mortality is given at 30 days. The TEG parameter R-time (reaction time) corresponds to RoTEM CT (clotting time) and is measured in seconds. K-time (kinetic time) corresponds to CFT (clot formation time), also measured in seconds. a-angleTEG corresponds to a-angleRoTEM and is measured in degrees (8). MA (maximum amplitude) corresponds to MCF (maximum clot formation) and is measured in millimetres. Base excess is measured in mequiv./L.
Fig. 1. Scatter plots showing the relation between the corresponding TEG1 and RoTEM1 measurements in each centre. (a) TEG reaction time (R) vs RoTEM clotting time (CT) in seconds, (b) TEG kinetic-time (K) vs RoTEM clot formation time (CFT) in seconds, (c) TEG a-angle vs RoTEM a-angle in degrees and (d) TEG maximum amplitude (MA) vs RoTEM maximum clot formation (MCF) in millimetres.
Please cite this article in press as: Hagemo JS, et al. Evaluation of TEG1 and RoTEM1 inter-changeability in trauma patients. Injury (2012), http://dx.doi.org/10.1016/j.injury.2012.11.016
G Model
JINJ-5208; No. of Pages 6 J.S. Hagemo et al. / Injury, Int. J. Care Injured xxx (2012) xxx–xxx
4
Table 2 Spearman correlation coefficients for corresponding TEG1 and RoTEM1 values, in each of three centres and total. Oslo (n = 65)
R-time vs CT K-time vs CFT a-angleTEG vs RoTEM MA vs MCF
Copenhagen (n = 79)
San Francisco (n = 38)
All centres (n = 182)
Correlation
p-Value
Correlation
p-Value
Correlation
p-Value
Correlation
p-Value
0.15 0.21 0.17 0.67
0.231 0.093 0.185 0.001
0.11 0.56 0.45 0.84
0.33 <0.001 <0.001 <0.001
0.06 0.57 0.62 0.86
0.72 <0.001 <0.001 <0.001
0.24 0.48 0.44 0.76
0.001 <0.001 <0.001 <0.001
The TEG parameters: R-time (reaction time), K-time (kinetic time), a-angleTEG and MA (maximum amplitude) corresponds to the RoTEM parameters: CT (clotting time), CFT (clot formation time), a-angleRoTEM, and MCF (maximum clot formation), respectively.
correlation coefficients in Table 2. No statistically significant correlation was found for R/CT-time for any centre individually. Only the correlation between MA and MCF was statistically significant within all centres. When data from all centres were combined, a statistically significant correlation was found for all four TEG1/RoTEM1 variables, ranging from 0.21 for R/CT-time to 0.76 for MA/MCF. The Bland–Altman plots for TEG1 and RoTEM1 measurements are shown in Fig. 2. There was a significant linear relationship between difference and average for paired measurements of R/CT, K/CFT and a-angle, but not for MA/MCF. LoAs are given in Table 3. Neither of the LoA, evaluated within each centre or overall, were within the predefined clinically acceptable range for R/CT, K/CFT or a-angleTEG/a-angleRoTEM. For MA/MCF, only measurements from
Copenhagen were within the predefined clinically acceptably limits. Discussion The main finding of this study is that the agreement between kaolin activated TEG1 assay and tissue factor activated RoTEM1 assay is limited in a clinical setting of trauma patients. We found that correlation was highly variable in the different stages of the clotting process, as well as between the different centres. The best correlation was found for MA/MCF with a correlation coefficient of 0.84 and 0.86 in Copenhagen and San Francisco, respectively. In the Oslo centre, where clinicians ran the analyses as PoC, only the MA/ MCF measurements correlated significantly, and with a markedly
Fig. 2. Bland–Altman difference-mean plot for corresponding TEG1 and RoTEM1 values. (a) TEG R-time vs RoTEM CT, (b) TEG K-time vs RoTEM CFT, (c) TEG a-angle vs RoTEM a-angle and (d) TEG MA vs RoTEM MCF. Regression lines for mean bias (middle line), and limits of agreement are indicated.
Please cite this article in press as: Hagemo JS, et al. Evaluation of TEG1 and RoTEM1 inter-changeability in trauma patients. Injury (2012), http://dx.doi.org/10.1016/j.injury.2012.11.016
G Model
JINJ-5208; No. of Pages 6 J.S. Hagemo et al. / Injury, Int. J. Care Injured xxx (2012) xxx–xxx
5
Table 3 Mean bias and limits of agreement for corresponding TEG1 and RoTEM1 measurements. Copenhagen
Oslo D a R-time vs CT K-time vs CFT a-angleTEG vs RoTEM MA vs MCF
1.94 0.82 0.90 0
LoA b 122.9 64.2 67.9 0.9
D a
67.6 111.5 21.8 13.0
1.78 0 0.24 0
San Francisco LoA
b 82.0 5.0 20.2 1.6
D a
49.4 48.4 10.7 5.7
1.9 1.13 0.65 0
Total LoA
b 107.4 87.7 51.1 0.5
D a
48.2 95.7 14.8 8.0
1.89 0.72 0.67 0
LoA b 106.5 61.2 51.6 1.13
57.1 92.9 17.1 9.4
The mean bias (D) is expressed as aA + b, where a is the linear regression derived coefficient, A is the average of two corresponding measurements and b is a constant. LoA: limits of agreement. R-time: reaction time, K-time: kinetic time, MA: maximum amplitude, CT: clotting time, CFT: clot formation time, a-angleRoTEM: MCF (maximum clot formation).
weaker correlation coefficient of 0.67. Bland–Altman analyses showed poor agreement between R/CT, K/CFT and the a-angles in all centres, with wider LoA than what was considered clinically acceptable. MA/MCF had the best agreement over all and in each centre individually, but only the measurements in Copenhagen showed acceptable LoA. In accordance with the weaker correlations, the LoA in Oslo was also broader than in the other two centres. A clinical comparison of TEG1 and RoTEM1 assays was examined by Venema et al. in a group of 46 patients scheduled for cardiac surgery.12 Their findings indicate that the kaolin activated TEG1 assay and the tissue factor activated RoTEM1 assay (ExTEM1) were inter-changeable for the a-angle and MA/MCF, but not for R/CT and K/CFT. The lack of interchangeability of R/CT and K/CFT is in accordance with our findings. In the same study, an evaluation of repeatability of measurements from two healthy volunteers was also performed. Their analyses showed variability for the R/CT and K/CFT variables of more than 10%, and consequently considered inappropriate for clinical purposes. The inter-changeability of the a-angle and MA/MCF reported in the study by Venema et al. is, however, partly in contrast to our findings. This may be due to at least two factors. First, our study population of trauma patients is expected to contain subjects with a significant degree of coagulopathy. The activating agent in our TEG1 measurements, kaolin, activates coagulation by directly binding to and activating factor XII, subsequently activating factors XI, IX and X. The last step however, depends on factor VIII as a cofactor.19 In trauma patients, studies have shown increased levels of activated protein C due to hypoperfusion and tissue injury.20 Activated protein C is a potent inhibitor of factor VIII.21 Hence, the rate of clot formation may be decreased when activated by kaolin compared to tissue factor in this group of patients, and subsequently increasing the discrepancy from tissue factor activated measurements. A direct comparison of TEG1 and RoTEM1 was performed in a porcine model by Tomori et al.22 In their study, blood was collected in three different stages of coagulopathy in combination with trauma. No activating agent was added to the blood sample. They found correlation coefficients higher than in our study, although the correlation for the a-angle was not statistically significant. Analyses of agreement were not performed in their study. Their findings support the fact that the activating agent may have implications for the correlation of the measurements. In a recent paper by Solomon et al.14, TEG and RoTEM assays with platelet inhibition by abciximab and cytochalasin D were evaluated. In their study both reagents were used on both machines, and the samples were subjected to hemostatically active medications, as well as different degrees dilution. Their most notable findings are that the MCF and MA differ depending on both reagent and device used. Moreover, the effect of high dose heparin affected the MA in the TEG assay when abciximab was used,
whereas the cytochalasin D in the RoTEM assay remained unchanged. The authors conclude that target thresholds for fibrinogen substitution should be reviewed for the each device separately. One previous study has shown that kaolin activated TEG1 and tissue factor activated RoTEM1 show better agreement than several alternative assays. We therefore found it reasonable to use kaolin and tissue factor for TEG1 and RoTEM1 respectively in our study.12 Regardless of whether kaolin or tissue factor is used, the addition of an activating agent may be a crucial part of the test procedure in TEG1 and RoTEM1. Only small errors of pipetting may affect the clot initiation time and clot formation rate. The discrepancies found for one of the centres in our study might be due to staff less experienced with laboratory techniques. This finding is nevertheless of clinical importance, as it indicates that agreement between the two VHA devices is subject to variations in operator skills. Other possible explanations for the wide LoA in our study could be the time factor from sampling to initiation of the analyses. The time from sampling to analyses differed between centres. In San Francisco and Oslo, the measurements were performed almost immediately after sampling, whereas in Copenhagen they were performed after 1 h. Previous studies examining the significance of time delay indicate that R-time may be reduced, and the a-angle increased by approximately 10% over a 60-min period after collection in a citrated tube. The MA value, on the other hand, was not affected.23 We find it unlikely that this would inflict a systematic bias to our results, since changes of comparable magnitude also occur when tissue factor was used as an activator, and corresponding TEG1 and RoTEM1 analyses in our study were performed simultaneously. We defined LoA constituting less than 10% of mean values as being clinically acceptable. This limit might be too strict. According to previous studies on method comparisons some authors argue that a lesser extent of agreement is tolerable.24 LoA for MA/MCF in the total population indicate that a deviation of 15.1% of the mean values would have to be accepted. Correspondingly, the acceptable deviation for R/CT would have to be 27.5%, for K/ CFT 85.4% and for a-angleTEG/a-angleRoTEM 25.4%. This study has some limitations. The difference between the numbers of patients from each centre affected the level of significance for the correlations. A more homogenous group of laboratory technicians might have given better agreement between measurements overall, but an evaluation of the practical use of VHA as PoC devices was considered to be one important aspect of this study. Transfusion algorithms based on TEG1 and RoTEM1 do exist, but there are no studies validating their use with regard to patient outcome.25 This study was not powered to determine whether the overall performance in detecting coagulation perturbations or predicting outcome differed between the two devices. The discrepancies found in our data, however, warrant further research to validate the two devices in the trauma setting.
Please cite this article in press as: Hagemo JS, et al. Evaluation of TEG1 and RoTEM1 inter-changeability in trauma patients. Injury (2012), http://dx.doi.org/10.1016/j.injury.2012.11.016
G Model
JINJ-5208; No. of Pages 6 J.S. Hagemo et al. / Injury, Int. J. Care Injured xxx (2012) xxx–xxx
6
TEG1 and RoTEM1 have become widely used in the trauma setting over the past few years. Their proposed advantages in being PoC devices have been a driving force. This study indicate that the inter-changeability between kaolin activated TEG1 and tissue factor activated RoTEM1 measurements is limited in the trauma setting. Agreement is worse when the devices are used as PoC devices by clinicians in our study. Development and validation of separate trauma treatment algorithms for TEG1 and RoTEM1 is required. Conflict of interest statement The research group have received support from both Haemonetics and TEM international for this study in the form of reagents for the analyses, and leasing of devices at reduced prices. KB has also performed consultancy work for Haemonetics. Acknowledgements All of the author institutes are affiliated members of the International Trauma Research Network (www.INTRN.org) and as such this work represents a combined output resulting from this international partnership. The authors would like to thank The Oslo University Hospital Trauma Registry for contributing with descriptive data in the Oslo population of the study. References 1. Sauaia A, Moore FA, Moore EE, Moser KS, Brennan R, Read RA, et al. Epidemiology of trauma deaths: a reassessment. Journal of Trauma 1995;38(February (2)):185–93. 2. Gruen RL, Jurkovich GJ, McIntyre LK, Foy HM, Maier RV. Patterns of errors contributing to trauma mortality: lessons learned from 2,594 deaths. Annals of Surgery 2006;244(September (3)):371–80. 3. Brohi K, Singh J, Heron M, Coats T. Acute traumatic coagulopathy. Journal of Trauma 2003;54(June (6)):1127–30. 4. MacLeod JBA, Lynn M, McKenney MG, Cohn SM, Murtha M. Early coagulopathy predicts mortality in trauma. Journal of Trauma 2003;55(July (1)):39–44. 5. Fries D, Innerhofer P, Schobersberger W. Time for changing coagulation management in trauma-related massive bleeding. Current Opinion in Anesthesiology 2009;22(April (2)):267–74. 6. Johansson PI, Stissing T, Bochsen L, Ostrowski SR. Thrombelastography and tromboelastometry in assessing coagulopathy in trauma. Scandinavian Journal of Trauma Resuscitation and Emergency Medicine 2009;17(Januar (1)):45. 7. Scho¨chl H, Nienaber U, Hofer G, Voelckel W, Jambor C, Scharbert G. Goaldirected coagulation management of major trauma patients using thromboelastometry (ROTEM)-guided administration of fibrinogen concentrate and prothrombin complex concentrate. Critical Care 2010;14(Januar (2)):R55.
8. Davenport R, Manson J, De’ath H, Platton S, Coates A, Allard S, et al. Functional definition and characterization of acute traumatic coagulopathy. Critical Care Medicine 2011;July. 9. Kashuk JL, Moore EE, Sawyer M, Le T, Johnson J, Biffl WL, et al. Postinjury coagulopathy management: goal directed resuscitation via POC thrombelastography. Annals of Surgery 2010;251(April (4)):604–14. 10. Johansson PI, Ostrowski SR, Secher NH. Management of major blood loss: an update. Acta Anaesthesiologica Scandinavica 2010;July. 11. Jackson GNB, Ashpole KJ, Yentis SM. The TEG1 vs the ROTEM1 thromboelastography/thromboelastometry systems. Anaesthesia 2009;64(Februar (2)): 212–5. 12. Venema LF, Post WJ, Hendriks HGD, Huet RCG, de Wolf JTW, De Vries AJ. An assessment of clinical interchangeability of TEG and RoTEM thromboelastographic variables in cardiac surgical patients. Anesthesia and Analgesia 2010;111(August (2)):339–44. 13. Coakley M, Reddy K, Mackie I, Mallett S. Transfusion triggers in orthotopic liver transplantation: a comparison of the thromboelastometry analyzer, the thromboelastogram, and conventional coagulation tests. Journal of Cardiothoracic and Vascular Anesthesia 2006;20(August (4)):548–53. 14. Solomon C, Sorensen B, Hochleitner G, Kashuk J, Ranucci M, Schochl H. Comparison of whole blood fibrin-based clot tests in thrombelastography and thromboelastometry. Anesthesia and Analgesia 2012;114(April (4)):721–30. 15. Cohen MJ, Call M, Nelson M, Calfee CS, Esmon CT, Brohi K, et al. Critical role of activated protein c in early coagulopathy and later organ failure infection and death in trauma patients. Annals of Surgery 2011;December. 16. Brohi K, Cohen MJ, Ganter MT, Matthay MA, Mackersie RC, Pittet JF. Acute traumatic coagulopathy: initiated by hypoperfusion: modulated through the protein C pathway? Annals of Surgery 2007;245(May (5)):812–8. 17. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986;1(Februar (8476)):307–10. 18. Bland JM, Altman DG. Measuring agreement in method comparison studies. Statistical Methods in Medical Research 1999;September (8):135–60. 19. Thorsen K, Ringdal KG, Strand K, Soreide E, Hagemo J, Soreide K. Clinical and cellular effects of hypothermia, acidosis and coagulopathy in major injury. British Journal of Surgery 2011;98(July (7)):894–907. 20. Cohen MJ, Call M, Nelson M, Calfee CS, Esmon CT, Brohi K, et al. Critical role of activated protein c in early coagulopathy and later organ failure. Infection and Death in Trauma Patients Annals of Surgery 2012;255(February (2)):379–85. 21. Esmon CT. The protein C pathway. Chest 2003;124(September (3 Suppl.)):26S– 32S. 22. Tomori T, Hupalo D, Teranishi K, Teranishi K, Michaud S, Hammett M, et al. Evaluation of coagulation stages of hemorrhaged swine: comparison of thromboelastography and rotational elastometry. Blood Coagulation and Fibrinolysis 2010;21(January (1)):20–7. 23. Johansson PI, Bochsen L, Andersen S, Viuff D. Investigation of the effect of kaolin and tissue-factor-activated citrated whole blood, on clot-forming variables, as evaluated by thromboelastography. Transfusion 2008;48(November (11)): 2377–83. 24. Critchley LA, Critchley JA. A meta-analysis of studies using bias and precision statistics to compare cardiac output measurement techniques. Journal of Clinical Monitoring and Computing 1999;15(February (2)):85–91. 25. Afshari A, Wikkelso A, Brok J, Moller AM, Wetterslev J. Thromboelastography (TEG) or thromboelastometry (ROTEM) to monitor haemotherapy versus usual care in patients with massive transfusion. Cochrane Database of Systematic Reviews 2011;(3):CD007871.
Please cite this article in press as: Hagemo JS, et al. Evaluation of TEG1 and RoTEM1 inter-changeability in trauma patients. Injury (2012), http://dx.doi.org/10.1016/j.injury.2012.11.016