Using data from the study with the largest sample size of the five studies provided the most precise measure of test accuracy.32 If the pretest probability of a TE event in a postoperative patient was 4% (the prevalence of this study), a positive TEG test would generate a new posttest probability of 8% and a negative TEG test would generate a posttest probability of 1%. In patients at higher risk, if the pretest probability was 45% (which is the highest prevalence reported in the 10 studies included in this analysis),35 a positive TEG test result increased the posttest probability to 63%, and a negative TEG test decreased the probability to 21%.
This is the first systematic review to examine the diagnostic value of TEG to predict DVT and other TE events in postsurgical patients. The majority of studies reported favorable results associated with the use of TEG testing. Although the MA value seemed to be the best parameter for defining a hypercoagulable state, there was no consensus regarding the minimum (cutoff) MA value that should be used for predicting TE events. Based on the 5 studies with measures of test accuracy, the diagnostic odds ratio ranged from 1.5 to 27.7, area under the curve ranged from 0.57 to 0.91, suggesting that TEG testing had considerable variability in its accuracy for predicting postoperative TE events. Of note, 3 studies showed that the TEG had diagnostic odds ratios of 5–6 for predicting postoperative TE events.28,32,33 Therefore, when TEG is used alone, the posttest probabilities did not change significantly from the pretest probabilities with respect to their ability to alter clinical decision-making.27,28,32,33 The TEG may, however, have some incremental value when combined with the Wells’ score43 or other decision-making aids for predicting TE events.
The TEG has been used to evaluate both preexistent and acquired hypercoagulable states.13–18 Compared to healthy reference subjects, patients with hyperhomocysteinaemia and heterozygosity for the factor V Leiden polymorphism had a significantly greater hypercoagulant whole blood coagulation clot signature as defined by a shortened clotting time and an accelerated maximum velocity of clot propagation.13 TEG can also serve as a useful adjunctive test for thrombophilia screening, particularly in patients in whom the regular thrombophilia screen tests were normal.44
Surgery itself is an important contributing factor to hypercoagulable states detected by TEG testing.16,45–48 The mechanisms responsible for surgery-induced hypercoagulation include tissue factors released from damaged vessels, local tissue trauma, increases in stress, hormone levels, inflammatory cytokine network activation and compromised fibrinolysis. Schreiber et al.45 reported that hypercoagulability after traumatic injuries was most prevalent during the first 24 h, and TEG testing was more sensitive than routine coagulation assays for the detection of a hypercoagulable state. Mahla et al.14 revealed independent contributions of platelets and procoagulatory proteins to clot strength. Spiel et al.15 reported that endotoxemia significantly shortened TEG clotting time and maximal lysis, which correlated with established in vivo markers of coagulation activation (F[1 + 2] and fibrinolysis [t-PA]), respectively. Several studies have confirmed the changes in fibrinolysis detected by TEG testing,49–52 but the relationship between hyperfibrinolysis and TE events requires further study.
In the last 26 yr, many changes have occurred in TEG methodology, which helps explain the heterogeneity observed in TEG test characteristics and reference ranges. Wide variability was reported in the mean values of TEG variables between patients with and without TE events. Of the six studies with a clear definition of hypercoagulability, only two had sufficient data for patients to be classified as either normal or hypercoagulable. There was no universal definition among the 10 studies for defining the value for normal or hypercoagulable state. Obviously, the heterogeneity of the methodology was the main reason for the lack of consensus in reference ranges. There were two commonly used TEG devices, namely TEG® (Hemoscope Corporation, Niles, IL) and the RoTEM® (Pentapharm GmbH, Munich, Germany).51 These two monitors use different reaction cups and activators, which helps to differentiate coagulation, platelet function, and fibrinolysis. Although most of the parameters are comparable and measure similar changes in the tracings (e.g., both the MA of TEG® and maximum clot firmness of RoTEM® represent the ultimate strength of the fibrin clot as defined as the maximal width of the tracing), one needs to be aware that differences may be found when comparing results using TEG® and RoTEM®. For example, a previous study53 found divergent results when these two devices were directly compared, and these differences could affect clinical decision-making. Additional comparative studies are clearly needed to develop clinically useful reference ranges. One multicenter study reported that the reference ranges were comparable among centers.54 Therefore, it is likely that with the introduction of more sophisticated techniques and reliable reference ranges, TEG testing will be more useful for evaluating hypercoagulable states in the future.
Although reagents for the individual pathways provide differentiation between the coagulation system, platelet, fibrinolysis and the influence of anticoagulant drugs, TEG testing by its nature is a “global assay” for the entire hemostasis system. The ability to detect the systematic changes caused by different coagulation disorders makes TEG a suitable tool to predict postoperative TE events because there are many factors affecting the hemostasis system during the perioperative period. Differentiated assays are more useful in attempting to understand the underlying mechanism(s) involved and the pathway(s) that are altered.
Theoretically, it would be optimal if preoperative TEG measurements could predict postoperative TE events. Earlier detection of hypercoaguable states (i.e., before the operation or at early stage after the operation) would facilitate intervention with appropriate anti-TE therapies. The results of this review suggest that evoked hypercoagulability in the early postoperative period is the most important finding for predicting TE complications. A more robust therapeutic response to early changes in a patient’s coagulation status as assessed by TEG might prove to be effective in preventing the development of postoperative TE events. Therefore, the early postoperative period may be the optimal sampling time, and 8 of the 10 studies obtained blood samples during this time period. It may be helpful in future studies to evaluate the diagnostic accuracy of preoperative versus early postoperative TEG values. Importantly, the clinical outcome variables should be standardized when evaluating the performance of the TEG as a diagnostic test.
Among the 10 studies, there was a high degree of heterogeneity in patient characteristics and reference standards, contributing to the wide variability observed in patient outcome measurements. The specific TE events included arterial, venous and vascular graft thrombosis, and the reference standards for these events also varied across all studies. The incidence of DVT varied regardless of the surgical population and included both symptomatic and asymptomatic patients. Given the clinical heterogeneity among studies, statistical pooling of test accuracy measures as part of a meta-analysis of these data would not have been appropriate.
The optimal design for assessing the accuracy of TEG testing as a predictor of TE complications is a prospective, blinded comparison of TEG and the reference test in a consecutive series of patients from a relevant (at risk) surgical patient population.55 The results of this systematic review demonstrated that the quality of the published studies is highly variable, ranging from poor (scored as 3) to good (scored as 7), and none of the studies completely fulfilled all the requirements for a high-quality diagnostic test study.37,38 Moreover, only a few studies described the study methodology in sufficient detail for replicating the findings. A majority of the studies lacked adequate details for interpretation of TEG and reference test results. The use of a blinded study design is crucial for preventing investigator bias in the clinical diagnostic evaluations, as unblinded assessments can inappropriately inflate the accuracy of a diagnostic test.56 The reference tests were not performed in all patients in 4 of the 10 studies.30,32,34,35 Since most of the patients are asymptomatic, verification bias might complicate the interpretation of these data and lead to a biased estimate of the diagnostic test’s accuracy. None of the studies performed an a priori sample size calculation and, therefore, there is uncertainty about the power of the negative studies. Of the 5 studies which provided information regarding sensitivity and specificity, 3 had relatively small sample sizes (10–65 patients).
The results of the level of diagnostic test evaluation suggested that most of the studies were poor-to-moderate quality. Four of the studies were focused on the technical feasibility (level 1), which is considered a poor diagnostic level.38 To be included in a meta-analysis of a diagnostic test, the studies should be homogeneous and ideally a minimum of a level 2 diagnostic test evaluation. Only one study31 evaluated the patient outcome impact of TEG, but it failed to reach statistical significance. Another study33 evaluated the impact of the therapeutic choice, which is considered a moderate level of diagnostic test evaluation.
In this literature review, there are some important deficiencies which must be acknowledged. First, the heterogeneity of study population and TEG characteristics was the greatest obstacle to performing statistical pooling of these data. There were insufficient data available concerning a single type of surgical procedure, patient characteristic or TEG device. Second, most of these studies used different definitions of hypercoagulability. Considering the heterogeneity of the TEG methodology, we used the definitions provided by the authors to calculate specificity and sensitivity, and then compared the results across the studies meeting the prespecified inclusion criteria. Another important confounding factor was that patients manifesting symptoms of hypercoagulability, as determined by an abnormal TEG variable, were promptly treated with standard TE prophylaxis regimens. This clinical practice would compromise the predictive value of TEG testing. Therefore, it is not surprising that sensitivity and specificity values of 50% or less were obtained in this analysis. Future prospective studies should be well-designed and more rigorously controlled cohort studies using the same treatment regimens for all patients.
In conclusion, the predictive accuracy of TEG measurements for postoperative TE events in adult postsurgical patients is highly variable; an increased MA seems to be a useful risk factor for postoperative TE events. Due to the lack of adequate data, it was not possible to perform a meta-analysis. The overall quality of the studies in this systematic literature review varied widely, and the diagnostic test evaluation was generally of a low level. High-quality diagnostic studies examining the performance of TEG as a predictor of TE events are needed in order to determine whether TEG testing is a clinically useful screening test for high risk surgical patients.
1.Nerurkar J, Wade WE, Martin BC. Cost/death averted with venous thromboembolism prophylaxis in patients undergoing total knee replacement or knee arthroplasty. Pharmacotherapy 2002;22:990–1000
2.Kakkar VV, Cohen AT, Edmonson RA, Phillips MJ, Cooper DJ, Das SK, Maher KT, Sanderson RM, Ward VP, Kakkar S. Low molecular weight versus standard heparin for prevention of venous thromboembolism after major abdominal surgery. The Thromboprophylaxis Collaborative Group. Lancet 1993;341:259–65
3.Anderson FA Jr, Spencer FA. Risk factors for venous thromboembolism. Circulation 2003;107:I9–I16
4.Bick RL, Haas S. Thromboprophylaxis and thrombosis in medical, surgical, trauma, and obstetric/gynecologic patients. Hematol Oncol Clin North Am 2003;17:217–58
5.Devereaux PJ, Goldman L, Cook DJ, Gilbert K, Leslie K, Guyatt GH. Perioperative cardiac events in patients undergoing noncardiac surgery: a review of the magnitude of the problem, the pathophysiology of the events and methods to estimate and communicate risk. CMAJ 2005;173:627–34
6.McCrath DJ, Cerboni E, Frumento RJ, Hirsh AL, Bennett-Guerrero E. Thromboelastography maximum amplitude predicts postoperative thrombotic complications including myocardial infarction. Anesth Analg 2005;100:1576–83
7.Kelley RE. Stroke in the postoperative period. Med Clin North Am 2001;85:1263–76
8.Breuer AC, Furlan AJ, Hanson MR, Lederman RJ, Loop FD, Cosgrove DM, Greenstreet RL, Estafanous FG. Central nervous system complications of coronary artery bypass graft surgery: prospective analysis of 421 patients. Stroke 1983;14:682–7
9.Patel R, Cook DJ, Meade MO, Griffith LE, Mehta G, Rocker GM, Marshall JC, Hodder R, Martin CM, Heyland DK, Peters S, Muscedere J, Soth M, Campbell N, Guyatt GH. Burden of illness in venous thromboembolism in critical care: a multicenter observational study. J Crit Care 2005;20:341–7
10.Lowe GD, Haverkate F, Thompson SG, Turner RM, Bertina RM, Turpie AG, Mannucci PM. Prediction of deep vein thrombosis after elective hip replacement surgery by preoperative clinical and haemostatic variables: the ECAT DVT Study. European Concerted Action on Thrombosis. Thromb Haemost 1999;81:879–86
11.Virchow RLK. Cellular pathology as based upon Physiololgical and Pathohistology. In: Virchow RLK, ed. Thrombosis and Emboli. Frankfurt: Meidinger Sohn & Co., 1856. Matzdorff AC, Bell WR (transl), ed. Canton: Science History Publications, 1998;5–11,110
12.Whitten CW, Greilich PE. Thromboelastography: past, present, and future. Anesthesiology 2000;92:1223–5
13.Hvitfeldt Poulsen L, Christiansen K, Sørensen B, Ingerslev J. Whole blood thrombelastographic coagulation profiles using minimal tissue factor activation can display hypercoagulation in thrombosis-prone patients. Scand J Clin Lab Invest 2006;66:329–36
14.Mahla E, Lang T, Vicenzi MN, Werkgartner G, Maier R, Probst C, Metzler H. Thromboelastography for monitoring prolonged hypercoagulability after major abdominal surgery. Anesth Analg 2001;92:572–7
15.Spiel AO, Mayr FB, Firbas C, Quehenberger P, Jilma B. Validation of rotation thrombelastography in a model of systemic activation of fibrinolysis and coagulation in humans. J Thromb Haemost 2006;4:411–6
16.Wilson D, Cooke EA, McNally MA, Wilson HK, Yeates A, Mollan RA. Changes in coagulability as measured by thrombelastography following surgery for proximal femoral fracture. Injury 2001;32:765–70
17.Abrahams JM, Torchia MB, McGarvey M, Putt M, Baranov D, Sinson GP. Perioperative assessment of coagulability in neurosurgical patients using thromboelastography. Surg Neurol 2002;58:5–11
18.Bell CR, Cox DJ, Murdock PJ, Sullivan ME, Pasi KJ, Morgan RJ. Thrombelastographic evaluation of coagulation in transurethral prostatectomy. Br J Urol 1996;78:737–41
19.Heather BP, Greenhalgh RM. Prophylaxis of early post operative deep venous thrombosis by monitored heparin infusion. Phlebologie 1982;35:173–85
20.Huang GS, Chang JH, Lee MS, Wu CC, Lin SP, Lin SL, Wong CS. The effect of anesthetic techniques on hemostatic function in arthroscopic surgery: evaluation by thromboelastography. Acta Anaesthesiol Sin 2002;40:121–6
21.Hepner DL, Concepcion M, Bhavani-Shankar K. Coagulation status using thromboelastography in patients receiving warfarin prophylaxis and epidural analgesia. J Clin Anesth 2002;14:405–10
22.Poston RS, White C, Gu J, Brown J, Gammie J, Pierson RN, Lee A, Connerney I, Avari T, Christenson R, Tandry U, Griffith BP. Aprotinin shows both hemostatic and antithrombotic effects during off-pump coronary artery bypass grafting. Ann Thorac Surg 2006;81:104–10
23.Iwade M, Iwade K, Nomura M, Ozaki M. Differences in perioperative coagulation between Japanese and other ethnic groups undergoing laparoscopic cholecystectomy. Surg Endosc 2003;17:2012–5
24.Gibbs NM, Crawford GP, Michalopoulos N. Thrombelastographic patterns following abdominal aortic surgery. Anaesth Intensive Care 1994;22:534–8
25.Quigley RL, Fried DW, Pym J, Highbloom RY. Off-pump coronary artery bypass surgery may produce a hypercoagulable patient. Heart Surg Forum 2003;6:94–8
26.Edwards CM, Morgan JD, Donnelly PK. Cyclosporine-related hypercoagulability: fact or fiction? Transplant Proc 1992;24:2616–7
27.Wen YR, Ho WY, Sun WZ, Or CH, Yeh M, Yao WC, Tai YT. Thromboelastographic study of thrombosis in the implantable central venous access device. Acta Anaesthesiol Sin 1997;35:223–8
28.Traverso CI, Arcelus JI. Prospective assessment of the risk of deep vein thrombosis in elective abdominal surgery. Predictive role of thromboelastography. Thromb Haemorrh Disord 1993;7:9–15
29.Jennings S, Cass AJ, Heather BP, Greenhalgh RM. Coagulation changes during major surgery and relationship to post-operative deep vein thrombosis. J Cardiovasc Surg (Torino) 1981;22:327–9
30.Tuman KJ, McCarthy RJ, March RJ, DeLaria GA, Patel RV, Ivankovich AD. Effects of epidural anesthesia and analgesia on coagulation and outcome after major vascular surgery. Anesth Analg 1991;73:696–704
31.Caprini JA, Arcelus JI, Laubach M, Size G, Hoffman KN, Coats RW, Blattner S. Postoperative hypercoagulability and deep-vein thrombosis after laparoscopic cholecystectomy. Surg Endosc 1995;9:304–9
32.McCrath DJ, Cerboni E, Frumento RJ, Hirsh AL, Bennett-Guerrero E. Thromboelastography maximum amplitude predicts postoperative thrombotic complications including myocardial infarction. Anesth Analg 2005;100:1576–83
33.Cerutti E, Stratta C, Romagnoli R, Schellino MM, Skurzak S, Rizzetto M, Tamponi G, Salizzoni M. Thromboelastogram monitoring in the perioperative period of hepatectomy for adult living liver donation. Liver Transpl 2004;10:289–94
34.Abrahams JM, Torchia MB, McGarvey M, Putt M, Baranov D, Sinson GP. Perioperative assessment of coagulability in neurosurgical patients using thromboelastography. Surg Neurol 2002;58:5–11
35.Wilson D, Cooke EA, McNally MA, Wilson HK, Yeates A, Mollan RA. Changes in coagulability as measured by thrombelastography following surgery for proximal femoral fracture. Injury 2001;32:765–70
36.Manchio JV, Gu J, Romar L, Brown J, Gammie J, Pierson RN III, Griffith B, Poston RS. Disruption of graft endothelium correlates with early failure after off-pump coronary artery bypass surgery. Ann Thorac Surg 2005;79:1991–8
37.Lijmer JG, Mol BW, Heisterkamp S, Bonsel GJ, Prins MH, van der Meulen JH, Bossuyt PM. Empirical evidence of design-related bias in studies of diagnostic tests. JAMA 1999;282:1061–6
38.Fryback DG, Thornbury JR. The efficacy of diagnostic imaging. Med Decis Making 1991;11:88–94
39.Deville WL, Buntinx F, Bouter LM, Montori VM, de Vet HC, van der Windt DA, Bezemer PD. Conducting systematic reviews of diagnostic studies: didactic guidelines. BMC Med Res Methodol 2002;2:9
40.Walter SD. Properties of the summary receiver operating characteristic (SROC) curve for diagnostic test data. Stat Med 2002;21:1237–56
41.Jones CM, Athanasiou T. Summary receiver operating characteristic curve analysis techniques in the evaluation of diagnostic tests. Ann Thorac Surg 2005;79:16–20
42.Moses LE, Shapiro D, Littenberg B. Combining independent studies of a diagnostic test into a summary ROC curve: data-analytic approaches and some additional considerations. Stat Med 1993;12:1293–316
43.Wells PS, Anderson DR, Rodger M, Forgie M, Kearon C, Dreyer J, Kovacs G, Mitchell M, Lewandowski B, Kovacs MJ. Evaluation of D-dimer in the diagnosis of suspected deep-vein thrombosis. N Engl J Med 2003;349:1227–35
44.O’Donnell J, Riddell A, Owens D, Handa A, Pasi J, Hamilton G, Perry DJ. Role of the Thrombelastograph as an adjunctive test in thrombophilia screening. Blood Coagul Fibrinolysis 2004;15:207–11
45.Schreiber MA, Differding J, Thorborg P, Mayberry JC, Mullins RJ. Hypercoagulability is most prevalent early after injury and in female patients. J Trauma 2005;58:475–80
46.Gibbs NM, Bell R. The effect of low-dose heparin on hypercoagulability following abdominal aortic surgery. Anaesth Intensive Care 1998;26:503–8
47.Goh KY, Tsoi WC, Feng CS, Wickham N, Poon WS. Haemostatic changes during surgery for primary brain tumours. J Neurol Neurosurg Psychiatry 1997;63:334–8
48.Burke GW III, Ciancio G, Figueiro J, Buigas R, Olson L, Roth D, Kupin W, Miller J. Hypercoagulable state associated with kidney-pancreas transplantation. Thromboelastogram-directed anti-coagulation and implications for future therapy. Clin Transplant 2004;18:423–8
49.Viana JS, Pereira MG, Lozano L, Vieira H, Palmeiro A, Lourenço M, Tavares CA, Seco C, Neves S, Bento C, Perdigoto R, Ferrer-Antunes C, Craveiro AL, Furtado L. Thrombelastographic evidence of hyperfibrinolysis during liver transplantation for familial amyloidotic polyneuropathy ATTR met 30. Transplant Proc 2000;32:2645–6
50.Miller BE, Guzzetta NA, Tosone SR, Levy JH. Rapid evaluation of coagulopathies after cardiopulmonary bypass in children using modified thromboelastography. Anesth Analg 2000;90:1324–30
51.Liu G, Bowkett J, Przybylowski G, Bellomo R, McNicol PL. Postoperative fibrinolysis diagnosed by thrombelastography. Anaesth Intensive Care 2000;28:77–81
52.Yamakage M, Tsujiguchi N, Kohro S, Tsuchida H, Namiki A. The usefulness of celite-activated thromboelastography for evaluation of fibrinolysis. Can J Anaesth 1998;45:993–6
53.Nielsen VG. A comparison of the Thrombelastograph and the ROTEM. Blood Coagul Fibrinolysis 2007;18:247–52
54.Lang T, Bauters A, Braun SL, Pötzsch B, von Pape KW, Kolde HJ, Lakner M. Multi-centre investigation on reference ranges for ROTEM thromboelastometry. Blood Coagul Fibrinolysis 2005;16:301–10
55.Jaeschke R, Guyatt G, Sackett DL. Users’ guides to the medical literature. III. How to use an article about a diagnostic test. A. Are the results of the study valid? Evidence-Based Med Working Group. JAMA 1994;271:389–91
56.Lijmer JG, Bossuyt PM, Heisterkamp SH. Exploring sources of heterogeneity in systematic reviews of diagnostic tests. Stat Med 2002;21:1525–37