Can the use of thromboelastography predict and decrease bleeding and blood and blood product requirements in adult patients undergoing cardiac surgery?
Date First Published:
September 22, 2005
Last Updated:
September 22, 2005
Report by:
Andrew Ronald, Consultant (Department of Cardiac Anaesthesia, Aberdeen Royal Infirmary)
Search checked by:
Joel Dunning, Department of Cardiac Anaesthesia, Aberdeen Royal Infirmary
Three-Part Question:
In [adult patients undergoing cardiac surgery], does [Thromboelastography] predict or decrease [bleeding and blood product requirements].
Clinical Scenario:
You start work in a new unit which routinely uses thromboelastography to manage coagulopathy and guide treatment with blood component therapy following surgery. As you have no experience of the technique you decide to review the literature to identify whether the technique is actually beneficial in decreasing exposure to allogeneic blood and blood component therapy.
Search Strategy:
Medline 1966 to June 2005 using the OVID interface and EMBASE 1980 to June 2005.
Search Details:
[CABG.mp OR exp Thoracic Surgery/OR Coronary art$ bypass.mp OR Cardiopulmonary bypass.mp OR exp Cardiovascular Surgical Procedures/OR exp Thoracic Surgical Procedures/OR exp Coronary Artery Bypass/] AND [thromboelastography.mp. OR exp Thrombelastography/OR TEG.mp] AND [bleeding.mp OR platelets.mp OR exp Blood Platelets/OR blood transfusion.mp OR exp Blood Transfusion/OR fresh frozen plasma.mp OR exp Plasma/ OR exp Blood Component Transfusion/OR exp Platelet Transfusion/OR blood component therapy.mp OR exp Erythrocyte Transfusion/].
Outcome:
A total of 170 papers were identified using the reported search of which 14 represented the best evidence on the subject. These studies are summarised in the table
Relevant Paper(s):
| Study Title | Patient Group | Study type (level of evidence) | Outcomes | Key results | Study Weaknesses |
|---|---|---|---|---|---|
| Thromboelastography as an indicator of post-cardiopulmonary bypass coagulopathies. Spiess BD, Tuman KJ, McCarthy RJ, DeLaria GA, Schillo R, Ivankovich AD. 1987, USA | 38 patients undergoing Cardiac Surgery (29 CABG, 9 valve replacements) Activated clotting times (ACT), Thromboelastography (TEG) and coagulation profiles performed pre-bypass, 45 min after institution of CPB and 30 min post-protamine Postoperative blood loss (1 ml/kg/h in first 8 h classed as abnormal) |
Cohort Study (level 2b) | Postop blood loss | Significant correlation between aPTT & TEG | Small study 1 patient returned to theatre with arterial bleeding from aortotomy site and normal TEG and excluded from further analysis |
| Correlation between TEG variables, ACT's and coagulation profiles, and comparison to postop blood loss to identify predictive capabilities of each test | All TEG's abnormal during CPB (no evidence coagulation)<br><br>29/37 (78%) TEG's normal post-Protamine<br><br>- no single coagulation test consistently abnormal<br><br>No correlation between abnormal TEG's and single abnormal coagulation test (moderate correlation between A60 and aPTT - r=–0.4708, P=0.01)<br><br>TEG 87% accurate as predictor of postoperative bleeding (ACT 30% and coagulation profiles 51% accurate, P=0.004) | ||||
| Plateletworks platelet function test compared to the thromboelastograph for prediction of postoperative outcomes. Ostrowsky J, Foes J, Warchol M, Tsarovsky G, Blay J. 2004, USA | 35 adults patients undergoing cardiac surgery with CPB Excluded if history of haemostatic disorder or requiring deep hypothermic circulatory arrest Plateletworks platelet function analyser (platelet count + platelet aggregation in presence of agonist) vs thromboelastography (TEG) Platelet assays - preincision, after removal of aortic cross-clamp, 1 h and 24 h postop using baseline, ADP and collagen reagent tubes TEG assays preop, post-protamine and 24 h postop |
Cohort Study (level 2b) | Plateletworks data | No correlation between plateletworks and blood product usage | Wide range of procedures CABG, Valve or combination Valve or combination of CPB/cross-clamp times and lowest core temperatures |
| TEG data | Significant change in alpha -angle from preop to postop TEG samples (p<0.035) | ||||
| Blood product use | No significant difference in k-time and MA at the time intervals | ||||
| Chest tube (CT) drainage | No correlation between TEG data and CT drainage<br><br>Correlation between preop TEG MA and blood product usage (P=0.016) | ||||
| Does the platelet-activated clotting test (HemoSTATUS) predict blood loss and platelet dysfunction associated with cardiopulmonary bypass? Ereth MH, Nuttall GA, Klindworth JT, MacVeigh I, Santrach PJ, Orszulak TA, Harmsen WS, Oliver WC Jr. 1997, USA | 200 adult patients undergoing cardiac surgery with CPB Comparison of platelet-activated clotting test (PACT HemoSTATUS)- vs ACT and clotting studies (PT and APTT) vs TEG to predict blood loss and platelet dysfunction associated with CPB Coagulation studies, platelet tests and TEG taken pre and ost CPB; ACT samples taken post-heparin and on CPB; PACT performed pre-induction, after 40 min CPB and 20-40 min post-protamine Sensitivity, specificity of tests analysed using receiver operating characteristics (ROC) analysis of system using excessive blood loss as abnormal state |
Cohort Study (level 2b) | Mediastinal blood loss (MBL) at 4 and 24 h | Post-protamine PACT correlated with 4-h MBL MBL (r=–0.30; P=0.014) | Design of study to assess PACT, HemoSTATUS test, not TEG Design of study specifically to correlate platelet function to bleeding Only small sub-group did not receive aprotinin |
| Abnormal blood loss level set at: >100 ml/h in 1st 4 h | TEG MA correlated with 4 h MBL (r=–0.32; P=0.003) | ||||
| Abnormal blood loss level set at: >200 ml/h in 1st 4 h | PACT sensitivity and specificity comparable to conventional coagulation tests in predicting blood loss | ||||
| Platelet activated clotting test (PACT) | TEG MA more predictive than PACT and routine coagulation tests in predicting excessive blood loss post-CPB | ||||
| Platelet count<br><br>PT;aPTT<br><br>TEG-R-Time;R+K-time; alpha-angle; MA; MA+30 | >100 ml/h blood loss in 4 h:<br><br>– PACT 80% – sensitivity 80%; specificity 56%; – TEG-MA 48 mm – sensitivity 80%; specificity 75%; – Platelet count 110x10(to the power of 9)/l – sensitivity 64%; specificity 61%; – PT 14 s – sensitivity 68%; specificity 68%; – aPTT 40 s – sensitivity 50%; specificity 55%; | ||||
| Comparison of thromboelastography to bleeding time and standard coagulation tests in patients after cardiopulmonary bypass. Essell JH, Martin TJ, Salinas J, Thompson JM, Smith VC. 1993, USA | 36 adult patients undergoing cardiopulmonary bypass Comparison of TEG to platelet function tests (bleeding time, platelet count, mean platelet volume) and standard coagulation tests (PT, aPTT, fibrinogen) Abnormal blood loss defined as loss >1500 ml blood in first 24 h postop or need to transfuse platelets of FFP to control haemorrhage (platelet of FFP ordered on basis of lab tests or clinical decision) |
Cohort Study (level 2b) | Patients classed as 'bleeders' or 'non-bleeders' on basis of mediastinal tube (MT) or chest tube (CT) drainage<br><br>Preop and postop bleeding time (BT)<br><br>Preop and postop PT;aPTT: platelet count; MPV; Hb; Hct; fibrinogen; fibrinogen split products (FSP's); thrombin time<br><br>TEG data – R-time; K-time; alpha-angle; MA; A60 (normal values derived from review of 'bleeders' and 'non-bleeders' on basis of mediastinal tube (MT) or chest tube (CT) drainage<br><br>Preop and postop bleeding time(BT)<br><br>Preop and postop PT; aPTT: platelet count; MPV; Hb; Hct; fibrinogen; fibrinogen split products (FSP's); thrombin time<br><br>TEG data – R-time; K-time; alpha-angle; MA; A60 normal values derived from review of 'bleeders' and 'non-bleeders' | TEG – sensitivity 71.4%; specificity 89.3%;<br><br>Bleeding time >9 min – sensitivity 71.4%; specificity 78.5%;<br><br>Platelet count <130x10 (to the power of 9) – sensitivity 100%; specificity 53.6%;<br><br>PT >13.55 s – sensitivity<br><br>85.7%; specificity 10.7%;<br><br>APTT>30.85 s – sensitivity 85.7%; specificity 21.4%;<br><br>Fibrinogen <175 – sensitivity 85.7%; specificity 32.1%;<br><br>Thrombin time <9.95 s –sensitivity 28.6%; specificity 78.6%; Hb <12 g/dl – sensitivity 85.7%; specificity 7.1%; | Small study 61% of patients having CABG surgery. Remainder mix of valve, ASD repair, bundle ablations – wide range of procedures Assessing TEG as predictor of bleeding rather than to guide to therapy Unclear whether 'bleeders' and 'non-bleeders' used to define TEG normal values were from study or other population Blood loss seems excessive compared to other studies 1/36 died 8 h postop 35/36 completed study |
| The predictive value of modified computerized thromboelastography and platelet function analysis for postoperative blood loss in routine cardiac surgery. Cammerer U, Dietrich W, Rampf T, Braun SL, Richter JA. 2003, Germany | 255 consecutive patients Comparison of modified TEG (ROTEG) using activated blood samples ± Abciximab vs platelet function analyser (PFA-100) using epinephrine and ADP coated cartridges post-induction of anaesthesia, during CPB after rewarming and post-protamine administration to predict postop blood loss Data for blood loss >750 ml in 6 h compared to group with blood loss <750 ml in 6 h Data for blood loss above and below 75th percentile for 6 h blood - 'enhanced blood loss' - also compared ROC curves used to identify/compare best predictors Significant variables with univariate analysis subjected to binary logistic regression analysis |
Cohort Study (level 2b) | Absolute blood loss | Blood loss >750 ml vs loss <750 ml | Bleeding threshold of 750 ml based on one other study and may be somewhat arbitary 75th percentile more useful Demonstrates low PPV and high NPV Low PPV means that coagulopathy identified by test may not necessarily cause bleeding – risk of over -treating if test result solely used to guide therapy |
| Comparison of ROTEG and PFA-100 data for blood loss groups >750 ml and <750 ml in 6 h | – TEG data – no significant difference – PFA data – significant differences for preop ADP PFA vs postop ADP PFA and epinephrine PFA (P<0.05) | ||||
| Comparison of ROTEG and PFA-100 data for blood loss groups more and less than 75th percentile in 6 h | Blood loss >500 ml vs. loss <500 ml – significant differences between ROTEG MA's on CPB and post-CPB (± Abciximab); ROTEG post-CPB; alpha-angle (± Abciximab); and post-CPB PFA's with ADP and epinephrine (P<0.05)<br><br>Post-protamine post-CPB samples best predictors for bleeding with univariate analysis – MA <57 mm – PPV 37%; NPV 77% – alpha-angle <71° – PPV 41%; NPV 82%– Abciximab MA <17 mm – PPV 41%; NPV 85% – Abciximab -angle <26° – PPV 35%; NPV 84% – PFA ADP <118 s – PPV 34%; NPV 76%<br><br>-angle of TEG best predictor of postop bleeding on logistic regression analysis with PFA-ADP post-CPB an additional significant factor | ||||
| Prediction of excessive bleeding after coronary artery bypass graft surgery: the influence of timing and heparinase on thromboelastography. Ti LK, Cheong KF, Chen FG. 2002, Singapore | 40 patients undergoing elective multivessel CABG surgery TEG samples (plain and heparinase-modified) taken pre-induction and at 10 min and 60 min post-protamine administration Subdivided into 'bleeders' (blood loss>1000 ml over 24 h, or 250 ml over 2 consecutive hours) and 'non-bleeders' |
Cohort Study (level 2b) | Abnormal TEG values:– K-time >16 mm – {alpha}-angle <40° – MA<40 mm – LY60 >15% in MA | Moderate correlation between TEG parameters, total blood loss & requirement for FFP &/or platelets<br><br>60 min TEG better than 10 min<br><br>TEG (sensitivity 100% vs. 70%) | Definition of degrees of postop bleeding somewhat arbitrary and based on 'local institutional modifications of definitions by previous authors' though no references cited 6/40 (15%) 10-min and 10/40 (25%) 60-min TEG's produced unreadable 'straight-line' traces and were excluded from analysis Small numbers in study |
| Absolute blood loss<br><br>Requirement for FFP and/or platelet transfusion | Use of heparinase- modified-TEG specificity (83% vs.40% @10 min, 73% vs. 20% @ 60 min); and PPV (58% vs. 28% @ 10 min, 55% vs. 29% @ 60 min)<br><br>'Bleeder' vs. 'non-bleeder' lab tests no significant difference | ||||
| Coagulation tests predict bleeding after cardiopulmonary bypass. Nuttall GA, Oliver WC, Ereth MH, Santrach PJ. 1997, USA | 82 adult patients undergoing elective cardiac surgery Agreed subjective assessment by anaesthetist and surgeon into 'bleeders' and 'non-bleeders' 10 min post-protamine Quantitative assessment of bleeding post-CPB by collection of blood and irrigation fluid by suction from operative field Comparison of coagulation tests, bleeding times, platelet counts, TEG and sonoclot data as predictors of bleeding |
Cohort Study (level 2b) | Identification of 'bleeders' | TEG data - – R time – 18 vs. 16 mm. (P=0.1475)) – sensitivity 47%; specificity 71%; | Diagnosis of bleeding subjective (but 'agreed' between surgeon and anaesthetist ROC analysis based on this subjective designation of 'bleeder' Lack of 'gold-standard' for diagnosis of bleeding may affect the ROC analysis |
| Thromboelastography – post-Protamine R & K times; alpha-angle; MA; MA+30 (A30)<br><br>Assessment of sensitivity, specificity, PPV and NPV achieved using Receiver Operator Characteristic (ROC) curves for bleeding | – R+K time – 29 vs. 23 mm P=0.0123) – sensitivity 53%; specificity 75%;<br><br>– R+K time – 29 vs. 23 mm (P=0.0123) – sensitivity 53%; specificity 75%;<br><br>alpha-angle – 40 vs. 46° (P=0.0138) – sensitivity 63%; specificity 71%;<br><br>– MA – 47 vs. 53 mm (P=0.0006) – sensitivity 60%; specificity 79%;<br><br>– MA+30 – 45 vs. 49 mm P=0.0614) – sensitivity 58%; specificity 67%; 30/82<br><br>36.6%) classed as 'bleeders' 10 min post-protamine<br><br>Median blood loss ('bleeders vs 'non-bleeders'): – 775 vs. 200 ml intraoperative (P<0.001) – 949 vs. 547 ml first 24 h postop (P<0.001)<br><br>Coagulation tests correlated best with intraoperative and postoperative blood loss (but predictive values outside normal ranges) | ||||
| Identification of patients at risk for excessive blood loss during coronary artery bypass surgery: thromboelastography vs. coagulation screen. Dorman BH, Spinale FG, Bailey MK, Kratz JM, Roy RC. 1993, USA | 60 patients presenting for CABG surgery Comparison of preoperative tests to predict blood loss: - Coagulation screen (aPTT, PT, platelet count, fibrinogen level; bleeding time) vs -ACT -TEG Blood/blood products given on basis of transfusion protocol |
Cohort Study (level 2b) | Absolute intraoperative blood loss<br><br>Packed cell and blood product requirements; Hb; Hct;Coag screen results –PT; aPTT; fibrinogen; platelet count; forearm bleeding time<br><br>ACT results<br><br>TEG results – R time; R+K time; MA; MA+30; MA+60; propto angle | PT; platelet count; bleeding time identified as significant contributing variables when blood loss considered dependant variable (r=0.75; P<0.05)<br><br>No significant relationship between any TEG variable and blood loss by step-wise linear regression analysis (r<0.25, P>0.78)<br><br>Significant linear relationship between propto-angle and MA(r=–0.54; P<0.05), propto-angle and R-time (r=–0.53; P<0.05),and propto-angle and R+K-time (r=–0.7; P<0.05)<br><br>Significant correlation between propto-angle and PT (r=–0.58; P<0.05). Poor correlation between other coagulation screen parameters and TEG variables | Looked at preoperative coagulation studies; ACT and TEG data only as predictors of intraoperative blood loss 2/60 TEG's repeated post-induction in viewof unsatisfactory pre-induction traces |
| Comparison of structured use of routine laboratory tests or near-patient assessment with clinical judgement in the management of bleeding after cardiac surgery. Avidan MS, Alcock EL, Da Fonseca J, Ponte J, Desai JB, Despotis GJ, Hunt BJ. 2004, USA/UK | 102 patients undergoing elective CABG surgery with cardiopulmonary bypass 'Point of Care' (POC) - hepcon + Thromboelastography (TEG) + platelet function analyser (PFA-100) - vs 'laboratory algorithm group' (LAG) management - rapidly available ACT, INR, APTT - to direct treatment algorithm POC & LAG transfusion trigger 8 g/dl Comparison to retrospective case-control group (n=108) - 'clinician directed therapy' |
PRCT with retrospective cohort control group (level 2b) | 24-h Blood loss | Median blood loss between two study and case-control and study groups not significant | Study assessed effects of combined POC monitors, not just TEG Variations in antifibrinolytic therapy between groups (tested groups given tranexamic acid; Case control group mostly given tranexamic acid or aprotinin at clinicians discretion Study questions accuracy of its own POC algorithm Time delays obtaining conventional clotting studies compared to rapidly available POC and LAG data might skew outcomes No transfusion trigger in case-control group |
| Blood transfusion requirement<br><br>Blood component requirement | Blood and component transfusion (Red cells, FFP and Platelets) in 'clinically directed' group > that in LAG or POCgroups (P<0.025)<br><br>No significant difference between POC and LAG groups –therefore POC no better than rapidly available coag tests<br><br>Abnormal TEG MA associated with increased postop bleeding. No results predicted increase in postop bleeding | ||||
| Changes in transfusion therapy and re-exploration rate after institution of a blood management program in cardiac surgical patients. Spiess BD, Gillies BS, Chandler W, Verrier E. 1995, USA | 1079 sequential patients for revascularisation and/or open ventricle procedures 2 groups Group 1 - 488 pre introduction of TEG based guidelines Group 2 - 591 post introduction of TEG based guidelines Baseline and post-protamine TEG analysis performed |
Cohort study (level 2b) | Number of pts transfused | – 13.7% pre-TEG vs.21.5% post-TEG received no transfusion (P=0.001) | Increased awareness of transfusion issues and philosophy and move towards less liberal transfusion and stricter transfusion triggers probably main contributor to decreased red cell transfusion during study. Introduction of TEG at the same time may have been coincidental Lack of rigid TEG protocol use and interpretation of TEG data not standardised and left up to individual clinicians Other coagulation data also available to clinical team if deemed necessary Much of study may show advantages of cooperative behaviour in monitoring changes rather than true experimental differences between groups |
| Number receiving platelet transfusion | – 38.8% pre-TEG vs.42.1% post-TEG received no transfusion in operating room (P=0.005) | ||||
| Number receiving FFP transfusion | – 83.2% pre-TEG vs. 73.9% post-TEG received red cell transfusion overall (P=0.0001) | ||||
| Number receiving cryoprecipitate transfusion | – 59.2% pre-TEG vs. 48.2% post-TEG received Platelets (P=0.0001) – 36.1% pre-TEG vs. 26.4% post-TEG received FFP (P=0.001) | ||||
| Incidence of massive transfusion | – 9.1% pre-TEG vs. 6.4% post-TEG received cryoprecipitate (P=0.091) – 10.2% pre-TEG vs. 9.5% transfusion (P=NS) | ||||
| Re-exploration rate | – 5.7% pre-TEG vs. 1.5% TEG-monitored patients required re- exploration (P=0.0001) | ||||
| Evaluation of post-cardiopulmonary bypass coagulation disorders by differential diagnosis with a multichannel modified thromboelastogram: a pilot investigation. Koster A, Kukucka M, Fischer T, Hetzer R, Kuppe H. 2001, USA | 30 adult patients undergoing surgery with cardiopulmonary bypass with prolonged ACT> 150 s after protamine reversal of heparin Treatment protocol based on analysis of data from multiple kaolin-activated TEG's: -Baseline -Heparinase -Heparinase/Abciximab -Heparinase/FFP-TEG ROTEG R-time <900 s and MA 40-50mm regarded as normal |
Case series (level 3b) | experimental differences and MA variables<br><br>Pre-treatment ACT<br><br>Post treatment ACT | Therapy guided by ROTEG<br><br>2/30 (6.7%) given protamine<br><br.5/30 (16.7%) given desmopressin<br><br>19/30 (63.3%) given FFP<br><br>2/30 (6.7%) given platelets<br><br>2/30 (6.7%) given platelets + FFP<br><br>Mean pre treatment ACT 162.2±7.8 s<br><br>Mean post-treatment ACT 127±8.3 s (P<0.05) | Small study No comparison to control group Mixed group of patients – CABG ± valve; multiple valves; aortic surgery; VAD implantation Mixed group of interventions Treatment algorithm more complicated than that used in other studies |
| Thromboelastography-guided transfusion algorithm reduces transfusions in complex cardiac surgery. Shore-Lesserson L, Manspeizer HE, DePerio M, Francis S, Vela-Cantos F, Ergin MA. 1999, USA | 107 adult patients undergoing cardiac surgery utilising cardiopulmonary bypass (105 completed study) Conventional vs TEG protocol in presence of bleeding 53/105 TEG protocol – protamine, platelets, FFP, EACA, cryoprecipitate given according to TEG data (R-time, MA and LY30), plateletcount and fibrinogen levels 52/105 Conventional protocol – protamine, platelets, FFP, and cryoprecipitate given according to ACT, plateletcount, PT and fibrinogen levels |
PRCT (level 1b) | Multiple time point analysis of: – Mediastinal tube drainage (MTD)<br><br>-Blood component transfusion requirements<br><br>-Lab coag tests (platelets, PT, aPTT, fibrinogen)<br><br>-TEG variables (R-time, alpha-angle, MA and LY30) | 34/52 (65.4%) conventional group vs 22/53 (41.5%) of TEG group transfused (P=0.01)<br><br>17/52 (33%) Conventional group vs. 7/53 (13%) TEG group received non-RBC transfusion (P=0.02)<br><br>16/52 (31%) Conventional group received FFP vs. 4/53 (7.5%) of TEG group (P<0.002)<br><br>FFP volume significantly more in conventional than TEG group (P<0.04)<br><br>15/52 (29%) Conventionalvs. 7/53 (13%) in TEG group received platelets (P<0.05). No significant difference in MTD between groups | Routine use of E-aminocaproic acid may modify TEG and make it more likely to trigger intervention Part of TEG protocol dependent on lab studies More options as part of TEG protocol Post-protamine lab tests used to guide 'Conventional protocol' vs. 'post-warm CPB heparinase modified samples' in TEG group. Ordering of component therapy based on this TEG sample but only actually given if patient bleeding therefore facilitated very early intervention for bleeding post-CPB |
| Reduced haemostatic factor transfusion using heparinase-modified thrombelastography during cardiopulmonary bypass. Royston D, von Kier S. 2001, UK | Series 1 60 patients undergoing complex cardiac surgery given haemostatic treatment at clinician's discretion/lab tests in presence of bleeding vs 'predicted' component requirements guided by heparinase modified celite activated TEG algorithm Series 2 60 patients randomised to 'clinician directed/lab tests' or TEG-directed component therapy using algorithm driven by R-time, MA and LY30 data. Products ordered as soon as clinicians judged necessary or test results available |
Series 1 -retrospective cohort study (level 2b) Series 2 -PRCT (level 1b) |
Mediastinal tube drainage (MTD) at 6 and 12 h post-op<br><br>Red cell usage<br><br>FFP usage<br><br>Platelet usage | Series 1 - 22/60 (37%) given blood components vs 7/60 (12%) 'predicted' to require component therapy (P<0.05)<br><br>38 units FFP and 17 platelet pools given vs 6 units FFP and 2 units pooled platelets in 'predicted' group (P<0.05)<br><br>Series 2 - 10/30 (33%) clinically managed patients received blood components vs 5/30 (17%) TEG patients (P<0.05)<br><br>16 units FFP and 9 platelet pools in clinically managed group vs 5 units FFP and 1 unit pooled platelets in TEG group (P<0.05)<br><br>MTD no significant differences between groups | Patients who returned to theatre for bleeding discarded from study and replaced by additional patient TEG algorithm facilitated early intervention possibly before associated with excessive bleeding Could TEG give false positive results leading to unecessary treatment? Algorithm issues - study acknowledges that absolute values used in algorithm are derived from reference values in normal population and there is no evidence that they are optimal for study group |
Author Commentary:
Thromboelastography (TEG) is a point-of-care whole blood coagulation monitor which provides information on specific aspects of coagulation including time to production of initial fibrin strands (R-time), time to develop clot (R-time, K-time), rate of fibrin build-up and cross linking (-angle), maximum clot strength (maximum amplitude–bMA) and measures of fibrinolysis (decreasing amplitude post-MA).
Abnormal TEG data may predict patients who will bleed. Spiess [2] found that TEG correlated well with ACT and coagulation profiles and whilst no coagulation test was consistently abnormal the TEG was the most accurate predictor of bleeding.
Ereth studied a 'Platelet-activated clotting test' (PACT HemoSTATUS), ACT and clotting studies, and TEG. Whilst PACT sensitivity and specificity was comparable to conventional coagulation tests in predicting blood loss, TEG was more predictive at both blood loss levels. Essell found that whilst the bleeding time and platelet count had sensitivities similar to the TEG, TEG specificity was greater. In addition, they suggested that patients with an abnormal TEG were at increased risk of bleeding but that excessive bleeding in the face of a normal TEG implied surgical bleeding and FFP and platelets should not simply be used empirically. Ti found moderate correlation between TEG parameters, total blood loss and requirements for FFP and/or platelets in their group of 'bleeders'.
In contrast, other studies did not find the TEG to be a useful predictor of blood loss. Nuttall reported that TEG values had a low sensitivity and specificity in predicting 'bleeders'. Dorman compared preoperative coagulation screens to ACTs and TEGs as predictors of blood loss but found no significant relationship between any TEG variable and losses.
A number of studies have used the TEG to guide transfusion management. Avidan Compared TEG to a Laboratory based algorithm. They concluded that whilst blood and blood product usage was significantly more in the laboratory group, there was no statistically significant difference between the study groups. Spiess analysed 1079 patients before and after the introduction of TEG as part of an overall transfusion management strategy. They identified significant changes in their practice with decreased usage of all blood and blood component therapies with the exception of cryoprecipitate. Their re-exploration rate also fell significantly. However, this study probably reflects the effects of education and co-operative behaviour in monitoring changes rather than a true experimental difference between groups.
Two randomised controlled trials have been performed. Shore-Lesserson compared 'TEG-based' and 'conventional' protocols to manage postoperative bleeding. Whilst there was no significant difference in mediastinal tube drainage between the groups, blood and blood component therapy was significantly less in the 'TEG' than in the 'conventional group'. However, the 'TEG protocol' did have more options than the conventional protocol and was also partly dependent on laboratory tests. In addition, blood products were ordered on the basis of a TEG taken at rewarm on cardiopulmonary bypass and given in the presence of continued bleeding following protamine, whereas, the conventional group required post-protamine tests to dictate intervention. This inevitably meant earlier intervention in the TEG group. Royston studied 60 patients who had undergone complex surgery comparing their actual blood/blood product use to a 'predicted usage' derived from a TEG-based algorithm. 'Predicted' blood/blood component transfusion was significantly less than 'actual' transfusion. They subsequently used this algorithm comparing it to conventional management in a further 60 patients. Again they demonstrated significantly less blood/blood component usage in the TEG-based group compared to the conventional 'clinician-directed' group with no excessive mediastinal bleeding. However, this study was designed to identify TEG-evidence of coagulation before physical evidence of microvascular bleeding and the authors acknowledge the fact that their protocol allowed much earlier intervention in the active than in the control limb.
A recent review by Samama has raised concerns that thromboelastography remains an unvalidated technique which fails to achieve the stringent standard quality-control procedures essential in laboratory-based tests, citing absence of a formal standard operating procedure taking into account factors such as gender and pregnancy differences, stability of blood samples, and sampling site. There is also no standardised technique and multiple modifications exist including plain versus heparinase samples; celite, kaolin or tissue factor activation; abximimib (Reopro) modified; modified multi-channel; and ROTEG have been described. Several studies acknowledge that TEG facilitates earlier intervention than standard coagulation tests [Avidan, Shore-Lesserson, Royston] thus making true comparisons difficult. Samama et al. conclude their review by suggesting that extended collaborative studies involving haematologists are required to further evaluate and validate thromboelastography.
Abnormal TEG data may predict patients who will bleed. Spiess [2] found that TEG correlated well with ACT and coagulation profiles and whilst no coagulation test was consistently abnormal the TEG was the most accurate predictor of bleeding.
Ereth studied a 'Platelet-activated clotting test' (PACT HemoSTATUS), ACT and clotting studies, and TEG. Whilst PACT sensitivity and specificity was comparable to conventional coagulation tests in predicting blood loss, TEG was more predictive at both blood loss levels. Essell found that whilst the bleeding time and platelet count had sensitivities similar to the TEG, TEG specificity was greater. In addition, they suggested that patients with an abnormal TEG were at increased risk of bleeding but that excessive bleeding in the face of a normal TEG implied surgical bleeding and FFP and platelets should not simply be used empirically. Ti found moderate correlation between TEG parameters, total blood loss and requirements for FFP and/or platelets in their group of 'bleeders'.
In contrast, other studies did not find the TEG to be a useful predictor of blood loss. Nuttall reported that TEG values had a low sensitivity and specificity in predicting 'bleeders'. Dorman compared preoperative coagulation screens to ACTs and TEGs as predictors of blood loss but found no significant relationship between any TEG variable and losses.
A number of studies have used the TEG to guide transfusion management. Avidan Compared TEG to a Laboratory based algorithm. They concluded that whilst blood and blood product usage was significantly more in the laboratory group, there was no statistically significant difference between the study groups. Spiess analysed 1079 patients before and after the introduction of TEG as part of an overall transfusion management strategy. They identified significant changes in their practice with decreased usage of all blood and blood component therapies with the exception of cryoprecipitate. Their re-exploration rate also fell significantly. However, this study probably reflects the effects of education and co-operative behaviour in monitoring changes rather than a true experimental difference between groups.
Two randomised controlled trials have been performed. Shore-Lesserson compared 'TEG-based' and 'conventional' protocols to manage postoperative bleeding. Whilst there was no significant difference in mediastinal tube drainage between the groups, blood and blood component therapy was significantly less in the 'TEG' than in the 'conventional group'. However, the 'TEG protocol' did have more options than the conventional protocol and was also partly dependent on laboratory tests. In addition, blood products were ordered on the basis of a TEG taken at rewarm on cardiopulmonary bypass and given in the presence of continued bleeding following protamine, whereas, the conventional group required post-protamine tests to dictate intervention. This inevitably meant earlier intervention in the TEG group. Royston studied 60 patients who had undergone complex surgery comparing their actual blood/blood product use to a 'predicted usage' derived from a TEG-based algorithm. 'Predicted' blood/blood component transfusion was significantly less than 'actual' transfusion. They subsequently used this algorithm comparing it to conventional management in a further 60 patients. Again they demonstrated significantly less blood/blood component usage in the TEG-based group compared to the conventional 'clinician-directed' group with no excessive mediastinal bleeding. However, this study was designed to identify TEG-evidence of coagulation before physical evidence of microvascular bleeding and the authors acknowledge the fact that their protocol allowed much earlier intervention in the active than in the control limb.
A recent review by Samama has raised concerns that thromboelastography remains an unvalidated technique which fails to achieve the stringent standard quality-control procedures essential in laboratory-based tests, citing absence of a formal standard operating procedure taking into account factors such as gender and pregnancy differences, stability of blood samples, and sampling site. There is also no standardised technique and multiple modifications exist including plain versus heparinase samples; celite, kaolin or tissue factor activation; abximimib (Reopro) modified; modified multi-channel; and ROTEG have been described. Several studies acknowledge that TEG facilitates earlier intervention than standard coagulation tests [Avidan, Shore-Lesserson, Royston] thus making true comparisons difficult. Samama et al. conclude their review by suggesting that extended collaborative studies involving haematologists are required to further evaluate and validate thromboelastography.
Bottom Line:
Thromboelastography can be used to predict bleeding in cardiac surgery, but it can also be used to guide transfusion therapy during postoperative bleeding using appropriate treatment algorithms where its use has been associated with significant decreases in blood and blood component transfusion. However, thromboelastography remains unvalidated compared to other laboratory-based routine coagulation studies and further large controlled studies involving haematology input are required to confirm that its use can be extrapolated to all types of cardiac surgery and also to define the 'ideal' treatment algorithm.
References:
- Spiess BD, Tuman KJ, McCarthy RJ, DeLaria GA, Schillo R, Ivankovich AD.. Thromboelastography as an indicator of post-cardiopulmonary bypass coagulopathies.
- Ostrowsky J, Foes J, Warchol M, Tsarovsky G, Blay J.. Plateletworks platelet function test compared to the thromboelastograph for prediction of postoperative outcomes.
- Ereth MH, Nuttall GA, Klindworth JT, MacVeigh I, Santrach PJ, Orszulak TA, Harmsen WS, Oliver WC Jr.. Does the platelet-activated clotting test (HemoSTATUS) predict blood loss and platelet dysfunction associated with cardiopulmonary bypass?
- Essell JH, Martin TJ, Salinas J, Thompson JM, Smith VC.. Comparison of thromboelastography to bleeding time and standard coagulation tests in patients after cardiopulmonary bypass.
- Cammerer U, Dietrich W, Rampf T, Braun SL, Richter JA.. The predictive value of modified computerized thromboelastography and platelet function analysis for postoperative blood loss in routine cardiac surgery.
- Ti LK, Cheong KF, Chen FG.. Prediction of excessive bleeding after coronary artery bypass graft surgery: the influence of timing and heparinase on thromboelastography.
- Nuttall GA, Oliver WC, Ereth MH, Santrach PJ.. Coagulation tests predict bleeding after cardiopulmonary bypass.
- Dorman BH, Spinale FG, Bailey MK, Kratz JM, Roy RC.. Identification of patients at risk for excessive blood loss during coronary artery bypass surgery: thromboelastography vs. coagulation screen.
- Avidan MS, Alcock EL, Da Fonseca J, Ponte J, Desai JB, Despotis GJ, Hunt BJ.. Comparison of structured use of routine laboratory tests or near-patient assessment with clinical judgement in the management of bleeding after cardiac surgery.
- Spiess BD, Gillies BS, Chandler W, Verrier E.. Changes in transfusion therapy and re-exploration rate after institution of a blood management program in cardiac surgical patients.
- Koster A, Kukucka M, Fischer T, Hetzer R, Kuppe H.. Evaluation of post-cardiopulmonary bypass coagulation disorders by differential diagnosis with a multichannel modified thromboelastogram: a pilot investigation.
- Shore-Lesserson L, Manspeizer HE, DePerio M, Francis S, Vela-Cantos F, Ergin MA.. Thromboelastography-guided transfusion algorithm reduces transfusions in complex cardiac surgery.
- Royston D, von Kier S.. Reduced haemostatic factor transfusion using heparinase-modified thrombelastography during cardiopulmonary bypass.
- Samama CM, Ozier Y.. Near-patient testing of haemostasis in the operating theatre: An approach to appropriate use of blood in surgery.