Department of Anaesthesia and Critical Care, Royal Brompton and Harefield NHS Trust, Harefield Hospital, Harefield, Middlesex UB9 6JH, UK*Corresponding author
Accepted December 7, 2000
![]() |
Abstract |
---|
![]() ![]() ![]() ![]() ![]() ![]() |
---|
Br J Anaesth 2001; 86: 5758
Keywords: surgery, cardiac; blood, coagulation
![]() |
Introduction |
---|
![]() ![]() ![]() ![]() ![]() ![]() |
---|
The thrombelastogram measures the rate and strength of clot formation.1 Initiation of clot formation is defined as the reaction or r-time and the strength of the clot by the maximum amplitude of the trace. During cardiopulmonary bypass, the heparinase-modified thrombelastogram will develop despite anticoagulation with heparin in doses of 300 IU kg1 and gives the same results as those obtained using blood when heparin has been antagonized with protamine.2
Previous reports3 4 have described algorithms that reduce the need for haemostatic blood component therapy and re-exploration. However, these studies waited until microvascular bleeding occurred before starting testing and intervention.
In this pilot study we investigated the predictive value and use of an algorithm using thrombelastogram measurements made during heart surgery using anticoagulated blood. The principal end-point for efficacy was reduced total exposure to haemostatic component therapies.
![]() |
Methods and results |
---|
![]() ![]() ![]() ![]() ![]() ![]() |
---|
The study plan allowed data from patients who returned to theatre for the control of surgical bleeding or who died within 48 h of the surgery to be discarded and replaced by measurements from an additional patient allocated to the same group. Three patients in the first series of patients and two in the second series (one each in the algorithm and control groups) returned to theatre. All these patients had drain losses of >400 ml h1 and a thrombelastogram result within the normal range. No patient died during the first 48 h after operation.
Patient details were not significantly different between groups but showed a wide distribution of values for age (2183 yr) and bypass time (48167 min).
We made standard thrombelastogram measurements (TEG®; Haemoscope Corporation, Niles, IL, USA) in 360 µl of whole blood activated with 1% Celite. Samples were taken and analysed at the following times. (i) Baseline, after induction of anaesthesia. This sample was to exclude any pre-operative coagulopathy. (ii) Bypass. This sample was taken during rewarming when the bypass system venous blood temperature was 35°C, or within 30 min of separation from bypass in patients when tepid bypass was used (nasopharyngeal temperature
34°C). The sample was developed in a cuvette containing heparinase. The result was used to define the need for haemostatic components using the algorithm shown in Table 1. (iii) Ten to fifteen minutes after protamine. The thrombelastogram was developed with and without heparinase, to detect any residual heparin effect. Some of this blood was sent for laboratory measurement of prothrombin time, activated partial thromboplastin time (APTT), platelet count and fibrinogen concentration.
|
Statistical and power analysis was performed using GB-StatTM version 6.5 for Macintosh (Dynamic Microsystems, Silver Spring, MD, USA) on a PowerMac. Statistical differences between groups were defined using the 2 test.
Series one
In the first 60 patients, haemostatic treatment (fresh frozen plasma and platelets) was determined using criteria obtained from conventional laboratory tests (prothrombin time and/or APTT ratio >1.5 times control, platelet count <50 000 dl1, fibrinogen concentration 0.8 mg dl1) in the presence of microvascular bleeding (>400 ml in the first hour after surgery or >100 ml h1 for four consecutive hours). We noted the number of component units actually given, and the amount predicted from the thrombelastogram algorithm. This was done so that the method could be used to estimate the likely reduction in component transfusion, and to define the size of sample in the subsequent series after power analysis.
On the basis of clinical assessment and laboratory tests, 22 of 60 patients received blood components. If we had used the treatment algorithm, the number of patients receiving component therapy would have been seven (P<0.05). The predicted number of components used (six units of fresh frozen plasma and two pools of platelets) would have been significantly less than the 38 units of fresh frozen plasma and 17 platelet pools actually transfused in these 60 patients (P<0.05).
Power analysis suggested that 27 patients per group would be required to demonstrate a fourfold reduction in component use with an of 0.05 and a power of >90%.
Series two
In the second part of the study, 60 patients were randomly allocated to have products ordered and administered on the basis of either the algorithm (group T) or the wishes of the clinician (group C). Patients were allocated to groups by means of a series of sealed envelopes. Haemostatic products were ordered and given as soon as test results were known, or as judged clinically by the surgeon or anaesthetist responsible for the case. The amounts of products given and volumes of chest drainage 6 and 12 h after surgery were recorded.
Ten of the 30 clinically managed patients received blood components according to clinical assessment and/or laboratory test results, compared with five of 30 managed by thrombelastogram data (P<0.05). The amount of product use in group T (five units of fresh frozen plasma and one pool of platelets) was significantly (P<0.05) less than the 16 units of fresh frozen plasma and nine platelet pools given to the 30 patients in group C.
Chest tube losses [median (lower and upper quartiles)] were not different between groups despite the fourfold difference in the use of haemostatic products. Twelve-hour losses were 470 (295, 820) in the group managed with the thrombelastogram and 390 (240, 820) in the group managed clinically.
![]() |
Comment |
---|
![]() ![]() ![]() ![]() ![]() ![]() |
---|
Others in the USA found the thrombelastogram useful for cardiac surgery, in which coagulopathy has different causes compared with hepatic surgery. In the presence of microvascular bleeding, thrombelastography was better than conventional tests of coagulation in defining the abnormality and in guiding treatment.6 Abnormal formation (principally r-time) implied reduced clotting factors, and reduced maximum amplitude deficiency implied reduced platelet numbers or function. A later study showed that this device reduced blood transfusions and also re-exploration for bleeding, which decreased fourfold during the study period.3 Other centres reported that the maximum amplitude, measured after protamine administration, was related to early losses from post-operative drains.4 7 This device has also been used in obstetric,8 trauma9 and peripheral vascular10 practice. In all these reports the patients blood was not anticoagulated at the time of the analysis and the test was being used to characterize rather than to predict and anticipate a coagulation abnormality.
The pilot study described here is the first to show that a heparinase-modified thrombelastogram in anticoagulated blood can predict, and reduce substantially, the need for haemostatic products without increasing post-operative bleeding in patients having heart surgery.
There are a number of potential concerns and weaknesses in the present study.
The first is related to the magnitude of the observed effect. We studied patients who were considered at high risk of requiring haemostatic products but in whom prophylactic drug therapies, such as aprotinin, would not be universally appropriate. Haemostatic components were given to about 30% of our control group patients, which is the same proportion reported from Montreal11 and New York.4 The second of these reports showed the benefit of an algorithm when microvascular bleeding was established. The authors described a four- to fivefold reduction in the use of frozen plasma, and our study supports this finding. In contrast to this earlier report,4 the present study demonstrates a statistically significant, eightfold reduction in platelet transfusion, suggesting that assessment of coagulation before the observation of microvascular bleeding can be valuable.
Part of this improvement may be related to the time the components are given. The algorithm we used allowed treatment shortly after the antagonism of heparin action by protamine. In contrast, the delay in obtaining laboratory data meant that patients in the control group received their component therapy after transfer to the intensive care unit. It is obvious that the magnitude of any benefit would be less for operations in which there is less need for haemostatic components.
Secondly, we did not want to use prophylactic drug therapies such as aprotinin and tranexamic acid, as they have different effects on thrombelastography. In particular, the r-time is prolonged by aprotinin therapy and shortened by tranexamic acid.12
This leads to the third weakness of the present algorithm and other algorithms,4 which is that the numerical values for each thrombelastogram parameter are based on reference ranges found in a standard population. There are no current data to show that the values chosen, or the treatment these values indicate, are optimal for any particular patient group. This is highlighted by the difference in the predicted requirement for transfusion of haemostatic components between the first and second cohorts despite similar patient characteristics, the same team of anaesthetists and surgeons, and very similar actual transfusions using conventional clinical criteria.
Finally, if the algorithm predictions were wrong, then the patients in group T might have had increased bleeding after the bypass. However, the accuracy of the prediction is not known as in neither group did we completely withhold products to determine any association with increased bleeding and the need for re-exploration. The values used in this algorithm may need modification to allow additional benefits and improve their applicability. This would require a larger, appropriately powered, multicentre study.
![]() |
References |
---|
![]() ![]() ![]() ![]() ![]() ![]() |
---|
2 Tuman KJ, McCarthy RJ, Djuric M, Rizzo V, Ivankovich AD. Evaluation of coagulation during cardiopulmonary bypass with a heparinase-modified thromboelastographic assay. J Cardiothorac Vasc Anesth 1994; 8: 1449[Medline]
3 Spiess BD, Gillies BS, Chandler W, Verrier E. Changes in transfusion therapy and reexploration rate after institution of a blood management program in cardiac surgical patients. J Cardiothorac Vasc Anesth 1995; 9: 16873[ISI][Medline]
4 Shore-Lesserson L, Manspeizer HE, DePerio M, et al. Thrombo elastography-guided transfusion algorithm reduces transfusions in complex cardiac surgery. Anesth Analg 1999; 88: 3129
5 Kang YG, Martin DJ, Marquez J, et al. Intraoperative changes in blood coagulation and thrombelastographic monitoring in liver transplantation. Anesth Analg 1985; 64: 88896[Abstract]
6 Spiess B, Tuman K, McCarthy R, et al. Thrombelastography as an indicator of post cardiopulmonary bypass coagulopathies. J Clin Monit 1987; 3: 2530[Medline]
7 Nuttall GA, Oliver WC, Ereth MH, Santrach PJ. Coagulation tests predict bleeding after cardiopulmonary bypass. J Cardiothorac Vasc Anesth 1997; 11: 81523[ISI][Medline]
8 Orlikowski CE, Payne AJ, Moodley J, Rocke DA. Thrombelastography after aspirin ingestion in pregnant and non-pregnant subjects. Br J Anaesth 1992; 69: 15961[Abstract]
9 Kaufmann CR, Dwyer KM, Crews JD, Dols SJ, Trask AL. Usefulness of thrombelastography in assessment of trauma patient coagulation. J Trauma 1997; 42: 71620[ISI][Medline]
10 Martin P, Greenstein D, Gupta NK, Walker DR, Kester RC. Systemic heparinization during peripheral vascular surgery: thromboelastographic, activated coagulation time, and heparin titration monitoring. J Cardiothorac Vasc Anesth 1994; 8: 1502[Medline]
11 Hardy JF, Perrault J, Tremblay N, et al. The stratification of cardiac surgical procedures according to use of blood products: a retrospective analysis of 1480 cases. Can J Anaesth 1991; 38: 5117[ISI][Medline]
12 Robbins P, von Kier S, Forrest M, Royston D. Activated clotting times are shortened by tranexamic acid. Anesthesiology 1998; 89 (3A): A962