Shen-Chih Wang contributed to the study design and the conduct of the study. Ho-Tien Lin contributed to the data collection and manuscript preparation. Kuang-Yi Chang contributed to the data analysis. M. Susan Mandell contributed to the data analysis and manuscript preparation. Chien-Kun Ting, Ya-Chun Chu, and Kwok-Hon Chan contributed to the conduct of the study. Che-Chuan Loong contributed to the conduct of the study and was the leading surgeon of the transplant team. Mei-Yung Tsou contributed to the study design.
Blood transfusions in liver transplant recipients are associated with an increased risk of infection and lung injury.1 This leads to increases in morbidity and mortality.2 Other outcome studies have shown that intraoperative transfusions of plasma products significantly reduce 1-year survival3-5 and increase morbidity in liver transplant recipients.6
Physicians have used the thromboelastogram (TEG) as a tool for identifying abnormal coagulation. TEG can discriminate between different phases of the coagulation system and is, therefore, useful for identifying clotting abnormalities and guiding the administration of blood products.
In 2007, we reported using less frozen plasma in transplant patients when transfusions were guided by standard TEG cutoff values instead of the international normalized ratio (INR)/prothrombin time and platelet counts.7 We concluded that TEG was a helpful tool for making decisions about transfusions. Our findings were independently confirmed by other investigators.8 Therefore, in 2007, we started using standard TEG values to guide the transfusion of coagulation factors in our liver recipients. The endpoint of our therapy was TEG values within normal limits.
However, investigators have postulated that TEG values higher than normal may not predispose transplant recipients to blood loss in the pre-anhepatic stage of surgery.9-11 If this premise is true, it may be possible to safely decrease the amount of coagulation products administered without effects on overall blood loss. The use of fewer blood products theoretically should improve patient outcomes.
Therefore, in 2009, we adopted a TEG-guided transfusion protocol using higher threshold values to initiate transfusions.12 The revised value for the reaction time (R time) was 15 minutes, and the maximum amplitude (MA) was 40 mm. The values differed by approximately 35% from established normal values reported for routine TEG analysis. The aim of the present study was to determine whether this change in our transfusion strategy made a difference in overall blood transfusions and in quality indicators of patient outcomes. Therefore, we measured blood loss and blood product use intraoperatively and postoperatively for 3 days to test our new protocol.
FFP, fresh frozen plasma; HBV, hepatitis B virus; HCV, hepatitis C virus; ICU, intensive care unit; INR, international normalized ratio; MA, maximum amplitude; R time, reaction time; TEG, thromboelastogram.
PATIENTS AND METHODS
Since 2007, our transfusion protocol was guided by standard TEG cutoff values. In 2009, we started to use a TEG-guided transfusion protocol with 35% higher threshold values. Our surgeons were blinded to the changes in the transfusion protocol. Data for each recipient began to be collected immediately after surgery and included 3 days of postoperative follow-up. After institutional board approval (201011029IC), we collected transfusion data from adult patients who underwent living donor liver transplantation between 2007 and 2010. We excluded 9 patients with Model for End-Stage Liver Disease scores higher than 30 because of their potentially higher rates of adverse events and mortality.13 One patient was excluded because of surgical complications that resulted in massive blood loss during surgery.
Data were collected and analyzed for 77 patients. The patients were divided into 2 groups according to the TEG values used to guide the transfusion of fresh frozen plasma (FFP) and platelets. Thirty-eight patients received transfusions according to standard TEG values (the control group), and 39 received transfusions according to higher TEG values (study group).
Patients were monitored with a 5-lead electrocardiogram, pulse oximetry, a noninvasive blood pressure monitor, and an Alaris auditory evoked potential monitor (version 1.4, Alaris, Hampshire, United Kingdom). General anesthesia was induced and maintained with total intravenous anesthesia according to the Schnider propofol infusion model.14 For induction, our propofol target concentration was set at 5 μg/mL, and fentanyl (5 μg/kg) was administered. Rocuronium (0.6 mg/kg) was used to facilitate intubation. After intubation, the propofol dose was adjusted to keep the auditory evoked potential between 15 and 25. After induction, a 20-G radial arterial catheter and a 9-Fr triple-lumen central venous catheter were inserted. The operating room temperature was kept at 22°C, and we used a hotline combined with a surface heater (Bair Hugger) to keep the body temperature higher than 35°C. Arterial blood gases were checked whenever we drew blood samples for TEG. The free calcium level was kept higher than 1.0 mmol/L.
All TEG analysis (5000 series, Haemoscope, Skokie IL) was performed with a kaolin active assay. Blood samples were routinely obtained in accordance with our previous study7 at the following time points: after induction; 1 hour after induction; 5 minutes before the anhepatic phase; 10 minutes into the anhepatic phase; 5 minutes before reperfusion; and 10, 30, and 60 minutes after reperfusion.
We used the transfusion strategy established by Kang et al.12 to guide the treatment of both groups of patients. Patients in our first group received 2 U of FFP if the R time was greater than 11 minutes. One unit of pheresis platelets (equivalent to 8-10 pooled units) was given to patients with an MA less than 55 mm. The second group of patients received 2 U of FFP if the R time was greater than 15 minutes and 1 U of apheresis platelets if the MA was less than 40 mm. Cryoprecipitate (6 pooled units) was given to both groups of patients when the rate of clot formation was slow (α < 45 degrees).
If uncontrollable bleeding was noticed by the surgeon, we transfused 2 U of FFP and 1 U of pheresis platelets into the patient and then rechecked the TEG values. The transfusion protocol for our study group is shown in Fig. 1. However, no patient in either group received transfusions for uncontrollable bleeding. Packed red blood cells were transfused to keep the hemoglobin level at 10 g/dL.
Blood loss was calculated as the amount of fluid in the suction bottle minus ascites and flushing fluid used on the table plus cell saver and the weight of blood-soaked sponges. After surgery, patients were kept intubated and ventilated and were transferred directly to the intensive care unit (ICU). Extubation and transfer to the surgical ward were based on a single set of criteria used by the ICU care team.15 In the ICU, we checked the hemoglobin level, INR, and platelet count daily. Packed red blood cells (1 U) were transfused if the hemoglobin level was less than 8 g/dL. FFP (2 U) was used when the INR was higher than 2. One unit of pheresis platelets was transfused if the platelet count was less than 30,000/mm3. We recorded the results of coagulation tests on postoperative day 1 and totaled the use of blood products on postoperative days 1 to 3.
The parametric and categorical demographic data were compared with the independent t test and the χ2 test, respectively. Perioperative fluid intake, blood loss, and blood product use were analyzed with the Mann-Whitney U test. A P value < 0.05 was considered statistically significant. The statistical treatment of all data was performed with SPSS 17.0 (SPSS, Inc., Chicago, IL).
The demographic profiles of our patient populations and their indications for liver transplantation are shown in Table 1. There were no significant differences in the ages, body mass indices, preoperative hemoglobin levels, platelet counts, INRs, total bilirubin levels, creatinine levels, or Model for End-Stage Liver Disease scores of the 2 groups of patients. There were also no differences in the number of patients with a history of major abdominal surgery, transarterial chemoembolization, or portal vein thrombosis.
Table 1. Demographic Data for the Patients
Study Group (n = 39)
Control Group (n = 38)
NOTE: The data are presented as means and standard deviations unless otherwise noted.
*P < 0.05.
52.61 ± 10.27
55.11 ± 7.79
Body mass index (kg/m2)
23.18 ± 5.42
23.94 ± 2.86
Indication for liver transplantation (n)
Acute hepatitis with liver failure
Liver cirrhosis without hepatocellular carcinoma
Non-B, non-C hepatitis
Liver cirrhosis with hepatocellular carcinoma
Non-B, non-C hepatitis
HBV and HCV
Medical history (n)
Major abdominal surgery
Portal vein thrombosis
0.92 ± 0.42
1.03 ± 0.44
Total bilirubin (mg/dL)
4.64 ± 6.97
4.88 ± 7.22
1.40 ± 0.54
1.36 ± 0.41
10.86 ± 2.05
10.47 ± 1.90
Model for End-Stage Liver Disease score
13.53 ± 6.26
12.87 ± 6.13
Fluid balance, blood product use, and blood loss results are shown in Table 2. Significantly lower volumes of FFP and platelets were used in patients for whom the critical TEG value to initiate therapy was higher. However, there were no significant differences in blood loss or packed red blood cell transfusions.
Table 2. Operation Times, Perioperative Fluid Intake, Blood Product Use, and Blood Loss
Historical Control Group
NOTE: The data are presented as medians and interquartile ranges.
The results for postoperative coagulation tests and blood product use on days 1 to 3 are shown in Table 3. For the study group, the INR was significantly higher, and the platelet count was significantly lower. However, none of the 77 patients included in this study experienced postoperative bleeding that required additional surgery. No significant difference was noted in the postoperative use of blood products.
Table 3. Postoperative Coagulation Tests and Blood Product Use on Days 1 to 3
NOTE: The data are presented as means and standard deviations.
The findings of our study suggest that transfusion protocols using TEG cutoff values 35% greater than the standard values are not associated with bleeding in liver transplant recipients. The reduction in the transfusion of plasma-containing products did not increase the risk of postoperative bleeding. Our findings provide evidence supporting the argument that abnormal clotting values measured by laboratory tests do not accurately predict blood loss during liver transplantation.
The results of coagulation tests are often abnormal in patients with liver disease because these patients fail to synthesize normal amounts of coagulation factors.16 However, studies have shown that there is a rebalancing of the coagulation system in patients with liver disease.16 The injured liver synthesizes less of naturally occurring anticoagulants such as protein C, protein S, and antithrombin III.17 Investigators have shown that the propensity to bleed due to coagulation defects is in part determined by the balance between naturally occurring coagulation factors and anticoagulants.17
TEG evaluates the 2 arms of the coagulation cascade simultaneously. This distinguishes it from serum-derived laboratory tests such as the INR. Normal TEG values were determined from the average clotting time of healthy volunteers.18 Although investigators have tested the correlation between the risk of bleeding and TEG values in various surgical populations,19, 20 it is possible that standard TEG cutoff values derived from a healthy population do not predict bleeding in patients with liver disease. This line of reasoning is supported by our finding that the titration of FFP and platelets to obtain normal TEG values is not cost-effective.
Blood loss was not increased in our liver recipients by the initiation of transfusions with higher critical R times and lower MA values from TEG. We cannot be sure that our findings can be explained by the rebalancing of the coagulation system in patients with chronic liver disease. The reason for this lack of correlation was not addressed by the design of our study. This observation suggests, however, that standard TEG values obtained from healthy volunteers may not represent normal values for patients with liver disease. It is also possible that the risk of bleeding increases only when the normal values are exceeded by a specified amount. Additional studies are required to examine these hypotheses.
We transfused 2 U of FFP for prolonged R times and 1 U of pheresis platelets for abnormal MA values. In Taiwan, each blood product unit is made from 250 mL of whole blood. In order to decrease unnecessary blood product use, we checked TEG values at fixed time points and treated abnormal data with minimum blood product use.
Our study was limited by the use of historical data. However, our surgeons are experienced in liver transplantation and have been a stable team since 2002. Thus, we do not think that our findings are biased by changes in practice over the time period of the study.
We used a single protocol for interventions to prevent treatments based on clinical impressions. This limited the number of therapeutic interventions that we used. It is possible that the therapeutic interventions used in this study require further modification to optimize outcomes. In addition, it is also possible that our higher TEG value protocols may not be as successful in different populations. These limitations will require further testing.
In conclusion, the transfusion of plasma and platelets to obtain normal TEG values is not cost-effective. The use of critical values that are 35% higher does not increase the risk of bleeding. Additional studies are needed to identify TEG values that are predictive of bleeding.