The Thromboelastograph (TEG) was first used during orthotopic liver transplantation by Kang et al.1 in the 1980s as a bedside monitor of coagulation, and it is a standard guide to transfusion of various blood components in many transplant centers.2 Typical changes seen in TEG on reperfusion of the graft liver include a lengthening of the reaction time (R time) and coagulation time (K time), a diminished maximum amplitude, and an increase in fibrinolysis.2 This phenomenon is partly explained by the heparin-like effect (HLE) because even a small amount of heparin in the circulation prolongs R time and, in some cases, K time.3, 4 An HLE seen on TEG was first described during human liver transplantation in 1985 by Kang et al.1 It was noted that the HLE corrected in vitro with protamine administration1 and resolved spontaneously in most patients within 1 to 2 hours without any treatment. The use of heparinase-modified TEG began in the 1990s. In 1997, Harding and colleagues demonstrated that the addition of heparinase I, an enzyme-neutralizing heparin, heparan sulfate, and dermatan sulfate5 to whole blood samples in vitro normalizes the TEG indicating the presence of one of these substances6 (Fig. 1). This gives valuable information regarding the coagulopathy seen in these patients and the presence of heparin and heparin-like substances. Despite this, there remain a number of centers that do not routinely use heparinase TEG.
The origin of the observed HLE is unclear. The heparin and heparin-like substances may originate from an exogenous source. In current practice in the United Kingdom, donor livers are flushed with 300 units/kg heparin in the preservation solution prior to clamping of the hepatic vessels during organ procurement. Although the liver is flushed with 4.5% human albumin solution prior to reperfusion, residual heparin is still present, particularly in large livers or where heparin binds to the donor hepatic vascular endothelium.7
It is also possible that the heparin and heparin-like substances have an endogenous source. Kettner and colleagues8 demonstrated an HLE before reperfusion of the donor graft in patients undergoing liver transplantation without exogenous heparin administration. The HLE was also present after reperfusion when the donor liver was not perfused with a heparin-containing solution. These findings indicate that an endogenous source of heparin may be present in the recipient, perhaps because of decreased clearance by the diseased liver.9 Early animal work in dogs during liver transplantation demonstrated a rise in the circulating heparin level that had not originated from the reperfused liver.10 Further work has shown that hepatoenteric ischemic perfusion injury is associated with an increase in circulating heparin and heparinoid activity and a prolonged R time on TEG.11 There is also the possibility that the presence of HLE is a marker of vascular endothelium damage, and it has been found to be present in patients with underlying sepsis.12
There are known differences shown on TEG in liver diseases of various etiologies, particularly in cholestatic liver disease. Ben-Ari and colleagues13 looked at the differences between patients with cholestatic and noncholestatic cirrhosis that were found on TEG and demonstrated hypercoagulability (defined in this study as R time < 19 mm, maximum amplitude > 60 mm, and alpha angle > 43°) in up to 32% of cholestatic patients versus less than 5% of noncholestatic liver failure patients.13 Another study demonstrated that those with cholestatic liver disease had higher fibrinogen levels and less fibrinolysis than those with nonbiliary disease.14 However, there is little data regarding the difference in HLE seen at transplantation in patients of different etiologies.
The clinical significance of HLE has been controversial. Although a case report described a reduction in blood loss with the administration of protamine,15 most studies fail to show an increase in blood loss with HLE.16, 17 Nevertheless, some clinicians have continued to suggest treating HLE with protamine to neutralize the HLE in an attempt to decrease perioperative blood loss.15, 18
A new development of TEG software technology enables the production of a thrombin generation curve (VCurve) based on the first derivative of the TEG curve19 (Fig. 2). VCurve has been shown to have a high correlation with the standard thrombin-antithrombin enzyme-linked immunosorbent assay method of measuring thrombin generation (r = 0.94, prediction equation of thrombin generation = 8.0465 + 0.0005 VCurve).19
The primary aim of this retrospective observational study was to determine the prevalence and extent of the HLE in 211 consecutive patients undergoing liver transplantation in our institution at various stages throughout the transplant procedure. The secondary aims were to analyze the prevalence of HLE with respect to various etiologies and to analyze the differences in thrombin generation observed.
HLE, heparin-like effect; K time, coagulation time; MA, maximum amplitude; MTG, maximum rate of thrombin generation; R time, reaction time; SD, standard deviation; TEG, Thromboelastograph; TMG, time to maximal generation.
PATIENTS AND METHODS
The data of 211 consecutive patients undergoing orthotopic liver transplantation in our institution from 2001 to 2006 were retrospectively analyzed. All potential causes of introducing exogenous heparin prior to graft reperfusion were eliminated as none of the subjects were given heparin or warfarin in the month prior to transplantation and all lines and flush bags during the transplant were nonheparinized. None of the patients received protamine.
Blood (360 μL) was collected from a dedicated nonheparinized arterial cannula into a plain syringe and pippetted into plain and heparinase-coated TEG cups. Samples were run simultaneously 3 minutes after blood collection according to the manufacturer's recommendation (Haemoscope Corp, Niles, IL).
Paired TEG traces (with and without heparinase) were examined at 5 stages of the orthotopic liver transplantation: (1) at baseline, (2) 1 hour into the dissection, (3) half an hour into the anhepatic period, (4) half an hour after reperfusion, and (5) at the end of the operation. The TEG traces and appropriate software were used to derive the total thrombin generation, and they were compared between blood samples with and without heparinase.
HLE was calculated with the percentage correction of the R+K times on TEG with the addition of heparinase as shown:
HLE was considered present when the correction of the R+K time was greater than 50%, and severe HLE was defined as a value greater than 80%.
The measurements produced from the VCurve software give an indication of the speed of the thrombin burst (time to maximal thrombin generation) and its intensity (maximal rate of thrombin generation; that is, the peak of the thrombin generation curve) as well as the total amount of thrombin generated (the area under the curve; Fig. 2).
Patients were analyzed with respect to their underlying etiology. They were divided into the following groups: acute liver failure (fulminant or primary graft nonfunction), chronic hepatocellular [viral hepatitis, alcoholic hepatitis, nonalcoholic steatohepatitis, hepatocellular carcinoma, and retransplant (nonemergency)], and cholestatic (primary biliary cirrhosis and sclerosing cholangitis).
Statistical analysis of these data was performed with SPSS 11.0 (SPSS, Chicago, IL). Wilcoxon signed-ranks tests were used to evaluate the difference between the patients during the various stages of the operation. Differences between continuous variables in patients with and without HLE were evaluated with the Mann-Whitney U test, and differences between dichotomous variables were evaluated with the Fischer exact test. Differences between patients of different etiologies were examined with 1-way analysis of variance.
The mean age of the patients was 51.4 years [standard deviation (SD) 10.4] with a mean Child-Pugh score of 8.4 (SD 2.1) and total bilirubin of 62.9 μmol/L (SD 96.9 μmol/L). At the start of the transplant, there was a mean platelet count of 81 × 109/L (SD 45 × 109/L) and prothrombin time of 20.8 seconds (SD 6 seconds).
The preoperative characteristics of the subjects were analyzed with respect to the severity of HLE both at the baseline and at reperfusion. There was no significant difference in the Child-Pugh score or baseline coagulation or liver function tests in the groups of patients with or without HLE.
Prevalence of HLE at Each Stage
The prevalence of HLE is shown in Table 1. At the start of the transplant, 31% of patients had demonstrable HLE (n = 65); however, this was severe in only 6% (n = 14). Over half of these patients continued to demonstrate HLE during the dissection phase and the anhepatic phase, although it disappeared in over 20 patients. No patient developed HLE de novo during the dissection or anhepatic period. After reperfusion, 75% of the patients had HLE, which was severe in over 35% of patients. This resolved spontaneously in 47%, although 53% had residual HLE at the end of the procedure, which was severe in 23%.
Table 1. Prevalence of a Heparin-Like Effect (HLE) in All Patients at Various Stages of the Liver Transplant
Stage of Transplant
With HLE [% (n)]
With Severe HLE [% (n)]
The relationship between those who had HLE at the start of the transplant and those who developed severe HLE post reperfusion was explored with analysis of variance. There was no relationship demonstrable between the 2 sets of patients; those who developed severe HLE were not necessarily those who had preexisting HLE prior to reperfusion.
HLE at Each Stage by Etiology
At baseline, there was no significant difference in the presence of HLE observed between the groups (not significant), as shown in Table 2, although there was a trend for those with acute liver failure to have a greater prevalence of HLE. After reperfusion of the donor graft, those with acute liver failure were significantly more likely to develop severe HLE (P = 0.08), although there was a similar prevalence of HLE at the 50% level in all 3 groups.
Table 2. Prevalence of a Heparin-Like Effect (HLE) at the Start of the Transplant and at Reperfusion of the Graft with Respect to the Etiology: Acute (Fulminant or Primary Graft Nonfunction), Chronic [Viral Hepatitis, Alcoholic Hepatitis, Nonalcoholic Steatohepatitis, Hepatocellular Carcinoma, and Retransplant (Nonemergency)], and Cholestatic (Primary Biliary Cirrhosis and Sclerosing Cholangitis)
With HLE [% (n)]
With Severe HLE [% (n)]
With HLE [% (n)]
With Severe HLE [% (n)]
Cholestatic (n = 40)
Acute (n = 147)
Chronic (n = 24)
Thrombin Generation and HLE
Differences in Thrombin Generation on Native Samples
The thrombin generation traces derived from the native samples of those with and without HLE were compared with the Mann-Whitney U test. Those with HLE at the start of the transplant had an increased time to maximal thrombin generation (6575 versus 2376 seconds, P = 0.03) and a decreased maximum rate of thrombin generation (2.8 versus 9.1 mm × 100/second, P = 0.05). However, the total thrombin generation did not differ significantly at this point.
At reperfusion, there were more marked differences. Patients with HLE at reperfusion had a significant decrease in total thrombin generation (1416 versus 3550 mm × 100, P = 0.02). In addition, they had an increase in the time to maximal thrombin generation (5933 versus 2713 seconds, P = 0.03) and decrease in the maximum rate of thrombin generation (4.0 versus 6.2 mm × 100/second, P = 0.05).
There was no difference in the blood transfusion requirement between those with and without HLE at any stage of the transplant.
An HLE has been noted for many years during liver transplantation, usually after reperfusion of the donor liver, although its significance has never been fully elucidated.6, 18 We found that HLE was present at the start of the transplant in over 30% of our patients. Kettner and colleagues8 were among the first to demonstrate that HLE was present to a variable extent prior to reperfusion of the graft during liver transplantation by comparing the baseline R and K times in native and heparinase samples. However, the HLE that we observed at the start of the transplant resolved by the anhepatic period in 30% of those in whom it was observed at baseline. The reasons for this are unclear: there was no difference in the transfusion of blood products in those with or without HLE that would account for this; it is possible that there was an underlying activation of coagulation that led to this change. The HLE that was observed prior to reperfusion was also less severe than that observed after reperfusion; only 12% of patients had severe HLE prior to reperfusion, whereas a severe HLE was observed in 39% of the subjects after reperfusion.
Definitions of HLE vary between studies, although a previous study similarly described the difference in native R+K values versus heparinase R+K values.8 The clinical significance of HLE has been debated in the past, particularly whether HLE is associated with an increase in blood loss in transplantation. A previous study looking at HLE found that there was an increase in blood loss only if the underlying coagulation was poor.6 There are numerous factors that affect blood loss during liver transplantation, including surgical difficulties and management of coagulation, which were not controlled in this retrospective analysis. Nevertheless, this study has demonstrated that although HLE is detected by TEG in many liver transplant patients both before and after reperfusion of the donor graft, its presence does not correlate with an increase in blood requirement.
The use of protamine in cases in which HLE is present is controversial. Kang,2 Pivalizza et al.,18 and Mallett et al.6 have all used protamine in vitro to demonstrate the reversal of HLE. One case report describes the use of protamine to neutralize the heparin effect in an attempt to decrease perioperative blood loss15; in vivo, however, there is little evidence that it affects intraoperative blood loss.
The origin of the HLE has been the subject of much study. There is a clear exogenous source of heparin: all our donors receive 300 μg/kg heparin prior to clamping of the hepatic vessels, as this is standard UK transplant procedure. Although the liver is flushed with 500 mL of 4.5% albumin prior to completion of the venous anastomoses, there is a well-documented residual effect, particularly in larger livers.7 Exogenous heparin in the preservation fluid may bind to the donor liver hepatic vascular endothelium and be subsequently released into the circulation after reperfusion. In addition, there may be release of donor endogenous heparin from the ischemic vascular endothelium and from activated macrophages. All hepatocytes and liver vascular endothelium contain a large amount of heparin, and it is to be expected that an HLE will become apparent after reperfusion of the graft with ischemic damage.
The presence of HLE prior to reperfusion would indicate that there is also an endogenous source in the patients with cirrhosis, which may be a marker of endothelial damage or infection.20 Heparan sulfate and dermatan sulfate (both heparin-like substances) are naturally occurring constituents of the vessel wall and may be bound by the hepatic endothelium. They are known to inhibit clot formation by interaction with anti-Xa and heparin cofactor II.5 A study investigating variceal bleeding has clearly found the presence of endogenous heparinoids during the time of bleeding.21 A separate study investigating the levels of heparinoids in portal hypertension postulated that the high levels found could be related to decreased hepatic clearance in liver disease.9 This may contribute to both baseline HLE and HLE seen after reperfusion when the donor liver is functioning suboptimally as it recovers from ischemia and reperfusion. In our study, 100 subjects continued to demonstrate HLE at the end of the procedure.
There were some significant differences between the patients with respect to the prevalence of HLE and the etiology of liver failure. There was a higher preponderance of HLE in those with acute liver failure. These groups often have the worst liver function, and the HLE could signify endothelial injury or reaction to the stress of illness. The significance of this is unclear; it may be related to endogenous heparin release and is worthy of further investigation. Interestingly, we also found that although patients with cholestatic and noncholestatic disease had markedly different baseline coagulation profiles, the prevalence of HLE was similar. Again, the reasons and potential clinical significance of this remain unclear.
Our study demonstrated that those with HLE at the start of the transplant had an increased time to maximum thrombin generation (P < 0.05) and a decreased maximum rate of thrombin generation (P = 0.05). However, the total thrombin generation did not differ at this point. This would indicate that those patients with HLE at baseline formed a clot slowly and with less intensity compared to those without HLE, but the total amount of the clot that formed over the time of the VCurve was the same.
At reperfusion, there were more marked differences. There was an increased severity of HLE that was associated with an increased time to maximum thrombin generation (P < 0.05) and a decrease in total thrombin generation in those with HLE (P = 0.02). This would indicate that after reperfusion, those with HLE were likely to form a clot more slowly and with less intensity and that the total amount of the clot would also be diminished. Heparin acts by binding with anti-thrombin III inhibiting thrombin generation.22 When given to patients, it demonstrates significant interindividual variability23 and behaves in a dose-dependent linear model with saturable pharmacokinetics.22 Those with more severe HLE may have greater inhibition of thrombin generation as they have greater antithrombin III binding. A previous study investigating thrombin-antithrombin complexes during liver transplantation showed thrombin generation to slightly increase after reperfusion.23 However, this study did not investigate those with HLE separately. In addition, thrombin is known to contribute to clot stability by activating thrombin activatable fibrinolysis inhibitor, which protects the fibrin clot against lysis. A reduction in thrombin may directly affect the protection of the clot and contribute to coagulopathy.24
This study has demonstrated that the prevalence and severity of HLE vary throughout the stages of liver transplantation. It is observed in up to a third of patients with chronic liver disease at baseline, with an even higher prevalence in acute liver failure. At reperfusion, it is an almost universal phenomenon, with over 80% of patients demonstrating HLE. The effects of the HLE on thrombin generation, as demonstrated on the TEG VCurve, indicate that at reperfusion the clot that has formed may be less solid and may exacerbate any underlying coagulopathy.6 This study confirms that there may be both endogenous and exogenous sources of the HLE; however, its clinical significance remains unclear.