The study was supported by a starter grant from the Department of Anesthesiology of the Penn State Milton S. Hershey Medical Center/Penn State College of Medicine and by a grant from the General Clinical Research Center of Pennsylvania State University.
These data were presented at the June 2010 meeting of the International Liver Transplantation Society in Hong Kong.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
The reperfusion of a transplanted liver graft is a critical step during liver transplantation (LT). The hemodynamic instability can be profound, and an infusion of catecholamines and the administration of fluids may be required to maintain homeostasis. Arrhythmias are frequent and may be difficult to manage. The amount of catecholamines required to maintain acceptable organ perfusion can vary widely and has no apparent correlation with the preoperative cardiovascular status of the patient. As part of the preoperative evaluation for LT, patients undergo an extensive cardiac workup. They usually have no or only minimal cardiac disease.
The cause of unpredictable hemodynamic events during LT is not well understood. Possible mechanisms include electrolyte abnormalities, cold preservation solution entering the patient's systemic circulation, and vasodilation from the release of nitric oxide.1, 2
Liver procurement and preservation are associated with a complex, systemic inflammatory response related to the release of various cytokines. Cytokines have messenger RNA expression times of 30 minutes to 3 hours.3, 4 Tumor necrosis factor α (TNF-α), which is produced by macrophages and Kupffer cells, exists as preformed, membrane-bound molecules.5 These are activated by the release of the matrix metalloproteinase TNF-α–converting enzyme.6 TNF-α–converting enzyme expression is mediated by monocytes released as part of an inflammatory reaction.7 The inflammatory cytokine interleukin-8 (IL-8) is released by various cell types, including vascular endothelium, neutrophils, macrophages, Kupffer cells, and fibroblasts, in response to IL-1β and TNF-α stimulation.8 The inflammatory response itself depends on changes in the metabolic activity of neutrophils and results in the production of free radicals that may contribute to the development of graft injury.9
It has been hypothesized that proinflammatory cytokines, produced in the liver graft and flushed into the patient's circulation during reperfusion, may contribute to negative inotropy and systemic vasodilatation.1, 10 This cytokine elevation could be the result of acute production during graft reperfusion or the release of preexisting cytokines generated in the graft. The release of preexisting cytokines could be a reaction to the cold preservation solution or a response to a mechanical or pressure phenomenon, such as clamping of the portal vein, during the anhepatic phase.
The objectives of this study were as follows:
1To determine whether the transplanted liver graft itself contributes to elevated levels of various cytokines (TNF-α, IL-1β, IL-2, IL-6, and IL-8) seen in the patient's systemic circulation.
2To determine whether elevated cytokine levels in the systemic circulation correlate with the amount of catecholamines required to maintain hemodynamic stability during graft reperfusion.
This prospective, observational study was approved by the institutional review board of the Penn State Milton S. Hershey Medical Center and Penn State College of Medicine. Seventeen consecutive adult patients undergoing orthotopic LT were enrolled in the study after written, informed consent was obtained.
Liver grafts from extended criteria donors (ECDs) were used for 11 patients [ECD criteria included hypernatremia > 155 mmol/L, cold ischemia time (CIT) > 15 hours, donor age > 60 years, steatosis > 30%, aspartate aminotransferase/alanine aminotransferase levels > 100 U/L, intensive care unit stay > 7 days, downtime requiring cardiopulmonary resuscitation, and hepatitis C positivity]. Six transplants were performed with grafts from standard criteria donors (SCDs). Donation after cardiac death donor grafts were not used. All patients included in the study, with the exception of 1 case with fulminant liver failure, had advanced cirrhosis.
Anesthetic and Surgical Procedures
Anesthesia was provided according to the institutional protocol. After the application of noninvasive monitors, anesthesia induction was performed with the administration of fentanyl (2 μg/kg), propofol (1-2 mg/kg), and succinylcholine, cisatracurium, or rocuronium. After tracheal intubation, the mechanical ventilation (35%-40% O2 in air) was adjusted to maintain an end-tidal carbon dioxide concentration of 35 to 38 mm Hg. Maintenance of anesthesia was accomplished with sevoflurane, a fentanyl infusion (2 μg/kg/hour), and intermittent boluses of cisatracurium (0.05-0.1 mg/kg) for muscle relaxation. Bilateral radial arterial lines were placed. Under ultrasound guidance, the right internal jugular vein was cannulated with a 9-Fr MAC introducer for central venous pressure monitoring, volume resuscitation, and possible infusion of catecholamines. One or 2 rapid infusion catheters (7.5 or 8.5 Fr) were inserted into the cephalic or basilic veins. A Belmont rapid infusion system (Belmont Instrument Corp., Boston, MA) was used for volume resuscitation.
Fifteen of 17 patients received a tranexamic acid infusion (10 mg/kg/hour). Two patients not receiving tranexamic acid had a history of thrombosis. The coagulation profile, which included the partial thromboplastin time, prothrombin time/international normalized ratio, fibrinogen level, platelet count, and thromboelastogram, was evaluated hourly.
No immunosuppression was given before surgery. At the beginning of the anhepatic phase, 1 g of methylprednisolone was infused over 30 minutes.
Before reperfusion, the liver graft was flushed with the patient's blood by temporary unclamping of the portal vein for the removal of the preservation solution (no crystalloid flush was used in any of the study patients). While suprahepatic and infrahepatic inferior vena cava (IVC) occlusion was maintained, the portal vein was unclamped. Flush blood was sampled through a catheter placed above the infrahepatic IVC clamp.
At the beginning of reperfusion, all patients had similar hemodynamic and physiological parameters; this was demonstrated by a central venous pressure of 7 to 10 cm H2O, a mean arterial pressure of at least 65 mm Hg, a normal blood potassium concentration, and a pH greater than 7.38.
Catecholamine and fluid management during liver graft reperfusion were performed with a uniform institutional protocol:
1All liver grafts were weighed before implantation. The amount of flush blood was standardized as 20 mL of blood per 100 g of liver graft (250-400 mL).
2If the mean arterial pressure, after liver graft reperfusion was begun, was lower than 65 mm Hg, a 10-μg bolus of norepinephrine (NE) was administered. If the blood pressure did not immediately improve, 2 subsequent doses of NE (10 μg each) were administered. If the target mean arterial pressure (65 mm Hg) was not reached after the initial NE dosing, NE (50 μg) was given with a 200-mL bolus of a colloid solution through the Belmont rapid infusion system. Boluses of NE (100 μg) were then used for all subsequent NE administrations.
3Boluses of colloid (albumin, blood, or fresh frozen plasma according to the clinical situation) equal to the amount of flush used were given before graft flushing was initiated. Additional boluses of a colloid solution (200 mL) were given if the administration of NE failed to produce the desired arterial blood pressure. We used a restrictive protocol for fluid management with a target central venous pressure of 7 to 10 cm H2O in order to avoid congestion of the liver graft.
None of the patients received a catecholamine infusion before or during graft reperfusion. In 3 cases, an NE infusion was started 1 to 1. 5 hours after the IVC was unclamped.
Samples and Catecholamine Requirement
Serial blood samples (8 mL placed in ethylene diamine tetraacetic acid tubes) were obtained before and immediately after graft reperfusion (see Fig. 1 for details). These time points were chosen to isolate the period of potential hemodynamic instability during graft reperfusion:
Sample 1 was obtained by the surgeon from the recipient portal vein just before completion of the portal vein anastomosis and graft irrigation.
Sample 1a was obtained at the same time from the arterial line.
Sample 2 was obtained from the flush blood at the beginning of anterograde liver graft perfusion with the patient's blood.
Sample 2a was obtained at the same time from the arterial line.
Sample 3 was obtained from the flush blood at the end of anterograde liver graft perfusion with the patient's blood just before the recipient's IVC was unclamped.
Samples 4a and 5a were obtained from the arterial line 10 and 20 minutes after reperfusion, respectively.
The total amount of NE necessary to maintain a mean arterial blood pressure higher than 65 mm Hg for the first 20 minutes after reperfusion was recorded and compared to the levels of cytokines in the systemic circulation.
All blood samples were placed on ice and processed with centrifugation. The plasma supernatant was frozen at −80°C. Cytokine assays were performed to determine the concentrations of TNF-α, IL-1β, IL-2, IL-6, and IL-8 (R&D Systems, Inc., Minneapolis, MN).
Calculations and Statistics
Cytokine levels were compared with the least squares mean method to detect any statistically significant increases in the concentration. Kendall's tau-b test and regression analysis were used to discern any correlation between elevated concentrations of cytokines in the systemic circulation after graft reperfusion and the amount of the catecholamine required to maintain hemodynamic stability in the first 20 minutes of the neohepatic phase. A P value less than 0.05 was considered statistically significant.
The demographic characteristics of the patients enrolled in the study are provided in Table 1. Cytokine concentrations at important time points are provided in Table 2. Positive values indicate increased cytokine levels, and negative values indicate decreased levels with respect to the reference sample.
Table 2. Differences in the Cytokine Concentration at Selected Sampling Points
Concentration Change (pg/mL)
Standard Error (pg/mL)
NOTE: Sample 1 was obtained by the surgeons from the recipient portal vein just before completion of the portal vein anastomosis and liver irrigation. Sample 1a was obtained at the same time from the arterial line. Sample 2 was obtained from the flush blood at the beginning of graft irrigation. Sample 2a was obtained at the same time from the arterial line. Sample 3 was obtained from the flush blood at the end of graft irrigation just before the redirection of the blood flow and the unclamping of the recipient IVC. Samples 4a and 5a were obtained from the arterial line 10 and 20 minutes after reperfusion, respectively.
The concentration of TNF-α in the flush blood (sample 2) was significantly higher than the concentrations in the blood samples obtained from the radial artery (P = 0.0002; sample 1a) and the portal vein (P = 0.0010; sample 1) before reperfusion. The level of TNF-α in the flush blood at the end of liver irrigation (sample 3) was significantly higher than the levels in the samples from the radial artery (P = 0.0067) and portal vein (P = 0.0003). Kendall's tau-b test confirmed the correlation between the amount of the catecholamine and the level of TNF-α (Table 3). The levels of samples obtained from the radial artery 10 and 20 minutes after reperfusion (samples 4a and 5a) were also higher, but they were not statistically significant (P = 0.0566 and 0.0482, respectively). Kendall's tau-b test (tau-b = 0.1310) and regression analysis also failed to demonstrate an association between the increased level of TNF-α in the systemic circulation after graft reperfusion at the same time points (10 and 20 minutes) and the amount of NE used (P = 0.64).
Table 3. Correlation Between the Differences in TNF-α Levels in Arterial Blood and Flush Blood at the End of Graft Irrigation and the Amount of NE Used to Maintain Hemodynamic Stability During Reperfusion
NOTE: Kendall's tau-b test was used.
Adjusted standard error
95% lower confidence limit
95% upper confidence limit
The concentration of IL-1β in the flush blood (sample 2) was significantly higher than the concentrations in the blood samples obtained from the radial artery (P = 0.0003; sample 1a) and portal vein (P = 0.0138; sample 1) before reperfusion. A significant increase in the IL-1β level was found in sample 3 at the end of graft irrigation (P = 0.0020); however, Kendall's tau-b test failed to demonstrate a correlation between the IL-1β level and the amount of the catecholamine used (tau-b = 0.1238). No statistically significant difference between the samples obtained from the radial artery before and after reperfusion (samples 4a and 5a) was demonstrated.
The concentration of IL-2 was increased in sample 3 (P = 0.0027). This did not correlate with the amount of NE used (tau-b = 0.1225).
The concentration of IL-6 was not significantly higher in the flush blood (samples 2 and 3) versus the baseline levels obtained from the radial artery, but it was significantly higher in the sample from the portal vein (sample 1) versus the sample from the radial artery (sample 1a; P = 0.0190).
The concentration of IL-8 in the flush blood (sample 2) was significantly higher than the concentration in the sample obtained from the radial artery (P = 0.0046; sample 1a) but was not significantly different from the concentration in the sample obtained from the portal vein (P = 0.0787; sample 1) before reperfusion. The IL-8 level of the sample obtained at the end of liver flushing (sample 3) was significantly higher than the level of the sample from the radial artery (sample 1a; P = 0.0140). Kendall's tau-b test failed to correlate this with the amount of the catecholamine used (tau-b = 0.1412). The levels of the samples obtained from the radial artery 10 and 20 minutes after reperfusion (samples 4a and 5a) were significantly higher (P = 0.0212 and P = 0.0126, respectively). Kendall's tau-b test demonstrated a correlation between the level of cytokines obtained from the radial artery after reperfusion and the amount of NE used (Table 4). A regression analysis, however, failed to demonstrate an association between the same parameters (P = 0.37).
Table 4. Correlation Between the Differences in IL-8 Levels in Arterial Blood Before and 20 Minutes After Liver Graft Irrigation and the Amount of NE Used to Maintain Hemodynamic Stability During Reperfusion
NOTE: Kendall's tau-b test was used.
Adjusted standard error
95% lower confidence limit
95% upper confidence limit
CIT and Catecholamine Requirement
A statistical analysis failed to demonstrate a correlation between any cytokine and CIT (P = 0.5-0.8).
A standard surgical technique (with cross-clamping of the IVC) was used in 9 patients. A piggyback technique with partial clamping of the IVC was used in 8 patients. None of the patients required venovenous bypass. A statistical analysis did not demonstrate a significant difference in the level of any cytokine and the surgical technique (P = 0.07-0.8).
Type of Liver Graft
There was no statistically significant difference in the levels of IL-8 and TNF-α between SCD and ECD grafts (P = 0.18 and P = 0.31, respectively).
Severe hypotension is a frequent event both during and immediately after reperfusion of a liver graft. This is usually described as postreperfusion syndrome (PRS). A diagnosis of PRS can be made when a decrease in the systolic blood pressure of at least 30% occurs in the first 5 minutes after graft reperfusion and lasts at least 1 minute.11, 12 In the present study, 16 of 17 patients developed PRS according to these criteria. The observed incidence of PRS is significantly higher than that previously reported. This is most likely related to the surgical technique (caval cross-clamping without a temporary shunt in the majority of cases) and to the fact that a patient's own blood was used for the initial flush of the liver graft, which produced relative hypovolemia before the reperfusion itself.
The challenges associated with managing PRS have made its prediction and treatment the target of several investigations. Paugam-Burtz et al.13 demonstrated that an absence of temporary portocaval shunts and the duration of CIT are independent predictors of PRS. In our study, a temporary portocaval shunt was not used in any patient. The present study failed to demonstrate a correlation between any cytokine levels and the duration of CIT (7-11 hours). Other factors that can contribute to the development of PRS include the severity of the medical condition of the recipient, renal failure, infection, the length of the intensive care unit stay, the surgical technique, the duration of reperfusion, and the quality of the liver graft.
All patients enrolled in our study were in stable medical condition without any known significant cardiac problems. One patient was admitted to the intensive care unit before transplantation because of fulminant liver failure. All other patients were admitted from home. None of the patients required mechanical ventilation or catecholamine administration before surgery. Two of 17 patients in the study developed renal failure preoperatively because of hepatorenal syndrome and required dialysis. No significant preoperative infections were identified in any patients in the study.
There were no significant deviations in the duration of reperfusion. Reperfusion lasted 3.5 to 4 minutes in all patients.
PRS is both patient-dependent and graft-dependent. In this study, we focused our attention on 2 potential graft problems:
1Identifying the graft as a source of cytokines during graft reperfusion.
2Correlating any acute release of cytokines from liver grafts with periods of hemodynamic instability in an effort to identify a possible cause for PRS.
It has been previously postulated that the liver graft itself could be a possible source of increased cytokine levels.14, 15 Until now, this has not been clearly demonstrated.
Duran et al.14 used a similar study design and found increased cytokine levels in blood obtained from reperfused grafts 5 minutes after unclamping. Other recently published studies have failed to demonstrate a significant increase in cytokine levels during graft reperfusion16 or a cytokine gene polymorphism associated with PRS.17 The majority of the trials demonstrated increased levels of cytokines later in the hospital course.9, 18 These studies found that cytokine levels increased at least 1 hour after graft reperfusion and persisted for several days after surgery.
In our study, cytokines were detected directly in the flush blood (washout blood obtained through a catheter placed within the graft vena cava to remove the preservation solution from the donor organ). These levels were compared with those obtained from the radial artery and portal vein before and after reperfusion. This technique is unique and has not been previously reported.
Levels of TNF-α, IL-1β, and IL-8 in the flush blood were indeed significantly increased. Especially notable were the data demonstrating a highly significant increase in the TNF-α level in the flush blood versus the portal vein and systemic circulation. Our results show an association between the levels of TNF-α released from the liver graft at the end of liver flushing and the amount of NE used to treat reperfusion-induced hypotension. The concentration of TNF-α in these samples represents the amount of TNF-α in the systemic circulation immediately after graft reperfusion. At this time, hemodynamic instability is most pronounced. An analysis of the cytokine dynamics has demonstrated that this TNF-α elevation (measured in peripheral arterial blood samples) persists for several minutes after IVC unclamping but decreases with time. A statistically significant correlation between the level of TNF-α 20 minutes after liver graft reperfusion and the administered amount of NE could not be demonstrated. The relationship between TNF-α and hemodynamic instability has been extensively studied. Caorsi et al.19 showed an association between increased levels of TNF-α and decreases in the systemic vascular resistance and the left ventricular stroke work index. These hemodynamic changes were thought to be mediated by nitric oxide release secondary to increased TNF-α levels.20-22
Our study did not demonstrate any difference in cytokine levels in the portal vein and systemic circulation with the exception of IL-6. This indicates that the liver graft itself is a major source of increased cytokine levels.
The present study demonstrated a significantly increased IL-6 level in the portal vein, and this suggests that cross-clamping of the portal vein and congestion in the splanchnic circulation might be potential causes of increased IL-6 levels. This finding is consistent with results shown in another investigation. Arranz et al.1 obtained samples from the portal vein, vena cava, and pulmonary artery, and the highest level of IL-6 was found in the portal vein. They demonstrated a significant correlation between basal levels of IL-6 and low systemic vascular resistance during hepatectomy, the anhepatic phase, and liver graft reperfusion in patients with cirrhosis. It is not completely understood why increases in IL-6 were not seen in the flush blood. IL-6 may be more prone to metabolic degradation and thus may be completely metabolized by either the warm flush blood or the liver graft itself during the flushing process.
An unusual pattern of IL-8 distribution during liver flushing was noted. Our data demonstrated equal and relatively low levels of IL-8 in the radial artery versus the portal vein before reperfusion and in the flush blood after the initiation of liver graft irrigation. Significant increases in the IL-8 level were detected at the end of graft flushing. The cause of this delay could be associated with the pattern of blood distribution in the different parts of the graft or with stronger binding of preformed cytokines to liver proteins. It is unlikely that the acute production of cytokines during graft reperfusion is responsible for PRS because of the very short time difference between the initiation of graft flushing and the appearance of cytokines in the flush blood.3, 4 Despite a previously described association between increased levels of IL-8 and arterial hypotension,23 we were unable to demonstrate a significant correlation between the levels of this cytokine and hemodynamic instability.
Our study did not show that longer preservation times led to increased cytokine levels. A correlation between the cytokine concentrations and CIT was also not demonstrated. Additional analysis revealed that the levels of cytokines were higher in ECD grafts versus SCD grafts, but we failed to demonstrate a statistically significant difference between these 2 groups. This suggests that SCD grafts are associated with reduced cytokine levels, although these data are insufficient for clear conclusions to be drawn.
PRS not only is responsible for hypotension during graft reperfusion but also seems to have an impact on long-term postoperative outcomes. In a study of 75 patients with cirrhosis undergoing LT, those who experienced PRS had a lower 5-year survival rate than those without PRS.13 Hilmi et al.24 found that the days in the intensive care unit, the total number of hospital days, and the days on the ventilator were greater in patients with PRS.24 This is one of the reasons that the identification of factors contributing to the development of PRS has become the target of several investigations.
We have demonstrated that flush blood from the liver graft during reperfusion contains significantly higher concentrations of cytokines in comparison with those found in the patient's systemic circulation before reperfusion. This excess of cytokines is not completely removed by graft irrigation before the completion of the IVC anastomosis. Our study has demonstrated that the release of TNF-α from a liver graft contributes to the development of PRS. If this finding is confirmed in future investigations, different flushing techniques or pharmacological interventions might be considered to prevent hemodynamic instability at the time of graft reperfusion.
The services provided by the General Clinical Research Center of Pennsylvania State University in analyzing blood cytokine levels are appreciated. In particular, the authors are grateful for the research assistance of Rick Ball (General Clinical Research Center, Pennsylvania State University).