Administration of vasopressors or inotropes during liver transplant surgery is almost universal, as this procedure is often accompanied by massive haemorrhage, acid–base imbalance, and cardiovascular instability. However, the actual agents that should be used and the choice between a vasopressor and an inotrope strategy are not clear from existing published evidence. In this prospective, randomised, controlled and single-blinded study, we compared the effects of a vasopressor strategy on intra-operative blood loss and acid–base status with those of an inotrope strategy during living donor liver transplantation. Seventy-six adult liver recipients with decompensated cirrhosis were randomly assigned to receive a continuous infusion of either phenylephrine at a dose of 0.3–0.4 μg.kg−1.min−1 or dopamine and/or dobutamine at 2–8 μg.kg−1.min−1 during surgery. Vascular resistance was higher over time in the phenylephrine group than in the dopamine/dobutamine group. Estimated blood loss was significantly lower in the phenylephrine group than in the dopamine/dobutamine group (mean (SD) 4.5 (1.8) l vs 6.1 (3.4) l, respectively, p = 0.011). Patients in the phenylephrine group had lower lactate levels in the late pre-anhepatic and the early anhepatic phase and needed less bicarbonate administration than those in the dopamine/dobutamine group (median (IQR [range]) 40 (0–100 [0–160]) mEq vs 70 (40–163 [0–260]) mEq, respectively, p = 0.018). Postoperative clinical outcomes and laboratory-measured hepatic and renal function did not differ between the groups. Increased vascular resistance and reduction of portal blood flow by intra-operative phenylephrine infusion is assumed to decrease the amount of intra-operative bleeding and thereby ameliorate the progression of lactic acidosis during liver transplant surgery.
Patients undergoing liver transplantation for decompensated liver cirrhosis invariably demonstrate pathophysiological circulatory changes. Elevated intrahepatic vascular resistance induces dilatation of extrahepatic vessels, which is aggravated by overproduction and impaired hepatic metabolism of endogenous vasodilators such as nitric oxide . Dilatation of extrahepatic vessels, including portosplanchnic vessels, redistributes body fluid from the central to the peripheral compartment, reducing effective blood flow to the major organs. Additionally, wide surgical dissection of porto-systemic collateral vessels and an inherent haemorrhagic tendency may result in massive blood loss. Clamping or unclamping of great vessels can also produce marked fluid shifts. Thus, effective oxygen delivery and utilisation at the tissue level are impaired, and acid–base balance is frequently abnormal .
Rational fluid and vasopressor/inotrope administration during liver transplant surgery is important to maintain haemodynamic stability and adequate perfusion to major organs. Inotropes such as dopamine and dobutamine have been mainstays for the management of haemodynamic instability in these patients. However, fluid administration in patients with cirrhosis leads to increased splanchnic volume , and therefore some centres advocate the preferential use of vasopressors [4, 5]. Therefore, α1-agonists such as noradrenaline and phenylephrine have now been adopted into clinical practice. Research has demonstrated that α1-agonists reduce portal pressure by reducing portal tributary blood flow . In an animal study, synephrine, a sympathomimetic compound with α-adrenoceptor-mediated vasoconstrictor activity, significantly ameliorated portal venous pressure and flow . Methoxamine, a potent α1-adrenoceptor, decreased hepatic venous pressure gradient and hepatic blood flow in a clinical study .
Based on these observations, vasopressor use during liver transplant surgery may potentially have beneficial effects compared with the use of inotropes, possibly because such a strategy may increase systemic vascular resistance (SVR), and redistribute hepato-splanchnic blood to the central compartment. These effects may therefore decrease blood loss during the procedure as well as potentially improving haemodynamic parameters, acid–base balance, and oxygen supply and demand. Therefore, we performed a prospective, randomised, controlled study to determine whether intra-operative vasopressor infusion using phenylephrine produces clinically beneficial effects over an inotrope infusion using dopamine and/or dobutamine during liver transplant surgery.
This prospective study was conducted in 101 patients with decompensated liver cirrhosis undergoing living donor liver transplantation (LDLT) at Seoul St. Mary’s Hospital between June 2007 and March 2009. The study protocol was approved by the Institutional Review Board and was registered at a publicly accessible clinical trial registration site. Written informed consent was obtained from all patients before enrolment. Exclusion criteria were: systemic vascular resistance index (SVRI) ≥ 1700 dyne.s.cm−5.m−2; pulmonary vascular resistance index ≥ 150 dyne.s.cm−5.m−2; a history of angina or myocardial infarction; evidence of cardiac decompensation; severe haemodynamic instability; pre-operative vasopressor administration; primary renal dysfunction or hepatorenal syndrome; retransplantation; and age under 18 years. Patients were randomly allocated into either a phenylephrine or a dopamine/dobutamine group via computer-generated random numbers. Patients and the operators were unaware of group assignments.
Right-lobe LDLT using a piggyback technique was performed by two operators. During dissection of the recipient’s liver, bleeding was controlled with haemostatic clips or hand ties. In all cases, portal vein anastomosis was followed by hepatic artery anastomosis and bile duct reconstruction. All donor liver grafts were prepared with histidine–tryptophan–ketoglutarate solution.
Anaesthetic technique (sodium thiopental, atracurium, isoflurane) was standardised for all patients. A 22-G angiocatheter (BD Angiocath Plus™; Becton Dickinson Medical Pte, Ltd., Singapore) was inserted into the right radial artery for continuous blood pressure monitoring and arterial blood gas analysis. A 9-Fr multi-lumen introducer (MAC™; Arrow International Inc., Reading, PA, USA) was inserted into the right internal jugular vein and attached to a rapid infusion system (FMS2000; Belmont Instrument Corp., Billerica, MA, USA) for infusion of blood and fluid. A pulmonary artery catheter (Arrow International) was inserted via the introducer and was used to monitor and measure haemodynamic variables. The right subclavian vein was catheterised with a two-lumen catheter (Arrow International) and was used for drug administration.
Intravenous fluids were given to maintain central venous pressure between 5 and 10 mmHg. Crystalloid solutions included 0.45% sodium chloride and Normosol-R pH-7.4 (Plasma solution A; CJ Pharma, Eumseong-gun, South Korea). Hydroxyethyl starch (6%; 130/0.4) solution (Voluven; Fresinius Kabi, Bad Hombourg, Germany) and 20% albumin were used when intravascular volume expansion was urgently needed. The maximum dosage of 6% hydroxyethyl starch permitted was 20 ml.kg−1. Packed red blood cells (PRBCs) were transfused to maintain a target haematocrit of 30%. Platelet concentrate, fresh frozen plasma (FFP) and cryoprecipitate were administered under the guidance of thromboelastography (TEG; Thromboelastograph Analyzer 5000; Haemoscope Corp., Niles, IL, USA) and a laboratory-based coagulation profile. Five units of platelet concentrate were transfused when the maximum amplitude of the TEG was <45 mm (normal range, 48–60 mm) and the platelet count was <40 × 109.l−1 . If bleeding persisted, two units of FFP were given when the reaction time of TEG was >30 mm (normal range, 19–28 mm) and the prothrombin time as the international normalised ratio was >2.0. Five units of cryoprecipitate were given when the α-angle of the TEG was <30° (normal range, 29–43°) and the fibrinogen level was <100 mg.dl−1, with persistent bleeding despite administration of platelet concentrate and FFP. Tranexamic acid was administered using a loading dose of 10 mg.kg−1, followed by infusion at 1 mg.kg−1.h−1 during the rest of the operation, when the percentage reduction in maximum amplitude after 60 min was >15% (normal range, <15%) . Severe metabolic acidosis was corrected with sodium bicarbonate when pH was <7.2 and the base deficit was >10 mmol.l−1. The amount of sodium bicarbonate administered was calculated as follows: ½ × (base deficit – 10) × 0.3 × body weight (mmol).
During surgery, either vasopressors or inotropes were infused to maintain haemodynamic stability. Phenylephrine at a starting dose of 0.3 μg.kg−1.min−1 up to 0.4 μg.kg−1.min−1 was continuously infused to the patients in the phenylephrine group Patients in the dopamine/dobutamine group received dopamine and/or dobutamine at a starting dose of 2 μg.kg−1.min−1 after induction of anaesthesia and up to 5–8 μg.kg−1.min−1 during surgery. In both groups, dosage was adjusted to maintain the mean arterial pressure >65 mmHg, along with adequate volume resuscitation. The total amount of vasopressor or inotrope infused was measured at the end of the procedure. If hypotension was unresponsive to the vasopressor or inotrope infusion, a 10–50 μg bolus of adrenaline was given. If mean arterial pressure was >100 mmHg and SVRI was >2600 dyne.s.cm−5.m−2, we stopped infusing vasopressors or inotropes. To minimise the effects on blood flow to the liver graft, the infusion of vasopressors or inotropes was stopped 2 h after reperfusion of the graft, except when haemodynamic instability necessitated sustained infusion of vasopressors or inotropes .
The volume of blood loss, fluid requirement, transfusion and urine output were recorded to assess intra-operative fluid shifts. Blood loss was estimated as follows: fluid in suction bottles – irrigated fluid – ascites + increased weight of gauze swabs. This was measured every 2 h by an anaesthetic nurse who was blind to the study drugs.
Haemodynamic variables included mean arterial pressure, heart rate, central venous pressure, cardiac index, SVRI, pulmonary vascular resistance index, left ventricular stroke work index, and right ventricular stroke work index. Cardiac output was measured by the thermodilution method using a pulmonary artery catheter. The other haemodynamic variables were calculated using standard formulae. The variables for acid–base balance were base excess and lactate concentration, which were determined by arterial blood gas analysis. The total amount of bicarbonate needed to overcome severe metabolic acidosis was also recorded. To evaluate global tissue oxygenation, arterial and mixed venous blood samples were analysed (ABL800 FLEX; Radiometer, Brønshøj, Denmark). Arterial oxygen tension, oxygen delivery index and oxygen consumption index were calculated from standard formulae. Data were collected at the following time points: immediately after induction of anaesthesia and before infusion of study drugs; 1 and 2 h after starting infusion of study drugs; 1 h after clamping the portal vein; 10 min before reperfusing the liver graft; and 1 and 2 h after reperfusing the liver graft. At each time point, we waited for haemodynamic stability for at least 1 min before taking measurements. We also assessed for the occurrence of post-reperfusion syndrome, which was defined as a >30% decrease in central venous pressure lasting longer than 1 min and occurring within 5 min of graft reperfusion .
Patients were followed up for 30 days after surgery. Death, duration of mechanical ventilation, ICU stay, re-operation, postoperative bleeding, graft-related complications, vascular thrombosis, cardiovascular events, renal dysfunction and biliary complications were included in the postoperative outcomes by application of respective clinically accepted definitions. Laboratory data for hepatic and renal function were collected before surgery, immediately after surgery, and on postoperative days 1, 2, 3, 7 and 30.
From our preliminary observations, we estimated the SVRI before dopamine/dobutamine or phenylephrine infusion to be 1475 dyne.s.cm−5.m−2 with a SD of 490 dyne.s.cm−5.m−2. Assuming a 25% difference in SVRI between the two groups after study drug infusion to be clinically significant, a minimum sample size of 34 patients was calculated for each group, with a type-1 error of 0.05 and a power of 0.8. To analyse intergroup differences, the unpaired t-test was used for other continuous variables and the chi-squared test or Fisher’s exact test for categorical variables. We used ANOVA to compare haemodynamics, acid–base balance, global tissue oxygenation, and peri-operative data for hepatic and renal function between groups over time. For any significant difference over time or any significant interaction of the two groups, an unpaired t-test with Bonferroni correction was performed. Statistical analyses were performed using SPSS for Windows version 18.0 (SPSS Inc., Chicago, IL, USA). In all analyses, p <0.05 was taken to indicate statistical significance.
Among 101 patients who underwent LDLT, 76 were randomly allocated into two groups of 38 each and were included in the analysis (Fig. 1). There was no difference in baseline data, Child–Turcotte–Pugh classification, model for end-stage liver disease score, indications for LDLT or dose of rescue adrenaline between the two groups (Tables 1 and 2). The groups were also similar in terms of duration of hepatectomy, graft ischemia, surgery, anaesthesia, graft-to-recipient weight ratio, and pre-operative and postoperative haemoglobin concentration/haematocrit (Table 3).
Table 1. Baseline characteristics of 76 patients who received either phenylepherine or dopamine/dobutamine infusion during living donor liver transplantation surgery. Values are number (proportion) or mean (SD).
Phenylephrine (n = 38)
Dopamine/dobutamine (n = 38)
CTP, Child–Turcotte–Pugh classification; MELD, model for end-stage liver disease score.
Indications for LDLT
Chronic active hepatitis B
Chronic active hepatitis C
Primary biliary cirrhosis
Acute hepatic failure
Cirrhosis of unknown origin
Table 2. Amount of vasoactive or inotropic drugs administered in 76 patients who received either phenylepherine or dopamine/dobutamine during living donor liver transplantation. Values are number (proportion), mean (SD) or median (IQR [range]).
Phenylephrine (n = 38)
Dopamine/dobutamine (n = 38)
Total dose; mg
Total dose; mg
Dopamine and dobutamine
Total dose of dopamine; mg
Total dose of dobutamine; mg
Total dose; mg
Bolus adrenaline dose; μg
Immediate post-reperfusion period
20 (0–100 [0–160])
40 (0–120 [0–180])
After immediate post-reperfusion period
10 (0–80 [0–200])
30 (0–80 [0–200])
Table 3. Intra-operative data for 76 patients who received either phenylepherine or dopamine/dobutamine during living donor liver transplantation. Values are mean (SD) or median (IQR [range]).
Phenylephrine (n = 38)
Dopamine/dobutamine (n = 38)
Duration of hepatectomy; min
Graft ischaemic time; min
Duration of surgery; min
Duration of anaesthesia; min
Graft-to-recipient weight ratio; %
Preoperative haemoglobin; g.dl−1
Postoperative haemoglobin; g.dl−1
Estimated blood loss; l
Fluid requirements; l
0.45% sodium chloride
6% hydroxyethyl starch
Transfused blood components; units
Packed red cells
Fresh frozen plasma
0 (0–5 [0–15])
0 (0–5 [0–15])
0 (0–5 [0–15])
0 (0–5 [0–15])
Urine output; l
Total dose of tranexamic acid; mg
Estimated blood loss and units of packed red blood cells and FFP transfused were significantly lower in the phenylephrine group than in the dopamine/dobutamine group (Table 3). Compared with inotrope infusion, phenylephrine infusion during LDLT decreased intra-operative blood loss and transfusion amounts by approximately 25%. Left ventricular stroke work index, right ventricular stroke work index, SVRI and pulmonary vascular resistance index were higher over time in the phenylephrine group than in the dopamine/dobutamine group (p = 0.016, 0.032, 0.013 and 0.015, respectively, Fig. 2), although the patterns were similar between the two groups. The SVRI in the phenylephrine group was higher in the late pre-anhepatic phase and the early anhepatic phase. Heart rate was lower in the phenylephrine group than in the dopamine/dobutamine group (p = 0.001); however central venous pressure and cardiac index did not differ (Fig. 2a).
Lactate levels in the late pre-anhepatic phase and early anhepatic phase were also lower in the phenylephrine group than in the dopamine/dobutamine group (Fig. 2b). Less bicarbonate was required to overcome severe metabolic acidosis in the phenylephrine group than in the dopamine/dobutamine group, median (IQR [range]) 40 (0–100 [0–160]) mmol vs 70 (40.0–162.5 [0–260]) mmol, respectively, p = 0.018. There was no difference in global tissue oxygenation between the two groups (Fig. 2c).
Post-reperfusion syndrome occurred in 15 patients (40%) in the phenylephrine group and 19 patients (50%) in the dopamine/dobutamine group (p = 0.411). One patient in the phenylephrine group and two patients in the dopamine/dobutamine group died within 30 days of the operation. Duration of ICU stay and postoperative mechanical ventilation of the lungs were similar between the two groups, as were hepatic and renal function (Table 4 and Fig. 3).
Table 4. Postoperative outcomes following living donor liver transplantation in 76 patients who received either phenylepherine or dopamine/dobutamine. Values are mean (SD) or number (proportion).
Phenylephrine (n = 38)
Dopamine/dobutamine (n = 38)
*Some patients developed more than one complication.
ICU stay; days
Mechanical ventilation; h
Early graft dysfunction
We have shown that infusion of phenylephrine, compared with dopamine and/or dobutamine, reduced the amount of intra-operative bleeding and transfusion requirements during LDLT and reduced the severity of metabolic acidosis.
Major surgery in patients who already have a tendency towards vasodilatation, such as cirrhotic patients undergoing liver transplantation, is associated with a frequent requirement for intra-operative administration of either inotropes or vasopressors [5, 13, 14]. Although a previous study has suggested that the response to vasoconstrictors may be blunted in cirrhosis , other more recent studies have disputed this [16, 17]. Noradrenaline and phenylephrine are the most commonly used vasopressors in modern clinical practice. Phenylephrine, a synthetic catecholamine, which is a powerful α1-agonist but has little β-effect, has been particularly widely used in septic shock. Compared with noradrenaline, phenylephrine causes more profound hepatosplanchnic vasoconstriction in such patients . Although there are few reports in the literature, phenylephrine is also frequently used during liver transplantation. It has been shown to cause vasoconstriction in mesenteric and hepatic arteries in cirrhotic patients .
In our study, phenylephrine did indeed increase vascular resistance, as indicated by the increased SVRI and the pulmonary vascular resistance index; however, these did not increase beyond the normal range. In decompensated cirrhotic patients with portal hypertension, optimising SVR with vasopressors such as phenylephrine has several potential benefits. In these patients, therapeutic expansion in central blood volume is not sustained and fluid is rapidly redistributed to the hepatosplanchnic circulation . During liver transplant surgery, clamping of the inferior vena cava is known to compromise cardiac function. At this point, increasing the SVR without the use of venovenous bypass is beneficial, because infusion of a significant volume of intravenous fluid may precipitate right heart dysfunction after reperfusion of the donor liver . Phenylephrine has been shown to increase cardiac preload in the absence of fluid administration, suggesting an effect on venous compliance . Although we failed to find a significant increase in cardiac preload, as estimated by the central venous pressure, phenylephrine infusion did elevate left and right ventricular stroke work index, which is consistent with a previous study carried out in patients with septic shock . Although the phenylephrine dose in this study (0.3–0.4 μg.kg−1.min−1) was lower than that used in septic shock, it was generally effective in optimising vascular resistance within the normal range in cirrhotic patients undergoing LDLT when we carried out a pilot study. We decided not to infuse a higher dose because of the risk of compromising graft perfusion or renal function.
According to several retrospective studies, intra-operative blood transfusion is independently associated with outcome after orthotopic liver transplantation [21–23]. However, little is known about pre-operative factors associated with increased blood loss . Placement of an intra-operative portacaval shunt has been found to reduce transfusion, and this was attributed to the reduction of portal pressure . In our study, phenylephrine infusion reduced both intra-operative bleeding and transfusion of red blood cells and FFP. We hypothesise that this was due to the reduction in portal blood flow and pressure in these patients compared with those who received dopamine and/or dobutamine. One possible confounding factor is that the amounts of bleeding and transfusion in our study were considerably higher than those in other reports [14, 22]. This may be attributed to the technical differences between LDLT and deceased donor liver transplantation and different strategies for fluid administration and blood transfusion. In addition, our study was not powered for this outcome, and therefore the significance of this finding is not certain.
In patients undergoing liver transplantation, the degree of metabolic acidosis is related to the accumulation of lactic acid, indicating that the major aetiology is lactic acidosis . Bleeding from surgical dissection of the intra-abdominal portal collaterals during the pre-anhepatic phase is associated with lactate accumulation, and this is further aggravated by loss of hepatic metabolic function during the anhepatic phase. A burst of acidic substances from congested splanchnic blood causes the pH to reach a nadir in the first few minutes after graft reperfusion. We have shown that patients managed with phenylephrine did not become as profoundly acidotic as those managed with dopamine and/or dobutamine. The observed differences in base excess and lactate occurred during the late anhepatic phase, and disappeared after portal reperfusion. We believe that the difference in the amounts of bleeding and transfusion between the two groups was the main cause of the difference in the severity of metabolic acidosis, and that the recovery of lactic acid metabolism after portal reperfusion led to normalisation of plasma pH. As a result, the amount of bicarbonate needed to overcome severe metabolic acidosis was significantly lower in the phenylephrine group than in the dopamine/dobutamine group. A lower amount of sodium bicarbonate is desirable because its high sodium content may be problematic in hyponatremic cirrhotic patients.
Global tissue oxygenation was not significantly different between the two groups. In humans, 25% of oxygen consumption results from hepatosplanchnic metabolism. Patients undergoing liver transplantation show decreased global oxygen consumption, which is due to low carbohydrate and protein metabolism in the liver . Decreasing hepatosplanchnic circulation by the use of vaspressors might be expected to optimise oxygen utilisation, resulting in better perfusion and oxygenation of other organs. However, we could not confirm this theory because we did not measure liver tissue oxygenation directly, as this may cause a number of complications [27, 28]. In addition, we did not observe a difference in the incidence of post-reperfusion syndrome, which might have been expected with a reduction in portal hypertension and splanchnic hyperaemia associated with a vasopressor strategy. Although factors causing hypotension after reperfusion of the liver graft are yet to be defined, the potent vasodilatory action of bradykinin, a product of the kallikrein–kinin system, is considered to be the most likely cause of post-reperfusion syndrome .
This study has a number of limitations that must be considered. First, the study was not performed in a double-blind manner. However, we adhered strictly to standardised intra-operative strategies, including maintaining central venous pressure and target haematocrit, and correcting abnormal thromboelastographic findings and coagulation profiles with fluids and blood products as indicated. The administration of additional inotropes and other drugs such as sodium bicarbonate and antifibrinolytic agents was also strictly controlled. Second, we did not insert a catheter for measurement of portal vein pressure. Future study of the pharmacological effect of phenylephrine on portal pressure and portal blood flow and the relationships between portal vein pressure and peri-operative variables is required. Third, the design of the study did not allow the determination of the optimal dose of phenylephrine infusion. Although excessive administration of vasopressors may have unfavourable effects on graft liver perfusion, further study to establish the optimal dose is required.
In conclusion, intra-operative infusion of phenylephrine, an α1-agonist, appears to be an effective and safe technique for reversing vasodilatation in cirrhotic patients undergoing LDLT. This technique is associated with reduced intra-operative bleeding, transfusion and metabolic acidosis compared with an inotrope-based strategy.
This study was funded entirely by the Department of Anesthesiology and Pain Medicine at Seoul St. Mary’s Hospital. No competing interests declared.