Hemodynamic profile and tissular oxygenation in orthotopic liver transplantation: Influence of hepatic artery or portal vein revascularization of the graft
We performed a prospective, randomized study of adult patients undergoing orthotopic liver transplantation, comparing hemodynamic and tissular oxygenation during reperfusion of the graft. In 30 patients, revascularization was started through the hepatic artery (i.e., initial arterial revascularization) and 10 minutes later the portal vein was unclamped; in 30 others, revascularization was started through the portal vein (i.e., initial portal revascularization) and 10 minutes later the hepatic artery was unclamped. The primary endpoints of the study were mean systemic arterial pressure and the gastric-end-tidal carbon dioxide partial pressure (PCO2) difference. The secondary endpoints were other hemodynamic and metabolic data. The pattern of the hemodynamic parameters and tissue oxygenation values during the dissection and anhepatic stages were similar in both groups At the first unclamping, initial portal revascularization produced higher values of mean pulmonary pressure (25 ± 7 mm of Hg vs. 17 ± 4 mm of Hg; P < 0.05) and wedge and central venous pressures. At the second unclamping, initial portal revascularization produced higher values of cardiac output and mean arterial pressure (87 ± 15 mm of Hg vs. 79 ± 15 mm of Hg; P < 0.05) and pulmonary blood pressure. Postreperfusion syndrome was present in 13 patients (42.5%) in the arterial group and in 11 patients (36%) in the portal group. During revascularization, the values of gastric and arterial pH decreased in both groups and recovered at the end of the procedure, but were more accentuated in the initial arterial revascularization group. In conclusion, we found that initial arterial revascularization of the graft increases pulmonary pressure less markedly, so it may be indicated for those patients with poor pulmonary and cardiac reserve. Nevertheless, for the remaining patients, initial portal revascularization offers more favorable hemodynamic and metabolic behavior, less inotropic drug use, and earlier normalization of lactate and pH values. Liver Transpl, 2006. © 2006 AASLD.
Provide feedback or get help You are viewing our new enhanced HTML article.
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 period of greatest hemodynamic instability during orthotopic liver transplantation occurs at graft reperfusion.1 After unclamping, marked vasodilatation of the splanchnic bed produces an insufficient preload and, consequently, hemodynamic instability,2, 3 which is magnified in those patients with lower cardiopulmonary reserve or with significant blood loss during the hepatectomy. Other factors possibly involved are substances coming from the graft, such as potassium,1 which are related to the venous stasis of the portal territory or to preservation injury of the graft. This hemodynamic instability may be harmful to patients,4 and may also contribute to ischemia-reperfusion injury, which leads to significant graft dysfunction in the postoperative period.5
A classic surgical technique used in orthotopic liver transplantation consists of unclamping the portal vein as soon as the anastomosis is complete, keeping portal occlusion time as brief as possible. After this, arterial anastomosis is performed and the hepatic artery is unclamped. By performing portal anastomosis first, the duration of splanchnic congestion and portal hypertension secondary to vena porta clamping is shortened, as is the entire anhepatic stage. In addition, as initial reperfusion through the portal vein alone is technically easier, it has been used to minimize cold and warm ischemia time.
Today, simultaneous arterial and portal anastomosis is feasible due to the hemodynamic improvement offered by the preservation of the vena cava during the anhepatic phase6; also, the practice of a portocaval shunt minimizes portal congestion even though the duration of the anhepatic stage is longer.7
As far as we know, few controlled studies comparing the hemodynamic effects of the order of graft revascularization have been performed. One study by a French group8 and 1 by a Scottish group have been published,9 but their data were inconclusive.
The aim of our study was to compare hemodynamic and tissular oxygenation at reperfusion of the graft in both situations, i.e., when the hepatic artery reperfusion of the graft was done first and when initial portal vein reperfusion was done first. Thus, we conducted a prospective, randomized, and controlled study on adult patients undergoing orthotopic liver transplantation, comparing hemodynamic and tissular oxygenation during reperfusion of the graft and at the end of the procedure.
MATERIALS AND METHODS
After the Institutional Review Board approved the study, written consent was obtained from all patients. A prospective, randomized study was performed on all consecutive patients undergoing orthotopic liver transplantation from June 2001 to December 2002 in a single adult liver transplantation center. Our exclusion criteria were: 1) primary familiar amyloidotic neuropathy, 2) Budd-Chiari syndrome, 3) acute liver failure, 4) early retransplantation (less than 1 month), and 5) simultaneous kidney and liver transplantation.
General anesthesia was induced with 2 mg/kg of propofol, 0.3 mg of fentanyl, and 0.6 mg/kg of rocuronium. After tracheal intubation, a continuous infusion of propofol and remifentanil was started to maintain the bispectral index analysis of the electroencephalogram (Aspect, Leiden, The Netherlands) between 40 and 60 and a continuous infusion of rocuronium to maintain the first stimulus response of the train of 4 ratio between 10 and 25% of the control. Mechanical ventilation was begun at 10 mL/kg with a respiration rate to obtain an end-tidal partial pressure of CO2 around 35 mm of Hg and an inspired oxygen fraction of 0.5 in air. Calcium was administered to maintain ionized calcium levels at approximately 1.2 mmol/L and sodium bicarbonate was administered to reach a pH greater than 7.30. A rapid infusion system attached to a 10 French catheter was used to infuse isotonic saline solution. Fluid infused (including drug administration) was kept at 7 mL/kg/hour. In those situations when a low preload was detected, a rapid infusion of gelatins was administered to maintain central venous pressure (CVP) above 8 mm of Hg. In case of high CVP values (>15 mm of Hg), a diuretic (20 mg of furosemide) was administered. Losses related to ascitis were corrected by the administration of 200 mL of albumin 20%. Packed red blood cells were administered in order to maintain hemoglobin level at 100 gm/L. Fresh frozen plasma was administered only when the international normalized ratio was −1.8. Platelets were administered to maintain a platelet count above 50 × 109/L and fibrinogen was administered to maintain fibrinogen levels above 1 gm/L. No intraoperative salvage of blood was used during surgery. All patients received a continuous infusion of 10 mg/kg/hour of Tranexamic acid, an antifibrinolytic drug, from the induction of anesthesia until 2 hours after graft reperfusion. All patients were placed on a warm blanket of convection air (Warm-Touch, Mallincrod Medical, St. Louis, MO), their lower limbs covered with cotton and had all intravenous fluids administered through a fluid warmer. A methylprednisolone bolus of 500 mg was administered at the time of first reperfusion, as part of the immunosuppressive regimen.
Liver allografts were preserved using University of Wisconsin solution.10 Elderly donors were admitted and ischemic time was intended to be less than 10 hours. In all surgical procedures, temporary portocaval anastomoses with preservation of caval vein flow was performed to minimize the stasis of the portal territory and to maintain hemodynamic stability during the anhepatic stage. Patients in whom at portocaval shunt was not technically possible were excluded from the study and not randomized.
In all cases the 2 anastomoses were performed before reperfusion of the graft. Hepatic artery anastomosis was done first followed by portal vein anastomosis, at which time the portocaval shunt was suppressed to complete the portal anastomosis.
Prior to reperfusion of the graft, the liver was flushed with 1,000 mL of Ringer lactate solution at 38°C to remove air and detritus through the entire inferior vena cava of the graft. Next, the distal end of the donor's vena cava was closed by a vascular stapler. Once we reached this point, the sealed randomization envelope was opened and patients were assigned to 1 of 2 groups. The initial arterial revascularization group, with reperfusion started first through the hepatic artery and 10 minutes later through the portal vein; and the initial portal revascularization group, with reperfusion started first through the portal vein and 10 minutes later through the hepatic artery. The time between the first reperfusion and the second reperfusion was identical in both groups: it took 10 minutes by protocol.
Management of Postreperfusion Syndrome
We defined postreperfusion syndrome (PRS) as when the mean arterial blood pressure was 30% down from the previous value and lasted for almost 1 minute within the first 5 minutes after unclamping.1 Before unclamping, a rapid infusion of gelatins (Haemocell) was administered to maintain CVP above 8 mm of Hg. After unclamping, patients whose mean blood pressure was lower than 60 mm of Hg were treated first with a bolus of 10 mg of ephedrine intravenously. In those patients that did not recover promptly, an infusion of dopamine at 10 μg/kg/minutes was started. In critical situations (bradycardia) or on nonresponse to the previous treatment, an intravenous bolus of 0.2 mg of epinephrine was administered as required. Depending on the cause of the hemodynamic disturbances, fluids and/or inotropic drugs were continued until complete restoration of the patient.
Data Collection and Endpoints
The primary endpoints of the study were the mean systemic arterial pressure (, a simple measure of organ perfusion) and the gastric-end-tidal carbon dioxide partial pressure (PCO2) difference (1 of the parameters derived from the gastric PCO2measure, an expression of tissular perfusion). These were considered the main variables of the study. The secondary endpoints were other hemodynamic and metabolic data, which consisted of: heart rate; mean pulmonary pressure; CVP; pulmonary capillary pressure; systemic vascular resistance, and pulmonary vascular resistance; cardiac output; hemoglobin level; oxygen consumption, and oxygen delivery; plasma lactate levels; end-tidal PCO2, gastric PCO2, arterial PCO2, arterial pH, and gastric mucosal pH.
Data were collected at the following points: T1, when surgery starts; T2, when the portocaval shunt was open; T3, when the portocaval shunt was closed; T4, during the first unclamping; T5, at the second unclamping (depending on the study group); and T6, at the end of the surgical procedure.
Additionally, end-tidal PCO2 and the hemodynamic profile were recorded continuously (Datex-Ohmeda S/5 monitor, Louisville, CO); semicontinuous air samples of gastric PCO2 were taken and measured by a tonometry catheter (Datex-Ohmeda TONO-14F); and derivative parameters and gradients were calculated. Samples were collected every 10 minutes throughout the procedure. Intraoperative nasopharyngeal temperature profile and blood temperature were recorded. Continuous cardiac output was collected every minute by a thermal catheter (Swan-Ganz Baxter 744HF75, Deerfield, IL). At each collection time, data were obtained when catheter thermal noise disappeared,11 usually between 5 and 8 minutes after each change. In all determinations, we choose the lowest values for mean systemic blood pressure and the highest values for mean pulmonary blood pressure within 10 minutes after each point of the study. Because of thermal noise, we selected the highest cardiac output values between minute 8 to minute 10 after each point of the study. Gastric PCO2 and biochemical data were taken simultaneously 10 minutes after each event. Also, vasoconstrictor and inotropic drugs during both unclamping periods were recorded.
Other variables recorded were recipient preoperative variables: age, gender, indication for orthotopic liver transplantation, Child-Turcotte-Pugh score, Model for End-Stage Liver Disease score, and United Network for Organ Sharing score, history of upper abdominal surgery, presence of portal thrombosis, coagulation profile, hematocrit and hemoglobin levels, platelet count, and presence of associated medical condition. The time between the start of anastomosis and the first unclamping was considered as warm ischemia time; therefore, the time between the start of anastomosis and both unclampings (including temporary clamps) was registered and named total warm ischemic time. Cold ischemia time, fluid reposition, blood product requirements, and sodium bicarbonate and calcium administration were also recorded. Hepatic artery and portal vein flows were measured after each reperfusion with a transit-time ultrasound (Transonic Flowmeter HT311, Ithaca, NY). Biliary strictures, ischemic injury to the biliary duct, biopsy-proven rejections, and bacterial infective processes were also recorded. The duration of postoperative mechanical ventilation, intensive care unit stay, and length of stay in the hospital were also recorded.
All groups were evaluated according to intent to treat. The number of patients to be studied was determined by the following parameters: mean arterial pressure difference of 20 mm of Hg at reperfusion, a 2-tailed alpha error of 0.05, and a beta error of 0.2, giving a total sample size of 30 patients per group.
Continuous variables are given as means and standard deviation. Because of nonhomogenous distribution, data from the acid-base profile are given as median and quartiles. Nonparametric test (Mann-Whitney U test) analysis was used to compare groups. Discontinuous variables were given in percentages and the chi-squared (χ2) test was used to compare the groups. Differences with probability values of 0.05 or less were considered significant.
During the period of study, 83 patients received an orthotopic liver transplant. Then patients did not have the portocaval shunt performed, and so were excluded from randomization. Thirteen patients were excluded because of acute hepatic failure (2 cases), early retransplantation (3 cases), primary amyloidotic neuropathy (6 cases), and simultaneous liver and kidney transplantation (2 cases). In all excluded patients, the order chosen for graft reperfusion was first portal, so that the arterial anastomoses were done once the portal vein was unclamped.
Sixty patients were included in the study: 30 in the arterial group and 30 in the portal group. Patients in both groups were demographically similar in their age, gender, body mass index, diagnosis, Model for End-Stage Liver Disease score and Child-Turcotte-Pugh score. Preoperative values of hemoglobin were 11 ± 2.2 gm/in the arterial group and 11.3 ± 2.4 gm/L in the portal group. Groups were also similar in graft cold and ischemia time, dissection, anhepatic time, and total surgery time (Table 1). Total graft warm ischemia time was longer than 60 minutes in 80% of all patients in both groups. Fluid and transfusion requirements were also similar (Table 1). A high percentage of patients required no packed red blood cells; 40% in the arterial group and 36% in the portal group. PRS was present in 24 of 60 patients studied, which was not related to total warm ischemia (76.6 ± 26.6 vs. 85 ± 33 minutes; P = 0.231). Also, PRS was present in 1 of 7 patients with graft steatosis >20%, while PRS was present in 23 of 53 patients with graft steatosis <20% (P = 0.230).
Table 1. Demographic data and Intraoperative Data
|Demographic Data|| || |
| Age (yr)*,†||55 (36-67)||53 (36-68)|
| BMI (Kg/m2)||27.7 ± 5.8||26.15 ± 3.7|
| Diagnosis‡|| || |
| Child-Turcotte-Pugh score‡|| || |
| A (5-6 points)||27%||17%|
| B (7-9 points)||50%||56%|
| C (10-15 points)||23%||27%|
| MELD score†||15.1 ± 5.8||15.5 ± 5|
| UNOS score‡|| || |
| At home||61%||70%|
| At hospital||39%||30%|
|Intraoperative data|| || |
| Graft cold ischemia (minutes)||477 ± 195||494 ± 169|
| Duration of hepathectomy (minutes)||164 ± 35||170 ± 52|
| Warm ischemia (minutes)||53 ± 13||57 ± 14|
| Duration of surgical procedure (minutes)||362 ± 64||368 ± 77|
| Fluids requirements (mL)||4392 ± 2491||4488 ± 2094|
| Sodium bicarbonate (mmol)||69 ± 91||72 ± 78|
| RBC (units)||2.1 ± 2.1||2.5 ± 3.5|
| FFP (units)||1 ± 1.9||1.5 ± 2.8|
| Platelets (units)||4.7 ± 5.1||6.2 ± 6.2|
Hemodynamic and Oxygenation Data
The pattern of hemodynamic parameters and tissue oxygenation values during the dissection and anhepatic phase were similar in the 2 groups, with no significant differences found. However, significant differences between the groups were found during graft reperfusion (Table 2). At the first unclamping, initial portal revascularization produced higher values of mean pulmonary pressure, wedge pressure, and CVP, and higher pulmonary vascular resistance. At the second unclamping, differences between the groups were related to cardiac output and mean systemic and pulmonary blood pressure. Therefore, the order of vascular unclamping caused different hemodynamic and oxygenation behavior. Portal and hepatic arterial flow during reperfusion, shown in Table 3, were also similar in the groups.
Table 2. Hemodynamic and Tissue Oxygenation Evolution
|HR (bpm)||79 ± 17||86 ± 16||96 ± 19||97 ± 16||101 ± 20||102 ± 14||123 ± 22||121 ± 18||120 ± 19||118 ± 16||108 ± 13||104 ± 23|
|MAP (mmHg)||77 ± 14||77 ± 14||83 ± 15||84 ± 13||81 ± 13||83 ± 16||76 ± 19||76 ± 16||79 ± 15*||87 ± 15*||81 ± 15||80 ± 16|
|CO (L/minute)||7.6 ± 2.53||9.03 ± 3.38||8.48 ± 3.61||8.69 ± 3.21||8.84 ± 3.84||9.61 ± 3.57||7.43 ± 3.55||8.43 ± 3.3||7.76 ± 3.29*||11.19 ± 3.75*||10.07 ± 3.17||11.39 ± 3.74|
|MPP (mmHg)||19 ± 5||20 ± 7||18 ± 5||21 ± 12||12 ± 5||13 ± 6||17 ± 4*||25 ± 7*||23 ± 5*||28 ± 10*||19 ± 6||21 ± 5|
|PCWP (mmHg)||17 ± 7||13 ± 6||11 ± 4||12 ± 7||7 ± 4||7 ± 6||12 ± 3*||16 ± 6*||16 ± 4||18 ± 6||13 ± 4||13 ± 5|
|CVP (mmHg)||10 ± 4||9 ± 6||8 ± 4||7 ± 5||6 ± 4||6 ± 5||11 ± 4*||13 ± 5*||12 ± 4||13 ± 5||10 ± 4||8 ± 4|
|SVR (dynes* seconds/cm5)||793 ± 334||749 ± 424||798 ± 285||851 ± 407||737 ± 261||747 ± 324||895 ± 450||717 ± 388||770 ± 360*||631 ± 339*||616 ± 249||551 ± 241|
|PVR (dynes* seconds/cm5)||58 ± 30||88 ± 99||75 ± 31||92 ± 80||48 ± 22||57 ± 32||65 ± 57*||98 ± 66*||81 ± 53||97 ± 110||59 ± 45||67 ± 51|
|SVO2 (%)||87 ± 5||88 ± 4||91 ± 3||93 ± 2||88 ± 5||89 ± 4||86 ± 7*||90 ± 4*||89 ± 3||89 ± 4||85 ± 7||86 ± 6|
|Hb (gm/dL)||11.09 ± 2.05||11.02 ± 1.34||10.48 ± 2.22||10.44 ± 2.07||9.68 ± 2.71||10.21 ± 1.92||10.15 ± 1.8||10.19 ± 1.87||9.83 ± 1.79||9.79 ± 1.85||9.58 ± 1.53||9.42 ± 2.17|
|VO2 (mL/minute)||191 ± 88||267 ± 170||173 ± 108||195 ± 186||243 ± 182||243 ± 143||176 ± 82||189 ± 123||181 ± 84*||326 ± 299*||258 ± 98||298 ± 160|
|DO2 (mL/minute)||1205 ± 409||1423 ± 754||1281 ± 442||1345 ± 697||1369 ± 525||1462 ± 743||1091 ± 476||1236 ± 441||1108 ± 465*||1523 ± 627*||1458 ± 481||1626 ± 605|
Table 3. Vascular Graft Flow
|First reperfusion||261 ± 170||—||—||3220 ± 1118|
|Second reperfusion||239 ± 149||2651 ± 1155||235 ± 105||2799 ± 904|
In the 2 unclampings, PRS was present in 13 patients (42.5%) in the arterial group and in 11 patients (36%) in the portal group. More patients in the arterial group received amine drugs to maintain blood pressure; also, epinephrine was only administered in this group during the entire process of revascularization of the graft. However, these differences were not significant. At the first unclamping, in the arterial group (initial arterial revascularization) 7 patients required ephedrine, 5 of these needed an associated infusion of dopamine and 1 additional patient required an intravenous bolus of epinephrine; in the initial portal revascularization group, 4 patients received a bolus of ephedrine and also required infusions of dopamine, and 1 other patient required an infusion of dopamine. At the second unclamping, in the initial arterial revascularization group), 2 patients required a bolus of ephedrine and 1 of these also needed a bolus of epinephrine; in the initial portal revascularization group, 3 patients received an infusion of dopamine.
There were no differences between groups for values of pH, lactate, or gastric end-tidal and arterial PCO2 at the end of the portocaval shunt. During revascularization, the value of graft gastric and arterial pH decreased in both groups and recovered at the end of the procedure. This decrease was more accentuated in the initial artery revascularization group (Table 4). However, values of gastric, end-tidal, and arterial PCO2 increased during revascularization and remained high at the end of the procedure, whereas CO2 gradients remained nearly constant during the procedure (Table 4).
Table 4. Acid-Base Evolution at Reperfusion of the Graft
|Pa CO2 (mm of Hg)||30 (32–35)||31 (28.1–33)||37 (33–39.75)||36 (34–38.75)||38.5 (35.25–41)||36 (34.5–42)||37 (33.25–37.75)||35.5 (33.75–3)|
|PEtCO2 (mm of Hg)||27 (25.5–29.5)||25.5 (23.75–28)||30 (26–32.5)||31 (28.75–34)||33 (31.75–35)||31 (29.75–35)||33 (29–35.5)||32 (28–34)|
|Pg CO2 (mm of Hg)||39 (34.5–42)||38 (33–45)||41 (37–48)||40.5 (35–47.5)||40.5 (38–46.25)||40 (35–46.5)||41.5 (38–46)||41.5 (37–48)|
|P(g–Et)CO2 (mm of Hg)||12 (9–15)||13 (8–17)||12 (7.5–21.5)*||8 (5–12)*||9 (6–12.5)||8.5 (5.5–12)||9 (5–13)||9.5 (7–12)|
|P(g–a)CO2 (mm of Hg)||6 (3–10)||7 (4–10)||4 (2–15)||3 (−1.75–8.75)||3.5 (0.62–10.25)||3 (−1–7.5)||4.5 (2–8)||5 (2.75–10)|
|Gastric pHi||7.27 (7.18–7.32)||7.31 (7.18–7.395)||7.215 (7.16–7.27)||7.245 (7.18–7.30)||7.20 (7.10–7257)*||7.27 (7.16–7.32)*||7.29 (7.22–7.33)||7.32 (7.27–7.3)|
|Arterial pH||7.39 (7.34–7.42)||7.40 (7.33–7.415)||7.31 (7.27–7.34)||7.30 (7.275–7.34)||7.28 (7.24–7.317)*||7.315 (7.277–7.35)*||7.365 (7.325–7.42)||7.385 (7.347–7.)|
|Lactate (mmol/L)||4.53 (3.15–5.6)||3.81 (3.25–5.45)||6.07 (4.67–6.95)||5.57 (4.295–6.58)||5.58 (4.195–6.58)*||4.08 (3.5–5.75)*||4.12 (3–5.56)||3.78 (2.76–4.7)|
|Temperature (°C)||36.1 (35.6–36.4)||36.1 (35.77–36.4)||36.1 (35.5–36.55)||36 (35.1–36.3)||36.05 (35.65–36.52)||36.25 (36–36.6)||36.8 (36–37)||36.7 (36.2–3)|
One patient in the arterial group needed reconstruction of the arterial anastomoses after unclamping, but, no patients needed reoperation due to technical vascular problems. One patient in each group required an intervention because of bleeding in the postoperative period.
Five patients (16%) in the arterial group and 4 patients (13%) in the portal group presented biliary strictures. Ischemic injury was present in 2 patients in the arterial group and none in the portal group. Rejection was detected and treated in 7 patients (23%) in the arterial group and in 8 patients (27%) in the portal group. Infective symptoms were detected in 13 patients (43%) in the arterial group and in 13 patients (43%) in the portal group. Mechanical ventilation was maintained during 32 ± 30 hours in the arterial group and 48 ± 127 hours in the portal group. Patients remained in the intensive care unit for 4.3 ± 2.38 days and 4.9 ± 5.7 days for arterial and portal groups, respectively. Most patients in both groups increased their creatinine value in the first postoperative week. Two patients in the arterial group and 1 patient in the portal group required hemofiltration because of postoperative acute tubular necrosis.
One patient in each group died within 30 days of the operation. No patients required retransplantation. The length of the hospital stay was 13.6 ± 6.7 days and 16.4 ± 11 days for the arterial and portal group, respectively.
In this prospective study, we found that graft reperfusion causes a decrease in mean systemic blood pressure in both groups at the first unclamping. However, at the second unclamping mean blood pressure recovered in the portal group but not in the artery group. This biphasic behavior depends on the order of unclamping. Initial reperfusion with portal vein leads to higher values of preload parameters (CVP and mean pulmonary and wedge pressures) and, thus, higher cardiac output and mean systemic arterial pressure values than initial reperfusion of the graft with the hepatic artery. Durcef et al.8 found similar results at the first reperfusion, but gave no information about the second unclamping. Nevertheless, hemodynamic and metabolic changes could have occurred on the second vascularization of the graft, as occurred in our study (Table 2).
Pulmonary blood volume increases significantly during unclamping.12 In our study, the central venous and mean pulmonary pressures at first reperfusion increased from preceding values (closure of portocaval shunt at anhepatic stage), which means a sudden increase in the blood return to right cardiac chambers. On comparing groups, at the first unclamping, portal revascularization achieved much higher pulmonary pressure than hepatic artery revascularization. These differences were maintained at the second unclamping. However, the rising of pulmonary pressure values at the second unclamping in the artery group indicates that the blood flow from the portal vein was responsible for most of the increase in venous return. This increase in pulmonary pressure was less marked in the artery group than the portal group, so it could protect against pulmonary edema, mainly in patients at risk, i.e., those with lower pulmonary compliance, diastolic failure, or pulmonary hypertension. Despite this, we found more favorable metabolic behavior for the portal group after unclamping, with even less inotropic drug needed. Against this, however, Walsh et al.9 found that the artery group required less inotropic support. This discrepancy between our study and Walsh et al.9 may be related to the methodology used: in their study, the order of the vascular reperfusion was not random. In addition, in Walsh et al.,9 in some cases a portocaval shunt was not added, differences in the duration of anhepatic stage and the time between the first and the second unclamping might be present, so making the results obtained less valuable for comparison.
The values of oxygen consumption and oxygen delivery were higher for the portal group mainly at the second unclamping. However, these values are directly related to cardiac output, and less so to metabolic status. Again, arterial pH, lactate, and venous oxygen saturation were better for the portal group. However, tissular oxygenation was maintained in both groups, and there were no differences in gastric mucosal PCO2 values or related gradients. After graft reperfusion, a metabolic monitor showed13 increased oxygen consumption associated with an increase in cardiac output and simultaneous acid release, with a rise in both sources of acid load, the graft and the ischemic tissues of the recipient. The same results were obtained by Ducerf et al.9 in a more recent study comparing the order of unclamping. Their data suggested a greater increase of arterial PCO2 after reperfusion in the portal vein group than in the artery group. In our study, we found a similar increase of arterial PCO2, independent of the order of unclamping; and parallel to changes in other CO2 values (end-tidal and gastric). An increase in tissular flow (including the splanchnic territory and the graft) may explain the sudden acid release found. In addition, pulmonary pressure and CVP were higher in the portal group and, thus cardiac output and portal flow were as well (Table 3). This could explain a more favorable arterial pH profile in the portal group, mainly once graft reperfusion was completed.
In comparison with the anhepatic stage, during revascularization of the graft the gastric PCO2 increased in both groups (parallel to the values found for arterial and end-tidal PCO2), but the gradients between gastric and arterial or end-tidal PCO2 were maintained or even decreased (Table 4). These results are similar to findings of Ronholm et al.,14 who also measured PCO2 from the portal vein and found no significant differences between the portal and systemic arterial PCO2. They concluded that gastric-intestinal perfusion was in the range of aerobic metabolism. In our study, the lactate and pH value profiles were characterized by an acidic state throughout the procedure in both groups. However, lactate and pH values normalized sooner in the initial portal revascularization group, mainly at the second unclamping. At the end of the procedure both groups were comparable in all parameters observed, so that these differences accounted for the intraoperative period of graft reperfusion.
At present, the cava vein preservation technique remains part of the standard procedure for orthotopic liver transplantation: it maintains venous return to the heart during the anhepatic stage and makes veno-venous bypass unnecessary. However, this technique does not avoid splanchnic congestion and portal hypertension secondary to vena cava clamping. Adding a temporary portocaval shunt, as we did in our patients, to the piggyback technique, minimizes hemodynamic and metabolic instability during the anhepatic stage, and is likely to lead to more favorable values of gastric mucosal pH and gastric CO2. Therefore, changes found in our study should be attributed to the different vascular unclamping order in the 2 groups and not to changes occurring during the anhepatic stage. Initial reports considered that the severity of PRS was closely linked to the surgical technique, mainly when the inferior vena cava was resected.15 However, most recent studies have shown that the preservation of caval flow, even the addition of a portocaval shunt, did not avoid systemic arterial hypotension during graft reperfusion.6, 7 The addition of a portocaval shunt has the advantage of maintaining portal pressure while clamping the portal vein, which makes it feasible to perform arterial anastomoses before unclamping. Therefore, the graft may be perfused by either arterial or portal flow or even simultaneously.
It has been suggested that simultaneous arterial and portal reperfusion of the graft may reduce biliary complications due to better vascular supply to the bile ducts,16 although this view was not validated in a recent study.17 One study favoring arterial revascularization first18 reported lower transfusion. We did not confirm this, as in our study no differences in blood product usage were found. However, in Polak et al.'s study17 simultaneous arterial-portal revascularization had a higher packed red blood cells transfusion rate than sequential portal first revascularization, which may be related to increased fibrinolysis because of the longer anhepatic stage.
Differences in portal venous and hepatic arterial circulation are substantially important, as only a small proportion of hepatic arterial blood drains into the hepatic sinusoids. Although the relative contributions of the 2 systems at reperfusion are unknown, the order of unclamping might influence postoperative graft function. It has been suggested that arterial revascularization of the graft produces a more visually uniformly-colored liver than portal revascularization.8 However, portal venous blood is well oxygenated, offering adequate hepatic oxygenation soon after reperfusion.19
Overall, no studies have demonstrated differences in postoperative outcome that depend on the order of vascular release. We agree with Sadler et al.20 that in terms of outcome, 300 patients would be required to test the hypothesis that both unclampings are similar.
We conclude that initial arterial revascularization of the graft increases pulmonary pressure less markedly, so it may be indicated in those patients with poor pulmonary and cardiac reserve. Nevertheless, for all other patients, initial portal revascularization offers more favorable hemodynamic and metabolic behavior, less inotropic drug use, and earlier normalization of lactate and pH values. However, in terms of outcome there are no differences between initial arterial and portal revascularization.