Temporary portocaval shunt in orthotopic liver transplantation: Need for a standardized approach?

The classical technique of orthotopic liver transplantation (LT) involved resection of the retrohepatic vena cava and cross-clamping of the porta hepatis, which led to complete interruption of the venous return to the heart from the inferior vena cava (IVC) and portal vein.1 In an attempt to maintain hemodynamic stability, reduce blood loss, and preserve renal function, venovenous bypass was introduced to bypass the splanchnic and caval blood into tributaries of the superior vena cava during the anhepatic phase. This technique enabled many surgeons to perform LT with good results, providing hemodynamically stable patients who were able to tolerate a more prolonged anhepatic phase. However, not all surgeons used venovenous bypass. The clamp-clamp technique continued to be successfully used by many skilful surgeons and anesthetists paying close attention to optimal fluid and pressure management. In 1989, Tzakis et al.2 described the piggyback technique, in which the donor suprahepatic IVC was anastomosed to the confluence of the recipient's hepatic veins with ligation of the donor infrahepatic IVC. Subsequent modification by Belghiti et al.,3 in which the suprahepatic IVC was anastomosed to the recipient's middle and left hepatic veins, avoided the need to include the right hepatic vein, which often caused caval obstruction, thus ensuring excellent caval flow during the anhepatic phase. Further modifications have led to a wider anastomosis by extension of the left and middle hepatic veins (with closure of the right hepatic vein) into the IVC.

Clamping of the portal vein and the subsequent rise in portal venous pressure, congestion of the splanchnic bed, and intestinal edema led to the description by Tzakis et al.4 of the use of a temporary portocaval shunt (TPCS), in which an end-to-side anastomosis is formed between the recipient's portal vein and infrahepatic IVC. The use of a TPCS has now become common practice. Anecdotally, it has been shown to improve retrohepatic dissection and lessen splanchnic congestion, but the evidence is scarce. The aim of this study was to systematically review the literature on the use of TPCS in LT and determine whether there is evidence to support its routine use.

Abbreviations

ALF, acute liver failure; AP, anhepatic phase; CLD, chronic liver disease; FFP, fresh frozen plasma; IVC, inferior vena cava; LT, liver transplantation; MAP I, mean arterial pressure prior to clamping; MAP PVC, mean arterial pressure at portal vein clamping; MAP PVD, mean arterial pressure at portal vein declamping; NA, not available; NAH, no statistically significant difference according to the author; OLT, orthotopic liver transplantation; Plat, platelet units; RBC, red blood cell; TPCS, temporary portocaval shunt.

PATIENTS AND METHODS

All published reports describing the use of a TPCS in adult LT were identified through an electronic search of the Medline and Cochrane library databases from 1963 to 2007. The search terms liver transplantation and temporary portocaval shunt or portocaval bypass were used as both subject headings (MeSH and Emtree) and truncated words. Articles were restricted to publications in English and Spanish. The initial search identified 4386 articles, of which 73 were potentially relevant. Further references were identified by manual cross-checking of the reference lists.

The criteria used to select studies for inclusion in the final analysis were performance of a TPCS during LT in human adult patients. Articles describing pediatric patients (<16 years of age), multivisceral grafts, retransplants, or auxiliary transplants were excluded. When repeated reports from the same institution were identified, only the most recent or most relevant study was included.

RESULTS

The searches in the Medline and Cochrane databases identified 4386 and 3 articles, respectively. After the exclusion of duplicates, 4386 titles were obtained; 19 articles met all the criteria, but only in 11 of those articles was the experience with patients reported. Three articles were excluded as they were written by the same group of authors5-8 in the same hospital during a similar time period, and although it was not specified, it was assumed that they described the results from the same cohort of patients who were more fully reported in other publications that were included in our analysis. Results from the 8 included articles are summarized in Tables 1 and 2, including a total of 2073 patients.1, 3, 8-13

The indication for transplantation was given for 647 recipients. The majority of patients had chronic liver disease (95.51%), with the remainder (4.49%) being transplanted for acute liver failure. The time course of the liver disease was not specified in 1426 patients.12, 13 Operative data were available for 712 patients and are summarized in Table 1. The duration of the operation in patients undergoing TPCS ranged from 330 ± 63 to 465 ± 86 minutes. A direct comparison of the operative times with and without TPCS was available in 3 studies.9, 10, 14 There was no difference in 2 studies,9, 14 and 1 study found that the operative time was shorter in patients undergoing TPCS.10

The intraoperative blood product transfusion requirement was described in 6 articles.1, 3, 8, 10, 13, 14 The number of red blood cell units transfused into patients with a TPCS ranged from 1.1 ± 2.3 to 98.4 ±4.3 units. A comparison of patients with a TPCS and patients without one was reported in 3 articles.10, 13, 14 Two of 3 studies found that a TPCS was associated with less red blood cell requirement.13, 14 The amount of fresh frozen plasma transfused into patients with a TPCS ranged from 1.4 ±2.7 to 19 ±14.7 units. Two of 3 studies found that TPCS was associated with less fresh frozen plasma transfusion.13, 14 Platelet transfusion was described in 3 studies and ranged from 5 to 17 units.8, 10, 14 One of 2 studies found that TPCS was associated with less platelet transfusion.10

Intraoperative hemodynamic variables were reported in 5 articles.1, 3, 8, 13, 14 The mean arterial pressure during portal vein clamping in patients with a TPCS ranged from 77.6 to 128 mm Hg. Two studies compared the mean arterial pressure during portal vein clamping in patients with and without a TPCS.13, 14 Both showed that the use of TPCS was associated with a higher mean arterial pressure at the time of portal vein clamping.

Postoperative renal function was reported in 3 studies, and all showed that TPCS did not appear to have an adverse effect on postoperative serum creatinine9, 13, 14

Perioperative morbidity and mortality was described in 3 studies and ranged from 2.5% to 18%1, 14 and from 10% to 12%,1, 3 respectively. No comparison was made between patients with and without a TPCS.

DISCUSSION

The hemodynamic and immunological consequences of portal vein clamping are poorly characterized. Portal vein clamping induces a reflexive increase in systemic vascular resistance that helps to maintain mean arterial pressure despite a fall in the cardiac index as a result of the interruption of portal flow. However, with re-establishment of portal flow, there is a reflex fall in blood pressure as well as the elaboration of vasoactive metabolites from the splanchnic circulation.15 This is usually short-lived and well tolerated, but with prolonged clamping and in the presence of hemodynamic instability, the re-establishment of portal flow causes problems. In animal models, an interruption of portal flow for up to 90 minutes results in increased permeability of the splanchnic vessels, interstitial edema of the gut, and the accumulation of acute inflammatory cells with evidence of mucosal cell damage.16, 17 With the release of portal clamping, inflammatory cytokines are released into the liver and subsequently the systemic circulation, leading to vasodilatation and end organ damage, including subpericardial hemorrhage, myocardial necrosis, and pulmonary infiltration and edema.16, 17 It has been postulated that a TPCS prevents splanchnic congestion and helps to maintain portal venous return, thereby abrogating the regional and systemic consequences associated with prolonged portal vein clamping.

The use of a TPCS was initially advocated for patients with fulminant or subacute liver failure with portal hypertension in the absence of portosystemic venous collaterals.4 Clamping of the portal vein during the anhepatic phase when portosystemic collaterals have not developed may lead to hemodynamic instability, cerebral edema, and splanchnic venous congestion with associated hemorrhage and intestinal edema. Only 1 article has reported the effects of TPCS in patients undergoing LT for fulminant or subacute liver failure.18 Belghiti et al.18 reported that the maintenance of portal and caval flow during hepatectomy and graft implantation avoids the fall in venous return and other hemodynamic changes resulting from the venous clamping, release of metabolites, and secondary effects of fluid management seen in patients undergoing LT for fulminant and subacute liver failure. In patients with cirrhosis, it is thought that the use of a TPCS is unnecessary, as the presence of spontaneous portosystemic venous collaterals will result in little hemodynamic change during the period of portal vein clamping. This, however, remains unproven. Muscari et al.19 published their experience with 84 patients who underwent piggyback LT without a TPCS and concluded that the routine use of a TPCS in patients with cirrhosis undergoing the piggyback technique was not justified. Nevertheless, they did not include a group for comparison who had a TPCS. A number of studies have demonstrated an improvement in hemodynamic parameters and blood product requirement with the use of a TPCS in patients with cirrhosis.10, 14 It is difficult to compare the findings from each of these studies, as there was no standardized method of patient selection or monitoring. What is apparent is that the routine use of a TPCS in patients with cirrhosis is not deleterious.

In a prospective randomized trial of 80 patients with cirrhosis, Figueras et al.14 demonstrated that only 55% of patients with a TPCS required a blood transfusion versus 78% of patients without a TPCS (P = 0.03). The clinical benefits of a TPCS appear to be greatest in patients with severe portal hypertension and high portal flow. In patients with a portal flow of greater than 1000 mL/minute and a portocaval gradient of greater than 16 mm Hg, only 33% of patients with a TPCS required a blood transfusion versus 85% of patients without a TPCS (P = 0.04).14 During the anhepatic phase, the decrease in cardiac output was significantly less with the use of a TPCS only in those patients with a high (>1000 mL/minute) portal blood flow.14 Similarly, Margarit et al.9 reported a beneficial effect of a TPCS in patients with cirrhosis when the portal flow was greater than 800 mL/minute.

The use of a TPCS does not prolong the duration of the operation. There were 3 articles in which a direct comparison of the operative time was possible: TPCS was associated with a shortening of the operative time in 1 article,10 whereas in the other 2 articles, the difference was not significant.9, 14 The lowering of portal pressure leads to less bleeding during dissection and hepatectomy, translating an often difficult step into a more straightforward procedure that is less time-consuming and thus compensating for the time taken to complete the TPCS. In addition, the improvement in operative exposure gives trainees the opportunity to perform the retrohepatic dissection under controlled circumstances.

Renal impairment is a common sequel to LT. Impaired renal perfusion, vascular instability, and the release of cytokines at the time of reperfusion contribute to a reduction in renal function.18 Preservation of caval flow with the use of the piggyback technique is associated with a lower incidence of postoperative acute renal failure.2 A number of studies have demonstrated an improvement in renal function with the use of a TPCS.9, 14 In a prospective randomized trial comparing patients with a TPCS and patients without one, Figueras et al.14 demonstrated that the urine output during transplantation was greater in all patients with a TPCS but reached statistical significance only in those patients with both high portal flow (>1000 mL/minute) and high portocaval gradient (>16 mm Hg). In addition, postoperative serum creatinine levels were stable in patients with a TPCS, whereas the serum creatinine increased over the first 48 hours post-transplant in those without one.14 Three patients, 2 of whom did not have a TPCS, required renal replacement therapy postoperatively. Similarly, Margarit et al.9 reported a significantly lower postoperative serum urea in patients with a high (>800 mL/minute) portal flow. The beneficial effect of a TPCS on renal function most likely reflects an improvement in hemodynamic stability, as the piggyback technique allows for preservation of renal perfusion.

From a technical point of view, the use of a TPCS minimizes the effects of prolonged portal vein clamping, reducing the urgency to implant the liver. This provides the opportunity to undertake the more time-consuming arterial anastomosis prior to re-establishing portal flow and undertaking initial arterial reperfusion. Although controversial, arterial reperfusion has been demonstrated in a number of clinical trials to have both hemodynamic and functional advantages over primary portal reperfusion.20 To date, the potential benefit of the combined use of TPCS and arterial reperfusion has not been compared to portal vein reperfusion, and a prospective randomized trial is needed to determine whether it is associated with an improved outcome, particularly in the use of marginal grafts.20

A prospective randomized clinical trial of patients undergoing piggyback LT with and without a TPCS is required. Given the improvements in perioperative care, organ preservation, and surgical technique over the past 20 years, it is likely that well-compensated patients would have similar outcomes and those recipients with portal hypertension or receiving marginal grafts would have the greatest benefit from a TPCS. We propose undertaking a prospective randomized trial looking at intraoperative cardiovascular parameters, postoperative measures of reperfusion injury, perioperative transfusion requirements, and renal function in all patients undergoing LT. In order to achieve 80% power for detecting a 5% difference in the perioperative variables, we estimate the minimum sample size would need to be 200 patients. Until this study is undertaken, we will continue to use a TPCS on the basis of the aforementioned evidence demonstrating that it is associated with less blood product transfusion, improvement in intraoperative hemodynamics, and ease of retrohepatic dissection with a shorter operative time.