Surgical Shunt Versus TIPS for Treatment of Variceal Hemorrhage in the Current Era of Liver and Multivisceral Transplantation Guilherme Costa, MD , Ruy J. Cruz Jr, MD, PhD , Kareem M. Abu-Elmagd, MD, PhD * KEYWORDS Portal hypertension Portomesenteric venous thrombosis Surgical shunt Transjugular intrahepatic portosystemic shunt Multivisceral transplant Hypercoagulable status Variceal hemorrhage occurs as a result of portal hypertension due to different under- lying diseases such as portal vein thrombosis (prehepatic), liver cirrhosis (hepatic), and Budd-Chiari syndrome (posthepatic). 1,2 The interplay between the portal hemody- namic changes and hepatic reserve ( Table 1 ) plays a major role in the short and long-term management of these complex patients. Of the important observed hemo- dynamic changes, is the increase in the intrahepatic vascular resistance compounded by the vasoconstrictor effect of a deficient state of intrahepatic nitric oxide. Of the major sequelae are the development of portosystemic collaterals, gastroesophageal varices, and a chronic state of systemic hyperdynamic syndrome. 3–7 The natural history of the underlying liver disease is another major factor that should be considered for the establishment of the management algorithm of these patients. Ethanol abuse, chronic viral hepatitis, and cholestatic/autoimmune disorders are the most common liver diseases that significantly influence the long-term outcome with the currently available different therapeutic modalities that are discussed herein. Another important factor that further complicates the management plan of these complex patients is the presence of extensive splenic and portomesenteric venous Intestinal Rehabilitation and Transplantation Center, Thomas East Starzl Transplantation Insti- tute, Department of Surgery, University of Pittsburgh Medical Center, UPMC Montefiore – 7 South, 3459 Fifth Avenue, Pittsburgh, PA 15213-2582, USA * Corresponding author. E-mail address: [email protected] Surg Clin N Am 90 (2010) 891–905 doi:10.1016/j.suc.2010.04.015 surgical.theclinics.com 0039-6109/10/$ – see front matter ª 2010 Elsevier Inc. All rights reserved. 892 Costa et al Table 1 Child-Pugh classification 1 2 3 Albumin (g/dL) >3.5 2.8–3.5 <2.8 Bilirubin (mg/dL) <2 2–3 >3 Prothrombin time a 1–3 4–6 >6 Encephalopathy None 1–2 3–4 Ascites None Slight Moderate Child-Pugh A 5 5–6 points: excellent hepatic reserve; Child-Pugh B 5 7–9 points: good hepatic reserve; Child-Pugh C 5 10–15 points: poor hepatic reserve. a Seconds prolonged. thrombosis that develops in cirrhotics and in hypercoagulable patients with normal hepatic parenchyma. Recent advances in pharmacologic and endoscopic management of gastroesoph- ageal variceal bleeding has significantly shifted the main paradigm of therapy from surgical to medical management. 8–14 However, medical failure does occur with an incidence of 20%, which requires radiologic or surgical intervention to decompress the portal system as a palliative but life-saving procedure. Nonetheless, organ replacement with liver alone or a multivisceral graft is the ultimate and definitive treat- ment for these patients. 13–15 In this article, a management algorithm is outlined to guide and optimize therapy for acute and recurrent variceal bleeding in these unique patients. EVALUATION GUIDELINES The standard evaluation process of patients with variceal bleeding has been compre- hensively addressed in the literature. 9–13 However, special emphasis should be placed on the history of the gastrointestinal hemorrhage including the number and severity of episodes. Documentation of the source of bleeding and the location of varices must be established to exclude other sources and guide therapy. The presence of gastric varices in the absence of Food and Drug Administration approval to use cyanoacrylate in the United States limits the therapeutic use of endoscopic therapy and calls for radiologic or surgical intervention. 11 The role of new endoscopic techniques including the recently introduced endoscopic capsule has increased the accuracy of variceal detection, particularly in patients with enteric and ectopic varices. The diagnosis of portal hypertensive gastropathy and/or colopathy, in the absence of gastrointestinal varices, may also shed some light on the source of bleeding and guide further therapy. The status of portal hypertension can be semiquantitatively assessed by the degree of the pancytopenia, splenomegaly, and radiologic evidence of intra-abdominal visceral collaterals in addition to the endoscopic documentation of gastrointestinal varices, gastropathy, and/or colopathy. Noninvasive radiologic studies could also be helpful as a screening test for the possible coexistence of splenic or portomesen- teric venous thrombosis. In persons with radiologic evidence of partial or complete visceral venous thrombosis, a hypercoagulable syndrome must be suspected and thoroughly evaluated. The evaluation process includes measurement of protein C, protein S, antithrombin III, and total homocysteine serum levels. In addition, genetic studies for factor V Leiden, prothrombin G20210A, and JAK-2 gene mutations should be conducted. Equally important are the diagnosis of paroxysmal nocturnal Surgical Shunt Versus Tips 893 hemoglobinuria and the detection of anticardiolipin, lupus anticoagulant, and anti- phospholipid antibodies. When portomesenteric venous thrombosis is suspected or the pattern of gastro- esophageal varices is unusual, particularly in the presence of normal liver paren- chyma, patients must undergo visceral angiographic studies to evaluate the extent of thrombosis and to map the collateral circulation, including the assessment of the extent of collateralization and direction of flow ( Fig. 1 ). The angiographic studies include superior mesenteric and splenic arterial injections with venous phases. On multiple occasions in their practice, the authors have observed unexpected angio- graphic findings such as arterial varices, arteriovenous malformations, and intrahe- patic arterioportal communications. The old technique of measuring the portal venous pressure through a transsplenic, transhepatic, or transvariceal approach has been abandoned. In selected patients, catheterization of the hepatic veins with measurement of the free and wedged hepatic venous pressure is used, particularly in those with gastroesophageal varices in the absence of significant liver disease or portomesenteric venous thrombosis. Finally, the etiology and extent of liver disease can be easily assessed in a system- atic way, including thorough medical history, biochemical and hematological studies, abdominal imaging and, if necessary, percutaneous or transjugular liver biopsy. The histopathologic examination is of utmost importance, particularly in Child A/B patients who could be managed by radiologic or surgical intervention. VARICEAL DECOMPRESSION WITHOUT ORGAN REPLACEMENT Surgical Shunts With a long-standing history of more than a half of a century, surgical shunts have played a major rule in the management of variceal bleeding by total, partial, selective, or super-selective decompression of the portal, mesenteric, splenic, and gastro- esophageal variceal venous system; respectively. Since its peak popularity during the 1960s through the 1980s, surgical shunts have been gradually used with less Fig. 1. Superior mesenteric arteriogram with delayed venous phase. Note complete occlu- sion of the superior mesenteric and portal veins. The patient was hypercoagulable and underwent a multivisceral transplantation. 894 Costa et al frequency because of the introduction or revisiting of new nonsurgical therapeutic modalities and the evolution of abdominal organ transplantation. Total portal systemic shunts are shunts that totally decompress the portal system with surgical anastomosis between the portal vein or one of its major branches, and the vena cava or one of its major tributaries. These shunts have usually been made as a side-to-side shunt with 10 or more millimeters of diameter, with decompression of the whole splanchnic portal hypertensive bed and the high hepatic sinusoidal pres- sure. Such shunts have a high patency rate, and excellent control of variceal bleeding as well as ascites. 16,17 However, most of the published series document a high inci- dence of encephalopathy and variable long-term survival outcome, which could be attributed to the differences in the severity and nature of the underlying liver disease among the selected study population. 18 Because of the potentially prohibitive risk of encephalopathy with total shunts, Sar- feh and colleagues 19 popularized partial portal systemic shunts in the 1980s and 1990s by reducing the diameter of the shunt to 8 mm. The rationale is to maintain some portal flow to the hepatocyte without compromising the therapeutic decompres- sion of the varices. 18–20 Technically, the shunt can be easily performed by placing an 8-mm polytetrafluoroethylene-reinforced graft between the portal vein and inferior vena cava ( Fig. 2 ). Data that were published in the 1990s documented excellent control of bleeding with a low risk of encephalopathy and acceptable rate of long- term survival. The shunt has also been compared with the transjugular intrahepatic portosystemic shunt (TIPS) procedure in a randomized trial, as discussed later. Selective decompression of the gastroesophageal varices was achievable by the clinical introduction of the distal splenorenal shunt (DSRS) by Warren and colleagues. 21–23 The shunt