Page |1

PORTAL HYPERTENSION
MU Khan Moderator Seminar No 27 Commentators. Dr I Seedat & Dr MC Mayet Date: 09/10/2010 1 Definition Portal hypertension is haemodynamically defined by a pathological increase of the portal pressure gradient (the pressure difference between the portal vein and the inferior vena cava) and by the formation of portalsystemic collaterals that shunt part of the portal blood flow to the systemic circulation bypassing the liver. Normal values of the portal pressure gradient are of 15 mm Hg. Clinically significant portal hypertension (CSPH) is diagnosed when clinical manifestations of the disease appear or when portal pressure gradient exceeds a threshold value of 10 mm Hg. Values of portal pressure gradient between 5 and 9 mm Hg correspond to pre-clinical portal hypertension. Anatomy 2,3,4 Embyology The portal venous system is formed from the two vitelline and two umbilical veins. The vitelline veins which drain blood from the yolk sac intercommunicate in the septum transversum, which is the site of development of the liver sinusoids. The extrahepatic portal system develops primarily from the left vitelline vein, while the intra-hepatic portal circulation is formed from the umbilical veins. The left umbilical vein also directly communicates to the sinus venosus which connects to the inferior vena cava, and allows most blood to bypass the liver in the fetal circulation. The portal vein The portal vein is formed behind the neck of the pancreas where the superior mesenteric joins at the splenic vein. The portal vein is approximately 6 to 8 cm long, and 1 to 1.2 cm in diameter, and it runs in the free edge of the lesser omentum from the pancreas to the right end of the porta hepatis. The entry of the tributaries to the portal vein are variable, and comprise primarily the left and right gastric veins, the gastroepiploic veins, the middle colic vein, and the inferior mesenteric vein. In portal hypertension the veins of importance are the left gastric (coronary vein), and the inferior mesenteric vein. These vary as to whether they enter the splenic vein, splenic/ portal junction, or the portal or superior mesenteric vein respectively. The portosystemic circulation

Page |2

The portosystemic collateral circulation that forms in portal hypertension can be described by the pattern of drainage (ie either into the superior vena cava or the inferior vena cava). Superior vena cava Left gastric (coronary) vein is the most commonly visualised systemic communication in portal hypertension. It drains the anterior and posterior of the stomach and ascends to the oesophageal hiatus to drain the lower oesophagus, then descends to the right to enter the portal vein superior to the duodenum. The left gastric vein traverses the lower oesophagus to drain into and azygous and hemiazygous veins. They may drain directly into the subclavian/ brachiocephalic system through the left pericardiophrenic vein or through the inferior phrenic vein. Some may also communicate with a pulmonary vein or with the spinal venous plexus. The clinical implication of these shunts is the potential for injected material to travel through these systems. Endoscopically detected oesophageal varies generally refer to dilated veins found in the wall of the oesophagus. These arise from the anterior branch of the left gastric vein. The posterior branch of the left gastric vein supplies the paraoesophageal veins, which are located outside the wall of the oesophagus. A third category of oesophageal veins, the perioesophageal veins, are thought to be located in the adventitia of the oesophagus. These communicate freely with the intraoesophogeal veins via perforating vein. Previously paraoesophageal and perioesophageal veins were described as distinct entities. These have now been commonly described as paraoesphageal veins. All three levels of veins appear to communicate through perforating veins which have important therapeutic implications in the treatment of this pathology. A left gastric vein diameter greater than 7 mm in diameter is associated with a portohepatic gradient greater than 10 mmHg. Short gastric veins course long the lateral border of the greater curvature and communicate with the splenic vein or its major tributaries at the splenic hilum. These drain the left half of the greater curvature and the gastric fundus. In portal hypertension these usually from a complex of tortuous veins at the splenic hilum and gastric fundus . Individual veins are delineated with difficulty. Posterior gastric vein (accessory left gastric vein) drains the area of the stomach between the left gastric and the short gastric veins. They are inconsistently visualised. These veins are more likely than the left gastric vein, to drain into the renal venous system. Venous anatomy of the gastrooesophageal junction can be divided in to four different zones, which are clinically important as this is the site where areas of portosystemic communication most commonly bleed. The gastric zone is 2 to 3 cm below the gastroesophageal junction, where the veins meet at the upper end of the cardia of the stomach, drain into short gastric and left gastric veins, and then drain into the splenic and portal veins, respectively. The palisade zone is 2 to 3 cm proximal to the gastric zone into the lower esophagus, where the veins communicate with extrinsic (periesophageal) veins in

Page |3

the distal esophagus. This zone forms the dominant watershed area between the portal and the systemic circulations. The perforating zone, is proximal to the palisade zone where a network of submucosal veins in the esophagus connects to the periesophageal veins, which drain into the azygous system and subsequently into the systemic circulation. The truncal zone is approximately 10 cm in length and is located proximally to the perforating zone in the esophagus. It typically has four longitudinal veins in the lamina propria.5 Inferior vena cava (IVC) Left renal vein. Communication between the splenoportal vein axis and the left renal vein may occur through the left gastic, the short gastric and the posterior gastric veins (gastrorenal shunt) or the splenic vein (splenorenal shunt). Communication between the left gastric and the left renal veins can be formed by the left inferior phrenic vein or the left adrenal vein. Paraumbilical and abdominal wall veins. The paraumbilical vein arises from the left portal vein. Portosystemic collaterals develop between recanalised veins of the ligamentum teres and falciform ligament of the portal system, into the IVC from inferior epigastric veins via external iliac veins . These vessels may also drain into the SVC through the superior epigastric and internal thoracic veins, resulting in the characteristic caput medusa. Cruveilhier-Baumgarten syndrome refers to a clinically detectable venous hum and caput medusae as a result of a significant postosystemic shut in patients with cirrhosis and portal hypertension. Other collaterals Left and right gastroepiploic veins are generally considered involved if the diameter is greater than 6mm. Perisplenic collaterals between the retroperitoneal veins branches of splenorenal shunts. Intrahepatic portal veins may form connections with branches of the hepatic vein or the left gastric vein, usually in the left lobe of the liver. The hepatic surface may be covered by a loose collateral system that often forms connections with pleural, pericardial and pulmonary veins referred to as pleuropericardial-peritoneal collaterals. The accessory portal system of Sappey, originally included vessels of the round ligament, nutrient vessels of the portal tract and veins of the suspensory ligament of the liver. The term vein of Sappey is regarded as a synonym for para-umbilical vein, but has historically been used to discribe diaphragmatic collaterals. Collateral pathways between the retroperitoneal vessels of the superior and inferior mesenteric veins and the systemic circulation are common in portal hypertension. The veins of Retzius refers to communications between these systems and the inferior vena cava. Classification and causes of portal hypertension
1

Page |4

Portal hypertension can be classified anatomically as prehepatic (involving the splenic,mesenteric or portal vein), intrahepatic (liver diseases) and posthepatic (diseases blocking the hepatic venous outflow). Approximately 90% of portal hypertension in westernised countries is due to cirrhosis, whereas the chief cause in developing countries is schistosomiasis. The remaining causes account for 10%, and are usually collectively referred to as non-cirrhotic portal hypertension. Prehepatic causes PORTAL VEIN THROMBOSIS Portal vein thrombosis is a common cause of prehepatic portal hypertension. It is usually as a result of omphalitis in children and as a result of thrombophillic diseases in adults. Additionally it may be acquired (myeloproliferative disease) or secondary to local factors (sepsis, abdominal trauma or surgery). Idiopathic portal vein thrombosis accounts for the remaining 30% of causes. Acute portal vein thrombosis is uncommon, and is usually asymptomatic. Clinical manifestations include abdominal pain and fever, variceal bleeding, hypersplenism, and growth retardation. Less frequent manifestations include venous infarction of intestines, hemobilia and ascites. Chronic PVT is characterized by the formation of collateral vessels, causing the appearance of the so-called portal cavernoma. These hepatopetal collaterals may provide sufficient portal flow to maintain normal liver function. In addition, hepatofugal esophagogastric and splenorenal collaterals develop aiming to decompress portal pressure. Patients with chronic PVT are frequently diagnosed after a first episode of variceal bleeding. Gastric, anorectal varices and ectopic varices and hypertensive gastropathy or colonopathy are more common than in cirrhotics. Varices in portal vein thrombosis rarely bleed, but when bleeding does occur, it can be life threatening and difficult to control endoscopically, especially in ectopic locations. Due to preserved liver function the survival of patients with variceal bleeding from noncirrhotic and nontumorous portal vein thrombosis is better than in cirrhosis. The mortality from variceal bleeding ranges between 1.5% and 25%. 6 Treatment options include anticoagulation in the acute phase, and medical and endoscopic therapy for portal hypertension complications in the chronic phase. Surgery consists of bypass and shunting procedures 6 Shunts require favourable extrahepatic portal pattern. A conduit between the IVC and a branch of the portal vain is usually used, in a end to side, a side to side or using an interposition shunt using an autologous venous graft usually the internal jugular vein. The patency of these shunts is excellent and hepatic encephalopathy rates are low. In those with unfavourable extrahepatic anatomy, a bypass procedure is considered. The mesenterico-portal bypass has been popularised in the last decade. First used in paediatric patients with portal vein thrombosis post liver transplant. A conduit between the left intrahepatic portal vein and the superior mesenteric vein, the splenic vein or a large pancreaticoduodenal vein is created. This allows restoration of hepatic blood flow and results in decompression of the portal system. Success requires large intra- and extrahepatic portal system.

Page |5

Problematic shunt occlusion is treated with radiological intervention or repeat surgery.

Box 1.Classification PH according to the anatomic site of increased resistance Prehepatic Splenic vein thrombosis Portal vein thrombosis Congenital stenosis of the portal vein Extrinsic compression of the portal vein Arteriovenous fistulae Intrahepatic Cirrhosis (viral, alcoholic, biliary, metabolic) Granulomatous diseases (schistosomiasis, sarcoidosis, tuberculosis, PBC) Partial nodular transformation* Nodular regenerative hyperplasia* Congenital hepatic fibrosis Peliosis hepatis Polycystic disease* Idiopathic portal hypertension* Hypervitaminosis A Arsenic, copper sulfate, vinyl chloride monomer poisoning Amyloidosis Mastocytosis Rendu-Osler-Weber syndrome Liver infiltration in hematologic diseases Acute fatty liver of pregnancy Severe acute viral and alcoholic hepatitis Chronic active hepatitis Hepatocellular carcinoma Cyanamide toxicity Veno-occlusive disease Posthepatic Hepatic vein thrombosis (Budd-Chiari syndrome) Congenital malformations & thrombosis IVC

Figure . 1 Three types of bypass procedures. (a) Mesenteric to left branch of portal vein. (b) Splenic to left branch of portal vein. (c) Pancreatico duodenal to-left branch of portal vein. The dotted line figures the obstructed main portal vein.7

Post hepatic causes BUDD-CHIARI SYNDROME Hepatic vein thrombosis, also referred to as Budd-Chiari syndrome, is one of the more common causes of post hepatic portal hypetension. Thrombosis can occur at the main hepatic veins or in suprahepatic inferior vena cava. A prothrombotic disorder is usually the cause, with myeloproliferative disorders being the most common. Major complications are ascites and gastrointestinal bleeding associated with a variable degree of liver failure. The disease can present as an acute, subacute or chronic disease. Diagnosis is usually made by imaging techniques. Treatment includes anticoagulation to prevent recurrence or extension of thrombosis, treatment of ascites and gastrointestinal bleeding, and procedures aiming at re-establishing hepatic blood outflow. Transjugular intrahepatic portosystemic shunt (TIPSS) has substituted shunt surgery in patients who do not improve with medical treatment. Sever liver failure will necessitate transplantation.

Intrahepatic causes Distinguishing the level of sinusoidal disease is possible using HPVG and includes :

Page |6

(a) pre-sinusoidal PH: normal wedged and free hepatic venous pressure (WHVP and FHVP); (b) sinusoidal PH: increased WHVP and normal FHVP; (c) post-sinusoidal PH: increased WHVP and FHVP. Some disorders may act at several sites, for example, in Schistosomiasis portal hypertension is the consequence of the formation of granulomas due to the deposition of parasite eggs in portal venules. The inflammatory response induces fibrosis and obliteration of portal venules (pre-sinusoidal PH), which later extends to sinusoids (sinusoidal PH). At this stage the clinical and haemodynamic alterations resemble liver cirrhosis. The acceptable options for treating hepatic schistosomiasis in patients with portal hypertension are an extensive esophagogastric devascularisation operation with splenectomy, or a distal splenorenal shunt. Any cause of chronic liver disease, except chronic cholestatic syndromes cause sinusoidal PH. Liver cirrhosis from a variety of causes is responsible for the vast majority of cases of portal hypertension, chiefly arising from viral hepatitis and alcohol related liver disease. As such, much of the evidence for the treatment of PH is based on studies in this patient population and is not necessarily applicable to PH due to other diseases. Pathophysiology 8,13 Determinants of portal pressure Portal pressure is the result of the relationship between the blood flow volume entering the portal system and the resistance to portal blood flow. The mathematical expression of this relationship is given by the Ohms formula: P = Q x R, where P represents change in pressure along the vessel, Q represents blood flow and R resistance to the flow. In normal liver, resistance is mainly located in terminal portal venules, in the sinusoids and in the roots of hepatic venules. The increase in resistance to portal flow is the main determinant of portal hypertension in cirrhosis (backward-flow theory). It is caused by structural alterations of hepatic microcirculation secondary to fibrosis and nodular regeneration, but also by an active vasoconstriction at the level of the intrahepatic vascular system. The increase in resistance to portal flow represents a stimulus for an increase in splanchnic inflow, which in turn maintains and worsens portal hypertension (forward-flow theory). 8 In cirrhosis, the primary factor leading to portal hypertension is an increased resistance to portal blood flow. Later on, an increase in portal venous inflow will help to maintain and aggravate portal hypertension. 9

Page |7

Figure 2. Pathophysiology of cirrhotic PH

10

Increased vascular resistance to portal blood flow (backward flow theory) This is due to fixed anatomical obstruction and a modifiable reversible component which may account for up to 40% of the increased intrahepatic resistance in cirrhosis. 9 Structural alterations The progressive deposition of collagen in the spaces of Disse narrows the sinusoidal lumen, and increases the distance between sinusoidal lumen and hepatocytes, causing an increase in resistance to portal flow. The further transformation of collagen into fibrotic tissue, together with regeneration of hepatocytes, leads to the derangement in hepatic vascular structure. The compression of centrolobular venules by regenerating nodules, granulomas, or portal inflammation are major further mechanisms responsible for the increase in hepatic resistance. 8 Dynamic/ functional component10 Hepatic stellate cells (HSC), found in the spaces if Disse, are thought to modulate hepatic blood flow and vascular resistance by contraction and relaxation. These HSCs are activated by a variety of neurohormonal signals, responding gradually to change morphology from the quiesant form to a myofibrocyte type, the so called activated hepatic stellate cell. These cells cause extracellular matrix deposition, release of growth factors and increased cell contractility. They are thus considered to be major contributors to connective tissue deposition and thus have important implications in the progression of portal hypertension. Proposed mediators of their action include endothelin-1 (ET-1) and nitric oxide. 11,12

Page |8

Vasocontrictor contribution to increased vascular tone. An imbalance in the levels of vasoactive mediators is seen in portal hypertension, promoting vasocontriction. A few of these with therapeutic implications will be discussed below. COX derived prostanoids The arachidonic acid breakdown pathway is responsible for the production of prostaglandins and thromboxanes, regulated by the enzyme, cyclooxygenase (COX). COX-1 derived prostanoids are vasoconstrictive and are involved in promoting increased intrahepatic resistance. The increased production of thromboxane A2(TXA2) by COX-1 is ameliorated by treatment with non-selective COX inhibitors, selective COX-1 inhibitors and TXA2 antagonists in cirrhotic livers. Nitric oxide (NO) NO is a potent vasodialator. In cirrhosis there is reduced NO bioavailability, which contributes to an increase in the intrahepatic vascular resistance. Exogenously administered nitrates have been shown to decrease portal pressures, but studies have not shown conclusive benifit, perhaps due to the lack of hepatic specificity, and the potential for generalized vasodilation. Initial studies with NCX-1000, a hepatic specific NO releasing derivative of urodeoxycolic acid, showed promise in reducing the hepatic portal venous gradient and attenuating the post prandial increases in portal pressure. However follow up studies have not confirmed this benefit. Statins, have been proposed to modulate intrahepatic venous resistance by increasing NO bioavailability. Their postulated mechanism of action is varied and includes increasing NO bioavailibility, decreasing HSC contraction, antifibrotic and anti-inflammatory effects. Initial reports are promising but further evaluation is warranted. Endothelin Endothelin-1 (ET-1) is a vasoconstrictor which has a stimulatory effect on the HSC and also contributes to the increases in intrahepatic vascular resistance. It may, however also contribute to decreasing hepatic fibrosis and accordingly the blockade of these receptors is not currently advocated. 8 Increase in splanchnic inflow (Forward flow theory) In the normal liver increases in flow do not modify the portal pressure. Obstruction to the portal outflow with an increase in portal inflow is responsible for an increase in portal in portal pressure. This is demonstratd by the therapeutic effect of non selective beta blockers, which decrease portal pressure by increasing splanchnic resistance, decreasing cardiac output, so decreasing portal inflow. Hyperdynamic circulatory syndrome is the term used to describe the decrease in systemic vascular resistance and an increased cardiac index. The increase in splanchnic inflow is the result of decreased splanchnic arterial resistance. The portosystemic collateral circulation in PH directly decreases systemic vascular resistance, and indirectly contributes to this clinical scenario by allowing vasoactive mediators to bypass the liver and enter the systemic circulation directly. The hyperdynamic circulatory syndrome is implicated in pathogenesis of the complications of cirrhosis (ie. Hepato-renal syndrome,

Page |9

hepato-pulmonary syndrome etc.) and for a number of symptoms of portal hypertension (ie tachycardia, warm skin, systemic arterial hypotension). An increased circulating blood volume is another finding in the hyperdynamic circulatory syndrome. This is a result of portal hypertension which causes splanchnic vasodilatation and consequently as a result of renal medullary hypoperfusion, hypervolaemia results due to the retention of sodium. This hypervolaemic state perpetuates the hyperdynamic syndrome by increasing the preload, which in turn increases cardiac contractility and thus the cardiac output. The hyperdynamic syndrome maintains and enhances portal hypertension. Longterm therapy with anti-aldosteronic drugs in cirrhotic patients without ascites decreases portal pressure by decreasing blood volume. Splanchnic vasodilatation 8 Due to the splanchnic vasodialation and increased blood flow through this system there is hypoperfusion in the region of the kidney. This results in vasoconstriction of the renal arteries by paracrine mediators such as prostaglandin (PGE) 2 in an attempt to preserve juxtamedullary blood flow. With progression of disease this increased splanchnic flow with relative hypoperfusion of other organs worsens. Renal blood flow is often improved by constriction of the splanhnic circulation by vasoconstrictors such as vasopressin, somatostatin and their analogues. The hyperdynamic syndrome and splanchnic vasodialation is caused by the alteration of neurohormonal factors involved in the regulation of vascular tone and sodium and water retension. Mediators that have been implicated in this scenario include glucagon, prostaglandinsprostacyclin, histamine, intestinal vasoactive peptide, substance P, colecystokinin, estrogens, ammonia, endotoxins, adenosine, biliary acids, NO, alpha-calcitonin gene-related peptide, adrenomedullin, vascular endothelial growth factor, and carbon monoxide. Nitric oxide also plays a key role in the vasodilation seen in these patients, thought to be due to increased levels of endothelial nitric oxide synthase. This may be further fuelled by bacterial translocation. Here the inducible form of NO (iNOS) is implicated as part of the sepsis syndrome. Endocannabinoids, which increase NO production, aggrevate the hyperdynamic syndrome. Autonomic dysfunction, leads to vascular hyporeactivity to the vasoconstrictor systems to vasoconstrictor systems such as the sympathetic nervous system, vasopressin, angiotensin-II and endothelin. This may be related to a change in receptor affinity, or to a downregulation of receptors. Angiogenesis Recently neoangiogenesis has been implicated in splanchnic hyperaemia and portosysemic collateralisation. Pathological angiogenesis is also thought to be responsible for the abnormal angioarchitecture of the cirrhotic liver and is intimately related to fibrogenesis thus also contributing to increased intrahepatic resistance to blood flow seen in cirrhosis and portal hypertension. Vascular endothelial growth factor (VEGF) and platelet-derived growth factor (PDGF), potent angiogenic factors, are integral in this process. Sorafenib, a tyrosine kinase inhibitor, has been used to block VEGF and PDGF activity in experimental models, showing benefits in decreasing portal pressures. It is currently used for renal and hepatocellular carcinoma. However due to serious side effects, optimal dosing and further evaluation is required. 10

P a g e | 10

PATHOPHYSIOLOGY OF VARICEAL HAEMORRHAGE 14 The factors regulating the tension exerted by the wall of varices to contain the expanding force can be described according to the Laplaces law : T=TP x r/w, in which T is the wall tension, TP the transmural pressure, r the radius and w the wall thickness of the varices. The transmural pressure is the difference between intravariceal pressure and the pressure in the oesophageal lumen. With progression of portal hypertension, the variceal blood flow and intravariceal pressure increases thus increasing the radius of the varix and decreasing the wall thickness, thereby increasing the transmural pressures. Areas with decreased tissue support (oesophagus & rectal in particular) have higher transmural pressures, as there is limited capacity to counter the expanding force.
Figure 3. Relative position and size of varices with respect to the oesophageal lumen. At equal transmural pressure(TP), the tension (T) developed will be greater in T1 than in T2 due to the increased radius(r)and smaller wall thickness(w).

Bleeding occurs probably when the expanding force can no longer be counter-balanced by the variceal wall tension; at this point the varices ruptures and bleed. Portal pressure and blood flow vary markedly with various physiological stimuli. Portal flow increases transiently after meals because of postprandial hyperaemia. The effect of octreotide is related its ability to limit this postprandial hyperaemia. Similarly, fluctuations in portal pressure has been observed during circadian rhythm, with portal pressure being higher at night and lower during afternoon and evening . These variations in portal pressure may influence the onset of bleeding in patients with already elevated baseline portal pressure. Clinical Manifestations & disease severity scoring Refer to previous seminars Diagnosis Portal hypertension is an increase of portal pressure above the upper limit of normal; ideally, it should be evaluated by actually measuring the pressure in the portal vein. However, direct measurement of portal pressure is invasive and impractical, and surrogate methods must be used. Hepatic portal venous gradient 15,16 In 1951, an indirect measurement method was described in which a small catheter was advanced into the hepatic vein and wedged so that it could not be advanced any further, and this formed a continuous static column of blood between the catheter and the hepatic sinusoids. This wedged hepatic vein pressure (WHVP) equals the pressure in the hepatic sinusoids, which in turn corresponds to the portal vein pressure (in the absence of presinusoidal portal

P a g e | 11

hypertension). To eliminate variations in abdominal pressure, the free hepatic vein pressure (FHVP-unwedged pressure in the hepatic vein) is currently used as an internal zero reference. The difference between WHVP and FHVP, the hepatic venous pressure gradient(HVPG), is a measure of the severity of portal hypertension. HVPG measurement is the gold standard technique to evaluate the presence and severity of portal hypertension. HPVG has a strong predictive power for mortality independent to severity of liver disease. It also predicts the advent of complications of cirrhosis. Patients with an HVPG below 10 mm Hg have a 90% chance of not decompensating over the next 4 years. Oesophageal varices develop only at an HVPG greater than 10 mm Hg and bleed at an HVPG greater than 12 mm Hg. In patients with acute variceal hemorrhage, an HVPG > 20 mm Hg is associated with worse outcomes and identifies patients who would benefit from early transjugular intrahepatic portosystemic shunt (TIPS) placement or other therapies. For primary prophylaxis, patients who decrease their HPVG by > 20% have a 5% risk of bleeding in the next 2 to 3 years and those in whom the HPVG is reduced to less than 12mmHg, the annual risk of bleeding is virtually zero. Therapeutic targets are assessed by the measurement of HPVG after institution of secondary prophylaxis. It has been shown that by decreasing the HPVG by > 20% of baseline or to less than 12 mmHg has shown to decrease the incidence of rebleeding. At present, HVPG measurement is the best surrogate of portal pressure. It provides valuable information for establishing diagnosis, for predicting prognosis, and, most importantly, for making therapeutic decisions in portal hypertensive patients. It is safe to perform and is accurate and there is good inter-observer correlation with measurements. Further, clearly defined targets for therapy have been established using the HPVG, a reduction of which has been proven to decrease the complications of portal hypertension. Abdominal ultrasound and Doppler 18,19 Signs seen on ultrasound to suggest portal hypertension are splenomegaly (spleen length >13 cm), portal vein dilatation, umbilical vein patency, the presence of other portosystemic collaterals and ascites. A portal vein diameter of greater than 13 mm generally signifies portal hypertension. Doppler ultrasound also allows the calculation of portal flow velocity, portal flow, demonstates reversal of flow and other haemodynamic parameters. The pros of ultrasound is that it is non invasive and easily repeatable. However, its use is limited by the fact that that evaluation is strongly dependant on the skills of the operator, and there is a high inter-observer and inter-equipment variability demonstrated. Oesophagogastroduodenoscopy (OGD) OGD is considered the gold standard for diagnosis of gastrooesophageal varices because it enables the endoscopist to obtain a direct view of the varices and decide on the management by assessing the size of the varices and the presence of red wale marks and cherry spots. Also, if prophylactic or therapeutic variceal banding is warranted, it can be performed at the same sitting. 19 In addition, endoscopy allows the identification of other potentially bleeding lesions unrelated to portal hypertension. Endoscopic findings unrelated to portal hypertension requiring a change in management were reported in 39.4% of asymptomatic cirrhotic patients undergoing endoscopy for the screening of portal

P a g e | 12

hypertension. Common lesions discovered were gastritis (22.6%), oesohagitis (5.2%), peptic ulcer disease (5.6%) and the most frequent vascular anomaly was GAVE (1.8%). This highlights the importance of endoscopy in these patients with lesions that would be missed by less invasive screening modalities such as CT, MRI or even video capsule endoscopy. 20 Capsule endoscopy (CE) 17,20 Capsule endoscopy is a less invasive method to evaluate the GI tract in patients with portal hypertension. The initial studies showed good correlation with OGD. Subsequently follow up studies have not mirrored these result with sensitivity and specificity ranging between 64.5-84% and 66.7-88% respectively, compared to OGD. There is also poor inter-observer correlation, which makes the results more subjective. This makes it clearly less effective than OGD. However it may be an option in patients who are unable to tolerate OGD. 21 Endosonography EUS appears to be as good as EGD for detection of clinically significant oesophageal varices,and may be better at identifying gastric varices as a results of its ability to differentiate between enlarged gastric folds and gastric varices. In the monitoring of therapy, EUS has been used for the prediction of the recurrence of varices. Other EUS features, such as higher cross-sectional area of varices, perforating veins greater than or equal to 3 mm in diameter before therapy, increase in diameter and lack of hepatopetal flow in the left gastric vein, and anterior branch dominance with rapid hepatofugal flow in the left gastric vein, predict a higher risk for recurrence of varices after obliteration. It may also be used to guide sclerotherapy. 22 Computed tomography (CT) 19 In the determination of the size of varix and thus the risk of bleeding, varices greater than 3 mm in diameter are considered to be large. CT grading of varices strongly correlates with endoscopic grading. CT oesophagography has been shown to have a sensitivity of >90% for large varices, although this falls to <70% if all varices are considered, but 87% sensitivity for detection of gastric varices. A significant number of additional gastric and periesophageal varices and extraluminal pathology are also identified on CT. Patients also prefer CT compared to unsedated upper endoscopy. Aside from the cost and radiation dose of the procedure, CT does not allow assessment of red signs. Box 2.EUS finding in Therefore, high-risk varices that are small or that have PH 22 red signs may be missed with this technique. Further, Dilation of portal, splenic as alluded to earlier, these patients frequently have and azygous veins pathologies other than those associated with portal Dilation of the thoracic hypertension at ODG, largely mucosal lesions, than duct cannot be detected by CT. Presence of: MRI 19 Paraoesophageal MRI with elastography is being extensively studied as a collateral veins noninvasive tool to determine fibrosis in the liver. MRI Perigastric collateral veins also provides an excellent view of the vascularity of the Paragastic collateral liver and the flow through the portal and azygous veins. veins In patients with portal the azygous blood flow is higher.
Intramural veins Perforating veins Thickness of gastric mucosa and submucosal layers

P a g e | 13

Gadolinium enhanced MRI has been shown to be 81% compared to oesophagogastroduodenoscopy. It has also been shown that MRI may be equivalent to EUS in the detection of fundal varices. Endoscopic grading of oesophageal varices19 Endoscopic grading of esophageal varices is quite subjective and interobserver variability exists. Three grading systems exist: by Dagradi (1972); by the Japanese Research Society for Portal Hypertension (JRSPH, 1980); and by the North Italian Endoscopy Club for the Study and Treatment of Esophageal varices (NIEC, 1988). The most widely used grading system is the NIEC The NIEC index takes into account the following: The Child-Pugh class of cirrhosis (A, B, or C) Variceal size (small, medium, or large) Presence of red color signs (absent, mild, moderate, or severe). The JRSPH and NIEC classifications were highly specific to predict variceal bleeding (93.4% and 94.8%, respectively) but not sensitive. All three systems had low positive predictive values. Screening Up to 2/3rd of patients with cirrhosis will develop eosophogeal varices. One third of these patients will have a variceal haemorrhage within a year of their diagnosis with an associated mortality of 20 to 40% per bleeding episode.23 The prevalence of oeosphageal varices in cirrhosis ranges from 0% in compensated cirrhotics to 80% in decompensated cirrhosis with the mean prevalence being 55%.The annual incidence of gastrointestinal bleeding is only 1-2% in patients without varices, 5% in those with small varices and 15-20% in patients with large oesophageal varices.24 The current consensus position is to screen all patients with a diagnosis of liver cirrhosis with OGD.25 As the prevalence of large oesophogeal varices is only 936% in patients with cirrhosis who have not bled, a large number of invasive endoscopic procedures turn out to be negative. Thus, there is a need for non-invasive means to diagnose or predict the presence or absence of large oesophogeal varices. Availability of such methods may help limit the number of endoscopic procedures performed for detection of large oesophogeal varices.24 Various non-invasive methods have been studied to predict the presense of varices in these patients, but as yet there is little consensus on a selective endoscopic screening policy. Criteria to suggest the presence of varices are, patients with prothrombin activity greater than 70%, a portal vein diameter greater than 13mm and a platelett count of less than 100 x 109/L. Other methods suggested are test to measure lever stiffness (Fibrotest, Fibroscan), the platelett spleen ratio of >909 and transient elastorgraphy. Validation of these methods is ongoing.26 Management

P a g e | 14

Pre primary prophylaxis (absence of varices on OGD) Treatment with a nonselective beta-blocker does not prevent the development of gastrooesphageal varices in unselected patients with cirrhosis and portal hypertension. The use of NSBB was related to an increased number of adverse effects compared to placebo including bradycardia, fatigue and syncope requiring diagnosis or intervention. In this patient population there is little or no significant change in HPVG with treatment, thus there is currently no role for its routine use in preprimary prophylaxis.27 While pharmacological therapy to reduce portal pressures has been proven ineffective for the prevention of the development of OVs, an alternative approach is to prevent the progression of cirrhosis (i.e. abstinence in alcoholics, anti-virals in viral cirrhosis, corticosteroids in autoimmune hepatitis, phlebotomies in haemochromatosis, copper chelators in Wilsons disease). For example, alcohol abstinence has been demonstrated to decrease the incidence of complications of cirrhosis and PHT resulting in a prolonged survival. Prevention of the first variceal bleed (primary prohylaxis) The yearly incidence of first variceal bleed in cirrhotic patients is estimated at 4%, but when medium-large varices are present this risk increases to 15%. The main risk factors predicting variceal bleeding are the size of the oesophageal varices, the severity of liver disease and the presence of red wale markings. Therapeutic options discussed Pharmoacotherapy Oesophageal varices bleed only when the HPVG is greater than 12 mmHG. A reduction in the HPVG below 12 mmHG or a decrease of > 20% of baseline markedly reduces the risk of variceal haemorrhage. This may be achieved with drugs or by shunting procedures. Non-selective B-blockers (NSBB). Propanolol, timolol and nadolol are non selective beta blockers that lower portal venous inflow by 2 receptor (vasodilator) inhibition, thus causing unopposed 1 vasoconstriction in the splanchnic circulation, and by decreasing the cardiac output by 1 inhibition. Additional benificial effects of beta blockade beyond its haemodynamic effects include a proposed decrease in bacterial translocation and the decrease in the incidence of spontaneous bacterial peritonitis in ascitic cirrhotic patients.28 Recently, carvedilol, an NSBB with mild anti-1 adrenergic (vasoconstrictive) effects, has been compared with EBL for the primary prophylaxis of VH. Conventional NSBB cause splanchnic vasoconstriction, but do not address the increased intrahepatic vascular resistance. This drug seeks to address the intrahepatic vasoconstriction, but may worsen systemic hypotesion in advanced disease.29 However, NSBBs do not protect all patients from variceal haemorrhage (VH). Only a third to half of patients with cirrhosis on NSBBs achieve a therapeutic reduction in HVPG below 12 mm Hg or >20% from baseline levels. Further difficulties include non-compliance or inadequate dosing. The routine evaluation of beta blockade efficacy by performing monthly to 3 monthly HVPG measurements may be warranted but is subject to further study. Therapeutic role. There is no current role for NSBB for pre-primary prophylaxis.27

P a g e | 15

In primary prophylaxis, beta blockers reduced the bleeding risk from 30 to 15%; in secondary prophylaxis, this risk decreased from 60 to 42% in the first year.30 Adverse Drug Effects. The rate of adverse is estimated at 17% over 2 years. Severe adverse events included cardiac failure, asthma, and symptomatic bradycardia. Less serious adverse events causing discontinuation include fatigue, insomnia, and sexual dysfunction. Absolute contraindications include obstructive pulmonary disease, aortic valve disease, second- and third-degree atrioventricular heart block, and peripheral arterial insufficiency. Relative contraindications are sinus bradycardia and insulin-dependent diabetes mellitus. Dosing. Propranolol is administered orally twice a day, increasing the dose every 2 days so that the titrating phase is as short as possible, ideally 2 weeks. If nadolol is employed, it is recommended that this be administered once a day because of its longer half-life.31 Therapy should ideally be targeted to the reduction in HPVG. The greatest therapeutic effect occurs if the HPVG decreases to less than 12mmHg, or a 20% drop from baseline. Because HVPG measurement is not widely available and because a reduction in heart rate does not correlate with reduction in HVPG, the dose of nonselective b-blockers (propranolol, nadolol) is adjusted to maximal tolerability, maintaining a resting pulse of approximately 55 beats per minute, or a reduction from the baseline heart rate of >25%, or the development of side effects. An important problem with -blockers is their variable effect on portal pressure and the consequent difficulty in predicting a clinical response. Intolerance to therapy (10%20%) and rebound portal hypertension if stopped suddenly are further problems. 32 Nitrates.Short-acting (nitroglycerin) or long-acting (isosorbide mononitrates) nitrates cause venodilatation, rather than arterial dilatation, and decrease portal pressure by decreasing portal venous blood flow. Although initial studies where promising in preventing bleeding, there is currently no evidence for their use in primary prophylaxis. The may be added to NSBB in secondary prophylaxis. Angiotensin receptor blockers. 33 In the cirrhotic patient, the rennin angiotensin system (RAS) is frequently activated, and plasma level of AT-II is elevated. AT-II induces sodium and fluid retention by the stimulation of aldosterone secretion, and induces contraction and fibrogenesis of the hepatic stellate cells thereby increasing intrahepatic vascular resistance. Initial studies using Losartan showed a 40 % decrease in portal pressures without significant effects on the systemic circulation. ARBs have a potent antifibrotic effect. Termisartan prevents liver fibrogenesis by inhibiting the HSCs activation and proliferation, thus decreasing progressive hepatic fibrosis. Further evaluation is required. Endoscopy Endoscopic band ligation (EBL) EBL was first reported in 1989. The technique involves the placement of rubber bands around a portion of the varix containing esophageal mucosa. The varix is sucked into a hollow, clear plastic cylinder attached to the tip of the endoscope. Once suctioned into the sheath, a trigger device allows deployment of the band

P a g e | 16

around the varix. The blood flow is completely interrupted, producing ischemic necrosis of the mucosa and submucosa. Later granulation takes place with sloughing of the rubber rings and necrotic tissue, leaving shallow mucosal ulcerations that heal in 14 to 21 days. Application of the bands is started at the gastroesophageal junction and progresses cephalad in a helical fashion. EBL sessions are repeated at approximately 2-week intervals until varices are obliterated, usually requiring 2 to 4 ligation sessions. Complications are less frequent than with ES. 32 Endoscopic band ligation (EBL) has been demonstrated to prevent bleeding effectively in patients with medium to large varices. This technique has been compared to b-blockers as a first-line option for primary prophylaxis of variceal bleeding in 16 trials and the global results have been reported in two meta-analyses, showing an advantage of EBL over b-blockers in terms of prevention of first variceal bleeding, but without differences in mortality. These data have led to an extensive debate on whether b-blockers should remain the first treatment option for preventing first variceal bleeding or shouldbe replaced by EBL. 35 Endoscopic band ligation vs. NSBB for primary prophylaxis 35

Endoscopic Sclerotherapy as primary prophylaxis Sclerotherpy as primary prophylaxis for variceal haemorrhage has been well studied. A direct comparison between trials is difficult to make as technique varied, with a variety of sclerosants in different doses injected intravariceally or perivariceally or both. Although some early trials showed a reduction in bleeding in the sclerotherapy group more recent and larger trials have shown either no value or a deleterious effect of sclerotherapy. 32

P a g e | 17

Endoscopic sclerotherapy is therefore not indicated as primary prophylaxis. Endoscopic treatment: how frequent, how to monitor the treatment 36 There is no agreement on how frequently the varices should be ligated in the initial course of eradication, the interval varying from 1 to 4 weeks. The current

Box 3.Prophylaxis of first variceal bleeding in patients with cirrhosis 25,34,38,39 Clinical situation Treatment Follow up endoscopy Comments
No varices Small varices benefit of None Every 3 yrs -blockers optional Every 2 yrs Treat underlying disease Long term

Small varices + red-blockers Signs on varices or Child-Pugh class B/C Large varices Endoscopic band Both therapies Ligation or -blockers patient

Every yr if Child-Pugh B/C -clockers not established Not needed Every 6 months effective Consider local expertise, preferences &

contraindications/ intolerance to -blockers

evidence, although weak, currently evidence favours monthly intervals. This does not apply to prophylaxis of recurrent bleeding (in which the risk of rebleeding is maximal in the first few weeks) where a 12 week interval might be more appropriate. Once the varices are eradicated, follow-up endoscopies should be performed at 13 months and every 6 months thereafter, and varices should be re-eradicated upon recurrence. This is in marked contrast with prophylaxis with beta-blockers, in which no follow-up endoscopies are needed. Acute Variceal haemorrhage (AVH) Mortality rates from AVH have dropped dramatically in the past 3 decades from 42% in the 1980s to the current rates of 15% to 20% probably due to improvements in intensive care and therapeutic advances in this period. Approximately one-third of deaths are a direct consequence of bleeding and the remaining are caused by liver failure, infections, and renal failure.39 General management

P a g e | 18

Patients who have suspected acute variceal hemorrhage should ideally be admitted to an ICU setting for resuscitation and management. Initial resuscitation involves basic measures, including assessing the patients airway, obtaining peripheral venous access, and initiating fluid resuscitiaiton. Large volumes of blood in the stomach have the potential to be aspirated. Emergent intubation should be considered especially in those with encephalopathy.40 Packed red cells are transfused conservatively to keep the target hemoglobin after transfusion around 8-9 gm/dL (hematocrit 25%30%), as overtransfusion may exacerbate portal hypertension and Box 4 . General measures for the management of active variceal increases the risk for rebleeding. This should hemorrhage 36 however be tailored to individual patients, with those patients with ischemic heart Airway protection disease and those with ongoing bleeding Endotracheal intubation if altered mental requiring higher transfusion goals. Fresh status or unconscious frozen plasma and platelets, although Gastric aspiration frequently used, do not reliably correct Hemodynamic resuscitation coagulopathy and can induce volume Crystalloids and blood transfusion overload. Correction of coagulopathy and Recombinant factor VIIa (rFVIIa) in patients thrombocytopenia with cirrhosis and gastrointestinal Antibiotic prophylaxis for spontaneous bacterial peritonitis hemorrhage failed to show a beneficial Blood cultures and diagnostic effect versus standard therapy. It has been paracentesis if ascites present shown to decrease bleeding in patients with Third-generation cephalosporin a Child-Pugh score >9 from days 5 to 42 intravenously and switch to oral after the index bleed. Due to its prohibitive quinolone when patient is stable and GI cost and failure to control bleeding in the tract is functional acute setting even in patient with advanced Renal support disease it cannot currently be recommended Maintain urine output >50 mL/h as routine therapy. In patients who are Avoid nephrotoxic drugs refractory to all other therapeutic Metabolic support modalities, it may be considered an option. Inject thiamine when indicated
41, 42

Prevention and treatment of infection

41,42

Cirrhosis is frequently associated with defects in both humoral and cellular host defense. Up to 20% of patients with cirrhosis and gastrointestinal bleeding have a bacterial infection at initial hospitalisation and further 50% may later during the hospitalisation develop infection. Presence of bacterial infections seem to be closely related to prognosis in bleeding cirrhotic patients, moreover bacterial infections are also associated with a higher risk of variceal rebleeding. Most frequent causes of bacterial infections are urinary tract infections (1229%) caused by gram-negative bacilli, spontaneous bacterial peritonitis (723%) caused

Monitoring blood glucose level Monitor and treat for delerium tremens Monitor and treat for acid base and electrolyte disturbances Neurologic support Monitor mental state Avoid sedation

P a g e | 19

by gram-negative bacilli and aerobic gram-positive cocci and Staphylococcus species. Short-term antibiotic prophylaxis has a significant beneficial effect on mortality, with a decrease of approximately 9%, as well as a decrease in the incidence of bacterial infections. The most studied antibiotic studied for this purpose have been quilones (norfloxacillin specifically), with intravenous and oral administration appearing equally efficacious (oral quinalones are well absorbed orally). Recent data suggests that intravenous cephalosporins (Ceftrioxone) may be better at preventing infections. Bacterial infections occurred in 33% of patients using Norfloxacillin compared to 11% of patients on ceftrioxone. Intravenous ceftriaxone should be considered in patients with advanced cirrhosis, in hospital settings with high prevalence of quinolone-resistant bacterial infections and in patients on previous quinolone prophylaxis.25 All cirrhotics with upper gastrointestinal bleeding should receive prophylactic treatment with antibiotics either Quinolones or intravenous cephalosporins(in high risk patients) for 57 days. Aminoglycosides should be avoided due to risk of renal toxicity. Pharmacotherapy 41 Vasoactive drugs can be administered in the hospital, at home or during transfer to the hospital, which may improve survival in patients with massive bleeding. They also facilitate endoscopy. The selection of the drug depends on the local resources. Terlipressin should be the first choice if available, since it is the only drug that has been shown to improve survival. Somatostatin and somatostatin analogues (octreotide or vapreotide) are second choice. If these drugs are not available vasopressin plus transdermal nitroglycerin is an acceptable option. Vasopressin is a posterior pituitary hormone that acts as a potent vasoconstrictor leading to marked decreases in splanchnic blood flow. It causes a decrease in the HPVG by 23%. Its use is associated with complications related to its vasoconstrictive and antidiuretic effects which may be be ameliorated with the concurrent use of nitroglycerin. These include myocardial infarction, peripheral vascular ischemia, bradycardia, hypertension, hyponatremia, and fluid retention. Terlipressin is a long-acting synthetic analogue of vasopressin that is slowly cleaved to lysine vasopressin in vivo. Compared with vasopressin, lysine vasopressin has a prolonged half-life, achieves lower blood concentrations of the active drug form, has a higher tissue penetration, and has a milder cardiovascular side-effect profile. Terlipressin modifies systemic haemodynamics with a decrease in cardiac output and an increase in the arterial blood pressure and the systemic vascular resistance. This leads to a decrease in splanchnic inflow, with a decrease in portal pressures of up to 20% seen within an hour. The effect is achieved within 30 min and is still significant 4 h after administration. This is the only treatment that has been shown to improve prognosis of variceal bleeding and may prevent rebleeding when administered for prolonged periods. The overall efficacy of terlipressin in controlling acute variceal bleeding at 48 h is of 7580%, and of 67% at 5-days. The most common side effect of this drug is abdominal pain. Serious side effects such as peripheral or myocardial ischemia occur in less than 3% of the patients.

P a g e | 20

Somatostatin inhibits the secretion of vasodilatory peptides from the gastrointestinal tract, including glucagon, which have been shown to contribute to the maintenance of portal hypertension. Its half life is 2 minutes. It is empirically used as an initial bolus of 250 g (may be repeated three times in 24 hrs) followed by a 250 g/h infusion that is maintained until a 24 h bleed-free period. Higher doses of somatostatin (500 g/h) may have increased clinical efficacy. Major side effects are rare. Minor side effects, such as nausea, vomiting and hyperglycemia occur in up 30% of patients. Somatostatin significantly improves the rate of control of bleeding compared with placebo, but has no proven mortality benifit. Somatostatin has been compared with terlipressin and no differences were found for failure to control bleeding, rebleeding, mortality or in the incidence of adverse events in both treatment groups. Somatostatin has also been compared with vasopressin in RCTs that showed equal efficacy but increased safety with somatostatin. Somatostatin analogues. The cyclic octapeptides, octreotide, lanreotide, and vapreotide, are synthetic analogues of somatostatin. Octreotide has a longer half-life (1.5 hrs) compared to somatostatin. This, however, is not associated with longer hemodynamic effects. Octreotide, thought to act through the inhibition of glucagon to blunt postprandial hyperemia and thereby limits increases in portal pressure. It is usually given as an initial bolus of 50 g, followed by an infusion of 25 or 50 g/h. Therapy should be maintained for 5 days to prevent early rebleeding. The efficacy of octreotide as a single therapy for variceal bleeding is controversial. No benefit from octreotide was found in the only trial using octreotide or placebo as initial treatment, which may be due to rapid development of tachyphylaxis. However, RCTs using octreotide after sclerotherapy have shown a significant benefit in terms of reducing early rebleeding. These results suggest that octreotide may improve the results of endoscopic therapy but has uncertain effects if used alone. When compared with other vasoactive drugs, octreotide was better than vasopressin and equivalent to terlipressin, again suggesting a clinical value from the use of octreotide. Although somatostatin and somatostatin analogues have similar side-effect profiles, pulmonary edema, paralytic ileus, liver toxicity, and rebound gallbladder hypermotility (following withdrawl of long-term use) have been reported with the use of the cyclic octapeptides. Endoscopy in acute variceal bleeding The gold standard for diagnosis is upper endoscopy, which may show one of the following: active blood spurting or oozing from a varix, which is present in nearly 20% of patients; white nipple or clot adherent to a varix; or presence of varices without other potential sources of bleeding in the stomach or duodenum. The 2 endoscopic methods available for AVH are endoscopic sclerotherapy (ES) and band ligation (EBL). Both are effective in controlling bleeding. Meta-analysis shows that EBL is better than sclerotherapy in the initial control of bleeding, and is associated with less adverse events and improved mortality. Additionally, sclerotherapy, but not EBL, may increase portal pressure. Therefore EBL is the endoscopic therapy of choice in acute variceal bleeding, though injection sclerotherapy is acceptable if band ligation is not available or technically difficult.1

P a g e | 21

Endoscopic sclerotherapy consists of the injection of a sclerosing agent into the variceal lumen (intravariceal) or adjacent to it (paravariceal). This causes thrombosis of the varix and inflammation of the surrounding mucosa that creates a scar over the esophageal wall. A variety of sclerosant solutions may be used, with non being superior to another. The most common are ethanolamine oleate (5%) or polidocanol (1%2%) in Europe, and sodium morrhuate (5%) in the United States.36,37 The first injection of 1 to 2 mL of the sclerosant should be placed right below the bleeding site. Afterward, 1 to 2ml injections of the remaining varices adjacent to the bleeding varix are performed. The main objective is to target the lower esophagus near the gastrooesophageal junction. Up to 10 to 15 mL of a sclerosant solution may be used in the session. In the acute setting, the paravariceal injection technique cannot be easily performed because of the ongoing bleeding and it is mostly reserved for elective sclerotherapy. Immediate problems include substernal chest pain, fever, dysphagia, and pleural effusion. Oesophageal ulcers are common and in 20% of patients they may cause bleeding. Bacteremia may occur in up to 35% and lead to other complications such as spontaneous bacterial peritonitis or distal abscesses. Other complications include esophageal strictures, perforations, mediastinitis, pericarditis, chylothorax, esophageal motility disorders, and acute respiratory distress syndrome. Endoscopic variceal ligation was developed as an alternative, with fewer complications than ES, for the treatment of esophageal varices. It consists of the placement of elastic bands on the varices to occlude the varix and cause thrombosis.39 This causes necrosis of the mucosa and the bands eventually fall off in a few days leaving a superficial mucosal ulceration that heals and eventually scars. Repeat sessions are performed at 2- to 4-week intervals to reduce the risk of rebleeding. Balloon tamponade 39,43 Balloon tamponade arrests AVH by a direct tamponade effect at either the gastooesophageal junction, to interrupt the left gastric supply, or in the oesophageal lumen to cause direct compression. Three types of balloon are available, namely the Sengstaken-Blakemore (SB) tube, the Minnesota(M) tube (both with 100-200ml gastric balloons) and the Linton-Nachlas (LN) tube (600ml gastric balloon). The gastric balloon, once in the stomach, is inflated with water, pulled back to sit at the GO junction, and secured to a 0.5kg weight to maintain this pressure. If this does not arrest the bleed, the pressure of the oesophageal balloon may be increased to 3040 mmHg, to cause a tamponade effect. The optimal duration of balloon tamponade is believed to be around 1224 hours. Effectiveness is about 8094%. Bleeding from OV is better controlled with the SB or M tubes, whereas bleeding GV are better managed with LN tube. However, the use of balloon tamponade is very unpleasant for the patients and the rebleeding rate is about 50% after the balloon is deflated. Thus, balloon tamponade should be reserved to patients with massive variceal hemorrhage, which cannot be arrested by vasoconstrictors or endoscopic therapy, or when endoscopists are unavailable for emergency endoscopic therapy. The rate of complications (i.e. nasal alar necrosis, esophageal ulcers, esophageal rupture, airway obstruction and aspiration pneumonia) with balloon tamponade is

P a g e | 22

approximately 15%, and 6% of complications were fatal. Once the balloon tamponade is removed, endoscopic therapy should be performed to check for and prevent early rebleeding as soon as possible. Covered Oesophageal stents39 A covered oesophageal self-expandable stent as been developed and used in patients with failed medical and endoscopic therapy. The stent is placed over a guide wire previously passed to the stomach. The stent has a distal balloon that inflated with a syringe to ensure proper location in the cardia and lower esophagus and does not require fluoroscopic for placement. The stent can be left in place for up to 14 days and it can be retrieved by endoscopy with a hook system. The pilot study recruiting 20 patients for the abovementioned purpose, showed a success rate of 100% without significant complication. Further studies are required to evaluate its role in AVH. Failure to control bleeding Although acute bleeding from varices may cease spontaneously in nearly half of patients, rebleeding rates are significantly high (30%40%) if patients are not treated. The highest risk occurs within the first 2 to 3 days after the index bleed and most of the rebleeding episodes occur within the first 14 days. After 6 weeks, the risk of further bleeding is similar to that before the index bleed. Initial failure to control bleeding occurs more commonly in patients with Child class C cirrhosis, bacterial infection, portal vein thrombosis, active spurting of a varix, and an HVPG >20mmHg. Failure to control bleeding: Baveno V 25 The time frame for the acute bleeding episode should be 120 h (5 days). Failure is defined as death or need to change therapy defined by one of the following criteria: Fresh hematemesis or NG aspiration of >100 ml of fresh blood 2 h after the start of a specific drug treatment or therapeutic endoscopy. Development of hypovolaemic shock. 3 g drop in Hb (9% drop of Ht) within any 24 h period if no transfusion is administered. (validation required) Rescue therapies 39 Despite urgent endoscopic and/or pharmacologic therapy, variceal bleeding cannot be controlled or recurs in about 10% to 20% of patients. An elevated HVPG of more than 20 mmHg has been shown to be predictive of treatment failure. Shunt therapy, either as shunt surgery (in Childs A patients) or transjugular intrahepatic portosystemic shunt (TIPSS), has clinical efficacy as salvage therapy. TIPSS 39 TIPSS is a radiologic procedure by which a tract is created between the hepatic vein and portal vein ( in essence a non-selective shunt) and kept open by deployment of a coated stent. It produces hemostasis in over 90% of cases. TIPSS-related complications include procedural complications (in about 10%) and later TIPS dysfunction or stenosis (decreasing with covered stents). The most

P a g e | 23

frequent complication encountered is that of worsening encephalopathy, which occurs in about 25% of patients following TIPSS placement. Contraindications to TIPS include severe congestive heart failure, severe pulmonary hypertension, severe hepatic failure, portal vein thrombosis with cavernomatous transformation, and polycystic liver disease. A recent randomised trial explored whether patients with poor prognostic indicators might benefit from a more aggressive therapeutic approach on initial presentation. Patients with high risk (with HVPG > 20 mm Hg) were randomised to receive standard therapy or TIPS. Those who underwent early TIPS had significantly less treatment failure and lower mortality than patients undergoing standard therapy. Surgical 39 The surgical options currently used to treat variceal bleeding are either shunt or nonshunt procedures. The shunt operations are either total, including portacaval shunt, narrow-diameter portacaval shunt, mesocaval shunt, and central splenorenal shunt, or selective, including the distal splenorenal shunt (DSRS) and coronary caval shunt. The nonshunt operations are primarily gastroesophageal devascularization or oesophageal transection. A large multicenter trial of TIPSS versus DSRS (the DIVERT study) showed similar rates of rebleeding, encephalopathy, and mortality in patients with Childs A and B cirrhosis who had failed pharmacologic/endoscopic therapy, with a higher rate of shunt dysfunction in the TIPS group. Because both procedures have equivalent outcomes, the choice is dependent on available local expertise. 39 Current recommendations for acute variceal haemorrhage : Baveno V 25 In suspected variceal bleeding, vasoactive drugs should be started as soon as possible, before endoscopy. Vasoactive drugs (terlipressin, somatostatin, octreotide, vapreotide) should be used in combination with endoscopic therapy and continued for up to 5 days. The early administration of a vasoactive drug facilitates endoscopy and improves control of bleeding and 5-day rebleeding. Conversely, the association of endoscopic therapy also improves the efficacy of vasoactive treatment. Patients with GI bleeding and features suggesting cirrhosis should have upper endoscopy as soon as possible after admission Endoscopic therapy is recommended in any patient who presents with documented upper GI bleeding and in whom esophageal varices are the cause of bleeding. Ligation (EVL) is the recommended form of endoscopic therapy for acute esophageal variceal bleeding, although sclerotherapy may be used in the acute setting if ligation is technically difficult. Balloon tamponade should only be used in massive bleeding as a temporary bridge until definitive treatment can be instituted (for a maximum of 24 h, preferably in an intensive care facility).

P a g e | 24 Uncontrolled data suggest that self-expanding covered esophageal metal

stents may be an option in refractory esophageal variceal bleeding, although further evaluation is needed. Persistent bleeding despite combined pharmacological and endoscopic therapy is best managed by TIPS with PTFE-covered stents . Re-bleeding during the first 5 days may be managed by a second attempt at endoscopic therapy. If re-bleeding is severe, PTFE-covered TIPS is likely the best option. Prevention of variceal rebleeding/Secondary prophylaxis 44,45 The risk of rebleeding is up to 50% within 6 weeks and up to 70 to 80% at 2 years. The mortality of each bleeding episode ranges between 20% and 40%. Factors that predict the likelihood of rebleeding are the severity of liver disease, presence of hepatocellular carcinoma, continued alcohol abuse and variceal size. Due to the high mortality associated with rebleeding, measures to decrease these events have been suggested. Therapeutic goals for secondary prophylaxis is a reduction in the HPVG to less than 12mmHg or a decrease > 20% of baseline. However the risk of rebleeding in patients who achieve these targets in portal pressure still remains around 1520%. Endoscopic band ligation has replaced injection sclerotherapy as first-line therapy because it has fewer complications, is easier to perform, requires fewer sessions and lowers the risk of rebleeding more efficiently than sclerotherapy and may have a mortality benefit. The combination of EBL and then sclerotherapy has no additional benefit and therefore should not be used. The use of combination therapy, a beta blocker plus a nitrate or combination of EVL and drug therapy is required due to the high risk of recurrence, even though the side effects will be greater with combination therapy than either treatment alone. Treatment should be started as soon as possible from day 6 of the index variceal bleed. Studies have shown that rebleeding rates are lower with combination therapy of EVL and drug treatment than with drug therapy alone. However there was no mortality benefit. In patients who were treated with either NSBB or EVL as primary prophylaxis and bled, should undergo therapy with a combination of the above. TIPSS for first line therapy as secondary prophylaxis is not rommended. While TIPSS significantly reduces portal pressures and decreases the rebleeding rate, and is well tolerated by patients, it does carry the risk hepatic encephalopathy and and deterioration of liver function. Further repeated reintervention and a trend toward lower mortality in patients not treated with TIPSS, limit its use as first line therapy for seconday prophylaxis. It use in rebleeding should be reserved for those patients that fail combination therapy or as a bridge to transplantation.

P a g e | 25

Figure 4. Prevention of variceal rebleeding: rebleeding rates with different treatment options. The bars represent median values from data reported in metaanalysis. Note the extremely low rebleeding rate in patients treated with beta-blockers and who are HVPG responders (fall in HVPG of at least 20% of baseline and/or to 12 mm Hg or below). Failure of secondary prophylaxis 25 Failure to prevent re-bleeding is defined as a single episode of clinically significant rebleeding from portal hypertensive sources after day 5. Clinically significant re-bleeding: recurrent melena or hematemesis resulting in any of the following: hospital admission, blood transfusion,3 g drop in Hb or death within 6 weeks. In those patient who bleed despite the combination therapy, should undergo TIPSS placement or derivative surgery stratified by the severity of their liver disease, patient preference and local expertise. In patients who a good risk candidates, transplantation, which provides good long term outcomes, should be considered. TIPSS may be used as a bridge to surgery. Gastric Varices 51 These varices are present in 5% to 33% of patients with portal hypertension with a reported incidence of bleeding of about 25% in 2 years, with a higher bleeding incidence for fundal varices. 46 Management of GV presents a challenge for several reasons: there is no consensus regarding the optimum treatment of GV, thus treatment tends to be empiric GV are not a homogenous entity, and accurate classification defines the natural history and should dictate the management; although GV bleed less frequently than EV, bleeding tends to be more severe, to require more transfusions, and to have a higher mortality rate than EV bleeding; after control of acute bleeding, GV have a high rebleeding rate (34% to 89%)

P a g e | 26

Classification of gastric varices The most commonly used classification system is that used by Sarin et al. This catagorises GV on their location and their relationship with oesophageal varices. Gastrooesphageal varices (GOV) refer to varices that extend from the oesophagus to the stomach. Type 1 GOV (GOV1) : gastric varices continuous with oesophageal varices and extending along the lesser curvature 2-5 cm below the gastrooesphageal junction. Type 2 GOV (GOV2) : varices, often long and tortuous, extending from the oesophagus below the gastrooesophageal junction toward the fundus. Isolated GV (IGV) Refer to gastric varices that occur in the absence of oesophageal varices Type 1 IGV (IGV1): varices located in the fundus that often are tortuous and complex in shape; and Type 2 IGV (IGV2): ectopic varices in the antrum, corpus, and around the pylorus. Duodenal varices may sometimes also be included in this group. A duodenal varix is usually single and is found in the first or second part of the duodenum. GV may also be considered as primary or secondary. Primary varices are those discovered on initial screening endoscopy or in patients who have not undergone variceal ligation or sclerotherapy previously. Secondary GVs refer to those varices that develop after endoscopic treatment for oesophageal varices. Prevalence GVs are present in approximately 20% of patients with portal hypertension and secondary GVs have been discovered in 7% of patients on follow up who have undergone endoscopic therapy from oesophageal varices. GV occur more frequently in patient with extrahepatic PH than in those with cirrhosis. GOV1 are the most common GV subtype occurin in 74% of cases. GOV2, IGV1 and IGV2 occurs in 16%, 8% and 2% of cases respectively. Due to their submucosal location, GV may be underestimated on endoscopy. They may also be difficult to distinguish from rugae of the stomach.

P a g e | 27

Figure 5. (a) Oesophageal varices (OVs) and gastro-oesophageal varices type 1 (GOV1) are formed by increased flow and dilatation of the left gastric vein (LGV). GOV1 are extensions of OVs, and therefore have the same venous drainage, and a relatively low risk of variceal haemorrhage (VH). (b) Fundal varices.are formed by the dilatation of the short (SGV) and posterior (PGV) gastric veins, branches of the splenic vein (SV). Isolated fundal varices are classified as IGV1, and if there are associated OVs they are classified as GOV2. IGV2 are isolated ectopic varices, present in the body and antrum of the stomach, and the duodenum. These have a low risk of VH. 47 Anatomy of GVs In the upper gastrointestinal tract the increased portal pressure is transmitted through 2 main venous pathways. First, through the right and left gastric veins, which drain varices around the distal oesophagus and cardia (GOV1) into the portal vein, or when flow is reversed the blood flows cephalad into the azygous system. The second pathway is via the short and posterior gastric veins, which under normal circumstances drain blood from the fundus into the splenic vein, but in PHT the flow often is reversed and blood drains from the spleen toward the stomach into fundal varices(GOV2 and IGV1). IGV2 often are caused by dilation of branches of the gastroepiploic veins. Gastric varices lie deep in the submucosa, under the gastric mucosa, which unlike the oesophageal mucosa is relatively thick. Spontaneous portosystemic splenorenal or gastrorenal shunts commonly develop, and communicate with the left renal vein vein via the inferior phrenic or left suprarenal vein. Such shunts, collectively termed gastrorenal shunts (GRS), are more common in GV (60% to 85%) of cases than OV (17% to 21% of cases). Precisely what determines the predominant collateral pathways that develop in a given individual with PH is unknown. Inpatient with left sided PH or sinistral portal hypertension isolated gastric varices often occur (IGV > IVG2). Here blood usually flows retrogadely through the short

P a g e | 28

and posterior gastric veins and the gastroepiploic veins resulting in the formation of GVs. From these GVs the blood flows hepatopetally through the left and right gartic veins to the portal veins. Risk of bleeding The risk of GV bleeding is not as well studied, however it would appear that the mean portal pressure is less in patients with GVs than those with OVs. This is probably the result of the formation of spontaneous gastrorenal shunts in this group of patients limiting the increases in portal pressure. Patients with GVs also present more commonly with encephalopathy due to these shunts. The two year risk of GV bleeding is 25% similar to OV at 20 40%. Fundal varices have a significantly higher incidence of bleeding (78% for IGV1 and 55% for GOV2), compared to GOV1 and IGV2 (10%). Similar to oesophageal varices, the risk of bleeding for gastric varices are increased in those with advancing Childs class, presence of red spots, increasing size of varices (> 5mm) and in those with fundal varices. Effects of EV eradication on GV The endoscopic eradication of oesophageal varices can have two distinct effects on GVs. On the one hand GVs develop in approximately 9 to 20% of patients who have undergone endoscopic variceal sclerotherapy or ligation. This figure may be as high as 26 to 43% if EUS is used for the evaluation. The alternate effect is endoscopic sclerotherapy for oesophageal varices can lead to the eradication of GVs. This may be the result of caudal seepage of sclerosant toward the GVs. Sclerotherapy for OVs led to the disappearance of 30-60% of GOV1 and 20% of GOV2 within 6 months. As a result of these observations it has been recommended that patients GOV1/2, oesophageal varices should first be treated. If at follow up at 6 months there is persistence of GVs, specific therapy should then be instituted. Primary Prophylaxis 49 Trails are lacking concerning the prophylactic treatment of gastric varices that have not bled. Thus current consensus position is to treat patients with gastric varices with NSBB as primary prophylaxis, despite the lack of specific data on the subject. 25 Since GOV1 see to behave similarly to OVs it would seem reasonable that they be treated alike. N-butyl cyanoacrylate injections endoscopically for primary prophylaxis have been studied alone and compared to NSBB or placebo, showing beneficial outcomes, but concerns over safety, small trial numbers and poor study design have prevented its widespread adoption for this purpose. 47, 48 Management of acute bleeding and the prevention of rebleeding 49 Bleeding should be considered to have arisen from GV if there is (1) active spurt or ooze, (2) adherent clot, or (3) presence of large GV, no EV, and no other source of bleeding evident. Current guidelines recommend endoscopic therapy as first-line treatment for bleeding gastric varices, with the option of TIPSS where endoscopic therapy is not available. Currently, the endoscopic therapeutic options for gastric variceal bleeding include band ligation, tissue adhesives, and thrombin. 49

P a g e | 29

Pharmacotherapy. Data on the ability of somatostatin, vasopressin or their analogues to control acute haemorrhage from gastric varices is limited. Due to the similar pathophysiology and anatomy, GOV1 should be treated as for OVs. Until good quality data is available on the efficacy of these drugs in acute GV bleeding, it would not be unreasonable to continue its use in these circumstances. Balloon tamponade. The Sengstaken-Blakemore or Minnesota tubes are usually not effective in controlling haemorrhage from fundal varices, owing to the small size of the gastric balloon (200ml). The Linton-Nachlas tube with its 600ml volume single gastric balloon seems more effective for this purpose, controlling variceal bleeding in up to 50% of cases. Rebleeding occurs in 20% of patients. Endoscopic therapy 50,51 Sclerotherapy is largely ineffective in standard doses for GV haemorrhage. This is thought to be a result of the higher rate of blood flow in gastric varices, which flushes away the sclerosant. GVs as a whole require more sclerosant than OVs to halt bleeding. Further fundal varices (GOV2 and IGV1) require significantly higher doses than GOV1 (an extension of OVs) to control bleeding. This higher dosing is associated with a higher incidence of side effects. Sclerotherapy may be used for the control of bleeding and secondary prophylaxis for GOV1. It is ineffective therapy for the other subtypes of GVs, due to its low rate of primary haemostasis, ineffectiveness in eradicating these varices and for an unacceptably high rebleeding rate. Endoscopic variceal obturation(EVO) refers to the injection of agents such as n-butyl-2-cyanoacrylate (Histoacryl), isobutyl-2-cyanoacrylate (Bucrylate), or thrombin, which solidify and/or induce thrombosis in the varix, with ultimate sloughing off of the glue cast weeks to months postinjection, resulting in late ulceration. EVO is the modality of choice for controlling acute gastric variceal haemorrhage and for secondary eradiacation of GV. EVO with cyanoacrylate is highly effective in achieving control of GV haemorrhage, with rates approaching 90% compared to 62% with endoscopic sclerotherpy and 40% with endoscopic variceal ligation. Rebleeding rates are reported to be between 22-25%. EVO using thrombin is similarly effective. The long term efficacy in GV eradication is variable. Eradication rates between 50 to 100% have been reported, with an average of 75%. Rebleeding usually occurs in the first year. This occurs usually in patients where the varices persists despite therapy and is reported in 23 and 50% of patients. Complications of tissue adhesives include embolization, with case reports of cerebral stroke, portal vein embolization, splenic infarction, coronary emboli, and a series demonstrating nonfatal pulmonary emboli in 4.6% of cases. Endoscopic variceal ligation(EVL). Despite almost equal efficacy with EVL compared to EBO in controlling acute bleeding from GV, it is not recommended for secondary prophylaxis due to higher rebleeding rates (54%). Although EVL may be successful in eradicating varices in the initial period, recurrences are almost invariable within 2 years of therapy. A combination of banding and sclerotherapy may be as effective EBO. Further studies are required. Radiological techniques
49,57

P a g e | 30

Balloon occluded retrograde transvenous obliteration (B-RTO). These are reserved for the management of GV with a know gastrorenal shunt, which occurs in 85% of patients with GV. The aim is to abolish the feeding shunts with a sclerosant thus achieving variceal obliteration. It is used mainly for the control of acute haemorrhage from GV and the eradication of GV as secondary prophylaxis. It has been shown to be effective in the control of acute haemorrhage with rates of up to 100% and has low rebleeding rates, close to 0% at 2 year follow up. There is concern however, regarding the appearance or worsening of oesophageal varices in up to 66%, and the large amount of sclerosant used with the potential to result in systemic embolization, hepatic and renal toxicity. Complications related to its use include haemoglobinuria, abdominal pain, pleural effusion and transient worsening of liver functions. Further studies are therefore required if widespread adoption of this technique is to be recommended. Balloon occluded endoscopic injection therapy (BO-EIS). This involves the occlusion of all veins using coils except the supplying vein of the fundal varix. The supplying vein is then occluded with a balloon tipped catheter. The varix is then injected endosopically using a sclerosant. The efficacy is close to B-RTO. The concern with the above techniques remains that in the process of obliterating spontaneous shunts and collateral vessels, where is all this blood diverted to? These effects can be seen by the alarming emergence of OVs in patients treated using these methods. Thus long term outcomes need further study. Transjugular intrahepatic portosystemic shunts (TIPSS). The control of acute GV haemorrhage with haemostasis achieved up to 90%. The rebleeding rates are between 11 and 31%. These shunts serve to decrease the portal pressure gradient leading to the cessation of bleeding in patients with GV haemorrhage. When comparing TIPS and cyanoacrylate injection therapy, both achieved similar results in acute haemorrhage control, with reduced rebleeding rates in those treated with TIPSS (35% vs. 20%). Therapy with TIPSS is however more morbid by virtue of a higher rate of encephalopathy (3-18%) and is more expensive to perform. There is no mortality benefit using either therapy.49 It should be considered in patients in whom cyanoacrylate injection therapy has failed as a salvage procedure. Surgery Although shunt surgery is associated with a significant reduction in the rate of variceal bleeding, it is associated with a significantly increased incidence of hepatic encephalopathy and mortality. Surgical shunts such as the splenorenal shunts have a rebleeding rate of 5-40% which is a reasonable option for well compensated patients. Splenectomy or splenic embolisation may be considered for those with left sided PH. Summary The optimal management of gastric variceal bleeding remains controversial, owing to a lack of controlled studies. It is important to emphasize that GOV1 can be treated as oesophageal varices. With careful attention to technique, cyanoacrylate is effective for the treatment of gastric varices and more economical than TIPSS. It is also associated with a small

P a g e | 31

but clear risk of mortality caused by embolization, despite expert hands. None the less it remains the de facto choice to the control of GV acute haemorrhage and its secondary prophylaxis. The use of TIPSS is limited to salvage therapy, but there is little doubt it is effective in this regard although mortality is unaffected and encephalopathy rates are high compared with endoscopic therapies with tissue adhesives. B-RTO is potentially useful in patients with gastrorenal shunts to eradicate gastric varices once haemostasis has been achieved. It may also have a use as secondary prophylaxis in situations where TIPSS cannot be used such as portal vein thrombosis or patients with severe encephalopathy. It is limited by its availability, lack of technical expertise and controlled trials outside Asia to duplicate its proposed success. Sclerotherapy is effective for the control of GV bleeding, but its use is marred by a high rebleeding rate and incidence of local complications. Surgical shunts may have a role in compensated patients, and splenectomy offers

a curative solution for patients with isolated splenic vein thrombosis. Figure 6. Algorithm for the management of acute gastric variceal bleeding Portal hypertensive gastropathy (PHG) Portal hypertensive gastropathy is a macroscopic finding of a characteristic mosaic like pattern of the gastric mucosa (mild PHG), red-point lesions, cherry red spots, and/or black-brown spots (severe PHG). These lesions, however, are non-specific, i.e. can occur in the absence of portal hypertension. In PHG there is

P a g e | 32

marked dilatation of the vasculature of the gastric mucosa and submucosa, together with an increased blood flow and tendency to decreased acid secretion. PHG is unrelated to Helicobacter pylori infection. The overall prevalence of PHG in patients with cirrhosis strongly correlates with the severity of the diseaseand ranges between 11% and 80%. The incidence of acute bleeding is low (less than 3% at 3 years) with a mortality of 12.5%, while the incidence of chronic bleeding is 1015% at 3 years. In acute bleeding from PHG beta-adrenergic blockers, somatostatin, octreotide, vassopressin, terlipressin and estrogens have been proposed based on their ability to decrease gastric perfusion in this condition. However, only somatostatin has been evaluated, in an uncontrolled study, in patients bleeding from PHG. Haemostasis was achieved in all patients. Nonselective beta-blockers effectively decrease chronic bleeding from PHG. Surgery in Portal Hypertension The use of portosystemic shunt surgery and devascularisation procedures to decompress portal hypertension and gastroesophageal varices has declined during the past decade in favor of alternative therapies. Orthotopic liver transplantation (OLT) is the only therapy which significantly prolongs long-term survival in patients with cirrhosis and variceal bleeding. However, about 2030% of patients with variceal bleeding refractory to medical or endoscopic treatment are in a Childs A class and may not require transplantation for many years. Others may not be candidates for OLT. 52 There are three surgical methods used in the management of portal hypertension: decompressive shunts devascularization procedures liver transplantation (see seminar on transplantation) Each of these has a specific role in the management of portal hypertension and are usually complementary, not competitive. DECOMPRESSIVE SURGICAL SHUNTS 2*, 53 Surgical shunts fall into three distinct groups: Total portal systemic shunts which decompress all portal hypertension. Partial portal systemic shunts which reduce portal pressure to 12 mmHg and allow some continuing prograde portal flow. Selective shunts which decompress gastroesophageal varices, but maintain portal hypertension and portal flow to the liver. Total portosystemic shunts These include any shunt greater than 10 mm in diameter between the portal vein or one of its main tributaries and the inferior vena cava or one of its main tributaries. All non selective shunts result in the elimination of hepatic portal perfusion, resulting in increased rates of post operative encephalopathy and postoperative liver failure are observed. Examples of nonselective shunts include the end-to-side portacaval shunt, the side-to-side portacaval shunt, the conventional splenorenal shunt (anastomosis of the proximal end of the splenic vein to the left renal vein in combination with splenectomy), and various largebore interposition (>10mm) shunts using a synthetic graft or autogenous vein to create a portacaval, mesocaval, or mesorenal interposition.

P a g e | 33

The Eck fistula is the classic end to side portocaval anastamosis. This historic shunt was first performed on eight dogs by Nikoli Eck in 1877. The portal vein is transacted close to the liver, the hepatic end is ligated and the splanchnic end is connected to the IVC in an end to side configuration. This results in total derivation of the portal blood flow into the systemic venous system. It effectively decompresses the splanhnic hypertension and controls variceal bleeding, but leaves high pressures in the hepatic sinusoids and thus does not relieve ascites. Its use is largely historic. Figure 7. A total side-to-side portacaval shunt with direct anastomosis (>10 mm) between the portal vein and inferior vena cava.

Side to side portocaval shunts are generally achieved by anastamosing the portal vein with the inferior vena cava in a side to side manner. Side to side shunts may also be achieved with a mesocaval or mesorenal shunt greater than 10mm in diameter. Both the portal vein and obstructed hepatic sinusoids decompress through this outflow tract, thus relieving both portal hypertension and ascites. Its modern role is limited to perhaps decompression of patients with intractable variceal bleeding with ascites. Its success in doing so is around 90% but due to progressive liver failure the encephalopathy rate is high at around 40-50%. Other patients that may benefit from this procedure are those with Budd-Chiari syndrome with ongoing hepatocyte necrosis by decompressing the sinusoids, and thus, halting further necrosis. Selective shunts Due to the preservation of portal blood flow to the liver, hepatic perfusion, although decreased is maintained. Thus, in contrast to nonselective shunts, these procedures have lower rates of post operative encephalopathy and liver failure but are less effective at relieving ascites. Examples of selective portosystemic shunts are the left gastric caval shunt(Inokuchi shunt) and the more widely used distal splenorenal shunt (Warren shunt). The distal splenorenal shunt has been the most commonly performed surgical shunt procedure. This decompresses the gastrooesophageal varices through the short gastric veins, the spleen, and the splenic vein to the left renal vein. This is the most used decompressive operation for refractory variceal bleeding currently, and is used primarily for patients who present with refractory bleeding and still have good liver function.

P a g e | 34

A high rate of portal vein thrombosis (10%) is associated with this procedure. Portalsplenic disconnection has been added to this procedure to decreases the formation of collaterals between the portal and splenic veins noted to occur soon after the introduction of this procedure (usually after 6mo). Distal splenorenal shunt gives a greater than 90% control of variceal bleeding and excellent initial maintenance of portal perfusion. In patients with non-alcoholic liver disease portal perfusion is maintained long-term. Contraindications for the distal splenorenal shunt include previous splenectomy, medically intractable ascites and a splenic vein with a diameter of less than 8 mm. Additionally good outcomes are the result of good patient selection, and those with Childs class A or B having the best outcomes. When DSRS was compared with nonselective shunts, the rate of postoperative encephalopathy was lower (15 vs.36%).

Figure 8. A. Distal splenorenal shunt anastomoses the divided splenic vein to the left renal vein. Portal hypertension is maintained in the superior mesenteric and portal veins. B. Partial shunt as an interposition portacaval graft of PTFE (8 mm).

Partial portosystemic shunts These shunts reduce the size of the anastomosis of a side-to-side shunt to 8 mm. The aim of partial shunts is to incompletely decompress the portal venous system to maintain hepatic portal perfusion. The preservation of hepatic perfusion allows detoxification and thus decreased the rate of encephalopathy when compared to nonselective shunts, yet achieving variceal decompression. Examples of partial shunts include small-diameter side-to-side portacaval shunts and various smallbore interposition shunts (interposition diameter 8 to 10 mm), using a synthetic graft or autogenous vein to create a portocaval, mesocaval, or mesorenal interposition. The ligation of portal collateral vessels is an integral part of these procedures as it directs greater portal flow to the liver. These shunts in general have decreased rates of encephalopathy when compared to nonselective shunts. The problems associated with these small calibre shunts is the increased risk of graft thrombosis, which is ameliorated attention to detail, and by increasing the size of the anastamosis by bevelling the ends. The rates of bleeding control are

P a g e | 35

excellent at 90% and it has a more favourable encephalopathy rate than the larger bore, 16mm, shunts. Contraindications to the partial portocaval shunt include previous operation in the right upper abdominal quadrant, Child C cirrhosis, and portal vein thrombosis. Devascularisation Procedures
54

Figure 9. Sugiura procedure includes: 1) splenectomy; 2) devascularisation of 8 to 10 cm of the esophagus; 3) transection and end-to-end anastomosis of the lower esophagus 4 to 5 cm above the gastroesophageal junction (EEA-stapler, introduced through anterior gastrostomy); 4) devascularization of the lesser and greater curvatures of the stomach; 5) pyloroplasty; 6) feeding jejunostomy (not shown). During the Sugiura procedure the left gastric vein (not shown) is carefully preserved to allow blood drainage into the azygous system.

Devascularization procedures have the components of splenectomy, gastric and oesophageal devascularization, and in some instances oesophageal transaction while maintaining the plexus of collaterals that connect the coronary vein to the azygous system. In particular the left gastric vein must be preserved during the gastric devascularization to allow drainage into the azygos vein. Practice has shifted from original description by Sugiura, which consisted of a two stage thoraco-abdominal approach for the oesophageal devascularisation, with many centres now performing a single abdominal approach. Although the initial results with this procedure were exceptional in terms of mortality, morbidity, and recurrent bleeding, surgeons outside of Japan have had difficulty reproducing these results. This may be due to the less stringent adherence to the procedure initially described, stressing the sparing of the left gastric vein to allow hepatofugal flow and thus variceal decompression. Also, the initial series Sugiura and Futagawa contained few patients with Childs-Pugh class C disease, and may represent a selection bias with their results. The procedure is associated with a low rate of rebleeding, hepatic encephalopathy and good long term results, and thus should be considered as a viable surgical technique in the treatment of PH. Patients with preserved liver function who are not candidates for transhepatic porto-systemic shunt (TIPS), orthotopic liver transplantation (OLT), or a distal

P a g e | 36

splenorenal shunt (DSRS) should be evaluated for the modified Sugiura procedure. 54 The role of oesophageal transection is more debatable. Particularly in patients who have undergone extensive and repeated sclerotherapy in whom the gastroesophageal junction is thickened, the ability to perform a satisfactory transection is limited. For these patients, it is probably unwise to add this component with the risk of an oesophageal leak.

Figure 10 . Simplified algorithm for the management of patients with variceal bleeding refractory to medical treatment.

The role of surgery55 While the advent of liver transplantation and TIPS may have decreased the use of surgical shunts in the management of portal hypertension and its complications, patients with either Childs A cirrhosis or noncirrhotic portal hypertension can expect good long term liver function and survival after receiving a surgical shunt. A very small proportion of these patients will eventually need liver transplantation. Surgery as primary prophylaxis Surgery should not be considered for primary prophylais for portal hypertension due to cirrhosis. Currently available therapy is as effective, and has a better risk benefit ratio (complications of surgery and risk of encephalopathy with surgery), although some studies from Japan in the early 1990s showing good results, these have not been reproduced. Surgery for acute variceal haemorrhage 56 For the acute bleeding episode, it is also generally accepted that better results can be obtained with pharmacotherapy and endoscopic therapy. Extreme cases with good liver function

P a g e | 37

that are refractory to other options (including second line treatment choices) should be considered for surgical treatment. Failure of medical and endoscopic therapy When compared to TIPPS, surgical shunts have a lower risk of rebleeding, requires less reintervention, with similar encephalopathy rates. Mortality in Chid-Pugh class C patients was higher in the surgically treated patients, whereas those with advanced liver disease the mortality rate was equally high (73-74%). Cost differences are negligible. Shunt surgery should be considered in patients with Childs-Pugh class A or B, ahead of TIPPS, in those who are refractory to first line therapies. In patients who require emergent surgery for variceal haemorrhage and are candidates for liver transplantation, an interposition mesocaval or mesorenal shunt is advocated. These can simply be ligated at the time of surgery. Traditional total portosystemic shunts which require hepatic hilar dissection may compromise future transplantation and thus are avoided. Emergent surgery may also be the only option for patients in whom initial bleeding control fails, as TIPSS may not ameliorate portal pressures in these patients. These include patients with BCS and those with portomesenteric thrombosis. Surgery for Secondary prophylaxis ChildPugh class A and good risk Childs B cases can be considered for surgical treatment. In many centers only cases that fail medical and endoscopic therapy are treated by means of surgery. It may be reasonable to consider surgery as first line treatment for low risk, Child-Pugh A patients, because the other options, as stated above, have a higher rebleeding risk and require more frequent follow up and reintervention. In poor-risk Child's B patients and Child's C patients, liver transplant offers the only realistic surgical option. Patients in whom long term non-compliance or geographical inaccessibility to medical follow up is anticipated may also be considered. TIPSS is generally used in these cases as a bridge to transplantation. References
1. Bosch J, Berzigotti A, Garcia-Pagan JC, Abraldes JG. The management of portal hypertension: rational basis, available treatments and future options. Journal of hepatology.2008;48 Suppl 1:S68-92. 2. Henderson JM. Surgical treatment of portal hypertension. Baillire's best practice & research. Clinical gastroenterology.2000;14(6):911-25. 3. Kim M, Mitchell DG, Ito K. Portosystemic collaterals of the upper abdomen: review of anatomy and demonstration on MR imaging. Abdominal Imaging. 2000;25(5):462-470. 4. Choi S, Chung TW, Seo JJ, et al. Multi Detector Row CT Portal Venography in the Evaluation of eral Vessels in Liver. Radiographics. 2002:1053-1061. 5. Toubia N, Sanyal AJ. Portal hypertension and variceal hemorrhage. The Medical clinics of North America. 2008;92(3):551-74, viii. 6. Wolff M, Hirner A. Current state of portosystemic shunt surgery. Langenbeck's archives of surgery / Deutsche Gesellschaft fr Chirurgie. 2003;388(3):141-9. 7. Gauthier F. Recent concepts regarding extra-hepatic portal hypertension. Seminars in pediatric surgery. 2005;14(4):216-25. 8. Gatta A, Bolognesi M, Merkel C. Vasoactive factors and hemodynamic mechanisms in the pathophysiology of portal hypertension in cirrhosis. Molecular aspects of medicine. 2008;29(1-2):119-29. 9. Rodrguez-vilarrupla A, Fernndez M, Bosch J, Garca-pagn JC. Artemisa Current concepts on the pathophysiology of portal hypertension. Gastroenterology. 2007;6(December 2006):28-36.

P a g e | 38 10. Bosch J, Abraldes JG, Fernndez M, Garca-Pagn JC. Hepatic endothelial dysfunction and abnormal angiogenesis: new targets in the treatment of portal hypertension. Journal of hepatology. 2010;53(3):558-67. 11. de Franchis RD. Stellate cells and the "reversible component" of portal hypertension. Digest liver dis. 2000;32:194-7. 12. Rockey DC. Hepatic fibrosis, stellate cells, and portal hypertension. Clinics in liver disease. 2006;10(3):459-79, vii-viii. 13. Gupta T K; L Chen; R J Groszman. Pathophysiology of portal hypertension. Baillieres Clinical Gastroenterology. Vol 11, No.2, June 1997 14. Jensen and Groszmann Complications of liver disease : pathophysiology and management. In Kaplowitz N. Liver and Biliary Diseases, p502. Baltimore: Williams & Wilkins. 1992 15. Groszmann R, Vorobioff JD, Gao H. Measurement of portal pressure: when, how, and why to do it. Clinics in liver disease. 2006;10(3):499-512, viii. 16. Burroughs AK, Thalheimer U. Hepatic venous pressure gradient in 2010: optimal measurement is key. Hepatology (Baltimore, Md.). 2010;51(6):1894-6. 17. Mehta G, Abraldes JG, Bosch J. Developments and controversies in the management of oesophageal and gastric varices. Gut. 2010;59(6):701-5. 18. Franchis RD, Era AD, Iannuzzi F. Diagnosis and treatment of portal hypertension. Digestive and Liver Disease. 2004;36:787-798. 19. Coelho-prabhu N, Kamath PS. Current Staging and Diagnosis of Gastroesophageal Varices. Clin Liver Dis. 2010;14:195-208. 20. Lisi SD, Peralta S, Arini A, Simone F, Crax A. Oesophagogastroduodenoscopy in patients with cirrhosis: Extending the range of detection beyond portal hypertension. Digestive and Liver Disease. 2010. 21. Schreibman I, Meitz K, Kunselman AR, et al. Defining the Threshold: New Data on the Ability of Capsule Endoscopy to Discriminate the Size of Esophageal Varices. Digestive diseases and sciences. 2010. 22. Gins A, Fernndez-Esparrach G. Endoscopic ultrasonography for the evaluation of portal hypertension. Clinics in liver disease. 2010;14(2):221-9. 23. Kovalak M, Lake J, Mattek N, et al. Endoscopic screening for varices in cirrhotic patients: data from a national endoscopic database. Gastrointestinal endoscopy. 2007;65(1):82-8. 24. de Franchis R, Dell'Era a, Iannuzzi F. Diagnosis and treatment of portal hypertension. Digestive and liver disease : official journal of the Italian Society of Gastroenterology and the Italian Association for the Study of the Liver. 2004;36(12):787-98. 25. de Franchis R. Revising consensus in portal hypertension: report of the Baveno V consensus workshop on methodology of diagnosis and therapy in portal hypertension. Journal of hepatology. 2010;53(4):762-8. 26. Sharma SK, Aggarwal R. Prediction of large esophageal varices in patients with cirrhosis of the liver using clinical, laboratory and imaging parameters. Journal of gastroenterology and hepatology. 2007;22(11):1909-15. 27. Groszman RJ, Garcia-Tsao G, Bosch J. Beta-blockers to Prevent Gastroesophageal Varices in Patients with Cirrhosis. N Engl J Med 2005;353:2254-61. 28. Senzolo M, Cholongitas E, Burra P, et al. beta-Blockers protect against spontaneous bacterial peritonitis in cirrhotic patients: a meta-analysis. Liver international : official journal of the International Association for the Study of the Liver. 2009;29(8):1189-93. 29. Tripathi D, Ferguson JW, Kochar N, et al. Randomized controlled trial of carvedilol versus variceal band ligation for the prevention of the first variceal bleed. Hepatology (Baltimore, Md.). 2009;50(3):825-33. 30. Lpez-mndez E, Uribe M. Beta blockers in portal hypertension. Are they really a good option ? Annals of Hepatology. 2006;(5(2):86-91. 31. Talwalkar Ja, Kamath PS. An evidence-based medicine approach to beta-blocker therapy in patients with cirrhosis. The American journal of medicine. 2004;116(11):759-66. 32. Sass Da, Chopra KB. Portal hypertension and variceal hemorrhage. The Medical clinics of North America. 2009;93(4):837-53, vii-viii. 33. Segawa M, Sakaida I. Diagnosis and treatment of portal hypertension. Hepatology research : the official journal of the Japan Society of Hepatology. 2009;39(10):1039-43. 34. Garcia-tsao G, Bosch J. Management of Varices and Variceal Hemorrhage in Cirrhosis. New England Journal of Medicine. 2010;(362):823-832.

P a g e | 39 35. Garcia-Pagan JC, De Gottardi a, Bosch J. Review article: the modern management of portal hypertension--primary and secondary prophylaxis of variceal bleeding in cirrhotic patients. Alimentary pharmacology & therapeutics. 2008;28(2):178-86. 36. Gow PJ. Modern management of oesophageal varices. Postgraduate Medical Journal. 2001;77(904):7581. 37. Khuroo MS, Khuroo NS, Farahat KL, Khuroo YS, Sofi AA, Dahab ST. Meta-analysis: endoscopic variceal ligation for primary prophylaxis of oesophageal variceal bleeding. Aliment Pharmacol Ther 2005;21:347-361. 38. Garcia-Pagan JC, Bosch J. Endoscopic band ligation in the treatment of portal hypertension. Nat Clin Pract Gastroenterol Hepatol 2005;2:526-535. 39. Crdenas A. Management of acute variceal bleeding: emphasis on endoscopic therapy. Clinics in liver disease. 2010;14(2):251-62. 40. Longacre AV, Garcia-Tsao G. A commonsense approach to esophageal varices. Clinics in liver disease. 2006;10(3):613-25 41. Bendtsen F, Krag A, Mller S. Treatment of acute variceal bleeding. Digestive and Liver Disease. 2008;40:328-336. 42. Sass Da, Chopra KB. Portal hypertension and variceal hemorrhage. The Medical clinics of North America. 2009;93(4):837-53, vii-viii. 43. Lo G. Management of Acute Esophageal Variceal Hemorrhage. The Kaohsiung Journal of Medical Sciences. 2010;26(2):55-67. 44. Garcia-tsao G, Bosch J. Management of Varices and Variceal Hemorrhage in Cirrhosis. New England Journal of Medicine. 2010;(362):823-832. 45. Heller J, Sauerbruch T. Prevention of recurrent haemorrhage. Best Practice & Research Clinical Gastroenterology. 2007;21(1):43-53. 46. Gastric Varices in Patients With Portal Hypertension Evaluation With Multidetector Row CT. K Zhu, MD, X Meng et al. J Clin Gastroenterol 2010;44:e108e115 47. Gautam Mehta, Juan G Abraldes, Jaime Bosch. Developments and controversies in the management of oesophageal and gastric varices. Gut 2010 59: 701-705 48. Yun Jung Chang Jong-Jae Park. Long-Term Outcomes of Prophylactic Endoscopic Histoacryl Injection for Gastric Varices with a High Risk of Bleeding. Dig Dis Sci (2010) 55:23912397 49. Tripathi D. Therapies for bleeding gastric varices: is the fog starting to clear? Gastrointestinal endoscopy. 2009;70(5):888-91. 50. RYAN BM, STOCKBRUGGER RW, RYAN J. A Pathophysiologic, Gastroenterologic, and Radiologic Approach to the Management of Gastric Varices. GASTROENTEROLOGY 2004;126:11751189 51. D. TRIPATHI et al. Review article: recent advances in the management of bleeding gastric varices. Aliment Pharmacol Ther 24, 117 52. Wolff M, Hirner A. Current state of portosystemic shunt surgery. Langenbeck's archives of surgery / Deutsche Gesellschaft fr Chirurgie. 2003;388(3):141-9. 53. Klempnaue J, Schrem H. Review: surgical shunts and encephalopathy. Metabolic brain disease. 2001;16(1-2):21-5. 54. Selzner M, Tuttle-Newhall JE, Dahm F, Suhocki P, Clavien Pa. Current indication of a modified Sugiura procedure in the management of variceal bleeding. Journal of the American College of Surgeons. 2001;193(2):166-73. 55. Knechtle SJ, D'Alessandro A, Armbrust M. Surgical portosystemic shunts for treatment of portal hypertensive bleeding: Outcome and effect on liver function. Surgery. 1999;126(4):0708-0713. 56. Costa G, Cruz RJ, Abu-Elmagd KM. Surgical shunt versus TIPS for treatment of variceal hemorrhage in the current era of liver and multivisceral transplantation. The Surgical clinics of North America. 2010;90(4):891-905. 57. Yoshida H, Mamada Y, Taniai N, Tajiri T. New methods for the management of gastric varices. World journal of gastroenterology : WJG. 2006;12(37):5926-31.

P a g e | 40

Balloon-Occluded Retrograde Transvenous Obliteration of Gastric Varices Comment By Dr MC Mayet Gastric fundal varices with hemorrhage are associated with a higher mortality rate than esophageal variceal bleeding, and therefore, optimal management of gastric varices requires a multidisciplinary approach. Generally, various treatment modalities such as pharmacotherapy, balloon tamponade, endoscopic procedures, endovascular treatment, and surgery have been performed. Transjugular intrahepatic portosystemic shunt (TIPS) has been widely used to treat bleeding esophageal and gastric varices. Although studies have demonstrated successful initial control of bleeding in both esophageal and gastric varices, rebleeding rates post-TIPS have been higher with gastric varices. In fact, gastric varices can still bleed despite portal pressure gradients below 12 mm Hg, and TIPS is not always effective in such patients with low initial portal pressure gradients. In addition, TIPS may not be tolerated in patients with encephalopathy or poor liver function. Balloon-occluded retrograde transvenous obliteration (BRTO) is an endovascular technique that was developed in Japan as a therapeutic adjunct or alternative to TIPS in the management of gastric varices. It is also an effective therapy for sclerosis of de novo portosystemic shunts complicated by hepatic encephalopathy. A BRTO procedure involves occlusion of outflow veins of the portosystemic shunt, such as a gastrorenal shunt, using an occlusion balloon followed by the injection of a sclerosing agent directly into the varix endovascularly. Stagnation of the sclerosant within the varix or shunt without reflux into either the portal or systemic vasculature is critical to the procedure because this can result in serious complications. To avoid these events, occlusion balloons are strategically placed to modulate the flow within the varix and/or shunt. Additionally, microcatheters and embolization coils are adjunctive tools that are used to administer the sclerosant in high concentration within the varix and prevent reflux to nontarget sites. The complexity of gastric varices cannot be overstated. The following is a review of the relevant anatomy and the integral relationship between the anatomy and the therapeutic approach, describing the technical aspects of the procedures. ANATOMIC CLASSIFICATION SYSTEM FOR VARICES TYPE A Defined by the presence of a single draining vein TREATMENT STEPS 1. Access of the right femoral vein using standard angiographic technique. 2. Catheterization of the gastrorenal shunt via the left renal vein is typically accomplished using a catheter with a mounted occlusion balloon. 3. Balloon occlusion venography performed to confirm type of varix.

P a g e | 41

4. Infusion of sclerosant until adequate to fill the full extent of the varix with minimal filling of the afferent vein/portal vasculature. TYPE B Defined by the presence of collateral draining veins. TREATMENT STEPS Standard BRTO technique as described for type A. Selective use of a microcatheter or advancement of the balloon catheter further into the draining vein can be performed to optimize sclerosant delivery. When high-flow collateral veins are present, these must first be occluded with microcoils delivered sequentially to each collateral vein.

TYPE C Defined by the presence of both a gastrocaval and gastrorenal shunt of asymmetric sizes/flow. TREATMENT STEPS Standard BRTO technique as described for type A. Embolize the smaller shunt with coils, then standard BRTO is performed. For larger shunts, occlusion balloons are manipulated into the outflow of both the gastrocaval and gastrorenal shunts. Type D has multiple small draining veins and no shunt.

TECHNIQUE BRTO has become the treatment of choice for gastric varices in many hospitals in Asia and is becoming a favorable option in the United States as well. The complexity of the variceal anatomy mandates a customized approach. The procedural main steps are: 1. Access of the right femoral or internal jugular vein using standard angiographic technique and placement of a 6- to 12-F sheath. Most of the reported cases in the literature have been described using the right femoral vein approach. Certain institutions have adopted the jugular vein approach exclusively. One approach is to review the preprocedure computed tomographic angiogram (CTA) or magnetic resonance image to decide the approach that provides the best angle for selecting the shunt. 2. Catheterization of the gastrorenal shunt via the left renal vein is typically accomplished using a catheter with a mounted occlusion balloon. Reverse-shaped balloon catheters are available in Asia and provide easier and stable

P a g e | 42

access into the shunt; however, such catheters are not readily available in the United States. Access into the shunt can be achieved by selective catheterization of the left renal vein using a forward-seeking catheter such as a Cobra-shaped catheter, which is then exchanged for an angled-tip catheter and can be used to select the shunt. A 0.035-inch stiff wire is then advanced into the shunt, followed by a standard occlusion balloon catheter (8.532 mm). The access sheath is usually positioned in the inferior vena cava or renal vein. 3. Balloon occlusion venography is performed to define the anatomy and type of varix. 4. Infusion of a sclerosant follows. The goal is filling of the full extent of the varix with the embolization endpoint being minimal filling of the afferent vein/portal vasculature. The injection of a sclerosing agent can be performed with or without the use of a microcatheter for more selective injection. The authors suggest advancing a microcatheter as deep as possible into the varix and injecting the sclerosant through the microcatheter. This provides the benefit of selective delivery of the sclerosant into the varix and minimizing the volume used, as well as minimizing the risk of balloon rupture when the sclerosant comes in close proximity to the balloon. In addition, leaking collateral veins, such as the inferior phrenic or paravertebral veins, are commonly present and prevent full opacification of gastric varices. These veins are occluded using coils or gelfoam pledgets through a selectively catheterized microcatheter. This would modulate flow in the veins in an effort to concentrate the sclerosant at the varix and minimize nontarget distribution in the portal or systemic vasculature. If the patient has more than one shunt (eg, gastrorenal and gastropericardiophrenic shunts), an additional occlusion balloon, usually from the jugular approach, is usually needed to occlude the second shunt. Ethanolamine oleate iopamidol (EOI), sodium tetradecyl sulfate (STS) as foam, and n-butyl-2cyanoacrylate are examples of agents that have been frequently used in therapy. EOI is the original agent used for this procedure and is the agent of choice in Asia. A 5% EOI solution consists of a mixture of 10% EOI and the same dose of a contrast medium. EOI causes hemolysis in the blood vessels. As a result, free hemoglobin is released, which may cause renal tubular disturbances and acute renal failure. To prevent renal insufficiency, 4,000 units of haptoglobin are administered intravenously and combine with free hemoglobin. Other complications associated with the use of EOI, including cardiogenic shock, pulmonary edema, and disseminated intravascular coagulation, have also been reported. Therefore, if possible, < 40 mL of EOI should be used in each procedure. Recently, other agents have been used in the United States, namely STS as foam. STS (3%) is mixed with lipiodol and gas (air or CO2) in a 2:1:3 ratio.17 The foam has the potential advantage of minimizing the dose of the sclerosant and may provide better contact with the variceal wall. Experience with this agent is limited, and no long-term results are available. Follow-up with CTA, magnetic resonance imaging, or endoscopic ultrasound before discharge is obtained to confirm obliteration of the varices and for the additional possible need for further interventions. Periodic follow-up with endoscopy is recommended to evaluate for recurrence or development of new gastric or esophageal varices.

P a g e | 43

OUTCOMES Technical success has been reported in 84% to 100% of cases. The complete obliteration rate ranges from 86% to 100%. Patients with partial obliteration may undergo subsequent BRTO procedures to obtain full obliteration. Failure to obliterate the varix can be attributed to several factors:
1. The gastrorenal shunt may not be fully occluded with the balloon catheter due to the large shunt

size and rapid flow through the shunt. Some authors suggest partial splenic embolization to decrease the flow in the shunt and reattempting the BRTO procedure 2 weeks later. 2. Lack of a defined gastrorenal or gastrocaval shunt or a very tortuous shunt, which may prevent catheterization and balloon occlusion. 3. Extensive leaking collateral veins that prevent full opacification of the varix. Selective catheterization of these collaterals may not be possible. Slow nonselective injection of gelfoam pledgets may help decrease flow into these collaterals and further opacify the varix. Advancement of a microcatheter deep into the varix may bypass these leaking collaterals; however, this can be difficult to perform without the varix being opacified. 4. Inadequate volume of sclerosant was administered. As stated previously, the endpoint should be filling of the entire varix and visualization of the afferent vein. Recurrence of gastric varices is rare, and rebleeding from gastric varices after technically successful procedures is uncommon (0%9%). However, worsening of esophageal varices has been reported in up to 68% of patients in one series, with a median followup of 15 months, and in 39% and 52% at 3 and 5 years, respectively, in another series. Bleeding from esophageal varices has been reported in 10% to 32% of patients. BRTO has been used to treat symptoms of encephalopathy. The reported efficacy for the treatment of gastric varices and hepatic encephalopathy is 87% to 100%, with a relapse rate of 0% to 10%. The overall cumulative survival rates were 90%, 75%, 68%, and 55% at 1, 3, 5, and 7 years after BRTO, respectively.

CONCLUSION BRTO is a safe, effective, and established technique to treat and prevent gastric variceal bleeding and improve hepatic encephalopathy symptoms. Knowledge of the anatomy and technical aspects of this procedure are crucial in obtaining the desired results.

P a g e | 44

ROLE OF SURGERY IN PORTAL HYPERTENSION Comment by Dr I Seedat With advances in pharmacologic & endoscopic management of gastroesophageal variceal bleeding, the mainstay of management has largely shifted from surgical to medical. However, medical failure does occur with an incidence of 20%, which requires radiological or surgical intervention to decompress the portal system. Candidates for surgical intervention should be carefully evaluated. Those with recurrent refractory bleeding with good liver function (Childs A or Childs B with 7 or 8 points) are more likely to survive & benefit from an operative procedure. Primary Prophylaxis Surgery has not shown to be effective in the prophylaxis of initial variceal bleed. Pharmacological therapy remains the mainstay as only 30% of patients are at risk of a life-threatening bleed. Acute Variceal Bleeds Surgical therapy in acute variceal bleeding is exceedingly limited. Here endoscopy proves effective with control of 90% of bleeding patients. Surgery may be useful in patients with severe acute variceal bleeding not responding to medical or endoscopic methods. However, TIPS also offers an attractive option as it is less invasive & provides acute decompression for the control of bleeding. This makes TIPS more favourable in the unstable patient with poor renal function. Devascularization strategies have also laid a claim in this setting, making open surgery more uncommon. Only onestudy by Orloff et al in1995 supports surgical intervention with a side-to-side porto-caval shunt in alcoholic patients with acute bleeding.

P a g e | 45

Prevention of Rebleeding Surgery may be indicated if first line modalities have failed to prevent rebleeding. A clear definition of what constitutes failure of first line therapy is lacking, but recurrence of bleeding or persistent high risk varices should prompt the need for surgical intervention. Shunt surgery is an excellent option in semi-emergent patients with good liver function. Some authors favour shunt surgery in patients with acute Budd-Chiari syndrome with ongoing hepatocyte necrosis, as it decompresses the high pressure sinusoids & halts this injury. Surgery as a bridge to transplantation In patients awaiting transplant with recurrent bleeding, the success of shunt surgery depends on the disease severity & comorbidities. TIPS is a viable option, but patients run the risk of venous thrombosis. Shunt surgery may be useful especially in patients with failed TIPS, with porto-mesenteric thrombosis or Budd-Chiari syndrome. Here, an interposition mesocaval or mesorenal shunt is favoured as it can easily be ligated during the transplant operation. References Henderson JM. Surgical treatment of portal hypertension. Clinical Gastroenterology 2000;14.6:911-925
1.

Guilherme C, Ruy J, Kareem AM. Surgical shunt versus TIPS for treatment of variceal hemorrhage in the current era of liver & multivisceral transplantation. Surg. Clin N Am 2010; 90:891-905.
2.

Hector O, Miguel AM et al. current role of surgery for the treatment of portal hypertension. Annals of Hepatology 2002; 1.4:175-178
3.

Orloff MJ, Orloff MS. Three decades of experience with emergency portocaval shunt for acutely bleeding esophageal varices in 400 patients withcirrhosis of the liver. Journ of Am college of Surg 1995; 180: 257-272
4.

Orloff MJ, Girad B. long term results of treatmentof Budd-Chiari syndrome by side-toside portocaval shunt. Surgery, Gynaecology & Obstetrics 1989; 168:33-41
5.

P a g e | 46