Portal Hypertension: Complications and Medical Management from a Gastroenterologist’s Perspective
Physiology
The portal vein is derived from the splenic vein and large mesenteric veins. As such, it carries blood draining from the splanchnic circulation to the liver. Indeed, under normal circumstances, the portal vein (rather than the hepatic artery) supplies the overwhelming majority (75%) of the nutrient blood flow to the liver. The liver, in turn, is the main site of resistance to portal blood flow and acts as a distensible vascular network with a low resistance. Thus, portal venous pressure is normally low (5 to 10 mm Hg). Portal hypertension occurs when portal venous pressure exceeds these values.
Changes in the pressure within the portal vein are regulated by Ohm’s law, which states that changes in pressure (P1 — P2) along a blood vessel are a function of blood flow (Q) and vascular resistance (R) where (P1 – P2) = Q × R. Increased resistance to portal blood flow is the principal driving force responsible for the pathogenesis of portal hypertension. Increased resistance to blood flow can develop anywhere along the portal venous system (i.e., in the portal vein itself or in its tributaries) before blood reaches the liver (prehepatic portal hypertension), the vascular spaces within the liver (intrahepatic portal hypertension), or the vascular compartments that receive the portal blood flow after it exits the liver (posthepatic portal hypertension). Examples of pre- and posthepatic portal hypertension are listed in Table 10–1.
Within the liver, obstruction to portal venous flow also can occur at multiple sites. For example, occlusion of small portal vein branches within portal triads may complicate liver diseases in which there is extensive portal and periportal inflammatory infiltration or fibrosis. When portal flow is obstructed at this level, the ensuing portal hypertension is termed presinusoidal. Alternatively, portal blood flow may be impeded by narrowing of the hepatic sinusoids as a result of collagen deposition or occlusion of the sinusoids by contractile cellular elements. Obstruction at this level leads to sinusoidal portal hypertension. Finally, portal blood flow may be obstructed at the level of the terminal hepatic venules. Such postsinusoidal portal hypertension may develop as a result of toxin-induced damage of small hepatic vein branches. Although many of the processes that impede portal blood flow presinusoidally, within the hepatic sinusoids, or postsinusoidally lead to cirrhosis, and cirrhosis is the most common cause of intrahepatic portal hypertension, it is important to recognize that obstruction to portal blood flow may occur at any level in the absence of cirrhosis (called noncirrhotic portal hypertension).
Pressure within the portal vein can be measured directly or indirectly. Direct measurement is performed by inserting a catheter into the portal vein itself (during laparatomy) or one of its branches (by percutaneous, transhepatic cannulation of an intrahepatic portal vein branch and advancement of the cannula into the main portal vein). Direct measurement within the portal vein is the most reliable way to estimate portal pressure when blood flow is obstructed before it reaches the liver. More commonly, however, portal vein pressure is estimated by indirect methods, which assume that the pressure across the sinusoidal bed within the liver has equilibrated so that the pressure in small tributaries of the portal vein (which deliver blood to the sinusoids) equals that of small branches of the hepatic vein (which carry blood away from the sinusoids). A catheter is inserted into an antecubital or femoral vein and advanced with fluoroscopic guidance into an hepatic vein. By “wedging” the catheter into a small branch of the hepatic vein, the pressure in the hepatic sinusoids (and thus indirectly in the portal vein) is recorded. The wedged hepatic vein pressure (WHVP) then is subtracted from the pressure, which is freely recorded in the hepatic vein (free hepatic venous pressure, or FHVP) to normalize for the effect of intraabdominal pressure. Thus, portal pressure is estimated by recording the pressure gradient between FHVP and WHVP. This procedure is extremely safe and simple and is therefore the most common method used to measure portal pressure; however, it may not be accurate if portal blood flow is blocked before it enters the sinusoids. In such cases, the hepatic venous pressure gradient would underestimate the severity of portal hypertension. To avoid confusion in this circumstance, another indirect method of measuring portal vein pressure can be used, which records the pressure in the splenic pulp. In practice, however, this method is rarely used because percutaneous puncture of the spleen carries a significant risk (1–2%) of bleeding and the technique has no method to normalize for intraabdominal pressure.
Prehepatic obstruction to portal blood flow Splenic vein thrombosis (pancreatitis, intraabdominal infection, abdominal trauma) Portal vein thrombosis (intraabdominal sepsis, abdominal trauma/surgery, umbilical vein catheterization) Intrahepatic obstruction to portal blood flow Presinusoidal (schistosomiasis, sarcoidosis, congenital hepatic fibrosis) Sinusoidal (injury-induced fibrosis, sickle cell crisis, hepatic infiltration with myeloid elements, amyloidosis) Postsinusoidal (alcohol-induced perivenular fibrosis, venoocclusive disease from antineoplastic agents) Posthepatic obstruction to portal blood flow Right heart failure Constrictive pericarditis Hepatic vein occlusion (thrombosis-Budd-Chiari syndrome, malignant, hepatic vein webs) |
Symptoms and Signs of Portal Hypertension
When portal blood flow is obstructed (either inside or outside of the liver), blood begins to pool in the vascular beds that normally empty into the portal vein. Congestion of omental tissues contributes to the pathogenesis of ascites. Sequestration of blood in the spleen causes splenic enlargement (splenomegaly) and increases splenic destruction of blood elements (hypersplenism) with secondary thrombocytopenia, neutropenia, and anemia. Eventually, a collateral circulatory system (i.e., varices) develops to decompress the portal bed by diverting portal blood flow into systemic veins. These collateral veins can become quite prominent in the esophagus, stomach, and distal intestine. Increased flow through such esophagogastric and hemorrhoidal varices can lead to vessel rupture, with devastating hemorrhage. If extensive variceal collaterals develop, significant amounts of portal blood flow are shunted away from the liver, and portal pressure may fall to within normal range. The end result, however, is that the liver is deprived of portal blood and must rely increasingly on hepatic artery blood for nurture; however, hepatic arterial blood is deficient in many of the splanchnically derived factors that are hepatotrophic. Thus, the liver’s normal capacity to regenerate is impaired and hepatic atrophy ensues, leading to a small, shrunken liver. In addition, portal–venous shunting diverts blood draining the intestines away from the liver, compromising hepatic clearance of gut-derived products. This has been implicated in the pathogenesis of hepatic encephalopathy. Increasing evidence suggests that portal-systemic shunting also permits systemic endotoxemia, which promotes the release of several proinflammatory cytokines, including tumor necrosis factor alpha (TNF) and endothelins. TNF contributes to the muscle wasting and endothelins to the renal dysfunction associated with advanced cirrhosis.
Hence, many physical findings in cirrhotic patients are caused by portal hypertension (Table 10–2). Conversely, detection of these signs in a patient with liver disease suggests that cirrhosis has developed. Furthermore, because many of the complications of portal hypertension present significant threats to survival in cirrhotic patients, they have been the focus of considerable diagnostic and therapeutic efforts. The following section briefly summarizes current knowledge about the pathogenesis, diagnosis, and treatment of the four most clinically important complications of portal hypertension: gastrointestinal bleeding, hepatic encephalopathy, ascites, and the hepatorenal syndrome.
Ascites Splenomegaly Collateral vessels (varices, “caput Medussa”) Hepatic atrophy Hepatic encephalopathy Cachexia Hepatopulmonary syndrome (cyanosis, clubbing, cor pulmonale) Feminization (gynecomastia, testicular atrophy, palmar erythema, spider telangiectasias) |
Clinical Complications of Portal Hypertension: Gastrointestinal Bleeding
Pathogenesis
Gastrointestinal bleeding is a feared complication of portal hypertension. Although many potential sources of bleeding (e.g., peptic ulcers, portal hypertensive gastropathy, portal–venous collaterals) may coexist in cirrhotic patients, the most clinically significant hemorrhage typically occurs when dilated venous collaterals (varices) rupture. Esophageal, gastric, and hemorrhoidal collaterals have the highest propensity to bleed profusely, although varices can form in many other locations (e.g., in surgical scars, ostomies, and around the umbilicus). Esophageal and gastric varices develop through connections between the coronary and short gastric veins (tributaries of the portal vein) and the azygos vein. Hemorrhoidal varices are the result of connections between the portal system and the middle and superior hemorrhoidal veins.
The prevalence of variceal collaterals among patients with cirrhosis is unknown, but it has been estimated that perhaps as few as a third of cirrhotic patients have esophageal varices. A pressure gradient between the portal and systemic venous systems (i.e., WHVP-FHVP) of at least 12 mm Hg is necessary, but not sufficient, for esophageal varices to develop. Other factors that appear to play a role in the development of esophageal varices are the extent of other collateral routes available to drain the portal bed and the presence or absence of conditions (e.g., lower esophageal sphincter pressure) that permit blood to flow into the azygos veins. Even once esophageal varices have formed, they usually do not bleed. In fact, the risk of an index episode of variceal bleeding has been estimated to be no greater than 30% per year. The risk of variceal rupture is more directly dependent on the wall tension of the varix than on the portal pressure per se. Wall tension is determined by the diameter of the varix and its wall thickness as well as the pressure within its lumen. Hence, for any given portal pressure, large, thin-walled veins within the esophagus are at greatest risk of rupture. Endoscopy is useful in identifying which esophageal varices have the highest risk of bleeding. The identification of large varices that protrude significantly into the esophageal lumen and the presence of red or blue discolored blebs overlying a varix correlate with an extremely high short-term risk of hemorrhage. Although vessel wall thickness appears to be important in determining which collaterals rupture, erosion of esophageal varices resulting from esophagitis or nasogastric tube placement is much less a threat to variceal integrity than rupture of the vessel wall due to elevations in intravariceal (i.e., portal) pressure. Hemorrhage from esophagogastric varices is typically a dramatic event with painless hematemesis or hematochezia. Bleeding is rarely insidious or chronic. Signs of hemodynamic instability (hypotension, tachycardia) are common. The risk of dying from variceal bleeding is variable and depends on the severity of the hemorrhage and the severity of the underlying liver disease. Patients with severely decompensated liver disease have at least a 50% chance of dying after hospitalization with a hemodynamically significant variceal hemorrhage. Furthermore, once varices have bled, they are at high risk (70 to 80% probability) of bleeding again, particularly within the first few weeks to months of the index hemorrhage.
Diagnosis
As noted, upper endoscopy offers a sensitive, specific means of identifying clinically threatening esophageal varices in cirrhotic patients who have never had overt gastrointestinal bleeding. Endoscopy is also diagnostically useful in patients who present with suspected gastroesophageal variceal bleeding; however, the latter patients first must be hemodynamically resuscitated. Optimal recuscitation requires intensive monitoring of blood losses and repletion of intravascular volume in an intensive care unit. In addition, before attempting to identify the source of gastrointestinal bleeding, patients with active hematemesis should be intubated electively to protect against inadvertent aspiration. Once hemodynamic rescuscitation and airway protection have been accomplished, upper endoscopy is performed to diagnose and treat the bleeding source. Even in patients with well-documented portal hypertension, endoscopic visualization of the upper gastrointestinal tract is essential because, as noted, nonvariceal sources of gastrointestinal bleeding also are common in cirrhotic patients, and both short- and long-term management will differ markedly, depending on the source of bleeding. Active bleeding from a varix, adherent clot on a varix, or signs of redness or thinning of a varix wall (dubbed “red or blue wale signs”) in the absence of other identifiable bleeding sources will implicate varices as the source of hemorrhage.
Treatment
Treatment of varices in the upper gastrointestinal tract can be subdivided into therapies for patients who have varices that have never bled (so-called prophylactic treatments) and therapies for patients whose varices have bled at least once. In the latter group of patients, treatment involves the acute management of active or recent bleeding and long-term strategies to prevent delayed rebleeding from varices. These are each summarized briefly in the following sections (Table 10–3).
Prophylactic (i.e., expectant) Beta-blockers titrated to reduce heart rate by 25% Therapy for acute hemorrhage from esophageal varices Hemodynamic stabilization/airway protection Endoscopic sclerosis or banding Octreotide or vasopressin + nitroglycerin Balloon tamponade Emergent coronary vein embolization/TIPS Emergent surgical portal decompressive shunt or esophageal transection Therapy to prevent rebleeding from esophageal varices Endoscopic sclerosis/banding to achieve variceal obliteration Beta-blockers (alone or in conjunction with endoscopic Rx) Elective TIPS Elective surgical portal–decompressive shunt Orthotopic liver transplantation |
Rx, medication; TIPS, transjugular intrahepatic portosystemic shunt.
Prophylactic Treatments to Prevent Index Hemorrhage from Esophageal Varices
Although vessel wall thickness is an important determinant for variceal rupture, there is little evidence that treatments that decrease esophagitis decrease the risk of bleeding from esophageal varices. On the other hand, treatments that reduce portal pressure (and thus lower intravariceal pressure) are effective in reducing the incidence of bleeding from esophageal varices. Greatest benefit has been demonstrated in patients who have varices that display endoscopic features of impending rupture. In such patients, treatment with beta-blockers to reduce the resting heart rate by 25% is generally well tolerated and significantly reduces the incidence of variceal bleeding. In contrast, other strategies to lower intravariceal pressure (i.e., portal–decompressive shunt surgery or endoscopic sclerotherapy) have proven too toxic to justify widespread use in patients who have never bled.
Management of Acute Variceal Bleeding