The most common cause of portal hypertension is cirrhosis. Vascular resistance and blood flow are 2 important factors in its development. The images below depict esophageal varices, which are responsible for the main complication of portal hypertension, massive upper gastrointestinal (GI) hemorrhage.
Large esophageal varices with red wale signs seen on endoscopy.
Uphill esophageal varices. Barium swallow demonstrates multiple serpiginous filling defects primarily involving the lower one third of the esophagus w....
Signs and symptoms of liver disease include the following:
Complications of portal hypertension in patients may present with the following symptoms:
Signs of portosystemic collateral formation include the following:
Signs of a hyperdynamic circulatory state include the following:
Other signs of portal hypertension and esophageal varices include the following:
See Clinical Presentation for more detail.
Other laboratory tests may include the following:
See Workup for more detail.
Treatment is directed at the cause of portal hypertension. Gastroesophageal variceal hemorrhage is the most dramatic and lethal complication of portal hypertension; therefore, the focus is on the treatment of variceal hemorrhage. Management of patients with liver cirrhosis and ascites but without hemorrhage includes a low-sodium diet and diuretics.
Surgery has no role in primary prophylaxis. Consider procedures, such as the following, for the prevention of rebleeding when pharmacologic and/or endoscopic therapy have failed:
See Treatment and Medication for more detail.
Many conditions are associated with portal hypertension, with cirrhosis being the most common cause of this disorder.
Two important factors—vascular resistance and blood flow—exist in the development of portal hypertension. Ohm law is V = IR, where V is voltage, I is current, and R is resistance. This can be applied to vascular flow; ie, P = FR, where P is the pressure gradient through the portal venous system, F is the volume of blood flowing through the system, and R is the resistance to flow. Changes in either F or R affect the pressure, although in most types of portal hypertension, both of these are altered. (See Anatomy and Etiology and Pathophysiology.)
Normal portal pressure is generally considered to be between 5 and 10 mm Hg. Once the portal pressure rises to 12 mm Hg or greater, complications can arise, such as varices and ascites. Indeed, esophageal varices are responsible for the main complication of portal hypertension, massive upper gastrointestinal (GI) hemorrhage (see Etiology and Pathophysiology, Prognosis, Presentation, and Workup). (See the images below.)
Uphill esophageal varices. Barium swallow demonstrates multiple serpiginous filling defects primarily involving the lower one third of the esophagus w....
Barium swallow demonstrating esophageal varices involving the entire length of the esophagus. This appearance may be seen in advanced uphill varices o....
The portal vein carries approximately 1500 mL/min of blood from the small and large bowel, the spleen, and the stomach to the liver. Obstruction of portal venous flow, whatever the etiology, results in a rise in portal venous pressure. (See Etiology and Pathophysiology.)
The response to increased venous pressure is the development of collateral circulation that diverts the obstructed blood flow to the systemic veins. These portosystemic collaterals form by the opening and dilatation of preexisting vascular channels connecting the portal venous system and the superior and inferior vena cava.[3, 4, 5] (See the image below.)
Computed tomography scan showing esophageal varices. Note the extensive collateralization within the abdomen adjacent to the spleen as a result of sev....
Although high portal pressure is the main cause of the development of portosystemic collaterals, other factors, such as active angiogenesis, may also be involved. The most important portosystemic anastomoses are the gastroesophageal collaterals, which include esophageal varices. The gastroesophageal collaterals drain into the azygos vein.
The portal vein drains blood from the small and large intestines, stomach, spleen, pancreas, and gallbladder. The superior mesenteric vein and the splenic vein unite behind the neck of the pancreas to form the portal vein. The portal trunk divides into 2 lobar veins. The right branch drains the cystic vein, and the left branch receives the umbilical and paraumbilical veins that enlarge to form umbilical varices in portal hypertension. The coronary vein, which runs along the lesser curvature of the stomach, receives distal esophageal veins, which also enlarge in portal hypertension. (See the image below.)
Portal vein and associated anatomy.
The initial factor in the etiology of portal hypertension is the increase in vascular resistance to the portal blood flow. Poiseuille’s law, which can be applied to portal vascular resistance, R, states that R = 8hL/pr4, where h is the viscosity of blood, L is the length of the blood vessel, and r is the radius of the blood vessel. The viscosity of the blood is related to the hematocrit. The lengths of the blood vessels in the portal vasculature are relatively constant.
Thus, changes in portal vascular resistance are determined primarily by blood vessel radius. Because portal vascular resistance is indirectly proportional to the fourth power of the vessel radius, small decreases in the vessel radius cause large increases in portal vascular resistance and, therefore, in portal blood pressure (P = F8hL/pr4, where P is portal pressure and F is portal blood flow).
Liver disease that decreases the portal vascular radius produces a dramatic increase in portal vascular resistance. In cirrhosis, the increase occurs at the hepatic microcirculation (sinusoidal portal hypertension). Increased hepatic vascular resistance in cirrhosis is not only a mechanical consequence of the hepatic architectural disorder; a dynamic component also exists due to the active contraction of myofibroblasts, activated stellate cells, and vascular smooth-muscle cells of the intrahepatic veins.
Endogenous factors and pharmacologic agents that modify the dynamic component include those that increase or decrease hepatic vascular resistance. Factors that increase hepatic vascular resistance include endothelin-1 (ET-1), alpha-adrenergic stimulus, and angiotensin II. Factors that decrease hepatic vascular resistance include nitric oxide (NO), prostacyclin, and vasodilating drugs (eg, organic nitrates, adrenolytics, calcium channel blockers).[4, 7]
Endothelin and nitric oxide
Studies have demonstrated the role of ET-1 and NO in the pathogenesis of portal hypertension and esophageal varices.[4, 7] ET-1 is a powerful vasoconstrictor synthesized by sinusoidal endothelial cells that has been implicated in the increased hepatic vascular resistance of cirrhosis and in the development of liver fibrosis.
NO is a vasodilator substance that is also synthesized by sinusoidal endothelial cells. In the cirrhotic liver, the production of NO is decreased, and endothelial nitric oxide synthase (eNOS) activity and nitrite production by sinusoidal endothelial cells are reduced. As a result, intrahepatic vasoconstriction occurs in cirrhotic liver and accounts for approximately 20-30% of the increased intrahepatic resistance.[7, 8, 9] Another major contribution to increased portal venous pressure is the concomitant splanchnic arteriolar vasodilation causing increased portal venous inflow.
Obstruction and increased resistance can occur at 3 levels in relation to the hepatic sinusoids, as follows (see the Table, below):
Table 1. Interpretation of Surrogate Portal Venous Pressure Measurements in the Differential Diagnosis of Portal Hypertension
With regard to the liver itself, causes of portal hypertension usually are classified as prehepatic, intrahepatic, and posthepatic.
Prehepatic causes of increased resistance to flow include the following:
Studies of hepatic microcirculation have identified several mechanisms that may explain increased intrahepatic vascular resistance to flow. These mechanisms may be summarized as follows :
More specifically, intrahepatic, predominantly presinusoidal causes of resistance to flow include the following:
Intrahepatic, predominantly sinusoidal causes of resistance include the following:
With regard to chronic active hepatitis, noncirrhotic portal fibrosis is observed with various toxic injuries, and one of these includes vitamin A toxicity. This probably is due to vascular injury. Excessive doses of vitamin A taken for months or years can lead to chronic hepatic disease. Intake of doses ranging from as small as 3-fold the recommended daily dose continued for years to doses as high as 20-fold the approved dose for a few months can lead to hepatic disease. The pericellular fibrotic characteristic of vitamin A toxicity may lead to portal hypertension.
Postsinusoidal obstruction syndrome and veno-occlusive disease of the liver are postsinusoidal causes of resistance.
Posthepatic causes of resistance to flow include the following:
The second factor that contributes to the pathogenesis of portal hypertension is an increase in blood flow in the portal veins. This increase is established through splanchnic arteriolar vasodilatation caused by an excessive release of endogenous vasodilators (eg, endothelial, neural, humoral).
The increase in portal blood flow aggravates the increase in portal pressure; the increased flow contributes to the ability of portal hypertension to exist despite the formation of an extensive network of portosystemic collaterals that may divert as much as 80% of the portal blood flow.
Manifestations of splanchnic vasodilatation include increased cardiac output, arterial hypotension, and hypervolemia. This explains the rationale for treating portal hypertension with a low-sodium diet and diuretics to attenuate the hyperkinetic state.
An elevated pressure difference between systemic and portal circulation (ie, HVPG) directly contributes to the development of varices.[8, 12, 13] HVPG is a surrogate marker of portal pressure gradient and is derived from WHVP corrected (subtracted) with free hepatic venous pressure (FHVP).
The hypertensive portal vein is decompressed by diverting up to 90% of the portal flow through portasystemic collaterals back to the heart, resulting in enlargement of these vessels. These vessels are commonly located at the gastroesophageal junction, where they lie subjacent to the mucosa and present as gastric and esophageal varices. Varices form when the HVPG exceeds 10 mm Hg; they usually do not bleed unless the HVPG exceeds 12 mm Hg (normal HVPG: 1-5 mm Hg).[8, 12, 13, 14]
Gastroesophageal varices have 2 main inflows. The first is the left gastric or coronary vein, and the second is the splenic hilum, through the short gastric veins. The gastroesophageal varices are important because of their propensity to bleed. (See the images below.)
Normal venous flow through the portal and systemic circulation. IMC = inferior mesenteric vein; IVC = inferior vena cava; SVC = superior vena cava.
Redirection of flow through the left gastric vein secondary to portal hypertension or portal venous occlusion. Uphill varices develop in the distal on....
Increased portal pressure contributes to increased varix size and decreased varix wall thickness, thus leading to increased variceal wall tension. Rupture occurs when the wall tension exceeds the elastic limits of the variceal wall. Varices are most superficial at the gastroesophageal junction and have the thinnest wall in that region; thus, variceal hemorrhage invariably occurs in that area.
The following are risk factors for variceal hemorrhage[8, 12, 15] :
Note that bacterial infection could also trigger variceal bleeding through a number of mechanisms, including the following:
Population-based prevalence data for portal hypertension in the United States are not available, but portal hypertension is a frequent manifestation of liver cirrhosis. According to the National Institute on Alcohol Abuse and Alcoholism (NIAAA), liver cirrhosis accounted for almost 30,000 deaths in the United States in 2007, making it the 12th leading cause of US deaths.
The international incidence of portal hypertension is also not known, although it is probably similar to that of the US, with differences primarily in the causes. In Western countries, alcoholic and viral cirrhosis are the leading causes of portal hypertension and esophageal varices; 30% of patients with compensated cirrhosis and 60-70% of patients with decompensated cirrhosis have gastroesophageal varices at the time of diagnosis.[8, 12] The frequency of gastroesophageal varices directly correlates with the severity of the liver disease from 40% in Child class A to 85% in Child class C.[8, 12]
The de novo rate of development of esophageal varices in US patients with chronic liver disease is approximately 8% per year for the first 2 years and 30% by the sixth year. The risk of bleeding from esophageal varices is 30% in the first year after identification. Bleeding from esophageal varices accounts for approximately 10% of episodes of upper gastrointestinal bleeding.
Hepatitis B is endemic in the Far East and Southeast Asia, particularly, as well as in South America, North Africa, Egypt, and other countries in the Middle East. Schistosomiasis is an important cause of portal hypertension in Egypt, Sudan, southern and sub-Saharan Africa, Southeast Asia, Caribbean, and South America. Nonalcoholic steatohepatitis (NASH) is becoming a major cause of liver cirrhosis in the United States as hepatitis C is becoming a major cause of liver cirrhosis worldwide.
Liver disease demonstrates a sex predilection, with males making up more than 60% of patients with chronic liver disease and cirrhosis.
In general, alcoholic liver disease and viral hepatitis are the most common causes for esophageal varices in both sexes. However, veno-occlusive diseases and primary biliary cirrhosis are more common in females; and in females with esophageal varices, alcoholic liver disease, viral hepatitis, veno-occlusive disease, and primary biliary cirrhosis are usually responsible. In males with esophageal varices, alcoholic liver disease and viral hepatitis are usually the cause.
Portal vein thrombosis and secondary biliary cirrhosis are the most common causes of esophageal varices in children. Cirrhosis is the most common cause of esophageal varices in adults.
Patients with severe and persistent upper gastrointestinal (GI) hemorrhage (ie, requiring transfusions of >5 U of packed red blood cells) have higher morbidity and mortality rate.
Variceal hemorrhage is the most common complication associated with portal hypertension. Almost 90% of patients with cirrhosis develop varices, and approximately 30% of varices bleed. The estimated mortality rate for the first episode of variceal hemorrhage is 30-50%.
Patients with a known diagnosis of esophageal varices have a 30% chance of variceal bleeding within the first year after the diagnosis. The mortality rate of the bleeding episode depends on the severity of the underlying liver disease.
Patients who have had 1 episode of bleeding from esophageal varices have a 60-80% chance of rebleeding within 1 year after the initial episode; approximately one third of further bleeding episodes are fatal.[8, 12, 20] The risk of death is maximal during the first few days after the bleeding episode and decreases slowly over the first 6 weeks. However, despite improvements in therapy, the mortality rate at 6 weeks is remains greater than 20%; this rate is higher when surgical intervention is needed.
Associated abnormalities in the renal, pulmonary, cardiovascular, and immune systems of patients with esophageal varices contribute to 20-65% of deaths in these individuals. Complications associated with portal hypertension and GI bleeding include the following:
Other complications include those related to blood transfusion(s) and/or those related to the therapeutic procedures used in the management of bleeding varices.
In a retrospective study (2002-2014) of 80 patients with portopulmonary hypertension, Mayo Clinic investigators noted that intrapulmonary vascular dilatations (IPVDs) were common and associated with reduced survival. The presence of IPVDs was detected by agitated saline contrast-enhanced transthoracic echocardiography (cTTE), traditionally more frequently associated with hepatopulmonary syndrome than with portopulmonary hypertension.
Patients with a hepatic venous pressure gradient (HVPG) of 20 mm Hg measured 24 hours after the onset of bleeding esophageal varices have a higher 1-year mortality rate. Other factors that can affect the prognosis of patients with esophageal varices include the following:
Several factors are known to influence the prognosis of esophageal bleeding. These include the following:
Educate patients about the benefits and disadvantages of available treatment options.
Alcohol intake should strongly be discouraged, especially in patients with alcoholic cirrhosis. Available resources for alcohol rehabilitation should be provided, along with any prophylaxis for alcohol withdrawal symptoms, when indicated.
Unless contraindicated, all patients with esophageal varices should take beta-blockers to reduce the risk of bleeding. Patients should also be educated about the adverse effects of beta-blockers and the possible risks of their abrupt discontinuation.
Advise patients who have ascites of the risk of spontaneous bacterial peritonitis during an episode of acute variceal bleeding.
For patient education information, see the Digestive Disorders Center and the Heartburn & GERD Center, as well as Gastrointestinal Bleeding, Cirrhosis, and Gastritis.
In obtaining the medical history of a patient with portal hypertension, attention should be directed toward determining the cause of the condition and, secondarily, to which complications are present.
Determining the cause of portal hypertension involves obtaining information on the following:
Despite conflicting studies, the most common causes of gastrointestinal (GI) bleeding are peptic ulcer disease, of which gastric ulcers are usually more common than duodenal ulcers, and a nonspecific mucosal abnormality (21-55%).[22, 23, 24] Bleeding from esophageal varices is responsible for 12-14% of upper GI bleeding; acute gastric erosions/hemorrhagic gastritis, Mallory-Weiss tears, gastric carcinoma, and Dieulafoy lesion, account for less than 10% of cases.[22, 23, 24]
Patient history of risk factors for upper GI bleeding, including the following, should also be assessed:
Symptoms of liver disease include the following:
The presence of complications of portal hypertension can be ascertained by determining whether the following are present:
Check the patient's blood pressure and pulse with the patient in the supine and seated positions.
Signs of portosystemic collateral formation include the following:
Signs of liver disease include the following:
Signs of a hyperdynamic circulatory state include the following:
Other signs of portal hypertension and esophageal varices include the following:
As noted under Physical Examination, rectal examination that reveals a black, soft, tarry stool on the gloved examining finger suggests upper gastrointestinal bleeding.
Ultrasonography is a safe, economical, and effective method of screening for portal hypertension. It can also demonstrate portal flow and helps in diagnosing cavernous transformation of the portal vein, portal vein thrombosis, and splenic vein thrombosis. Ultrasonography of the upper abdomen may be indicated in patients with esophageal varices, especially if biliary obstruction or liver cancer is suspected. Computed tomography (CT) scanning and magnetic resonance imaging (MRI) can be used when ultrasonographic findings are inconclusive.
When bleeding is obscure and the source is unclear, a bleeding scan or angiography may be warranted; angiography can also provide therapeutic intervention in an acute bleeding episode. Selective angiography of the superior mesenteric artery or splenic artery with venous return phase can also be performed in patients with portal hypertension.
On liver biopsy, histologic findings are varied and depend not only on the cause of liver disease but also on the cause of portal hypertension. Zone 3 necrosis can be observed in portal hypertension secondary to congestive heart failure and Budd-Chiari syndrome. In cases of normal liver parenchyma, investigate for prehepatic causes of portal hypertension.
Laboratory studies are directed towards investigating the etiologies of cirrhosis, which is the most common cause of portal hypertension. The rate and volume of bleeding in the patient should be assessed.
Gain venous access and obtain blood for immediate hematocrit measurement. Obtain a type and cross-match for possible blood product transfusion. Measure the platelet count and prothrombin time (PT), send blood for renal and liver function tests (LFTs), and measure serum electrolyte levels.
The presence of anemia, leukopenia, and thrombocytopenia may be present in patients with cirrhosis. Anemia may be secondary to bleeding, nutritional deficiencies, or bone marrow suppression secondary to alcoholism. Pancytopenia can result from hypersplenism, a common complication in patients with portal hypertension. Serial monitoring of the hemoglobin and hematocrit value is useful in patients with suspected ongoing gastrointestinal bleeding.
Abnormal liver function can be approached as a transaminitis (an elevation of the plasma activity of aspartate aminotransferase [AST] and alanine aminotransferase [ALT]) or cholestasis (an elevation of bilirubin, especially conjugated bilirubin, with or without increased alkaline phosphatase [ALP] activity), both of which may occur in cirrhosis. However, normal liver function studies do not exclude liver disease, as a "burned out" liver (ie, one that loses features of disease activity) may not give rise to aminotransferase activity.
Transfusion with packed red blood cells (RBCs) and fresh frozen plasma (FFP) are usually required in patients with massive variceal bleeding.
Coagulation studies include PT, partial thromboplastin time (PTT), and international normalized ratio (INR). Because the synthetic function of the liver is impaired in cirrhotic patients, coagulopathy with prolonged PT and PTT is expected; INR is also used to assess the severity and prognosis of the liver disease through Model for End-Stage Liver Disease (MELD) score calculation (see the MELD Score calculator). Prolonged INR is suggestive of impaired hepatic synthetic function. See also the Medscape Reference articles Cirrhosis and Liver Transplantation.
Blood urea nitrogen (BUN) and creatinine levels may be elevated in patients with esophageal bleeding; BUN is also used in calculating the Blatchford bleeding score in the initial evaluation, and serum creatinine results are used in calculating the MELD score.
A high anion gap may suggest hyperlactatemia or hyperammonemia.
Obtain viral hepatitis serologies, particularly hepatitis B and C. These may help in assessing the cause of liver cirrhosis.
Other laboratory tests may include the following:
On duplex Doppler ultrasonography, features suggestive of hepatic cirrhosis with portal hypertension include the following:
Limitations of ultrasonography include the fact that the reproducibility of data is problematic and that many variables, such as circadian rhythm, meals, medications, and the sympathetic nervous system, affect portal hemodynamics. Moreover, significant interobserver and intraobserver variation exist in quantitative ultrasonographic measurement.
Computed tomography (CT) scanning is a useful qualitative study when ultrasonographic evaluations are inconclusive. CT scanning is not affected by the patient’s body habitus or the presence of bowel gas. With improvement of spiral CT scanning and 3-dimensional (3-D) angiographic reconstructive techniques, portal vasculature may be visualized more accurately. (See the image below.)
Computed tomography scan showing esophageal varices. Note the extensive collateralization within the abdomen adjacent to the spleen as a result of sev....
Findings suggestive of portal hypertension include collaterals arising from the portal system and dilatation of the inferior vena cava (IVC).
Limitations of CT scanning include the fact that it cannot demonstrate the venous and arterial flow profile and that intravenous contrast agents cannot be used in patients with renal failure or contrast allergy.
Magnetic resonance imaging (MRI) provides qualitative information similar to that from CT scanning when Doppler ultrasonographic findings are inconclusive. MRI angiography detects the presence of portosystemic collaterals and obstruction of portal vasculature. MRI also provides quantitative data on portal venous and azygos blood flow.
Liver-screen scanning is described for historical interest only, because this technique has been superseded by ultrasonography and CT scanning. Liver-spleen scans use technetium sulfur colloid, which is taken up by cells in the reticuloendothelial system. A colloidal shift from the liver to the spleen or bone marrow is suggestive of increased portal pressure.
Limitations of these scans include the fact that portal hypertension cannot be ruled out in the absence of this shift. In addition, liver-spleen scans lack spatial resolution.
Direct portal measurements are usually not performed, due to their invasive nature, the risk of complications, and the interference of anesthetic agents with portal hemodynamics.
More commonly, measurement of the hepatic venous pressure gradient (HVPG) is performed; this is an indirect measurement that closely approximates portal venous pressure. Monitoring HVPG is useful in assessing the patient's response to treatment, progression of the disease, and prognosis. Reduction in HVPG of greater than 20% of baseline or to less than 12 mm Hg significantly reduces mortality and the risk of recurrent variceal hemorrhage.
A fluid-filled balloon catheter is introduced into the femoral or internal jugular vein and advanced under fluoroscopy into a branch of the hepatic vein. Free hepatic venous pressure (FHVP) is then measured. The balloon is inflated until it is wedged inside the hepatic vein, occluding it completely and thus equalizing the pressure throughout the static column of blood. The occluded hepatic venous pressure (ie, wedged hepatic venous pressure) minus the unoccluded, or free, portal venous pressure (ie, FHVP) is the HVPG.
In cirrhotic patients, measurements of peripheral endothelin-1 (ET-1) and transforming growth factor-beta1 levels may be used as noninvasive markers of portal hypertension and liver insufficiency. Wereszczynka-Siemiatkowska et al found that peripheral levels of these mediators are strongly correlated with their hepatic levels. Among their findings were that patients with cirrhosis had significantly higher levels of peripheral ET-1 but decreased levels of transforming growth factor-beta1, as well as before and after treatment, peripheral and hepatic ET-1, transforming growth factor-beta1 and 2 levels correlated significantly with liver failure indicators (eg, laboratory parameters, Child-Pough and MELD scores) and pressure gradient values.
Endoscopy (esophagogastroduodenoscopy [EGD]) is an essential diagnostic and therapeutic tool at an early stage to formulate the management plan for patients with esophageal varices. If active variceal bleeding or an adherent clot is observed, variceal hemorrhage can be diagnosed confidently. The presence of any of the following risk factors warrants a screening endoscopy to search for varices :
The presence of variceal red color signs (eg, cherry red spots, red wale markings [longitudinal red streaks on varices], blue varices) and the "white nipple sign" (platelet fibrin plug overlying a varix, resembling a white nipple) indicates an increased risk of rebleeding.[12, 27, 28]
Perform upper endoscopy, as appropriate, to screen for varices in every patient with suggestive findings of portal hypertension. This procedure allows not only direct visual evaluation of the size, location, and bleeding stigmata of the lesion, but it can also provide prompt therapeutic intervention.
At the initial diagnosis of cirrhosis, all of these patients should be considered for the presence of varices; however, endoscopy may be omitted in patients who are already on a nonselective beta-blocker for other indications. Gastroesophageal varices confirm the diagnosis of portal hypertension; however, their absence does not rule it out. At times, gastroesophageal varices are incidental findings in patients undergoing upper endoscopy for other reasons (eg, dyspepsia refractory to medications, dysphagia, weight loss); these patients should undergo further investigations for etiologies of portal hypertension.
Various indirect indices, such as platelet count, spleen size, albumin, and Child-Pugh score, have been studied to help diagnose varices without endoscopy. A case review study, however, revealed that some of these predictors are unreliable. For the time being, endoscopy remains the criterion standard for screening patients with cirrhosis for varices.
Periodic surveillance endoscopy should be performed in patients with cirrhosis as follows :
Although measurement of hepatic venous pressure gradient (HVPG) and upper endoscopy are considered the criterion standards for assessment of portal hypertension, ultrasonography-based transient elastography is a novel noninvasive technology to detect clinically significant portal hypertension. Further studies are being conducted to validate this.
Korean investigators have indicated that shear wave elastography is a reliable noninvasive study for predicting clinically significant and severe portal hypertension. In 92 patients with cirrhosis, real-time shear wave elastography measurement of liver stiffness appeared to be strongly correlated with HVPG regardless of the presence/absence of ascites.
Treatment is directed at the cause of portal hypertension. Gastroesophageal variceal hemorrhage is the most dramatic and lethal complication of portal hypertension; therefore, most of the following discussion focuses on the treatment of variceal hemorrhage. Medical care includes emergent treatment, primary and secondary prophylaxis, and surgical intervention.
Pharmacologic therapy for portal hypertension includes the use of beta-blockers, most commonly propranolol and nadolol. Brazilian investigators have suggested that the use of some statins (eg, simvastatin) may lower portal pressure and potentially improve liver function. In a 3-month prospective, triple-blind randomized trial with simvastatin 40 mg/day and placebo in 34 patients with cirrhotic portal hypertension, 55% of those who received simvastatin showed a clinically relevant decrease in hepatic venous pressure gradient compared to none in the placebo group. Moreover, simvastatin response rates were greater in those with medium to large esophageal varices and previous variceal bleeding.
Endoscopic procedures such as sclerotherapy and variceal ligation can be used to prevent the recurrence of variceal hemorrhage. Surgical care includes the use of decompressive shunts, devascularization procedures, and liver transplantation. Decompressive shunts and devascularization procedures are mainly rescue therapies.
Management of patients with liver cirrhosis and ascites but without hemorrhage includes a low-sodium diet and diuretics.
In patients with hemodynamically significant upper gastrointestinal (GI) tract bleeding, a nasogastric tube should remain in place for 24 hours to assist in identifying any rebleeding. Gastric lavage may be performed frequently through the nasogastric tube, and the volume and appearance of material aspirated from the stomach should be recorded. Do not allow any food by mouth.
Initial volume resuscitation with or without blood product transfusion, together with medical treatment to reduce portal pressure (ie, anti-secretory agent infusion) should be promptly initiated in the emergency department. A transfer may be necessary if endoscopic treatment and/or surgical treatment are not readily available. If possible, transfer patients with uncontrollable bleeding from portal hypertension; these individuals should be sent to a tertiary center with a liver transplantation service.
Surgery has no role in primary prophylaxis. Its role in acute variceal bleeding is exceedingly limited, because therapy with endoscopic treatment controls bleeding in 90% of patients. A transjugular intrahepatic portosystemic shunt (TIPS) is a viable option and is less invasive for patients whose bleeding is not controlled. However, if TIPS is not available, then staple transection of the esophagus is an option when endoscopic treatment and pharmacologic therapy have failed.
Consider surgery for the prevention of rebleeding when pharmacologic and/or endoscopic therapy have failed. As per the Baveno II consensus conference on portal hypertension, failure is defined as a single episode of clinically significant rebleeding (transfusion requirement of 2 U of blood or more within 24 h, a systolic blood pressure < 100 mm Hg or a postural change of >20 mm Hg, and/or a pulse rate greater than 100 bpm).
Surgical interventions include the following:
Surgical shunts provide better control of rebleeding when compared to the combination therapy of beta-blocker and endoscopic variceal ligation (EVL). However, these shunts are associated with higher incidence of hepatic encephalopathy and should be reserved for Child class A patients with recurrent bleeding despite adequate combination therapy. Decompressive shunts include total portal systemic shunts, partial portal systemic shunts, and other selective shunts.
Total portal systemic shunts include any shunt larger than 10 mm in diameter between the portal vein (or one of its main tributaries) and the inferior vena cava (IVC) (or one of its tributaries).
Eck fistula (a classic end-to-side portacaval shunt; described for historical interest only) was performed by Eck in dogs in the late 19th century. The portal vein is divided close to the liver, the hepatic end of the portal vein is ligated, and the splanchnic end is anastomosed to the IVC. This controls variceal bleeding and decompresses splanchnic hypertension but leaves high pressure in the hepatic sinusoids; thus ascites is not relieved.
For the side-to-side portacaval shunt, the portal vein and the infrahepatic IVC are mobilized after dissection and anastomosed. All portal flow is directed through the shunt, with the portal vein itself acting as an outflow from the obstructed hepatic sinusoids. Excellent control of bleeding and ascites is achieved in more than 90% of patients. Encephalopathy (rate of 40-50%) and progressive liver failure are possible. The procedure has relatively limited indications, which include massive variceal bleeding with ascites or acute Budd-Chiari syndrome without evidence of liver failure.
Partial portal systemic shunts reduce the size of the anastomosis of a side-to-side shunt to 8 mm in diameter. Portal pressure is reduced to 12 mm Hg, and portal flow is maintained in 80% of patients.
The operative approach is similar to that for side-to-side portacaval shunts, except the interposition graft must be placed between the portal vein and the IVC.
Two prospective, randomized, controlled trials revealed a 90% rate for control of bleeding. Maintenance of some portal flow has decreased the incidence of encephalopathy and liver failure.
Selective shunts provide selective decompression of gastroesophageal varices to control bleeding while at the same time maintaining portal hypertension to maintain portal flow to the liver. One example is the distal splenorenal shunt, which is the most commonly used decompressive operation for refractory variceal bleeding; it is used primarily in patients who present with refractory bleeding and continue to have good liver function. The distal splenorenal shunt decompresses the gastroesophageal varices through the short gastric veins, the spleen, and the splenic vein to the left renal vein.
Portal hypertension is maintained in the splanchnic and portal venous system, and the shunt maintains portal flow to the liver. This type of shunt provides the best long-term maintenance of some portal flow and liver function, with a lower incidence of encephalopathy (10-15%) compared with total shunts. The operation produces ascites because the retroperitoneal lymphatics are diverted.
Devascularization is rarely performed but may have a role in patients with portal and splenic vein thrombosis who are not suitable candidates for shunt procedures and who continue to have variceal bleeding despite endoscopic and pharmacologic treatment.
Devascularization procedures consist of the transabdominal devascularization of the lower 5 cm of the esophagus and the upper two thirds of the stomach, with staple gun transection of the lower esophagus) (eg, splenectomy, gastroesophageal devascularization, and esophageal transection [at times]).
The incidence of liver failure and encephalopathy is low following devascularization procedures, presumably because of better maintenance of portal flow. However, these procedures are rarely performed but may have a role in patients with portal and splenic vein thrombosis who are not suitable candidates for shunt procedures and who continue to have variceal bleeding despite endoscopic and pharmacologic treatment.
The spleen is one of the major inflow paths to gastroesophageal varices. Splenectomy allows better access to the gastric fundus and the distal esophagus to complete the devascularization.
Portal vein thrombosis of as much as 20% is reported following splenectomy. Ascites is a frequent early postoperative complication because portal hypertension is maintained.
Gastroesophageal devascularization should devascularize the whole greater curve of the stomach from the pylorus to the esophagus and the upper two thirds of the lesser curve of the stomach. The esophagus should be devascularized for a minimum of 7 cm.
In patients who have undergone extensive and repeated sclerotherapy, the gastroesophageal junction is thickened and the ability to perform a satisfactory transection is limited.
Liver transplantation should be considered for patients with end-stage liver disease (eg, cirrhotic patients with Child-Pugh score =7 or Model for End-Stage Liver Disease [MELD] score =15 [see the MELD Score calculator]). The selection of candidates is dictated by the patient's clinical status, etiology of cirrhosis (viral hepatitis, alcoholic, nonalcoholic steatohepatitis, cholestatic liver disease), abstinence from alcohol, and availability of a donor organ.
Liver transplantation is the ultimate shunt, because it relieves portal hypertension, prevents variceal rebleeding, and manages ascites and encephalopathy by restoring liver function. It is the treatment modality that has significantly improved the outcome of patients with Child-Pugh class C disease and variceal bleeding.
In most patients, it is impractical to use liver transplantation to treat portal hypertension, because these individuals can be managed successfully with lesser methods. Therefore, the use of transplantation must be based on appropriate patient selections, as follows :
A study by Burger-Klepp et al indicated that, despite concerns that transesophageal echocardiography (TEE) can cause esophageal and gastric variceal hemorrhage, TEE is a relatively safe means of monitoring cardiac performance in patients with varices who are undergoing OLT. Of 287 patients in the study who underwent OLT and who had esophageal (82.2%), gastric (4.2%), or esophagogastric (13.6%) varices, only 1 major incidence of hemorrhage occurred.
Secondary prophylaxis is used to prevent rebleeding. Variceal hemorrhage has a 2-year recurrence rate of approximately 80%.
Propranolol and nadolol significantly reduce the risk of rebleeding and are associated with prolongation of survival. Studies comparing propranolol with sclerotherapy in the prevention of variceal rebleeding demonstrated comparable rates of variceal rebleeding and survival, but sclerotherapy was associated with significantly more complications.
Endoscopic sclerotherapy is usually performed at weekly intervals. Approximately 4-5 sessions are required for the eradication of varices, which is achieved in nearly 70% of patients.
Endoscopic variceal ligation (EVL) is considered the endoscopic treatment of choice in the prevention of rebleeding. Sessions are repeated at 7- to 14-day intervals until variceal obliteration (which usually requires 2-4 sessions). This procedure is associated with lower rebleeding rates and a lower frequency of esophageal strictures. Fewer sessions are required to achieve variceal obliteration than are required for sclerotherapy. See the video below.
Band ligation of esophageal varices.
A randomized trials demonstrated that EVL plus nadolol plus sucralfate is more effective in preventing variceal rebleeding than is EVL alone. In a more recent study by the same investigators, combined EVL and pharmacotherapy was marginally more effective than pharmacotherapy alone for preventing variceal rebleeding. These studies appear to indicate that combining EVL with beta-blockers is reasonable for patients in whom pharmacologic therapy has failed.
However, another report had different results. In a study by Kumar et al that compared the effectiveness of EVL alone with that of combination therapy consisting of EVL, propranolol, and isosorbide mononitrate (ISMN) for secondary prophylaxis in patients with previous variceal bleeding, no difference between the groups was observed for rebleeding 2 years after initial therapy. The investigators concluded that EVL by itself is sufficient to prevent variceal rebleeding. Moreover, they found that the addition of propranolol and ISMN to EVL may increase the risk for adverse effects.
Despite the contrasting findings above, combination of beta-blocker therapy with EVL is considered to the best option for secondary prophylaxis of variceal hemorrhage. Rather than titrating beta-blockers to goal reduction in heart rate, doses should be titrated to the maximal tolerated dose, because a goal reduction in heart rate may not correlate to a reduction in hepatic venous pressure gradient (HVPG). EVL should be repeated every 1-2 weeks until complete variceal obliteration occurs; then, endoscopy can be repeated every 3-6 months to evaluate for recurrence and for the need to repeat EVL. .
Complications of EVL are not frequent (14%) and usually minor, including transient dysphagia, chest discomfort, and small ulcer around the variceal base. Treatment with a proton-pump inhibitor for 10 days after EVL can reduce the size of these ulcers.
Complications related to the therapeutic procedures used in management of bleeding esophageal varices include the following:
Consider early consultation with a gastroenterologist and a surgeon, particularly for patients with active bleeding from esophageal varices. Consultation with a hepatologist and transplant surgery should be considered in patients with Child class B or C disease or a high Model for End-Stage Liver Disease (MELD) score. Good coordination among gastroenterologists, interventional radiologists, critical care team, and surgeons is essential.
To prevent recurrent variceal hemorrhage, patients with portal hypertension should have endoscopic variceal ligation (EVL) sessions scheduled until complete obliteration of varices is achieved. EVL sessions are repeated at 7- to 14-day intervals. These usually require 2-4 sessions for complete obliteration of varices.
As noted in Upper Gastrointestinal Endoscopy, periodic surveillance endoscopy should be performed in patients with cirrhosis as follows[8, 12, 19] :
Ferreira et al suggested that a high portal blood flow velocity can indicate progression of gastroesophageal varices and the need to include a patient in a postoperative, on-demand, endoscopic follow-up program of varices eradication (rather than in a prophylactic program). Their study findings included significantly higher values of portal blood flow velocity in patients with variceal progression. Those with a portal flow velocity of greater than 15.5 cm/s at the first postoperative year not only had progression of gastroesophageal varices but also were at higher risk for rebleeding.
These investigators analyzed data on portal vein Doppler ultrasonography for postoperative follow-up in 146 patients with schistosomal portal hypertension and a previous history of upper digestive bleeding from gastroesophageal varices rupture, who had undergone a gastroesophageal devascularization procedure with splenectomy. At each of 4 postoperative time points—1, 2, and 5 years and up to 10 years—patients were separated into 2 groups according to gastroesophageal varices progression; diameter and mean blood flow velocity were measured at these time points.
Promptly resuscitate and restore the circulating blood volume in patients with suspected cirrhosis and variceal hemorrhage (bleeding esophageal varices can be fatal). Rapid initial evaluation with measurement of vital signs, including orthostatic hypotension, assessment of the rate and volume of bleeding, mental status, and patient's ability to protect the airway is mandatory.
Establish 2 large-bore venous accesses for blood transfusion. While results are pending for the complete blood count (CBC), prothrombin/partial thromboplastin time (PT/PTT), and international normalized ratio (INR) and while the blood is being cross-matched, start rapid infusion of 5% dextrose and colloid solution until the blood pressure is restored and the urine output is adequate. Obtain other laboratory tests (eg, serum electrolyte levels, including calcium, especially when a large transfusion is required; serum creatinine levels; liver function tests [LFTs]) (see Laboratory Studies).
Blood should be replaced at a modest target of the hematocrit of 25-30%. The goal is to maintain hemodynamic stability and hemoglobin of approximately 8 g/dL.[8, 40, 41] As soon as the acute bleeding episode is adequately controlled, it is critical to initiate therapy to prevent recurrent bleeding. Fluid resuscitation should be made with caution: Avoid vigorous saline and blood infusion due to the risk of rebound increased portal pressure precipitating recurrent variceal hemorrhage and ascetic fluid accumulation. If indicated (eg, patients with severe coagulopathy with/without significant thrombocytopenia [< 50,000/mm3 platelets]), correct clotting factor deficiencies with fresh frozen plasma, fresh blood, and vitamin K.
Establish airway protection in patients with massive upper gastrointestinal (GI) tract bleeding, especially if the patient is not fully conscious.
Insert a nasogastric tube to assess the severity of the bleeding, to decompress the stomach, and to lavage the gastric contents to improve visualization during endoscopy.
Variceal bleeding may cease spontaneously in as many as 50% of patients. Each episode of variceal bleeding is associated with a 30% mortality rate. Such episodes occur mostly in patients with severe liver disease and in those with early rebleeding. Rebleeding occurs in 40% of patients within 6 weeks.
Following resuscitation, treatment of acute variceal bleeding includes control of bleeding (24 h without bleeding within the first 48 h following the start of therapy) and prevention of early recurrence.
All patients with cirrhosis and upper GI bleeding are at a high risk for developing severe bacterial infections. These infections are associated with early rebleeding.
A short course of prophylactic antibiotics has been demonstrated to decrease both the rate of bacterial infections and mortality rates. The improvement in the survival rate with antibiotic prophylaxis has been attributed to a decrease in early rebleeding.[8, 15] Thus, prophylactic antibiotic use (norfloxacin 400 mg PO bid for 7 d; alternatively, PO ciprofloxacin or other broad-spectrum antibiotics) in the setting of acute bleeding is recommended, including cirrhotic patients with upper GI bleeding with/without ascites.[8, 43] If oral administration is not possible, intravenous (IV) ciprofloxacin may be used. IV ceftriaxone should be considered in patients with advanced cirrhosis and in centers with documented quinolone-resistant bacteria.[8, 44]
Combination endoscopic and pharmacologic therapy minimizes the risk of complications, especially within the period when the risk of rebleeding is the greatest (ie, within 5 days of initial episode).
Beta-blocker therapy is not recommended in the setting of acute bleeding owing to its potential to cause hypotension, further diminishing the compensatory tachycardia to hemorrhage.
Somatostatin (not available in the United States) is an endogenous hormone that at pharmacologic doses decreases portal blood flow by splanchnic vasoconstriction, without significant systemic adverse effects.
Octreotide is the pharmacologic agent of choice in acute variceal bleeding and is used in conjunction with endoscopic therapy.[15, 45, 46] This agent is a synthetic analogue of somatostatin that is usually administered at a constant infusion of 50 mcg/h. Octreotide has been shown not only to be effective in reducing the complications of variceal bleeding after emergency sclerotherapy or variceal ligation, but it is also superior to vasopressin, particularly in its side effect profile.
Vasopressin is the most potent splanchnic vasoconstrictor; it reduces blood flow to all splanchnic organs, decreasing portal venous inflow and portal pressure. This agent should not be administered via a central line, especially in elderly patients or patients with coronary artery disease, because of possible coronary vasospasm and subsequent myocardial infarction (MI). Use of vasopressin is also limited by adverse effects related to splanchnic vasoconstriction (eg, bowel ischemia) and systemic vasoconstriction (eg, hypertension, myocardial ischemia). Continuous infusion of 0.2-0.4 IU/min (not to exceed 0.8 IU/min) is recommended.
Vasopressin always should be accompanied by intravenous nitroglycerin at a dose of 40 mcg/min (not to exceed 400 mcg/min) to maintain a systolic blood pressure of greater than 90 mm Hg. Adding nitrates to vasopressin therapy significantly improves efficacy, although the adverse effects of combination therapy are higher than those associated with terlipressin or somatostatin.
Terlipressin (not approved by the US Food and Drug Administration [FDA] for use in the United States) is a synthetic analogue of vasopressin that has longer biologic activity and significantly fewer adverse effects than vasopressin. Therapy should be continued for up to 5 days following the initial variceal hemorrhage to reduce the risk of recurrent bleeding.
A randomized, controlled trial showed that octreotide only transiently reduced portal pressure and flow, whereas the effects of terlipressin were sustained. These findings suggest that terlipressin may have more sustained hemodynamic effects in patients with bleeding varices.
Vasoactive drugs are safe and effective alternative therapy whenever endoscopic ligation (banding) therapy is not promptly available. In addition, pharmacotherapy seemed to create less adverse events than emergency sclerotherapy.
Perform endoscopy as soon as possible after the patient has been resuscitated. The aim is to establish the cause of and to control the bleeding, and endoscopic therapy has the advantage of allowing specific treatment to be provided at the time of diagnosis. This procedure has largely replaced balloon tamponade as the initial nonpharmacologic hemostatic modality for variceal bleeding.
Efficacy in achieving hemostasis is higher than 80% with endoscopic therapy, but its effectiveness declines to 70% at day 5 due to very early rebleeding in some patients. Failures in endoscopic treatment may be managed with a second session of such therapy, but no more than 2 sessions should be allowed before deciding to perform a transjugular intrahepatic portosystemic shunt (TIPS) procedure or surgery.
Endoscopic variceal ligation
EVL is the preferred endoscopic therapy in acute esophageal variceal bleeding.[8, 15] It has superiority over sclerotherapy in all major outcomes (recurrent bleeding, local complications including ulceration and stricture formation, time to variceal obliteration, and survival).
This procedure is based on the widely used technique of rubber-band ligation of hemorrhoids. EVL is performed using a banding device attached to the tip of the endoscope. The varix is aspirated into the banding chamber, and a trip wire dislodges a rubber band carried on the banding chamber, ligating the entrapped varix. One to 3 bands are applied to each varix, resulting in thrombosis. Ultimately, strangulation, sloughing, and fibrosis obliterate the varices.
EVL and sclerotherapy have achieved similar rates of initial hemostasis in patients whose varices were actively bleeding at the time of treatment. However, a meta-analysis of 10 randomized controlled trials patients showed an almost statistically signi?cant bene?t of EVL in the initial control of bleeding relative to sclerotherapy. Moreover, EVL requires fewer treatment sessions than does sclerotherapy, and local complications are less common with EVL than with sclerotherapy. For example, esophageal strictures have been found to be less common with EVL than with sclerotherapy, and rebleeding has occurred less frequently with the ligation technique (26%) than with sclerotherapy (45%).
Systemic complications, however, such as pulmonary infections and bacterial peritonitis, have not been found to significantly different between the 2 techniques. (However, a trend toward a decrease in these 2 complications in patients treated with ligation has been observed.) Moreover, there is no overall survival benefit to EVL over injection sclerotherapy.
A limitation of endoscopic ligation is that it requires placement of an opaque cylinder over the end of the endoscope, which decreases the endoscopic field of view and may allow pooling of blood. Thus, in patients with active bleeding, visualization may be impaired more with ligation than with sclerotherapy. In addition, EVL has the same limitations as injection sclerotherapy regarding availability, cost, and difficulty in treating gastric varices.
Although ligation has come to be considered the treatment of choice for esophageal varices, the choice of technique should hinge on the experience of the operator, as well as the particular circumstances found during endoscopic therapy.
Endoscopic injection sclerotherapy
Endoscopic injection sclerotherapy is a very effective emergency treatment for acute variceal bleeding (but it is not optimal for patients bleeding from gastric fundal varices).[8, 40] This procedure is an alternative procedure when EVL is not technically feasible, but sclerotherapy has higher complication rates relative to EVL.
Treatment involves injecting a sclerosant solution into the bleeding varix, obliterating the lumen by thrombosis, or into the overlying submucosa, producing inflammation followed by fibrosis. Several different sclerosants are available, including 5% sodium morrhuate, 1-3% sodium tetradecyl sulfate, and 5% ethanolamine oleate. The typical volume used per injection is 1-2 mL of sclerosant, with the total volume ranging from 10 to 15 mL.
Hemorrhagic control should be obtained with 1-2 sessions. Patients continuing to bleed after 2 sessions should be considered for alternative methods to control their bleeding. Note that HVPG remains elevated 5 days following sclerotherapy, whereas it returns to baseline 48 hours following endoscopic variceal ligation (EVL).[40, 51]
In the United States, sodium tetradecyl sulfate or sodium morrhuate has generally been used as a sclerosant, whereas polidocanol or ethanolamine has been more popular in Europe. Variations in the technique or the sclerosant used have not been shown to influence the outcome.
Complications of endoscopic injection sclerotherapy, which are more frequent in acute bleeding than in elective situations, are related to the toxicity of the sclerosant and include transient fever, esophageal stricture formation, dysphagia, esophageal perforation (rarely), chest pain, mediastinitis, mucosal ulceration, and pleural effusion.[8, 15] Serious complications related to sclerotherapy have been reported in 15-20% of patients.
Approximately 5-10% of patients with esophageal variceal hemorrhage have conditions that cannot be controlled by endoscopic and/or pharmacologic treatment. Balloon tamponade (eg, Minnesota tube, Sengstaken-Blakemore [S-B] tube, Linton-Nachlas tube) may be used in the management of these patients; it achieves hemostasis in 60-90% of variceal bleedings.
However, this technique should be employed only in patients with massive bleeding and should serve only as a temporizing measure (should be used for < 24 h owing to risk of esophageal rupture/necrosis) (ie, bridging therapy) until definitive treatment (eg, TIPS, surgical intervention) can be instituted.
Moreover, balloon-tube tamponade must be performed by experienced personnel because the procedure is potentially dangerous. An endotracheal tube should be placed to protect the airway before attempting to place the balloon tube.
Complications of balloon-tube tamponade are esophageal and gastric ulceration, aspiration pneumonia, and esophageal perforation. Continued bleeding during balloon tamponade indicates an incorrectly positioned tube or bleeding from another source.
The Minnesota tube is an adaptation of the S-B tube, the difference being that the S-B tube does not have an esophageal suction port to prevent aspiration. The Minnesota tube has 4 lumens, including 1 for gastric aspiration, 2 to inflate the gastric and esophageal balloons, and 1 above the esophageal balloon to suction secretions in order to prevent aspiration. The tube is inserted through the mouth, and its position within the stomach is checked by auscultation while air is injected through the gastric lumen.
The gastric balloon is inflated with 200 mL of air. Once fully inflated, the gastric balloon is pulled up against the gastroesophageal junction, using approximately 0.5 kg of traction, compressing the submucosal varices. The esophageal balloon rarely is required.
Percutaneous transhepatic embolization (PTE) of gastroesophageal varices involves catheterization of the gastric collaterals that supply blood to varices via the transhepatic route. A variety of agents had been used, with varying degrees of success in controlling acute bleeding.
Generally, PTE is less effective than endoscopic sclerotherapy for treatment of variceal hemorrhage, and it is much less effective compared with medical and surgical options. Thus, PTE should be reserved for situations in which acute variceal bleeding is not controlled by pharmaceutical treatment, endoscopic sclerotherapy, or endoscopic variceal ligation and in which contraindications for surgical management are present.
Endoscopic administration of cyanoacrylate monomer (superglue) in gastric varices is another intervention. The occurrence of complications after gastric variceal obliteration with butyl cyanoacrylate is low, with a complication-related mortality rate of 0.53%.
TIPS is a useful procedure for patients in whom bleeding has continued despite medical and endoscopic treatment, for patients with Child class C disease, and for selected patients with Child class B disease. It is effective only in portal hypertension of hepatic origin.
Under local anesthesia, with sedation via the internal jugular vein, the hepatic vein is cannulated and a tract is created through the liver parenchyma, from the hepatic to the portal vein, with a needle. This is performed under ultrasonographic and fluoroscopic guidance. The tract is dilated, and an expandable metal stent is introduced, connecting the hepatic and portal systems. Blood from the hypertensive portal vein and sinusoidal bed is shunted to the hepatic vein.
An international, multicenter study revealed that patients who received a TIPS early (within 72 h of presentation) had a significantly better chance of remaining free from bleeding than did patients who received standard care (fluid resuscitation, antibiotic prophylaxis, vasoactive drugs, early endoscopy with ligation, or sclerosis of varices).
Accepted indications for a TIPS procedure (in which the efficacy of TIPS has been established in controlled trials) include: (1) active variceal bleeding despite emergency endoscopic and/or pharmacologic treatment and (2) recurrent variceal bleeding despite adequate endoscopic treatment.
Potential indications (in which the efficacy of the TIPS procedure has been proven but has not been adequately compared with that of existing therapies) include: (1) isolated bleeding from gastric fundic varices and (2) refractory ascites.
Experimental indications (in which efficacy has not been established in large-scale trials) include the following:
Recurrence of portal hypertension
Causes of recurrent portal hypertension and bleeding after a TIPS procedure include the following:
TIPS complications related to technique include the following:
TIPS complications related to portosystemic shunting include: (1) hepatic encephalopathy (approximately 30%), (2) increased susceptibility to bacteremia, and (3) liver failure. Other complications may include TIPS-associated hemolysis (approximately 10%), stent infection, hemorrhage, and acute kidney injury associated with intraprocedural intravenous contrast administration.
In patients with small varices (< 5 mm or minimally elevated veins above the esophageal mucosal surface), surveillance is preferred over other therapeutic modalities. In patients with medium to large varices (>5 mm or esophageal vein raised beyond mucosal surface occupying esophageal lumen) without a high risk of bleeding, a nonselective beta-blocker is the preferred first line treatment, although esophageal varices ligation (EVL) may be considered.[8, 12]
Patients at high risk for bleeding have large varices, red wale markings on the varices, and severe liver failure; either nonselective beta-blockers or EVL can be used as the primary prophylaxis.[3, 8, 12, 55, 56, 57, 58]
In patients with medium or large varices with bleeding stigmata regardless of the size, and patients with decompensated cirrhosis, nonselective beta-blockers are preferred as they have been shown to decrease the number of bleeding episodes. If contraindications, patient intolerance, or patient noncompliance exist regarding the use of nonselective beta-blockers, EVL should be considered.[8, 12]
If patients are on selective beta-blocker (eg, atenolol, metoprolol) for other indications, switching to a nonselective beta-blocker (eg, propanolol, nadolol) is necessary. Selective beta-blockers have been shown to be less effective than nonselective beta-blockers for primary prophylaxis of variceal hemorrhage.
All patients with liver cirrhosis should undergo a screening upper gastrointestinal (GI) endoscopy to determine their risk for bleeding. Patients without varices should have a follow-up upper GI endoscopy (surveillance esophagogastroduodenoscopy [EGD]) after 2 years, or sooner if they have signs of clinical decompensation (see Upper Gastrointestinal Endoscopy). Nonselective beta-blockers may be considered in those with decompensated cirrhosis (particularly when compliance with EGD surveillance is a concern), but these agents are not recommended in patients with compensated cirrhosis.
Patients with small varices should have repeat endoscopy annually[8, 59] ; surveillance EGD is preferred over beta-blockers as they do not prevent the development of varices and are associated with a higher incidence of adverse effects. However, because EGD requires sedation and is expensive, it is not mandatory in patients who are already on nonselective beta-blockers for other reasons.
Noncardioselective beta-blockers are used most commonly for primary prophylaxis of variceal bleeding, and they include propranolol and nadolol. These nonselective beta-blockers reduce portal and collateral blood flow as well as have smaller effects on the increase in portal resistance and decrease on portal pressure. Reduction in cardiac output (via blockade of beta1 adrenoreceptors) occurs, as does splanchnic vasoconstriction (via blockade of vasodilatory adrenoreceptors of the splanchnic circulation).[3, 55, 56, 57, 58]
Nonselective beta-blockers have been shown to prevent decrease the risk of initial bleeding by approximately 45-50%, and they reduce bleeding in more than 50% of patients with medium or large varices. A meta-analysis of 11 trials evaluating nonselective beta-blockers in the prevention of first variceal bleeding showed that the bleeding rate in controls (25%) was significantly reduced (to 15%) in patients treated with beta-blockers after a median follow-up of 24 months. The mortality rate also was lower in the beta-blocker group; however, the difference did not achieve statistical significance.
In the same meta-analysis, evaluation of the effects of beta-blockers as a function of variceal size showed the risk of first variceal bleeding in patients with medium to large varices was 30% in controls and 14% in patients treated with beta-blockers. In patients with small varices, a tendency existed for reduction in the first bleeding episode; however, the number of patients and the rate of first bleeding were too low to achieve statistical significance.
Propranolol is administered at a dose of 20 mg every 12 hours, which is increased or decreased every 3-4 days until a 25% reduction in the resting heart rate occurs or the heart rate is down to 55 beats per minute (bpm). The average dose of propranolol is usually 40 mg twice daily. Administering more than 320 mg/day is not recommended. Nadolol dosing is half the daily dose of propranolol, administered once daily.
Reduction in heart rate may not lead to a reduction in the hepatic venous pressure gradient (HVPG); therefore, it is recommended that the dose should be titrated to the maximal tolerable dose until any adverse effects develop. Response to treatment is monitored by a reduction of the portal pressure gradient by more than 20% of the baseline value or less than 12 mm Hg. However, checking the HVPG response in primary prophylaxis is not mandatory, because 60% of patients who do not achieve these targets still do not bleed at 2-year follow-up evaluations.
Propranolol is contraindicated in patients with asthma, chronic obstructive pulmonary disease (COPD), atrioventricular (AV) block, intermittent claudication, and psychosis. The most frequent adverse effects are lightheadedness, fatigue, dyspnea upon exertion, bronchospasm, insomnia, impotence, and apathy. Reducing the dose of propranolol frequently controls these adverse effects.
Beta-blockers are best continued for the patient's lifetime, because the risk of variceal hemorrhage returns to that of the untreated population once beta-blockers are withdrawn.
The vasodilator isosorbide mononitrate (ISMN) may be considered as a second-line agent for secondary prophylaxis for variceal bleeding. The available evidence does not support the use of this agent as monotherapy for primary prophylaxis, even in patients with contraindications or intolerance to beta-blockers. One study reported ISMN to be as effective as propranolol in preventing first variceal bleeding, but long-term follow-up showed a higher mortality rate in patients older than 50 years in the ISMN group.
Vasodilators also reduce esophageal variceal pressure. The primary concern in patients with advanced cirrhosis is that vasodilators can reduce arterial blood pressure and promote the activation of endogenous vasoactive systems that may lead to sodium and water retention. Although ISMN has been demonstrated to reduce HVPG markedly in acute administration, it provides significantly less reduction after long-term administration (due to probable development of patient tolerance).
Although many authorities recommended a combination therapy of pharmacologic treatment and EVL as the first-line treatment for secondary prophylaxis,[8, 12, 20] emerging evidences suggests that EVL alone is as effective as the combination therapy.
Theoretically, combination therapy with beta-blockers and ISMN should offer better reduction in portal pressure, but this has not shown statistical significance in preventing rebleeding episodes in the clinical setting. A large, double-blind, placebo-controlled trial was unable to demonstrate a significantly lower rate of first hemorrhage in the group treated with combination therapy versus those given beta-blockers alone. In addition, combination therapy appears to be associated with increased adverse effects and a higher rate of ascites.[8, 12]
Combination therapy cannot be recommended presently until further studies prove its efficacy. However, addition of ISMN should be considered when single pharmacotherapy fails.
Sclerotherapy has no role in primary prophylaxis. Randomized, controlled trials investigating the use of sclerotherapy for primary prophylaxis produced divergent results, with some studies showing a worse outcome in patients who underwent this therapy than in controls. Moreover, combination treatment with sclerotherapy and nonselective beta-blockers offer no advantages over the use of beta-blockers alone for the prevention of esophageal variceal hemorrhage.
Prophylactic EVL currently cannot be recommended as a routine measure for primary prevention as it offers no advantage over the use of beta-blockers alone for preventing esophageal variceal bleeding. However, this procedure may be an option in secondary prophylaxis for patients with grade 3 varices who have contraindications to or cannot tolerate beta-blockers.[8, 61]
EVL has been demonstrated to be more effective than the administration of no treatment in preventing a first variceal bleed. It has also been shown to have an efficacy similar to that of beta-blockers in the prevention of first variceal bleeds, but with increased adverse effects. (However, a study by Lay et al suggested that EVL is as safe as propranolol therapy in primary prophylaxis. The study involved 100 patients with cirrhosis—of whom half were treated with EVL and half received propranolol—who were at high risk of variceal bleeding. )
The goals of pharmacotherapy are to reduce mortality and morbidity, and prevent complications associated with acute bleeding related to portal hypertension. Two main categories of drugs, vasoconstrictors and vasodilators, are used.
The main advantages to using vasoactive agents include the ability of these drugs to treat variceal bleeding in the emergency department, lower portal pressure, and offer the endoscopist a clearer view of varices because of less active bleeding. Vasoactive agents represent an ideal treatment for sources of portal hypertensive bleeding other than esophageal varices (eg, gastric varices >2 cm below the gastroesophageal junction or portal hypertensive gastropathy).[8, 15]
The vasoconstrictors somatostatin and octreotide are used to treat acute bleeding in patients with portal hypertension before performing endoscopy.[45, 49] Intravenous infusions of octreotide will lower portal blood pressure and can prevent rebleeding during the patient's initial hospitalization. Vasodilators such as isosorbide mononitrate (ISMN) reduce intrahepatic vascular resistance without decreasing peripheral or portal-collateral resistance.
Beta-blockers, which include propranolol, nadolol, and timolol, are used to provide primary and secondary prophylaxis. Beta-blockers lower cardiac output (via blockade of beta1 adrenoreceptors) and cause splanchnic vasoconstriction (via blockade of vasodilatory adrenoreceptors of the splanchnic circulation), reducing portal and collateral blood flow.
Clinical Context: Octreotide, a synthetic octapeptide, acts primarily on somatostatin receptor subtypes II and V. It inhibits growth hormone secretion and has a multitude of other endocrine and nonendocrine effects, including the inhibition of glucagon, vasoactive intestinal peptide, and GI peptides. Octreotide has greater potency and a longer duration of action than somatostatin.
Somatostatin, an orphan drug, is a naturally occurring tetradecapeptide isolated from the hypothalamus and from pancreatic and enteric epithelial cells. Through vasoconstriction, somatostatin diminishes blood flow to the portal system, thus decreasing variceal bleeding. It has effects similar to those of vasopressin but does not cause coronary vasoconstriction. Somatostatin has an initial half-life of 1-3 minutes and is rapidly cleared from the circulation.
Somatostatin analogs inhibit the secretion of hormones involved in vasodilation. Octreotide is a synthetic octapeptide. Compared with somatostatin, octreotide has similar pharmacologic actions with greater potency and longer duration of action. In the US, octreotide is used off-label for the management of variceal hemorrhage.
Clinical Context: Propranolol is a noncardioselective beta-blocker that reduces portal pressure through the reduction of portal and collateral blood flow. It competes with adrenergic neurotransmitters (eg, catecholamines) at sympathetic receptor sites. Similar to atenolol and metoprolol, propranolol blocks sympathetic stimulation mediated by beta1-adrenergic receptors in the heart and vascular smooth muscles.
Clinical Context: Nadolol is a noncardioselective beta-blocker that reduces portal pressure through the reduction of portal and collateral blood flow.
Clinical Context: Timolol is a noncardioselective beta-blocker that reduces portal pressure through the reduction of portal and collateral blood flow.
Nonselective beta-blocking agents decrease hepatic arterial and portal venous perfusion. Beta-adrenergic blockers may block the effect of vasodilators, decrease platelet adhesiveness and aggregation, and increase the release of oxygen to tissues.
Nonselective beta-blockers have been shown to prevent bleeding in more than 50% of patients with medium or large varices. These agents exert a moderate effect on the reduction of portal flow, and smaller effects on the increase in portal resistance and decrease on portal pressure.
Propranolol is used off-label for primary prophylaxis — in combination with endoscopic variceal ligation (EVL) — for esophageal varices. This agent is also indicated for secondary prophylaxis for esophageal varices.
Clinical Context: Vasopressin has vasopressor and antidiuretic hormone (ADH) activity. It increases water resorption at the distal renal tubular epithelium (ADH effect) and promotes smooth muscle contraction throughout the vascular bed of the renal tubular epithelium (vasopressor effects). However, vasoconstriction is also increased in splanchnic, portal, coronary, cerebral, peripheral, pulmonary, and intrahepatic vessels. Vasopressin decreases portal pressure in portal hypertension.
A notable adverse effect of this agent is coronary artery constriction, which may dispose patients with coronary artery disease to cardiac ischemia. This can be prevented with the concurrent use of nitrates. Vasopressin is rarely used.
Vasoconstrictors reduce portal blood flow and/or increase resistance to variceal blood flow inside the varices. Therefore, these drugs reduce blood flow in the gastroesophageal collaterals because of their vasoactive effects on the splanchnic vascular system. When used in combination with nitrates, the efficacy and safety of vasoconstrictors have been shown to improve. However, their use may be limited as the risk of adverse events is higher with combination therapy.
In the US, vasopressin is used off-label for the management of acute variceal bleeding.
Terlipressin is widely used in Europe but has not received FDA approval for use in the United States. This is a synthetic analogue of vasopressin. It is the only pharmacologic agent shown to reduce mortality from variceal bleeding. Terlipressin has longer biologic activity than vasopressin. It significantly reduces portal and variceal pressure and azygos flow. The drug is beneficial when combined with sclerotherapy. Terlipressin also has the advantage of preserving renal function, which is a particularly important feature in patients with cirrhosis.
Clinical Context: Nitroglycerin causes relaxation of vascular smooth muscle by stimulating intracellular cyclic guanosine monophosphate production. The result is a decrease in blood pressure.
Vasodilators have been shown to exert a small effect on the reduction of portal flow, an increase in portal resistance, and decrease on portal pressure. These agents reduce intrahepatic vascular resistance without decreasing peripheral or portal-collateral resistance.
Nitrates, however, technically work by decreasing resistance. They decrease portal flow by decreasing mean arterial pressure. Oral nitroglycerin is used off-label for the management of variceal bleeding.
Power Doppler sonogram through the spleen shows varices at the hilum of an enlarged spleen. The final diagnosis was hepatitis C cirrhosis, hepatocellular carcinoma of the left hepatic lobe (which had ruptured into the peritoneum), and portoarterial fistula (which had developed inside the ruptured tumor, giving rise to severe portal hypertension).
Duplex spectral Doppler sonogram of the portal vein (same patient as in the previous image) shows a bidirectional flow within the vein. The final diagnosis was hepatitis C cirrhosis, hepatocellular carcinoma of the left hepatic lobe (which had ruptured into the peritoneum), and portoarterial fistula (which had developed inside the ruptured tumor, giving rise to severe portal hypertension).
Digital subtraction selective common hepatic artery angiogram shows immediate filling of the portal venous radicles in the left lobe of the liver (straight arrow) and early filling of portal vein (curved arrow), suggestive of hepatic arterial-portal vein fistula. The final diagnosis was hepatitis C cirrhosis, hepatocellular carcinoma of the left hepatic lobe (which had ruptured into the peritoneum), and portoarterial fistula (which had developed inside the ruptured tumor, giving rise to severe portal hypertension).
Delayed venous phase of a selective common hepatic angiogram (same patient as in the previous image) shows the portal vein (P), with filling of the coronary vein caused by retrograde flow feeding gastric and lower esophageal varices (arrows). Retrograde flow in enlarged umbilical veins also is seen. The final diagnosis was hepatitis C cirrhosis, hepatocellular carcinoma of the left hepatic lobe (which had ruptured into the peritoneum), and portoarterial fistula (which had developed inside the ruptured tumor, giving rise to severe portal hypertension).
Digital subtraction venous phase of a superior mesenteric artery angiogram (same patient as in the previous 2 images) shows retrograde flow into the coronary vein (curved arrow) and the inferior mesenteric vein (straight arrow). Note the flow defect of the distal portal vein caused by retrograde flow (open arrowhead). The final diagnosis was hepatitis C cirrhosis, hepatocellular carcinoma of the left hepatic lobe (which had ruptured into the peritoneum), and portoarterial fistula (which had developed inside the ruptured tumor, giving rise to severe portal hypertension).
Etiology of Portal Hypertension WHVP FHVP HVPG Prehepatic Normal Normal Normal Intrahepatic Presinusoidal Normal Normal Normal Sinusoidal Increased Increased Increased Postsinusoidal Increased Normal Increased Posthepatic Budd-Chiari syndrome N/A Hepatic vein cannot be cannulated N/A Other posthepatic causes Increased Increased Normal FHVP = free hepatic venous pressure; HVPG = hepatic venous pressure gradient; N/A = not applicable; WHVP = wedged hepatic venous pressure.