Medical management of bleeding varices: Primary and secondary prophylaxis for variceal bleeding

Introduction

The presence of portal hypertension is a prerequisite for the development of, and bleeding from, varices. Portal hypertension is defined by a portal venous pressure exceeding the inferior vena cava pressure by more than 5 mm Hg (see Chapters 74 , 76 , and 79 ). Measuring the free hepatic vein pressure and subtracting this from the wedged hepatic vein pressure, commonly known as the hepatic venous pressure gradient (HVPG), is the most common approach for establishing an increased gradient. A normal HVPG is between 3 and 5 mm Hg. An HVPG greater than 10 mm Hg is designated “clinically significant portal hypertension” (CSPH) With a gradient greater than 10 mm Hg. CSPH is a harbinger for the development of varices and gradients greater than 10 correlate with bleeding risk. Varices may occur anywhere in the portal venous circulation with esophageal varices the most common and the most likely to bleed. The likelihood of esophageal varices being present runs parallel to the severity of liver disease, as measured by Child-Turcotte-Pugh (CTP) classification; 43% of CTP class A and 75% of CTP class C have esophageal varices (see Chapters 4 , 5 , and 79 ).

Although varices generally develop in the presence of cirrhosis, varices may also develop in the absence of intrinsic liver disease. This most often occurs when obstruction to the portal vein, splenic vein, or another large mesenteric vein occurs generally from thrombosis, most often secondary to phlebitis following an infection (see Chapter 74 ). Table 80.1 details the causes of portal hypertension. In those with liver disease, the likelihood of variceal bleeding increases as liver function declines. In addition, the presence of other complications of cirrhosis, like spontaneous bacterial peritonitis (SBP), hepatocellular carcinoma, and portal vein thrombosis (PVT) independently increase the risk for bleeding form varices.

TABLE 80.1

Locations and Pathologies Leading to Portal Hypertensions

LOCATION	PATHOLOGY
Prehepatic	Portal vein thrombosis Mesenteric or splenic vein thrombosis Congenital stenosis of the portal vein Extrinsic portal vein compression Arteriovenous fistula of the mesenteric circulation Splenomegaly
Intrahepatic-presinusoidal	Schistosomiasis Hepatoportal sclerosis Primary biliary cirrhosis Sarcoidosis Sclerosing cholangitis Congenital hepatic fibrosis Hepatic arterioportal fistula Intrahepatic portal vein obstruction (e.g., tumor or cyst associated)
Intrahepatic-sinusoidal	Cirrhosis Nodular regenerative hyperplasia Steatohepatitis Acute hepatitis Toxin (e.g., arsenic, vinyl chloride)
Intrahepatic-Postsinusoidal	Budd-Chiari syndrome Sinusoidal obstruction syndrome/veno-occlusive disease (SOS/VOD)
Extrahepatic	Restrictive cardiomyopathy Constrictive pericarditis Congenital heart disease Pulmonary hypertension

When bleeding occurs, the event contributes morbidity to an often already compromised state. Bleeding events in patients with end-stage liver disease are common causes of acute-on-chronic liver failure (ACLF), and each bleeding event contributes independently to mortality. The risk for mortality associated with a bleed is greatest in those bleeding with higher Model for End-Stage Liver Disease (MELD) scores, and historically the likelihood for mortality approached 30% to 50% (see Chapter 4 ). Current data fortunately illustrate the 6-week mortality following a portal hypertensive bleed to be substantially less than this. Changes to several aspects of clinical care and near universal adoption of standardized evidence-based approaches to acute event management and to secondary prophylaxis has likely led to reductions in mortality. The current 6-week mortality risk following a variceal bleed is now approximately 15%.

In studies designed specifically to assess whether timing of endoscopy impacts outcome with acute bleeding events, early endoscopy, defined as occurring in less than 6 hours, has not shown diminished mortality, although early endoscopy may diminish cost. Endoscopy in most instances should be undertaken within 24 hours of hospitalization and following the establishment of hemodynamic stability (see Chapter 81 ). There are some instances during which early endoscopy (i.e., within 6 hours) may benefit particular subgroups at high risk for early mortality and in which certain therapeutic options may be limited. Examples of such patient groups include those with a MELD score greater than 18, those with a CTP score greater than 12, those with unfavorable anatomy for interventional radiology, and those with significant premorbid encephalopathy. ^, Endoscopic management of acute bleeding from esophageal varices is accomplished with band ligation, and Following the initial endoscopy, timely follow-up endoscopy at 2 weeks is important along with repeat endoscopy preferably every 2 weeks until the varices are completely ablated. Best illustrating the role and success of this and other available medical therapies in the primary and secondary prophylaxis against variceal bleeding are diminishing mortality rates in patients with cirrhosis and the decline in liver transplant waiting list mortality despite rising MELD scores in those listed for transplantation. ^, These findings also underscore the utility of an early diagnosis of esophageal varices through aggressive screening strategies, careful surveillance for small varices, and the role of prophylactic therapies with nonselective β-blockade and/or band ligation in the management of varices (see Chapter 81 ).

Natural history of varices

The diagnosis of esophageal varices should be considered in any patient with cirrhosis (see Chapters 5 and 74 ). The likelihood of discovering varices in any patient correlates with the severity of the underlying liver disease. The prevalence of esophageal varices among patients with compensated cirrhosis is about 30% to 40%, whereas it approaches 85% among patients with decompensated cirrhosis. In patients with compensated cirrhosis without varices, esophageal varices develop at a rate of 7% to 8% per year and progress from small to large varices at a rate of 10% to 12% per year. The propensity for developing varices increases with the development of clinical decompensation. The incidence of bleeding in those with varices is approximately 10% to 15% per year in patients with high-risk varices. Factors independently associated with variceal bleeding, beyond liver disease severity, include the size of the varices, coloring on the varices, a history of previous bleeding, HVPG, ACLF, PVT, liver stiffness measured by transient elastography, and active alcohol use. ^, Bleeding from varices is rare if the HPVG is less than 12 mm Hg. ^, Albeit uncommonly measured endoscopically, the degree of pressure within the varix and the wall tension of the varix correlate with the probability for bleeding. Figure 80.1 illustrates the evolution of portal hypertension complication.

FIGURE 80.1, Evolution of portal hypertension complication.

Management of the acute hemorrhage (see Chapter 81 )

When bleeding occurs, resuscitation is immediately initiated with care to avoid over-resuscitation. Excessive resuscitation increases plasma volume and thereby portal pressure, which may contribute to ongoing bleeding risk. ^, However, adequate and timely resuscitation are critical to diminish the likelihood of developing an acute kidney injury. For transfusion, in those with cirrhosis, a hemoglobin threshold of 7 g/dL is suggested. With active bleeding, in particular when hemodynamic instability is present, immediate cross-matching is imperative and earlier initiation of transfusion will often occur. However, a restrictive transfusion strategy should be employed with a hematocrit or hemoglobin target of 21% to 24% or 7 to 8 g/dL, respectively. These targets, in high-quality studies, are associated with increased survival and lower HVPGs when compared with liberal transfusion strategies. ^, Patients with portal hypertension have preexisting hypotension and in resuscitating those with bleeding a target systolic blood pressure between 90 and 100 mm Hg and target heart rate less than 100 beats/min are recommended. Because of the use of nonselective β-blockers (NSBBs) in some, tachycardia is not always present in these patients. Accordingly, targets for resuscitation during an active bleed include restoration of the hematocrit to 24%, maintaining fibrinogen levels greater than or equal to 120 mg/dL, where possible increase platelet counts to 50,000, administer intravenous antibiotics and vasoactive medication (octreotide or terlipressin) during the resuscitation, and undertake endeavors to optimize renal function. Interestingly, across the majority of studies examining variceal bleeding, the use of therapeutic anticoagulation before a bleeding event was not identified as an independent risk factor variceal hemorrhage, although no study has been performed to examine bleeding risk in those on anticoagulation in contrast to peptic ulcer bleeding. Anticoagulation is less commonly used in those with cirrhosis.

Sites and the propensity to variceal bleeding

Although variceal bleeding is most common in cirrhotic patients, it is important to remember other etiologies, such as peptic ulcer disease, are also common causes of acute upper gastrointestinal bleed. When variceal in origin, the most common source is esophageal varices (70%) and the most common location is just above the gastroesophageal junction. Signs of the site of bleeding include a clot overlying a varix, a white plug also known as a fibrin plug on the varix, or active extravasation of blood from the side of a varix during endoscopy. It is not uncommon, however, to fail to identify the specific site of bleeding at the time endoscopy. In instances where an acute gastrointestinal bleed occurs, in the presence of varices and in the absence of an alternative site to explain blood loss, band ligation starting at the gastrointestinal junction should occur when varices are present. It is also important to note that when bleeding occurs as an initial event, and in the absence of other complications of cirrhosis, the 5-year mortality is relatively low at 20%. In contrast, when bleeding occurs in the presence of other complications of cirrhosis, the transplant-free survival falls to less than 20% at 5 years.

Gastric varices are present in 20% of patients with cirrhosis. They bleed less commonly than esophageal varices, and a greater likelihood for these being present occurs with aberrancies to the portal circulation (for example, splenic vein thrombosis). Gastric varices account for 5% to 10% of all variceal hemorrhages and bleeding from these is associated with significantly greater mortality than esophageal varices. In contrast, rectal varices are far more common than gastric varices and are present in 44% of patients with cirrhosis and in 59% with a history of esophageal varices (59%); however, they only account for less than 5% of all portal hypertensive bleeds.

Ectopic varices are an uncommon cause of variceal bleeding and account for up to 5% of all bleeding events. Ectopic varices that are associated with bleeding are most often located in a stoma or surgical anastomosis and then, in decreasing order, the duodenum, jejunum/ileum, rectum, and colon. ^, Ectopic varices are more common when extrahepatic thrombosis within the portal circulation occurs. Portal biliopathy, an uncommon clinic complication of portal hypertension, occurs most commonly when cirrhosis exists in the presence of an extrahepatic PVT and these rarely bleed in the absence of a surgical intervention such as an endoscopic retrograde cholangiopancreatography (ERCP). Finally, another rare site of ectopic bleeding associated with portal hypertension includes vessels in the peritoneal cavity or those associated with a recanalized umbilical vein. ^, Bleeding from these sites is very rare and generally associated with trauma or iatrogenic injury. The presence of varices in these locations should be considered when planning a procedure, such as a paracentesis; however, the presence of varices will rarely preclude a procedure.

Stages of portal hypertension

Cirrhosis, which is often the consequence of a chronic and progressive liver disease, is the most common cause of portal hypertension (see Chapter 74 ). After developing cirrhosis, patients move through stages of portal hypertension. Table 80.2 outlines the stages of portal hypertension and details the clinical objectives for each stage. For all chronic liver diseases, the primary objective is to arrest and, where possible, reverse the process that caused cirrhosis. Examples of this include eradication or suppression of viral hepatitis in those with hepatitis C or hepatitis B (see Chapter 68 ), abstinence from alcohol in those with alcoholic liver disease, and weight loss in those with nonalcoholic steatohepatitis (see Chapter 69 ). When portal hypertension is present (HVPG >5 mm Hg), patients are initially compensated, meaning cirrhosis in the absence of any decompensating event (ascites, encephalopathy, variceal hemorrhage) and without CSPH (HVPG 5–10 mm Hg) or with CSPH (HPVG >10 mm Hg). The final phase is decompensated cirrhosis, and commonly this population has an HPVG greater than 12 mm Hg. In each phase of portal hypertension associated with cirrhosis the therapeutic objectives and approaches differ.

TABLE 80.2

Stages of Portal Hypertension in Cirrhosis, Clinical Manifestations, and Goals of Therapy Disease

Adapted from Chandra R, Kapoor D, Tharakan A, Chaudhary A, Sarin SK. Portal biliopathy. J Gastroenterol Hepatol . 2001;16(10):1086–1092.

DISEASE STAGE	COMPENSATED CIRRHOSIS			DECOMPENSATED CIRRHOSIS
HVPG	<10 mm Hg	≥10 mm Hg (CSPH)		≥12 mm Hg
Varices	Absent	Absent	Present	Present
Complications of portal hypertension	Absent	Absent	Absent	Acute variceal bleed	Prior variceal bleed without complications	Prior variceal bleed with complications
Goals of therapy	Prevent CSPH	Prevent decompensation	Prevent decompensation	Control bleed	Prevent further bleeds and decompensation	Prevent further decompensation and mortality

CSPH, Clinically significant portal hypertension; HVPG, hepatic venous pressure gradient.

Compensated cirrhosis without clinically significant portal hypertensions

In those with compensated cirrhosis without CSPH the goal of therapy is largely confined to management of the underlying cause of cirrhosis. In compensated cirrhosis without CSPH, and in contrast with those with cirrhosis and CSPH, patients have a normal cardiac index as opposed to an increased cardiac index, which is seen in decompensated disease. Accordingly, those with compensated cirrhosis without CSPH (HPVG <10) obtain no benefit from medical treatments for portal hypertension as the treatments are designed to diminish the cardiac output; in patients with cirrhosis without CSPH the cardiac output is normal.

Compensated cirrhosis with clinically significant portal hypertension

Identifying those with CSPH commences with the clinical examination. CSPH is present in all patients with signs of decompensation (ascites, hepatic encephalopathy) and in those clinical signs of portal hypertension (splenomegaly, caput medusae). In these patients variceal screening is always indicated. In those with cirrhosis and without signs of decompensation or portal hypertension, a liver stiffness measurement less than 20 kPa by transient elastography and platelet count greater than 150,000/mm ³ defines a group unlikely to have CSPH and patients not requiring screening endoscopy. ^, However, all patients with CSPH (HPVG >10 mm Hg), and in all with a liver stiffness >20 kPa and platelet count less than 150,000/mm, screening endoscopy should be offered.

The gold standard for screening for esophageal varices is upper gastrointestinal endoscopy. Albeit other tests may provide similar information, like capsule endoscopy or endoscopic ultrasound (EUS), which both have a limited role, although EUS has some very specific therapeutic roles in the management of gastric and ectopic varices. ^, In all patients with evidence of CSPH, an ultrasound of the liver with Doppler of the portal vein is indicated. Whenever possible, a triple-phase contrast cross-sectional imaging study of the abdomen should be considered to evaluate patients with portal hypertension and, in all patients with PVT, a contrast imaging study is required to define the portal venous anatomy. Contrast imaging studies better exclude pathologies that might lead to thrombosis, like hepatocellular carcinoma. The contrast imaging studies also provide a road map for potential intervention in instances where interventional radiology may be considered to manage refractory bleeding or portal vein recanalization.

In those with cirrhosis with CSPH and without varices, timolol, an NSBB, was examined in a multicenter study by the Portal Hypertension Collaborative Group. This placebo-controlled study failed to demonstrate a statistical difference in the development of varices between groups, although it did show a greater number of serious adverse reactions in the treatment arm when compared with the placebo. Accordingly, NSBBs should not be given to patients with cirrhosis and without varices. ^,

You're Reading a Preview

Become a Clinical Tree membership for Full access and enjoy Unlimited articles

Become membership

If you are a member. Log in here