General Considerations

Chronic liver disease and cirrhosis are the 12th leading causes of mortality in the United States. Portal hypertension and its consequences are progressively debilitating complications of cirrhosis (Table 48–1). Variceal hemorrhage, spontaneous bacterial peritonitis, and the hepatorenal syndrome are chiefly responsible for the high morbidity and mortality rates in patients with cirrhosis.

Esophageal varices develop at a rate of 5–8% per year in patients with cirrhosis and portal hypertension, and up to 80% of patients with cirrhosis will eventually develop this complication. Variceal hemorrhage occurs in 25–35% of patients with cirrhosis and large esophagogastric varices. The majority of bleeding episodes occur within the first year of diagnosis of varices. Bleeding from esophageal varices is associated with 15–20% early mortality and accounts for one-third of all deaths. If no long-term therapy is instituted after control of acute hemorrhage, 60–70% of patients will experience recurrent variceal hemorrhage. Most of these episodes occur within 6 months of the index bleed.

Pathogenesis

Portal hypertension develops as a result of two main factors: (1) an increase in intrahepatic resistance, and (2) an increase in portal blood flow. In cirrhosis, the initiating event is an increase in hepatic and portocollateral resistance. The increased resistance occurs, in part, from sinusoidal encroachment, collagen deposition, vascular tree pruning, and nodular regeneration. These elements, together with the overexpression of endogenous vasoconstrictors (eg, endothelins and leukotrienes) and the underproduction of endogenous vasodilators (primarily nitric oxide), are responsible for the increase in intrahepatic and portocollateral resistance. Systemic vasodilation follows with the increased release of neurohormonal vasodilators (eg, nitric oxide, glucagon, tumor necrosis factor, prostaglandins, and other cytokines) and results in a hyperdynamic circulatory system. This is further complicated by angiogenesis, which increases splanchnic blood flow, exacerbates portal pressure elevation, induces neovascularization, and enhances the development of portosystemic collateral circulation including the development of esophageal varices. These factors also lead to the development of nonvariceal complications of portal hypertension including the development of ascites, hydrothorax, and the hepatorenal syndrome. The goal of therapy is to interrupt the process by decreasing portal venous blood flow and/or intrahepatic and portocollateral resistance.

Natural History of Cirrhosis

Cirrhosis is no longer considered a static entity but may progress from a compensated state to a decompensated state, with the prognosis varying with the stage. This progression is dynamic, not necessarily relentless and potentially reversible. In a classic review by D’Amico et al, involving 843 patients, the estimated 1-year mortality was 1% for stage 1, 3.4% for stage 2, 20% for stage 3, and 57% for stage 4 (Figure 48–1). Median survival for compensated cirrhosis was >12 years compared to 2 years for compensated cirrhosis.

Clinical Findings

A. Symptoms and Signs

Patients with cirrhosis have symptoms that are nonspecific for the presence of portal hypertension. Physical findings in cirrhosis that may suggest the presence of portal hypertension include muscle wasting, spider angiomata, jaundice, splenomegaly, ascites, abdominal collateral vessels, and an altered mental status.

B. Laboratory Findings

Laboratory findings include hyperbilirubinemia, hypoalbuminemia, thrombocytopenia, and a prolonged prothrombin time. Other abnormalities that may coexist include anemia, elevated creatinine level, and hyponatremia. Although the presence of these abnormalities may indicate the presence of portal hypertension, these values often remain normal in patients with compensated or early cirrhosis.

C. Noninvasive Studies for Predicting Cirrhosis

Radiographic studies that strongly suggest cirrhosis include a small, nodular liver, ascites, splenomegaly, intra-abdominal varices, or portal and hepatic vein thrombosis; however, no test is considered a diagnostic gold standard. The current best test for diagnosing cirrhosis is liver biopsy. However, transient elastography as a noninvasive alternative has excellent predictive value for diagnosing cirrhosis and hepatic decompensation.

1. Abdominal ultrasound

Abdominal ultrasound findings that support a diagnosis of cirrhosis include a nodular liver, with increased echogenicity. In patients with more advanced cirrhosis and portal hypertension, findings of ascites, splenomegaly, and intra-abdominal varices may be detected. Unfortunately, ultrasonography is limited by interoperator variability, with a diagnostic accuracy of 85–91%. The addition of portal and hepatic vein flow Doppler images enables the assessment of hemodynamic changes that occur with cirrhosis. The reversal of portal vein flow occurs with increased hepatic resistance. This resistance results in the diversion of flow from the portal vein through portosystemic collaterals. This mechanism has been shown to be present in patients with advanced portal hypertension. Interobserver variability, patient position, phase of respiration, cardiac output, and timing of meals limit the accuracy of Doppler ultrasound.

2. Computed tomography scan and magnetic resonance imaging

These imaging studies are limited in their ability to detect changes associated with early cirrhosis but can accurately demonstrate later changes in liver architecture, ascites, and varices. Computed tomography angiography and magnetic resonance angiography can assess portal vein patency. Magnetic resonance elastography is very good for evaluating liver and spleen stiffness and identifies patients with cirrhosis and esophageal varices. However, performance of this test is complicated and it is not recommended for routine clinical evaluation.

3. Fibroscan

Transient elastography (Fibroscan,) is a technique that uses pulse-echo ultrasound to measure liver stiffness as a way of detecting fibrosis. This method of measuring fibrosis is reported to have a low interobserver variability and correlates well with the severity of fibrosis and the presence of portal hypertension. It is probably the best noninvasive test for determining the presence of cirrhosis. Spleen stiffness can be measured by either transient elastography or acoustic radiation force impulse imaging. However the heterogeneity of the results precludes its value for routine clinical practice.

D. Noninvasive Predictors of Esophageal Varices

Current practice guidelines recommend endoscopic screening for the presence of esophageal varices in all patients with cirrhosis. If varices are not present, screening endoscopy should be repeated within 2–3 years or sooner if there is evidence of hepatic decompensation. Several studies have recently attempted to identify noninvasive predictors of esophageal varices. A low platelet count has been associated with the presence of varices, although the discriminating threshold for the presence of varices ranges between 68,000 and 160,000/mm². A recent study found that patients with esophageal varices had a lower mean platelet count and a greater rate of reduction of platelets over time compared with those who did not have varices. However, several patients with cirrhosis developed varices despite having a normal (>150,000) platelet count. Noninvasive serum markers (indices combining a number of biochemical tests) have been useful in identifying cirrhotic patients in whom the risk of developing clinically significant esophageal varices is low. However, their positive or negative predictive values are insufficient to avoid screening endoscopy. Other noninvasive findings, such as splenomegaly, enlarged portal vein diameter greater than 13 mm on ultrasound imaging, and advanced Child-Pugh class, have not been reproducible predictors of esophageal varices. The predictive accuracy of suggested noninvasive markers remains low, and no markers that are currently available replace the need for endoscopic diagnosis of varices.

Promising predictors of esophageal varices are the platelet count/spleen diameter ratio as measured by abdominal ultrasound and liver stiffness as measured by transient elastography. In patients with compensated cirrhosis, the higher the platelet count/spleen ratio, the less likely it is that a patient will have varices. The use of the Fibroscan is limited by its lack of availability outside of academic medical centers. However, as fibroscan (transient elastography) becomes more widely available, it has the greatest promise for identifying patients with cirrhosis who are at very low risk for the development of esophageal varices. In these patients, screening endoscopy might be avoided.

Wireless capsule endoscopy has been compared with conventional upper endoscopy in identifying and characterizing esophageal varices. High sensitivity and specificity have been reported for the ability of wireless capsule endoscopy to determine the presence of esophageal varices (84–96% accuracy), the size of esophageal varices, and the presence of red wale signs. The potential benefits of capsule endoscopy include decreased study time, better patient tolerance, avoidance of intravenous conscious sedation, and possibly decreased costs. Wireless capsule endoscopy, however, is limited by inability to insufflate the esophagus, difficulty in measuring the length of varices, and image quality artifacts.

E. Invasive Studies to Measure Portal Hypertension: Hepatic Venous Pressure Gradient

Hepatic venous pressure gradient (HVPG) measurements provide information for diagnosis, prognosis, and management of portal hypertension. The HVPG is the difference between the wedged or occluded hepatic vein pressure and the free hepatic vein pressure. Normal portal pressure (HVPG) ranges from 1 to 5 mm Hg with greater than 5 mm Hg indicating the presence of portal hypertension. In patients with sinusoidal cirrhosis, the HVPG accurately predicts the portal venous pressure gradient. An HVPG greater than 10 mm Hg identifies patients with clinically significant portal hypertension and is predictive of the development of varices. Pharmacologically reducing the HPVG more than 10% at 1 year compared with the baseline measurement significantly lowers the risk of developing varices. Similarly, a 10% increase in HPVG significantly increases the risk of developing varices.

Determining a patient’s HPVG also predicts variceal bleeding, development of hepatic decompensation, development of hepatocellular carcinoma, determination of response to clinical treatment, and estimations of patient survival. Variceal bleeding can occur in patients with HVPG greater than 12 mm Hg. In patients with acute or ongoing bleeding, an HVPG greater than 20 mm Hg is associated with early rebleeding or uncontrolled bleeding, longer intensive care unit stay, prolonged hospital stay, higher transfusion requirements, and a lower probability of survival.

Patients achieving a reduction in HVPG to less than 12 mm Hg or a reduction in HVPG of 20% after pharmacologic therapy are less likely to develop recurrent esophagogastric variceal bleeding, ascites, spontaneous bacterial peritonitis, hepatorenal syndrome, and hepatic encephalopathy. Unfortunately, only 35–45% of treated patients respond with a 20% decrease in HVPG. Patients, who do not achieve an HVPG of less than 12 mm Hg, or a 20% reduction from their baseline HVPG after pharmacologic treatment, have a high risk for recurrent variceal bleeding and a greater risk of developing complications of portal hypertension and a higher probability of death. Repeat HPVG measurements obtained 1–3 months after initiating treatment help guide therapeutic decisions.

Treatment