96 Portal Hypertension

Anatomy and Physiology of the Portal System

The term portal system refers to a venous system that begins and ends in capillaries. The portal venous system commences in the capillaries of the intestine and ends in the hepatic sinusoids. The portal venous system drains blood from the gastrointestinal (GI) tract, pancreas, gallbladder, and spleen. The portal vein originates from the confluence of the splenic vein and the superior mesenteric vein. The inferior mesenteric vein and short gastric veins drain into the splenic vein. The superior mesenteric vein drains all the blood from the small bowel and the right colon, while the inferior mesenteric vein drains the blood from the remainder of the colon and most of the rectum. Flow in the portal vein is normally about 1 L/min (approximately 20% cardiac output) with a mean pressure of 7 mm Hg. Although the blood in the portal vein is the outflow from capillary beds and therefore has relatively low oxygen content, 70% of hepatic oxygenation is derived from portal flow. The blood flowing through the hepatic artery supplies the remainder of hepatic oxygen consumption and is the primary blood supply to the biliary tree. The portal vein carries a high concentration of nutrients and hormones, facilitating the liver’s central role in fat, carbohydrate, drug, and protein metabolism. Toxic substances are removed by hepatocytes, and bacteria (and bacterial products) are removed by Kupffer cells. Portal venous blood and hepatic arterial blood mix at the sinusoidal level, and there exists an adenosine-mediated local hepatic arterial autoregulatory “buffer response” that increases arterial inflow in response to low portal flow; however, total hepatic flow is not preserved when hepatic arterial flow is decreased. This buffer response is also dysregulated in sepsis.¹

Postsinusoidal blood drains through hepatic venules into hepatic veins and then into the inferior vena cava to return to the systemic circulation. A variety of pathologic processes can result in portal venous flow becoming “obstructed.” Regardless of the cause (i.e., intra- or extrahepatic obstruction), this resistance to portal flow increases portal pressure and leads to the development of what is called the portal hypertension syndrome, which is characterized by the formation of portosystemic collaterals. Under these circumstances, only a portion of the blood flow that originates within the portal system reaches the liver; the remainder is diverted through collaterals and enters the systemic circulation directly.

The major sites of collateral remodeling are the gastroesophageal region, between the inferior mesenteric vein and the hemorrhoidal vein, the umbilical veins and cutaneus veins of the abdominal wall, and via retroperitoneal systems into the azygous system and the vena cava. Collateral vessels (varices) also may develop at the sites of previous surgery, trauma, or adhesions and may similarly be found at ileostomy or colostomy stomas (ectopic varices). In addition to the formation of discrete collateral vessels, there are also more generalized changes within the GI tract, leading to vascular ectasia or so-called portal hypertensive enteropathy. Bleeding may result from varices or portal hypertensive enteropathy.

Patients with portal hypertension exhibit characteristic splanchnic and systemic circulatory changes. Key to these manifestations is abnormal vasodilatation. Decreased arteriolar tone in the splanchnic vessels leads to splanchnic hyperemia and hypervolemia, but also a reduction in effective central blood volume, with the majority of the excess blood volume being within the splanchnic bed. These circulatory changes prompt homeostatic systemic responses, with activation of the vasoconstrictor and sodium-retaining mechanisms. Overall, these changes comprise a hyperdynamic circulation characterized by increased cardiac output and heart rate to maintain blood pressure in the face of decreased systemic vascular resistance, with an overall increase in total plasma volume.

Pathophysiology

According to Ohm’s law (as applied to the cardiovascular system rather than an electrical circuit), the pressure within a vessel is determined by the flow of the blood in that vessel divided by the resistance. Apparent resistance depends upon a number of factors, including the length of the vessel, the radius of the vessel, and the viscosity of the blood. Since length and blood viscosity remain relatively constant, changes in radius are of paramount importance for determining changes in apparent resistance. An increase in blood flow in the portal vein and hepatic artery are important to the development of portal hypertension in some cases, but the increase in resistance seems to be the most important factor and is used to classify portal hypertension (PHT). The origin of PHT can be divided into cirrhotic and noncirrhotic and presinusoidal, sinusoidal, and postsinusoidal (Table 96-1). In response to PHT, vascular collaterals develop, and vascular resistance drops in the splanchnic bed, leading to the development of a hyperdynamic circulation. As a consequence, splanchnic and portal venous inflow increases, and PHT persists even with the development of vascular collaterals. As the pressure within the portal system continues to rise, portal blood flow decreases and hepatic perfusion deteriorates. The liver is deprived of portal blood, and this tends to accelerate the progression of liver disease. The hyperdynamic circulation and PHT also contribute to the development of portopulmonary syndrome (pulmonary hypertension and PHT), hepatopulmonary syndrome (hypoxia and intrapulmonary shunting in association with PHT), cirrhotic cardiomyopathy, ascites, and hepatorenal syndrome.

TABLE96-1 Etiology of Portal Hypertension

Increased vascular resistance in the portal system is the most important factor in the development of the PTH syndrome. Disorders as diverse as splenic vein or portal venous thrombosis, cirrhosis, or constrictive pericarditis can result in PHT even though their clinical manifestations differ. The site of increased resistance in cirrhosis was initially thought to be post-sinusoidal in nature, but increasingly it is recognized that sinusoidal and pre-sinusoidal factors contribute. The increased resistance to flow has both a static and a dynamic component. The static component in cirrhotic livers is due to distortion of liver architecture with reduced hepatic microcirculation. These vascular changes, together with compression of the portal vein branches by regenerative nodules, increase vascular resistance. Furthermore, fibrosis initially develops in the space of Disse, where hepatic stellate cells produce collagen and further compress the sinusoids. In clinical studies, measured portal pressure correlates with the extent of fibrosis in liver biopsy specimens. Also, the size of the hepatocytes is important. Treatments which are known to increase the size of hepatocytes increase portal pressure. This observation may explain, at least partially, why portal pressure frequently falls with abstinence from alcohol consumption. The dynamic component is mostly related to endothelins and nitric oxide. After activation, hepatic stellate cells, or Ito cells, become myofibroblasts and express endothelin (ETa and ETb) receptors and contract after exposure to endothelin-1, resulting in an increase in portal pressure. The dynamic nature of portal pressure changes and Ito cell function are important, both in terms of the pathogenesis of acute bleeding and as a target for pharmacotherapy.

Diagnosis of Portal Hypertension

PHT is defined as a portal pressure that is 5 mm Hg greater than the pressure measured in the inferior vena cava or a pressure of more than 15 mm Hg in the splenic vein or portal pressure measured at surgery. If the gradient is greater than 10 mm Hg, then clinically significant PHT is present. The direct consequences of PHT are formation of portosystemic collaterals and splenomegaly. Portosystemic collaterals can become clinically apparent as gastric or esophageal varices, umbilical vein recanalization, retroperitoneal collaterals, and/or rectal or ileostomy varices. The complications of PHT are variceal bleeding, ascites, spontaneous bacterial peritonitis, hepatic encephalopathy, hyperdynamic circulation, and hypersplenism. Varices are rarely (maybe never) seen if the gradient is less than 10 mm Hg.² Variceal bleeding is not observed if the pressure gradient is less than 12 mm Hg, and protection from variceal bleeding is gained if the pressure gradient can be manipulated to less than 12 mm Hg or a 20% reduction in pressure is achieved.³

Direct measurement of the hepatic vein wedge pressure or the portal venous pressure requires invasive means, most often transjugular catheterization. The advantage of this approach is that caval and hepatic venous pressures can be measured during the same procedure. Less frequently, a transhepatic approach is adopted, and rarely, portal venous pressures are measured directly (by cannulating a branch of the superior mesenteric vein) at the time of a surgical procedure. Very rarely, splenic pulp pressure is measured.

Indirect measurements also can be used to assess the portal pressure gradient. This procedure involves measurement of the free and wedged hepatic venous pressure using catheterization of the right hepatic vein. Wedged hepatic venous pressure (measured using a balloon-tipped catheter) reflects the pressure in a static column of blood from the hepatic vein to the sinusoid. It is an assessment of sinusoidal pressure rather than portal venous pressure and therefore may underestimate the portal pressure gradient in disease states characterized by pre-sinusoidal hypertension (see Table 96-1). The free hepatic venous pressure is obtained with the catheter in the hepatic vein and gives an assessment of caval pressure. Free hepatic venous pressure is not elevated in patients with diseases characterized by pre-sinusoidal and sinusoidal PHT, but it is characteristically raised in post-hepatic (or extrahepatic postsinusoidal) etiologies. The gradient between the two measurements is called the hepatic venous pressure gradient and is the most commonly quoted parameter in the medical literature regarding management of PHT. Both the absolute value of hepatic venous pressure gradient and the change in hepatic venous pressure gradient with pharmacotherapy have prognostic significance related to the risk of variceal bleeding.⁴

It can be appreciated that even indirect methods of measuring portal pressure are not readily available in most settings. Instead, most clinicians rely on the clinical manifestations of PHT: esophagogastric varices, splenomegaly, edema, and ascites.

Complications of Portal Hypertension

Varices

Bleeding from varices is a major cause of morbidity and mortality in patients with significant PHT. Life expectancy after variceal bleeding is considerably curtailed, both as an immediate consequence of hemorrhage (and related complications, such as sepsis and renal failure) and in the longer term due to rebleeding.

Two main mechanisms have been implicated in the pathogenesis of variceal hemorrhage in patients with established varices and PHT: erosions secondary to acid reflux and spontaneous rupture. Effects related to ascites and changes in plasma volume also have been implicated in the genesis of bleeding. Ascites may be a factor in variceal hemorrhage, because ascites can transmit intraabdominal pressure and thereby increase the pressure inside the esophageal varices. Some studies have shown a decrease up to 10% in the hepatic vein pressure gradient (HVPG) and a decrease in portal flow with paracentesis. Drainage of large volume of ascites can decrease intraabdominal pressure, leading to splanchnic dilation and increased blood flow against a fixed resistance in the liver.⁵ This circumstance would increase portal pressure and, hence, the risk of bleeding. The pressure inside the varices does seem to be affected by intraabdominal pressure, but whether this affects the risk of bleeding is not known. Erosions secondary to esophagitis also have been suggested as an important factor for the bleeding process.^6,⁷ However, there was no evidence of acid reflux in patients studied with a pH electrode, and treatment with cimetidine did not affect the rates of rebleeding from varices.⁸

The pressure inside the varices is directly dependent on the portal pressure and also on the radius of the varix. The pressure within the varix is inversely proportional to wall thickness. Therefore, varices are more likely to bleed when they are larger and have a thinner wall. Large, thin-walled varices are generally located near the gastroesophageal junction where the veins are more superficial and less surrounded by other tissues. There is a relationship between the risk of bleeding and portal pressure. If the HVPG is greater than 20 mm Hg after an initial bleed, it is a poor prognostic sign, and there is a substantial risk of rebleeding and mortality.⁹ In a recent study, patients underwent portal pressure measurement after initial endoscopy and control of bleeding. Those whose pressure was above 20 mm Hg were randomized to early transjugular intrahepatic portosystemic shunt (TIPS) or conventional treatment. The group who underwent TIPS shunt insertion had an improved outcome compared to the standard-of-care high-risk group and similar to that of the low-risk group. Recent work has also suggested that portal pressure may be equally predicted by Child-Pugh score, and thus in this group of patients, consideration should be given to early TIPS shunt insertion.

It may be equally possible to measure variceal pressure at the time of endoscopic band ligation. This appears to be feasible and safe, although it remains at present an experimental method, and more studies are required.¹⁰

The main risk factors for rebleeding and mortality according to the North Italian Group of Portal Hypertension are the Child-Pugh class, the size of the varix, and the presence or absence of red wale markings and/or cherry spot at endoscopy.

There is increasing evidence to suggest that an episode of infection is the precipitating event in most cases of variceal bleeding. Infection is thought to trigger the release of cytokines and other proinflammatory mediators, resulting in increased hepatic resistance and possibly increased portal venous flow, with a sudden rise in portal pressure. A postprandial increase in splanchnic blood flow also may contribute to an acute rise in portal pressure.

Gastroesophageal varices are present in approximately 50% of cirrhotic patients and are large in 20%. Approximately one-third of patients with varices have bleeding complications. The patients at greatest risk of bleeding are those with advanced liver disease and large varices and high-risk stigmata. It is important, therefore, to identify the population at risk and modify their risk of bleeding. In patients with cirrhosis, the incidence of varices is 5% per year.

Gastric varices can be classified into four different groups: GOV1 is the gastric varix that is a direct continuation of an esophageal varix; GOV2 are also a continuation of a esophageal varix, but are more extensive and reach the fundus of the stomach; IGV1 are fundal varices that are not in continuity with an esophageal varix; and IGV2 are gastric varices that do not originate from a esophageal varix and are located in the body of the stomach. GOV2 and particularly IGV1 bleed more commonly than other gastric varices.

Portal hypertensive gastropathy is a complication of PHT that causes flow and pressure changes in the gastric mucosa. Mild or chronic bleeding is observed in 35% of the patients with mild gastropathy and 90% of those with severe gastropathy. Overt bleeding happens in 30% of those with mild and in 60% of those with severe gastropathy.¹¹ The administration of propranolol decreases gastric mucosal blood flow and is effective in reducing bleeding in portal hypertensive gastropathy.^12,¹³ TIPS and transplant are effective modes of treatment.

Another form of gastropathy is gastric antral vascular ectasia (GAVE). This lesion is localized at the antrum, and is associated with poor hepatic function. GAVE carries a high risk of bleeding. It is best managed with endoscopic therapy, generally with sessions of argon plasma coagulation (APC) until hepatic transplantation can be performed.

Diagnosis of Variceal Bleeding

When a patient with a possible hepatic disorder presents with hematemesis or melena, the most common cause of bleeding is from varices.¹⁴ Sometimes it is useful to insert a nasogastric tube followed by lavage of the stomach, both as a diagnostic tool and also to clear the gastric cavity before endoscopy. Since patients with PHT also can bleed from gastritis, esophagitis, Mallory-Weiss tears, and peptic ulcers, the most accurate method for the diagnosis of bleeding varices is upper endoscopy; accuracy exceeds 90%. Frequently one or two doses of 250 mg of erythromycin are administered before the procedure to help clear the stomach before the procedure.