Mr. Norman’s history indicates a recent and perhaps ongoing hemorrhage. Acute blood loss is indicated by the history and vital signs, which are best explained by significant hypovolemia. A priority is to resuscitate him by placing large-bore intravenous lines and rapidly infusing isotonic saline solution. Blood is drawn for typing and cross-matching. Additional tests include a complete blood count, platelet count, clotting studies (prothrombin time and partial thromboplastin time), and determination of electrolytes, blood urea nitrogen, and creatinine levels. Because of his history, liver function should be studied, including determination of albumin, total protein, bilirubin, alkaline phosphatase, aspartate transaminase, and alanine transaminase. The amylase level should also be checked.

The pulse rate, blood pressure, and urine formation should be evaluated. If the ability to urinate is in doubt, a Foley catheter should be placed to guide resuscitation. With complete resuscitation, his blood pressure will rise to whatever is his normal, probably a systolic above 120 mm Hg; the heart rate will drop, probably to below 100 beats per minute; and his urine output will be more than 30 to 50 mL per hour. He would undoubtedly have orthostatic hypotension on admission to the emergency department, and this would disappear with full resuscitation. Note that he may manifest a tachycardia because of anemia.

Upper GI endoscopy should be performed as soon as possible.

Endoscopic sclerotherapy should be performed if possible. The technique of sclerotherapy is to visualize the varices directly and to inject a sclerosant in and around them. As early as 1939, Crafoord considered using specially modified rigid esophagoscopes to inject esophageal varices with material that would cause them to scar and thrombose (3). The procedure did not become prominent because of its technical difficulty in comparison with shunting, but when fiberoptic endoscopes became clinically available in the 1970s, the procedure was revived. Because of its ease, most centers used sclerotherapy as the main procedure for early control of bleeding varices, especially in poor-risk patients (4).

The agents available for sclerotherapy in the United States is ethanolamine oleate. The procedure is successful in the initial control of bleeding varices in 80% to 90% of patients, in combination with other measures for medical management. In patients with acute bleeding, such as Mr. Norman, technical success depends on a clear view of the site of hemorrhage without interference from bleeding sites or persistent blood in the gastroesophageal lumen. Bleeding from gastric varices is more difficult to control with sclerotherapy because the upper stomach cannot be fixed in position for injection and because there are more veins that can bleed. Gastropathy from portal hypertension, including submucosal hyperemia, cannot be treated with sclerotherapy (5). The outcome is influenced by the Child classification of the patient at the time of sclerotherapy (6). An alternative surgical procedure, early portacaval shunting, has produced better results in the short- and long-term management of people with bleeding varices than has sclerotherapy (7). Additional studies favorably compare esophageal transection with sclerotherapy for control of acute bleeding (8).

In normal conditions, blood flows through the portal veins under low pressure (less than 10 mm Hg), which is consistent with the low resistance in the portal circulation of the liver. Significant increases in portal venous blood flow can be accommodated (e.g., during digestion) without an increase in portal venous pressure. In a diseased liver, with alteration of the microscopic portal architecture, obstruction to portal venous flow leads to elevation of portal venous pressure. The physiologic response to this rise in pressure is the formation of alternative channels for blood flow from the intestine to the central venous system, which bypass the liver altogether. These varices occur in the retroperitoneum, along the esophagus and stomach, around the anal canal, and around the umbilicus. Patients with symptoms of portal hypertension have portal venous pressures typically higher than 20 mm Hg, sometimes higher than 40 mm Hg.

Alcohol abuse is common in the general population; practitioners should therefore check for warning signs by asking each patient about the quantity of alcohol used, binge drinking if any, and any societal adverse events such as being away from work or being arrested for drunk driving (9). Often, the most important history for the quantity of alcohol consumption comes from the patient’s family. Almost always associated with alcohol abuse is smoking. The physical examination often shows a general loss of muscle mass, and wasting can be severe in the end-stage patient. The liver feels firm and is often enlarged below the costal margin. Ascites may be present, although it may be difficult to differentiate fluid in the abdomen from mesenteric adiposity combined with muscle laxity. Spider angiomata over the chest and upper back can occur, as can gynecomastia. Testicular atrophy is common. There are no direct signs of portal engorgement other than the rare appearance of caput medusa from periumbilical collateral venous channels.

Child classified patients with alcoholic cirrhosis in terms of the expected morbidity and mortality from surgery for portal hypertension (2). The original intent was to stratify outcomes on the basis of a worst-case analysis. This classification was later modified by Pugh (2) to provide a general classification of the severity of disease. It describes the degree of liver damage in terms of protein synthetic function and ability to process digestive products (Table 17.1). A patient’s classification is based on the worst criterion. For example, a patient who has fixed ascites but all other factors in the B group range is still placed in group C.

No single test assesses liver function overall in a manner analogous to the creatinine clearance for renal function or ejection fraction for cardiac function. Other tests that describe liver function include prothrombin time as a measure of synthetic function and the bromsulphalein clearance test as a measure of the ability to process products of digestion (10). A cofactor in the prognosis of these patients is alcoholic hepatitis as established by liver biopsy. Biopsy is not routinely performed for patients with portal hypertension, but when tissue is obtained, the architectural changes in the liver and the prominence of Mallory bodies correlate with the outcome of surgery (11).

The specific reason varices that have been stable for years suddenly hemorrhage is unknown. Portal pressure varies significantly over the day, especially with changes in the degree of inflammation of the liver. Other factors include esophageal or gastric erosion over varices. A prospective study of the predictors of bleeding from extant varices indicated that dominant risk factors are Child class, size of varices, and any abnormal markings over the varices (red wale markings) (12). Other factors that correlated with bleeding in other studies include the location of varices, other markings over the varices (cherry-red spots and hematocystic spots), color of varices, esophagitis, ascites, and prothrombin time.