+++
TEXTBOOK PRESENTATION
++
A patient with known cirrhosis presents with heavy upper GI bleeding (hematemesis or melena). Stigmata of chronic liver disease as well as a history of previous hemorrhages are frequently present. Laboratory data demonstrate liver biochemical tests consistent with cirrhosis and thrombocytopenia.
++
Esophageal varices are portosystemic collaterals that dilate when portal pressures exceed 12 mm Hg.
Although varices are the second most common cause of upper GI bleeding, they account for 80–90% of GI bleeds in patients with cirrhosis.
Gastroesophageal varices are present in about 50% of patients with cirrhosis.
The prevalence of varices depends on the severity of the cirrhosis.
The Child-Pugh system classifies patients based on the severity of their cirrhosis.
The system takes into account the presence of encephalopathy, ascites, hyperbilirubinemia, hypoalbuminemia, and clotting deficiencies (Table 19-5).
40% of patients with Child-Pugh grade A disease have varices, while 85% of patients with grade C disease have varices.
Approximately 33% of patients with varices experience hemorrhage.
Varices may develop from cirrhosis of any cause.
Of all GI bleeds, those from varices carry the worst prognosis.
Nearly 33% of patients die at the time of their first variceal hemorrhage.
Up to 60% of survivors have recurrent bleeding in the first year.
A variceal bleed carries a 32–80% 1-year mortality and a 6-week mortality of 15–20%.
A hepatic venous pressure gradient > 20 mm Hg predicts poor outcomes.
Esophageal varices are, by far, the most lethal type of GI bleeding.
++
+++
EVIDENCE-BASED DIAGNOSIS
++
Esophageal varices are diagnosed with endoscopy.
Screening for varices
Because variceal bleeding carries such a high mortality, the goal is to detect varices before they bleed so that prophylactic treatment can be initiated.
All patients with cirrhosis should undergo screening endoscopy every other year.
Patients with cirrhosis but without splenomegaly or thrombocytopenia are at the lowest risk for having varices (about 4%). Endoscopy may be delayed in these patients.
Of all causes of GI bleeding, varices are probably the easiest to predict. One study has the sensitivity and specificity of physicians predicting variceal hemorrhage at 82% and 96%, respectively.
++
Primary prophylaxis (patients with varices but no previous bleeding)
Nonselective beta-blockers (usually propranolol or nadolol) effectively decrease portal pressures.
Patients at higher risk for bleeding should also undergo band ligation of the varices.
Secondary prophylaxis (patients who have ...