OBJECTIVE: Recently it has been recommended that all cirrhotic patients without previous variceal hemorrhage undergo endoscopic screening to detect varices and that those with large varices should be treated with β-blockers (American College of Gastroenterology guidelines). However, endoscopic screening only of patients at highest risk for varices may be the most cost effective. METHODS: Ninety-eight patients without a history of variceal hemorrhage underwent esophagogastroduodenoscopy as part of a liver transplant evaluation. Univariate/multivariate analysis was used to evaluate associations between the presence of varices and patient characteristics including etiology of liver disease, Child-Pugh class, physical findings (spider angiomata, splenomegaly, and ascites), encephalopathy, laboratory parameters (prothrombin time, albumin, bilirubin, BUN, creatinine, and platelets), and abdominal ultrasound findings (portal vein diameter/flow, splenomegaly, and ascites). RESULTS: The causes of cirrhosis among the 67 men and 31 women (mean age, 48 yr) included 28% Hepatitis C/alcoholism, 25% Hepatitis C, 13% alcoholism, 9% primary sclerosing cholangitis/primary biliary cirrhosis, 9% cryptogenic, 6% Hepatitis B, 1% Hepatitis B and C, and 9% other. Patients were Child-Pugh class A 34%, B 51%, and C 15%. Endoscopic findings included esophageal varices in 68% of patients (30% were large), gastric varices in 15%, and portal hypertensive gastropathy in 58%. Platelet count <88,000 was the only parameter identified by univariate/multivariate analysis (p < 0.05) as associated with the presence of large esophageal varices (odds ratio 5.5; 95% confidence interval 1.8-20.6) or gastric varices (odds ratio 5; 95% confidence interval 1.4-23). CONCLUSIONS: Platelet count <88,000 is associated with the presence of esophagogastric varices. A large prospective study is needed to verify and validate these findings and may allow identification of a group of patients who would most benefit from endoscopic screening for varices.
ASJC Scopus subject areas