Objective: The incidence of esophageal and gastric varices and portal hypertensive gastropathy (PHG) has been well studied in cirrhotic patients. Because little is known of the prevalence of other upper and lower gastrointestinal tract pathology in pre-liver transplant candidates, we retrospectively studied the prevalence of and factors associated with these findings.
Methods: One hundred and twenty pre-liver transplant candidates underwent esophagogastroduodenoscopy to evaluate for varices, and 71 of them also underwent flexible sigmoidoscopy to screen for colorectal carcinoma. The association of upper and lower GI tract pathology with Child-Pugh Class, etiology of cirrhosis, and signs of portal hypertension, including presence and size of esophageal varices, presence of gastric varices, PHG, ascites, and splenomegaly, was analyzed using univariate and multivariate analysis.
Results: Etiology of cirrhosis among 87 men and 33 women (mean age, 52 yr) included 25% hepatitis C, 27% hepatitis C/alcohol, 15% alcohol, 10% primary sclerosing cholangitis/primary biliary cirrhosis, 9% cryptogenic, 8% metabolic, and 6% hepatitis B. Prevalence of Child-Pugh Classes A, B, and C were 34%, 49%, and 17%, respectively; 73% of patients had esophageal varices (23% were large), 62% PHG (23% were severe), and 16% gastric varices. Excluding varices and PHG, endoscopic findings in the upper GI tract (n = 120) included: 13% esophagitis/ulcers, 7.5% gastritis, 8% duodenitis, 2% Barrett's esophagus, 3% duodenal ulcers, and 2% gastric ulcers. Findings in the lower gastrointestinal tract (n = 71) included 21% adenomatous polyps, 21% internal hemorrhoids, 15% diverticulosis, 7% rectal varices, 3% colopathy, and 3% vascular ectasias. Univariate analysis revealed that there was a significant association between rectal varices and severe PHG (p < 0.05). This association was not maintained when multivariate analysis was performed.
Conclusions: Among all the findings, only rectal varices and colopathy were of higher prevalence in the pre-liver transplant population than that reported for the general population. No significant associations were found between these gastrointestinal tract lesions and patient characteristics.