Incidence of large oesophageal varices in patients with cirrhosis: application to prophylaxis of first bleeding.

Because several studies have suggested that beta blockers are effective in the prophylaxis of first variceal bleeding in cirrhosis, screening for oesophageal varices might be appropriate. We prospectively studied 84 cirrhotic patients without obvious evidence of large oesophageal varices and previous bleeding during a mean follow up of 16 months. At entry to the study 41 patients had no oesophageal varices and in 43 these were grade 1. The subsequent percentages of patients without large oesophageal varices were 74% at one year and 52% at two years. Univariate analysis showed that a longer duration of cirrhosis (p less than 0.05) and grade 1 oesophageal varices at entry (p less than 0.001) were predictive factors for the occurrence of large oesophageal varices, whereas, multivariate analysis showed that the initial size of the oesophageal varices (p less than 0.001), a high initial Child-Pugh score, and a smaller improvement in Child-Pugh score during the study were independent risk factors. Among patients with grades 0 and 1 oesophageal varices at the start of the study the proportions with large oesophageal varices at two years were 31% and 70% respectively. We have calculated that, accepting a maximum risk of first bleeding of 10% without prophylactic treatment, a patient without oesophageal varices should be screened endoscopically every other year, while a patient with grade 1 disease should benefit from one annual upper gastrointestinal endoscopy.