Two hundred patients affected by liver cirrhosis and oesophageal varices were studied in 9 Gastrointestinal Units in Lombardy (Northern Italy) in order to assess factors possibly related to variceal bleeding. Only patients without any previous episode of gastrointestinal bleeding were included in the prospective evaluation. For each patient demographic data, aetiology of cirrhosis, various clinical and biochemical parameters able to group patients into the three Child-Pugh Classes, endoscopic items for calculation of Beppu's and of NIEC prognostic scores were recorded on computerized cards. Patients were regularly interviewed every three months for one year and underwent an upper gastrointestinal endoscopy at enrollment, after six months and in case of bleeding. Within the twelve-month follow-up period, 29 out of the 200 patients (14%) bled and 52 out of 200 died (26%). In 16 of the 52 patients who died (59% of bleeding patients) death was directly related to gastrointestinal bleeding. Bleeding from oesophageal varices was endoscopically proven in 19/29 patients, in another 9 bleeding was classified as from unknown source and in one patient a bleeding gastric ulcer was diagnosed. Univariate analysis of all the recorded clinical, biochemical and endoscopic parameters, performed by Chi-square method and Fisher exact test showed that the presence of RWM (p < 0.001) was the only factor significantly associated to variceal bleeding within one year. Relationship between size of varices and bleeding was very close to the statistical significance but did not achieve it (p = 0.058).