To determine if prophylaxis with cefazolin produces a significant reduction in infections in cirrhotic patients with gastrointestinal bleeding when compared with ciprofloxacin. Randomized clinical trial. Patients 18 years or older with diagnosis of cirrhosis, gastrointestinal bleeding and no clinical or laboratory evidence of infection who were admitted to the gastrointestinal bleeding unit of HNERM between July 2008 and July 2010 were included. Patients were allocated to receive either i.v. ciprofloxacin 200 mg bid or i.v. cefazolin 1 gm tid for 7 days. 98 patients were included, 53 in the cefazolin group and 45 in the ciprofloxacin one. Age average was 66 +/- 10 years, 61% were male, 59,2% had ascites. Overall rate of infections was 14,3% (14/98). Rebleeding rate was 8,1% and mortality 4,1%. There were no differences in age, sex, Child Pugh score, ascites, hepatic encephalopathy nor in billirubin, albumin, PT and creatinine levels between the study groups. Infection rate in the cefazolin groups was 11,3% while in the ciprofloxacin one 17,8% (p=0,398).When Child-Pugh A and patients without ascites were excluded of the analysis, the cefalozin group had 22,2% of infections and 26,9% in the ciprofloxacin one (p=0,757). there were no differences in infection rates between patients with prophylaxis with cefazolin and ciprofloxacin after an episode of gastrointestinal bleeding.