Prophylactic antibiotics have been routinely recommended for cirrhotic patients with upper gastrointestinal bleeding recently. However, the regimen and duration of its use remain an inconclusive issue. Quinolones and 3rd generation cephalosporins have been more often used for prophyalxis recently. The duration for antibiotic usage were variable in the literatures, ranged from 4-10 days. The latest guideline from AASLD in 2007 was 7 days. In the survey of infections in cirrhotic patients with UGI bleeding performed by Bernard et al, most infections occurred in the first 5 days and half within the first 48 hours. Therefore, considering the cost-effectiveness and drug resistance issues, the necessity for such prophylaxis for 7 days may need to be re-evaluated.

The purpose of our study is to investigate the antibiotic prophylaxis duration for cirrhotic patients with acute gastro-esophageal variceal bleeding. We will enroll those patients suffering from variceal bleeding documented by endoscopic examination and without apparent evidence of infection. Those who have received antibiotics within 2 weeks, are less than 18 years old, get pregnant, have malignancy other than HCC, have allergy to ceftrioxone are excluded. After receiving well explanation and giving consent, these patients are randomly allocated to 2 groups and receive prophylactic antibiotic just after endoscopic examination; Group I: receiving ceftriaxone 500 mg iv bolus stat and then q12h for 3 days, Group II: receiving ceftriaxone 500 mg iv bolus stat and then q12h for 7 days. They will receive appropriate endoscopic treatment for gastro-esophgeal varices and glypressin 1mg q6h for 3 days. They will start to feed on the 2nd day if not contraindicated. The 2nd endoscopic treatment for varices will be performed 2 weeks later. We record the demographic data, vital signs, transfusion amount; check hemogram, U/A, CXR, ascites routine (with apparent ascites), classification of variceal size and Child-Pugh classification. We monitor the events of rebleeding & infection, transfusion amount and hospitalization days We use rebleeding rate within 14 days as the primary end point. It is defined as the following events after initial stabilization of vital signs for 24 hours; (1): recurrence of hematemesis or bloody stool (2); need of transfusion more than 2 unit of blood and systolic pressure < 100 mmHg or pulse rate > 100/mn. We use infection rate during admission and mortality rate within 28 days as secondary end points.

1. Patient including & excluding Criteria Including: Those were diagnosed to have gastro-esophageal variceal bleeding after endoscopic examination. No evidence suggestive of infection was found including: 1)fever more than 37.5。C, 2)leucocytosis with WBC > 15000 mm3 or immature neutrophils > 500 mm3, 3)PMN of ascites more than 250/mm3，4)urinalysis, WBC > 15 WBC/HPF, 5)suspected pneuminia in CXR.

   Excluding:1)no consent 2)already has infection 3) antibiotic usage within 14 days 4) less than 18 years old 5) pregnancy 6) has malignancy other than HCC 7) allergy to ceftriaxone

2. Management, Allocation and Prophylaxis Those with esophageal variceal bleeding receive edoscopic variceal ligation; those with gastric variceal bleeding receive Histoacryl injection therapy. They all recieve glypressin 1mg q6h injection for 3 days and start to feed on the second day if not contraindicated. We will perform 2nd session of endoscopic treatment 14 days later. These patients will randomly alloacted to 2 groups and recieve prophylactic antibiotic after 1st endoscopic treatment; Group I: ceftriaxone 500 mg iv bolus stat and then q12h for 3 days, Group II: ceftriaxone 500 mg iv bolus stat and then q12h for 7 days. If rebleeding occurs after allocation, further treatment was decided by attending physician. If active infection occurs during prophylaxis period and no improvement was observed for 24 hours, ceftriaxone use will be stopped and further treatment was decided by attending physician.

3. Data record and Surveillance Data Record: We will record the basic data of the patietns: including: age, sex, co-morbidity (CAD, heart failure, COPD, chronic liver and renal diseases, malignancy, etc);hemogram before allocation,U/A, transfusion amount,ascites routine if significant ascites noted,CXR, endoscopic finding of varices; lowest blood pressure before allocation, Child-Pugh classification of the patient Surveillance: rebleeding and infection during admission, Transfusion amount and hospitalization day after allocation, rebleeding rate within 14 days & mortality rate within 28 days.

4. End points Primary end points: Rebleeding rate within 14 days after initial endoscopic treatment. Definition of rebleeding: the following manifestation after stabilization of vital sign for 24 hours,(1) appearance of hematemesis and bloody stool(2) need of transfusion more than 2 units of blood and vital sign change (systolic pressure < 100 mmHg or heart rate >100 beats/min).

Secondary end points: infection during admission, mortality rate within 28 days. Definition of infection: 1) bacteremia：positive blood culture, but no definite focus indentified, 2) spontaneous bacterial peritonitis : ascites routine: PMN> 250/mm3, 3)urinary tract infection: U/A WBC > 15/HPF and positive urine culture, 4)other infections: suspected by clinical, radiological or other bacterialogical examination, 5)respiratory tract infection:CXR changes with clinical evidence, 6)possible infection: fever (>37.5。C more than 6 hours) and leucocytosis,WBC > 15000 mm3 ，but with negative blood culture

Inclusion Criteria:

• those patients suffering from variceal bleeding documented by endoscopic examination and without apparent evidence of infection.

Exclusion Criteria:

• received antibiotics within 2 weeks, less than 18 years old, get pregnant, have malignancy other than HCC, have allergy to ceftrioxone