Antibiotic Prophylaxis for Upper Gastrointestinal Bleed in Liver Cirrhosis; Less May Be More.

Abstract

Administration of antibiotics in patients with cirrhosis and upper gastrointestinal bleeding has been shown to improve outcomes. Little is known regarding optimum duration of prophylactic antibiotics. Seven days of antibiotics are generally recommended but very few studies have compared antibiotic duration to clinical outcomes in current available scientific literature. The goal of our study was to study the effect of shorter antibiotic duration on patient outcomes. We conducted a retrospective cohort study of patients with cirrhosis presenting with upper GI bleeding at our institute from 2010 to 2018. Patients were divided into three cohorts based on duration of antibiotic administration for prophylaxis: 1-3days of antibiotics, 4-6days of antibiotics and 7days or more of antibiotics. Rates of infection diagnosis within 30days, rebleeding, and mortality were compared between the three groups with Chi square, Fisher Exact and Kruskall-Wallace tests. Multivariable analysis was conducted to evaluate independent risk factors for infection. Medical charts of 980 patients with cirrhosis and upper GI bleeding during the study period were reviewed. A total of 303 with upper gastrointestinal bleeding were included in the final sample, of these 243 patients received antibiotics for prophylaxis and were included for analysis. Seventy-seven patients received antibiotic therapy for 3days or less, 69 patients for 4-6days, and 97 patients longer than 6days. The three groups were well matched in demographic and clinical variables. Twenty-seven patients developed infections within 30days of bleeding. MELD-Na score at presentation and presence of ascites were associated with infection within 30days. Rates of infection were not statistically different between the three antibiotic groups (p = 0.78). In the thirty days following the GI bleed, pneumonia was the most diagnosed infection (eleven patients) followed by urinary tract infections (eight patients). Four patients developed spontaneous bacterial peritonitis and three were diagnosed with bacteremia. There was no difference in time to infection (Kruskall Wallace test p = 0.75), early re-bleeding (p = 0.81), late re-bleeding (p = 0.37) and in-hospital mortality (p = 0.94) in the three groups. Six patients in the cohort developed C. Difficile infection; no patient in the short antibiotic group developed C. Difficile infection. Short course of antibiotics for prophylaxis (3days) appears safe and adequate for prophylaxis in patients with cirrhosis with upper gastrointestinal bleeding if there is no active infection.