Antibiotic prophylaxis to prevent infection in patients with Child-Pugh A cirrhosis with upper gastrointestinal bleed: an open label randomised controlled trial.

Background and aims: Although beneficial in reducing the risk of bacterial infections in patients with advanced decompensated cirrhosis after upper gastrointestinal (GI) bleed, the utility of prophylactic antibiotics in those with Child-Pugh A cirrhosis is not known. We studied if prophylactic antibiotics can be withheld in this cohort.

Methods: This was a single-centre, open-label randomised-controlled-trial with non-inferiority design. Patients of Child-Pugh A cirrhosis with upper-GI bleed and hemodynamic stability were randomised to receive either no prophylactic antibiotics (test-group) or ceftriaxone [standard of care (SOC)] for 72 h alongside standard medical management. The primary outcome was infection at day-5 in both arms. Secondary outcomes included failure to control bleed, mortality at day-5, and at 6 weeks.

Results: Eligible patients (n = 180; mean age 45.1 ± 13.1 years, 76.9% males; median MELDNa 9 [interquartile-range: 7-12]) of predominant non-viral etiology (alcohol: 43.4%; non-alcoholic steatohepatitis: 21.7%) were randomised, of whom outcomes could be reliably assessed for 172 and 140 patients at 5-day and 6-week follow-up, respectively. Rate of day-5 infections in test-group [7.0% (95% CI 2.8-15.1%)] was non-inferior to SOC arm [11.6% (95% CI 6.02-20.8%); absolute risk difference: -4.7% (95% CI -13.3% to 4.0%; non-inferior at 10% margin)]. Spontaneous bacterial peritonitis following post-bleed ascites was the most common site of infection in both groups (10/16; 66.7%). Rates of failure to control bleed [0% vs 4.9; absolute-risk-difference: -4.6% (95% CI -9.1% to 0.2%)], day-5 mortality [0% vs 2.5%; absolute-risk-difference: -2.3% (-5.5% to 0.9%)], and 6-week mortality [1.4% vs 2.5%; absolute-risk-difference: -1.6% (-6.5% to 3.2%)] were comparable in both arms.

Conclusion: Among patients with Child-Pugh A cirrhosis with hemodynamic stability, withholding prophylactic antibiotics after upper GI bleed was not associated with increased risk of post-bleed infections.