Antibiotics in Cirrhosis with Hemorrhage

Published

Fernández J, et al. "Norfloxacin vs. ceftriaxone in the prophylaxis of infections in patients with advanced cirrhosis and hemorrhage". Gastroenterology. 2006. 131(4):1049-56.
PubMed • Full text • PDF

Clinical Question

Among cirrhotic patients who present with GI bleed, how does ceftriaxone compare to norfloxacin in reducing the rate of bacterial infections?

Bottom Line

Among cirrhotic patients with GI bleed, ceftriaxone reduces the rate of bacterial infection by 67% but does not confer an increased survival benefit when compared to norfloxacin.

Major Points

In 1992, a landmark trial by Soriano et al demonstrated that norfloxacin reduces the incidence of bacterial infections and is associated with a nonsignificant trend towards improved survival among cirrhotic patients with GI bleed. Administration of antibiotics to these patients became considered standard-of-care soon thereafter. Subsequent epidemiologic studies revealed changes in antimicrobial resistance patterns, with many quinolone-resistant bacteria causing infections in these patients. This prompted the current study, which demonstrated the superiority of ceftriaxone over norfloxacin in reducing the rate of bacterial infections among cirrhotic patients with GI bleed. However, a 2010 Cochrane review did not find benefit of one antibiotic over another and, therefore, recommended antibiotic choice based upon local sensitivities.

Guidelines

AASLD Esophageal Varices (2007)

• Patients with cirrhosis and GI hemorrhage should receive up to 7 days of antibiotic prophylaxis (class I level A)
• Norfloxacin 400 mg PO BID or ciprofloxacin IV (no dose given) is recommended (class I level A)
• If the center has high quinolone resistance or the patient has advanced cirrhosis, ceftriaxone 1g IV daily may be preferable (class I level B)

Design

• Multicenter, double-blinded, parallel-group, randomized controlled trial
• N=111
  • Ceftriaxone (n=54)
  • Norfloxacin (n=57)
• Setting: 4 centers in Spain
• Enrollment: 2000-2004
• Analysis: Intention-to-treat
• Primary outcome: Proven infections plus possible infections

Population

Inclusion Criteria

• Age 18-80 years
• Hematemesis and/or melena within 24 hours of randomization
• Advanced cirrhosis, defined as 2 or more:
  • Severe malnutrition with muscle wasting
  • Bilirubin >3mg/dL
  • Ascites confirmed by paracentesis
  • Hepatic encephalopathy

Exclusion Criteria

• Allergy to study agent
• Infectious signs:
  • Fever >37.5 C
  • WBC >15
  • Immature neutrophils >500
  • Ascitic fluid ANC >250
  • Urine leukocytes >15/hpf
  • CXR compatible with PNA
• Antibiotics within 2 weeks of hemorrhage (excluding oral norfloxacin for SBP prophylaxis)
• Prior HCC
• HIV

Baseline Characteristics

• Demographics: Age 58 years, male 77%
• Alcoholism as etiology of cirrhosis: 68%
• Mean laboratory measurements: Bilirubin 4.4 mg/dL, albumin 2.6 mg/dL, INR 1.6
• Degree of liver dysfunction:
  • Child-Pugh score: 9.8 (47% B and 53% C)
  • MELD score: 17
  • >2 signs of liver failure: 41%
• Complications: Ascites 79%, severe malnutrition 72%, hepatic encephalopathy 37%, HCC 18%
• PMH: Renal failure 12%, T2DM 20%
• Already receiving norfloxacin for SBP prophylaxis: 9%
• Etiology of bleed:
  • Esophageal varices: 64%
  • Gastric varices: 5%
  • Peptic ulcer: 10%
  • Portal hypertensive gastropathy: 8%
  • Other: 13%
• Adjunctive therapies:
  • Terlipressin or somatostatin: 77%
  • Emergency sclerotherapy or banding: 61%
  • Balloon tamponade: 6%
  • TIPS: 7%
  • Surgical shunt: 5%

Interventions

• Randomized to norfloxacin 400mg orally bid vs. ceftriaxone 1g IV daily
• Central line used in most patients
• Somatostatin or terlipressin was given to patients with hemorrhage from variceal or portal hypertensive gastropathy
• Evaluated daily for development of proved or possible infection
• Follow-up 10 days

Outcomes

Comparisons are ceftriaxone vs. norfloxacin.

Primary Outcomes

Proven infections plus possible infections

11% vs. 33% (P=0.003)

Secondary Outcomes

Proven infections

11% vs. 26% (P=0.03)

Spontaneous bacteremia or SBP

2% vs. 12% (P=0.03)

10-day mortality

11% vs. 9% (P=NS)

Inpatient mortality

15% vs. 11% (P=NS)

Criticisms

• Not powered to detect mortality
• Antibiotics compared were delivered via different routes (PO vs. IV), which may favor ceftriaxone, particularly when the majority of patients had nasogastric tubes in place

Funding

Funded by Fondo de Investigación Sanitaria and the Instituto de Salud Carlos III.