• 
• 

Three Part Question

In [cirrhotic patients presenting with an upper GI bleed], should [antibiotic prophylaxis] be [routinely administered to prevent an infectious complication]?

Clinical Scenario

A 67-year-old man known for alcoholic cirrhosis comes in with an upper GI bleed. While you order the appropriate management therapies to control the bleed, one of your colleagues comes in and suggests that you add prophylactic antibiotic coverage to prevent an infectious complication. You have not heard of such a practice and wonder if this should be done routinely in cirrhotic patients presenting with an upper GI bleed. You perform a thorough search of the available literature.

Search Strategy

A. The website clinicaltrials.gov was searched for an ongoing trial on the topic. None was found.

B. MEDLINE (PubMed)
[“Gastrointestinal hemorrhage”[MeSH term] OR “gastrointestinal bleeding” OR “gastrointestinal hemorrhage”] AND [“Liver cirrhosis”[MeSH term] OR « cirrhosis »] AND [“Antibiotic prophylaxis” OR “antibiotic prophylaxis” OR “antibiotic”] LIMITS Humans, English or French, Clinical Trial, RCT, Meta-analysis, Systematic review

C. EMBASE
[“gastrointestinal hemorrhage”/exp] AND [“liver cirrhosis”/exp] AND [“antibiotic prophylaxis”/exp OR “antibiotic agent”/exp] LIMITS Humans, French or English, RCT, meta-analysis, systematic review
PubMED: 30 abstracts were screened for relevance.
EMBASE: 41 abstracts were screened for relevance.

Search Outcome

1 meta-analysis was found to be relevant to the question.

Relevant Paper(s)

Author, date and country	Patient group	Study type (level of evidence)	Outcomes	Key results	Study Weaknesses
Chavez-Tapia, et al. June 2011 Mexico	17 studies including a total of 1891 patients. Of those, 12 studies were comparing antibiotics to no intervention or placebo (1241 patients). Other studies were comparing different antibiotic regimens (650 patients). All the included studies were RCT. They all included cirrhotic patients with UGIB.	Meta-analysis	Mortality	RR 0.79 (0.63-0.98 ) (I2 = 0%), BWC: RR 0.48 (0.38-0.60), WBC: RR 1.45 (1.04-2.02)	High drop-out rate has a major impact on mortality outcomes as demonstrated by WBC and BWC analyses, but no effect on bacterial infection rates Heterogeneity among the studies regarding the number of bacterial infections outcome but results still statistically significant Heterogeneity in the different antibiotic regimens Blinding was not present in all of the studies; bias in data interpretation is therefore possible
			Number of patients that developed bacterial infections	RR 0.35 (0.26-0.47) (I2 = 32%), BWC: RR 0.27 (0.17-0.45), WBC: RR 0.70 (0.57-0.86)

Comment(s)

Most guidelines already include antibiotic prophylaxis in the management of upper gastrointestinal bleeding in cirrhotic patients. Despite heterogeneity among the studies selected for this meta-analysis, the results show a decrease in the number of bacterial infections in cirrhotic patients suffering from an upper gastrointestinal bleeding who are prophylactically treated with antibiotics (NNT for infectious complication, all sources = 4). However, the overall effect on mortality remains unclear. The delay within which the antibiotics should be given was not stated in the studies.

Clinical Bottom Line

Antibiotic prophylaxis should be used to prevent infectious complications in cirrhotic patients suffering from an upper gastrointestinal hemorrhage.

References

1. N. C. Chavez-Tapia, T. Barrientos-Gutierrez, F. Tellez-Avila, K. Soares-Weiser, N. Mendez-Sanchez, C. Gluud– & M. Uribe Meta-analysis: antibiotic prophylaxis for cirrhotic patients with upper gastrointestinal bleeding – an updated Cochrane review