Prophylactic antibiotics for penetrating abdominal trauma

  • Protocol
  • Intervention



This is the protocol for a review and there is no abstract. The objectives are as follows:

To quantify the effectiveness and safety of prophylactic antibiotics in penetrating abdominal trauma.


Penetrating abdominal trauma occurs when the peritoneal cavity is breached. Routine laparotomy for penetrating abdominal injuries began in the 1800s, while antibiotics were first used in World War II to combat septic complications associated with such injuries (Poole 1944). In 1972 Fullen noted a 7-11% post surgical infection rate with pre-operative antibiotic administration, a 33-57% post surgical infection rate with intra-operative antibiotic administration, and a 30-70% post surgical infection rate with only post-operative antibiotic administration (Fullen 1972). Approximately 20% of all penetrating abdominal wounds result in colonic injuries, and, consequently, aerobic and anaerobic flora are frequently cultured from septic complications. The most common organisms include Escherichia coli (E. coli), enterobacter species, Klebsiella species, Bacteriodes fragilis (B. fragilis) and Clostridia species (Fabian 1993). Dellinger observed that antibiotic regimens that excluded any anaerobic coverage resulted in incidences of infection of between 20-26%, versus an incidence of 7-14% when anaerobic cover was given (Dellinger 1991). According to current guidelines, there is sufficient class I evidence to support the use of a single pre-operative broad-spectrum antibiotic dose (Luchette 2000), with aerobic and anaerobic cover (Thadepalli 1973), with continuation of up to only 24 hours, in the event of a hollow viscus perforation found at exploratory laparotomy (Luchette 2000).

Why it is important to do this review

The aim of this review is to assess the benefits and harms of prophylactic antibiotics administered for penetrating abdominal injury in order to reduce the incidence of septic complications, such as septicaemia, intra-abdominal abscesses and wound infections. In undertaking this review we also aim to determine whether or not current guidelines are evidence-based, and to what extent further research is required.


To quantify the effectiveness and safety of prophylactic antibiotics in penetrating abdominal trauma.


Criteria for considering studies for this review

Types of studies

We will include randomised controlled trials only.

Types of participants

Patients who have an isolated penetrating abdominal wound and are not receiving antibiotics. Patients must have no evidence of intra-abdominal sepsis or any other focus of infection on presentation at the emergency department.

Types of interventions

Trials that compare prophylactic antibiotic administration against no pre-operative administration of antibiotics will be eligible for inclusion. Trials will be included regardless of the type, dose, or route of administration of the antibiotics.

Types of outcome measures

Primary outcomes
  • Intra-abdominal abscesses.

  • Laparotomy wound infections.

Secondary outcomes
  • Mortality.

  • Septicaemia/bacteraemia in the absence of extra-abdominal infections.

  • Drain-tract infections.

Search methods for identification of studies

We will attempt to identify all relevant trials regardless of language, year of publication, or publication status.

Electronic searches

We will search the following electronic databases:

  • The Cochrane Injuries Group Controlled Trials Register;

  • The Cochrane Central Register of Controlled Trials (CENTRAL) in The Cochrane Library (to the latest issue);

  • MEDLINE (1950 to present);

  • EMBASE (1980 to present);

  • Science Citation Index Expanded (1945 to present).

A search strategy designed for use in MEDLINE is described in Appendix 1. This strategy will be used as a basis for all other searches and will be adapted, where necessary, for use in other electronic databases.

Searching other resources

We will check the reference lists of all relevant studies and contact experts in the field in order to identify ongoing and completed studies. We will run general Internet searches, using terms selected from the search strategy, to identify grey literature.

Data collection and analysis

We will conduct the review according to the recommendations of The Cochrane Collaboration (Higgins 2008), as well as the Cochrane Injuries Group. We will conduct the review and analyses using Review Manager 5 (Review Manager (RevMan)).

Selection of studies

MAB and ANG will independently identify trials for inclusion using the criteria specified above, as well as searching the references of these studies for further relevant trials. We will list excluded studies and will provide the reasons for exclusion. We will not apply any language or publication status restrictions.

Data extraction and management

MAB and ANG will independently extract the following data from identified trials:

  • Year and language of publication.

  • Year of study.

  • Inclusion and exclusion criteria.

  • Sample size.

  • Mechanism of injury.

  • Type, dose, and route of prophylactic antibiotic.

  • Incidence of complications.

Assessment of risk of bias in included studies

Methodological quality

MAB and ANG will assess the methodological quality of each trial independently. MAB and ANG will register whether or not the authors of the trials made their analysis using an intention-to-treat method. Any information that is unclear or missing will be clarified through contact with the authors of the specific trial. Differences in opinion between the review authors during extraction of data will be resolved through discussion. JAG will serve as arbitrator, should differences in opinion persist.

MAB and ANG will assess the methodological quality of the trials independently, without masking of trial names. The review authors will follow the instructions given in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2008). Due to the risk of biased overestimation of intervention effects in randomised trials that are of inadequate methodological quality (Kjaergard 2001; Moher 1998; Schulz 1995), the review authors will look at the influence of the methodological quality of the trials on the results by evaluating the reported randomisation and follow-up procedures in each trial. Where information is not available in the published trial reports, MAB and ANG will contact the authors in order to assess the trials correctly.

The review authors will assess generation of allocation sequence, allocation concealment, blinding, and follow up. This information will be presented in the risk of bias table in the review.

Allocation concealment

Allocation concealment will be judged to have been:

Adequate: if performed by means of centralised, or pre-numbered containers administered serially to patients, or an on-site computer with allocations in a locked unreadable file, or sequentially-numbered, sealed, opaque envelopes.

Unclear: if the trial is described as randomised, but the trial report fails to describe the method of allocation concealment.

Inadequate: if a completely transparent procedure was used, e.g. case record numbers, dates of birth, or an open list of random numbers.

Allocation sequence generation

Allocation sequence generation will be judged to have been:

Adequate: if a computer-generated or random-number table was used.

Unclear: if the trial is described as randomised, but the trial report fails to describe the method of allocation sequence.

Inadequate: if patients were allocated according to names, dates, admittance numbers, etc. Trials using these methods are known as quasi-randomised trials, and will be excluded from the review.


Blinding will be judged to have been:

Adequate: if the trial was described as double blind, and the method of blinding was described.

Unclear: if the trial was described as double blind, but the method of blinding was not described.

Not performed: if the trial was not double blind.

Follow up

Follow up will be judged to have been:

Adequate: if the numbers and reasons for dropouts and withdrawals in all intervention groups were described, or if it was specified that there were no dropouts or withdrawals.

Unclear: if the trial report gave the impression that there had been no dropouts or withdrawals, but this was not specifically stated.

Inadequate: if the numbers or reasons for dropouts and withdrawals were not described.

MAB and ANG will record sample sizes and durations of follow up.

We will use a funnel plot to explore bias (Macaskill 2001). We will perform the linear regression approach described by Egger et al to determine funnel plot asymmetry (Egger 1997).

Measures of treatment effect

Dichotomous data will be analysed for relative risk (RR), odds ratio (OR), and the absolute effects will be measured with risk differences (RD). Confidence intervals (CI), at 95%, will be calculated for these measures of effect. Treatment effect will also be considered by intention-to-treat analysis, using available case analysis and analysis by imputation. The Mantel-Haenszel method will be used for the meta-analysis (Greenland 1985; Mantel 1959). Results will be presented as a forest plot.

Subgroup analysis and investigation of heterogeneity

Statistical heterogeneity will be tested using the I2 test (Higgins 2002), and chi-squared test, with a P-value of 0.10 representing statistical significance. If heterogeneity is identified, we will consider performing subgroup analyses. Subgroups that will then be considered include: injury severity scores, time to surgical intervention, organs injured, different antibiotics, and trials with a low risk of bias (adequate generation of allocation sequence, allocation concealment, blinding, and follow up) compared to trials with a high risk of bias (one or more of the four components of methodological quality judged as being inadequate or unclear). If the results in the fixed-effect and random-effects models do not differ, we will report the fixed-effect model. Otherwise, we will report the results of both models.


The Cochrane Injuries Group editorial-base staff for their support, and advice in writing the protocol and review.


Appendix 1. Search strategy

MEDLINE (March 2008)
1. exp Abdominal Injuries/
2. ((abdominal or abdomen) adj3 (injur* or trauma* or perforat* or penetrat*)).ab,ti.
3. ((splenic or spleen) adj3 rupture*).ab,ti.
4. ((stomach or gastric) adj3 (rupture or perforation or injur* or burst*)).ab,ti.
5. ((stab* or gunshot or shot or penetrat* wound*) adj3 (abdomen* or abdominal or stomach or splenic or spleen)).ab,ti.
6. exp Wounds, Stab/
7. exp Wounds, Gunshot/
8. exp Rupture/
9. 6 or 7 or 8
10. (abdomen* or abdominal or stomach or splenic or spleen).ti,ab.
11. 9 and 10
12. 1 or 2 or 3 or 4 or 5 or 11
13. exp Antibiotic Prophylaxis/
14. exp Gentamicins/
15. exp Cefazolin/
16. (antibiotic* adj5 (prophylaxis or prophylactic* or premedication*)).ab,ti.
17. (Cefazolin or cephazolin or cefamezine or cephamezine or ancef or cefamedin or gramaxin or kefzol or totacef or Tazobactam piperacillin or tazocin or Gentamicin* or gentamycin* or g?myticin or gmyticin or gentacycol or gentavet or genticin or garamycin* or co?amoxiclavulanic acid* or augmentin).ab,ti.
18. 13 or 14 or 15 or 16 or 17
19. 12 and 18
20. (randomised or randomized or randomly or random order or random sequence or random allocation or randomly allocated or at random or controlled clinical trial$).tw,hw.
21. clinical
22. randomized controlled
23. 20 or 21 or 22
24. exp models, animal/
25. exp Animals/
26. exp Animal Experimentation/
27. exp Disease Models, Animal/
28. exp Animals, Laboratory/
29. or/24-28
30. Humans/
31. 29 not 30
32. 23 not 31
33. 19 and 32

What's new

6 April 2009AmendedCorrection of text formatting problem


Protocol first published: Issue 4, 2008

Contributions of authors

MAB and ANG were both responsible for the writing of the protocol as well as devising the search strategy.

Declarations of interest

None known.