The use of antibiotics in the initial treatment of open fractures of the extremities is almost universal in all high-income countries, where benefits are assumed to outweigh potential risks. In the USA, a 100% compliance level is one of many required criteria for hospital re-credentialing by the Joint Committee for Accreditation of Healthcare Organisations (JCAHO). In many lower-income countries antibiotics may not be used routinely for various reasons: costs, lack of knowledge, low levels of suspicion of infection or poor case-recognition from first-line health care workers, availability, accessibility, and use of traditional or alternative health care. In some of these countries, the use of antibiotics for open fractures is delayed until the patient is seen in the secondary or even the tertiary care centres. This delay often exceeds the accepted "golden period" for treatment of six to eight hours used in high income countries. Since an open fracture is by definition contaminated, the use of antibiotics is therapeutic, not prophylactic, and aims at preventing subsequent infectious problems such as cellulitis, myositis, acute or chronic osteomyelitis (bone infection), infected non-unions, recurrent abscesses, chronic drainage with fistula formation, and their associated impairment and disability.
A related systematic review (Gillespie 2001) concluded that antibiotic prophylaxis should be offered to those undergoing surgical treatment of closed hip and other long bone fractures. Our purpose is to review the evidence for use of antibiotics in the treatment of open fractures of the extremities.
To review the evidence for the effectiveness of antibiotics in the initial treatment of open fractures of the limbs.
Criteria for considering studies for this review
Types of studies
Randomised and quasi-randomised (e.g. date of birth, alternation) controlled trials.
Types of participants
People of any age with open fractures of the limbs.
Types of interventions
Antibiotic administered before or at the time of primary treatment of the open fracture compared with placebo or no antibiotic. Trials comparing different antibiotics, different antibiotic dosages, route of administration or differences in timing or duration of administration were excluded.
Types of outcome measures
- early wound infection (as defined in individual study reports)
- chronic wound discharge
- acute or chronic osteomyelitis
- delayed union or non-union
Search methods for identification of studies
We searched the Cochrane Bone, Joint and Muscle Trauma Group Specialised Register (26 July 2009), the Cochrane Central Register of Controlled Trials/Clinical Trials (The Cochrane Library 2009, Issue 3), MEDLINE (1950 to July Week 3 2009), EMBASE (1980 to 2009 Week 30), Latin American and Caribbean Literature on the Health Sciences (LILACS) (1982 to July 2009), and International Pharmaceutical Abstracts (1970 to July 2009).
The search strategy for MEDLINE combined a subject-specific section with the Cochrane Highly Sensitive Search Strategy for identifying randomized trials in MEDLINE (sensitivity-maximizing version) (Lefebvre 2008), and was modified for use in other databases (see Appendix 1). The MEDLINE search strategy for the previous version is shown in Appendix 2.
We searched the WHO International Clinical Trials Registry Platform (July 2009) for ongoing trials.
Searching other resources
Proceedings of meetings of the American Academy of Orthopaedic Surgeons (1980 to 2001), the Orthopaedic Trauma Association (1990 to 2001) and the Société Internationale de Chirurgie Orthopedique et Traumatologique (1980 to 2001) were handsearched. We also searched reference lists of articles and contacted published researchers in the field.
No language restriction was applied.
Data collection and analysis
Selection of studies
Two reviewers independently screened each record for eligibility by initially examining titles, abstracts and key words. Reports identified by either reviewer were retrieved. We searched the reference lists of relevant trials and reviews in which the largest number of eligible trials had been published. We contacted the authors of eligible trials and reviews to ask about unpublished trials. Reports of potentially eligible trials were obtained and two reviewers independently assessed each one for eligibility.
Data extraction and management
Two reviewers independently extracted data from included reports using a standard pro forma. Data were extracted on the type of participants, the type of antibiotic used, the number randomised to intervention or control groups, the quality of allocation concealment, and the outcome measures stated in the protocol. We wrote to the authors of reports if relevant information was missing. Trials in which we could not confirm that random or quasi-random allocation had been used to allocate participants were excluded.
Assessment of risk of bias in included studies
Methodological quality for each study was independently assessed by two reviewers using a schedule derived from the former generic evaluation tool developed by the Cochrane Bone, Joint and Muscle Trauma Group (see Table 1), from which a risk of bias assessment was derived. Disagreements were resolved by consensus. The scores for individual items were designed to give readers a general impression of trial quality and were not used in any quantitative manner.
For each intervention, we estimated the pooled risk ratio in a fixed-effect model. We calculated 95% confidence intervals for each outcome. Selection bias was assessed using Egger's weighted regression method and Begg's rank correlation test and funnel plot. Heterogeneity was assessed using visual inspection of overlap in the forest plot and consideration of the Chi² test and I² statistic. We planned three subgroup analyses exploring whether the effect differed between placebo-controlled and no-placebo studies, whether the effect of antimicrobials differed depending on the location of the fracture (specifically comparing use in phalangeal fractures in the hand with use in fractures of major limb bones), and whether the timing of antibiotic administration was a critical factor.
Description of studies
The updated search strategy identified 13 trials in which the use of any antibiotics for open fractures of the limbs was compared with placebo or no antibiotics. Five studies were excluded (see Characteristics of excluded studies) either because there was no randomisation (two studies), patients were randomised after an initial dose of antibiotics (one), or it was impossible to disaggregate data on open fractures and soft tissue injuries only (two). The remaining eight studies were included and are described in detail in Characteristics of included studies.
There were three randomised double-blinded trials (Bergman 1982; Braun 1987; Stevenson 2003), three quasi-randomised trials (Rojczyk 1983; Dickey 1989; Suprock 1990), and two trials where the randomisation methods were unclear (Patzakis 1974; Sloan 1987).
Three studies were limited to open fractures of fingers (Sloan 1987; Stevenson 2003; Suprock 1990), while four excluded hand and finger fractures (Bergman 1982; Braun 1987; Dickey 1989; Rojczyk 1983). Patzakis 1974 included all fractures of the extremities. Four studies included only patients in the ill-defined 'adult age group' (Bergman 1982; Braun 1987; Dickey 1989; Sloan 1987). Two studies included patients of all ages (Patzakis 1974; Rojczyk 1983) and one study did not mention age (Suprock 1990). Participants in Stevenson 2003 were aged over 16 years.
Four studies used a placebo (Bergman 1982; Braun 1987; Sloan 1987; Stevenson 2003). Antibiotic regimens differed but all employed penicillin derivatives or first generation cephalosporins, active against gram-positive organisms: Penicillin/streptomycin and cephalothin (Patzakis 1974); penicillin and dicloxacillin (Bergman 1982); cloxacillin (Braun 1987); flucloxacillin (Stevenson 2003); cefazolin (Dickey 1989; Rojczyk 1983); first generation cephalosporin, dicloxacillin or erythromycin (Suprock 1990); or cephradine (Sloan 1987). There were also differences in the duration and route of administration of treatment.
Risk of bias in included studies
The risk of bias in the included studies varied (see Figure 1 and Figure 2). Only Stevenson 2003 provided information on sequence generation for randomisation. Concealment of allocation and blinding were judged adequate in three studies (Bergman 1982; Braun 1987; Stevenson 2003). Hence, these three trials were considered to be at a lower risk of bias than the other five trials. The trialists' reported definitions of outcome varied, as did length of follow-up periods. Wound infection required microbiological confirmation in five studies (Bergman 1982; Braun 1987; Patzakis 1974; Rojczyk 1983; Sloan 1987).
|Figure 1. Risk of bias summary: review authors' judgements.|
|Figure 2. Funnel plot of comparison: 1 Any antibiotic treatment versus control, outcome: 1.1 Early wound infection.|
Effects of interventions
Although there was no uniform definition of the primary outcome measure, we considered that those measures used in each study were clinically sufficiently consistent to permit pooling. The data did not permit the analysis of secondary outcomes (osteomyelitis, chronic drainage, infected non-union, amputation for infection, or infection-related death).The antibiotic regimens varied significantly in terms of agents used and duration, but we found each regimen likely to be effective at the time of the study against the gram-positive organisms usually associated with early infection in open fractures.
The data from 1106 participants in the eight studies were pooled ( Analysis 1.1), subgrouped by lower and higher risk of bias (see Figure 1). In the overall analysis, antibiotics significantly reduced the incidence of wound infection (risk ratio (RR) 0.43, 95% confidence interval (CI) 0.29 to 0.65). The absolute risk of wound infection was 0.11 (53/461) in the controls (no antibiotic) and 0.05 (33/645) in those receiving any antibiotics, giving an absolute risk reduction of 0.07 (95% CI 0.03 to 0.10) ( Analysis 1.2). In the subgroup analysis, the lower risk of bias group had an absolute risk of wound infection of 0.13 (22/169) for the controls and 0.04 (9/201) for those receiving antibiotics (risk difference -0.09 (95% CI -0.03 to -0.15), while the no-placebo group had an absolute risk of 0.11 (31/292) for the controls and 0.05 (24/444) for those receiving antibiotics (risk difference -0.05 (95%CI -0.01 to -0.10). The difference between these subgroups was not statistically significant.
Subgroup analysis ( Analysis 1.3) explored the importance of fracture location and found that in open finger fractures, there was no evidence of significant benefit from antibiotics (3 trials, 367 participants; RR 0.56, 95% CI 0.26 to 1.23). In trials which did not include open finger fractures, antibiotics significantly reduced early infection (4 trials, 472 participants, RR 0.37, 95% CI 0.21 to 0.68). However, the difference between the two groups was not statistically significant.
Subgroup analysis based on the timing of antibiotic administration was not possible as the necessary data were not available.
Summary of main results
Despite the risk of bias and relative lack of power of most studies, meta-analysis supports the effectiveness of antibiotics active against gram-positive organisms in reducing the incidence of early infection when administered before or at the time of primary treatment of an open limb fracture, at least in the short term. All authors agree that antibiotic therapy is an adjunct to, not a replacement for, a comprehensive open fracture management protocol that includes early lavage, surgical debridement, fracture stabilisation when appropriate, and bone grafting and wound coverage if necessary. Subgroup analysis indicates that antibiotic prophylaxis administered to people with open finger fractures may not reduce the incidence of early infection, but the difference between subgroups is not significant. We did not include trials comparing one antimicrobial agent against another, or comparing single dose or short course administration with longer periods of prophylaxis.
Overall completeness and applicability of evidence
Although the number of included studies is small, and the review has a number of limitations in overall completeness and applicability, it establishes with some confidence the effectiveness of antibiotic prophylaxis directed at gram-positive organisms in reducing early infection after open fracture in the extremities.
We feel confident that our search strategy was thorough and comprehensive, although limited to Europe and the Americas. We did not apply any language restriction, but we were unable to search Chinese or Japanese databases. We cannot rule out the possibility of publication bias.
The limitations in completeness and applicability arise from the limited scope of the included studies as much as from their quality. They include:
1. Data from the included studies does not permit the analysis of a number of longer-term outcomes, such as chronic osteomyelitis, infected non-union, amputation for infection, or infection-related death.
2. Although raising the hypothesis that antibiotics may possibly not be effective in the case of relatively low energy injuries (phalangeal fractures in the finger), the available data did not allow exploration of whether effectiveness differs with the severity or grade of the open wound.
3. Available data did not allow exploration of the optimal duration of prophylaxis in respect of either effectiveness or adverse effects.
4. This review does not address some of the important issues faced in the management of major trauma in the first decade of the 21st century, in which the treatment of infection caused by antimicrobial resistant organisms, both gram-positive and gram-negative, plays an important part. The included studies did not address the question of whether antibiotic prophylaxis contributes to the development of antimicrobial resistance, as a number of studies in the last decade have suggested. Nor did they address the question of whether prophylaxis should take into account the increasing incidence of multiple antibiotic resistance amongst gram-positive organisms, particularly Staphylococcus aureus. Nor did they provide useful data on the incidence of hypersensitivity reactions to the antibiotic agents used.
5.The extent to which results from studies carried out in developed countries, where antibiotics are part of a comprehensive open fracture management protocol, are applicable to developing countries remains debatable. We found no studies from developing countries comparing antibiotics as the main, or even the only initial treatment of open fractures of the limbs, to placebo or no treatment. Nevertheless, it seems reasonable to assume that, at least to some extent, the findings of this review would apply to low-resource settings. If good data on open fracture prevalence and incidence were available for a given population, it could be possible to estimate the morbidity burden that might be avoidable. An economic analysis taking the local context into account could then evaluate the costs and benefits.
Although it seems somewhat unlikely that further placebo-controlled trials of antibiotic prophylaxis for open fracture management could be justified, future studies could certainly address longer-term outcomes and duration of prophylaxis (probably single dose versus a longer but restricted regimen). Dellinger 1991 listed recommendations for the conduct and reporting of clinical trials on the management of open fractures which should improve both study quality and applicability of data.
Quality of the evidence
Only three placebo-controlled included studies (Bergman 1982; Braun 1987; Stevenson 2003 ) were judged to have both adequate allocation concealment and blinding. The subgroup analysis which shows a larger effect in these studies with lower risk of bias than in those with higher risk of bias is reassuring. Another potential weakness comes from the relatively short periods of follow-up in most studies, although it is impossible to estimate in which direction this would bias the results. On the other hand, very few patients were lost to follow-up, and all data analyses appear to have been intention-to-treat analyses.
The precision of the results of this systematic review may be decreased by the relatively small numbers of included studies, and participants. This is unlikely to change in the future. It is doubtful whether further trials using placebo or no antibiotics will be justifiable, as the use of antibiotics has been part of the standard care of open fractures of the extremities since the mid 1970s, despite some controversy about its use in open fractures of the fingers.
Agreements and disagreements with other studies or reviews
The Eastern Association for the Surgery of Trauma (EAST 2000) has published a literature review leading to "practice management guidelines for prophylactic antibiotic use in open fractures" which is in agreement with the findings of this review. More recently, Hauser 2006, although noting that "current antibiotic management of open fractures is based on a small number of studies that generally are more than 30 years old and do not reflect current management priorities in trauma and critical care", concluded, as this review does, that a short course of an agent effective against gram-positive organisms such as a first-generation cephalosporin "begun as soon as possible after injury, significantly lowers the risk of infection when used in combination with prompt, modern orthopedic fracture wound management".
A recent systematic review and economic analysis (Cranny 2008) has addressed the issue of whether the choice of agent for routine antibiotic prophylaxis for surgery should take account of multiple antibiotic resistance in Staphylococcus aureus (MRSA) by a switch from non-glycopeptide to glycopeptide prophylaxis. It concluded that "There is insufficient evidence to determine whether there is a threshold prevalence of MRSA at which switching from non-glycopeptide to glycopeptide antibiotic prophylaxis might be clinically effective and cost-effective. Future research needs to address the complexities of decision-making relating to the prevention of MRSA and infection control in general. Research including evidence synthesis and decision modelling comparing a full range of interventions for infection control, which extends to other infections, not just MRSA, is needed."
The problem of adverse effects of antibiotic prophylaxis, in particular hypersensitivity reactions, the development of antibiotic associated diarrhoea, and the development of antibiotic resistance to which this review was unable to contribute any evidence has recently been reviewed in a contemporary Practice Guideline on antibiotic prophylaxis in surgery (SIGN 2008).
Implications for practice
The use of a prophylactic antibiotic regimen effective against gram-positive organisms (for example a narrow spectrum beta-lactam agent such as dicloxacillin or flucloxacillin, or a first generation cephalosporin), begun as soon as possible after injury, significantly lowers the risk of early infection after an open fracture when used in combination with good wound management.
Although no relevant data could be provided by this review, health professionals and people at risk should ensure their awareness of the adverse effects of antibiotic prophylaxis.
Implications for research
1. Further placebo controlled trials to evaluate the effectiveness of antibiotics for open fractures of the limbs proximal to the phalanges are unwarranted. There may be a case for further evaluation of the effectiveness of antibiotic prophylaxis in open finger fractures.
2. For those populations for which antibiotics are not universally used in the management of open fractures, reliable epidemiological data on the incidence and prevalence of open fractures, prevalence of antibiotic utilisation, and incidence and prevalence of infectious complications would allow quantifying the avoidable burden associated with this problem.
3. Both in these low-resource settings, where management of open fractures is inconsistent, and in settings where antibiotic administration is routine, there may be justification for further studies comparing the benefits of single versus multiple doses of antibiotics, and which antibiotic by which route.
4. Economic analysis would then allow the most cost-effective intervention regimen to be determined, and to value the costs and benefits associated with burden reduction.
The authors would like to acknowledge the following for their help in the preparation of the original review, and this update: Ms Leeann Morton, Mrs Lesley Gillespie and Ms Frances Bunn. We would also like to thank the following for their useful comments during the editorial review process: Professor John Stothard, Professor Marc Swiontkowski, Professor Rajan Madhok, Associate Professor Peter Herbison and Dr Janet Wale.
Data and analyses
- Top of page
- Authors' conclusions
- Data and analyses
- What's new
- Contributions of authors
- Declarations of interest
- Index terms
Appendix 1. Search strategies
The Cochrane Library (Wiley InterScience interface)
#1 MeSH descriptor Antibiotic Prophylaxis, this term only
#2 MeSH descriptor Anti-Bacterial Agents explode all trees
#3 (antibiotic* or antimicrob*):ti,ab,kw in Clinical Trials
#4 (#1 OR #2 OR #3)
#5 MeSH descriptor Infection, this term only
#6 MeSH descriptor Wound Infection explode all trees
#7 MeSH descriptor Sepsis explode all trees
#8 (infect*):ti,ab,kw in Clinical Trials
#9 (#5 OR #6 OR #7 OR #8)
#10 MeSH descriptor Fractures, Bone explode all trees
#11 MeSH descriptor Fracture Fixation explode all trees
#12 (fractur*):ti,ab,kw in Clinical Trials
#13 (#10 OR #11 OR #12)
#14 (#4 AND #9 AND #13) (99 records identified)
MEDLINE (Ovid interface)
1 Antibiotic Prophylaxis/
2 exp Antibiotics/
3 (antibiotic$ or antimicrob$).tw.
6 exp Wound Infection/
10 exp Fractures, Bone/
11 exp Fracture Fixation/
15 Randomized Controlled Trial.pt.
16 Controlled Clinical Trial.pt.
19 drug therapy.fs.
24 exp Animals/ not Humans/
25 23 not 24
26 and/14,25 (515 records identified)
EMBASE (Ovid interface)
1 exp Antibiotic Agent/
2 (antibiotic$ or antimicrob$).tw.
4 exp Infection/
5 Infection Prevention/
6 Infection Complication/
8 exp Fracture/
9 exp Fracture Treatment/
13 exp Randomized Controlled Trial/
14 exp Double Blind Procedure/
15 exp Single Blind Procedure/
16 exp Crossover Procedure/
17 Controlled Study/
19 ((clinical or controlled or comparative or placebo or prospective$ or randomi#ed) adj3 (trial or study)).tw.
20 (random$ adj7 (allocat$ or allot$ or assign$ or basis$ or divid$ or order$)).tw.
21 ((singl$ or doubl$ or trebl$ or tripl$) adj7 (blind$ or mask$)).tw.
22 (cross?over$ or (cross adj1 over$)).tw.
23 ((allocat$ or allot$ or assign$ or divid$) adj3 (condition$ or experiment$ or intervention$ or treatment$ or therap$ or control$ or group$)).tw.
26 limit 25 to human
27 and/12,26 (524 records identified)
LILACS (Clinical Trials in LILACS)
((Pt RANDOMIZED CONTROLLED TRIAL OR Pt CONTROLLED CLINICAL TRIAL OR Mh RANDOMIZED CONTROLLED TRIALS OR Mh RANDOM ALLOCATION OR Mh DOUBLE-BLIND METHOD OR Mh SINGLE-BLIND METHOD OR Pt MULTICENTER STUDY) OR ((tw ensaio or tw ensayo or tw trial) and (tw azar or tw acaso or tw placebo or tw control$ or tw aleat$ or tw random$ or (tw duplo and tw cego) or (tw doble and tw ciego) or (tw double and tw blind)) and tw clinic$)) AND NOT ((CT ANIMALS OR MH ANIMALS OR CT RABBITS OR CT MICE OR MH RATS OR MH PRIMATES OR MH DOGS OR MH RABBITS OR MH SWINE) AND NOT (CT HUMAN AND CT ANIMALS)) [Palavras] and (antibiotic$ OR antimicrob$) [Palavras] and (fractur$) [Palavras] (1 record identified)
International Pharmaceutical Abstracts (Ovid interface)
1. (antibiotic$ or antimicrob$ or infect$ or septic or sepsis).tw.
2. (fractur$ and open).tw.
3. and/1-2 (10 records identified)
Appendix 2. Search strategy for MEDLINE (OVID WEB) in previous versions
1. Antibiotic Prophylaxis/
2. exp Antibiotics/
3. (antibiotic$ or antimicrob$).tw.
6. exp Wound Infection/
11. Fractures, Open/
12. exp Fractures/
13. (open or compound).tw.
15. (infect$ adj3 (bone$ or fracture$)).tw
Last assessed as up-to-date: 27 July 2009.
Protocol first published: Issue 3, 2002
Review first published: Issue 1, 2004
Contributions of authors
All reviewers contributed to the protocol development and the editing of the review. One reviewer (RG) executed the search strategy. All reviewers independently assessed trial quality and extracted data. WJG contributed to the drafting of this update. Richard Gosselin is the guarantor of the review.
Declarations of interest
Medical Subject Headings (MeSH)
MeSH check words
* Indicates the major publication for the study