Antibiotic prophylaxis is frequently administered in severe trauma. However, the risk of selecting resistant bacteria, a major issue especially in critical care environments, has not been sufficiently investigated. The aim of the present study was to provide guidelines for antibiotic prophylaxis for four different trauma-related clinical conditions, taking into account the risks of antibiotic-resistant bacteria selection, thus innovating previous guidelines in the field.
The MEDLINE database was searched for studies comparing antibiotic prophylaxis to controls (placebo or no antibiotic administration) in four clinical traumatic conditions that were selected on the basis of the traumatic event frequency and/or infection severity. The selected studies focused on the prevention of early ventilator associated pneumonia (VAP) in comatose patients with traumatic brain injury, of meningitis in severe basilar skull fractures, of wound infections in long-bone open fractures. Since no placebo-controlled study was available for deep surgical site-infections prevention in abdominal trauma with enteric contamination, we compared 24-hour and 5-day antibiotic prophylaxis policies. A separate specific research focused on the question of antibiotic-resistant bacteria selection caused by antibiotic prophylaxis, an issue not adequately investigated by the selected studies. Randomised trials, reviews, meta-analyses, observational studies were included. Data extraction was carried out by one author according to a predefined protocol, using an electronic form. The strength of evidence was stratified and recommendations were given according to the Grading of Recommendations Assessment, Development and Evaluation (GRADE) criteria.
Uncertain evidence deserving further studies was found for two-dose antibiotic prophylaxis for early VAP prevention in comatose patients. In the other cases the risk of resistant-bacteria selection caused by antibiotic administration for 48 hours or more, outweighed potential benefits.
Citation: Poole D, Chieregato A, Langer M, Viaggi B, Cingolani E, Malacarne P, et al. (2014) Systematic Review of the Literature and Evidence-Based Recommendations for Antibiotic Prophylaxis in Trauma: Results from an Italian Consensus of Experts. PLoS ONE 9(11): e113676. https://doi.org/10.1371/journal.pone.0113676
Editor: Zhi Zhou, Purdue University, United States of America
Received: June 24, 2014; Accepted: October 27, 2014; Published: November 20, 2014
Copyright: © 2014 Poole et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: The authors confirm that all data underlying the findings are fully available without restriction. All relevant data are within the paper and its Supporting Information files.
Funding: The authors have no support or funding to report.
Competing interests: The authors have declared that no competing interests exist.
In the attempt to investigate antibiotic prophylaxis in several trauma-related clinical conditions, taking into account the potential ecological risk, innovating previous approaches, in February 2013 a meeting of experts was held in Bologna in the context of the 8th edition of the Trauma Update and Organization conference.
Infection is the consequence of contamination due to the contact between normally sterile tissues and microorganisms from the external environment or from internal sites. The inoculation occurs when the trauma breaks natural barriers against infections. All attempts to reduce the bacterial charge in the tissues should be made. Cleaning, irrigation, disinfection of wounds, followed by the appropriate early surgical treatment when indicated, are crucial for infection prevention. Antimicrobial prophylaxis is only one element integrating these interventions and should be limited to the period immediately after the trauma.
One major drawback of exposure to antibiotics is the emergence of drug-resistant bacteria strains, as the result of selection pressure on resistant colonizing strains or on bacteria with newly acquired mutations. Unfortunately, most studies dealing with antibiotic prophylaxis have not adequately investigated this issue, which is particularly relevant in critical care settings where bacterial resistance to antibiotics has become a major problem. Here, we focused on studies comparing antibiotic prophylaxis and placebo in reducing infections in several trauma-related conditions taking into account the risk of antibiotic-resistant bacteria selection to produce a fair risk/benefit ratio assessment, in contrast with other guidelines addressing similar issues , .
In the present paper, we report the contents of the review, the analysis and discussion of the literature, and the conclusions of the conference.
The works was organized in three sequential steps:
Step 1: The Coordinating Committee formed by five intensivists selected an initial list of topics. Head, thorax, abdomen, and limbs trauma were considered. Because of time constraints (there were only two days to discuss the issues during the convention) only four clinical conditions were selected based on traumatic event frequency and/or infection severity. The scenarios included antibiotic prophylaxis for 1) early VAP in comatose patients, 2) meningitis in severe skull fractures, 3) wound infection in long-bone open fractures (we initially focused on osteomyelitis but found no study specifically investigating this issue), 4) deep surgical-site infections in abdominal trauma with hollow viscus perforation. We also considered but discarded pneumonia in thoracic contusion, empyema after chest tube insertion, surgical-site infection for long-bone closed fractures, infection after soft tissues open injuries, abdominal trauma without bowel perforation, and inhalation pneumonia in pre-hospital intubation.
The Coordinating Committee then generated the queries for the literature search and selected the articles fulfilling the criteria established for each subject. The fundamental criterion, shared by all the investigated issues, was the comparison between adult patients subjected to antibiotic prophylaxis and controls. When such studies were not available comparison between groups receiving different pharmacological protocols were considered.
The MEDLINE database was searched using the free PubMed provider. Randomised trials, reviews, meta-analyses, or observational studies were selected. Pediatric and animal studies, and articles not written in English, were excluded. The MEDLINE search covered a period between January 1970 and January 2014. This search was integrated with the consultation of reference lists of retrieved articles.
The Coordinating Committee members unanimously selected the articles after reviewing the abstracts, with the help of one of the authors with specific expertise in statistics and research methodology. Full-texts were then retrieved and a further selection was carried out.
The statistical and methodological expert also performed data extraction according to a predefined plan (the protocol was not pre-registered), using a dedicated electronic form that automatically performed all planed computations and plot generation. The electronic form was developed ad hoc for this revision, and tested and refined using the studies on early VAP prevention with antibiotic prophylaxis.
Data collected included study inclusion criteria, antibiotic class and dose, duration of prophylaxis, antibiotic-resistant bacteria selection when investigated, infection rates, and mortality rates when available, along with information needed for the GRADE evidence quality assessment.
Data were presented as event rates in treatment arms and controls, absolute risks, absolute risk reductions, and relative risks. Adjusted odds ratios were reported when dealing with multivariate analysis results in observational studies. Numbers needed to treat were also calculated. Confidence intervals were computed for all the above measures. Relative risks were graphically represented in Forest plots.
Design heterogeneity among selected studies advised against their combination in meta-analyses. For the same reason statistical approaches to investigate presence of asymmetry and potential publication bias were not carried out .
The queries and the results of the summarised results were submitted to the External Scientific Advisory Panel, formed by four intensivists with expertise in the field of infections and trauma in critical care settings, who discussed them together with Coordinating Committee members.
Evidence provided by randomised controlled trials (RCTs) and observational studies was ranked on the basis of the Grading of Recommendations Assessment, Development and Evaluation (GRADE) criteria. – The GRADE provides a framework to guide reviewers for the assessment of the strength of evidence, taking into account that different reviewers or reviewing groups can evaluate the same evidence differently.  Moreover, it eliminates automatic links between the strength of the recommendation and the strength of the evidence, leaving to the reviewers the possibility of grading recommendations based on other variables relevant to clinical decisions, besides evidence.
The GRADE stratifies quality of evidence on a four-level scale, ranging from high to very low, , in favour or against intervention. In the middle of this scale an absence of evidence level was added. This event occurs frequently in studies with small sample size, where the alternative hypothesis and statistical power have not been adequately taken into account. In our opinion, negative results from such studies should thus not be considered indicative of no effect, but rather of an unresolved issue .
The reporting of RCTs was evaluated on the basis of currently shared standards.  We considered quasi-randomised studies, in which treatment allocation is based on non-random methods such as alternation, leading to severe bias in regards to sequence generation and allocation concealment.
Our evaluation of meta-analyses was restrictive, according to the principle that heterogeneity among studies in terms of inclusion criteria, study design, study treatment, type of outcome, which are not accounted for by statistical methods focused on heterogeneity of outcome rates,  should contraindicate the merging of their results.  When dealing with observational studies we discarded al investigations that reported only crude-data comparisons without performing adjustment for potential confounders.
Step 2: The results were presented to the physicians invited to the symposium - intensivists, infectious disease specialists, surgeons, emergency physicians, radiologists, and physiatrists – and the issues were collectively discussed. The discussion was audio-recorded and at a later stage revised by members of the Coordinating Committee to identify and highlight the main opinions that emerged from the discussion.
Step 3: the scientific advisors and the Coordinating Committee members formed the jury that provided recommendations that were presented as preliminary results on the final day of the conference, but discussed and refined in the following months by e-mail. A final meeting was held one year after the conference, to discuss diverging opinions until unanimity was reached. To develop the recommendations the panel followed the GRADE framework. This is based on four points: quality of evidence, balance between desirable and undesirable effects, values and preferences, uncertainty about whether the intervention represents a wise use of resources .
The GRADE initially attributes a high default level of evidence to RCTs and a low level to observational studies.  RCTs can then be up- or downgraded according to the following criteria: study limitations due to insufficient internal validity, indirectness of evidence, imprecision of results, inconsistency between studies, and selective publication of studies. – While the first three are related to single studies, the last two points take into account the relation with other studies in terms of inconsistent results and publication bias, respectively. We added an other criterion which included other weak features of the study, such as the unbalance in important variables between study arms in RCTs that may occur when the sample size is small.
For RCTs the first criterion, i.e. bias, takes into account allocation concealment and blinding, completeness of follow-up and adherence to the intention-to-treat principle, premature study interruptions for benefit, and selective reporting of outcomes.  The last criterion takes into account the existence and availability of a study protocol or on the clear omission to report important outcomes.  We were not able to comply with this criterion because, for most RCTs, protocols were not available. For each of these items, concerns can be serious or very serious, down-scoring the evidence by one and two points respectively. For example, a RCT with an initial high-evidence grading can be downgraded to moderate or low, or even very low, if there are other weak points.
Limits in observational studies can include lack of controls, as in case series, or, when controls are present, lack of adjustment for important covariates, which may be unevenly distributed between study groups .
Initially, evidence is rated only one point up when the relative risk is between 0.5 and 0.2 or between 2 and 5, and two points up when lower than 0.2 or higher than 5, obviously when dealing with statistically significant results. However, even in the presence of large effects, when imprecision is strong (i.e. wide 95% confidence intervals) and negligible effects cannot be ruled out, upgrading is not performed. The rating can also be increased when a dose-response is evident or plausible potential confounders would only increase the confidence in the results , .
The strength of recommendations can be strong or weak, in favour or against interventions.  Only when an intervention with a strong likelihood of effectiveness had not been sufficiently investigated (i.e. absence of evidence as defined above) we considered the possibility of not giving any recommendation, although we usually gave a weak negative recommendation in most cases of absence of evidence, in the light of the limited resources available.
The conference dealt with four different issues of broad clinical interest but focusing especially on relevance for critically ill patients and from the point of view of intensive care unit (ICU) and emergency department physicians.
To investigate the main undesirable effects of prophylaxis, we preliminarily performed a systematic research on the effect of short antibiotic administration on bacterial antibiotic-resistance selection, to provide a fair assessment of the balance between benefits and risks. The results of this investigation integrated, when appropriate, the evidence concerning antibiotic prophylaxis in the clinical conditions we explored.
A transparent report of the reviewing process and of evidence grading is provided for each query in the supporting information.
We tried to find an answer to the following question: Does short antibiotic prophylaxis cause resistant bacteria selection? To do this we explored the influence of short antibiotic course on resistant bacteria selection at two levels: at the patient’s level, investigated before and after exposure, and at the centre level, comparing restrictive and liberal antibiotic prescription strategies between centres or within a single centre before and after the policy change.
We also assumed that there is strong evidence that prolonged antibiotic exposure (say, 7–10 days) increases the risk of colonization with resistant bacterial strains,  and without formally investigating this issue with the GRADE approach, we attributed to it a default high level of evidence.
Selection of antibiotic-resistant bacteria in patients exposed to prophylaxis
The contemporaneous combination of the mutant bacteria and antibiotic exposure determines the spread of antibiotic-resistant bacteria.  Although patients admitted to the hospital after trauma are usually young and healthy, given the presence of resistant bacteria also within the community, the possibility that they may be already colonized by such strains cannot be ruled out. ,  Moreover, colonization in non-colonized patients can occur because of newly acquired bacterial mutations, selection of bacterial resistant strains among heterogeneous populations,  and horizontal transfer,  the latter being particularly relevant in ICU environments.  We made an extensive literature revision focusing on antibiotic-resistant colonization in patients undergoing short antibiotic treatment, but RCTs in this field were unavailable. Thus, we had to rely on the few observational studies found (Table S5 in File S5).
In a study carried out on a cohort of 2,641 patients receiving antibiotic prophylaxis for cardiac surgery, of 1,094 undergoing cultures 426 were shown to harbour bacteria resistant to the antibiotic they were exposed to. Logistic regression, adjusting for many possible confounders, indicated an increased risk of resistance selection in case of antibiotic prophylaxis lasting more than 48 hours.  Overall the description of the statistical procedures was scanty, hindering quality assessment. One major limitation was that only part of the patients were submitted to cultures and that physicians in the centre were “encouraged” to keep antibiotic prophylaxis duration within 48 hours, potentially introducing selection biases that logistic regression may have not accounted for. Moreover, colonization status was not assessed before prophylaxis was administered, so it is not known whether the patients had previous colonization or acquired the resistant pattern during therapy or during the postoperative stay in a high-risk environment as the ICU.
In a before/after study involving 5 wards, the rate of acquisition within 30 days from initial antibiotic exposure, was 3% for MRSA (22/724), 2% for vancomycin-resistant enterococci (VRE) (10/430), and 1% (11/840) for ciprofloxacin-resistant Pseudomonas Aeruginosa (CRPA).  Forty per cent of newly acquired colonization with resistant bacteria was detected 48 hours from the first antibiotic dose. The initial colonization rate was 16% of 864 patients for MRSA, 50% for VRE, and 3% for ciprofloxacin-resistant Pseudomonas Aeruginosa. Cephalosporins, quinolones, and macrolides were the antibiotics more frequently administered (31, 29, 19% of cases, respectively). Absence of control (i.e. patient not exposed to antibiotics) although not irrelevant was not a major issue, in our opinion, because we do not expect factors other than antibiotic pressure to affect the colonization rate within a 48-hour period. A low colonization rate, as in this study, is still clinically relevant if we account for the duration and possibility of horizontal spread of resistant bacterial colonization, especially in environments such as the ICU , .
Another study investigated 29 patients undergoing cardiac surgery and 10 coronary angioplasty, receiving prophylaxis for three days with a first or a second-generation cephalosporin.  Cultures were collected at time 0 and on the third day. The investigators detected colonization by methicillin-resistant coagulase-negative staphylococci in 36 of 43 cultured sites (84%) previously not colonized and a quantitative colonization increase in 17 of 28 sites (61%) harbouring low-level colonization before antibiotic administration. The main limitation of the study was that multiple samples were collected from each patient that may have borne the same resistant bacteria, inflating the final count of true events. A reliable quantitative estimation of resistant bacteria selection was hence impossible. This study was also limited by its small sample size.
Finally, in a study investigating the effect of a five-day course with clarithromycin, 5 patients colonised by susceptible enterococcus strains few days after the end of treatment developed resistant types, which in 3 cases persisted for one year or more.  Although, almost a case report, this study is indicative of how antibiotic pressure can select resistant strains and how long the effects can last over time.
Effect of liberal vs. aggressive antibiotic prescriptive strategies
The limitation of antibiotics use is widely recognized as an effective strategy for the reduction in bacterial resistance. ,  However, the role of antibiotic prophylaxis policies, corresponding to strong limitations in antibiotic use, have never been investigated in relation to antibiotic-resistance acquisition. Indeed a recent review concerning the influence of antibiotic exposure on MRSA diffusion, the shortest period of antibiotic administration considered was 7 days.  In a before/after study dealing with late-generation cephalosporin administration policy, surgical prophylaxis was among the exceptions to restriction, indicating that the authors did not feel it was significant for the purposes of their policy .
Conclusions: there is low evidence indicating that antibiotic prophylaxis lasting 48 hours or more is capable of determining the colonization with resistant bacterial strains, that is the undesirable effect reported in the text and tables for each of the interventions under scrutiny.
Question 1: is antibiotic prophylaxis for the prevention of ventilator-associated lower respiratory tract infections effective in intubated patients with traumatic head injury (TBI) and coma, i.e. Glasgow Coma Scale (GCS) ≤8?
Results: the MEDLINE search provided only three studies, two RCTs and one observational study, – all complying with our inclusion criteria. Although, the studies did not deal exclusively with TBI patients, the results can be reasonably extended to this category of patients. The two RCTs are described in Table 1, Table S1 in File S1, and Figure S1 and tested respectively a 3-day and a single-day antibiotic prophylaxis. Both RCTs reported a statistically significant reduction of early-VAP, but no reduction of late-onset pneumonia and ICU mortality. Besides other flaws, major limitations of both RCTs were the absence of blinding, the small sample size and the lack of power to detect increases in antibiotic-resistant infections in the treatment group (outcome investigated only in the first RCT).  Moreover, the second RCT was not specifically focused on comatose patients.  The premature interruption of the first RCT after the first interim analysis when only 38 patients were randomised, on the basis of the reduction of early-onset pneumonia incidence in the treatment arm, while no difference in mortality was detected, is a questionable choice. Finally, the beneficial effect of the postponement of pneumonia in a phase of minor vulnerability of the brain in severe brain injury is only theoretical and was not demonstrated with a robust outcome, such as the long-term mortality/severe disability rate. The main undesirable effect, i.e. antibiotic-resistant bacteria selection, according to the evidence we collected, was likely to occur after three days of antibiotic administration, while probably irrelevant when only two doses are given.
The results of a recent observational study, including 129 patients suggested that a single dose of ceftriaxone was an independent predictor of reduced incidence of early-onset pneumonia.  The regression analysis was performed including a propensity score and the variable “antibiotic prophylaxis” in a regression model, but the number of outcomes (overall 15 cases of early-onset VAP) did not justify the inclusion of the two above variables according established rules. ,  Moreover, the sample size and the number of events were small, and the model was underpowered to include potential important covariates capable of influencing the significance of ceftriaxone administration in the predictive model. The observational study appeared to be, hence, unreliable and was not taken into account in our final evaluation.
Conclusions: The two RCTs had substantial different interventions and outcomes, thus we provided recommendations for two different queries (Table 2).
Three-day course with a beta-lactamase-protected penicillin for the prevention of ventilator-associated early-VAP in intubated patients with TBI and coma:
- Level of evidence in favour: Very low evidence.
- Level of evidence against: Low evidence.
- Recommendation: Weak against intervention.
Question 2: is antibiotic prophylaxis effective in patients with basilar skull fractures from non-penetrating head trauma to decrease the occurrence of meningitis?
Results: of 146 studies found only four were RCTs. One was discarded because only 10 patients were randomised while 129 were the object of a retrospective evaluation and therefore included among observational studies.  The remaining were single-centre trials, including 109, 37, and 52 patients respectively, all providing negative results, but the three studies were underpowered to demonstrate clinically relevant differences (Table 3, Table S2 in File S2, Figure S2). – The 37 patients enrolled in the second study with basilar fractures were a subset of 196 with either basilar or skull open fractures. Thus the findings of interest for our review were the result of a subgroup analysis, and as such biased.  The other two trials also suffered from methodological bias. Most important long-lasting antibiotic prophylaxis was administered (5-day ceftriaxone, 3-day ceftriaxone or ampicillin-sulbactam, and 8-day average penicillin course, respectively) and no monitoring of antibiotic-resistant bacteria selection was investigated. This was a major undesirable effect of such prolonged antibiotic prophylaxis, as supported by the literature revision we carried out.
The negative evidence regarding this issue are summarised in a recent Cochrane review and meta-analysis  that also includes the study of the randomised ten patients mentioned above and a RCT published as correspondence.  Although the studies were not heterogeneous in terms of meningitis odds ratios, they were heterogeneous in terms of meningitis absolute rates, inclusion criteria, duration of antibiotic treatment, ceftriaxone dose in the two studies that used this antibiotic, ,  and number of patients lost to follow-up, advising against their inclusion in a meta-analysis. Two other non-recent meta-analyses were biased by the merging of RCTs and observational studies , .
We found four observational studies matching the inclusion criteria, none of which performed adjusted comparisons, thus not providing any acceptable evidence in favour of or against antibiotic prophylaxis , –.
Conclusions: The three studies under scrutiny all investigated prolonged antibiotic administration, that was thus the object of our query (Table 4).
Question 3: is antibiotic prophylaxis beneficial for the reduction of wound infection rate in patients with long-bone open fractures?
Results: of 174 studies fulfilling the search criteria, only three trials, – one observational study,  and one meta-analysis  matched the inclusion criteria. The studies were carried out in between the 1970s and the 1980s. In the first RCT, patients with open-fractures were only a subset of the randomised sample,  the findings thus being the result of a subgroup analysis and as such inadequate to provide conclusive answers.  The second study  was of higher level, but still the reporting was insufficient to allow a thorough assessment of its quality according to current standards.  In these two studies, the diagnosis required positive cultures. The third study was very low quality, with no blinding and no placebo, and a partial report of the results. It was carried out on open fractures due to gunshot, with infections diagnosis not based on cultures.  This study was negative, since 2 cases out of 32 were reported in the single-day cefazolin group and 2 out of 35 in the group not receiving antibiotic. We considered this study methodologically too weak to be considered a reliable source of evidence.
The antibiotic tested in the first two RCTs (Table 5, Table S3 in File S3, Figure S3) were instead dicloxacillin or benzyl penicillin,  and cloxacillin,  respectively. In the first study treatment was continued for 48 hours and, in the second, for ten days. The influence of prophylaxis on bacterial resistance to antibiotics was not investigated in any of these studies. We assumed that a 10-day antibiotic administration would certainly greatly increase the risk of antibiotic-resistant bacteria selection. We attributed, instead, a lower level of evidence to a 48-hour treatment (Table 6).
A 13.3 and 22.6% wound-infection rate decrease was found in the antibiotic-prophylaxis arms in the two studies, corresponding respectively to an almost statistically significant result and a fully statistically significant result. Interestingly, in the second study, infections occurring within six weeks were considered early infections, including infections that, reasonably, should have not been influenced by early antibiotic prophylaxis. In the first study, instead, the duration of the “window” period was not specified.
The observational study that matched the inclusion criteria suggested a protective effect of antibiotic prophylaxis, but since it was based only on the crude infection rate, it was not included in the analysis .
The meta-analysis included, besides the three RCT mentioned above, two other studies that we excluded because one was written in German  and the other including hand, foot, and fingers fracture, pediatric patients and gunshot injuries, without performing any subgroup analysis on long-bone fractures.  The meta-analysis concluded that antibiotic prophylaxis is effective in reducing wound infections. However, study designs, inclusion criteria, treatment protocols, and quality of the trials, were so heterogeneous to question the reliability of their merging in a single sample. We hence considered this study a potential source of misleading information.
We did not find any study comparing no antibiotic prophylaxis with 24-hours antibiotic prophylaxis, an option recommended by other guidelines presumably on the basis of indirect evidence , , , .
Conclusions: Two different antibiotic strategies were adopted in the two studies that were included in the final evaluation, leading to two answers to two different queries (Table 6).
2-day antibiotic administration for the prophylaxis of wound infections (not specifically osteomyelitis).
- Level of evidence in favour: No evidence.
- Level of evidence against: Low evidence.
- Recommendation: Weak against intervention.
Question 4: is antibiotic prophylaxis indicated to reduce the risk of deep surgical site infections in patients with abdominal trauma and enteric abdominal contamination submitted to emergent surgery?
Results: out of 504 articles resulting from the literature search, none compared treatment and no-treatment arms. Our results were consistent with those of two recent Cochrane meta-analyses that did not find any RCTs investigating the effectiveness of antibiotic prophylaxis compared to placebo in reducing infections following penetrating abdominal trauma , .
Since no direct evidence exists with regard to the effectiveness of antibiotic prophylaxis in abdominal trauma compared to placebo, under the assumption that antibiotic prophylaxis is effective in preventing surgical site infections,  we oriented the query towards the duration of antibiotic prophylaxis as it was studied by three RCTs focused on penetrating abdominal trauma and deep surgical infections occurrence, – and excluded a fourth one because it did not report the rate of deep surgical site infections.  The main limitation of the studies was that only a minority of the patients had associated hollow viscus injury. Moreover, the infection rate for those who bore intestinal perforations was not reported, making it impossible to analyse subgroup performance. Thus, it was only possible to study the number of events in relation to the overall sample, including those without perforation. Another weakness was that no evidence concerning blunt trauma, a common event in abdominal trauma, was available.
The first two studies dealt almost exclusively with deep surgical site infections, i.e. intra-abdominal abscess, peritonitis, and necrotizing fasciitis, while the third one reported disaggregated data concerning deep and superficial surgical site infections. Thus, we were able to analyse data concerning deep surgical infections. The studies recruited 317, 515, and 300 patients respectively, and infections rates were 10 vs. 8% in the 5-day and 24-hour treatment arms in the first two studies, and 6% in both arms in the third one (Table 7). Although the studies were underpowered to detect clinically meaningful differences in the outcome rate (the power to detect a 5% difference ranged between 40 and 60% in the three RCTs), looking at the range of the absolute-difference 95%-confidence intervals, there is little chance that a clinically relevant superiority of 5-day prophylaxis is present (Table 7, Table S4 in File S4, Figure S4). Moreover, on the basis of our specific evidence assessment, five days of antibiotic administration have a high probability of determining resistant-bacterial selection unlike a 24-hour therapy.
In the studies, surgical management was carried out within few hours from the injury, and early surgery appears to be a reasonable condition for short (i.e. ≤24 hours) antibiotic prophylaxis. We adopted the Infective Disease Society of America (IDSA) definition of early surgery (i.e. carried out within 12 hours from trauma).  There is currently no clear proof in favour of this cut-off but it, nevertheless, seems reasonable.
Conclusions: The review was focused on different antibiotic strategies under the assumption that antibiotic prophylaxis is effective in reducing infections compared to placebo (Table 8).
Administration of 24-hour is equivalent to 5-day antibiotic prophylaxis for surgical abdominal trauma with hollow viscus injury and surgical repair within 12 hours of trauma:
- Level of evidence in favour: Very low evidence.
- Level of evidence against: No evidence.
- Recommendation: Weak in favour intervention.
The main limitation of our study is that the literature search was limited to the MEDLINE database and only to publications in English, because of financial constraints. However, the issues we have dealt with have been largely investigated in previous reviews, and the guidelines included in the literature we have scrutinized wee used as an alternative source of references. We are therefore confident that we did not miss important publications in this field.
It is our firm belief that our recommendations, and guidelines in general, should not be clinical rules that impose specific clinical behaviours on doctors, but only indications applicable to an average patient, and should never replace the complex and individualised decisional process that physicians follow for individual patients .
In studying the role of antibiotic prophylaxis in trauma we found that in many cases the studies did not provide sufficient information for GRADE categories to be fully applied. Specifically, the assessment of the overall benefit/risk profile requires the comparison between desirable and undesirable effects. In the case of prolonged antibiotic prophylaxis (in several scrutinized studies antibiotic were administered for three days or more), the undesirable effect is the selection of antibiotic-resistant bacteria, an outcome not investigated in most of the studies.  Thus we had to integrate studies investigating exclusively infection prevention with evidence from other sources dealing with selection of resistant bacteria, a novel approach compared to other experiences that have proposed a different interpretation of evidence in this field , .
The risk of resistant mutant selection due to short antibiotic pressure has not been investigated with RCTs. We had therefore to rely on before-after studies investigating either the colonization of single patients after treatment or the effect of antibiotic policies changes in single centres.
We were able to find observational studies suggesting that a 2/3-day antibiotic course could already result in the selection of antibiotic-resistant bacteria. We graded this evidence as low or very low, due to several limitations of the studies. However, in the face of similar strength of evidence in favour and against the intervention, as in the case of the 3-day antibiotic prophylaxis for wound infections in open-fractures or early-VAP prevention in comatose patients, we considered the antibiotic-resistant resistant selection the main concern in terms of clinical impact and social relevance of the outcome., Antibiotic-resistant bacteria spread constitutes an emergency for the global health-care system, given its high burden of mortality and resource consume. – Since resistant mutations are widespread in hospitals and extending within the community, the solution to this problem appears to be the restoration of susceptible bacteria by strongly limiting antibiotic use.  Our concern is that the common use of antibiotics in prophylaxis, even for short periods, might significantly worsen this problem. Our recommendations go in this direction.
In conclusion, we would like to mention a striking report from a neurosurgical centre that in the late 1960s had to face an outbreak of resistant Klebsiella Aerogenes, which at its peak caused urinary infections in 1 of 4 patients, respiratory infections in 1 of 8, and 9 cases of meningitis 8 of which were fatal. – Though antibiotic treatment was targeted on sensitivity tests and despite the major increase in antibiotic consumption in the ward between 1966 ad 1969 the outbreak was out of control. As the authors state, “in desperation”, antibiotic use whether for treatment or prophylaxis was abandoned. This resulted in an immediate reduction of infections and the elimination of the Klebsiella from the ward without any serious adverse event being reported as the consequence of this strategy. This report although deviating from the paradigm of evidence base medicine (we are dealing with a single-centre epidemiological descriptive report with no statistical analysis performed and no precise design) nevertheless provides a demonstration of the effectiveness of a drastic antibiotic containment strategy, a potential choice when “the going gets tough”.
Absolute proportions differences and relative risks for the studies concerning the first query.
Absolute proportions differences and relative risks for the studies concerning the second query.
Absolute proportions differences and relative risks for the studies concerning the third query.
Absolute proportions differences and relative risks for the studies concerning the fourth query.
MEDLINE database search, flow diagram illustrating the literature selection process, and Table S1 illustrating evidence assessment for the first query.
MEDLINE database search, flow diagram illustrating the literature selection process, and Table S2 illustrating evidence assessment for the second query.
MEDLINE database search, flow diagram illustrating the literature selection process, and Table S3 illustrating evidence assessment for the third query.
MEDLINE database search, flow diagram illustrating the literature selection process, and Table S4 illustrating evidence assessment for the fourth query.
MEDLINE database search, flow diagram illustrating the literature selection process, and Table S5 illustrating evidence assessment for antibiotic-resistant bacteria (undesirable effect).
The authors would like to thank Fulvia Zerri for her help in the organization of the meetings and for her extraordinary support to the works of the panel.
Conceived and designed the experiments: DP AC ML BV. Analyzed the data: DP. Contributed to the writing of the manuscript: DP AC ML BV. Interpretation of data: DP AC ML BV EC PM FM GN EN LR IT AV CC GG. Drafting the work and revising it critically for important intellectual content: DP AC ML BV EC PM FM GN EN LR IT AV CC GG.
- 1. Hauser CJ, Adams CA Jr, Eachempati SR, Council of the Surgical Infection S (2006) Surgical Infection Society guideline: prophylactic antibiotic use in open fractures: an evidence-based guideline. Surg Infect (Larchmt) 7: 379–405.
- 2. Luchette FA, Bone LB, Born CT, DeLong WG, Hoff WS, et al.. (2000) EAST Practice Management Guidelines Workgroup: Practice management guidelines for prophylactic antibiotic use in open fractures. Eastern Association for the Surgery of Trauma.
- 3. Ioannidis JP, Trikalinos TA (2007) The appropriateness of asymmetry tests for publication bias in meta-analyses: a large survey. CMAJ 176: 1091–1096.
- 4. Atkins D, Eccles M, Flottorp S, Guyatt GH, Henry D, et al. (2004) Systems for grading the quality of evidence and the strength of recommendations I: critical appraisal of existing approaches The GRADE Working Group. BMC Health Serv Res 4: 38.
- 5. Guyatt G, Oxman AD, Akl EA, Kunz R, Vist G, et al. (2011) GRADE guidelines: 1. Introduction-GRADE evidence profiles and summary of findings tables. J Clin Epidemiol 64: 383–394.
- 6. Guyatt GH, Oxman AD, Vist GE, Kunz R, Falck-Ytter Y, et al. (2008) GRADE: an emerging consensus on rating quality of evidence and strength of recommendations. BMJ 336: 924–926.
- 7. Balshem H, Helfand M, Schunemann HJ, Oxman AD, Kunz R, et al. (2011) GRADE guidelines: 3. Rating the quality of evidence. J Clin Epidemiol 64: 401–406.
- 8. Altman DG, Bland JM (1995) Absence of evidence is not evidence of absence. BMJ 311: 485.
- 9. Moher D, Schulz KF, Altman DG, Consort G (2001) The CONSORT statement: revised recommendations for improving the quality of reports of parallel-group randomized trials. Ann Intern Med 134: 657–662.
- 10. Higgins JP, Thompson SG, Deeks JJ, Altman DG (2003) Measuring inconsistency in meta-analyses. BMJ 327: 557–560.
- 11. Feinstein AR (1995) Meta-analysis: statistical alchemy for the 21st century. J Clin Epidemiol 48: 71–79.
- 12. Andrews JC, Schunemann HJ, Oxman AD, Pottie K, Meerpohl JJ, et al. (2013) GRADE guidelines: 15. Going from evidence to recommendation-determinants of a recommendation's direction and strength. J Clin Epidemiol 66: 726–735.
- 13. Guyatt GH, Oxman AD, Kunz R, Brozek J, Alonso-Coello P, et al. (2011) GRADE guidelines 6. Rating the quality of evidence–imprecision. J Clin Epidemiol 64: 1283–1293.
- 14. Guyatt GH, Oxman AD, Kunz R, Woodcock J, Brozek J, et al. (2011) GRADE guidelines: 8. Rating the quality of evidence–indirectness. J Clin Epidemiol 64: 1303–1310.
- 15. Guyatt GH, Oxman AD, Kunz R, Woodcock J, Brozek J, et al. (2011) GRADE guidelines: 7. Rating the quality of evidence–inconsistency. J Clin Epidemiol 64: 1294–1302.
- 16. Guyatt GH, Oxman AD, Montori V, Vist G, Kunz R, et al. (2011) GRADE guidelines: 5. Rating the quality of evidence–publication bias. J Clin Epidemiol 64: 1277–1282.
- 17. Guyatt GH, Oxman AD, Vist G, Kunz R, Brozek J, et al. (2011) GRADE guidelines: 4. Rating the quality of evidence–study limitations (risk of bias). J Clin Epidemiol 64: 407–415.
- 18. Guyatt GH, Oxman AD, Sultan S, Glasziou P, Akl EA, et al. (2011) GRADE guidelines: 9. Rating up the quality of evidence. J Clin Epidemiol 64: 1311–1316.
- 19. Andrews J, Guyatt G, Oxman AD, Alderson P, Dahm P, et al. (2013) GRADE guidelines: 14. Going from evidence to recommendations: the significance and presentation of recommendations. J Clin Epidemiol 66: 719–725.
- 20. Liberati A, Altman DG, Tetzlaff J, Mulrow C, Gotzsche PC, et al. (2009) The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate health care interventions: explanation and elaboration. PLoS Med 6: e1000100.
- 21. Tacconelli E, De Angelis G, Cataldo MA, Pozzi E, Cauda R (2008) Does antibiotic exposure increase the risk of methicillin-resistant Staphylococcus aureus (MRSA) isolation? A systematic review and meta-analysis. J Antimicrob Chemother 61: 26–38.
- 22. Levy SB (2002) The 2000 Garrod lecture. Factors impacting on the problem of antibiotic resistance. J Antimicrob Chemother 49: 25–30.
- 23. Naimi TS, LeDell KH, Como-Sabetti K, Borchardt SM, Boxrud DJ, et al. (2003) Comparison of community- and health care-associated methicillin-resistant Staphylococcus aureus infection. JAMA 290: 2976–2984.
- 24. Valverde A, Grill F, Coque TM, Pintado V, Baquero F, et al. (2008) High rate of intestinal colonization with extended-spectrum-beta-lactamase-producing organisms in household contacts of infected community patients. J Clin Microbiol 46: 2796–2799.
- 25. Schentag JJ, Hyatt JM, Carr JR, Paladino JA, Birmingham MC, et al. (1998) Genesis of methicillin-resistant Staphylococcus aureus (MRSA), how treatment of MRSA infections has selected for vancomycin-resistant Enterococcus faecium, and the importance of antibiotic management and infection control. Clin Infect Dis 26: 1204–1214.
- 26. Lipsitch M, Samore MH (2002) Antimicrobial use and antimicrobial resistance: a population perspective. Emerg Infect Dis 8: 347–354.
- 27. Bonten MJ, Slaughter S, Ambergen AW, Hayden MK, van Voorhis J, et al. (1998) The role of “colonization pressure” in the spread of vancomycin-resistant enterococci: an important infection control variable. Arch Intern Med 158: 1127–1132.
- 28. Harbarth S, Samore MH, Lichtenberg D, Carmeli Y (2000) Prolonged antibiotic prophylaxis after cardiovascular surgery and its effect on surgical site infections and antimicrobial resistance. Circulation 101: 2916–2921.
- 29. Tacconelli E, De Angelis G, Cataldo MA, Mantengoli E, Spanu T, et al. (2009) Antibiotic usage and risk of colonization and infection with antibiotic-resistant bacteria: a hospital population-based study. Antimicrob Agents Chemother 53: 4264–4269.
- 30. Filius PM, Gyssens IC, Kershof IM, Roovers PJ, Ott A, et al. (2005) Colonization and resistance dynamics of gram-negative bacteria in patients during and after hospitalization. Antimicrob Agents Chemother 49: 2879–2886.
- 31. Kernodle DS, Barg NL, Kaiser AB (1988) Low-level colonization of hospitalized patients with methicillin-resistant coagulase-negative staphylococci and emergence of the organisms during surgical antimicrobial prophylaxis. Antimicrob Agents Chemother 32: 202–208.
- 32. Sjolund M, Wreiber K, Andersson DI, Blaser MJ, Engstrand L (2003) Long-term persistence of resistant Enterococcus species after antibiotics to eradicate Helicobacter pylori. Ann Intern Med 139: 483–487.
- 33. Gould IM (1999) A review of the role of antibiotic policies in the control of antibiotic resistance. J Antimicrob Chemother 43: 459–465.
- 34. Tacconelli E (2009) Antimicrobial use: risk driver of multidrug resistant microorganisms in healthcare settings. Curr Opin Infect Dis 22: 352–358.
- 35. Rahal JJ, Urban C, Horn D, Freeman K, Segal-Maurer S, et al. (1998) Class restriction of cephalosporin use to control total cephalosporin resistance in nosocomial Klebsiella. JAMA 280: 1233–1237.
- 36. Acquarolo A, Urli T, Perone G, Giannotti C, Candiani A, et al. (2005) Antibiotic prophylaxis of early onset pneumonia in critically ill comatose patients. A randomized study. Intensive Care Med 31: 510–516.
- 37. Sirvent JM, Torres A, El-Ebiary M, Castro P, de Batlle J, et al. (1997) Protective effect of intravenously administered cefuroxime against nosocomial pneumonia in patients with structural coma. Am J Respir Crit Care Med 155: 1729–1734.
- 38. Valles J, Peredo R, Burgueno MJ, Rodrigues de Freitas AP, Millan S, et al. (2013) Efficacy of single-dose antibiotic against early-onset pneumonia in comatose patients who are ventilated. Chest 143: 1219–1225.
- 39. Babyak MA (2004) What you see may not be what you get: a brief, nontechnical introduction to overfitting in regression-type models. Psychosom Med 66: 411–421.
- 40. Peduzzi P, Concato J, Kemper E, Holford TR, Feinstein AR (1996) A simulation study of the number of events per variable in logistic regression analysis. J Clin Epidemiol 49: 1373–1379.
- 41. Ignelzi RJ, VanderArk GD (1975) Analysis of the treatment of basilar skull fractures with and without antibiotics. J Neurosurg 43: 721–726.
- 42. Klastersky J, Sadeghi M, Brihaye J (1976) Antimicrobial prophylaxis in patients with rhinorrhea or otorrhea: a double-blind study. Surg Neurol 6: 111–114.
- 43. Demetriades D, Charalambides D, Lakhoo M, Pantanowitz D (1992) Role of prophylactic antibiotics in open and basilar fractures of the skull: a randomized study. Injury 23: 377–380.
- 44. Eftekhar B, Ghodsi M, Nejat F, Ketabchi E, Esmaeeli B (2004) Prophylactic administration of ceftriaxone for the prevention of meningitis after traumatic pneumocephalus: results of a clinical trial. J Neurosurg 101: 757–761.
- 45. Assmann SF, Pocock SJ, Enos LE, Kasten LE (2000) Subgroup analysis and other (mis)uses of baseline data in clinical trials. Lancet 355: 1064–1069.
- 46. Ratilal BO, Costa J, Sampaio C, Pappamikail L (2011) Antibiotic prophylaxis for preventing meningitis in patients with basilar skull fractures. Cochrane Database Syst Rev: CD004884.
- 47. Hoff JT, Brewin A, U HS (1976) Letter: Antibiotics for basilar skull fracture. J Neurosurg 44: 649.
- 48. Brodie HA (1997) Prophylactic antibiotics for posttraumatic cerebrospinal fluid fistulae. A meta-analysis. Arch Otolaryngol Head Neck Surg 123: 749–752.
- 49. Villalobos T, Arango C, Kubilis P, Rathore M (1998) Antibiotic prophylaxis after basilar skull fractures: a meta-analysis. Clin Infect Dis 27: 364–369.
- 50. Dagi TF, Meyer FB, Poletti CA (1983) The incidence and prevention of meningitis after basilar skull fracture. Am J Emerg Med 1: 295–298.
- 51. Zrebeet HA, Huang PS (1986) Prophylactic antibiotics in the treatment of fractures at the base of the skull. Del Med J 58: 741–748.
- 52. Frazee RC, Mucha P Jr, Farnell MB, Ebersold MJ (1988) Meningitis after basilar skull fracture. Does antibiotic prophylaxis help? Postgrad Med 83: 267–268, 273–264.
- 53. Bergman BR (1982) Antibiotic prophylaxis in open and closed fractures: a controlled clinical trial. Acta Orthop Scand 53: 57–62.
- 54. Braun R, Enzler MA, Rittmann WW (1987) A double-blind clinical trial of prophylactic cloxacillin in open fractures. J Orthop Trauma 1: 12–17.
- 55. Dickey RL, Barnes BC, Kearns RJ, Tullos HS (1989) Efficacy of antibiotics in low-velocity gunshot fractures. J Orthop Trauma 3: 6–10.
- 56. Patzakis MJ, Wilkins J, Moore TM (1983) Use of antibiotics in open tibial fractures. Clin Orthop Relat Res: 31–35.
- 57. Gosselin RA, Roberts I, Gillespie WJ (2004) Antibiotics for preventing infection in open limb fractures. Cochrane Database Syst Rev: CD003764.
- 58. Rojczyk M (1983) Treatment results in open fractures, aspects of antibiotic therapy. Hefte Unfallheilkd 162: 33–38.
- 59. Patzakis MJ, Harvey JP Jr, Ivler D (1974) The role of antibiotics in the management of open fractures. J Bone Joint Surg Am 56: 532–541.
- 60. Hoff WS, Bonadies JA, Cachecho R, Dorlac WC (2011) East Practice Management Guidelines Work Group: update to practice management guidelines for prophylactic antibiotic use in open fractures. J Trauma 70: 751–754.
- 61. Hospenthal DR, Murray CK, Andersen RC, Bell RB, Calhoun JH, et al. (2011) Guidelines for the prevention of infections associated with combat-related injuries: 2011 update: endorsed by the Infectious Diseases Society of America and the Surgical Infection Society. J Trauma 71: S210–234.
- 62. Brand M, Goosen J, Grieve A (2009) Prophylactic antibiotics for penetrating abdominal trauma. Cochrane Database Syst Rev: CD007370.
- 63. Brand M, Grieve A (2013) Prophylactic antibiotics for penetrating abdominal trauma. Cochrane Database Syst Rev 11: CD007370.
- 64. Bratzler DW, Houck PM, Surgical Infection Prevention Guidelines Writers W, American Academy of Orthopaedic S (2004) American Association of Critical Care N, et al (2004) Antimicrobial prophylaxis for surgery: an advisory statement from the National Surgical Infection Prevention Project. Clin Infect Dis 38: 1706–1715.
- 65. Fabian TC, Croce MA, Payne LW, Minard G, Pritchard FE, et al. (1992) Duration of antibiotic therapy for penetrating abdominal trauma: a prospective trial. Surgery 112: 788–794 discussion 794–785.
- 66. Kirton OC, O'Neill PA, Kestner M, Tortella BJ (2000) Perioperative antibiotic use in high-risk penetrating hollow viscus injury: a prospective randomized, double-blind, placebo-control trial of 24 hours versus 5 days. J Trauma 49: 822–832.
- 67. Bozorgzadeh A, Pizzi WF, Barie PS, Khaneja SC, LaMaute HR, et al. (1999) The duration of antibiotic administration in penetrating abdominal trauma. Am J Surg 177: 125–131.
- 68. Cornwell EE III, Dougherty WR, Berne TV, Velmahos G, Murray JA, et al. (1999) Duration of antibiotic prophylaxis in high-risk patients with penetrating abdominal trauma: a prospective randomized trial. J Gastrointest Surg 3: 648–653.
- 69. Solomkin JS, Mazuski JE, Bradley JS, Rodvold KA, Goldstein EJ, et al. (2010) Diagnosis and management of complicated intra-abdominal infection in adults and children: guidelines by the Surgical Infection Society and the Infectious Diseases Society of America. Clin Infect Dis 50: 133–164.
- 70. Hurwitz B (1999) Legal and political considerations of clinical practice guidelines. BMJ 318: 661–664.
- 71. Guyatt GH, Oxman AD, Kunz R, Atkins D, Brozek J, et al. (2011) GRADE guidelines: 2. Framing the question and deciding on important outcomes. J Clin Epidemiol 64: 395–400.
- 72. de Kraker ME, Davey PG, Grundmann H (2011) group Bs (2011) Mortality and hospital stay associated with resistant Staphylococcus aureus and Escherichia coli bacteremia: estimating the burden of antibiotic resistance in Europe. PLoS Med 8: e1001104.
- 73. de Kraker ME, Wolkewitz M, Davey PG, Koller W, Berger J, et al. (2011) Burden of antimicrobial resistance in European hospitals: excess mortality and length of hospital stay associated with bloodstream infections due to Escherichia coli resistant to third-generation cephalosporins. J Antimicrob Chemother 66: 398–407.
- 74. de Kraker ME, Wolkewitz M, Davey PG, Koller W, Berger J, et al. (2011) Clinical impact of antimicrobial resistance in European hospitals: excess mortality and length of hospital stay related to methicillin-resistant Staphylococcus aureus bloodstream infections. Antimicrob Agents Chemother 55: 1598–1605.
- 75. Price DJ, Sleigh JD (1970) Control of infection due to Klebsiella aerogenes in a neurosurgical unit by withdrawal of all antibiotics. Lancet 2: 1213–1215.
- 76. Price D, Sleigh JD (1970) Infection due to Klebsiella aerogenes in a neurosurgical unit. J Clin Pathol 23: 378.
- 77. Price DJ, Sleigh JD (1972) Klebsiella meningitis–report of nine cases. J Neurol Neurosurg Psychiatry 35: 903–908.