Clinical Practice Guidelines for Antimicrobial Prophylaxis in Surgery

Strength of evidence and grading of recommendations

The primary literature from the previous ASHP Therapeutic Guidelines on Antimicrobial Prophylaxis in Surgery [1] was reviewed together with the primary literature published between the date of the previous guidelines, 1999, and June 2010, identified by searches of MEDLINE, EMBASE, and the Cochrane Database of Systematic Reviews. Particular attention was paid to study design, with greatest credence given to randomized, controlled, double-blind studies. There is a limited number of adequately powered randomized controlled trials evaluating the efficacy of antimicrobial prophylaxis in surgical procedures. Guidelines development included consideration of the following characteristics: validity, reliability, clinical applicability, flexibility, clarity, and a multidisciplinary nature as consistent with ASHP's philosophy on therapeutic guidelines [4]. The limitations of the evidence base are noted within each individual procedure section of the guidelines. Published guidelines with recommendations by experts in a procedure area (e.g., American College of Obstetricians and Gynecologists [ACOG]) and noted general guidelines (e.g., U.S. Centers for Disease Control and Prevention [CDC], Scottish Intercollegiate Guidelines Network, Medical Letter, SIS, SHEA/IDSA) were also considered [2,3,5–11].

Recommendations for the use of antimicrobial prophylaxis are graded according to the strength of evidence available. The strength of evidence represents only support for or against prophylaxis and does not apply to the antimicrobial agent, dose, or dosage regimen. Studies supporting the recommendations for the use of antimicrobial therapy were classified as follows:

• Level I (evidence from large, well-conducted, randomized, controlled clinical trials or a meta-analysis),

This system has been used by the Agency for Healthcare Research and Quality, and ASHP, IDSA, SIS, and SHEA support it as an acceptable method for organizing strength of evidence for a variety of therapeutic or diagnostic recommendations [4]. Each recommendation was categorized according to the strength of evidence that supports the use or nonuse of antimicrobial prophylaxis as category A (levels I–III), category B (levels IV–VI), or category C (level VII).

When higher-level data are not available, a category C recommendation represents a consensus of expert panel members based on their clinical experience, extrapolation from other procedures with similar microbial or other clinical features, and available published literature. In these cases, the expert panel also extrapolated general principles and evidence from other procedures. Some recommendations include alternative approaches in situations in which panel member opinions were divided.

A major limitation of the available literature on antimicrobial prophylaxis is the difficulty in establishing significant differences in efficacy between prophylactic antimicrobial agents and controls (including placebo, no treatment, or other antimicrobial agents) due to study design and low SSI rates for most procedures. A small sample size increases the likelihood of a Type II error; therefore, there may be no apparent difference between the antimicrobial agent and placebo when in fact the antimicrobial has a beneficial effect [12]. A valid study is placebo-controlled and randomized with a sufficient sample in each group to avoid a Type II error. Of note, prophylaxis is recommended in some cases due to the severity of complications of postoperative infection (e.g., an infected device that is not easily removable) necessitating precautionary measures despite the lack of statistical support.

Summary of key updates

These guidelines reflect substantial changes from the guidelines published in 1999 [1]. Highlights of those changes are outlined here.

Application of guidelines to clinical practice

Recommendations are provided for adult (age 19 years or older) and pediatric (age 1–18 years) patients. These guidelines do not specifically address newborn (premature and full-term) infants. While the guidelines do not address all concerns for patients with renal or hepatic dysfunction, antimicrobial prophylaxis often does not need to be modified for these patients when given as a single preoperative dose before surgical incision.

The recommendations herein may not be appropriate for use in all clinical situations. Decisions to follow these recommendations must be based on the judgment of the clinician and consideration of individual patient circumstances and available resources.

These guidelines reflect current knowledge of antimicrobial prophylaxis in surgery. Given the dynamic nature of scientific information and technology, periodic review, updating, and revisions are to be expected.

Special patient populations

Common principles and procedure-specific guidelines

The Common Principles section has been developed to provide information common to many surgical procedures. These principles are general recommendations based on currently available data at the time of publication that may change over time; therefore, these principles need to be applied with careful attention to each clinical situation. Detailed information pertinent to specific surgical procedures is included in the procedure-specific sections of these guidelines.

In addition to patient- and procedure-specific considerations, several institution-specific factors must be considered by practitioners before instituting these guidelines. The availability of antimicrobial agents at the institution may be restricted by local antimicrobial-use policy or lack of approval for use by regulatory authorities. Medications that are no longer available or not approved for use by the U.S. Food and Drug Administration (FDA) are so noted. Local resistance patterns should also be considered in selecting antimicrobial agents and are discussed in the colonization and resistance patterns section of the Common Principles section.

Quality improvement efforts

National, state, local, and institutional groups have developed and implemented collaborative efforts to improve the appropriateness of surgical antimicrobial prophylaxis. Various process and outcomes measures are employed, and results are disseminated. Institutional epidemiology and infection-control programs, state-based quality-improvement campaigns (e.g., the Michigan Surgical Quality Collaborative, the Washington State Surgical Clinical Outcomes Assessment Program [39,40]), CDC, NHSN, the National Surgical Quality Improvement Program, The Joint Commission, and the National Quality Forum have been instrumental in developing programs to prevent SSIs.

Over the past decade or more, several organizations, payers, and government agencies, including the Centers for Medicare and Medicaid Services (CMS), have established national quality-improvement initiatives to further improve the safety and outcomes of health care, including surgery [41–47]. One area of focus in these initiatives for patients undergoing surgical procedures is the prevention of SSIs. The performance measures used, data collection and reporting requirements, and financial implications vary among the initiatives. The Surgical Care Improvement Project (SCIP) began in 2002 as the Surgical Infection Prevention (SIP) project, focusing on the timing, selection, and duration of prophylactic antimicrobial agents [41,42]. The SIP project was expanded to SCIP to include additional process measures surrounding patient safety and care during surgical procedures, including glucose control, venous thromboembolism prophylaxis, hair removal, and temperature control. Similar measures have been adopted by The Joint Commission [43]. The Physicians Quality Reporting System was established in 2006 to provide financial incentives to physicians meeting performance standards for quality measures, including surgery-related measures similar to those reported for SCIP and the Joint Commission [44]. Data are required to be collected by institutions and reported to payers [42,44,46]. Data for CMS and the Physicians Quality Reporting System measures are displayed on public websites to allow consumers to compare performance among hospitals. Institutional data collection and reporting are required, with financial incentives tied to performance to varying degrees, including payment for reporting, payment increases for meeting or exceeding minimum levels of performance, payment reduction for poor performance, and lack of payment for the development of surgical complications, such as mediastinitis.

Quality improvement initiatives and mandated performance reporting are subject to change, so readers of these guidelines are advised to consult their local or institutional quality-improvement departments for new developments in requirements for measures and data reporting that apply to their practice.

Cost containment

Few pharmacoeconomic studies have addresed surgical antimicrobial prophylaxis; therefore, a cost-minimization approach was employed in developing these guidelines. The antimicrobial agent recommendations are based primarily on efficacy and safety. Individual institutions must consider their acquisition costs when implementing these guidelines.

Additional cost savings may be realized through collaborative management by pharmacists and surgeons to select the most cost-effective agent and minimize or eliminate postoperative dosing [48–50]. The use of standardized antimicrobial order sets, automatic stop-order programs, and educational initiatives has been shown to facilitate the adoption of guidelines for surgical antimicrobial prophylaxis [51–58].

Trends in microbiology

The causative pathogens associated with SSIs in U.S. hospitals have changed over the past two decades. Analysis of National Nosocomial Infections Surveillance (NNIS) System data found that the percentage of SSIs caused by gram-negative bacilli decreased from 56.5% in 1986 to 33.8% in 2003 [68]. Staphylococcus aureus was the most common pathogen, causing 22.5% of SSIs during this time period. NHSN data from 2006 to 2007 revealed that the proportion of SSIs caused by S. aureus increased to 30%, with MRSA comprising 49.2% of these isolates [61]. In a study of patients readmitted to U.S. hospitals between 2003 and 2007 with a culture-confirmed SSI, the proportion of infections caused by MRSA increased significantly from 16.1% to 20.6% (p<0.0001) [69]. The MRSA infections were associated with higher mortality rates, longer hospital stays, and higher hospital costs compared with other infections.

Spectrum of activity

Antimicrobial agents with the narrowest spectrum of activity required for efficacy in preventing infection are recommended in these guidelines. Alternative antimicrobial agents with documented efficacy are also listed herein. Individual health systems must consider local resistance patterns of organisms and overall SSI rates at their site when adopting these recommendations. Resistance patterns from organisms causing SSIs—in some cases procedure-specific resistance patterns—should take precedence over hospitalwide antibiograms.

Vancomycin

In 1999, HICPAC, an advisory committee to CDC and the Secretary of the Department of Health and Human Services, collaborated with other major organizations to develop recommendations for preventing and controlling vancomycin resistance [70]. The recommendations are echoed by these and other guidelines [6,7,41,71]. Routine use of vancomycin prophylaxis is not recommended for any procedure [8]. Vancomycin may be included in the regimen of choice when a cluster of MRSA cases (e.g., mediastinitis after cardiac procedures) or methicillin-resistant coagulase-negative staphylococci SSIs have been detected at an institution. Vancomycin prophylaxis should be considered for patients with known MRSA colonization or at high risk for MRSA colonization in the absence of surveillance data (e.g., patients with recent hospitalization, nursing-home residents, hemodialysis patients) [5,41,72]. In institutions with SSIs attributable to community-associated MRSA, antimicrobial agents with known in vitro activity against this pathogen may be considered as an alternative to vancomycin.

Each institution is encouraged to develop guidelines for the proper use of vancomycin. Although vancomycin is commonly used when the risk for MRSA is high, data suggest that vancomycin is less effective than cefazolin for preventing SSIs caused by methicillin-susceptible S. aureus (MSSA) [73,74]. For this reason, vancomycin is used in combination with cefazolin at some institutions with both MSSA and MRSA SSIs. For procedures in which pathogens other than staphylococci and streptococci are likely, an additional agent with activity against those pathogens should be considered. For example, if there are surveillance data showing that gram-negative organisms are a cause of SSIs for the procedure, practitioners may consider combining vancomycin with another agent (cefazolin if the patient does not have a β-lactam allergy; an aminoglycoside [gentamicin or tobramycin], aztreonam, or single-dose fluoroquinolone if the patient has a β-lactam allergy). The use of vancomycin for MRSA prophylaxis does not supplant the need for routine surgical prophylaxis appropriate for the type of procedure. When vancomycin is used, it can almost always be used as a single dose due to its long half-life.

Colonization and resistance

A national survey determined that S. aureus nasal colonization in the general population decreased from 32.4% in 2001–02 to 28.6% in 2003–04 (p<0.01), whereas the prevalence of colonization with MRSA increased from 0.8% to 1.5% (p<0.05) during the same time periods [75]. Colonization with MRSA was independently associated with health care exposure among men, having been born in the United States, age of >60 years, diabetes, and poverty among women. Similarly, children are colonized with S. aureus and MRSA, but colonization varies by age. Children under five years of age have the highest rates, mirroring rates seen in patients over age 60 years [76]. The rates drop in children between five and 14 years of age and gradually increase to rates seen in the adult population. Lo et al. [77]. reported that in a large cohort of children, 28.1% were colonized with S. aureus between 2004 and 2006. Between 2007 and 2009, 23.3% of children were colonized with S. aureus, but the proportion of children colonized with MRSA had increased from 8.1% in 2004 to 15.1% in 2009.

Surgical antimicrobial prophylaxis can alter individual and institutional bacterial flora, leading to changes in colonization rates and increased bacterial resistance [78–84]. Surgical prophylaxis can also predispose patients to Clostridium difficile-associated colitis [81]. Risk factors for development of C. difficile-associated colitis include longer duration of prophylaxis or therapy and use of multiple antimicrobial agents [85]. Limiting the duration of antimicrobial prophylaxis to a single preoperative dose can reduce the risk of C. difficile disease.

The question of what antimicrobial surgical prophylaxis to use for patients known to be colonized or recently infected with multidrug-resistant pathogens cannot be answered easily or in a manner that can be applied uniformly to all patient scenarios. Whether prophylaxis should be expanded to provide coverage for these pathogens depends on many factors, including the pathogen, its antimicrobial susceptibility profile, the host, the procedure to be performed, and the proximity of the likely reservoir of the pathogen to the incision and operative sites. While there is no evidence on the management of surgical antimicrobial prophylaxis in a patient with past infection or colonization with a resistant gram-negative pathogen, it is logical to provide prophylaxis with an agent active against MRSA for any patient known to be colonized with this gram-positive pathogen who will have a skin incision; specific prophylaxis for a resistant gram-negative pathogen in a patient with past infection or colonization with such a pathogen may not be necessary for a purely cutaneous procedure. Similarly, a patient colonized with vancomycin-resistant enterococci (VRE) should receive prophylaxis effective against VRE when undergoing liver transplantation but probably not when undergoing an umbilical hernia repair without mesh placement. Thus, patients must be treated on a case-by-case basis, taking into account multiple considerations.

Patients receiving therapeutic antimicrobials for a remote infection before surgery should also be given antimicrobial prophylaxis before surgery to ensure adequate serum and tissue levels of antimicrobials with activity against likely pathogens for the duration of the operation. If the agents used therapeutically are appropriate for surgical prophylaxis, administering an extra dose within 60 min before surgical incision is sufficient. Otherwise, the antimicrobial prophylaxis recommended for the planned procedure should be used. For patients with indwelling tubes or drains, consideration may be given to using prophylactic agents active against pathogens found in these devices before the procedure, even though therapeutic treatment for pathogens in drains is not indicated at other times. For patients with chronic renal failure receiving vancomycin, a preoperative dose of cefazolin should be considered instead of an extra dose of vancomycin, particularly if the probable pathogens associated with the procedure are gram-negative. In most circumstances, elective surgery should be postponed when the patient has an infection at a remote site.

Allergy to β-lactam antimicrobials

Allergy to β-lactam antimicrobials may be a consideration in the selection of surgical prophylaxis. The β-lactam antimicrobials, including cephalosporins, are the mainstay of surgical antimicrobial prophylaxis and are also the most commonly implicated drugs when allergic reactions occur. Because the predominant organisms in SSIs after clean procedures are gram-positive, the inclusion of vancomycin may be appropriate for a patient with a life-threatening allergy to β-lactam antimicrobials.

Although true Type 1 (immunoglobulin E [IgE]-mediated) cross-allergic reactions between penicillins, cephalosporins, and carbapenems are uncommon, cephalosporins and carbapenems should not be used for surgical prophylaxis in patients with documented or presumed IgE-mediated penicillin allergy. Confusion about the definition of true allergy among patients and practitioners leads to recommendations for alternative antimicrobial therapy with the potential for a lack of efficacy, increased cost, and adverse events [86,87]. Type 1 anaphylactic reactions to antimicrobials usually occur 30–60 min after administration. In patients receiving penicillins, this reaction is a life-threatening emergency that precludes subsequent use of penicillins [88]. Cephalosporins and carbapenems can safely be used in patients with an allergic reaction to penicillins that is not an IgE-mediated reaction (e.g., anaphylaxis, urticaria, bronchospasm) or exfoliative dermatitis (Stevens-Johnson syndrome, toxic epidermal necrolysis), a life-threatening hypersensitivity reaction that can be caused by β-lactam antimicrobials and other medications [88,89]. Patients should be carefully questioned about their history of antimicrobial allergies to determine whether a true allergy exists before selection of agents for prophylaxis. Patients with allergies to cephalosporins, penicillins, or both have been excluded from many clinical trials. Alternatives to β-lactam antimicrobials are provided in Table 2 based mainly on the antimicrobial activity profiles against predominant procedure-specific organisms and available clinical data.

Timing of initial dose

Successful prophylaxis requires the delivery of the antimicrobial to the operative site before contamination occurs. Thus, the antimicrobial agent should be administered at such a time to provide serum and tissue concentrations exceeding the minimum inhibitory concentration (MIC) for the probable organisms associated with the procedure, at the time of incision, and for the duration of the procedure [41,90]. In 1985, DiPiro et al [91]. demonstrated that higher serum and tissue cephalosporin concentrations at the time of surgical incision and at the end of the procedure were achieved when the drugs were given intravenously at the time of anesthesia induction compared with administration in the operating room. The average interval between antimicrobial administration and incision was 17–22 minutes [91] (Dellinger EP, personal communication, May, 2011).

A prospective evaluation of 1,708 surgical patients receiving antimicrobial prophylaxis found that preoperative administration of antimicrobials within 2 h before surgical incision decreased the risk of SSI to 0.59%, compared with 3.8% for early administration (2–24 h before surgical incision) and 3.3% for any postoperative administration (any time after incision) [92]. In a study of 2,048 patients undergoing coronary bypass graft or valve replacement surgery receiving vancomycin prophylaxis, the rate of SSI was lowest in those patients in whom an infusion was started 16–60 min before surgical incision [93]. This time interval (16–60 min before incision) was compared with four others, and the rates of SSIs were significantly lower when compared with infusions given 0–15 min before surgical incision (p<0.01) and 121–180 minutes before incision (p=0.037). The risk of infection was higher in patients receiving infusions 61–120 min before incision (odds ratio [OR], 2.3; 95% confidence interval [CI], 0.98–5.61) and for patients whose infusions were started more than 180 min before surgical incision (OR, 2.1; 95% CI, 0.82–5.62) [93].

In a large, prospective, multicenter study from the Trial to Reduce Antimicrobial Prophylaxis Errors (TRAPE) study group, the timing, duration, and intraoperative redosing of antimicrobial prophylaxis and risk of SSI were evaluated in 4,472 patients undergoing cardiac surgery, hysterectomy, or hip or knee arthroplasty [94]. The majority of patients (90%) received antimicrobial prophylaxis per the SCIP guidelines [41]. Patients were assigned to one of four groups for analysis. Group 1 (n=1,844) received a cephalosporin (or other antimicrobial with a short infusion time) administered within 30 min before incision or vancomycin or a fluoroquinolone within one hour before incision. Group 2 (n=1796) received a cephalosporin 31–60 min before incision or vancomycin 61–120 min before incision. Group 3 (n=644) was given antimicrobials earlier than recommended, and group 4 (n=188) received their initial antimicrobial doses after incision. The infection risk was lowest in group 1 (2.1%), followed by group 2 (2.4%) and group 3 (2.8%). The risk of infection was highest in group 4 (5.3%, p=0.02 compared with group 1). When cephalosporins and other antimicrobials with short infusion times were analyzed separately (n=3,656), the infection rate with antimicrobials administered within 30 min before incision was 1.6% compared with 2.4% when antimicrobials were administered 31–60 min before incision (p=0.13).

In a multicenter Dutch study of 1,922 patients undergoing total hip arthroplasty, the lowest SSI rate was seen in patients who received the antimicrobial during the 30 min before incision [95]. The highest risk for infection was found in patients who received prophylaxis after the incision.

It seems intuitive that the entire antimicrobial dose should be infused before a tourniquet is inflated or before any other procedure that restricts blood flow to the surgical site is initiated; however, a study of total knee arthroplasties compared cefuroxime given 10–30 min before tourniquet inflation with cefuroxime given 10 min before tourniquet deflation and found no significant difference in SSI rates between the two groups [96].

Overall, administration of the first dose of antimicrobial beginning within 60 min before surgical incision is recommended [41,94,97]. Administration of vancomycin and fluoroquinolones should begin within 120 minutes before surgical incision because of the prolonged infusion times required for these drugs. Because these drugs have long half-lives, this early administration should not compromise serum levels of these agents during most surgical procedures. Although the recent data summarized above suggest lower infection risk with antimicrobial administration beginning within 30 min before surgical incision, these data are not sufficiently robust to recommend narrowing the optimal window to begin infusion to 1–30 min before surgical incision. However, these data do suggest that antimicrobials can be administered too close to the time of incision. Although a few articles have suggested increased infection risk with administration too close to the time of incision [93,96,97], the data presented are not convincing. In fact, all of these articles confirm the increased rate of SSI for antimicrobials given earlier than 60 min before incision. In one article, the infection rate for patients given an antimicrobial within 15 min of incision was lower than when antimicrobials were given 15–30 min before incision [97]. In another article, small numbers of patients were reported, and an assertion of high infection rates for infusion within 15 min of incision was made, but no numeric data or p values were provided [98]. In a third article, only 15 of over 2,000 patients received antimicrobials within 15 minutes before incision [93]. Earlier studies found that giving antimicrobials within 20 min of incision and as close as 7 min before incision resulted in therapeutic levels in tissue at the time of incision [41,90,91,94,97,98].

Dosing

To ensure that adequate serum and tissue concentrations of antimicrobial agents for prophylaxis of SSIs are achieved, antimicrobial-specific pharmacokinetic and pharmacodynamic properties and patient factors must be considered when selecting a dose. One of the earliest controlled studies of antimicrobial prophylaxis in cardiac surgery found a lower rate of infection in patients with detectable concentrations of the drug in serum at the end of surgery compared with patients in whom the drug was undetectable [99]. In another study, higher levels of antimicrobial in atrial tissue at the time of starting the pump for open-heart surgery were associated with fewer infections than were lower antimicrobial concentrations [100]. In patients undergoing colectomy, infection levels were inversely related to the serum gentamicin concentration at the time of surgical closure [17]. In general, it seems advisable to administer prophylactic agents in a manner that will ensure adequate levels of drug in serum and tissue for the interval during which the surgical site is open.

Duration

The shortest effective duration of antimicrobial administration for preventing SSI is not known; however, evidence is mounting that postoperative antimicrobial administration is not necessary for most procedures [6,7,41,122–124]. The duration of antimicrobial prophylaxis should be less than 24 hours for most procedures. Cardiothoracic procedures for which a prophylaxis duration of up to 48 hours has been accepted without evidence to support the practice is an area that remains controversial. The duration of cardiothoracic prophylaxis in these guidelines is based on expert panel consensus because the available data do not delineate the optimal duration of prophylaxis. In these procedures, prophylaxis for the duration of the procedure and certainly for less than 24 hours is appropriate.

A 1992 meta-analysis of studies comparing first-generation cephalosporins and antistaphylococcal antimicrobials (e.g., penicillins) with second-generation cephalosporins in patients undergoing cardiothoracic surgery found a reduction in the rate of SSI with second-generation cephalosporins but no benefit from continuing surgical prophylaxis beyond 48 h [125]. Reports published in 1980 [126], 1993 [127], 1997 [128], and 2000 [129] involving seven studies that compared single-dose prophylaxis or prophylaxis only during the operation with durations of one to four days failed to show any reduction in SSIs with the longer durations of prophylaxis. In a more-recent observational four-year cohort study of 2,641 patients undergoing coronary artery bypass graft (CABG) surgery, the extended use of antimicrobial prophylaxis (>48 h) instead of a shorter duration of prophylaxis (<48 h) failed to reduce the risk of SSI (OR, 1.2; 95% CI, 0.8–1.6) [130]. Moreover, prolonged prophylaxis was associated with an increased risk of acquired antimicrobial resistance (cephalosporin-resistant Enterobacteriaceae and VRE) compared with short-term prophylaxis (OR, 1.6; 95% CI, 1.1–2.6). There are no data to support the continuation of antimicrobial prophylaxis until all indwelling drains and intravascular catheters are removed [19,31,32,41,131–134].

Background

Cardiac procedures include CABG procedures, valve repairs, and placement of temporary or permanent implantable cardiac devices, including ventricular assist devices (VADs). SSIs, including mediastinitis and sternal SSI, are rare but serious complications after cardiac procedures. In patients undergoing CABG, the mean frequency of SSIs depending on NHSN SSI risk index category ranges from 0.35 to 8.49 per 100 operations when donor sites are included [165]. The mean frequency of SSIs depending on NHSN SSI risk index category for patients undergoing CABG with only chest incisions ranges from 0.23 to 5.67 per 100 operations [165]. Most of these infections are superficial in depth. Patient-related and procedure-related risk factors for SSIs after cardiac procedures have been identified from several single-center cohort and case–control studies [117,128,166–176]. These include diabetes [166,169,171–175], hyperglycemia [177–182], peripheral vascular disease [171,172,174], chronic obstructive pulmonary disease [166,174,175], obesity (BMI of >30 kg/m²) [166–168,171,173–176], heart failure [171,172], advanced age [117,128,166,172], involvement of internal mammary artery [168–172], reoperation [169–171], increased number of grafts [171], long duration of surgery [117,166,167,176], and S. aureus nasal colonization [146,160].

Patients requiring extracorporeal membrane oxygenation (ECMO) as a bridge to cardiac or lung transplantation should be treated with a similar approach. If there is no history of colonization or previous infection, the general recommendations for SSI antimicrobial prophylaxis for the specific procedure should be followed. For ECMO patients with a history of colonization or previous infection, changing the preoperative antimicrobial prophylaxis to cover these pathogens must be considered, weighing whether the pathogen is relevant to SSIs in the planned procedure.

Organisms

Almost two thirds of organisms isolated in both adult and pediatric patients undergoing cardiac procedures are gram-positive, including S. aureus, coagulase-negative staphylococcus, and, rarely, Propionibacterium acnes. Gram-negative organisms are less commonly isolated in these patients and include Enterobacter species, Pseudomonas aeruginosa, Escherichia coli, Klebsiella pneumoniae, Proteus mirabilis, and Acinetobacter species [93,139,146,183–192].

Efficacy

The SSI rate in cardiac procedures is low, but there are potential consequences if infection occurs. Multiple studies have found that antimicrobial prophylaxis in cardiac procedures lowers the occurrence of postoperative SSI up to five-fold [125].

Pediatric efficacy

The rate of SSI in pediatric cardiac procedures is sometimes higher than in adult patients [20,31,221]. Significant risk factors in pediatric patients with a mediastinal SSI included the presence of other infections at the time of the procedure, young age (newborns and infants), small body size, the duration of the procedure (including CPB time), the need for an intraoperative blood transfusion, an open sternum postoperatively, the need for a reexploration procedure, the length of stay in the intensive care unit, an NNIS/NHSN risk score of 2, and the performance of emergency procedures [20,31,221].

The organisms of concern in pediatric patients are the same as those in adult patients [20,21,31,221]. However, MRSA is rarely a concern in this population as a risk factor for SSI [221]. Pediatric patients considered at high risk for MRSA infection are those with preoperative MRSA colonization or a history of MRSA infection, neonates younger than one month of age, and neonates under three months of age who have been in the hospital since birth or have a complex cardiac disorder [21]. Strategies such as intranasal mupirocin and changes in antimicrobial prophylactic agent to vancomycin led to decreased rates of MRSA carriage and the absence of MRSA infections in one time-series evaluation; however, the overall clinical impact of these efforts is still unclear [21,221].

No well-controlled studies have evaluated the efficacy of antimicrobial prophylaxis in pediatric patients undergoing cardiac procedures. Therefore, the efficacy of antimicrobial prophylaxis is extrapolated from adult studies and should be considered the standard of care for pediatric cardiac surgery patients [19].

No well-designed studies or consensus has established the appropriate doses for common antimicrobial prophylactic agents for use in pediatric cardiac patients. Antibiotic doses have been extrapolated from guidelines for the prevention of bacterial endocarditis [11]. In recent evaluations, doses of cefazolin have ranged from 25 to 50 mg/kg [19–21,31], and vancomycin doses have ranged from 10 to 20 mg/kg [19–21,31,222–226]. Gentamicin doses used in studies have included 2.5 [20] and 5 mg/kg [22]; however, the study authors [22] felt that the higher dose was excessive. The expert panel recognizes that the usual total daily dose for pediatric patients older than six months can be 6.5–7.5 mg/kg and that dosing schedules for younger patients may be complicated.

Recommendations

For patients undergoing cardiac procedures, the recommended regimen is a single preincision dose of cefazolin or cefuroxime with appropriate intraoperative redosing (Table 2). Currently, there is no evidence to support continuing prophylaxis until all drains and indwelling catheters are removed. Clindamycin or vancomycin is an acceptable alternative in patients with a documented β-lactam allergy. Vancomycin should be used for prophylaxis in patients known to be colonized with MRSA. If organizational SSI surveillance shows that gram-negative organisms cause infections for patients undergoing these operations, practitioners should combine clindamycin or vancomycin with another agent (cefazolin if the patient is not β-lactam-allergic; aztreonam, aminoglycoside, or single-dose fluoroquinolone if the patient is β-lactam-allergic). Mupirocin should be given intranasally to all patients with documented S. aureus colonization. (Strength of evidence for prophylaxis=A.)

Background

Antimicrobial prophylaxis is the standard of care for patients undergoing cardiac implantable device insertion (e.g., pacemaker implantation) [227]. Based on available data and perceived infection risk, antimicrobial prophylaxis is not routinely recommended for cardiac catheterization or transesophageal echocardiogram [228].

The NHSN has reported a mean SSI rate after pacemaker placement of 0.44 per 100 procedures [165]. This rate may underestimate the risk of late SSI and complications [229]. Risk factors for device-related infection after implantation of cardioverter–defibrillator systems or pacemakers identified in two large, prospective, multicenter cohort studies [230,231] and a large case–control study [232] included fever within 24 h before implantation, temporary pacing before implantation, and early reintervention for hematoma or lead replacement [230]; corticosteroid use for more than one month during the preceding year and more than two leads in place compared with two leads [232]; and development of pocket hematoma [231]. In all of the evaluations, antimicrobial prophylaxis was found to be protective against device-related infection [230–232]. Limited data are available on the efficacy and optimal dose and duration of antimicrobial prophylaxis in patients undergoing implantation of a new pacemaker, pacing system, or other cardiac device.

A meta-analysis of 15 prospective, randomized, controlled, mainly open-label studies evaluated the effectiveness of systemic antimicrobial prophylaxis compared with controls (no antimicrobials) on infection rates after pacemaker implantation [227]. Antibiotics included penicillins or cephalosporins with a duration ranging from a single preoperative dose to four days postoperatively. A consistent and significant protective effect of antimicrobial prophylaxis was found and encouraged the routine use of antimicrobial prophylaxis in patients undergoing permanent pacemaker implantation. A prospective, single-center cohort study found a low rate (1.7%) of SSI complications with a single 2 g dose of cefazolin in patients undergoing implantation of a new pacemaker, pulse-generator replacement, or upgrading of a preexisting pacing system [233]. A notable limitation of the study was the exclusion of patients with temporary percutanous cardiac stimulators who are at high risk of infection.

A large, randomized, double-blind, placebo-controlled study found a significantly lower rate of SSI with a single 1 g dose of cefazolin (0.64%) compared with placebo (3.28%) (p=0.016) given immediately before device implantation or generator replacement in a permanent pacemaker, implantable cardioverter defibrillator, or cardiac resynchronization device in a surgical operating room [231]. The expert panel noted that the cefazolin dose was not adjusted for patient weight. Recently, AHA produced evidence-based guidelines that recommend the use of a single dose of a preoperative antimicrobial [229].

Ventricular assist devices are increasingly used to bridge patients to transplantation or to support individuals who do not respond to medical therapy for congestive heart failure. Very limited data exist on infection rates, and there are no published studies that demonstrate the effectiveness of preoperative antimicrobial therapy. Using 2006–08 data from the Interagency Registry for Mechanically Assisted Circulatory Support, Holman and colleagues [234] reported that most infections related to mechanical cardiac support devices were bacterial (87%), with the remainder associated with fungal (9%), viral (1%), protozoal (0.3%), or unknown (2%) causes. Driveline infections are primarily caused by staphylococcal species from the skin. Fungal organisms also play an important role in VAD infections, most notably Candida species, and carry a high risk of mortality. A recent survey of antimicrobial surgical prophylaxis with VADs illustrates the variability and lack of consensus with regimens, using anywhere from one to four drugs for a duration of 24–72 h [235]. Immediate postoperative infections are caused by gram-positive organisms. Complications from long-term infections should not be confused with immediate postprocedure SSIs [236]. Based on the consensus of the expert panel, antimicrobial prophylaxis for replacement of a VAD due to ongoing or recent infection should incorporate coverage directed at the offending organism or organisms. While many centers use vancomycin plus ciprofloxacin plus fluconazole, this practice is not based on the published evidence.

Recommendation

A single dose of cefazolin or cefuroxime is recommended for device implantation or generator replacement in a permanent pacemaker, implantable cardioverter defibrillator, or cardiac resynchronization device. (Strength of evidence for prophylaxis=A.) There is limited evidence to make specific recommendations for VADs, and each practice should tailor protocols based on pathogen prevalence and local susceptibility profiles. Clindamycin or vancomycin is an acceptable alternative in patients with a documented β-lactam allergy. Vancomycin should be considered for prophylaxis in patients known to be colonized with MRSA.

Background

Noncardiac thoracic procedures include lobectomy, pneumonectomy, thoracoscopy, lung resection, and thoracotomy. In addition to SSIs, postoperative nosocomial pneumonia and empyema are of concern after thoracic procedures [237].

NHSN has reported that the rate of infection associated with thoracic surgery ranges from 0.76% to 2.04% [165]. Studies have found that the reported rate of SSIs after thoracic procedures in patients receiving antimicrobial prophylaxis ranged from 0.42% to 4% [238–241]. One study found an SSI rate of 14% when prophylaxis was not used [239]. The reported rates of pneumonia and empyema with antimicrobial prophylaxis are 3–24% and 0–7%, respectively [237,239–244].

Video-assisted thoracoscopic surgery (VATS) is commonly used for thoracic procedures. In some settings, VATS constitutes one third or more of all thoracic surgical procedures [245]. Since VATS uses small incisions, the rate of SSIs is lower compared with the rate associated with open thoracic surgical procedures [246]. A prospective cohort study (n=346) confirmed a low rate of SSIs (1.7%) after minimally invasive VATS procedures [240]. An additional prospective study of 988 lung resection patients confirmed that the SSI rate was significantly lower (5.5%) in VATS patients than in open thoracotomy patients (14.3%) [247]. Furthermore, SSI correlated with the duration of surgery, serum albumin, concurrent comorbidity, age, and forced expiratory volume in one second. Antimicrobial prophylaxis recommendations in this section refer to both open thoracotomy and VATS procedures. Based on available data and perceived infection risk, antimicrobial prophylaxis is not routinely recommended for chest tube insertion.

Results of a prospective cohort and case–control study revealed the following independent risk factors for pneumonia after thoracic procedures: extent of lung resection, intraoperative bronchial colonization, chronic obstructive pulmonary disease, BMI of >25 kg/m², induction therapy (chemotherapy, radiotherapy, or chemoradiotherapy), advanced age (≥75 years old), and stage III or IV cancer [243,244].

Organisms

The organisms reported from SSIs in patients undergoing thoracic procedures were S. aureus and S. epidermidis [237]. Organisms isolated in patients with postoperative pneumonia included gram-positive (Streptococcus and Staphylococcus species), gram-negative (Haemophilus influenzae, Enterobacter cloacae, K. pneumoniae, Acinetobacter species, P. aeruginosa, and Moraxella catarrhalis), and fungal (Candida species) pathogens [237,239–243].

Efficacy

Antimicrobial prophylaxis is the standard of care for patients undergoing noncardiac thoracic surgery, including pulmonary resection [11,201,237]. One randomized, double-blind, placebo-controlled, single-center study of patients in Spain undergoing pulmonary resection, persistent pneumothorax without thoracotomy tube before surgery, and nonpulmonary thoracic surgical procedures, excluding those involving the esophagus and exploratory thoracotomies, compared a single dose of cefazolin 1 g i.v. and placebo given 30 minutes before the procedure [239]. The study was stopped early due to the significant difference in SSI rates between groups (1.5% with cefazolin versus 14% with placebo, p<0.01). No differences in the rates of pneumonia and empyema were seen between groups, but these were not endpoints of the study.

Recommendations

In patients undergoing thoracic procedures, a single dose of cefazolin or ampicillin– sulbactam is recommended (Appendix B). Clindamycin or vancomycin is an acceptable alternative in patients with a documented β-lactam allergy. Vancomycin should be used for prophylaxis in patients known to be colonized with MRSA. If organizational SSI surveillance shows that gram-negative organisms are associated with infections during these operations or if there is risk of gram-negative contamination of the surgical site, practitioners should combine clindamycin or vancomycin with another agent (cefazolin if the patient is not β-lactam-allergic; aztreonam, aminoglycoside, or single-dose fluoroquinolone if the patient is β-lactam-allergic). (Strength of evidence for prophylaxis for VATS=C; strength of evidence for prophylaxis for other thoracic procedures=A.)

Background

The gastroduodenal procedures considered in these guidelines include resection with or without vagotomy for gastric or duodenal ulcers, resection for gastric carcinoma, revision required to repair strictures of the gastric outlet, percutaneous endoscopic gastrostomy (PEG) insertion, perforated ulcer procedures (i.e., Graham patch repair), pancreaticoduodenectomy (Whipple procedure), and bariatric surgical procedures (gastric bypass, gastric banding, gastroplasty, other restrictive procedures, biliopancreatic diversion). Studies specifically addressing antimicrobial prophylaxis for gastroesophageal reflux disease procedures (Nissen fundoplication) or highly selective vagotomy for ulcers (usually done laparoscopically) could not be identified. Antireflux procedures and highly selective vagotomy are clean procedures in contrast to essentially all other gastroduodenal procedures that are clean-contaminated. Other procedures that are generally performed using laparoscopic or endoscopic techniques (e.g., endoscopic retrograde cholangiopancreatography) are not specifically discussed in this document. Natural orifice transluminal endoscopic surgery (NOTES) is a developing operative technique using natural orifices (e.g., vagina, anus, mouth, stomach) for entry into the abdomen that leaves no visible scar [248]. No studies on antimicrobial prophylaxis using NOTES have been published. SSI rates reported in patients not receiving antimicrobial prophylaxis were 6% after vagotomy and drainage, 13% after gastric ulcer procedures, 6.8–17% after procedures for gastric cancer [249–253], 8% for pancreaticoduodenectomy [254], and 23.9–26% after PEG insertion [255,256].

The stomach is an effective barrier to bacterial colonization; this is at least partially related to its acidity. The stomach and the duodenum typically contain small numbers of organisms (<10⁴] colony-forming units [CFU]/mL), the most common of which are streptococci, lactobacilli, diphtheroids, and fungi [257,258]. Treatment with agents that increase gastric pH increases the concentration of gastric organisms [259–261]. Alterations in gastric and duodenal bacterial flora as a result of increases in gastric pH have the potential to increase the postoperative infection rate [262,263].

The risk of postoperative infection in gastroduodenal procedures depends on a number of factors, including the gastroduodenal procedure performed. Patients who are at highest risk include those with achlorhydria, including those receiving pharmacotherapy with histamine H₂-receptor antagonists or proton-pump inhibitors [264], gastroduodenal perforation, decreased gastric motility, gastric outlet obstruction, morbid obesity, gastric bleeding, or cancer [265]. Similar to other types of surgical procedures, risk factors for SSIs related to gastroduodenal procedures include long procedure duration [252,266,267], performance of emergency procedures [250,261], greater than normal blood loss [251,252], American Society of Anesthesiologists (ASA) classification of ≥3, and late administration of antimicrobials [268].

Organisms

The most common organisms cultured from SSIs after gastroduodenal procedures are coliforms (E. coli, Proteus species, Klebsiella species), staphylococci, streptococci, enterococci, and occasionally Bacteroides species [101,269–276].

Efficacy

Randomized controlled trials have shown that prophylactic antimicrobials are effective in decreasing postoperative infection rates in high-risk patients after gastroduodenal procedures. The majority of available studies were conducted in single centers outside of the United States. Relative to other types of gastrointestinal tract procedures, the number of clinical trials evaluating antimicrobial prophylaxis for gastroduodenal procedures is limited. In placebo-controlled trials, infection rates ranged from 0% to 22% for patients receiving cephalosporins or penicillins and from 1.7% to 66% for patients receiving placebo [270,271,273–275,277–284]. The difference was significant in most studies.

Data support antimicrobial prophylaxis for patients undergoing PEG insertion [264,285–287]. A Cochrane review of systemic antimicrobial prophylaxis for PEG procedures that included 11 randomized controlled trials and 1,196 patients found a statistically significant reduction in peristomal infections with antimicrobial prophylaxis (OR, 0.35; 95% CI, 0.23–0.48) [288]. Two meta-analyses found statistically significant decreases in SSIs with antimicrobial prophylaxis compared with placebo or controls, from 23.9–26% to 6.4–8%, respectively [255,256]. Most well-designed, randomized controlled studies found a significant decrease in postoperative SSIs or peristomal infections with single i.v. doses of a cephalosporin or penicillin, ranging from 11% to 17%, compared with from 18% to 66% with placebo or no antimicrobials [279–282,288]. Conflicting results have been seen in studies evaluating the use of preoperative patient MRSA screening, decontamination washes and shampoos, five-day preoperative treatment with intranasal mupirocin, and single-dose teicoplanin preoperative prophylaxis to decrease postoperative MRSA infections during PEG insertion [289,290].

While there have been no well-designed clinical trials of antimicrobial prophylaxis for patients undergoing bariatric surgical procedures, treatment guidelines support its use based on morbid obesity and additional comorbidities as risk factors for postoperative infections [264,291]. There is no consensus on the appropriate antimicrobial regimen; however, higher doses of antimicrobials may be needed for adequate serum and tissue concentrations in morbidly obese patients [13,268,291].

A notable risk factor for SSIs after esophageal and gastroduodenal procedures is decreased gastric acidity and motility resulting from malignancy or acid-suppression therapy [264,276]. Therefore, antimicrobial prophylaxis is indicated for patients undergoing gastric cancer procedures (including gastrectomy) and gastroduodenal procedures related to gastric and duodenal ulcer disease or bariatric surgery or pancreaticoduodenectomy. Evaluations of practice for pancreaticoduodenectomy show that antimicrobials are typically given due to concerns of bile contamination. Prophylaxis for gastroduodenal procedures that do not enter the gastrointestinal tract, such as antireflux procedures, should be limited to high-risk patients due to lack of data supporting general use in all patients. Furthermore, laparoscopic antireflux procedures are associated with very low SSI rates (0.3%) compared with open antireflux procedures (1.4%), just as laparoscopic gastric bypass procedures are associated with lower rates than in open procedures (0.4% versus 1.2%) [292].

Recommendations

Antimicrobial prophylaxis in gastroduodenal procedures should be considered for patients at highest risk for postoperative infections, including risk factors such as increased gastric pH (e.g., patients receiving acid-suppression therapy), gastroduodenal perforation, decreased gastric motility, gastric outlet obstruction, gastric bleeding, morbid obesity, ASA classification of ≥3, and cancer.

A single dose of cefazolin is recommended in procedures during which the lumen of the intestinal tract is entered (Table 2). (Strength of evidence for prophylaxis=A.) A single dose of cefazolin is recommended in clean procedures, such as highly selective vagotomy, and antireflux procedures only in patients at high risk of postoperative infection due to the presence of the above risk factors. (Strength of evidence for prophylaxis=C.) Alternative regimens for patients with β-lactam allergy include clindamycin or vancomycin plus gentamicin, aztreonam, or a fluoroquinolone. Higher doses of antimicrobials are uniformly recommended in morbidly obese patients undergoing bariatric procedures. Higher doses of antimicrobials should be considered in significantly overweight patients undergoing gastroduodenal and endoscopic procedures.

Background

Biliary tract procedures include cholecystectomy, exploration of the common bile duct, and choledochoenterostomy. These guidelines pertain only to patients undergoing biliary tract procedures with no evidence of acute biliary tract infection and to patients with community-acquired acute cholecystitis of mild-to-moderate severity. As noted in the Common Principles section, patients receiving therapeutic antimicrobials for an infection before surgery should be given additional antimicrobial prophylaxis before surgery.

These guidelines do not address patients requiring biliary tract procedures for more-severe infections, including community-acquired acute cholecystitis with severe physiological disturbance, advanced age, or immunocompromised state; acute cholangitis; and health-care-associated or nosocomial biliary infections. These biliary tract infections are treated as complicated intraabdominal infections [303]. All patients with a suspected biliary tract infection who undergo biliary tract surgery should receive preoperative i.v. antimicrobials.

The majority of published literature regarding SSIs in biliary tract procedures focuses on cholecystectomy. The overall reported rate of postoperative infection in open biliary tract procedures with antimicrobial prophylaxis is 1–19% [292,304–311]. Infection rates after laparoscopic cholecystectomy range from 0% to approximately 4% in patients without antimicrobial prophylaxis [308,312–320] and from 0% to 7% with prophylaxis [292,304–323]. Several studies found that laparoscopic cholecystectomy SSI rates were significantly lower than those associated with open cholecystectomy [292,306–311].

Risk factors associated with postoperative SSIs after biliary procedures include performance of emergency procedures [305], diabetes [305,306,311,315,317], longer procedure duration (over 120 minutes) [305,317,324], intraoperative gallbladder rupture [305], age of >70 years [6,311,315,317,325], open cholecystectomy [7,311], conversion of laparoscopic to open cholecystectomy [7], higher ASA classification (≥3) [306,310,317], episode of biliary colic within 30 days before the procedure [315,316], reintervention in less than a month for noninfectious complications [310], acute cholecystitis [6,7,306], bile spillage [7], jaundice [6,7,306], pregnancy [7], nonfunctioning gallbladder [6], and immunosuppression [7].

The biliary tract is usually sterile. Patients with bacteria in the bile at the time of surgery may be at higher risk of postoperative infection [305,326,327]; however, some studies have found no association between the presence of bacteria in the bile and infection [305,315,316,319,321]. Obesity (a BMI of >30 kg/m²) was found to be a risk factor in some studies [306] but not in others [315,319]. Laparoscopic cholecystectomy was associated with a significantly decreased risk for SSI [292,310,324,325].

Organisms

The organisms most commonly associated with infection after biliary tract procedures include E. coli, Klebsiella species, and enterococci; less frequently, other gram-negative organisms, streptococci, and staphylococci are isolated [305,306,312,315,316,318,319,321,326,328–338]. Anaerobes are occasionally reported, most commonly Clostridium species.

Recent studies have documented increasing antimicrobial resistance in the causative pathogens in biliary tract infections and other intra-abdominal infections, with up to 40% of E. coli isolates resistant to ampicillin–sulbactam and fluoroquinolones [339–341]. Due to this increasing resistance of E. coli to fluoroquinolones and ampicillin–sulbactam, local population susceptibility profiles should be reviewed to determine the optimal antimicrobials for SSI prevention in biliary tract procedures.

Efficacy

Numerous studies have evaluated the use of prophylactic antimicrobials during biliary tract procedures, with a focus on laparoscopic cholecystectomy. Laparoscopic cholecystectomy has replaced open cholecystectomy as the standard of practice because of the reduction in recovery time and shorter hospital stay. The majority of studies of antimicrobial prophylaxis for laparoscopic cholecystectomy were underpowered and varied in control groups used (placebo, active, or no treatment), follow-up (from 30 to 60 days, while some studies did not clearly define length of time), and how SSIs were detected and reported [308,312–316,318,319,321,322]. Some studies included patients who were converted from laparoscopic to open cholecystectomy and others did not.

A large, multicenter, quality assurance study in Germany assessed the effectiveness of antimicrobial prophylaxis in laparoscopic and open cholecystectomies [308]. This study included 4,477 patients whose antimicrobial choice and dosage regimens were at the discretion of the medical center and surgeon. Antimicrobials used included first-, second-, and third-generation cephalosporins or penicillins alone or in combination with metronidazole, gentamicin, or both metronidazole and gentamicin. The most common cephalosporin used was ceftriaxone, allowing its data to be separated from data for other antimicrobials. Antimicrobial prophylaxis was administered to 2,217 patients (ceftriaxone [n=787 laparoscopic and n=188 open] and other antimicrobials [n=229 laparoscopic and n=229 open]); none was given to 1,328 laparoscopic and 932 open cholecystectomy patients. Significantly lower overall infectious complications occurred in patients receiving antimicrobial prophylaxis (0.8% ceftriaxone and 1.2% other antimicrobials), compared with 5% of those who received no prophylaxis (p<0.05). The overall rates of infectious complications were 0.6%, 0.8%, and 3.3% in patients undergoing laparoscopic cholecystectomy receiving ceftriaxone, other antimicrobials, and no prophylaxis, respectively, and 1.6%, 3.9%, and 7.4%, respectively, for patients undergoing open cholecystectomy. Significantly lower rates of SSIs and postoperative pneumonia were noted in patients receiving antimicrobials compared with those who did not receive prophylaxis (p<0.05). Rates of SSI were significantly decreased in laparoscopic cholecystectomy patients who received ceftriaxone (0.1%) or other antimicrobials (0.2%) compared with those who received no antimicrobial prophylaxis (1.6%). SSI rates were significantly decreased in open cholecystectomy patients who received ceftriaxone (1.0%) or other antimicrobials (2.6%) compared with those who received no antimicrobial prophylaxis (4.4%). The study authors concluded that antimicrobial prophylaxis should be administered to all patients undergoing cholecystectomy, regardless of approach. The study had several limitations, including lack of randomization, lack of adequate controls, and lack of clear definition of patient selection for the antimicrobial regimens. The statistical analysis was not clearly defined. The study appears to have compared only the use and lack of use of antimicrobials (with ceftriaxone and other antimicrobials combined for analysis) and did not specifically compare the laparoscopic and open approaches.

The findings of this study contrast with those of several other published studies. A meta-analysis of 15 randomized controlled studies evaluated the need for antimicrobial prophylaxis in elective laparoscopic cholecystectomy for patients at low risk of infection [313]. Low risk was defined as not having any of the following: Acute cholecystitis, a history of acute cholecystitis, common bile duct calculi, jaundice, immune suppression, and prosthetic implants. A total of 2,961 patients were enrolled in the studies, including 1,494 who received antimicrobial prophylaxis, primarily with cephalosporins, vancomycin, fluoroquinolones, metronidazole, and amoxicillin–clavulanic acid, and 1,467 controls receiving placebo or no treatment. No significant difference was found in the rates of infectious complications (2.07% in patients receiving antimicrobial prophylaxis versus 2.45% in controls) or SSIs (1.47% in patients receiving antimicrobial prophylaxis versus 1.77% in controls). The authors of the meta-analysis concluded that antimicrobial prophylaxis was not necessary for low-risk patients undergoing elective laparoscopic cholecystectomy. An additional meta-analysis of nine randomized controlled trials (n=1,437) also concluded that prophylactic antimicrobials do not prevent infections in low-risk patients undergoing laparoscopic cholecystectomy [342].

A small, prospective, nonrandomized study compared the use of cefotaxime 1 g i.v. during surgery with an additional two i.v. doses given eight hours apart after surgery (n=80) with no antimicrobial prophylaxis (n=86) in patients undergoing elective laparoscopic cholecystectomy with accidental or incidental gallbladder rupture and spillage of bile [317]. Patients who had spillage of gallstone calculi or whose operations were converted to open operations were excluded from the study. The rate of SSIs did not significantly differ between treatment groups (2.5% with antimicrobials versus 3.4% without antimicrobial prophylaxis). Based on results of multivariable analysis, routine antimicrobial prophylaxis was not recommended for these patients unless they were diabetic, were older than 60 years, or had an ASA classification of ≥3 or the duration of the procedure exceeded 70 min.

Current data do not support antimicrobial prophylaxis for low-risk patients undergoing elective laparoscopic cholecystectomies or those with incidental or accidental gallbladder rupture. Antimicrobial prophylaxis should be considered for patients at high risk of infection, including those undergoing open cholecystectomy, as described above, or who are considered to be at high risk for conversion to an open procedure.

Recommendations

A single dose of cefazolin should be administered in patients undergoing open biliary tract procedures (Table 2). (Strength of evidence for prophylaxis=A.) Alternatives include ampicillin–sulbactam and other cephalosporins (cefotetan, cefoxitin, and ceftriaxone). Alternative regimens for patients with β-lactam allergy include clindamycin or vancomycin plus gentamicin, aztreonam, or a fluoroquinolone; or metronidazole plus gentamicin or a fluoroquinolone.

Antimicrobial prophylaxis is not necessary in low-risk patients undergoing elective laparoscopic cholecystectomies. (Strength of evidence against prophylaxis for low-risk patients=A.) Antimicrobial prophylaxis is recommended in patients undergoing laparoscopic cholecystectomy who have an increased risk of infectious complications. Risk factors include performance of emergency procedures, diabetes, anticipated procedure duration exceeding 120 min, risk of intraoperative gallbladder rupture, age of >70 years, open cholecystectomy, risk of conversion of laparoscopic to open cholecystectomy, ASA classification of ≥3, episode of biliary colic within 30 days before the procedure, reintervention in less than a month for noninfectious complications of prior biliary operation, acute cholecystitis, anticipated bile spillage, jaundice, pregnancy, nonfunctioning gallbladder, and immunosuppression. Because some of these risk factors cannot be determined before the surgical intervention, it may be reasonable to give a single dose of antimicrobial prophylaxis to all patients undergoing laparoscopic cholecystectomy. (Strength of evidence for prophylaxis for high-risk patients=A.)

Background

Cases of appendicitis can be described as complicated or uncomplicated on the basis of the pathology. Patients with uncomplicated appendicitis have an acutely inflamed appendix. Complicated appendicitis includes perforated or gangrenous appendicitis, including peritonitis or abscess formation. Because complicated appendicitis is treated as a complicated intra-abdominal infection [303], it has not been addressed separately in these guidelines. All patients with a suspected clinical diagnosis of appendicitis, even those with an uncomplicated case, should receive appropriate preoperative i.v. antimicrobials for SSI prevention, which, due to the common microbiology encountered, requires similar antimicrobial choices to those used to treat complicated appendicitis.

Approximately 80% of patients with appendicitis have uncomplicated disease [59]. SSI has been reported in 9–30% of patients with uncomplicated appendicitis who do not receive prophylactic antimicrobials, though some reports suggest lower complication rates in children with uncomplicated appendicitis [165,360–365]. Mean SSI rates for appendectomy reported in the most recent NHSN report (2006–08) were 1.15% (60 of 5211) for NHSN risk index categories 0 and 1 versus 3.47% (23 of 663) for NHSN risk index categories 2 and 3 [165]. Laparoscopic appendectomy has been reported to produce lower rates of incisional (superficial and deep) SSIs than open appendectomy in adults and children in multiple meta-analyses and several randomized clinical trials [292,310,366–371]. However, the rate of organ/space SSIs (i.e., intra-abdominal abscesses) was significantly increased with laparoscopic appendectomy.

Organisms

The most common microorganisms isolated from SSIs after appendectomy are anaerobic and aerobic gram-negative enteric organisms. Bacteroides fragilis is the most commonly cultured anaerobe, and E. coli is the most frequent aerobe, indicating that the bowel flora constitute a major source for pathogens [59,372,373]. Aerobic and anaerobic streptococci, Staphylococcus species, and Enterococcus species also have been reported. Pseudomonas aeruginosa has been reported infrequently.

Efficacy

Antibiotic prophylaxis is generally recognized as effective in the prevention of postoperative SSIs in patients undergoing appendectomy when compared with placebo [374].

Pediatric efficacy

In pediatric patients, as with adults, preoperative determination of complicated versus uncomplicated appendicitis is difficult. A comprehensive review is not provided here, but this topic has been addressed by SIS [388].

Two pediatric studies demonstrated no difference in SSI rates between placebo and several antimicrobials. The first study compared metronidazole, penicillin plus tobramycin, and piperacillin [389]. The second study compared single-dose metronidazole and single-dose metronidazole plus cefuroxime [390]. A meta-analysis including both adult and pediatric studies found that for pediatric patients, antimicrobial prophylaxis trended toward being beneficial, but the results were not statistically significant [374]. A retrospective chart review questioned the routine need for antimicrobial prophylaxis in children with simple appendicitis, due to relatively low infection rates in children not receiving prophylaxis [365]. However, these and other study authors have suggested antimicrobial prophylaxis may be considered due to the morbidity associated with infectious complications (e.g., prolonged hospitalization, readmission, reoperation) and due to the inability to preoperatively identify appendicitis.

As a single agent, metronidazole was no more effective than placebo in two double-blind studies that included children 10 years of age or older [360] and 15 years of age or older [363]. In a randomized study that included pediatric patients, ceftizoxime and cefamandole were associated with significantly lower infection rates and duration of hospitalization than placebo [391]. Both cefoxitin and a combination of gentamicin and metronidazole were associated with a lower rate of postoperative infection in a randomized study that included pediatric patients younger than 16 years [378]. Second-generation cephalosporins with anaerobic activity (cefoxitin or cefotetan) and third-generation cephalosporins with anaerobic activity (cefotaxime) were effective, with postoperative infection rates of <5% in two studies that included pediatric patients younger than 12 years [364,378,379]. A single dose of gentamicin with clindamycin was found to be safe and effective in children with simple appendicitis [392].

Recommendations

For uncomplicated appendicitis, the recommended regimen is a single dose of a cephalosporin with anaerobic activity (cefoxitin or cefotetan) or a single dose of a first-generation cephalosporin (cefazolin) plus metronidazole (Table 2). For β-lactam-allergic patients, alternative regimens include (1) clindamycin plus gentamicin, aztreonam, or a fluoroquinolone and (2) metronidazole plus gentamicin or a fluoroquinolone (ciprofloxacin or levofloxacin). (Strength of evidence for prophylaxis=A.)

Background

Small intestine procedures, or small bowel surgery as defined by NHSN, include incision or resection of the small intestine, including enterectomy with or without intestinal anastomosis or enterostomy, intestinal bypass, and strictureoplasty; it does not include small-to-large bowel anastomosis.

The risk of SSI in small bowel surgery is variable. The Surgical Site Infection Surveillance Service in England (data collected by 168 hospitals in 13 categories of surgical procedures between 1997 and 2002) reported an SSI rate of 8.9% (94 of 1,056) [393]. Mean SSI rates for small bowel procedures reported in the most recent NHSN report (2006–08) were 3.44% for NHSN risk index category 0 versus 6.75% for NHSN risk index categories 1, 2, and 3. A study of 1,472 patients undergoing bowel surgery (small bowel and colon) at 31 U.S. academic medical centers between September and December 2002 found an SSI rate of 8.7% for all wound categories. For patients with clean-contaminated wounds, the SSI rate was 7.9%; for those with contaminated or dirty-infected wounds, the SSI rates were 12.0% and 20.4%, respectively [394].

In a study of 178 penetrating stomach and small bowel injuries, 94% of which were operated on within six hours of presentation, SSIs occurred in nearly 20% of cases. When associated colon injuries were excluded, SSIs occurred in 16% of gastric injuries and 13% of small bowel injuries. Although 74% of patients received antimicrobials, the specific timing of antimicrobial administration was not provided [395]. Other studies of small bowel injury confirm similar SSI rates [396–400].

Antimicrobial prophylaxis is recommended for small bowel surgery, based on inferring effectiveness from other clean-contaminated procedures. No specific prospective randomized studies could be identified that addressed antimicrobial prophylaxis for small bowel surgery. Antimicrobial prophylaxis for small bowel surgical procedures related to a diagnosis of complicated intra-abdominal infection is not addressed separately in these guidelines, as antimicrobial therapy for established intraabdominal infection should be initiated preoperatively.

Organisms

The most common microorganisms isolated from SSIs after small bowel surgery are aerobic gram-negative enteric organisms. Among the species isolated from patients with SSI after small intestine surgery are gram-negative bacilli of gastrointestinal enteric origin (aerobic and anaerobic) and gram-positive species, such as streptococci, staphylococci, and enterococci, which is consistent with similar studies [401]. Escherichia coli is the most frequently identified aerobe, indicating that the bowel flora constitute a major source of pathogens. Aerobic and anaerobic streptococci, Staphylococcus species, and Enterococcus species also have been reported.

The microbiology of 2,280 SSIs after upper or lower abdominal surgery conducted from 1999 to 2006 was described in the Prevalence of Infections in Spanish Hospitals (EPINE) study [402]. The most frequent microorganisms isolated were E. coli (28%), Enterococcus species (15%), Streptococcus species (8%), P. aeruginosa (7%), and S. aureus (5%; resistant to methicillin, 2%). The microbiology of SSIs after upper abdominal tract surgery did not show any significant differences compared with SSIs of the lower tract, though there were relatively more staphylococci, K. pneumoniae, Enterobacter species, Acinetobacter species, and Candida albicans isolates and fewer E. coli, B. fragilis, and Clostridium species in the upper abdominal surgery group [402].

Efficacy

Antibiotic prophylaxis is generally recognized as effective in the prevention of postoperative SSIs in patients undergoing small bowel surgery when compared with placebo. However, there are no prospective placebo-controlled trials to definitively establish the efficacy of prophylactic antimicrobials in this patient population.

Pediatric efficacy

In pediatric patients, as with adults, antimicrobial prophylaxis for SSI prevention in small bowel surgery is recommended.

Recommendations

For small bowel surgery without obstruction, the recommended regimen is a first-generation cephalosporin (cefazolin) (Table 2). For small bowel surgery with intestinal obstruction, the recommended regimen is a cephalosporin with anaerobic activity (cefoxitin or cefotetan) or the combination of a first-generation cephalosporin (cefazolin) plus metronidazole. For β-lactam-allergic patients, alternative regimens include (1) clindamycin plus gentamicin, aztreonam, or a fluoroquinolone and (2) metronidazole plus gentamicin or a fluoroquinolone (ciprofloxacin or levofloxacin). (Strength of evidence for prophylaxis=C.)

Background

All patients who undergo hernioplasty (prosthetic mesh repair of hernia) or herniorrhaphy (suture repair of hernia) should receive appropriate preoperative i.v. antimicrobials for SSI prevention. The risk of SSIs is higher in hernioplasty compared with herniorrhaphy [403]. There is a significant risk of requiring prosthetic mesh removal in hernioplasty patients who develop an SSI, and determination of whether mesh placement will be required for hernia repair is not always possible in the preoperative period.

Mean SSI rates for herniorrhaphy reported in the most recent NHSN report (2006–08) were 0.74% (21 of 2,852) for NHSN risk index category 0, 2.42% (81 of 3,348) for NHSN risk index category 1, and 5.25% (67 of 1,277) for NHSN risk index categories 2 and 3 [165].

A Cochrane meta-analysis of 17 randomized trials (n=7,843; 11 hernioplasty trials, 6 herniorrhaphy trials) in elective open inguinal hernia repair reported SSI rates of 3.1% versus 4.5% in the antimicrobial prophylaxis and control groups, respectively (OR, 0.64; 95% CI, 0.50–0.82) [404]. The subgroup of patients with herniorrhaphy had SSI rates of 3.5% and 4.9% in the prophylaxis and control groups, respectively (OR, 0.71; 95% CI, 0.51–1.00). The subgroup of patients with hernioplasty had SSI rates of 2.4% and 4.2% in the prophylaxis and control groups, respectively (OR, 0.56; 95% CI, 0.38–0.81).

A meta-analysis of nine randomized trials of open hernioplasty for inguinal hernia documented SSI rates of 2.4% (39 of 1,642) in the antimicrobial group and 4.2% (70 of 1,676) in the control group. Antibiotics showed a protective effect in preventing SSI after mesh inguinal hernia repair (OR, 0.61; 95% CI, 0.40–0.92). Antibiotic prophylaxis did reduce the rate of SSI in hernia patients undergoing mesh hernioplasty [405].

Based on the results of these two systematic reviews, preoperative antimicrobial prophylaxis for SSI prevention is recommended for both herniorrhaphy and hernioplasty. Compared with open hernia repair, laparoscopic hernia repair has been reported to produce lower rates of incisional (superficial and deep) SSIs in randomized clinical trials [406–408]. In a recent multicenter randomized trial of laparoscopic versus open ventral incisional hernia repair (n=162), SSI was significantly less common in the laparoscopic group than in the open repair group (2.8% versus 21.9%; OR, 10.5; 95% CI, 2.3–48.2; p=0.003) [409]. A meta-analysis of eight randomized trials comparing laparoscopic and open incisional or ventral hernia repair with mesh revealed that laparoscopic hernia repair was associated with decreased SSI rates (relative risk, 0.22; 95% CI, 0.09–0.54) and a trend toward fewer infections requiring mesh removal [410].

Organisms

The most common microorganisms isolated from SSIs after herniorrhaphy and hernioplasty are aerobic gram-positive organisms. Aerobic streptococci, Staphylococcus species, and Enterococcus species are common, and MRSA is commonly found in prosthetic mesh infections [411].

Efficacy

Antibiotic prophylaxis is generally recognized as effective when compared with placebo in the prevention of postoperative SSIs in patients undergoing herniorrhaphy and hernioplasty.

Recommendations

For hernioplasty and herniorrhaphy, the recommended regimen is a single dose of a first-generation cephalosporin (cefazolin) (Table 2). For patients known to be colonized with MRSA, it is reasonable to add a single preoperative dose of vancomycin to the recommended agent. For β-lactam-allergic patients, alternative regimens include clindamycin and vancomycin. (Strength of evidence for prophylaxis=A.)

Background

Surgical site infections have been reported to occur in approximately 4–10% of patients undergoing colon procedures, 3–7% in small bowel procedures, and 3–27% in patients after rectal procedures, based on the risk index [165]. However, when patients are followed carefully in clinical trials, rates tend to be considerably higher (17–26%) [412]. Other septic complications, such as fecal fistula, intra-abdominal abscesses, peritonitis, and septicemia, are serious concerns but are much less common [413]. Infectious complication rates range from 30% to 60% without antimicrobial prophylaxis [59,414] and are <10% with appropriate antimicrobial prophylaxis. A pooled analysis of clinical trials of antimicrobial prophylaxis in colon procedures demonstrated that antimicrobial use significantly reduced mortality rates (11.2% for control versus 4.5% for treatment) and SSI rates [415].

The type and duration of the procedure can affect the risk of infection. Rectal resection is associated with a higher risk of infection than is intraperitoneal colon resection [416–418]. Other risk factors include extended procedure duration (e.g., >3.5 hours) [59,412,418,419], impaired host defenses [418], age of >60 years [418], hypoalbuminemia [419,420], bacterial or fecal contamination of the surgical site [418,420], inadvertent perforation or spillage [412,421], corticosteroid therapy [419], perioperative transfusion of packed red blood cells [394,418], hypothermia [422], hyperglycemia [423,424], and obesity [412,418].

Organisms

The infecting organisms in colorectal procedures are derived from the bowel lumen, where there are high concentrations of organisms. Bacteroides fragilis and other obligate anaerobes are the most frequently isolated organisms from the bowel, with concentrations 1,000–10,000 times higher than those of aerobes [425]. Escherichia coli is the most common aerobe. Bacteriodes fragilis and Escherichia coli comprise approximately 20–30% of the fecal mass. They are the most frequently isolated pathogens from infected surgical sites after colon procedures.

Efficacy

Results from randomized controlled trials and a Cochrane review of 182 studies of over 30,000 patients support the routine use of prophylactic antimicrobials in all patients undergoing colorectal procedures [426].

Pediatric efficacy

No well-controlled studies have evaluated the efficacy of antimicrobial prophylaxis in pediatric patients undergoing colorectal procedures. However, there is no reason to suspect that prophylaxis efficacy would be different. The safety, efficacy, tolerability, and cost-effectiveness of intestinal lavage have been demonstrated in two studies of 20 and 21 pediatric patients [488,489].

Recommendations

A single dose of second-generation cephalosporin with both aerobic and anaerobic activities (cefoxitin or cefotetan) or cefazolin plus metronidazole is recommended for colon procedures (Table 2). In institutions where there is increasing resistance to first- and second-generation cephalosporins among gram-negative isolates from SSIs, the expert panel recommends a single dose of ceftriaxone plus metronidazole over routine use of a carbapenem. An alternative regimen is ampicillin–sulbactam. In most patients, MBP combined with a combination of oral neomycin sulfate plus oral erythromycin base or oral neomycin sulfate plus oral metronidazole should be given in addition to i.v. prophylaxis. The oral antimicrobial should be given as three doses over approximately 10 hours the afternoon and evening before the operation and after the MBP. Alternative regimens for patients with β-lactam allergies include (1) clindamycin plus an aminoglycoside, aztreonam, or a fluoroquinolone and (2) metronidazole plus an aminoglycoside or a fluoroquinolone. Metronidazole plus aztreonam is not recommended as an alternative because this combination has no aerobic gram-positive activity [385]. (Strength of evidence for prophylaxis=A.)

Background

Elective procedures of the head and neck are predominantly clean or clean-contaminated [490]. Clean procedures include thyroidectomy and lymph node excisions. Clean-contaminated procedures include all procedures involving an incision through the oral or pharyngeal mucosa, ranging from parotidectomy, submandibular gland excision, tonsillectomy, adenoidectomy, and rhinoplasty to complicated tumor-debulking and mandibular fracture repair procedures requiring reconstruction. The frequency of SSIs reported for clean procedures without antimicrobial prophylaxis is <1% [491,492]. In contrast, infection rates in patients undergoing complicated head and neck cancer surgery are quite high, with infection occurring in 24–87% of patients without antimicrobial prophylaxis [493–497]. While many of these head and neck cancer procedures are clean-contaminated, these procedures can fall into different wound classifications. Head and neck cancer patients often have many of the risk factors for infection mentioned below [498].

Postoperative SSI rates are affected by age, nutritional status, and the presence of concomitant medical conditions such as diabetes mellitus, anemia, and peripheral vascular disease [496,499–504]. Use of tobacco [498,505], alcohol [505,506], or drugs of abuse [507] has also been associated with a higher risk of postoperative infection, particularly in patients with mandibular fracture. The hospital course, including length of hospitalization before operation, duration of antimicrobial use before operation, length of operation, presence of implants, and previous tracheotomy can also affect postoperative SSI rates [496,497,501–504,508]. In patients with cancer, preoperative radiation and chemotherapy as well as the stage of the malignancy may also affect infection risk [497,498,502–504]. Procedure-related risk factors for infection include radical or bilateral neck dissections [501,508] and reconstruction with myocutaneous flaps or microvascular-free flaps [497–499,508].

Organisms

The normal floras of the mouth and the oropharynx are responsible for most infections that follow clean-contaminated head and neck procedures [6,8,496,498,499,506,509–519]. Anaerobic and aerobic bacteria are abundant in the oropharynx. As a result, postoperative SSIs are usually polymicrobial and involve both aerobic and anaerobic bacteria. The predominant oropharyngeal organisms include various streptococci (aerobic and anaerobic species), other oral anaerobes including Bacteroides species (but not B. fragilis), Peptostreptococcus species, Prevotella species, Fusobacterium species, Veillonella species, Enterobacteriaceae, and staphylococci. Nasal flora includes Staphylococcus species and Streptococcus species.

Efficacy

Recommendations

Background

Nosocomial central nervous system (CNS) infections do not often occur but have potentially serious consequences and poor outcomes, including death [536]. One of the greatest risks for these infections in children and adults is undergoing a neurosurgical procedure. A classification system for neurosurgery, validated by Narotam et al. [537], divides procedures into five categories: clean, clean with foreign body, clean-contaminated, contaminated, and dirty. Risk factors for postoperative infections after neurological procedures include an ASA classification of ≥2 [538], postoperative monitoring of intracranial pressure [538,539] or ventricular drains [536,538] for five or more days, cerebrospinal fluid (CSF) leak [539–541], procedure duration of more than two to four hours [540,542–544], diabetes [544], placement of foreign body [536], repeat or additional neurosurgical procedures [538,541–543], concurrent (remote, incision, or shunt) or previous shunt infection [536,539,545,546], and emergency procedures [542,545].

Organisms

Data from most published clinical trials indicate that SSIs are primarily associated with gram-positive bacteria, S. aureus, and coagulase-negative staphylococci [6,8,537–545,547–554]. Several cohort studies revealed high rates (up to 75–80% of isolates) of MRSA [540–543,548–552] and coagulase-negative staphylococci among patients undergoing a variety of neurosurgical procedures [539,540,543,549]. Other skin organisms such as P. acnes may be seen after CSF shunt placement, craniotomy, and other procedures [536,555,556]. Gram-negative bacteria have also been isolated as the sole cause of postoperative neurosurgical SSIs in approximately 5–8% of cases and have been isolated in polymicrobial infections [537–539,541–545,547–550,552,553].

Efficacy

Efficacy for CSF shunting procedures

Antimicrobial prophylaxis is recommended for adults undergoing placement of a CSF shunt [7]. Prophylaxis in patients undergoing ventriculostomy or intraventricular prophylaxis at the time of ventriculoperitoneal shunt insertion has shown some benefit in reducing infection but remains controversial due to limited evidence [6,7].

Because CNS infections after shunting procedures are responsible for substantial mortality and morbidity, especially in children, the possible role of prophylactic antimicrobials in such procedures has been studied in numerous small, well-conducted, randomized controlled trials [564–571]. Meticulous surgical and aseptic techniques and short procedure times were determined to be important factors in lowering infection rates after shunt placement. Although the number of patients studied in each trial was small, two meta-analyses of these data demonstrated that antimicrobial prophylaxis use in CSF-shunting procedures reduced the risk of infection by approximately 50% [572,573].

Intrathecal pump placement involves the implantation of a permanent intrathecal catheter to allow instillation of medication. Central nervous system infections may occur after these procedures, which are performed in both pediatric and adult populations. Several retrospective series have reported infection rates of 4.5–9% after intrathecal baclofen pump placement [574–576]. There are minimal published trial data regarding appropriate prophylaxis for intrathecal pump procedures. It has been suggested that prophylaxis for intrathecal pump procedures be managed similarly to prophylaxis for CSF shunting procedures [577].

There is no consensus on the use of antimicrobial prophylaxis in patients with extraventricular drains (EVDs) or intracranial pressure monitors [134]. An international survey of neurosurgeons and critical care medicine and infectious diseases specialists illustrates the difference in practices. The majority of neurosurgeons used or recommended the use of antimicrobial prophylaxis with EVDs (73.5%) and other monitoring devices (59%), compared with rates of 46–59% for critical care medicine specialists and 35% for infectious diseases specialists. The majority of specialists did not recommend or use antimicrobial-coated EVD catheters.

Two randomized controlled studies comparing antimicrobial-impregnated shunts to standard, non-antimicrobial-impregnated shunts along with antimicrobial prophylaxis with i.v. cephalosporin found a decrease in rates of shunt infections [549] and a significant decrease in CSF infection with antimicrobial-impregnated shunts [545]. At this time, routine use of antimicrobial-impregnated devices is not recommended; additional well-designed studies are needed to establish their place in therapy [7,578].

Pediatric efficacy for CSF shunting procedures

Antimicrobial prophylaxis is recommended for children undergoing a CSF-shunting procedure [7]. The efficacy of antimicrobial prophylaxis is extrapolated from adult studies.

A retrospective pediatric study of 384 CSF shunting procedures found a lower infection rate in patients who received antimicrobials (2.1%) compared with those who did not (5.6%), but this difference failed to reach statistical significance [581]. Two randomized, prospective studies that included pediatric patients did not demonstrate a significant difference in infection rates between the control group and the groups that received cefotiam [571] (not available in the United States) or methicillin [568]. A randomized, double-blind, placebo-controlled study that included pediatric patients undergoing ventriculoperitoneal shunt surgeries failed to demonstrate that the use of perioperative sulfamethoxazole–trimethoprim reduced the frequency of shunt infection [564].

Other studies have demonstrated efficacy for prophylactic antimicrobials [566,582]. A single-center, randomized, double-blind, placebo-controlled trial of perioperative rifampin plus trimethoprim was performed in pediatric patients [582]. Among patients receiving rifampin plus trimethoprim, the infection rate was 12%, compared with 19% in patients receiving placebo. The study was ended because of the high infection rates before significance could be achieved. Infection rates at the study institution had been 7.5% in the years before the study. An open-label randomized study, including pediatric patients, demonstrated a lower infection rate in a group receiving oxacillin (3.3%) than in a control group (20%) [566].

Recommendations

A single dose of cefazolin is recommended for patients undergoing clean neurosurgical procedures, CSF-shunting procedures, or intrathecal pump placement (Table 2). Clindamycin or vancomycin should be reserved as an alternative agent for patients with a documented β-lactam allergy (vancomycin for MRSA-colonized patients). (Strength of evidence for prophylaxis=A.)

Background

Approximately 1.2 million infants are born by cesarean delivery in the United States annually [583]. The infection rate after cesarean delivery has been reported to be 4–15% [583], though recent NHSN data showed an infection rate of 2–4% [165].

Postpartum infectious complications are common after cesarean delivery. Endometritis (infection of the uterine lining) is usually identified by fever, malaise, tachycardia, abdominal pain, uterine tenderness, and sometimes abnormal or foul-smelling lochia [584]. Fever may also be the only symptom of endometritis.

Endometritis has been reported to occur in up to 24% of patients in elective cesarean delivery and up to approximately 60% of patients undergoing nonelective or emergency section [584,585]. Risk factors for endometritis include cesarean delivery, prolonged rupture of membranes, prolonged labor with multiple vaginal examinations, intrapartum fever, and low socioeconomic status [585,586]. Patients with low socioeconomic status may have received inadequate prenatal care.

The factor most frequently associated with infectious morbidity in postcesarean delivery is prolonged labor in the presence of ruptured membranes. Intact chorioamniotic membranes serve as a protective barrier against bacterial infection. Rupture of the membrane exposes the uterine surface to bacteria from the birth canal. The vaginal fluid with bacterial flora is drawn into the uterus when it relaxes between contractions during labor. Women undergoing labor for more than six to eight hours in the presence of ruptured membranes should be considered at high risk for developing endometritis [587]. Other risk factors for SSIs after cesarean delivery include systemic illness, poor hygiene, obesity, and anemia [587,588].

Organisms

The normal flora of the vagina include staphylococci, streptococci, enterococci, lactobacilli, diphtheroids, E. coli, anaerobic streptococci (Peptococcus species and Peptostreptococcus species), Bacteroides species (e.g., Bacteroides bivius, B. fragilis), and Fusobacterium species [584,587,589–592]. Endometritis infections are often polymicrobial and include aerobic streptococcus (particularly group B β-hemolytic streptococcus and enterococci), gram-negative aerobes (particularly E. coli), gram-negative anaerobic rods (particularly B. bivius), and anaerobic cocci (Peptococcus species and Peptostreptococcus species). Ureaplasma urealyticum has been commonly isolated from endometrial and surgical-site cultures. Additional commonly isolated organisms from SSIs include Staphylococcus species and enterococci.

Efficacy

While the use of antimicrobial prophylaxis in low-risk procedures (i.e., those with no active labor and no rupture of membranes) has been brought into question by the results of several randomized, placebo-controlled studies that found no reduction in infectious complications (fever, SSI, urinary tract infection, or endometritis) with the use of prophylaxis, the majority of these evaluations were underpowered and included administration of antimicrobial prophylaxis at cord clamping [593–599]. However, the efficacy of antimicrobial prophylaxis in cesarean delivery has been shown in several studies and two meta-analyses for both elective and nonelective procedures. Therefore, prophylaxis is recommended for all patients undergoing cesarean delivery [584,592].

One meta-analysis that reviewed seven placebo-controlled randomized trials in low-risk elective cesarean delivery found that prophylaxis was associated with a significant decrease in endometritis and fever [592]. A larger meta-analysis of 81 randomized trials with 11,937 women undergoing both elective and nonelective cesarean delivery found that antimicrobial prophylaxis was associated with a significant reduction in risk of fever, endometritis, SSI, urinary tract infection, and serious infection [585]. The relative risk for endometritis in elective cesarean section was 0.38 (95% CI, 0.22–0.64) in those receiving antimicrobial prophylaxis compared to those receiving no prophylaxis.

Recommendation

The recommended regimen for all women undergoing cesarean delivery is a single dose of cefazolin administered before surgical incision (Table 2). (Strength of evidence for prophylaxis=A.) For patients with β-lactam allergies, an alternative regimen is clindamycin plus gentamicin.

Background

Hysterectomy is second only to cesarean delivery as the most frequently performed major gynecologic procedure in the United States, with over 600,000 hysterectomies performed annually [609]. Uterine fibroid tumors account for 40% of all presurgical diagnoses leading to hysterectomy [609]. Other common diagnoses are dysfunctional uterine bleeding, genital prolapse, endometriosis, chronic pelvic pain, pelvic inflammatory disease, endometrial hyperplasia, and cancer.

Hysterectomy involves the removal of the uterus and, occasionally, one or two fallopian tubes, the ovaries, or a combination of ovaries and fallopian tubes [610]. Radical hysterectomy entails removal of the uterus, fallopian tubes, and ovaries and extensive stripping of the pelvic lymph nodes in patients with extension of their cancer. Hysterectomies are performed by a vaginal or abdominal approach using a laparoscopic- or robot-assisted method. During a vaginal hysterectomy, the procedure is completed through the vagina with no abdominal incision. Abdominal hysterectomy involves an abdominal incision. Laparoscopic and robotic methods involve small incisions and require additional equipment, increased operator experience, and increased length of procedures [611,612]. In the United States, between 2000 and 2004, the abdominal approach for hysterectomy was used in 67.9% of surgical procedures and the vaginal approach in 32.1%. Of hysterectomies performed via the vaginal approach, 32.4% also used laparoscopy [609]. The ACOG Committee on Gynecologic Practice recommends vaginal hysterectomy as the approach of choice for benign disease, based on evidence of better outcomes and fewer complications [613]. Laparoscopic abdominal hysterectomy is an alternative when the vaginal route is not indicated or feasible [613,614]. Of note, ACOG has stated that the supracervical approach—removal of the uterus with preservation of the cervix—should not be recommended as a superior technique for hysterectomy due to the lack of advantage in postoperative complications, urinary symptoms, or sexual function and the increased risk of future trachelectomy to remove the cervical stump [615].

Infections after hysterectomy include superficial and organ/space (vaginal cuff infection, pelvic cellulitis, and pelvic abscess) SSIs [589]. The reported SSI rates between January 2006 and December 2008 in the United States, based on NNIS risk index category, were 0.73–1.16 per 100 procedures for vaginal hysterectomy and 1.10–4.05 per 100 procedures for abdominal hysterectomy [165]. A multicenter surveillance study found a mean infection rate of 2.53% associated with all types of hysterectomy and a significantly lower mean rate of infection with laparoscopic versus abdominal hysterectomies (1.15% versus 3.44%, respectively) [325].

Risk factors for infection after vaginal or abdominal hysterectomy include longer duration of surgery, young age, diabetes, obesity, peripheral vascular disease, collagen disease, anemia, transfusion, poor nutritional status, and previous history of postsurgical infection [590,616–622]. The depth of subcutaneous tissue is also a significant risk factor for infection after abdominal hysterectomy [623]. Additional risk factors for infection after radical hysterectomy for cervical cancer include the presence of malignancy, prior radiation therapy, and the presence of indwelling drainage catheters [619,620].

Organisms

The vagina is normally colonized with a wide variety of bacteria, including gram-positive and gram-negative aerobes and anaerobes. The normal flora of the vagina includes staphylococci, streptococci, enterococci, lactobacilli, diphtheroids, E. coli, anaerobic streptococci, Bacteroides species, and Fusobacterium species [589,624]. Postoperative vaginal flora differs from preoperative flora; the amount of enterococci, gram-negative bacilli, and Bacteroides species increases postoperatively. Postoperative changes in flora may occur independently of prophylactic antimicrobial administration and are not by themselves predictive of postoperative infection [589,625,626]. Postoperative infections associated with vaginal hysterectomy are frequently polymicrobial, with enterococci, aerobic gram-negative bacilli, and Bacteroides species isolated most frequently. Postoperative SSIs after abdominal and radical hysterectomies are also polymicrobial; gram-positive cocci and enteric gram-negative bacilli predominate, and anaerobes are frequently isolated [626,627].

Efficacy

A meta-analysis of 25 randomized controlled trials demonstrated the efficacy of antimicrobial prophylaxis, including first- and second-generation cephalosporins and metronidazole, in the prevention of infections after abdominal hysterectomy [628]. The infection rates were 21.1% with placebo or no prophylaxis and 9.0% with any antimicrobial. Another meta-analysis found that the rate of postoperative infection (surgical and pelvic sites) in women undergoing vaginal hysterectomy who received placebo or no prophylactic antimicrobial ranged from 14% to 57%, which was significantly higher than the 10% rate reported with antimicrobials [629].

Malignant disease as the reason for hysterectomy is a common exclusion from studies of antimicrobial prophylaxis. Older, prospective, placebo-controlled studies found a lower rate of SSIs with antimicrobial prophylaxis after radical hysterectomy [619,630–633]. The applicability of these results is limited by small sample size and the inclusion of antimicrobials not available in the United States. Radical hysterectomy is primarily completed through an abdominal approach but can also be performed by a vaginal approach and using laparoscopic or robotic methods [634]. Therefore, antimicrobial prophylaxis would be warranted, regardless of approach. No placebo-controlled studies have been conducted to evaluate the efficacy of antimicrobial prophylaxis when used for laparoscopic hysterectomy.

Recommendation

The recommended regimen for women undergoing vaginal or abdominal hysterectomy, using an open or laparoscopic approach, is a single dose of cefazolin (Table 2). Cefoxitin, cefotetan, or ampicillin–sulbactam may also be used. Alternative agents for patients with a β-lactam allergy include (1) either clindamycin or vancomycin plus an aminoglycoside, aztreonam, or a fluoroquinolone and (2) metronidazole plus an aminoglycoside or a fluoroquinolone. (Strength of evidence for prophylaxis=A.)

Background

Ophthalmic procedures include cataract extractions, vitrectomies, keratoplasties, intraocular lens implantation, glaucoma procedures, strabotomies, retinal detachment repair, laser in situ keratomileusis, and laser-assisted subepithelial keratectomy. Most of the available data regarding antimicrobial prophylaxis involve cataract procedures. The goal of prophylaxis is primarily to reduce acute postoperative endophthalmitis, defined as severe intraocular inflammation due to infection, which can lead to loss of vision if untreated [666]. Since 2000, the reported frequency of endophthalmitis after ophthalmic procedures is low worldwide, ranging from zero to 0.63% [667–680]. The reported time from procedure to diagnosis of endophthalmitis ranges from one day to six weeks, with the majority of infections identified within one week [666,669,671,673,674,681–683].

Potential risk factors for postoperative ophthalmic infections include preoperative factors such as diabetes [666], active ocular infection or colonization [666,684], lacrimal drainage system infection or obstruction, age of >85 years [685], and immunodeficiency [684]. Procedure-related risk factors include clear corneal incisions (as opposed to scleral tunnel incisions) [680,686], any surgical complication, vitreous loss [684], posterior capsule tear [681,684,685], silicone intraocular lens implantation [677,680], and the nonuse of facemasks in the operating theater [681].

Organisms

Among organisms isolated from patients developing postoperative endophthalmitis after cataract procedure, approximately 25–60% were coagulase-negative Staphylococcus species, primarily S. epidermidis [668,670,671,673,674,678,683,684,686]. Other gram-positive organisms identified included S. aureus, Streptococcus species, Enterococcus species, P. acnes, and Corynebacterium species. Gram-negative organisms isolated included Serratia species, Klebsiella species, P. mirabilis, and P. aeruginosa. These organisms represent the normal flora isolated preoperatively in a number of studies [675,687–693].

Efficacy

Data on antimicrobial prophylaxis efficacy in ophthalmic procedures to prevent endophthalmitis are limited; however, prophylaxis is common [684]. The low rate of postoperative endophthalmitis makes it difficult to complete an adequately powered study to show efficacy of antimicrobial prophylaxis in ophthalmic procedures; therefore, surrogate markers of eradication of normal flora bacteria and reduction of bacterial count on the conjunctiva, lower and upper eyelids, eyelashes, and inner canthus (corner of the eye) preoperatively and postoperatively are used. Many of the available studies are flawed with retrospective or uncontrolled design, inadequate follow-up, variations in surgical techniques (including disinfection, antimicrobial prophylaxis strategies, and methods for performing procedures), and limited reporting of clinical outcomes.

The large, randomized, partially-masked, placebo-controlled, multinational, multicenter study conducted by the European Society of Cataract and Refractive Surgeons (ESCRS) compared the rate of postoperative endophthalmitis in over 16,600 patients undergoing routine cataract procedures at 24 centers in Europe randomized to one of four perioperative prophylaxis groups [679,680,694]. Patients received no antimicrobial prophylaxis, intracameral cefuroxime at the end of the procedure alone, perioperative levofloxacin 0.5% ophthalmic solution given within the hour before the procedure, or both intracameral cefuroxime and perioperative levofloxacin. All patients had the eye area disinfected with povidone–iodine 5% preoperatively and received topical levofloxacin postoperatively. The study was stopped after an interim analysis due to results of a multivariable analysis indicating that patients not receiving intracameral cefuroxime were approximately five times more likely to develop endophthalmitis. The study has been questioned for its high rate of endophthalmitis, selection of cefuroxime due to gaps in gram-negative coverage, unknown drug concentrations in the aqueous humor, risks of hypersensitivity, the lack of a commercially available preparation, the lack of a subconjunctival cefuroxime treatment group, selection of topical levofloxacin, and methods for statistical analysis [695–697].

Two single-center, historical-controlled studies in hospitals in Spain reported decreases in acute postoperative endophthalmitis among patients undergoing a cataract procedure with intracameral cefazolin added to the previous routine prophylaxis of preoperative eyelid cleansing with soap for three days [670] and povidone–iodine eye area preparation [670,674], topical antimicrobial, and corticosteroid preparations given at the end of the procedure and postoperatively. One study found a significant decrease and a relative risk reduction of 88.7% in postoperative endophthalmitis with intracameral cefazolin [670]. The other found a decrease from 0.63% to 0.055% in postoperative endophthalmitis with intracameral cefazolin [674]. No statistical analysis was performed in this study.

A retrospective cohort study of patients undergoing cataract procedure at one center in Canada between 1994 and 1998 found no significant difference in the rate of postoperative endophthalmitis with preoperative topical antimicrobials compared with none [668]. A significant decrease in endophthalmitis was seen with subconjunctival administration of antimicrobials at the end of the procedure compared with no antimicrobials.

Several prospective studies have shown decreases in ocular flora, measured by bacterial isolate and CFU counts, with preoperative antimicrobial irrigation [675], topical antimicrobials [687,688,691,692,698–700], and intracameral antimicrobials [682]. These studies did not report rates of endophthalmitis, limiting the application of the results.

Recommendation

Due to the lack of robust data from trials, specific recommendations cannot be made regarding choice, route, or duration of prophylaxis. As a general principle, the antimicrobial prophylaxis regimens used in ophthalmic procedures should provide coverage against common ocular pathogens, including Staphylococcus species and gram-negative organisms, particularly Pseudomonas species.

Preoperative antisepsis with povidone–iodine is recommended, based on available evidence. Appropriate topical antimicrobials include commercially available neomycin–polymyxin B–gramicidin solution or fluoroquinolones (particularly fourth-generation agents) given as one drop every 5–15 min for five doses within the hour before the start of the procedure (Table 2). The addition of subconjunctival cefazolin 100 mg or intracameral cefazolin 1–2.5 mg or cefuroxime 1 mg at the end of the procedure is optional. While some data have shown that intracameral antimicrobials may be more effective than subconjunctival antimicrobials, there are no commercially available antimicrobials approved for these routes of administration. (Strength of evidence for prophylaxis=B.)

Background

Orthopedic procedures considered in these guidelines include clean orthopedic procedures (not involving replacement or implantations), spinal procedures with or without instrumentation, repair of hip fractures, implantation of internal fixation devices (screws, nails, plates, and pins), and total-joint-replacement procedures. Grade III open fractures (extensive soft tissue damage and crushing) are often associated with extensive surgical site contamination and are routinely managed with empirical antimicrobial treatment and surgical debridement, for which guidelines have been published separately [720]. Available guidelines recommend that antimicrobial prophylaxis in grade I (clean wound with ≤1 cm laceration) and grade II (clean wound with >1 cm laceration without extensive soft tissue damage) open fractures be handled similarly to other clean orthopedic procedures [721–724].

Between 2006 and 2008, SSIs were reported nationally, based on risk category, in approximately 0.70–4.15 per 100 procedures for patients undergoing spinal fusion, 0.72–2.30 per 100 procedures in patients undergoing laminectomy, 0.67–2.40 per 100 procedures in patients undergoing hip prosthesis, and 0.58–1.60 per 100 procedures in patients undergoing knee prosthesis [165]. Postoperative SSI is one of the most costly complications of orthopedic procedures due to hospital readmissions, extended hospital length of stay, the need for additional procedures (often removal and reimplantation of implanted hardware), convalescent or nursing home care between procedures, and significant increases in direct hospital costs (e.g., prolonged antimicrobial therapy) [725,726]. Studies have found that the estimated economic impact of one deep SSI was $100,000 in hospital cost alone after hip arthroplasty and $60,000 after knee arthroplasty [727–731].

In light of the serious consequences, antimicrobial prophylaxis is well accepted in procedures involving the implantation of foreign materials [8,732]. Prophylaxis is also indicated in spinal procedures without instrumentation, where an SSI would pose catastrophic risks [726,733–738].

Organisms

Skin flora are the most frequent organisms involved in SSIs after orthopedic procedures. The most common pathogens in orthopedic procedures are S. aureus, gram-negative bacilli, coagulase-negative staphylococci (including S. epidermidis), and β-hemolytic streptococci [739–743]. Spinal procedures may be complicated by polymicrobial infection that includes gram-negative bacteria [740].

A contributing factor to SSIs in arthroplasty is the formation of bacterial biofilm, particularly with S. aureus and S. epidermidis, on inert surfaces of orthopedic devices. Bacterial biofilm confers antimicrobial resistance and makes antimicrobial penetration difficult [744–748].

There is increasing concern regarding the emergence of SSIs due to resistant microorganisms, specifically VRE and MRSA in surgical patients. Several studies have investigated MRSA colonization and SSIs and evaluated the effect of decolonization, including the use of topical mupirocin, in orthopedic procedures [150,157,741,749–753]. Mupirocin decolonization protocols as an adjunct to i.v. cephalosporin prophylaxis in orthopedic patients resulted in significant decreases in nasal MRSA carriage [150,751] and overall SSIs [157,750–752]. Preoperative decolonization with intranasal mupirocin may have utility in patients undergoing elective orthopedic procedures who are known to be colonized or infected with either MRSA or MSSA [150,151,157,741,749–755]. Readers are referred to additional discussion in the Common Principles section of these guidelines.

Background

In clean orthopedic procedures, such as knee, hand, and foot procedures, and those not involving the implantation of foreign materials, the need for antimicrobial prophylaxis is not well established [738,749,756]. Antimicrobial prophylaxis in patients undergoing diagnostic and operative arthroscopic procedures is controversial [6,757–760]. The risks of SSI and long-term sequelae are low for procedures not involving implantation.

Efficacy

The efficacy of antimicrobial prophylaxis in clean orthopedic procedures was first investigated in the middle part of the 20th century. A number of these studies and reviews have since been found to be flawed, as patients were not randomized to treatment groups and the timing and duration of antimicrobial prophylaxis were not studied [761,762]. Further, patients were administered prophylactic antimicrobials after the surgical procedure, which may have led to invalid results. The low rate of infection and absence of serious morbidity failed to justify the expense or potential for toxicity and resistance associated with routine use of antimicrobial prophylaxis in the setting of clean orthopedic procedures.

Recommendations

Antimicrobial prophylaxis is not recommended for patients undergoing clean orthopedic procedures, including knee, hand, and foot procedures, arthroscopy, and other procedures without instrumentation or implantation of foreign materials. (Strength of evidence against prophylaxis=C.) If the potential for implantation of foreign materials is unknown, the procedure should be treated as with implantation.

Background

Data support the use of antimicrobial prophylaxis for orthopedic spinal procedures with and without instrumentation, including fusions, laminectomies, and minimally invasive disk procedures, to decrease the rate of postoperative spinal infection [8,543,563,732,733,739,763–766]. Surgical site infections after orthopedic spinal procedures, including minimally invasive disk procedures, are associated with high morbidity. Invasion of the epidural space in organ/space SSIs is of particular concern after spinal procedures [8,145,767].

Surgical site infection rates vary with the complexity of the procedure. One retrospective, multicenter study of 1274 adult patients found an overall SSI rate of 0.22% with antimicrobial prophylaxis after minimally invasive spinal procedures (i.e., any spinal procedures performed through a tubular retractor-type system) [768]. Procedures included simple decompressive procedures (such as microscopic or endoscopic discectomy or foraminotomy or decompression of stenosis), minimally invasive arthrodeses with percutaneous instrumentation, and minimally invasive intradural procedures. The SSI rate in patients receiving antimicrobial prophylaxis undergoing spinal procedures with instrumentation has ranged from 2.8% to 9.7% [165,764,765,769,770]. Monosegmental instrumentation has a reported SSI rate of <2%, compared with 6.7% for instrumentation at multiple levels [771].

Several case–control studies of adults undergoing spinal procedures with and without instrumentation have found the following notable patient-related risk factors for SSI: prolonged preoperative hospitalization [771], diabetes [767,772–775], elevated serum glucose concentration (>125 mg/dL preoperatively [within 30 days] or >200 mg/dL postoperatively) [773], older age [767,776], smoking and alcohol abuse [776], previous procedure complicated by infection [774–776], and obesity [770–775,777]. Procedure-related risk factors include extended duration of procedure (defined in studies as two to five hours or greater than five hours [775], greater than three hours [771], and greater than five hours [776]), excessive blood loss (>1 L) [771,775], staged procedure [776], multilevel fusions [777], foreign-body placement (e.g., screw, rod, plate) [767], combined anterior and posterior fusion [776], and suboptimal antimicrobial timing (>60 min before or after incision) [773]. A significant decrease in SSIs was seen with procedures at the cervical spine level [772,773] or with an anterior surgical approach [775].

Efficacy

Despite the lack of comparative studies evaluating prophylaxis for spinal procedures with and without instrumentation (implantation of internal fixation devices), antimicrobial prophylaxis is recommended due to the associated morbidity and assumed costs of SSIs [771]. A meta-analysis of six studies with 843 patients undergoing spinal procedures (types of procedures were not differentiated in the analysis) demonstrated an overall effectiveness of antimicrobial prophylaxis [732]. Antimicrobials studied included single-dose or multidose regimens of <24 h duration of cephaloridine (a first-generation cephalosporin no longer available in the United States), vancomycin and gentamicin, cefazolin with and without gentamicin, piperacillin, and oxacillin. The pooled SSI rate with antimicrobial prophylaxis was 2.2%, compared with 5.9% in controls (OR, 0.37; 95% CI, 0.17–0.78; p<0.01). One randomized controlled study of 1237 adult patients undergoing spinal procedures to repair a herniated disk (hemilaminectomy, laminectomy, flavectomy, spondylosyndesis) found no significant difference in the rate of SSIs between single-dose cefuroxime 1.5 g i.v. (1.3%) and placebo (2.9%) given within 60 min before surgical incision. No significant difference was seen between treatment groups for incisional SSIs (0.98% and 1.12%, respectively) or deep SSIs (0.33% and 0.32%, respectively), but the difference in organ/space infections was significant between groups (0% and 1.44%, respectively; p<0.01) [778].

Pediatric efficacy

While no studies have evaluated the efficacy of antimicrobial prophylaxis in pediatric patients undergoing spinal procedures with or without instrumentation, the incidence and risk factors for SSIs in this population have been reported. The frequencies of SSIs in pediatric patients undergoing spinal fusion were 3.5% (<18 years old) [783], 3.8% (<19 years old) [784], 4.4% (ages 1–22 years old), and 5.2% (<17 years old) [764] for varying conditions, including Scheuermann kyphosis [784], myelodysplasia [764], idiopathic scoliosis [783,785], neuromuscular scoliosis [785], kyphosis [783], and spondylolisthesis [783]. The majority of patients in studies reporting antimicrobial prophylaxis received cefazolin, vancomycin, or clindamycin [764,783,785].

Risk factors for SSIs after spinal procedures with instrumentation in a pediatric population include myelodysplasia [764], procedure at the sacral spine, obesity [785], ASA classification of >2, a complex medical condition (including spinal bifida, cerebral palsy, Marfan syndrome, achondroplasia, osteogenesis imperfecta, other unspecified genetic disease, muscular dystrophy, spinal muscular atrophy, or other debilitating myopathies) [783], and previous spinal procedures. One study found a decreased risk of infection with hypothermia (core body temperature of <35.5 °C for the duration of the procedure) [785].

Two studies found suboptimal antimicrobial prophylaxis as a risk factor for SSIs in spinal procedures [764,783]. Optimal antimicrobial prophylaxis was defined as cefazolin 20 mg/kg (up to 2 g) given within 30 min [764] or 60 min [783] before surgical incision, vancomycin 10 mg/kg (up to 1 g) given within 60 min [783] or 150 min [764] before surgical incision, or clindamycin 10 mg/kg (up to 600 mg) given within 60 min before surgical incision [783]. Intraoperative redosing was defined as appropriate for cefazolin if administered for procedures lasting more than four hours and for vancomycin or clindamycin for procedures lasting more than six hours in one study [783] and for cefazolin administered every eight hours in the other study [764]. A third study found that use of clindamycin as the perioperative antimicrobial increased the risk of SSI [785].

Recommendations

Antimicrobial prophylaxis is recommended for orthopedic spinal procedures with and without instrumentation. The recommended regimen is cefazolin (Table 2). (Strength of evidence for prophylaxis in orthopedic spinal procedures=A.) Clindamycin and vancomycin should be reserved as alternative agents as described in the Common Principles section. If there are surveillance data showing that gram-negative organisms are a cause of SSIs for the procedure, practitioners may consider combining clindamycin or vancomycin with another agent (cefazolin if the patient is not β-lactam allergic; aztreonam, gentamicin, or single-dose fluoroquinolone if the patient is β-lactam allergic). Mupirocin should be given intranasally to all patients known to be colonized with S. aureus.

Background

Data support the use of antimicrobial prophylaxis for hip fracture repair to reduce the rate of SSIs, particularly in procedures that involve internal fixation (e.g., nails, screws, plates, wires). Surgical site infections after hip fracture repair can result in extensive morbidity, including prolonged and repeated hospitalization, sepsis, persistent pain, device replacement, and possible death [726,739,786–790].

Efficacy

The efficacy of antimicrobial prophylaxis in hip fracture repair has been illustrated in two meta-analyses [787,788]. One meta-analysis of 15 hip fracture procedure trials (the majority of procedures involved closed, proximal femoral, or trochanteric fractures with internal fixation) demonstrated that any dose and duration of prophylaxis are superior to no prophylaxis with respect to preventing SSIs (deep and superficial SSIs were analyzed together) [787]. The rate of SSIs was 10.4% in controls versus 5.39% in treatment groups. A second meta-analysis of 22 studies reiterated the efficacy of antimicrobial prophylaxis in fracture procedures [788]. The analysis included the same hip fracture studies examined in the first meta-analysis, with additional studies of long-bone fracture repair (i.e., closed ankle fracture and other closed fractures, some noted with internal fixation). This second meta-analysis reviewed 10 studies of 1,896 patients receiving a preoperative and two or more postoperative doses of a parenteral antimicrobial compared with a placebo or with no treatment. The authors found a relative risk of deep SSIs of 0.36 (95% CI, 0.21–0.65) and a relative risk of superficial SSIs of 0.48 (95% CI, 0.28–0.81) associated with antimicrobial use.

Recommendations

The recommended regimen in hip fracture repair or other orthopedic procedures involving internal fixation is cefazolin. Clindamycin and vancomycin should be reserved as alternative agents, as described in the Common Principles section. If there are surveillance data showing that gram-negative organisms are a cause of SSIs for the procedure, practitioners may consider combining clindamycin or vancomycin with another agent (cefazolin if the patient is not β-lactam-allergic; aztreonam, gentamicin, or single-dose fluoroquinolone if the patient is β-lactam-allergic). Mupirocin should be given intranasally to all patients with documented colonization with S. aureus. (Strength of evidence for prophylaxis=A.)

Background

In 2005, more than 750,000 hip or knee replacements were performed in the United States [793]. The reported frequency of SSIs complicating hip, knee, elbow, ankle, or shoulder replacement ranges from 0.6% to 12% [743,786,794–797]. Rates of SSI as high as 11% after hip replacement and 12% after elbow replacement have been reported [786,797]. However, for hip and knee replacements, the most common joint arthroplasties, infection rates are typically less than 2% [165].

The introduction of antimicrobial prophylaxis, stringent infection-control protocols, and the use of ultraclean operating rooms has led to a substantial reduction in SSI rates (to ≤1%) [734,786,796,798,799]. Postoperative prosthetic joint infection is an organ/space SSI that occurs early (within 3 mos postoperatively), is delayed (3–12 mos postoperatively), or occurs late (>12 mos after surgery) [748]. These infections frequently require removal of the prosthesis, a prolonged course of antimicrobials, and one- or two-stage reimplantation of the prosthesis and may result in permanent disability [796,800]. Studies have shown an estimated economic impact of one deep SSI of $100,000 in hospital cost alone after hip arthroplasty and $60,000 after knee arthroplasty [727–731].

Common risk factors for prosthetic joint infection [748] include advanced age; obesity; diabetes mellitus; corticosteroid use; malignancy; rheumatoid arthritis; previous arthroplasty on the same joint; arthroplasty undertaken to treat a fracture; type of joint replaced (e.g., risk is greater for the knee than the hip); perioperative surgical-site complications, including superficial SSI; hematoma; and persistent surgical-site drainage. Operative risk factors include ASA classification of ≥3, duration of procedure exceeding the 75th percentile for the procedure or exceeding three hours, surgical site classified as contaminated or dirty, and no systemic antimicrobial prophylaxis. Excluding the presence of a systemic antimicrobial, patients with these operative risk factors are at the greatest risk of developing an SSI. A contributing factor to SSIs in arthroplasty is the formation of bacterial biofilm, particularly with S. aureus and S. epidermidis, on inert surfaces of orthopedic devices to confer antimicrobial resistance and difficulty in antimicrobial penetration [744–748].

Efficacy

The majority of studies that have evaluated antimicrobial prophylaxis in joint replacements have been conducted in patients undergoing total hip or total knee arthroplasty [801]. There is a lack of efficacy data involving elbow, shoulder, and ankle arthroplasty; however, the same antimicrobial prophylaxis principles can be applied. In light of the serious potential consequences, antimicrobial prophylaxis is well accepted in procedures involving the implantation of foreign materials [8,543,732,733].

A meta-analysis supports the use of antimicrobial prophylaxis for SSI reduction in patients undergoing total joint replacement [801]. Of the 26 randomized controlled studies examined, 24 included patients undergoing total hip or total knee arthroplasty. The meta-analysis noted that the studies did not clearly state if the arthroplasties were primary or revision. The SSIs were defined as visible purulent exudates at the surgical site (deep or superficial) in the included studies. Seven studies (n=3,065 patients) pooled to compare antimicrobial prophylaxis with placebo found a relative risk reduction of SSIs of 81%.

Recommendations

The recommended regimen for patients undergoing total hip, elbow, knee, ankle, or shoulder replacement is cefazolin. Clindamycin and vancomycin should be reserved as alternative agents, as described in the Common Principles section. If there are any surveillance data showing that gram-negative organisms are a cause of SSIs for the procedure, practitioners may consider combining clindamycin or vancomycin with another agent (cefazolin if the patient is not β-lactam allergic; aztreonam, gentamicin, or a single-dose fluoroquinolone if the patient is β-lactam allergic). Mupirocin should be given intranasally to all patients with documented colonization with S. aureus. (Strength of evidence for prophylaxis=A.)

Background

The goals of antimicrobial prophylaxis in urologic procedures are the prevention of bacteremia and SSIs and the prevention of postoperative bacteriuria [59]. Postoperative urinary tract infections (UTIs) are the main concern for morbidity in patients after urologic procedures [817,818]. Bacteriuria, defined as >10³ or >10⁴ CFU/mL in symptomatic UTI and >10⁵ CFU/mL in asymptomatic bacteriuria, within 30 d postoperatively is a frequent primary outcome in urologic procedure studies [819–825]. The benefits of preventing postoperative bacteriuria are not clearly known [825].

In addition to general risk factors discussed in the Common Principles section of these guidelines, urologic-specific risk factors include anatomic anomalies of the urinary tract, [818] urinary obstruction, [826] urinary stone, [817,825,826] and indwelling or externalized catheters [817,818,822,826]. Preoperative UTI, particularly if recurrent, is recognized as a high-risk factor for postoperative infection, which is typically treated before procedures and is a common exclusion criterion from studies of efficacy of antimicrobial prophylaxis in urologic procedures [817,826–828]. Additional urologic operation-specific risk factors include length of postoperative catheterization, [829] mode of irrigation (closed versus open), and postoperative pyuria [821].

Organisms

Escherichia coli is the organism most commonly isolated in patients with postoperative bacteriuria; however, other gram-negative bacilli and enterococci may also cause infection [818,821,827,830–839]. Organisms such as S. aureus, coagulase-negative Staphylococcus species, and group A Streptococcus species are also a concern in procedures entering the skin with or without entering the urinary tract [818,827,830–832,838,840,841]. There is also some concern with biofilm-forming bacteria (S. epidermidis and P. aeruginosa) in patients with prosthesis implantation [842].

Efficacy

The efficacy of antimicrobial prophylaxis in select urologic procedures has been investigated in several clinical trials. Of note, many of these placebo-controlled studies have excluded patients with risk factors for infection, those requiring antimicrobial prophylaxis for another indication (e.g., infective endocarditis), and those with preoperative UTI or bacteriuria.

The efficacy of antimicrobial prophylaxis in clean procedures among patients at low risk of complications has been variable. One randomized, placebo-controlled study of oral antimicrobials in 2,083 patients undergoing flexible cystoscopy found a positive urine culture (bacteriuria with >10⁵ CFU/mL) in 9.1% of patients receiving placebo, 4.6% of patients receiving trimethoprim, and 2.8% of patients receiving ciprofloxacin [839]. The rates of bacteriuria compared with baseline were significantly higher with placebo and significantly lower with use of antimicrobials compared with placebo. A randomized, placebo-controlled study of 517 patients undergoing prostate brachytherapy found no significant difference in postimplantation epididymitis with or without antimicrobial prophylaxis (0.4% and 1.5%, respectively) [843]. A meta-analysis of eight randomized, placebo-controlled or no-treatment-controlled studies with 995 patients undergoing urodynamic studies found a decrease in bacteriuria with antimicrobial prophylaxis (OR, 0.39; 95% CI, 0.24–0.61) [820]. The number needed to treat was 13 to prevent one episode of asymptomatic bacteriuria using a pooled rate of 13.7% for bacteriuria. One study found that not using antimicrobial prophylaxis was a significant risk factor for bacteriuria caused by urinary dynamic studies [821].

Antimicrobial prophylaxis has been studied in urologic procedures involving entry into the gastrointestinal tract, with the majority of the literature on transurethral resection of the prostate (TURP) and prostate biopsy. Two large meta-analyses have suggested prophylactic antimicrobials may be effective in all patients undergoing TURP, including low-risk patients and those with preoperatively sterile urine [844,845]. One meta-analysis of 32 trials with 4,260 patients found that prophylactic antimicrobials decreased the combined bacteriuria (>10⁵ CFU/mL) event rate from 26% to 9.1%, for a relative risk reduction of 65% (95% CI, −55 to −72), and the combined clinical septicemia episode rate from 4.4% to 0.7% in TURP patients, including low-risk patients [846]. Another meta-analysis of 28 trials that included a total of 4,694 patients found prophylactic antimicrobials decreased the post-TURP rate of bacteriuria, fever, and bacteremia, as well as the need for additional postoperative antimicrobials [847]. An additional multicenter, open-label, randomized, active- and placebo-controlled trial in patients with sterile urine undergoing TURP found a decreased rate of bacteriuria (≥5 CFU/mL) with antimicrobial prophylaxis (21% with levofloxacin and 20% with sulfamethoxazole– trimethoprim) compared with placebo (30%) (p=0.009) [822].

Three randomized, placebo-controlled studies of patients undergoing transrectal needle biopsy of the prostate found significant differences in infectious complications (including bacteriuria, positive urine cultures, and UTI) in patients treated with single doses of oral antimicrobial prophylaxis compared with placebo [819,837,838]. These three studies support the routine use of antimicrobial prophylaxis in all patients undergoing transrectal needle biopsy of the prostate. Of note, all patients undergoing transrectal needle biopsy of the prostate received a cleansing enema before the procedure [819,837,838]. Use of MBP has been reported in urologic procedures that involve entering the gastrointestinal tract (e.g., urinary diversion) [844,846].

The use of antimicrobial prophylaxis in patients undergoing extracorporeal shock wave lithotripsy (ESWL) and ureterorenoscopy is supported by the results of a meta-analysis [847] and a small randomized controlled trial [848]. The meta-analysis included eight randomized controlled trials with 885 patients and six clinical case series involving 597 patients undergoing ESWL [845]. The overall rate of UTI in the randomized controlled trials ranged from 0% to 7.7% with antimicrobial prophylaxis and from 0% to 28% in the control groups (relative risk, 0.45; 95% CI, 0.22–0.93). A randomized, placebo-controlled study of 113 patients undergoing ureterorenoscopy found a rate of postoperative bacteriuria of 1.8% with antimicrobial prophylaxis and 12.5% without (p=0.0026) [848]. No patients had symptomatic UTI or inflammation complications of the urogenital tract postoperatively. There are no studies of antimicrobial prophylaxis in major open or laparoscopic procedures (cystectomy, radical prostatectomy, and nephrectomy); therefore, data have been extrapolated from other major intraabdominal procedures.

Pediatric efficacy

Limited data on antimicrobial prophylaxis are available for pediatric patients undergoing urologic procedures. One prospective, open-label, nonrandomized study of boys undergoing hypospadias repair with tabularized incision plate urethroplasty allocated patients to receive cefonicid (no longer available in the United States) with one i.v. dose before the procedure only or the addition of oral cephalexin three times daily starting on postoperative day 1 until two days after catheter removal (median, 8.3 days) [833]. More patients in the single-dose group had bacteriuria and complications (urethrocutaneous fistula and meatal stenosis); however, the rate of infection and infection-related complications did not significantly differ between groups.

Recommendations

No antimicrobial prophylaxis is recommended for clean urologic procedures in patients without risk factors for postoperative infections. Patients with preoperative bacteriuria or UTI should be treated before the procedure, when possible, to reduce the risk of postoperative infection. For patients undergoing lower urinary tract instrumentation with risk factors for infection, the use of a fluoroquinolone or trimethoprim–sulfamethoxazole (oral or i.v.) or cefazolin (i.v. or intramuscular) is recommended (Table 2). For patients undergoing clean urologic procedures without entry into the urinary tract, cefazolin is recommended, with vancomycin or clindamycin as an alternative for those patients allergic to β-lactam antimicrobials. For patients undergoing clean urologic procedures with entry into the urinary tract, cefazolin is recommended, with alternative antimicrobials to include a fluoroquinolone, the combination of an aminoglycoside plus metronidazole, or an aminoglycoside plus clindamycin. For clean-contaminated procedures of the urinary tract (often entering the gastrointestinal tract), antimicrobials as recommended for elective colorectal surgery are recommended. This would generally include the combination of cefazolin with or without metronidazole, cefoxitin, or, for patients with β-lactam allergy, a combination of either a fluoroquinolone or aminoglycoside given with either metronidazole or clindamycin. The medical literature does not support continuing antimicrobial prophylaxis until urinary catheters have been removed. See the colorectal procedures section of these guidelines for recommendations pertaining to procedures entering the gastrointestinal tract. (Strength of evidence for prophylaxis=A.)

Background

Infection after vascular procedures occurs with low frequency but can be associated with extensive morbidity and mortality [867,868]. Postoperative infections involving vascular graft material can result in limb loss and life-threatening conditions [868]. As a result, antimicrobial prophylaxis is widely used in procedures that involve implantation of prosthetic material and procedures for which there is greater risk of infection, such as aneurysm repair, thromboendarterectomy, and vein bypass [6,41,867,869]. Patients undergoing brachiocephalic procedures (e.g., carotid endarterectomy, brachial artery repair) without implantation of prosthetic graft material do not appear to benefit from routine antimicrobial prophylaxis [6,41,867,870].

Risk factors for postoperative SSI in patients undergoing vascular procedures include lower-extremity sites, delayed procedures after hospitalization, diabetes mellitus, and a history of vascular or aortocoronary bypass procedures [871,872]. Currently, prospective data from well-designed studies on prophylaxis for endovascular stenting do not exist. However, if prophylaxis is desired, the same antimicrobials and short duration of therapy used for open vascular procedures should be given. Risk factors that warrant consideration of prophylaxis in patients undergoing endovascular stenting include prolonged procedures (more than two hours), reintervention at the surgical site within seven days, vascular stent placement in the groin through a hematoma or sheath, procedures in immunosuppressed patients, and the presence of another intravascular prosthesis [873–877].

Organisms

The predominant organisms involved include S. aureus, S. epidermidis, and enteric gram-negative bacilli. MRSA is an emerging organism of concern. Several studies evaluated the rate of colonization, carriage, and infection with MRSA in patients undergoing various vascular procedures [878–884]. Independent risk factors for MRSA infection included MRSA colonization, open abdominal aortic aneurysm, tissue loss, and lower-limb bypass [878]. Patients who have or develop MRSA infections before vascular procedures have increased risk of inhospital death, intensive care unit admission, repeat surgeries, increased length of stay, and delayed wound healing, compared with patients without infections [880–883].

Efficacy

Prophylactic antimicrobials decrease the rate of infection after procedures involving the lower abdominal vasculature and procedures required to establish dialysis access. The follow-up time for patients with late surgical site complications was at least once after hospital discharge (not further defined) for most studies [829,865,871,885–887], at one month [869,871,888,889], at six months [872], and at three years [138].

A meta-analysis of 10 randomized controlled trials in patients undergoing peripheral arterial reconstruction with biological or prosthetic graft procedures found an overall consistent reduction in SSIs with systemic antimicrobial prophylaxis compared with placebo (relative risk, 0.25; 95% CI, 0.17–0.38; p<0.00001) [890]. An overall reduction was found among 5 studies evaluating early graft infection (relative risk, 0.31; 95% CI, 0.11–0.85; p=0.02), though no individual study found a significant reduction in SSIs.

The largest study included in the meta-analysis above was a randomized, prospective, double-blind, placebo-controlled study of patients undergoing peripheral vascular procedures (n=462). The infection rate was significantly lower with cefazolin than with placebo (0.9% and 6.8%, respectively) [885]. Four deep graft infections were observed in the placebo group; none occurred in the patients who received cefazolin. No infections were observed in patients who underwent brachiocephalic (n=103), femoral artery (n=56), or popliteal (n=14) procedures.

Patients undergoing vascular access procedures for hemodialysis may benefit from the administration of antistaphylococcal antimicrobials. A placebo-controlled study of 408 patients undergoing permanent vascular access placement demonstrated an upper-extremity prosthetic polytetrafluoroethylene graft infection rate of 6% with placebo compared with 1% with vancomycin (p=0.006). [869]

Recommendations

The recommended regimen for patients undergoing vascular procedures associated with a higher risk of infection, including implantation of prosthetic material, is cefazolin (Table 2). (Strength of evidence for prophylaxis=A.) Clindamycin and vancomycin should be reserved as alternative agents as described in the Common Principles section of these guidelines. If there are surveillance data showing that gram-negative organisms are a cause of SSIs for the procedure, practitioners may consider combining clindamycin or vancomycin with another agent (cefazolin if the patient is not β-lactam-allergic; aztreonam, gentamicin, or single-dose fluoroquinolone if the patient is β-lactam-allergic), due to the potential for gastrointestinal flora exposure.

Background

Solid-organ transplant recipients are at high risk for infections due to the complexity of the surgical procedures, donor- or recipient-derived infections, reactivation of recipient-associated latent infections, preoperative recipient colonization, exposure to community pathogens, and opportunistic infections due to immunosuppression [894–897]. Infections occur more frequently in the first year after transplantation, due to aggressive immunosuppression. Transplant recipients with infections are commonly asymptomatic or have nonspecific symptoms or sequelae of infection, which makes detection and diagnosis of infections difficult [855,857,894]. Postoperative infections caused by bacterial, viral, and fungal pathogens, including SSIs, UTIs, blood stream infections, and pneumonia, are of greater concern within the first month after transplantation [895–897]. Opportunistic infections that result from immunosuppression typically occur after the first month of transplantation. It is routine for transplant recipients to receive antimicrobial prophylaxis to prevent opportunistic infections [894–897]. A discussion of the prophylactic strategies for prevention of cytomegalovirus (CMV) infection, herpes simplex virus infection, pneumocystis, UTI in kidney transplant recipients, Aspergillus infection in lung transplant recipients, and other opportunistic infections outside of the immediate posttransplantation period is beyond the scope of these guidelines.

Few well-designed, prospective, comparative studies of antimicrobial prophylaxis have been conducted with patients undergoing solid-organ transplantation, and no formal recommendations are available from expert consensus panels or professional organizations; however, there are reviews that provide guidance [8,41,894].

The recommendations given for each of the solid-organ transplant procedures are intended to provide guidelines for safe and effective surgical prophylaxis based on the best available literature. Antimicrobial surgical prophylaxis practice will vary considerably among transplantation centers throughout the United States, based on the organ involved, preexisting recipient and donor infections, and local antimicrobial susceptibilities [894–897].

Heart transplantation

Lung and heart–lung transplantation

Background

Liver transplantation is a lifesaving procedure for many patients with end-stage hepatic disease for whom there are no other medical or surgical options [932,933]. In 2007, UNOS reported that 6,494 liver transplantations were performed in the United States, 96% of which had a cadaveric donor and 4% had a living-related donor source. [934] These liver transplantations were performed in 5,889 adults and 605 pediatric (<18 years old) patients. Reported one-year patient survival rates for adults ranged from 76.9% to 95% [932,935–938] and from 80% to 91.7% for pediatric patients [934,939–942]. Survival at three and five years ranged from 68.5% to 80.9% [934] and from 61.6% to 76.5% [932,933] in adult patients, respectively. In pediatric patients, three- and five-year survival ranged from 73.2% to 86% [897,934,941] and from 69.2% to 80.1% [934], respectively. One-year graft survival rates ranged from 74.2% to 94% in adults [934–936,938] and from 72.1% to 86.1% in pediatric patients [934,941,942]. Graft survival at three and five years ranged from 58.9% to 75.5% and from 51.6% to 70.5%, respectively, in adults and from 62.5% to 77.6% and from 68.4% to 71.4%, respectively, in pediatric patients [934,941]. No significant differences were noted in graft or patient survival between cadaveric and living-related donors in adult and pediatric liver transplant recipients [934]. Infection remains a major cause of morbidity and mortality in liver transplant recipients. Infections may occur in 31–83% of patients within three months of transplantation and are the cause of death in 4–53% of patients [934,936,940,943–950]. These rates are highly variable and do not seem to have changed despite advances in surgical technique and medical management. SSIs within 30 days after transplantation ranged from 4% to 48% with antimicrobial prophylaxis in several cohort and controlled studies [935–938,941,942,948,949,951–964]. Superficial SSIs are seen most often within the first two to three weeks postoperatively, whereas organ/space infections and deep infections are seen after three to four weeks.

Liver transplantation is often considered to be the most technically difficult of the solid-organ transplantation procedures. Surgical procedures lasting longer than 8–12 hours have been consistently identified as one of the most important risk factors for early infectious complications, including SSIs, intraabdominal infections, and biliary tract infections [896,938,939,945,947,957]. Other important risk factors for infectious complications related to liver transplantation surgery include previous hepatobiliary surgery [896,939,945,947,952,963], previous liver or kidney transplantation [937,951,952,965], and surgical complications such as anastomotic leakage [896,938,939,945,947,951,952]. Patient-related risk factors for infection after liver transplantation include antimicrobial use within three to four months before transplantation [935,954], low pretransplantation serum albumin concentration [938,958,963], high pretransplantation serum bilirubin concentration [939,945,947], ascites [938], obesity [963], diabetes, and hemochromatosis [966]. Procedure-related risk factors for infection include transfusion of >4 units of red blood cells [896,951], bacterial contamination due to entry into the gastrointestinal tract [963], surgical incision method [963], and use of muromonab-CD3 within the first week after transplantation [938].

Organisms

The pathogens most commonly associated with early SSIs and intraabdominal infections are those derived from the normal flora of the intestinal lumen and the skin. Aerobic gram-negative bacilli, including E. coli [935,937,939,940,942,945,947–949,951,967,968], Klebsiella species [933,936,937,939,940,945,947–949,967–969], Enterobacter species [936,939,940,942,945,947,952,959,964,967,968], A. baumannii [935–937,942,951], and Citrobacter species [939,940,945,947,952,959,967,968], are common causes of SSIs and intraabdominal infections and account for up to 65% of all bacterial pathogens. Infections due to P. aeruginosa may also occur but are much less common in the early postoperative period [936,937,939,940,942,945,947,948,952,959,969]. Enterococci are particularly common pathogens and may be responsible for 20–46% of SSIs and intraabdominal infections [894,933,935,937,938,940,943,945–947,951,952,955,964,965,969]. Staphylococcus aureus (frequently MRSA) and coagulase-negative staphylococci are also common causes of postoperative SSIs [936–938,940,942,943,945–949,955,957–961,964,965,970,971]. Candida species commonly cause both early and late postoperative infections [933,936,937,940,942,943,945–947,949,951,969].

Several studies have noted increasing concern about antimicrobial resistance based on detection of resistant organisms, including E. coli [935,937], Enterococcus species [933,937,964,965], Enterobacter species [964], Klebsiella species [933,937], coagulase-negative staphylococci [937,964], and S. aureus [937,948,957–961,970]. General information on antimicrobial resistance is provided in the Common Principles section of these guidelines. Of specific concern to the transplantation community is the emergence of multidrug-resistant A. baumannii [972], carbapenem-resistant Enterobacteriaceae [973,974], K. pneumoniae carbapenemase-producing organisms [975], and C. difficile [976–978].

Efficacy

Although there remains a high rate of infection directly related to the liver transplantation procedure, there are few well-controlled studies concerning optimal antimicrobial prophylaxis. In evaluating the efficacy of prophylactic regimens, it is important to differentiate between early infections (occurring within 14–30 days after surgery) and late infections (occurring more than 30 days after surgery). Infections occurring in the early postoperative period are most commonly associated with biliary, vascular, and abdominal surgeries involved in the transplantation procedure itself and are thus most preventable with prophylactic antimicrobial regimens [939,940,943,945]. The frequency of these infections varies from 10% to 55% despite antimicrobial prophylaxis [939,940,943,945,979]. It is difficult to assess the efficacy of prophylactic regimens in reducing the rate of infection, because prophylaxis has been routinely used in light of the complexity of the surgical procedure; therefore, reliable rates of infection in the absence of prophylaxis are not available. No controlled studies have compared prophylaxis with no prophylaxis.

Pediatric efficacy

There are few data specifically concerning antimicrobial prophylaxis in liver transplantation in pediatric patients. The combination of cefotaxime plus ampicillin has been reportedly used in children undergoing living-related donor liver transplantation; the efficacy of this regimen appeared to be favorable [946]. A small, retrospective, single-center cohort study reported outcomes of children undergoing liver, heart, small bowel, or lung transplantation receiving piperacillin–tazobactam 120–150 mg/kg/day beginning before surgical incision and continuing for 48 h postoperatively and found favorable results, with a superficial SSI rate of 8% and no deep SSIs [992].

Recommendations

The recommended agents for patients undergoing liver transplantation are (1) piperacillin–tazobactam and (2) cefotaxime plus ampicillin (Table 2). (Strength of evidence for prophylaxis=B.) For patients who are allergic to β-lactam antimicrobials, clindamycin or vancomycin given in combination with gentamicin, aztreonam, or a fluoroquinolone is a reasonable alternative. The duration of prophylaxis should be restricted to 24 h or less. For patients at high risk of Candida infection, fluconazole adjusted for renal function may be considered. (Strength of evidence for prophylaxis=B.)

Background

Pancreas transplantation is an accepted therapeutic intervention for type 1 diabetes mellitus; it is the only therapy that consistently achieves euglycemia without dependence on exogenous insulin [993–997]. Simultaneous pancreas–kidney (SPK) transplantation is an accepted procedure for patients with type 1 diabetes and severe diabetic nephropathy. In 2007, UNOS reported that 469 pancreas transplantations and 862 SPK transplantations were performed in the United States, of which 60 and four patients, respectively, were under age 18 years [998]. Pancreas graft one-year survival rates ranged from 70.2% to 89%, and the three-year rates ranged from 48% to 85.8% [998–1002]. Patient survival with pancreas transplantation has been reported between 75% and 97% at one year and between 54% and 92.5% at three years [998]. Allograft survival is higher in recipients of SPK transplantations, with allograft survival rates of 86.1–95.1% at one year and 54.2–92.5% at three years. Reported patient survival rates in SPK are 91.7–97.6% at one year and 84.4–94.1% at three years. During pancreas transplantation, surgical complications with portal-hepatic drainage significantly decreased the one-year and three-year survival rates to 48% and 44%, respectively, in one cohort study [999].

Infectious complications are a major source of morbidity and mortality in patients undergoing pancreas or SPK transplantation; the frequency of SSI is 7%–50% with antimicrobial prophylaxis [993–997,1000–1009]. The majority of SSIs occurred within the first 30 d to three months after transplantation [1000–1002,1005,1008,1009]. Urinary tract infections are also a significant concern during the same time frame, with rates ranging from 10.6% to 49% in pancreas transplant recipients who received antimicrobial prophylaxis, and are much more common in recipients with bladder drainage compared with enteric drainage [1000–1008].

Pancreas and SPK transplantation patients may be at increased risk of SSIs and other infections because of the combined immunosuppressive effects of diabetes mellitus and the immunosuppressive drugs used to prevent graft rejection [995,1000]. Other factors associated with increased SSI rates include prolonged operating and ischemic times (>4 h), organ donor age of >55 years, and enteric rather than bladder drainage of pancreatic duct secretions [895,995,1000]. Prolonged organ preservation time (>20 h) was shown to increase the risk of complications, including duodenal leaks and decreased graft survival in cadaveric pancreas transplant recipients [1003]. Risk factors for UTI are reviewed in the kidney transplant section.

Organisms

A majority of superficial SSIs after pancreas or SPK transplantation are caused by Staphylococcus species (both coagulase-positive and coagulase-negative) and gram-negative bacilli (particularly E. coli and Klebsiella species) [993–997,1000–1002,1004–1006,1009–1011]. Deep incisional SSIs also are frequently associated with gram-positive (Enterococcus species, Streptococcus species, and Peptostreptococcus species) and gram-negative organisms (Enterobacter species, Morganella species, and B. fragilis), as well as Candida species [993–997,1000–1002,1004–1006,1009–1011]. Although anaerobes are occasionally isolated, the necessity for specific treatment of anaerobes in SSIs after pancreas transplantation remains unclear.

Efficacy

Although no placebo-controlled studies have been conducted, several open-label, noncomparative, single-center studies have suggested that antimicrobial prophylaxis substantially decreases the rate of superficial and deep SSIs after pancreas or SPK transplantation. SSI rates were 7–33% with various prophylactic regimens [995,1000–1002,1004,1005], compared with 7–50% for historical controls in the absence of prophylaxis [1009,1010]. The reason for the wide disparity in infection rates observed with prophylaxis is not readily apparent but may include variations in SSI definitions, variations in antimicrobial prophylaxis, immunosuppression protocols, and variations in surgical techniques [999–1002,1005,1007,1008].

Recommendations

The recommended regimen for patients undergoing pancreas or SPK transplantation is cefazolin (Table 2). (Strength of evidence for prophylaxis=A.) For patients who are allergic to β-lactam antimicrobials, clindamycin or vancomycin given in combination with gentamicin, aztreonam, or a fluoroquinolone is a reasonable alternative. The duration of prophylaxis should be restricted to 24 hours or less. The use of aminoglycosides in combination with other nephrotoxic drugs may result in renal dysfunction and should be avoided unless alternatives are contraindicated. (Strength of evidence for prophylaxis=C.) For patients at high risk of Candida infection, fluconazole adjusted for renal function may be considered.

Background

In 2007, UNOS reported that 16,628 kidney transplantations were performed in the United States; of these, 796 patients were younger than 18 years [998]. The rate of postoperative infection after this procedure has been reported to range from 10% to 56%, with the two most common infections being UTIs and SSIs [1004,1014–1024]. Graft loss due to infection occurs in up to 33% of cases [1017,1023]. One study of adult and pediatric kidney transplant recipients (both living-related and cadaveric donor sources) found patient survival rates at seven years after transplantation of 88.9% and 75.5%, respectively, and graft survival of 75% and 55.5%, respectively [1025]. No patients developed an SSI. Mortality associated with postoperative infections is substantial and ranges from approximately 5% to 30% [1015,1017,1019,1022,1026,1027].

The frequency of SSIs in kidney transplant recipients has ranged from zero to 11% with antimicrobial prophylaxis [1023–1025,1028,1029] to 2% to 7.5% without systemic prophylaxis [1030,1031]. The majority of these infections were superficial in nature and were detected within 30 days after transplantation [1023,1028–1030]. Risk factors for SSI after kidney transplantation include contamination of organ perfusate [1027]; pretransplantation patient-specific factors, such as diabetes [1029,1030], chronic glomerulonephritis [1030], and obesity [1027,1030,1032]; procedure-related factors, such as ureteral leakage and hematoma formation [1027]; immunosuppressive therapy [1024,1027,1029]; and postoperative complications, such as acute graft rejection, reoperation, and delayed graft function [1030]. In one study, the frequency of SSI was 12% in patients receiving immunosuppression with azathioprine plus prednisone but only 1.7% in patients receiving cyclosporine plus prednisone [1033]. A significant difference in SSI rates was noted after kidney transplantation between immunosuppression regimens including mycophenolate mofetil (45 [3.9%] of 1150 patients) versus sirolimus (11 [7.4%] of 144 patients) [1029]. Sirolimus-containing immunosuppression was found to be an independent risk factor for SSIs. These recommendations refer to kidney transplant recipients; recommendations for living kidney donors can be found in the discussion of nephrectomy in the urologic section.

Organisms

Postoperative SSIs in kidney transplant recipients are caused by gram-positive organisms, particularly Staphylococcus species (including S. aureus and S. epidermidis) and Enterococcus species, gram-negative organisms, E. coli, Enterobacter species, Klebsiella species, P. aeruginosa, and yeast with Candida species [1004,1014–1021,1023,1024,1026,1028,1030,1034]. One study site in Brazil reported a high level of antimicrobial resistance [1030]. Organisms recovered from infections included MRSA (77%), methicillin-resistant coagulase-negative Staphylococcus (53.5%), extended-spectrum β-lactamase-producing K. pneumoniae (80%), and carbapenem-resistant P. aeruginosa (33.3%). Another center in Brazil reported a significant difference in resistance to broad-spectrum antimicrobials in pathogens isolated in UTIs from cadaveric kidney transplant recipients (n=21, 19.1%) compared with living-related donor kidney transplant recipients (n=2, 3.7%) (p=0.008) [1024]. One center in the United States reported 94% susceptibility to vancomycin of Enterococcus species within the first month after transplantation, while E. coli, cultured most commonly more than six months after transplantation, was 63% resistant to sulfamethoxazole–trimethoprim [1023]. This resistance may be related to the routine use of sulfamethoxazole–trimethoprim in prophylaxis of Pneumocystis jerovici pneumonia and UTI.

Efficacy

A number of studies have clearly demonstrated that antimicrobial prophylaxis significantly decreases postoperative infection rates in patients undergoing kidney transplantation. These have included at least one randomized controlled trial [1014] and many prospective and retrospective studies comparing infection rates with prophylaxis and historical infection rates at specific transplantation centers [1015–1018,1021,1033–1035]. Based on the available literature, the routine use of systemic antimicrobial prophylaxis is justified in patients undergoing kidney transplantation.

Two studies that evaluated a triple-drug regimen consisting of an aminoglycoside, an antistaphylococcal penicillin, and ampicillin found infection rates of <2%, compared with 10–25% with no antimicrobial prophylaxis [1018,1019]. More specifically, infection rates in patients without antimicrobial prophylaxis (45 cadaveric and 44 living-related donors) were 10.1% in total (8.9% and 11.4%, respectively), compared with 1.5% in total (1.5% and zero, respectively) with antimicrobial prophylaxis [1018]. Infection rates were as high as 33% in living-related patients with no antimicrobial prophylaxis and zero to 1% in both cadaveric and living-related transplant recipients with antimicrobial prophylaxis [1021]. Piperacillin plus cefuroxime was also shown to be efficacious; infection rates were 3.7%, compared with 19% in cadaveric transplant recipients not receiving prophylaxis [1018]. Several studies have shown that single-agent prophylaxis with an antistaphylococcal penicillin [1029,1034], a first-generation cephalosporin [1016,1017,1023,1024,1029], a second-generation cephalosporin [1028,1035,1036], or a third-generation cephalosporin (e.g., cefoperazone, cefotaxime, ceftriaxone) [1024,1029,1033,1037] can reduce postoperative infection rates to zero to 8.4%. All studies included cadaveric transplant recipients, whereas living-related transplant recipients were also included in select studies [1017,1024,1028,1036]. Where compared directly, infection rates between cadaveric and living-related transplant recipients receiving antimicrobial prophylaxis were not statistically different [1024].

Pediatric efficacy

Although pediatric patients were included in studies demonstrating the efficacy of antimicrobial prophylaxis, there are few data specific to pediatric patients. One cohort of 96 pediatric patients who underwent 104 renal transplants (63% cadaveric and 37% living-related donors) ranged in age from six months to 18 years (mean age, 8.2±5.5 years) [1040]. Antimicrobial prophylaxis included one dose of cefotaxime 30-mg/kg i.v. bolus at the start of the procedure and cefotaxime 90 mg/kg/day in three divided doses during the intensive care unit stay, which averaged one to two days. No SSIs were reported.

Recommendations

The recommended agent for patients undergoing kidney transplantation is cefazolin (Table 2). (Strength of evidence for prophylaxis=A.) For patients who are allergic to β-lactam antimicrobials, clindamycin or vancomycin given in combination with gentamicin, aztreonam, or a fluoroquinolone is a reasonable alternative. The duration of prophylaxis should be restricted to 24 h or less. The use of aminoglycosides in combination with other nephrotoxic drugs may result in renal dysfunction and should be avoided unless alternatives are contraindicated. (Strength of evidence for prophylaxis=C.) For patients at high risk of Candida infection, fluconazole adjusted for renal function may be considered.

Background

Plastic surgery encompasses a broad range of procedures focused on reconstructive, dermatological, and cosmetic procedures [1041]. The primary goal of these procedures is to restore function to the affected area, with a secondary goal of improving appearance. The scope of procedures ranges from simple primary surgical site closure, skin grafts, and skin flaps to composite tissue transplantations. Composite tissue transplantation for tissue reconstruction of the knee joint, larynx, uterus, abdominal wall, hand, face, and penis has been performed in a small number of patients [1042,1043].

Most dermatological, breast (reduction and reconstructive), clean head and neck, and facial procedures have an associated SSI rate of <5% [1044–1053]. Oral procedures, such as wedge excision of lip or ear, flaps on the nose [1046,1054], and head and neck flaps, have SSI rates of approximately 5–10% [1053,1055–1060]. In addition to general risk factors as described in the Common Principles section, factors that increase the risk of postoperative infectious complications for plastic surgery procedures include implants [1061], skin irradiation before the procedure, and procedures below the waist [1062,1063].

Organisms

The most common organisms in SSIs after plastic surgery procedures are S. aureus [1045,1049,1050,1053,1054,1056,1063–1068], other staphylococci, and streptococci [1045,1054,1064,1066,1067]. Procedures involving macerated, moist environments (e.g., under a panus or axilla of an obese individual), below the waist, or in patients with diabetes are associated with a higher rate of infection with gram-negative organisms such as P. aeruginosa, [1068] Serratia marcescens, or Enterobacteriaceae, including E. coli [1065,1068], Klebsiella species [1068], and P. mirabilis [1065].

Efficacy

The efficacy of antimicrobial prophylaxis in select plastic surgery procedures has been investigated in several clinical trials and cohort studies. Most placebo-controlled and retrospective studies for many clean plastic surgery procedures have found that antimicrobial prophylaxis does not significantly decrease the risk of infection. These studies have evaluated head and neck procedures (facial bone fracture, tumor excision and reconstruction, radical neck dissection, rhinoplasty) [1049], flexor tendon injury repairs [1051], augmentation mammoplasty using periareolar submuscular technique [1052], carpal tunnel [1069], and breast procedures (reduction mammoplasty, lumpectomy, mastectomy, axillary node dissection) [1056,1058,1070,1071].

However, a Cochrane review of seven randomized, placebo-controlled trials of 1984 patients undergoing breast cancer procedures (axillary lymph node dissection and primary nonreconstructive surgery) evaluated the effectiveness of preoperative or perioperative antimicrobial prophylaxis (n=995) compared with placebo or no treatment (n=989) in reducing the rate of postoperative infections [1072]. Pooled study results revealed a significant difference in SSI rates with antimicrobial prophylaxis (80 [8%] of 995), compared with 10.5% (104 of 989) for no antimicrobial prophylaxis (relative risk, 0.72; 95% CI, 0.53–0.97). Review authors concluded that antimicrobial prophylaxis is warranted to decrease the risk of SSIs in nonreconstructive breast cancer procedures.

Guidelines also support no antimicrobial prophylaxis in patients undergoing clean facial or nasal procedures without an implant [7]. For patients undergoing facial or nasal procedures with an implant, antimicrobial prophylaxis should be considered [7].

A randomized, double-blind, controlled trial of 207 patients evaluated the use of three antimicrobial prophylaxis regimens in patients undergoing abdominoplasty procedures [1066]. The reported SSI rates were 13% for patients receiving no antimicrobial prophylaxis, 4.3% for those receiving preoperative antimicrobials only, and 8.7% for those receiving one preoperative dose and three days of postoperative antimicrobials. There was a significantly lower infection rate in the group receiving preoperative antimicrobials only compared with the placebo group (p<0.05). The infection rate was slightly but not significantly higher in patients who received postoperative antimicrobials.

Pediatric efficacy

Limited data on antimicrobial prophylaxis are available for pediatric patients undergoing plastic surgery procedures. There is no consensus among surgeons regarding the use of antimicrobial prophylaxis in the repair of cleft lip and palate [1074]. The occurrence of postoperative infections after these procedures is 1.3% [1075]. No controlled trials have evaluated the use of antimicrobial prophylaxis in these procedures.

Recommendations

Antimicrobial prophylaxis is not recommended for most clean procedures in patients without additional postoperative infection risk factors as listed in the Common Principles section of these guidelines and the background discussion of this section. Although no studies have demonstrated antimicrobial efficacy in these procedures, expert opinion recommends that patients with risk factors undergoing clean plastic procedures receive antimicrobial prophylaxis. The recommendation for clean-contaminated procedures, breast cancer procedures, and clean procedures with other risk factors is a single dose of cefazolin or ampicillin–sulbactam (Table 2). (Strength of evidence for prophylaxis=C.) Alternative agents for patients with β-lactam allergy include clindamycin and vancomycin. If there are surveillance data showing that gram-negative organisms cause SSIs for the procedure, the practitioner may consider combining clindamycin or vancomycin with another agent (cefazolin if the patient is not β-lactam-allergic; aztreonam, gentamicin, or single-dose fluoroquinolone if the patient is β-lactam-allergic). Postoperative duration of antimicrobial prophylaxis should be limited to less than 24 h, regardless of the presence of indwelling catheters or drains.