perioperative antibiotics

Perioperative Antibiotics

Liaison: Erik Hansen MDLeaders: Katherine Belden MD, Randi Silibovsky MD (US), Markus Vogt MD (International)Delegates: William Arnold MD, PhD, Goran Bicanic MD, PhD, Stefano Bini MD, Fabio CataniMD, Jiying Chen MD, PhD, Mohammad Ghazavi MD, FRCSC, Karine M. Godefroy MD, PaulHolham MD, Hamid Hosseinzadeh MD, Kang II Kim MD, PhD, Klaus Kirketerp-Møller MD,Lars Lidgren MD PhD, Jian Hao Lin MD, Jess H Lonner MD, Christopher C Moore MD,Panayiotis Papagelopoulos MD, Lazaros Poultsides MD MSc PhD, R Lor Randall MD, BrianRoslund PharmD, Khalid Saleh MD MSC FRCSC MHCM, Julia V Salmon MD, Edward SchwarzPhD, Jose Stuyck MD, Annette W Dahl MD, Koji Yamada MD

Published online in Wiley Online Library (wileyonlinelibrary.com). DOI 10.1002/jor.22549

� 2014 Orthopaedic Research Society. Published by Wiley Periodicals, Inc. J Orthop Res 32:S31–S59, 2014.

Question 1: What is the optimal timing of thepreoperative dose of antibiotics?

ConsensusThe preoperative dose of antibiotics should be adminis-tered within 1h of surgical incision; this can beextended to 2h for vancomycin and fluoroquinolones.Furthermore, surveillance measures are critical inensuring clinician compliance with this objective.

Delegate VoteAgree: 97%, Disagree: 2%, Abstain: 1% (Strong Con-sensus)

JustificationThe scientific rationale for antibiotic prophylaxis is toinhibit or eliminate contaminating microorganismsthat gain access to the surgical site during theprocedure, which reduces the probability of an estab-lished infection. Thus, the goal of administeringpreoperative antibiotics is to allow for adequate tissue(blood, soft tissue, and bone) concentrations by thetime of incision. These antibiotics should exceedthe minimum inhibitory concentration (MIC) for theorganisms likely to be encountered for the durationof the operation. This depends on the antibioticused. There are a number of studies, which validatethe importance of the preoperative dose of antibioticsin decreasing periprosthetic joint infection (PJI)and surgical site infection (SSI) in total joint arthro-plasty (TJA). However, there are conflicting opinionsas to the optimal timing of this dose. Some studiessuggest that within 2 h of incision is best, whileothers recommend scheduling the dose as close tosurgical incision as possible. There are several insti-tutional guidelines which support a 1-h preoperativedose of antibiotics as a Surgical Care ImprovementProject (SCIP) measure. In addition to these guide-lines, it is critically important to have surveillancemeasures in place to document compliance with theseprotocols.

The American Academy of Orthopaedic Surgeons(AAOS), the Centers for Disease Control (CDC), andSCIP guidelines recommend that prophylactic anti-

biotics be completely infused within 1h before thesurgical incision.1 The AAOS recommendation for theuse of intravenous antibiotic prophylaxis in primaryTJA, recommendation 2, states that “timing anddosage of antibiotic administration should optimizethe efficacy of the therapy. Prophylactic antibioticsshould be administered within 1h before skin inci-sion.” Due to extended infusion time, vancomycin andfluoroquinonlones should be started within 2h beforeincision. When a proximal tourniquet is used, theantibiotic must be completely infused before inflationof tourniquet.2 The US advisory statement recom-mends that antimicrobial prophylaxis be administeredwithin 1h before incision and discontinued within 24hafter the end of the operation,3 while European guide-lines recommend a single dose within 30min beforeincision.4

Timing< 2 hThe seminal article on this subject studied the timingof administration of prophylactic antibiotics and therisk of surgical wound infections in clean and clean-contaminated cases at a large community hospital.5 Ina study of 2,847 patients, 313 (11%) received TJA. Theauthors found that the rate of infection was lowest forpatients who received an antibiotic from 0 to 2h beforethe incision.5 Specifically, of the 1,708 patients whoreceived prophylactic antibiotics during this timeframe, only 10 (0.6%) subsequently developed SSIcompared to 14 (3.8%) of 369 patients who receivedantibiotics 2–24h preoperatively, 4 (1.4%) of 282patients who received antibiotics within 3h afterincision, and 16 (3.3%) of 488 patients who receivedantibiotics 3–24h following incision. However, thisstudy was conducted in 1985–1986, when there wasconsiderable variation in timing of administration ofthe prophylactic antibiotic, and only 35% of patientsreceived their dose within the contemporary standardof 1 hour prior to incision. Furthermore, the study didnot find a significant difference in SSI rates whenantibiotics were administered within 1–2h prior toincision compared with antibiotics administered 0–3hpostoperatively.

JOURNAL OF ORTHOPAEDIC RESEARCH JANUARY 2014 S31

Timing <1 hThe leadership of the Medicare National SurgicalInfection Prevention Projected hosted the SurgicalInfection Prevention Guideline Writers Workgroup(SIPGWW) meeting and utilized the available litera-ture to draft a consensus paper. The position of theSIPGWW is that the infusion of the first antimicrobialdose should begin within 60min before incision.3,6

Galandiuk et al. combined the results of twoprospective randomized controlled trials (RCT) thatcompared antibiotic prophylaxis (either single-dosepiperacillin with multi-dose cefoxitin) in elective surgi-cal procedures of the gastrointestinal tract. Theauthors found that among other negative predictors,administration of an antibiotic for longer than 60minpreoperatively was associated with a higher rate ofinfectious complications.7

In a large, retrospective cohort study using Nation-al Veterans Affairs data on prophylactic antibiotics of32,459 surgical procedures from 2005 to 2009, Hawnet al. found that higher SSI rates were observed forantibiotic administration more than 60min prior toincision (unadjusted odds ratio (OR) 1.34, 95% confi-dence interval (CI) 1.08–1.66) compared with proce-dures in which antibiotics were administered within1h of incision. However, in generalized additive mod-els adjusted for patient, procedure, and antibioticvariables, no significant association was seen betweenprophylactic antibiotic timing and SSI.8

Timing 30–60minIn a prospective cohort study at a single academichospital analyzing the incidence of SSI by the timingof antimicrobial prophylaxis in a consecutive series of3,836 surgical procedures, Weber et al. determinedthat administration of single-shot prophylactic cefur-oxime is more effective when given 30–59min beforeincision than administration during the last 30min.The overall SSI rate for this mixed cohort of general,vascular, and orthopaedic surgeries was 4.7% (180),and antimicrobial prophylaxis was administered with-in the final 30min in 59% of all procedures. Multivari-able logistic regression analysis showed a significantincrease in the odds of SSI when antimicrobial prophy-laxis was administered fewer than 30min (crude OR2.01; adjusted OR 1.95, 95% CI, 1.4–2.8; p<0.001) and60 to 120min (crude OR 1.75; adjusted OR 1.74; 95%CI 1.0–2.9, p¼0.035) when compared with the refer-ence interval of 30–59min before incision.9

Timing <30minIn a large, prospective, multicenter observational studyexamining the relationship between antibiotic timingand SSI risk, Steinberg et al. determined that SSI riskincreased incrementally as the interval of time be-tween antibiotic infusion and creation of the incisionincreased. The authors analyzed the antimicrobialprophylaxis of 4,472 randomly selected cardiac, hip, orknee arthroplasty, and hysterectomy cases from 29

contributing hospitals, and ascertained SSI throughthe National Nosocomial Infections Surveillance sys-tem methodology. When antibiotics requiring longinfusion times (e.g., vancomycin) were excluded, theinfection risk following administration of antibioticswithin 30min was 1.6% compared with 2.4% associatedwith administration of antibiotic between 31–60minprior to surgery (OR 1.74; 95% CI 0.98–3.04).10

In another recent multicenter, observational studyfrom the Netherlands assessing risk factors for postop-erative infections in 1,922 total hip arthroplasty (THA)cases, the authors found a similar pattern with adecreased rate of infection in those who receivedprophylaxis within 30min prior to incision, although itdid not reach statistical significance.4 These authorscollected data about SSI and potential risks factorsrelated to prophylaxis, the patient, and procedure from11 hospitals that participated in the Surgical Prophy-laxis and Surveillance Intervention project and usedmultivariate logistic regression analysis to identifythose variables that were predictive of SSI. Althoughthere was a non-significant trend for the lowest SSIrate in those patients who received prophylaxis 30minbefore surgery, the highest odds ratios for SSI werefound in patients who received prophylaxis afterincision (2.8, 95% CI 0.9–8.6, p¼0.07) and prolongedduration of surgery was the only statistically signifi-cant risk factor for SSI following THA.

Timing with Tourniquet UseIn an RCT of 22 patients in which cefuroxime prophy-laxis was administered at various intervals (5, 10, 15,or 20min) before inflation of the tourniquet for totalknee arthroplasty (TKA), Johnson et al. measuredantibiotic levels of bone and subcutaneous fat through-out the operation. They found that an interval of10min prior to tourniquet inflation was necessary toobtain adequate prophylaxis. While the patientsobtained adequate levels in bone at 5min, an intervalof 10min or more was required for patients to havetherapeutic levels in the subcutaneous fat.11

In another similar RCT, 24 patients undergoingTKA were randomized to receive cefazolin 1, 2, or 5minbefore tourniquet inflation. Serum, soft tissue, andbone samples were measured for adequate cefazolinconcentration (defined as 4�MIC 90 (MIC 90¼1mg/ml). The median percentage of cefazolin penetrationinto soft tissue and bone for the 5-, 2-, and 1-mingroups was 14.5% and 4.6%, 6.7% and 3.0%, and 5.9%and 4.6%, respectively. The authors also noted that thepercentage of patients achieving the ratio of 4�MIC 90for soft tissue and bone was highest in the 5-min groupcompared with either the 2- or 1-min groups.12

In another prospective study by Soriano et al., 908patients undergoing TKA were randomized to receiveeither 1.5 g of cefuroxime 30min before inflation oftourniquet and placebo 10min before release of tourni-quet (standard group) or placebo 30min before infla-tion of tourniquet and 1.5 g cefuroxime 10min before

S32 JOURNAL OF ORTHOPAEDIC RESEARCH VOLUME 32 SUPPLEMENT 1

JOURNAL OF ORTHOPAEDIC RESEARCH JANUARY 2014

release of tourniquet. There was no difference amongthe patients with regard to various risk factors forSSI/PJI. The authors did not find a significant differ-ence in the incidence of infection at 3.6% for thestandard group and 2.6% for the control group at12 months. The authors concluded that administrationof antibiotics just prior to release of tourniquet wasnot inferior to a standard prophylactic regimen.13

Surveillance MeasuresIn a study evaluating the impact of a new nationalproject meant to reduce infections in arthroplastysurgery in Sweden, Dahl et al. found that only 57% ofpatients received preoperative antibiotics during therecommended time frame. In 2009, following theintroduction of the World Health Organization surgi-cal checklist and a new Swedish Knee ArthroplastyRegister (SKAR) reporting form, which included thetime for administration of preoperative antibiotics, thenumber of patients receiving appropriately-timeddoses of preoperative antibiotics increased to 69% in2009 and 79% in 2010.14

Question 2: Is there an optimal antibiotic that shouldbe administered for routine perioperative surgicalprophylaxis?

ConsensusA first or second generation cephalosporin (cefazolin orcefuroxime) should be administered for routine periop-erative surgical prophylaxis. Isoxazolyl penicillin isused as an appropriate alternative.


JustificationA first or second generation cephalosporin should beadministered for routine perioperative surgical prophy-laxis because of its broad spectrum of action, cost-effectiveness, and the need to preserve newer and moreexpensive therapies for drug-resistant microorganismsand emerging pathogens. These antibiotics cover gram-positive organisms and clinically important aerobicgram-negative bacilli and anaerobic gram positiveorganisms.6 Additionally, they have excellent distribu-tion profiles in bone, synovium, muscle, and hemato-mas.15 Many studies have documented that minimumbactericidal concentrations for most non methicillin-resistant Staphylococcus aureus (MRSA) organisms areachieved rapidly in these tissues-ie within minutesafter their administration.16,17 The optimal prophylac-tic antibiotic should be bactericidal (penicillin, cephalo-sporin, vancomycin, or aminoglycosides), not simplybacteriostatic (clindamycin, which is a lincosamide).The agent should also have a half-life that covers thedecisive interval (the first 2h after incision or contami-nation) with therapeutic concentrations from time ofincision to wound closure. Failure to maintain tissue

concentrations above the MIC increases the risk ofwound infection.18 In Scandinavia and elsewhere,isoxazolyl penicillin, such as cloxacillin, flucloxacillin,nafcilin, or oxacillinis, is used as an appropriatealternative. Some institutions administer carbapenems(namely imipenem/cilastin and meropenem) to patientswith penicillin allergy, as they felt that the potentialfor cross-reactivity between carbapenems and penicillinis less than traditionally believed.19

In a multicenter, placebo RCT, Hill et al. convincing-ly demonstrated the efficacy of cefazolin for antimicro-bial prophylaxis in reducing the risk of PJI. In 2,137THA patients randomized to either 5 days of cefazolinor placebo antibiotic prophylaxis reduced the incidenceof deep infection from 3.3% to 0.9% (p<0.01).20

Tyllianakis et al. performed an RCT comparingcefuroxime to two specific antistaphylococcal agents(fusidic acid and vancomycin) for prophylaxis inTHA and TKA in an institution where MRSA andmethicillin-resistant S. epidermidis (MRSE) preva-lence exceeded 25% of orthopaedic infections. In 435patients (260 hips and 175 knees) followed for aminimum of 2 years, the authors found no statisticallysignificant difference between the treatment groupsfor either THA or TKA, although the authors concedethat the power to detect meaningful statistical differ-ences between the groups was low and it was thereforedifficult to provide any definitive conclusions.21

The efficacy of 1 day of cefuroxime versus 3 days ofcefazolin on postoperative wound infections was studiedby Mauerhan et al. in a double-blind, multicenter trialof 1,354 patients undergoing hip and knee arthroplasty.The authors found no statistically significant differencebetween the two regimens. For the TKA patients, therate of PJI was 0.6% (1/178) for those receivingcefuroxime versus 1.4% (3/207) for those receivingcefazolin. For the THA patients, the rate of PJI was0.5% (1/187) for those receiving cefuroxime as comparedto 1.2% (2/168) for those receiving cefazolin.22

In a study investigating the bacterial coloniza-tion and resistance patterns of a cohort of patientsundergoing primary joint arthroplasty in Sweden,Stefansdottir et al. noted that in Scandinavia,isoxazolylpenicillin derivative cloxacillin is the mostcommonly used prophylactic antibiotic. Moreover,these b-lactams were effective against 99% of theS. aureus strains and 80% of the coagulase-negativeStaphylococcus (CNS) strains colonizing patients un-dergoing primary TJA. Furthermore, the gentamicin-laden bone cement used in many of these cases coversagainst most of the additional CNS strains.23

Question 3: What is the choice of antibiotic inpatients who have pre-existing prostheses such asheart valves?

ConsensusThe choice of antibiotics for patients with pre-existingprostheses such as heart valves is the same as that forroutine elective arthroplasty.

PERIOPERATIVE ANTIBIOTICS S33



JustificationPatients with preexisting prostheses such as heartvalves are at risk for infective endocarditis due tobacteremia, which is relatively rare but can lead tocatastrophic complications and death. Guidelines for theprevention of infective endocarditis have been publishedby the American Heart Association (AHA) for more than50 years. The first 9 guidelines (published between 1955and 1997) were based on low-level evidence; only morerecently have the guidelines been stratified based onlifetime risk of infective endocarditis. Similar to thechange in recommendations regarding dental prophy-laxis for patients undergoing TJA, the 2007 antibioticprophylaxis guidelines for infective endocarditis fromthe AHA and the Infectious Disease Society of America(IDSA) recommend antibiotic prophylaxis only forpatients at the highest risk of infective endocarditis andonly for selected dental procedures (e.g., those thatinvolve manipulation of gingival tissue or the periapicalregion of teeth or perforation of the oral mucosa).24

Infections that complicate heart valve replacementand prosthetic joint replacement have several features incommon. S. aureus and S. epidermidis are commonpathogens and infection rates are similar.23–25 It isgenerally accepted that antimicrobial prophylaxis reducesthe frequency of early postoperative infections; however,when such infections do occur, they are difficult to controlwithout removing the prosthesis. The antibiotics that arerecommended for endocarditis prophylaxis are similar tothat of prophylaxis against PJI. Similarly, if an infectionis known or suspected to be caused by S. aureus, theantibiotic regimen should contain an antistaphylococcalpenicillin or a cephalosporin; whereas vancomycin shouldbe used in those in whom an infection is known orsuspected to be caused by MRSA.25

While there is literature to support the use of prophy-lactic antibiotics up to 48h postoperatively in cardiacsurgery, this is to prevent deep and superficial sternalwound infection and is not relevant to our discussion ofTJA surgery in a patient with a preexisting heartvalve.26,27 Interestingly, there have been some studiesshowing an increase in the routine use of vancomycin forroutine valve surgery prophylaxis over the past years.Haydon et al.28 reviewed the national practice patternsfor antibiotic prophylaxis in cardiac surgery in Australiaand found that between 2004 and 2008, there was adoubling in the proportion of cardiac units using vancomy-cin for routine prophylaxis from 31% to 62% (p< 0.001).

Question 4: What alternatives are available for routineprophylaxis when cephalosporins are not an option?

ConsensusCurently teicoplanin and vancomycin are reasonablealternatives when routine antibiotic prophylaxis can-not be administered.


JustificationTeicoplanin has proven to be an effective and safeprophylactic agent in prosthetic implant surgery inEurope, but is not yet available in the US, Canada, orChina.29–32 Due to the increased frequency of MRSAand MRSE infections in recent years, the prophylacticuse of alternative antibiotics such as glycopeptides(vancomycin and teicoplanin) in hospitals whereMRSA/MRSE are prevalent may be justified.33 Asvancomycin is more difficult to administer and has ashorter half-life and poorer tolerability profile thanteicoplanin, the latter may be a better choice in thesesettings.34 Teicoplanin is notable for having a longhalf-life (32–176h), low toxicity, and good tissue pene-tration, which allows it to achieve therapeutic concen-trations in bone and surrounding soft tissues.33,35

Ceftaroline (fifth generation cephalosporin) has thesame spectrum of activity as ceftriaxone with addition-al MRSA activity. The US Food and Drug Administra-tion and the European Medicines Agency haveprovided indications for the use of ceftaroline fortreatment of complicated skin and soft tissue infec-tions only and not for prophylaxis.

In one multicenter RCT, Periti et al. comparedadministration of a single dose of teicoplanin (400mgintravenous (IV) bolus at time of anesthesia) versusthat of 5 doses of cefazolin over a 24-h period (2 g atinduction and 1 g every 6h postoperatively) as prophy-laxis in patients undergoing TJA. They randomized846 patients and noted that six patients (1.5%) in theteicoplanin group and seven patients (1.7%) in thecefazolin group developed a surgical wound infectionduring their hospital stay, which was a non-significantdifference. Additionally, a non-significant difference inadverse events was recorded in the two groups, withthree (0.7%) of the teicoplanin patients and nine(2.1%) of the cefazolin patients.32

Question 5A: What antibiotic should be administered ina patient with a known anaphylactic penicillin allergy?

ConsensusIn a patient with a known anaphylactic reaction topenicillin, vancomycin or clindamycin should be ad-ministered as prophylaxis. Teicoplanin is an option incountries where it is available.


Question 5B: What antibiotic should be administeredin a patient with a known non-anaphylactic penicillinallergy?

ConsensusIn a patient with a reported non-anaphylactic reactionto penicillin, a second-generation cephalosporin can be



used safely as there is limited cross-reactivity. Penicil-lin skin testing may be helpful in certain situationsto clarify whether the patient has a true penicillinallergy.


JustificationWhen patients present with a penicillin allergy, fur-ther information should be obtained to determinewhether an Immunoglobulin E(IgE)-mediated re-sponse (anaphylaxis) occurred. In patients with adocumented IgE-mediated response to penicillin, thirdand fourth generation cephalosporins can be used.First and second generation cephalosporins with R1side chains similar to that of penicillin (cefaclor,cefadroxil, cefatrizine, cefprozil, cephalexin, or cephra-dine) should be avoided; first and second generationcephalosporins with different R1 side chains can begiven.

Vancomycin and clindamycin are recommended asalternative agents for patients who have a true type Ib-lactam allergy, manifested by immediate urticaria,laryngeal edema, or bronchospasm.3 Clindamycin is apreferred alternative for persons with an establishedb-lactam allergy or with contraindications to its useand at institutions with low rates of MRSA infection.Clindamycin has good bioavailability and at 30minafter infusion has been shown to exceed the MICs forS. aureus in both animal and human cortical bonesamples.36 However, clindamycin is a bacteriostaticagent. In addition vancomycin alone has a relativelypoor activity against Staphylococcus aureus and clini-cal studies implicate that vancomycin as prophylaxisalone increases the risk for SSI. Therefore a secondagent should be considered (levofloxacine, moxi-floxa-cine) in addition to vancomycin.8

Cross-reactivity between penicillin and cephalospo-rin is overestimated and much lower than reported inearlier studies. The 10% estimate of risk of allergicreactions to cephalosporins in penicillin-allergicpatients is based on data collected and reviewed in the1960s and 1970s. It is due in large part to the widelyreferenced reviews of Dash and Petz, which reportedallergic reactions in 7.7% and 8.1%, respectively ofpenicillin-allergic patients (allergy was based on pa-tient history) and only included first generation ceph-alosporins and second generation cefamandole.37,38

The high cross-reactivity found in earlier studies maybe due in part to contamination of the study drugswith penicillin during the manufacturing process.39,40

Moreover, the authors of the early studies had abroader definition of allergy and did not account forthe fact that penicillin-allergic patients have an in-creased risk of adverse reactions to any medica-tion.41,42 Skin testing in penicillin-allergic patientscannot reliably predict an allergic response to a

cephalosporin, particularly to compounds with dissimi-lar side chains.43 However, skin testing may be usefulin determining whether a true allergy to penicillinexists.44

Twenty-seven articles on the topic of the cross-reactivity of penicillin and cephalosporin werereviewed, of which 2 were meta-analyses, 12 wereprospective cohorts, 3 were retrospective cohorts, 2were surveys, and 9 were laboratory studies. Theauthors demonstrated that penicillin has a cross-allergy with first generation cephalosporins (OR 4.8;CI 3.7–6.2) and a negligible cross-allergy with secondgeneration cephalosporins (OR 1.1; CI 0.6–2.1). More-over, laboratory and cohort studies indicate that theR1 side chain, not the b-lactam ring, is responsible forthis cross-reactivity. The authors conclude that theoverall cross-reactivity between penicillin and cephalo-sporin is lower than previously reported, at 10%,although there is a strong association between amoxi-cillin and ampicillin with first and second generationcephalosporins that share a similar R1 side chain. Theoverall cross-reactivity between penicillin and cephalo-sporin in individuals who report a penicillin allergyis approximately 1% and in those with a confirmedpenicillin allergy 2.55%. For penicillin-allergicpatients, the use of third or fourth generation cephalo-sporin or cephalosporins (such as cefuroxime andceftriaxone) with dissimilar side chains than theoffending penicillin carries a negligible risk of crossallergy.45

A similar review of 44 articles on the evidence ofcross-reactivity between cephalosporin and penicillinin human and animal studies supports the findingthat cephalosporin can be safely prescribed to apatient with a non-life threatening reaction to penicil-lin (including type I anaphylaxis, Stevens–Johnsonsyndrome, toxic epidermal necrolysis, and angioe-dema).46 The relative risk of an anaphylactic reactionto cephalosporin ranges from 1:1,000 to 1:1,000,000and this risk is increased by a factor of 4 in patientswith a history of penicillin allergy.47

Based on an analysis of nine articles that compareallergic reactions to a cephalosporin in penicillin-allergic and non-penicillin-allergic subjects, Pichicheroet al. found that first generation cephalosporins have across-allergy with penicillin, but cross-allergy is negli-gible with second and third generation cephalosporins.Specifically, a significant increase in allergic reactionsto cephalothin (OR 2.5, 95% CI 1.1–5.5), cephaloridine(OR 8.7, 95% CI 5.9–12.8), and cephalexin (OR 5.8, CI3.6–9.2) and all first generation cephalosporins pluscefamandole (OR 4.8, CI 3.7–6.2) were observed inpenicillin-allergic patients; no increase was observedwith second generation cephalosporin (OR 1.1, CI 0.6–2.1) or third generation cephalosporin (OR 0.5, CI 0.2–1.1).41,42

In a retrospective cohort of 2,933 patients whoreceived a cephalosporin (usually cefazolin) duringtheir procedure, including 413 who were allergic to



penicillin, only one of the penicillin-allergic patientsmay have had an allergic reaction to the cephalospo-rin; and one of the non-penicillin-allergic patientsdeveloped a rash while the antibiotic was infused,requiring discontinuation of the antibiotic.48

In a large, retrospective review of 534,810 patientswho received penicillin followed by a cephalosporin atleast 60 days later, Apter et al. noted that a total of3,877 patients had an allergic-like event (ALE) afterpenicillin administration, but only 43 (1.1%) experi-enced a second ALE after receiving cephalosporin(unadjusted risk ratio (RR) 10.0; 95% CI 7.4–13.6).Interestingly, in a separate analysis reviewing sulfon-amide antibiotics, 1.6% of penicillin-sensitive patientsexperienced a second ALE after receiving a sulfon-amide (7.2; 95% CI 3.8–12.5), suggesting that patientswho are allergic to penicillin are at a higher likelihoodof being allergic to other medications in general, notnecessarily indicating that cross-reactivity had oc-curred.49

Park et al. performed a retrospective cohort studyto determine whether patients with a penicillin allergywere at an increased risk of adverse drug reactionswhen administered cephalosporin. Eighty-five patientswith a history of penicillin allergy and positive penicil-lin skin test and 726 patients with a history of penicillinallergy and a negative penicillin skin test were adminis-tered a first generation cephalosporin. Five (6%) of 85cases had an adverse drug reaction to cephalosporincompared to five (0.7%) of 726 of the control population(p¼0.0019). The rate of presumed IgE-mediated adversedrug reactions to the cephalosporin among the cases was2 (2%) of 85 compared to 1 (0.1%) of 726 among thereference population (p¼0.03).50

Question 6: What are the indications foradministration of vancomycin?

ConsensusVancomycin should be considered for patients who arecurrent MRSA carriers or have anaphylactic allergy topenicillins.

Consideration should be given to screening highrisk patients such as:

- Patients in regions with a high prevalence ofMRSA.

- Institutionalized patients (nursing home residents,dialysis-dependent patients, and those who havebeen in the intensive care unit).

- Healthcare workers.


JustificationThe AAOS recommendation for the use of IV antibioticprophylaxis in primary TJA, recommendation 2, statesthat “vancomycin may be used in patients with known

colonization with MRSA or in facilities with recentMRSA outbreaks.”51 Similarly, the consensus positionof the Medicare National Surgical Infection PreventionProject’s SIPGWW meeting was that “for patients withknown MRSA colonization, vancomycin should beconsidered the appropriate antimicrobial agent forprophylaxis.”6 Additionally, the Society for HealthcareEpidemiology of America recently recommendedroutine surveillance cultures at the time of hospitaladmission for patients at high risk for carriage ofMRSA.52

Question 7: Is there evidence to support the routineuse of vancomycin for preoperative prophylaxis?

ConsensusNo. Routine use of vancomycin for preoperative pro-phylaxis is not recommended.


JustificationCurrent data suggest that the role of vancomycin inorthopaedic surgery prophylaxis should be limited.There is ample evidence that vancomycin is inferioragainst methicillin-sensitive strains of staphylococcalspecies when compared to cephalosporin and penicil-linase-resistant penicillin.8,53

Several systematic analyses concluded that no clearbenefit in clinical or cost effectiveness has beendemonstrated for the routine use of vancomycin com-pared with cephalosporin for prophylaxis. However,most of these studies were conducted before theincreasing prevalence of MRSA and may not accurate-ly reflect the current environment. In some hospitals,community-associated MRSA (CA-MRSA) strains arenow responsible for a significant portion of SSIs.54,55

However, there is no consensus about what constitutesa high prevalence of methicillin resistance and noevidence that routine use of vancomycin for prophylax-is in institutions with perceived high risk of MRSAinfection results in fewer SSIs than the use of acephalosporin. Although two RCTs have been con-ducted in institutions with a high MRSA prevalence,the differences in SSI rates and outcomes wereconflicting. Similarly, several studies have utilizeddecision analysis models to calculate MRSA prevalencethresholds for which vancomycin would have clinicalbenefit and be more cost-effective than cephalosporinfor surgical prophylaxis. However, these studies allsuffer from the lack of randomization to providebaseline probabilities for the clinical effectiveness ofeach treatment at different rates of MRSA prevalence.

While there is a growing body of evidence to supportthe routine use of vancomycin for preoperative prophy-laxis, this should be tempered by the fact that there isan increasing threat of colonization and infection withvancomycin-resistant enterococci (VRE)56 and an in-



creased prevalence of MRSA strains with reducedsusceptibility to vancomycin.57,58

The choice of drug prophylaxis should take intoaccount the antibiotic resistance patterns in hospitalsystems. In a recent study by Fulkerson et al.,59 thesusceptibilities of S. epidermidis and S. aureus tocefazolin at two high-volume academic centers in NewYork and Chicago were only 44% and 74%, respective-ly. Of the most common organisms infecting patientsundergoing TJA at these hospitals, 26%–56% wereresistant to the standard recommended prophylacticagent. Thirty-three of the 194 infections were diag-nosed within a month after the surgery. Of these,8 were due to S. epidermidis and 16 were due toS. aureus. Of these, only 2 of the 8 (25%) of theS. epidermidis infections and 11 of the 16 (69%) ofthe S. aureus infections were sensitive to cefazolin.However, these infections were 100% susceptible tovancomycin.

In a study of deep infections following hip and kneearthroplasty over a 15-year period at The RoyalOrthopaedic Hospital and Queen Elizabeth Hospital inEngland, 22 of 75 hip and knee infections (29%) werecaused by microorganisms that were resistant to theantibiotic used for prophylaxis (cefuroxime). Theseincluded all 3 MRSA infections, all 3 Pseudomonasaeruginosa infections, and 11 coagulase-negativestaphylococcus infections.60,61 Wiesel and Esterhai62

recommend administration of vancomycin in institu-tions where the prevalence of MRSA is greater than10% to 20%.

In a hospital with a high prevalence of MRSA,Merrer et al. conducted a prospective, observationalstudy comparing the incidence of SSI after vancomycinor cefazolin prophylaxis before femoral neck fracturesurgery, as well as the impact of antibiotic prophylaxison the emergence of VRE and Staphylococcus aureus.The authors found no significant difference in the rateof SSI, as a total of 8 (3%) occurred, 4% in the cefazolingroup and 2% in the vancomycin group (p¼0.47). At1 week after surgery, there were a total of six patients(2%) who had hospital-acquired MRSA, correspondingto 0.7% in the cefazolin group and 5% in the vancomy-cin group (p¼ 0.04), none of which were resistantto glycopeptides. Additionally, three patients (1%)acquired VRE, all of which were in the cefazolin group(p¼ 0.27).63

Cranny et al. used a combination of systematicreviews and economic modeling in order to answerquestions about whether there is a level of MRSAprevalence at which a switch from non-glycopeptide toglycopeptide antibiotics for routine prophylaxis isindicated in surgical environments with a high risk ofMRSA infection. The effectiveness reviews identified16 RCTs with a further three studies included foradverse events only. They found no evidence tosupport that glycopeptides are more effective thannon-glycopeptides in preventing SSI. Most of the trialsdid not report either the baseline prevalence of MRSA

at the participating surgical units or MRSA infectionsas an outcome. The cost-effectiveness review includedfive economic evaluations of glycopeptide prophylaxis.Only one study incorporated health-related quality oflife and undertook a cost-utility analysis. In conclu-sion, the authors indicate that there is currentlyinsufficient evidence to determine whether there is athreshold prevalence of MRSA at which switchingfrom non-glycopeptide to glycopeptide antibiotic pro-phylaxis might be cost effective.64

Bolon et al. performed a meta-analysis of sevenRCTs published in the cardiothoracic surgery litera-ture that compared SSIs in subjects receiving glyco-peptide prophylaxis with those who received b-lactamprophylaxis. While neither agent proved to be superiorfor prevention of the primary outcome, occurrence ofSSI at 30 days (RR 1.14, 95% CI 0.91–1.42), vancomy-cin prophylaxis was superior for the prevention of SSIcaused by methicillin-resistant gram-positive bacteria(RR, 0.54; 95% CI 0.33–0.90) at 30 days after surgery.65

The AAOS recommendations for the use of IVantibiotic prophylaxis in primary TJA, recommenda-tion 2, states that “vancomycin may be used inpatients with known colonization with MRSA or infacilities with recent MRSA outbreaks.”1 The HospitalInfection Control Practices Advisory Committee guide-line also suggests that a high frequency of MRSAinfection at an institution should influence the use ofvancomycin for prophylaxis but acknowledges thatthere is no consensus about what constitutes a highprevalence of methicillin resistance.66

Two prospective RCTs have evaluated antibioticprophylaxis in hospitals with a high prevalence ofMRSA. Tacconnelli et al. randomized patients under-going surgery for cerebrospinal shunt placement toreceive either vancomycin or cefazolin. The prevalenceof MRSA in 2001 for a 1,700-bed university hospitalwas reported as one new case of MRSA infection per100 hospital admissions. Shunt infections developed in4% of patients receiving vancomycin (4/88) and 14%receiving cefazolin (12/88, RR, 0.22; 95% CI 0.11–0.99,p¼ 0.03). The infecting pathogen was MRSA in 2 of 4patients (50%) receiving vancomycin and 9 of 12 (75%)patients receiving cefazolin.67 Finkelstein et al. ran-domized 855 patients undergoing cardiothoracic sur-gery to either a vancomycin or cefazolin group. Theprevalence of new cases of MRSA infection in thecardiac surgery ward was reported to be 3.0 and 2.6per 100 admissions in 1995 and 1996, respectively.The overall rates of SSI were similar in both groups(9.5% for vancomycin and 9.0% for cefazolin). A trendtoward more methicillin-resistant gram-positive infec-tions was observed in the cefazolin group (4.2% vs.2.0%; p¼0.09), while more methicillin-sensitive staph-ylococcus infections were seen in patients receivingvancomycin (3.7% vs. 1.3%; p¼ 0.04).68

Three other clinical studies have used pre- andpost-intervention periods to assess the effect of switch-ing to vancomycin for surgical prophylaxis in patients



undergoing cardiothoracic or orthopaedic surgery.Garey et al. demonstrated that a change from cefurox-ime to vancomycin prophylaxis decreased the averagemonthly SSI rate by 2.1 cases/100 coronary arterybypass graft (CABG) procedures when compared withpatients undergoing cardiac valve replacement sur-gery. This was attributed to a lower rate of infectionscaused by MRSA and CNS during this 4-year study ofnearly 6,500 patients.69 Similarly, Spelman et al.reported a decrease in SSI rates from 10.5% to 4.9%(p<0.001) after switching the antibiotic prophylaxisregimen from cefazolin to vancomycin plus rifampin in1,114 CABG procedures. This was attributed to adecrease in the incidence of MRSA infections from67% during the 1 year pre-intervention period to 0% inthe 1 year post-intervention period.70 Smith et al.retrospectively reviewed total and MRSA PJI in 5,036primary TJAs as well as the cure rate of PJI in a2 year pre-intervention period when cefazolin was theantibiotic prophylaxis of choice to the 2 year post-intervention period when vancomycin was the antibi-otic prophylaxis of choice. They found that with theuse of vancomycin the total rate of PJI was significant-ly reduced (1.0% vs. 0.5%, p¼0.03) and the rate ofMRSA PJI was also reduced (0.23% vs. 0.07%,p¼ 0.14). Furthermore, PJIs were more successfullytreated with irrigation and debridement only, notrequiring antibiotic spacers (76.9% vs. 22.2%,p¼ 0.002).71

A study published on Australian Surveillance Data(Victorian Healthcare Associated Surveillance System)of over 20,000 cardiac and arthroplasty proceduresidentified 1,610 cases in which vancomycin was admin-istered as compared to 20,939 cases in which a b-lactam was used. The adjusted OR for an SSI withmethicillin-sensitive Staphylococcus aureus (MSSA)was 2.79 (95% CI 1.6–4.9) when vancomycin prophy-laxis was administered (p< 0.001), whereas the unad-justed OR for an SSI with MRSA was 0.44 (OR 0.19–1.004; p¼0.05).72

Several recent studies have developed decisionanalysis models to determine the threshold of MRSAprevalence at which vancomycin would minimize theincidence and cost of SSI. For CABG surgery, theauthors of two studies have recommended a MRSAprevalence threshold of 3% among infections caused byS. aureus.73–75. Miller et al. suggested that lower ratesof MRSA prevalence (e.g., 3–10%) were within theerror of their model and that surgical prophylaxis withvancomycin would have a modest effect in reducingthe incidence of SSI. For vascular surgery, a MRSAprevalence of 50% was suggested before a b-lactamagent is replaced with vancomycin for surgical prophy-laxis.76 The authors also suggested that an amino-glycoside should be added to the prophylactic regimenonce the prevalence of MRSA reaches 10%, which is inagreement with the recent guidelines from the BritishSociety of Antimicrobial Chemotherapy.77 Elliot et al.developed an economic model to explore the cost-

effectiveness of vancomycin and/or cephalosporin forsurgical prophylaxis in patients undergoing THA.Vancomycin was recommended when the rate ofMRSA SSIs is �0.15% and the rate of non-MRSA SSIsis �0.1%, or when the rate of MRSA infections is�0.2% and the rate of other infections is >0.2%.78

Each of these decision analysis studies noted that theirbiggest limitation was the lack of available evidencefrom RCTs, with a high prevalence of MRSA infectionsas one of the most important factors that influencedmodeling assumptions.

Question 8: Is there a role for routine prophylacticuse of dual antibiotics (cephalosporins andaminoglycosides or cephalosporins andvancomycin)?

ConsensusRoutine prophylactic use of dual antibiotics is notrecommended.

Delegate VoteAgree: 85%, Disagree: 14%, Abstain: 1% (StrongConsensus)

JustificationClinical studies have used pre- and post-interventionperiods to assess the effect of switching to vancomycinfor surgical prophylaxis in patients undergoing cardio-thoracic surgery. Walsh et al. implemented a compre-hensive MRSA bundle program in which vancomycinwas added to the routine cefazolin prophylaxis regi-men for patients who tested positive for nasal MRSAcarriage. Other components of the program includeddecolonization of all cardiothoracic staff who screenedpositive for nasal MRSA, application of nasal mupir-ocin ointment for 5 days in all patients starting 1 daybefore surgery, application of topical mupirocin to exitsites after removal of chest and mediastinal tubes, andrescreening of patients for MRSA colonization at thetime of hospital discharge. This program resulted in asignificant reduction in the SSI rate (2.1–0.8%,p< 0.001) as well as a 93% reduction in postoperativeMRSA wound infections (from 32 infections/2,767procedures during the 3-year pre-intervention periodto 2 infections/2,496 procedures during the 3-yearpost-intervention period).79

Dhadwal et al. conducted a double-blind RCT tocompare the efficacy of a 48h, weight-based dosing ofvancomycin plus gentamicin and rifampin versus a24-h cefuroxime regimen for antibiotic prophylaxisof sternal wound infections in a high-risk group ofpatients undergoing CABG surgery. The infectionrates significantly decreased from 23.6% (25/106) inthe cefuroxime group to 8.4% (8/95) in the combinationvancomycin group (p¼0.004).80 Patrick et al. con-ducted an RCT to compare cefazolin and combinationsof cefazolin and either vancomycin or daptomycin in181 low-risk patients undergoing vascular surgery.



Only six postoperative MRSA infections were reported(two in the cefazolin group, four in the vancomycinplus cefazolin group, and 0 in the daptomycin pluscefazolin group), making the interpretation of thedifferences between antibiotic regimens difficult.81

Sewick et al. retrospectively reviewed 1,828 prima-ry TJAs that received either a dual antibiotic regimenof cefazolin and vancomycin or received cefazolin alonein order to determine the rate of SSI as well as themicrobiology of subsequent SSI. There was a total of22 SSIs (1.2%) with no significant difference in theinfection rate between the dual antibiotic prophylaxisgroup compared to the single antibiotic regimen (1.1%and 1.4%, respectively, p¼0.636), while the preva-lence of subsequent MRSA infection was significantlylower (0.002% vs. 0.08%, p¼0.02).82 Ritter et al.83

administered a single prophylactic dose of vancomycinand gentamicin in a cohort of 201 consecutive TJApatients and documented bactericidal blood concentra-tions during and for 24h after surgery with nopostoperative infections.

Elliot et al. developed an economic model to explorethe cost effectiveness of vancomycin and/or cephalospo-rin for surgical prophylaxis in patients undergoingTHA. Combination therapy (such as vancomycin plusa cephalosporin) was recommended when the rate ofMRSA SSIs is �0.25% and the rate of non-MRSA SSIsis �0.2%).78

Thus, based on the available literature, this work-group feels that dual antibiotics may be utilized toallow broad coverage in institutions or regions wherethere is a high rate of MRSA infection for whichprophylactic vancomycin use is deemed appropriateunder question 6 above.

Question 9: What should be the antibiotic of choicefor patients with abnormal urinary screening and/oran indwelling urinary catheter?

ConsensusThe presence of urinary tract symptoms should triggerurinary screening prior to TJA. Asymptomatic patientswith bacteriuria may safely undergo TJA providedthat routine prophylactic antibiotics are administered.Patients with acute urinary tract infections (UTI) needto be treated prior to elective arthroplasty.


JustificationThere is sparse literature on the risk of deep jointinfection in patients with abnormal perioperativeurinalysis. While several case reports in the 1970slinked postoperative UTIs to PJI,84,85 the literaturesupporting the correlation between preoperative UTIsand PJI following TJA is inadequate.86 Only threestudies have directly addressed the relationship be-

tween preoperative bacteriuria and PJI following TJA,none of which observed a positive correlation.87–89 Toour knowledge there are no studies of patients withsymptomatic UTI undergoing TJA with routine periop-erative prophylactic antibiotics. There is no evidenceeither in support of or against proceeding with surgeryin this cohort of patients.

The presence of UTI symptoms should serve as apreliminary screening tool for surgical clearance of theTJA candidate. Symptoms can then be classified aseither irritative or obstructive. Irritative symptoms(such as dysuria, urgency, or frequency) may or maynot be related to bacteriuria and a noncentrifugedclean catch midstream urine sample should be evalu-ated for white blood cells (WBCs) in these patients. Inpatients with >104WBC/ml, a bacterial count andculture should be obtained and in patients with >4WBC/high power field and bacterial count >103/mL,surgery should be postponed until an appropriatecourse of microbe-specific antibiotics is administeredand repeat urinalysis is obtained. On the other hand,asymptomatic patients with bacteriuria may safelyundergo TJA provided routine prophylactic antibioticsare administered. Patients with obstructive symptomsshould undergo urologic evaluation before arthro-plasty, as postoperative urinary retention has beenshown to be a risk factor for PJI.86,90,91

In a prospective, multicenter study of 362 knee and2,651 hip arthroplasty cases, the authors reported adeep joint infection rate of 2.5% for knee and 0.64% forhip cases at 1-year follow-up. While univariate analy-sis showed no association between deep joint infectionand preoperative UTI (>105CFU/ml), multivariateregression analysis indicated that postoperative UTIincreased the risk of hip PJI.88

Of 1,934 surgical cases (1,291 orthopaedic surger-ies) performed at a Veterans Administration hospital,a preoperative urine culture was obtained in 25%(489) of cases. Of these, bacteriuria was detected in 54(11%) patients, of which only 16 received antimicrobialdrugs. The incidence of SSI was similar between thosewith bacteriuria and those without (20% vs. 16%,p¼ 0.56), while the rate of postoperative UTI wasmore frequent among patients with bacteriuria thanthose without (9% vs. 2%, p¼ 0.01). Among the 54patients with a positive urinary culture, treated anduntreated patients were compared. Unexpectedly, agreater proportion of treated patients developed anSSI (45% vs. 14%, p¼ 0.03). This effect was greatestamong patients with high count bacteriuria(>105CFU/ml), with SSI occurring in 4 of 8 (50%) oftreated versus 1 of 15 (7%) of untreated (p¼ 0.03).These results led the authors to conclude that in thissystem preoperative urinary cultures were inconsis-tently ordered and that when they were, they wererarely positive for bacteriuria. Even when bacteriuriawas detected, it was usually not treated. The authorsnoted that treating bacteriuria associated with SSI islikely confounded by factors that contributed to the



initial decision to administer antimicrobials in the firstplace.92

A retrospective study of 274 THAs found that fivepatients with PJI had perioperative UTIs. However,the same organism was isolated from the urinary tractand hip in only three patients. Of these, only one hada documented preoperative urinalysis.93 A retrospec-tive analysis of 277 patients (364 TJAs) showed that35 patients had evidence of preoperative or periopera-tive UTI with colony counts greater than 105CFU/mlon preoperative clean-catch urine specimens. Onlythree patients (1.1%) developed joint infections at 9,19, and 45 months, respectively, and none was thoughtto be due to perioperative UTI.87 Another retrospectiveanalysis found 57 (55 asymptomatic, 2 symptomatic) of299 arthroplasty patients had bacteriuria on admis-sion. Twenty of the 57 patients went to surgery beforethe routine culture results were available, but postop-eratively received appropriate antibiotics for treat-ment of the UTI. Another 18 patients underwentsurgery during their treatment course for preopera-tively-diagnosed UTI, while the other 19 patientscompleted an appropriate antibiotic course prior tosurgery. None of the patients developed a PJI, whichled the authors to conclude that a treatment course ofantibiotics can be implemented at any time perioper-atively once culture data are obtained.89

The incidence of bacteriuria rises from 0.5% to 1%for a single in-and-out catheterization, 10–30% forcatheters in place for up to 4 days, and up to 95% forcatheters in place for 30 days or more.94,95

Question 10: Should the preoperative antibioticchoice be different in patients who have previouslybeen treated for another joint infection?

ConsensusThe type of preoperative antibiotic administered to apatient with prior septic arthritis or PJI should coverthe previous infecting organism of the same joint. Inthese patients, we recommend the use of antibiotic-impregnated cement, if a cemented component isutilized.


JustificationThere is no evidence that septic arthritis or a PJI canbe completely cured. Jerry et al. conducted a study of65 patients who underwent TKA and had a historyof prior sepsis or osteomyelitis around the knee.They reported rates of deep PJI of 4% and 15%,respectively.96

Lee et al. studied a consecutive series of 20 primaryTKAs in 19 patients with a history of prior septicarthritis or osteomyelitis around the knee. Theyperformed a preoperative workup to evaluate for

infection that included serologies and plain radio-graphs in all patients, while eight patients additionallyhad tagged WBC scans and seven patients had aknee aspiration. Intraoperatively, frozen section forevidence of acute inflammation was used to guidedecisions on whether the procedure was done as asingle or staged procedure. All TKA components wereimplanted with antibiotic cement containing 1 g ofvancomycin and 1.2 g of tobramycin/batch of Simplexbone cement. Of the 17 patients with a minimum of2 years follow-up, only one developed a PJI approxi-mately 3.5 years from the index arthroplasty. Of note,this was one of the two patients that had been treatedin a staged manner and additionally had immunosup-pressive comorbidities, including rheumatoid arthritis,insulin-dependent diabetes mellitus, and was takingdaily doses of prednisone.97

Larson et al. performed a retrospective matchedcase control study to review the clinical results of 19patients who underwent TKA after infected tibialplateau fractures, comparing them to 19 control sub-jects matched for age, gender, and arthroplasty year,who underwent TKAs for tibial plateau fractureswithout a history of infection. Of the 19 case patients,13 underwent one-stage TKA, while the remainderunderwent a staged TKA with either an antibioticspacer or debridement and intravenous antibiotictherapy. Antibiotic cement was used in the majority ofpatients. Previously infected knees were 4.1 timesmore likely to require additional procedures for com-plications compared with knees with no previousinfection (95% CI 1.2–18.3, p¼0.02). The 5 year infec-tion-free survival was 73%�10% in the case groupcompared with 100% in the control group (p¼0.023).The authors recommended that in patients at highrisk less than 1 year since active evidence of infection,a two-stage TKA be performed, with antibiotic therapyand a 4–6 week delay between procedures.98

Question 11: Should postoperative antibiotics becontinued while a urinary catheter or surgical drainremains in place?

ConsensusNo. There is no evidence to support the support thecontinued use of postoperative antibiotics when uri-nary catheter or surgical drains are in place. Urinarycatheters and surgical drains should be removed assoon as safely possible.


JustificationShort-term use of an indwelling catheter after surgeryreduces the incidence of urinary retention andbladder over-distension without increasing the rate ofUTI and is therefore common practice in many hospi-



tals.99 However, it has been shown that there is anincreased risk of UTIs when a catheter is employed formore than 48h.100,101 Urinary retention as well ascatheterization can both lead to bacteriuria,101–103

which increases the risk of deep PJI from 3 to 6times.87,88,104,105

Literature in the field of surgical oncology demon-strates that bacterial colonization of surgical drainsused in breast and axillary procedures is a significantrisk factor for the development of SSI and the micro-organisms that caused SSIs were the same as thosethat colonized the drainage tube in 83% of cases.106

Other studies have demonstrated that there is anassociation between longer duration of drain use andincreased incidence of SSI.107

The AAOS recommendations for the use of IVantibiotic prophylaxis in primary TJA, recommenda-tion 3, states that the “duration of prophylacticantibiotic administration should not exceed the 24-hpostoperative period. Prophylactic antibiotics shouldbe discontinued within 24h of the end of surgery. Themedical literature does not support the continuation ofantibiotics until all drains or catheters are removedand provides no evidence of benefit when they arecontinued past 24h.”2

Colonization of drains by skin organisms can cer-tainly occur, but in only 10% of cases with positivedrain tip culture does overt infection develop.108

Michelson et al. conducted an RCT of 100 TJA patientsusing two methods of bladder management: short term(<24h) indwelling catheters and intermittent cathe-terization. All patients received the same perioperativecefazolin prophylaxis. The authors reported a lowerincidence of urinary retention in the indwelling cathe-ter group (27% vs. 52%, p<0.01) and a lower rate ofbladder distension (7% vs. 45%; p<0.01). Moreover,patients who had an indwelling catheter for more than48h had a significantly higher rate of bladder infection(35%) than patients who were straight catheterizedand/or who had an indwelling catheter for fewer than48h (6%, p<0.01).99

Van den Brand et al. performed a prospective RCTto determine whether an indwelling catheter for 48hor intermittent catheterization leads to less postopera-tive bacteriuria or a UTI with a single dose of cefazolinprophylaxis in primary hip and knee arthroplasties. Intheir protocol, patients received 48h of IV prophylacticcefazolin during the postoperative period. Patientswho had an indwelling catheter in place after the IVantibiotics were completed were treated with oralantibiotic prophylaxis (nitrofurantoin) until catheterremoval. Of the 99 patients who completed the study,14 patients (5 men, 9 women) developed postoperativebacteriuria. The indwelling catheter group hada bacteriuria rate of 24% (11/46) compared with 6% (3/53) in the intermittent catheterization group (p¼0.018).109

Similar findings were reported by Oishi et al., whoreviewed 95 consecutive patients who had been man-

aged with either an indwelling catheter (72h) orintermittent catheterization. Patients who weretreated with an indwelling catheter had significantlylower incidences of urinary retention (7% vs. 84%respectively; p<0.005) and bladder distension (7% vs.41%; p< 0.005) than those who were treated withstraight catheterization. While not statistically signifi-cant, though no patient in the indwelling cathetergroup developed infection, in the intermittent cathe-terization group one patient (2%) had bacteriuria andone patient (2%) had a UTI (p>0.1).110

Koulouvaris et al. performed a retrospective casecontrol study to determine whether a treated preoper-ative or postoperative UTI or asymptomatic bacteri-uria increases the risk of deep PJI and whether theorganisms are the same for the UTI and PJI. Theauthors matched 58 patients who had wound infec-tions with 58 patients who did not develop woundinfection based on age, gender, surgeon, joint, year ofsurgery, and length of follow-up. The authors found noassociation between preoperative UTI and woundinfection (OR 0.34; 95% CI 0.086–1.357, p¼ 0.13), andno association between postoperative UTI and woundinfection (OR 4.22; 95% CI 0.46–38.9, p¼0.20). Onlyone patient had the same bacteria (E. faecalis) cul-tured in the urine and the wound.111

In a survey of the members of the American Societyof Breast Surgeons regarding the use of perioperativeantibiotics for breast operations requiring drains,respondents continued antibiotic prophylaxis for 2–7days or until all drains were removed (38% and 39%,respectively) in cases without reconstruction, while inreconstruction cases 33% of respondents continuedantibiotic prophylaxis for 2–7 days or until all drainswere removed.112 A similar study surveying the Amer-ican and Canadian societies of Plastic Surgeonsregarding drain use and perioperative antibiotic pro-phylaxis in cases of breast reconstruction found that72% of plastic surgeons prescribed postoperative out-patient antibiotics in reconstruction patients withdrains, with 46% continuing antibiotics until drainswere removed.113

Question 12: What is the evidence for the optimalduration of postoperative antibiotics in decreasingSSI or PJI?

ConsensusPostoperative antibiotics should not be administeredfor greater than 24h after surgery.


JustificationMany studies across surgical specialties have beenperformed to compare durations of antibiotic prophy-laxis and the overwhelming majority have not shownany benefit in antibiotic use for more than 24h in



clean elective cases.114–116 Prolonged postoperativeprophylaxis should be discouraged because of thepossibility of added antimicrobial toxicity, selection ofresistant organisms, and unnecessary expense.24

The AAOS recommendations for the use of IVantibiotic prophylaxis in primary TJA, recommenda-tion 3, states that “duration of prophylactic antibioticadministration should not exceed the 24-h postopera-tive period. Prophylactic antibiotics should be discon-tinued within 24h of surgery.”1

Mcdonald et al. performed a systematic reviewacross surgical disciplines to determine the overallefficacy of single versus multiple dose antimicrobialprophylaxis for major surgery. They included onlyprospective RCTs which used the same antimicrobialin each treatment arm whose results were publishedin English. Regardless of fixed models (OR 1.06, 95%CI 0.89–1.25) or random effects (OR 1.04; 95% CI0.86–1.25), there was no significant advantage ofeither single or multiple dose regimens in preventingSSI. Furthermore, subgroup analysis showed no signif-icant differences in the type of antibiotic used, lengthof the multiple dose arm (>24h vs. �24h), or type ofsurgery (obstetric-gynecological vs. other).117

Mauerhan compared the efficacy of a one-day regi-men of cefuroxime with a 3-day regimen of cefazolin ina prospective, double-blinded, multicenter study of1,354 patients treated with arthroplasty and concludedthat there was no significant difference in the preva-lence of wound infections between the two groups. Inthe group treated with primary THA, the prevalenceof deep wound infection was 0.5% (1/187) for thosetreated with cefuroxime compared with 1.2% (2/168)for those who had received cefazolin. In the grouptreated with a primary TKA, the rate of deep woundinfection was 0.6% (1/178) for those treated withcefuroxime compared with 1.4% (3/207) for those whohad received cefazolin.22

Heydemann and Nelson, in a study of hip and kneearthroplasty procedures, initially compared a 24-hregimen of either nafcillin or cefazolin with a 7-dayregimen of the same and found no difference in theprevalence of infection. They then compared a singlepreoperative dose with a 48-h regimen and againfound no difference in infection prevalence. A total of466 procedures was performed during the 4-yearstudy. No deep infections developed in either the one-dose or 48-h antibiotic protocol group. A deep infectiondeveloped in one (0.8%) of the 127 patients in the 24-hprotocol group and in two (1.6%) of the 128 patients inthe 7-day protocol group for an overall infection rate of0.6% (3/466). The authors recognized that as a resultof the small sample sizes, the study lacked the powerto compare the one dose and the more than one dosecategories.118

Stone et al. performed two separate prospective,placebo RCTs of variable-duration antibiotic prophy-laxis in patients undergoing elective gastric, biliary, orcolonic surgery and then in patients undergoing

emergency laparotomy and found that in both cases nosignificant difference was seen in the rate of SSI.Specifically, in a prospective RCT of 220 patientsundergoing elective general surgery who were ran-domized to either perioperative cefamandole plus5 days of placebo or perioperative plus 5 postoperativedays of cefamandole, there was no significant differ-ence in the rate of wound infection (6% and 5%,respectively). In a second prospective RCT of patientsundergoing emergent laporatomy in which cephalothinwas utilized perioperatively, there was no significantdifference in the rate of peritoneal infection betweenthose who received perioperative therapy only (8% and4%, respectively) compared to those who had 5 to7 days of additional postoperative therapy (10% and5%, respectively).119

In a retrospective review of 1,341 TJAs, Williamsand Gustilo found no difference in deep infection ratesbetween a 3 day and 1 day course of prophylacticantibiotics, but emphasized the importance of thepreoperative dose, which was 2 g of cefazolin.120

Clinical studies have used pre- and post-interven-tion periods to assess the effect of antibiotic durationfor surgical prophylaxis. One institution launched asurgical wound infection surveillance program to mon-itor all orthopaedic surgeries and changed the prophy-lactic antibiotic regimen from intravenous cefuroxime(one preoperative and two postoperative doses every8h) to one single preoperative dose of intravenouscefazolin for all clean orthopaedic surgeries. Theauthors of this study found no significant difference inthe superficial and deep wound infection rates in 1,367primary arthroplasties performed with a single preop-erative dose of cefazolin versus three doses of cefurox-ime. The deep wound infection rate for THA was 1.1%(95% CI, 0–3.3%) in the cefuroxime group and 1.1%(95% CI, 0–2.2%) in the cefazolin group (p¼1.0). Thedeep wound infection rate of TKA was 1.6% (95% CI,0–3.8%) in the cefuroxime group and 1.0% (95% CI,0.3–1.7%) in the cefazolin group (p¼ 0.63).121

Question 13: Until culture results are finalized, whatantibiotic should be administered to a patient with apresumed infection?

ConsensusIn a patient with a presumed infection when cultureresults are pending, empiric antibiotic coverage shoulddepend on the local microbiological epidemiology.Culture data should assist in the tailoring of antibioticregimens.


JustificationGuidelines based on individual institutional microbio-logical epidemiology should be developed.122 In the



US, vancomycin is recommended for gram-positivecoverage due to a high rate of resistance to methicillinin many cases and gentamicin or a third or fourthgeneration cephalosporin is recommended for gram-negative coverage. However, in areas with low MRSAprevalence, vancomycin should not be recommendedas the first choice of drug until culture results areobtained and other antibiotics should be choseninstead.

Sharma et al. classified the spectrum and antibioticsusceptibility of bacteria isolated from revision hip andknee arthroplasty specimens in order to recommendappropriate empiric perioperative antibiotics beforedefinitive cultures are obtained. They identified147 patients with positive specimens, yielding 248microorganisms from 195 tissue specimens, 43 fluidspecimens, and 10 swabs. Of the 248 isolated micro-organisms, staphylococcus species was the most com-mon genus encountered (53%), followed by gram-negative isolates (24%). Eighty-eight percent of gram-negative organisms were detected within 48h ofinoculation and 94% of gram-positive organisms with-in 96h. Overall, 46% of isolates were susceptible tocephalothin, while only 35% of CNS were sensitive tocephalothin. No gram-positive vancomycin resistancewas encountered. Therefore the authors concludedthat empiric prophylactic antibiotics for revision hipand knee arthroplasty should include vancomycin forgram-positive organisms and gentamicin for gram-negative bacteria; and if infection is suspected, vanco-mycin and gentamicin should be continued postopera-tively for 96 and 48h, respectively, unless culture orhistology results suggest otherwise.123

KneeIn a retrospective review of 121 patients who under-went revision TKA for infection between 1994 and2008 in the United Kingdom, the most common organ-ism was CNS (49%) and S. aureus (13%). The preva-lence of CNS appears to be increasing, while that of S.aureus and other organisms is decreasing. Vancomycinand teicoplanin were the most effective antibiotics,with overall sensitivity rates of 100% and 96%, respec-tively. Also, the authors reported that based on theirtheoretical model of comparing microorganism sensitiv-ities against specific antibiotics, gentamicin combinedwith vancomycin or teicoplanin is the most effectiveempirical regimen. While the authors recognized thepotential serious nephrotoxic side effects, these anti-biotics may be added to bone cement relatively safely.The authors also suggested that this empirical regimencan potentially allow for a one-stage revision procedureto be conducted when deep infection arises.124

In early, delayed, and late infections observed fromdata from the SKAR from 1986 to 2000 in 426surgically revised cases, CNS was most prevalent (105/299, 35.1%) and twice as common as S. aureus (55/299,18.4%). In hematogenous infections, S. aureus was thedominating pathogen (67/99, 67.7%), followed by strep-

tococci and gram-negative bacteria. Methicillin resis-tance was found in 1/84 tested isolates of S. aureusand 62/100 tested isolates of CNS. During the studyperiod of 1986–2000, methicillin resistance amongCNS increased (p¼0.002). Gentamicin resistance wasfound in 1/28 tested isolates of S. aureus and 19/29tested CNS isolates. Therefore, the authors concludethat empiric antibiotics should cover CNS, as mostearly infections were caused by this organism. Theyalso raised the concern that due to high rate ofgentamicin resistance among CNS in infected TKA,other antibiotics should be used in bone cement atrevision.23

Data from the SKAR have previously been used toreport on the microbiology of 357 TKA infections inpatients operated on before 1986. S. aureus was themost common pathogen (45.4%) followed by CNS(18%).122 In later studies, staphylococci continued tobe the most common pathogens, with S. aureusreported to account for 13–51% of the infections andCNS accounting for 15–49%.124,125,126

HipRafiq et al. retrospectively reviewed the microbiologyof 337 one-stage revision hip replacements for deepinfection and found that CNS was the predominantorganism (67%) and that Staphylococcus (13%) isbecoming more prevalent. The authors also noted anincrease in antimicrobial resistance (24% resistance togentamicin), which lead the authors to suggest thatother antibiotics such as erythromycin or fusidic acidbe added to bone cement during these procedures.127

In a study examining the microbiology of contami-nating bacteria during primary THA, Al-maiyah et al.cultured the gloved hands (n¼ 627 impressions) of thesurgical team in 50 THA cases after draping, at 20-min intervals, and then before cementation. Theyfound contamination present in 57 (9%) of impressionsand a total of 106 bacterial isolates, with CNS beingthe most frequent (68.9%), micrococcus (12.3%) anddiptheroids (9.4%) following, and S. aureus only repre-senting 6.6% of cases. Interestingly, only half (52%) ofthe CNS isolates were sensitive to cefuroxime, theinstitutional prophylactic agent of choice, suggestingalternate agents may be indicated.128

Phillips et al. reviewed the microbiology of deepinfection following hip and knee arthroplasty at aspecialist orthopaedic hospital in the United Kingdomover a 15-year period. At their institution, CNS wasthe most common infecting organism (36%), followedby S. aureus (25%), enterococcus (9%), and MRSA(4%). Of the infecting organisms, 72% were sensitiveto routine prophylactic agents. There was no signifi-cant change in microbiology over that time period atthis institution.129

Timing of InfectionA retrospective analysis of 146 patients who had atotal of 194 positive cultures obtained at time of



revision total hip or knee arthroplasty was performed.Seventy percent of the infections were classified aschronic, 17% as acute postoperative, and 13% as acutehematogenous. Gram-positive organisms caused themajority of the infections (87% or 168/194). The micro-organisms were sensitive to cefazolin in 61% of cases,gentamicin in 88% of cases, and vancomycin in 96% ofcases. The most antibiotic-resistant bacterial strainswere from patients in whom prior antibiotic treatmenthad failed. Acute postoperative infections had a great-er resistance profile than did chronic or hematogenousinfections. Bacteria isolated from a hematogenousinfection had a high sensitivity to both cefazolin andgentamicin. This led to the following recommenda-tions:

� Until final cultures are available, acute hematoge-nous infections should be treated with cefazolinand gentamicin.

� All chronic and acute postoperative infections withgram-positive bacteria and all cases in which agram stain fails to identify bacteria should bemanaged with vancomycin.

� Infections with gram-negative bacteria should bemanaged with third or fourth generation cephalo-sporin.

� Infections with mixed gram-positive and gram-negative bacteria should be managed with a combi-nation of vancomycin and third or fourth genera-tion cephalosporin.

� As 93% (180) of the 194 cultures tested positive bythe fourth postoperative day, the authors recom-mend that if culture results are not positive by thefourth postoperative day, termination of empiricantibiotic therapy should be considered.59

In a retrospective review of 97 patients (106 infec-tions in 98 hips), Tsukayama et al. noted that aerobicgram-positive cocci accounted for 109 (74%) of the 147isolates; gram-negative bacilli, 21 (14%); and anae-robes, 12 (8%). Of the CNS species 27 (48%) wereoxacillin-resistant, while all 33 (100%) of the coagu-lase-positive staph species were sensitive to oxacillin.The authors noted that most of the gram-negativeisolates came from the early postoperative and latechronic infections, while isolates from the acute hema-togenous infections were exclusively gram-positivecocci.130

Irrigation and Debridement (I&D)A retrospective review was conducted to describe themicrobiological spectrum of PJI in 112 patients man-aged with I&D or arthroscopic washout of infectedprosthetic joints between 1998 and 2003 in order toguide the choice of empirical antibiotics. Overall, themost frequently isolated organisms were CNS (47%)and methicillin-sensitive Staphylococcus aureus(MSSA) (44%), while 8% were MRSA and 7% wereanaerobes. In their series, 60% of CNS isolates wereresistant to methicillin. Most gram-negative isolates

were resistant to cefuroxime and all were sensitive tomeropenem. Based on the high rate of early polymicro-bial infection, cephalosporin resistance among gram-negative organisms, b-lactamase resistance amonggram-negative organisms, and b-lactam resistanceamong CNS, the authors recommend glycopeptideswith a carbapenem in the initial regimen, withmodification when culture and sensitivity results areavailable.131

Question 14: What is the appropriate preoperativeantibiotic for a second-stage procedure?

ConsensusThe appropriate preoperative antibiotic for the secondstage should include coverage of the prior organism(s).Cemented arthroplasty components should be insertedwith antibiotic-laden bone cement.


JustificationPatients undergoing reimplantation surgery followinga two-stage exchange procedure are at risk of develop-ing recurrent infection.132,133 The recurrent infectionmay be either due to incomplete eradication of theprior bacteria during the antibiotic spacer exchangeor to a new infection. In order to properly addressboth potential scenarios, the appropriate preoperativeantibiotics should include coverage of the prior organ-ism as well as the most common infecting micro-organisms.

Antibiotic-laden bone cement has been shown todecrease septic failure following TJA in high-riskindividuals and it is US Food and Drug Administra-tion-approved for use during reimplantation of compo-nents in a two-stage exchange. While there is noevidence to support the practice, it makes theoreticalsense to add antibiotics that are effective in treatingthe index infection.

In a systematic review of 31 studies that comparedthe clinical outcomes achieved with one- and two-stagerevision TKA with different types of spacers, theauthors noted that after the index revision for infec-tion, deep joint infection was detected in 0–31% ofcases. Of these, the infection was considered recurrentin 0–18% of cases, while new infection rates variedfrom 0 to 31%. While the length of follow-up did notappear to influence the rate of recurrent infections,the studies with <4 years of clinical follow-up hadfewer new infections.134

Azzam et al. retrospectively reviewed 33 patientswho had failed an initial two-stage exchange arthro-plasty, of whom 18 eventually went on to undergo asecond two-stage procedure. Of this cohort, the isolat-ed organism was different from the previous infectingorganism in only one of 18 patients.132



In a similar study, Kalra et al. retrospectivelyreviewed 11 patients who developed reinfection aftertwo-stage revision for infected THA and were subse-quently treated with a two-stage re-revision. In theirseries, the infecting microorganisms were polymicro-bial in three patients and only two had reinfection bythe initial offending microbe.133

In a review of the outcomes of 69 patients with PJIin TKA, Mont et al. determined that in eight of ninecases reinfections were from the organism that hadcaused the initial infection, although in six of the eightpatients the sensitivity of the organism to antibioticshad changed.126

Kubista et al. published results on 368 patientstreated with a two-stage revision for infected TKA. Ofthis cohort, 58 (15.8%) developed reinfection and acausative organism was identified in 47/58 (81%) ofpatients.135

In a retrospective review of 117 patients whounderwent two-stage exchange arthroplasty for PJI ofthe knee, 33 of 117 patients (28%) required reopera-tion for infection. At the time of reimplantation,antibiotic-laden bone cement (1.2 g tobramycin and 1 gvancomycin per 40 g of cement) was used for fixationof the prosthesis, but there was no note of theparenteral or perioperative antibiotics utilized at thesecond stage.136

Question 15: For surgeries of longer duration, whenshould an additional dose of antibiotic beadministered intraoperatively?

ConsensusAn additional dose of antibiotic should be administeredintraoperatively after two half-lives of the prophylacticagent. The general guidelines for frequency of intra-operative antibiotic administration are provided. Werecommend that re-dosing of antibiotics be consideredin cases of large blood volume loss (>2000 cc) and fluidresuscitation (>2000 cc). As these are independentvariables, re-dosing should be considered as soon asthe first of these parameters are met.


JustificationIn cases of large blood volume loss and fluid resuscita-tion there is a remarkable loss of the prophylacticagent that can result in levels below the MIC. Thesame is true for longer surgeries that extend beyondthe half-life of the agent. Thus, additional antibiotictreatment is needed to re-establish antibiotic levelsthat exceed the MIC. An additional dose of antibiotichas been shown to reduce SSI rates in cardiac patientsand should be administered intraoperatively after twohalf-lives of the prophylactic agent.3,74,75

The AAOS recommendations for the use of IVantibiotic prophylaxis in primary TJA, recommenda-

tion 2, states that “timing and dosage of antibioticadministration should be such to optimize the efficacyof the therapy.”1 Both the IDSA and AAOS state that“Additional intraoperative doses of antibiotic are ad-vised when the duration of the procedure exceeds oneto two times the antibiotic’s half-life or when there issignificant blood loss during the procedure.” Thegeneral guidelines for frequency of intraoperativeantibiotic administration are as follows: cefazolinevery 2–5 (4) h, cefuroxime every 3–4h, clindamycinevery 3–6h, isoxazoyl penicillin every 3h, and vanco-mycin every 6–12h.2,137,138

In a prospective multicenter study exploring therelationship between timing, duration, and intra-operative redosing of surgical antimicrobial prophylax-is and the risk of SSI, Steinberg et al. determined thatintraoperative dosing was associated with a lowerinfection risk only when the preoperative antibioticwas given in the recommended time frame. In 1,062(24%) cases, the surgical procedure lasted for at least4 h. Because of a longer half-life and the reduced needfor redosing, cases that received vancomycin or fluo-roquinolones were excluded from the analysis of theimpact of redosing on infection risk (n¼ 372). Intra-operative redosing was given in 21% of 690 of theselong operations. Of the group that had a surgicalprocedure with a duration of >4h and who receivedthe preoperative dose within 1h, 2 of 112 (1.8%)patients who were redosed intraoperatively developedinfection, compared to 22 of 400 (5.5%) of those whowere not re-dosed (OR 3.08, p¼0.06).10

Scher et al. randomized 801 patients undergoingclean contaminated operations to one of three antibiot-ic regimens: 1 g of cefazolin preoperatively, 1 g ofcefazolin preoperatively and another dose 3h later,and 1 g of cefotetan preoperatively. While all regimensdemonstrated similar wound infection rates for surger-ies lasting less than 3h, for those that exceeded 3h,the group that only received the single preoperativecefazolin dose had a statistically significant higherwound infection rate than those who received thesecond cefazolin dose (6.1% vs. 1.3%, p<0.01).139

Shapiro et al. performed a placebo-controlled RCTto test the efficacy of perioperative cefazolin in pre-venting infection after abdominal or vaginal hysterec-tomy. The authors sub-analyzed the effect of surgeryduration on the efficacy of perioperative prophylaxisby calculating adjusted relative odds of infection withand without prophylaxis for different durations ofsurgery and found that the efficacy of prophylaxisdiminishes rapidly with increasing length of surgery;by 3h, 20min prophylaxis had no measurable effect(OR¼1).140

Polk et al. prospectively analyzed the antibioticlevels of three cephalosporins (cefazolin, cephaloridine,and cephalothin) given as a single preoperative doseand found that acceptable concentrations of cefazolinwere maintained near the incision site until 3 h post-administration, whereas cephalothin did not maintain



wound levels consistent with effective antimicrobialactivity.141

Ohge et al. prospectively examined the pancreatictissue concentrations of cefazolin in 10 patients under-going pancreatectomy and determined the optimalintraoperative time to repeat the dose of cefazolin.Based on their results, the authors recommended asecond dose of kefzol be given 3h after first adminis-tration in order to maintain adequate levels of antibi-otic activity. They measured MIC for 4 bacterialspecies, namely 360 isolates of MSSA, 204 isolates ofK. pneuomoniae, 314 isolates of E. coli, and 30 isolatesof streptococci species; and measured tissue levels ofcefazolin. Antibiotic concentrations in adipose tissueand peritoneum 3h after administration of kefzol werelower than the MIC 80 for K. pneumoniae, E. coli, andstreptococcal species.142

In a retrospective review of 131 patients withprimary colorectal cancer in prolonged operationsexceeding 4h, the surgical wound infection rates were8.5% and 26.5%, respectively for those with (n¼ 47)and without (n¼ 49) intraoperative repeated dosing,which were significantly different based on both aunivariate (p¼0.031) and a multivariate analysis(p¼0.008).143

Zanetti et al. retrospectively compared the risk ofSSIs in 1,548 patients who underwent cardiac surgerylasting >240min after preoperative administration ofcefazolin prophylaxis. The overall risk of SSI wassimilar among patients with (43 [9.4%] of 459) andwithout (101 [9.3%] of 1089) intraoperative redosing(OR 1.01, 95% CI 0.7–1.47). However, redosing wasbeneficial in procedures lasting >400min; infectionoccurred in 14 (7.7%) of 182 patients with redosingand in 32 (16.0%) of 200 patients without (adjustedOR 0.44, 95% CI 0.23–0.86). Intraoperative redosing ofcefazolin was associated with a 16% reduction in theoverall risk for SSI after cardiac surgery, includingprocedures lasting >240min.74,75

Blood LossSwoboda et al. attempted to determine the effect ofintraoperative blood loss on prophylactic cefazolin andgentamicin serum and tissue concentration in a pro-spective study of elective spinal surgical procedureswith expected large blood loss. At 60min after theincision, blood loss correlated with cefazolin tissueconcentrations (r¼�0.66, p¼0.05) and the clearanceof gentamicin from the tissues (r¼0.82, p¼ 0.01).Based on their measured pharmacokinetic values,additional doses of cefazolin should be administeredwhen the operation exceeds 3h and blood loss isgreater than 1500ml. A dose of gentamicin greaterthan 1.8mg/kg should be administered more than30min prior to the surgical incision.144

Blood Loss/Volume ReplacementMarkantonis et al. investigated the effects of surgicalblood loss and fluid volume replacement on gentami-

cin concentrations in serum and in three tissuetypes (subcutaneous fat, epiploic fat, and colonicwall) in patients in undergoing colorectal surgery.Gentamicin was administered at a standard dose of2mg/kg and blood and tissue samples were obtainedconcurrently at specific times throughout each proce-dure. The mean concentration at first surgical incisionwas 7.83 (0.82)mg/ml and decreased to 2.60 (0.28)mg/ml at skin closure, resulting in borderline effective-ness even for susceptible gram-negative microorgan-isms (MIC-1.0). A strong negative correlation wasfound between the intravenously-administered fluidsand gentamicin concentrations in serum and tissues(p�0.04).145

Klekamp et al. prospectively studied orthopaedicpatients with either large or small blood loss who alsoreceived vancomycin prophylaxis to determine theeffect of intraoperative volume shifts on serum vanco-mycin concentrations. There were 6 index patientsin the large blood loss group (greater than 2L) and 7in the control group (less than 2L), with meanestimated blood loss for index and controls was 4.4and 1.0L; and the mean intraoperative fluid resusci-tation, excluding blood products, was 12.4 and 5.1L.respectively. There was a modest inverse correlationbetween blood loss and the intraoperative serum half-life of vancomycin. Although controls maintainedslightly higher intraoperative vancomycin concentra-tions at each time point, there was no statisticallysignificant difference between the groups with regardto absolute concentrations or rate of decline. After8h, the serum concentration of vancomycin exceededthe MIC-90 for S. aureus by approximately eightfold inall but one case patient, who was morbidly obese andhad massive blood loss. Thus blood loss during ortho-paedic procedures has a minimal effect on the intra-operative kinetics of vancomycin and administeringvancomycin every 8–12h seems appropriate for mostpatients.146

Two well-controlled studies of surgical prophylaxiswith cefazolin similarly demonstrated minimal effectsof blood loss on drug concentrations during THA andspine fusion procedures. Meter et al. examined theeffect of intraoperative blood loss and volume resusci-tation during THA on serum levels of cefazolin in18 patients. At 4 h after administration, the serumlevel of cefazolin was 45mcg/ml, which far exceededthe MIC for S. aureus (0.5mcg/ml), despite an aver-age intraoperative blood loss of 1,137��436ml. Thisled the authors to conclude that even with bloodlosses of 2 L, it is not necessary to redose cefazolinany earlier than 4 h in order to maintain the MIC formost common infecting organisms.147 The authorsrepeated the study in 19 patients undergoing instru-mented posterior spinal fusion and found that therewas no significant difference between preoperativeand intraoperative cefazolin clearance and therewas no correlation between blood loss and cefazolinlevel.148



Question 16: Should preoperative antibiotic doses beweight-adjusted?

ConsensusPreoperative antibiotics have different pharmacokinet-ics based on patient weight and should be weight-adjusted.


JustificationBecause of the relative unpredictability of pharmacoki-netics in obese individuals, doses are best estimatedon the basis of specific studies for individual drugscarried out in this population. Only a few antibiotics(aminoglycosides, vancomycin, daptomycin, and line-zolid) have been studied in the obese population.

AAOS recommendation for the use of IV antibioticprophylaxis in primary TJA, recommendation 2, statesthat “timing and dosage of antibiotic administrationshould optimize the efficacy of the therapy. Doseamount should be proportional to patient weight; forpatients >80kg, the doses of cefazolin should bedoubled.”2

The recommended dose of cefazolin is based onpatient’s body mass index (BMI), with 1.0 g for peoplewho weigh <80kg and 2.0 g for those who weigh>80kg. The adult dose of cefuroxime is 1.5 g. Therecommended dose of clindamycin is 600–900mg.61

The recommended dose of vancomycin, which is basedon BMI, is 10–15mg/kg, up to a limit of 1 g, in patientswith normal renal function.149 However, there isliterature to support the use of higher doses ofvancomycin, with emphasis that doses >4 g/day havebeen associated with increased risk of nephrotoxicity.A trough level is obtained prior to the fourth scheduleddose and in certain occasions there may be a need toshorten dosing interval to maintain therapeutic troughlevel (e.g., q12h–q8h dosing).

Because 30% of adipose is water, an empiricalapproach is to use the Devine formula to calculateideal body weight (IBW), to which is added a dosingweight correction factor (DWCF) of 0.3 times thedifference between actual body weight (ABW) andIBW (IBWþ0.3� [ABW�IBW]) to arrive at a weighton which to base dosage of hydrophilic antibiotics. Nostudies confirm this approach for b-lactam drugs.Clinical studies suggest a DWCF of 0.4 for aminoglyco-sides and 0.45 for quinolones.150

For aminoglycosides, some suggest using ABWusing a dosing correction factor,151,152 while otherssuggest dosing based on lean body weight (LBW) withappropriate monitoring with the first dose.153 Currentguidelines for vancomycin administration are basedon loading doses of vancomycin on the total bodyweight (TBW) of the patient and maintenance doseson the calculated creatinine clearance (CrCl) of the

patient.152,154 However, deciding whether to base CrClcalculations on ABW, IBW, or another measure is stillto be determined. As a general rule, obese and morbid-ly obese patients require higher doses of cephalosporinto achieve similar outcomes; however, there are fewerabsolute dosing recommendations. At least one studydemonstrated that a dose of 2 g of cephazolin shouldprovide adequate levels for at least 4 h, even in supermorbid obesity (MO) (BMI� 50kg/m2).155

Other studies confirm that vancomycin should begiven on the basis of ABW, with dosage adjustmentsbased on serum concentrations156 whereas aminoglyco-side dosing requires calculation of adjusted bodyweight via a correction factor.157

Forse et al. conducted a prospective RCT in MOpatients undergoing gastroplasty and found that theblood and tissue levels of cefazolin were significantlylower for all MO patients who received 1 g cefazolincompared with the blood and tissue levels of the drugfound in normal weight patients who received asimilar dose of antibiotic. Moreover, the MO patientswho only received 1 g of cefazolin had antibiotic levelsbelow the MIC of 2mcg/ml for gram-positive cocci and4mcg/ml for gram-negative rods. The serum and tissueconcentrations were adequate only when 2 g of cefazo-lin were administered. Also, relative to 1 g, the admin-istration of cefazolin 2 g decreased the wound infectionrate from 16.5% to 5.6% in these MO patients.18

Van Kralingen et al. studied the influence of bodyweight measures and age on pharmacokinetic param-eters and evaluated unbound cefazolin concentrationsover time in obese patients. Twenty MO patients (BMI38–79kg/m2) were studied following the administra-tion of 2 g of cefazolin at induction of anesthesia. Bloodsamples were collected up to 4h post dosing todetermine the total and unbound plasma cefazolinconcentrations. Cefazolin clearance was 4.2� 1.0L/h(mean� standard deviation) and showed a negativecorrelation with age (p¼0.003) but not with bodyweight measures (p> 0.05). In all patients, unboundcefazolin concentrations remained above 1mg/L (MIC90) of MSSA until 4 h post dosing.158

Ho et al. attempted to determine an optimal dosingregimen for cefazolin as a prophylactic antibiotic insurgery for patients with MO. Twenty-five patientsundergoing elective surgical procedures were given asingle dose of cefazolin: 10 with MO (BMI 40–50kg/m2) received 2 g via intravenous push (IVP), 5 withMO received 2 g via 30-min infusion, 5 with supermorbid obesity (SMO, BMI >50 kg/m2) received 2 g viainfusion, and 5 with SMO received 3 g via infusion.The protective duration, determined using a pharma-codynamic target for fT>MIC of 70%, was 5.1 h forMO2-IVP, 4.8 h for MO2-INF, 5.8 h for SMO2-INF,and 6.8 h for SMO3-INF. The authors concluded thata single 2 g dose of cefazolin appears to provideantibiotic exposure sufficient for most common gener-al surgical procedures of <5h duration regardless ofBMI.155



In contrast, Edmiston et al. concluded that 2 g ofcefazolin may not be sufficient for patients with a BMI>50kg/m2, based upon measurements of total serumconcentrations in morbidly obese patients undergoinggastric bypass. The authors assigned 38 patients toone of three BMI groups: (A) BMI¼ 40–49kg/m2

(n¼17), (B) BMI¼50–59kg/m2 (n¼ 11), and (C) BMI>¼ 60kg/m2 (n¼ 10) and measured serum and tissueconcentrations of cefazolin. They determined thattherapeutic tissue levels were only achieved in 48.1%,28.6%, and 10.2% in groups A, B, and C, respectively.The authors measured concentrations in the serumskin, adipose tissue, and omentum, but did not evalu-ate unbound cefazolin concentrations, which may beexpected to migrate across tissues rapidly.159

Question 17A: What type of perioperative antibioticprophylaxis is recommended for current MRSAcarriers?

ConsensusFor current MRSA carriers, vancomycin or teicoplaninis the recommended perioperative antibiotic prophy-laxis.


Question 17B: Should patients with prior history ofMRSA be re-screened? What should the choice ofperioperative prophylactic antibiotics be in thesepatients?

ConsensusPatients with prior history of MRSA should be re-screened preoperatively. If patients are found to benegative for MRSA, we recommend routine periopera-tive antibiotic prophylaxis.


JustificationImplementation of a MRSA prevention program maysignificantly reduce MRSA SSIs. However, it is unlike-ly that any single MRSA-specific intervention (such asadding or switching to vancomycin) can optimallyprevent SSIs. Several studies provide convincing dataon the clinical effectiveness of vancomycin in prevent-ing SSIs when MRSA prevalence is high.69,70,79 Fur-ther research is needed to determine whichcomponents of a MRSA prevention program are essen-tial in successfully preventing MRSA SSIs.160 It isuncertain whether decontamination should alter thetype of antibiotic prophylaxis, as few studies haveretested patients’ MRSA status immediately prior tosurgery.

The AAOS recommendations for the use of IVantibiotic prophylaxis in primary TJA, recommenda-tion 2, states that “vancomycin may be used inpatients with known colonization with MRSA or infacilities with recent MRSA outbreaks.”1 Additionally,the Society for Healthcare Epidemiology of Americarecently recommended routine surveillance cultures atthe time of admission to the hospital for patients athigh risk of MRSA.52

Walsh et al. implemented a comprehensive MRSAprogram in which vancomycin was added to theroutine cefazolin prophylaxis regimen for patients whotested positive for nasal MRSA carriage. Other compo-nents of the program included decolonization of allcardiothoracic staff who screened positive for nasalMRSA carriage, application of nasal mupirocin oint-ment for 5 days in all patients starting one day beforesurgery, application of topical mupirocin to exit sitesafter removal of chest and mediastinal tubes, and

Table 1. Recommended Dosing of Preoperative Antibiotics by Weight

AntimicrobialActual Body

Weight (ABW; kg)Recommended

Dose (mg)Perioperative

Redosing Schedule Indication

Cefazolin <60 1,000 Every 4h Primary perioperativeprophylaxis60–120 2,000 Every 4h

>120 3,000 Every 4hCefuroxime No adjustments 1,500 Every 4h Primary perioperative

prophylaxisVancomycin Weight based dosing

recommended15mg/kg(Maximum dose2,000mg)

One dose pre-op,one dose12h post-op, onedose 24h post-op

Perioperative prophylaxisfor current MRSAcarriers and/or patientswith b-lactam allergy

Clindamycin No adjustments 900 Every 3h Perioperative prophylaxisfor patients withb-lactam allergy

Teicoplanin No adjustments 400 NA Perioperative prophylaxisfor current MRSAcarriers and/or patientswith b-lactam allergy



rescreening of patients for MRSA colonization at thetime of hospital discharge. This program resulted in asignificant reduction in the SSI rate (2.1% vs. 0.8%,p< 0.001) as well as a 93% reduction in postoperativeMRSA wound infections (from 32 infections/2,767procedures during the 3 year pre-intervention periodto 2 infections/2,496 procedures during the 3-yearpostintervention period). The data suggest that abundled approach to preventing MRSA SSIs may bemore critical than a single intervention.79

Pofahl et al. published on the impact of introducingMRSA screening programs and treatment of subse-quent MRSA SSIs. After a MRSA surveillance pro-gram was instituted, the rate of MRSA SSI decreasedfrom 0.23% to 0.09%, with the most pronouncedreduction seen in TJA procedures (0.30–0%, p¼ 0.04).However, the authors note that changes in periopera-tive antibiotics in MRSA-positive patients was at thediscretion of the attending surgeon.161

Question 18: What is the recommended prophylaxisin patients undergoing major orthopaedicreconstructions for either tumor or non-neoplasticconditions using megaprosthesis?

ConsensusUntil the emergence of further evidence, we recom-mend the use of routine antibiotic prophylaxis forpatients undergoing major reconstruction.


JustificationDeep infection has been reported as being one of themost common complications following endoprostheticreplacement of large bone defects, ranging between 5%and 35% in some series.162–166 Reinfection rates afterrevision surgery for endoprosthetic infection have beenreported as high as 43%.165 Despite this there isinsufficient evidence to suggest that a different periop-erative antibiotic regimen is warranted. Recently amulticenter, blinded, randomized, controlled trial,using a parallel two-arm design has been set up thatwill evaluate 920 patients from Canada and the USAwho are undergoing surgical excision and endopros-thetic reconstruction of a primary bone tumour. Thepatients will receive either short (24h) or long (5 days)duration postoperative antibiotics. The primary out-come will be rates of deep postoperative infections ineach arm. Secondary outcomes will include type andfrequency of antibiotic-related adverse events, patientfunctional outcomes and quality-of-life scores, reopera-tion and mortality.167

Another area of development involves silver coatingof foreign materials, such as heart valves, cardiaccatheters, and urinary catheters, that has shown theability to reduce the infection rate of medical devices;

therefore, a logical extension of this work was totranslate this concept to the field of endopros-thetics.168,169 Both basic science and clinical researchsuggests a decreased incidence of SSI and PJI inendoprostheses coated with silver. Recently iodine-supported titanium implants have been also effectivefor preventing and treating infections after majororthopaedic surgery.170,171

In a rabbit study, the infection rate of silver-coatedversus noncoated prostheses after inoculation withStaphylococcus aureus was determined and the silverconcentrations in blood, urine, and organs with possi-ble toxic side effects were documented. The authorsconvincingly demonstrated that megaprostheses coat-ed with silver showed a significantly lower infectionrate (7% vs. 47%, p<0.05) in comparison with atitanium group.172 Furthermore, measurements of C-reactive protein, neutrophilic leukocytes, rectal tem-perature, and body weight showed significantly lower(p< 0.05) signs of inflammation in the silver group. Ina second study, authors analyzed the potential toxico-logical side effects of these implants and found thatthe silver concentration in blood (median 1.883 partsper billion [PPB]) and in organs (0.798–86.002PPB)showed elevated silver concentrations, without patho-logic changes in laboratory parameters and withouthistologic changes of organs.173

In a prospective observational study, Hardes et al.compared the infection rate in 51 patients with sarco-ma (proximal femur, n¼ 22; proximal tibia, n¼ 29)who underwent placement of a silver-coated megapros-thesis to 74 patients (proximal femur, n¼33; proximaltibia, n¼41) in whom an uncoated titanium megapros-theses was used. The authors reported a substantialreduction in the infection rate from 17.6% in thetitanium group compared to 5.9% in the silver group(p¼ 0.06). Furthermore, while 38.5% of patients ulti-mately underwent amputation when PJI developed,this was not necessary in any case in the study group.However, the authors note that the operating timerequired for the proximal tibia replacement wassignificantly shorter in the silver-coated prosthesisgroup (p¼ 0.034) and that prolonged operating timewas associated with a higher rate of PJI (p¼0.025).

The same group reported a lack of toxicological sideeffects of silver-coated megaprostheses in 20 patientswith bone metastases.172 They reported that silverlevels in the blood did not exceed 56.4PPB and can beconsidered non-toxic. They further excluded significantchanges in liver and kidney function based on labora-tory values; and histopathologic examination of theperiprosthetic environment in two patients showed nosigns of foreign body granulomas or chronic inflamma-tion, despite effective silver concentrations up to1,626PPB directly related to the prosthetic surface.172

Tsuchiya et al. reported that iodine-supportedimplants were used to prevent infection in 257patients with compromised status. Acute infectiondeveloped only in three tumor cases and one diabetic



foot among the 257 patients. Abnormalities of thyroidgland function were not detected. None of the patientsexperienced loosening of the implant. Excellentbone ingrowth was found around all hip and tumorprostheses. The results indicate that iodine-supportedtitanium has favorable antibacterial activity, biocom-patibility, and no cytotoxicity.170

Gosheger reviewed 197 patients with megaprosthe-ses and discovered that those with cobalt chromeimplants had more infections than those with titaniumimplants.174 Reviewing 197 patients (77 patients witha cobalt chrome alloy system and 120 patients witha titanium alloy system) who underwent lower extrem-ity reconstruction with a megaprosthesis, the authorsreported a 31.2% infection rate in the cobalt chromegroup compared to 14.2% in the titanium group(p<0.01). When they performed a secondary analysismatching two identical subgroups, the cobalt chromegroup was still associated with a significantly higherinfection rate, with five infections of 26 megaprosthe-ses versus one infection of 36 titanium megaprostheses(p<0.05).175

Question 19: Should antibiotic prophylaxis bedifferent in patients who have reconstruction by bulkallograft?

ConsensusWe recommend the use of routine antibiotic prophylax-is in patients who have reconstruction by bulk allo-graft.


JustificationThe periprosthetic area is inherently a locus minorisresistance. Bulk allograft is in essence is a largeforeign body and therefore represents a nidus for deepinfection following surgery, apart from the prostheticcomponents. Additionally, bulk allografts are usedmost often in the setting of revision arthroplasty whenthere is frequently additional local soft tissue andvascular compromise, which compounds the risk forinfection. Therefore, it would seem reasonable to wantto modify the perioperative antibiotic protocol toprotect these reconstructions. Unfortunately, there isinsufficient literature to support altering antibioticregimens, as most studies on the use of bulk allograftdo not indicate or detail the antibiotic regimensutilized. Even if this data were available, it would notbe accurate to properly compare the infection rates ofdifferent clinical series based on their perioperativeantibiotic protocols because of the heterogeneity ofpatient populations. However, there is a growing bodyof literature to support the use of antibiotic-impregnat-ed allograft in the revision setting as a means ofdecreasing infection rates. In addition, there areseveral reports of using antibiotic-impregnated graft

substitute or grafts as a way to fill bony defects andpromote bony ingrowth while delivering suprathera-peutic doses of antibiotics to the local environment incases of osteomyelitis. While there is no currentliterature applying this technology to the use of bonedefects in infected revision arthroplasty, it may be apromising technique.

Witso et al. used netilmicin-impregnated allograftsfor reconstruction in revision hip and knee surgeryand found no adverse effects.176 Buttaro et al.174,177

favorably used vancomycin-supplemented cancellousgrafts for reconstruction after infected THA. Michalaket al.178 and Khoo et al.179 impregnated segmentalallografts with gentamicin and flucloxacillin, respec-tively. However, all these groups used antibioticimpregnated grafts only in the second stage of a two-stage revision, after resolution of clinical and laborato-ry evidence of infection.

Winkler et al. performed 37 one-stage uncementedrevision THAs using cancellous allograft bone impreg-nated with antibiotics and noted a 92% success rate,defined as recurrent infection at a mean follow-up of4.4 years (range 2–8 years). In addition, no adverseeffects were seen and the incorporation of bone graftwas comparable to unimpregnated grafts.180

In a similar series, Buttaro analyzed the incidenceof infection after one-stage aseptic revision hip recon-struction using acetabular and/or femoral vancomy-cin-impregnated impacted bone allograft and a THAfixed with cement containing no antibiotic. In 75consecutive patients (80 hips), followed for a mean of36 months (range 24–59 months), deep infectionoccurred in one patient for an incidence of infectionof 1.25%, which occurred 2 years after the indexprocedure and was thought to be hematogenous inorigin.181

Cancellous bone allograft can store and release highinitial local amounts of vancomycin withoutcompromising incorporation of the graft, and somefavorable results have been published following two-stage revision of infected THA with this tech-nique.176,177,182–184

Question 20: Do patients with poorly controlleddiabetes, immunosuppression, or autoimmunedisease require a different perioperative antibioticprophylaxis?

ConsensusNo. Routine antibiotic prophylaxis is recommended inthese patients.


JustificationSeveral studies have demonstrated that diabetes mel-litus (DM), especially uncontrolled DM, is a risk factorfor postoperative infection in THA and TKA.185–188 A



recent retrospective cohort study within the KaiserHealthcare system found no significant increase inrisk of revision or deep infection or revision whetherpatients had controlled (HbA1c<7%) or uncontrolleddiabetes (HbA1c>7%). Specifically, compared withpatients without DM, there was no association be-tween controlled DM and risk of revision (OR 1.32;95% CI 0.99–1.76). Similarly, compared to patientswithout DM, there was no association between uncon-trolled DM and risk of revision (OR 1.03; 95% CI 0.68–1.54).189

Obesity has also been associated with a significantincrease in rate of postoperative infection followingTJA.190–192

Human immunodeficiency virus (HIV) has also beenassociated with an alarming rate of postoperativecomplications, including infection. Parvizi et al.reported on six deep infections in 21 HIV-positivepatients undergoing TJA. The authors remarked thatthe immune status of the patients was related to theirrisk of deep PJI, in that five of the six patientsultimately developed Acquired Immune DeficiencySyndrome (AIDS) and the CD4 count was significantlylower at 239�112ml at latest follow-up for patientswho developed infection compared to 523� 171ml forthe study population as a whole (p< 0.001). In thisstudy, the authors reported using prophylactic anti-biotics (cephalosporins) preoperatively and three dosespostoperatively and added antibiotic powder (vancomy-cin and tobramycin) to the cement in two patientsthought to be at high risk for infection.193

Similarly, Ragni et al. found a very high postopera-tive infection rate (26.5%) in 34 TJA in HIV-positivehemophiliacs, all of whom had CD4 counts less than200/ml at time of surgery.194 Haberman et al. noted aninfection rate of 12.7% in their cohort of 41 patientswith HIV undergoing TJA, but did not identify anydifference in the outcomes relating to CD4 count.195

Their perioperative antibiotic protocol was a 5 daycourse of cefuroxime and in all procedures antibiotic-containing cement (Palacos R, Zimmer, Warsaw, IN)was used. In a smaller series of six HIV-infectedpatients undergoing TJA, Wang et al. noted no infec-tious or other complications. The authors again usedantibiotic (vancomycin)-impregnated bone cement inall cemented cases.196 Unger et al.197 evaluated theresults of 26 TKAs in HIV-positive hemophiliacs andfound no cases of deep infection, but it is interesting tonote that the average CD4 count of these patients was463ml.

Hemophilia has historically been considered a riskfactor for PJI, due in part to its relation to HIV andAIDS, but also as an independent risk factor. Anarticle by Silva et al. reviewed the long-term results ofprimary TKA in patients with hemophilia and notedan overall prevalence of PJI of 16% with a rate ofinfection in HIV-positive and HIV-negative patients of17% and 13%, respectively (p¼0.5). The authors’perioperative protocol included 3–5 days of prophylac-

tic antibiotics and antibiotic cement was not used.198

In contrast, Rodriguez-Marchan199 reported an infec-tion rate of only 3% of 35 TJA in hemophiliac patients,but used antibiotic-laden bone cement and 2 days ofperioperative antibiotic prophylaxis.

Asplenic patients are at increased risk of infectionby encapsulated bacteria; and although there is evi-dence to support vaccinations and penicillin prophy-laxis in patients under 16 and over 50 years of age,there is no consensus on the appropriate perioperativemanagement of these immunocompromised patients.In a single case report by Shaarani et al.200 of anasplenic patient who underwent a TKA, the patientultimately developed a MRSA infection. In this casestandard polymethylmethacrylate (PMMA) was usedfor cementing components and the patient receivedintravenous prophylactic dose of second generationcephalosporin preoperatively.

Renal disease (including renal failure, dialysisdependence, and renal transplant) has been implicatedas increasing the risk of PJI. McCleery et al. analyzedthe Scottish Arthroplasty Registry in order to deter-mine the rates of PJI in patients with renal failure,those undergoing dialysis, and those with a renaltransplant. They found that patients with renal failurehad a significantly increased risk of early infection(1.6%, RR 1.52, p¼ 0.02) and late infection (4.47%, RR2.2, p< 0.001). Patients on dialysis had a significantlyincreased risk of late infection (8.0%, RR 3.99,p< 0.001) and early revision (3.7%, RR 4.4, p< 0.001).Renal transplant patients had a significantly increasedrisk of late infection, despite whether the transplanta-tion occurred before TKA (9.1%, RR 4.5, p¼0.03) or atany time (8.0%, RR 4.0, p¼ 0.05).201 Liebermanet al.202 documented a deep infection rate of 19% in 16chronic renal dialysis patients and more favorableoutcomes in renal transplant patients. Sakalkaleet al.203 reported a deep infection rate of 13% in 12patients with end-stage renal failure on dialysis whounderwent THA. In this study, perioperative prophy-lactic antibiotics were administered for 2–5 days. Incontrast, other authors have reported no increasedrate of infection in patients on chronic hemodialysisundergoing THA.204,205

Similarly, liver disease has been associated withincreased morbidity following TJA. Pour et al. per-formed a case control study of 71 non-cirrhotic patientswith hepatitis C undergoing TJA and found that thiscohort had higher rates of wound drainage followingTHA when compared to matched controls (15% vs.3.8%, p¼0.03).206 Orozco et al. recently published acase control study to analyze the effect of fibrosis andthrombocytopenia on the diagnosis of hepatitis C andclinical outcomes. Analyzing 72 patients (77 jointreplacements), the authors found that fibrotic hepatitisC patients had higher deep infection rates (21% vs.0%, p¼0.047) and rates of cellulitis (21% vs 0%,p¼ 0.047), while thromobocytopenia showed a trendtowards greater infection.207



Solid organ transplant (SOT) is a risk factor for PJIdue to the need for chronic use of immunosuppressantmedications. Vergidis et al. performed a case controlstudy of patients with SOT who developed PJI andcompared them to non-infected controls matched bytransplant type, prosthetic joint type, and order oforgan transplantation or joint implantation. Of 367patients with both a joint replacement and SOT, therewere 12 cases of PJI, of which 8 were renal trans-plants, three were liver transplants, and 1 was a hearttransplant patient. Eight infections were caused bygram-positive organisms, 2 were caused by nontuber-culous mycobacteria, and the remaining 2 were cul-ture-negative. Of note, patients received perioperativecefazolin, or in cases of colonization or prior infectionwith MRSA, vancomycin.208 Tannenbaum et al.reported results on 35 TJA in 19 patients with renal orliver transplant and documented an infection in 5patients who had the joint replacement after thetransplantation. There were no infections in patientswho had TJA before the organ transplantation. In thisseries, prophylactic antibiotics were administered forat least 48h or until the drains were removed andbone cement when used was not impregnated withantibiotics.209

Question 21A: Should preoperative antibiotics bedifferent for primary and revision TJA?

ConsensusNo. Perioperative antibiotic prophylaxis should be thesame for primary and uninfected revision arthro-plasty.


Question 21B: Should preoperative antibiotics bedifferent for hips and knees?

ConsensusPerioperative antibiotic prophylaxis should be thesame for hips and knees.


JustificationPatients undergoing revision TJA are at higher risk ofdeveloping PJI than primary arthroplasty and thoseundergoing revision knee procedures are at even high-est risk.210–212 One recent study has effectively demon-strated targeting infection prevention programs athigh-risk surgical patients that take into account aninstitution’s local epidemiology and antibiogram.213

Liu et al. determined the impact of adding vanco-mycin to cefazolin as antimicrobial prophylaxis in 414

patients undergoing revision TKA based on a notableincrease in PJI in revision TKA patients, with manybeing methicillin-resistant. Following introduction ofvancomycin to the routine preoperative antibioticprophylaxis, the infection rate decreased from 7.89%to 3.13% (p¼0.046). In particular, a significant reduc-tion in PJI resulting from methicillin-resistant organ-isms over this time period was seen (4.2–0.9%,p¼ 0.049).214

Question 22: What is the best antibiotic prophylaxisto choose in patients with colonization bycarbapenem resistant enterobacteriaceae ormulti-drug resistant (MDR)-Acinetobacter spp?

ConsensusThere is insufficient data to recommend expandedantibiotic prophylaxis in patients known to be colo-nized or recently infected with MDR pathogens.


JustificationThere is an increasing awareness of the threat posedby K. pneumoniae strains with decreased susceptibilityto carbapenems worldwide.215 This resistance is con-ferred by K. pneumo carbapemenase (KPC), which is ab-lactamase that also confers resistance to broad-spectrum cephalosporins, as well as commerciallyavailable b–lactam/b-lactamase inhibitor combina-tions.216 As there are few antimicrobial options, pre-vention of K. pneumo carbapemenase K. pneumoniae(KPC-KP) has become a major priority of thosestudying nosocomial infections.217

While there is no evidence on the management ofsurgical antimicrobial prophylaxis in a patient withpast infection or colonization with a resistant gram-negative pathogen, it is logical to provide prophylaxiswith an agent active against MRSA for any patientknown to be colonized with this gram-positive patho-gen who will have a skin incision; specifically, prophy-laxis for a resistant gram-negative pathogen in apatient with past infection or colonization with such apathogen may not be necessary for a purely cutaneousprocedure.

In a literature review, KPC-producing microbesare resistant to many non-b-lactam molecules. Mostisolates are resistant to fluoroquinolones, aminoglyco-sides, and co-trimoxazole. Some isolates are suscepti-ble to amikacin and gentamicin and most aresusceptible to colistin and tigecycline.214,215,218,219

In a prospective RCT, De Smet et al. studied theelimination of colonization with MDR organisms usingselective oropharynegeal and/or digestive tract decon-tamination (SOD/SDD) in a multicenter crossoverstudy using cluster randomization of 5,939 intensivecare unit patients in the Netherlands. SOD included



4 days of intravenous cefotaxime and topical applica-tion of tobramycin, colistin, and amphotericin B in theoropharynx and stomach. SDD consisted of oropharyn-geal application only of the same antimicrobials. Usinga random effects logistic regression analysis, the ORfor death at Day 28 in the SOD and SDD group, ascompared with the standard care group, were 0.86(95% CI 0.74–0.99) and 0.83 (95% CI 0.72–0.97),respectively.220

Perez et al. used a mouse model to examine theeffect of antibiotic treatment on the establishment andelimination of intestinal colonization of KPC-KP. Theyadministered 3 days of antibiotics (clindamycin, zosyn,tigecycline, ertapenem, cefepime, and ciprofloxacin)before KPC-KP was administered orogastrically. Theauthors reported that of the four antibiotics withminimal activity against the KPC-KP strain (MIC>16mcg/ml), those that suppressed total anaerobes andBacteroides (i.e., clindamycin and zosyn) promotedcolonization by KPC-KP (p<0.001), while agents thatdid not suppress total anaerobes and bacteroides (i.e.,ciprofloxacin and cefepime) did not (p¼0.35). Of theantibiotics with moderate activity against KPC-KP,ertapenem (MIC 4mcg/ml) did not promote coloniza-tion by KPC-KP, while tigecycline (MIC 3mcg/ml) did(p< 0.001), despite not reducing levels of total anae-robes and bacteroides. Orgogastric administration ofgentamicin and polmyxin E-suppressed KPC-KP wasat undetectable levels in the majority of mice. Theauthors posited that antibiotics that disturb the intes-tinal anaerobic microflora lack significant activityagainst KPC-KP promote colonization, while the ad-ministration of non-absorbed oral antibiotics may bean effective strategy to suppress colonization with thismicroorganism.221

REFERENCES1. Recommendations for the use of intravenous antibiotic

prophylaxis in primary total joint arthroplasty. http://www.aaos.org/about/papers/advistmt/1027.asp. Accessed2013.

2. Prokuski L. 2008. Prophylactic antibiotics in orthopaedicsurgery. J Am Acad Orthop Surg 16:283–293.

3. Bratzler DW, Houck PM. 2004. Antimicrobial prophylaxisfor surgery: an advisory statement from the NationalSurgical Infection Prevention Project. Clin Infect Dis38:1706–1715.

4. van Kasteren ME, Mannien J, Ott A, et al. 2007. Antibioticprophylaxis and the risk of surgical site infections followingtotal hip arthroplasty: timely administration is the mostimportant factor. Clin Infect Dis 44:921–927.

5. Classen DC, Evans RS, Pestotnik SL, et al. 1992. Thetiming of prophylactic administration of antibiotics and therisk of surgical-wound infection. N Engl J Med 326:281–286.

6. Bratzler DW, Houck PM. 2005. Antimicrobial prophylaxisfor surgery: an advisory statement from the NationalSurgical Infection Prevention Project. Am J Surg 189:395–404.

7. Galandiuk S, Polk HC Jr, Jagelman DG, et al. 1989. Re-emphasis of priorities in surgical antibiotic prophylaxis.Surg Gynecol Obstet 169:219–222.

8. Hawn MT, Richman JS, Vick CC, et al. 2013. Timing ofsurgical antibiotic prophylaxis and the risk of surgical siteinfection. JAMA Surg 148:649–657.

9. Weber WP, Marti WR, Zwahlen M, et al. 2008. The timingof surgical antimicrobial prophylaxis. Ann Surg 247:918–926.

10. Steinberg JP, Braun BI, Hellinger WC, et al. 2009. Timingof antimicrobial prophylaxis and the risk of surgical siteinfections: results from the Trial to Reduce AntimicrobialProphylaxis Errors. Ann Surg 250:10–16.

11. Johnson DP. 1987. Antibiotic prophylaxis with cefuroximein arthroplasty of the knee. J Bone Joint Surg Br 69:787–789.

12. Friedman RJ, Friedrich LV, White RL, et al. 1990.Antibiotic prophylaxis and tourniquet inflation in totalknee arthroplasty. Clin Orthop Relat Res 260:17–23.

13. Soriano A, Bori G, Garcia-Ramiro S, et al. 2008. Timing ofantibiotic prophylaxis for primary total knee arthroplastyperformed during ischemia. Clin Infect Dis 46:1009–1014.

14. A WD, Robertsson O, Stefansdottir A, et al. Timing ofpreoperative antibiotics for knee arthroplasties: improvingthe routines in Sweden. Patient Saf Surg 5:22.

15. Neu HC. 1984. Cephalosporin antibiotics as applied insurgery of bones and joints. Clin Orthop Relat Res 190:50–64.

16. Oishi CS, Carrion WV, Hoaglund FT. 1993. Use of paren-teral prophylactic antibiotics in clean orthopaedic surgery.A review of the literature. Clin Orthop Relat Res 296:249–255.

17. Schurman DJ, Hirshman HP, Kajiyama G, et al. 1978.Cefazolin concentrations in bone and synovial fluid. J BoneJoint Surg Am 60:359–362.

18. Forse RA, Karam B, MacLean LD, et al. 1989. Antibioticprophylaxis for surgery in morbidly obese patients. Surgery106:750–756, discussion 756–757.

19. Sodhi M, Axtell SS, Callahan J, et al. 2004. Is it safe to usecarbapenems in patients with a history of allergy topenicillin? J Antimicrob Chemother 54:1155–1157.

20. Hill C, Flamant R, Mazas F, et al. 1981. Prophylacticcefazolin versus placebo in total hip replacement. Report ofa multicentre double-blind randomised trial. Lancet 1:795–796.

21. Tyllianakis ME, Karageorgos A, Marangos MN, et al. 2010.Antibiotic prophylaxis in primary hip and knee arthro-plasty: comparison between cefuroxime and two specificantistaphylococcal agents. J Arthroplasty 25:1078–1082.

22. Mauerhan DR, Nelson CL, Smith DL, et al. 1994. Prophy-laxis against infection in total joint arthroplasty. One dayof cefuroxime compared with three days of cefazolin. JBone Joint Surg Am 76:39–45.

23. Stefansdottir A, Johansson D, Knutson K, et al. 2009.Microbiology of the infected knee arthroplasty: report fromthe Swedish Knee Arthroplasty Register on 426 surgicallyrevised cases. Scand J Infect Dis 41:831–840.

24. Enzler MJ, Berbari E, Osmon DR. 2011. Antimicrobialprophylaxis in adults. Mayo Clin Proc 86:686–701.

25. Wilson W, Taubert KA, Gewitz M, et al. 2007. Preventionof infective endocarditis: guidelines from the AmericanHeart Association: a guideline from the American HeartAssociation Rheumatic Fever, Endocarditis, and KawasakiDisease Committee, Council on Cardiovascular Disease inthe Young, and the Council on Clinical Cardiology, Councilon Cardiovascular Surgery and Anesthesia, and the Quali-ty of Care and Outcomes Research Interdisciplinary Work-ing Group. Circulation 116:1736–1754.

26. Eagle KA, Guyton RA, Davidoff R, et al. 2004. ACC/AHA2004 guideline update for coronary artery bypassgraft surgery: a report of the American College of Cardiolo-



http://www.aaos.org/about/papers/advistmt/1027.asp


gy/American Heart Association Task Force on PracticeGuidelines (Committee to Update the 1999 Guidelines forCoronary Artery Bypass Graft Surgery). Circulation 110:e340–437.

27. Edwards FH, Engelman RM, Houck P, et al. 2006. TheSociety of Thoracic Surgeons Practice Guideline Series:antibiotic prophylaxis in cardiac surgery, part I: duration.Ann Thorac Surg 81:397–404.

28. Haydon TP, Presneill JJ, Robertson MS. 2010. Antibioticprophylaxis for cardiac surgery in Australia. Med J Aust192:141–143.

29. Mollan RA, Haddock M, Webb CH. 1992. Teicoplanin vscephamandole for antimicrobial prophylaxis in prostheticjoint implant surgery: (preliminary results). Eur J SurgSuppl 567:19–21.

30. Lazzarini L, Pellizzer G, Stecca C, et al. 2001. Postopera-tive infections following total knee replacement: an epide-miological study. J Chemother 13:182–187.

31. Periti P, Pannuti F, Della Cuna GR, et al. 1991. Combina-tion chemotherapy with cyclophosphamide, fluorouracil,and either epirubicin or mitoxantrone: a comparativerandomized multicenter study in metastatic breast carcino-ma. Cancer Invest 9:249–255.

32. Periti P, Stringa G, Mini E. 1999. Comparative multi-center trial of teicoplanin versus cefazolin for anti-microbial prophylaxis in prosthetic joint implant surgery.Italian Study Group for Antimicrobial Prophylaxis inOrthopedic Surgery. Eur J Clin Microbiol Infect Dis18:113–119.

33. Brogden RN, Peters DH. 1994. Teicoplanin. A reappraisalof its antimicrobial activity, pharmacokinetic propertiesand therapeutic efficacy. Drugs. 47:823–854.

34. Wood MJ. 1996. The comparative efficacy and safety ofteicoplanin and vancomycin. J Antimicrob Chemother37:209–222.

35. Wilson AP. 1988. Antibiotic prophylaxis in cardiac surgery.J Antimicrob Chemother 21:522–524.

36. Darley ES, MacGowan AP. 2004. Antibiotic treatment ofgram-positive bone and joint infections. J AntimicrobChemother 53:928–935.

37. Dash CH. 1975. Penicillin allergy and the cephalosporins. JAntimicrob Chemother 1:107–118.

38. Petz LD. 1978. Immunologic cross-reactivity between pen-icillins and cephalosporins: a review. J Infect Dis 137(Suppl):S74–S79.

39. Kelkar PS, Li JT. 2001. Cephalosporin allergy. N Engl JMed 345:804–809.

40. Saxon A, Beall GN, Rohr AS, et al. 1987. Immediatehypersensitivity reactions to beta-lactam antibiotics. AnnIntern Med 107:204–215.

41. Pichichero ME. 2007. Use of selected cephalosporins inpenicillin-allergic patients: a paradigm shift. Diagn Micro-biol Infect Dis 57(3 Suppl):13S–18S.

42. Pichichero ME, Casey JR. 2007. Safe use of selectedcephalosporins in penicillin-allergic patients: a meta-analy-sis. Otolaryngol Head Neck Surg 136:340–347.

43. Audicana M, Bernaola G, Urrutia I, et al. 1994. Allergicreactions to betalactams: studies in a group of patientsallergic to penicillin and evaluation of cross-reactivity withcephalosporin. Allergy 49:108–113.

44. Solensky R, Earl HS, Gruchalla RS. 2002. Lack of penicillinresensitization in patients with a history of penicillinallergy after receiving repeated penicillin courses. ArchIntern Med 162:822–826.

45. Campagna JD, Bond MC, Schabelman E, et al. 2012. Theuse of cephalosporins in penicillin-allergic patients: aliterature review. J Emerg Med 42:612–620.

46. DePestel DD, Benninger MS, Danziger L, et al. 2008.Cephalosporin use in treatment of patients with penicillinallergies. J Am Pharm Assoc (2003) 48:530–540.

47. Platt R. 1982. Adverse effects of third-generation cephalo-sporins. J Antimicrob Chemother 10:135–140.

48. Goodman EJ, Morgan MJ, Johnson PA, et al. 2001.Cephalosporins can be given to penicillin-allergic patientswho do not exhibit an anaphylactic response. J Clin Anesth13:561–564.

49. Apter AJ, Kinman JL, Bilker WB, et al. 2006. Is therecross-reactivity between penicillins and cephalosporins?Am J Med 119:354.e1–359.e1.

50. Park MA, Koch CA, Klemawesch P, et al. 2010. Increasedadverse drug reactions to cephalosporins in penicillinallergy patients with positive penicillin skin test. Int ArchAllergy Immunol 153:268–273.

51. Advisory statement. Recommendations for the use ofintravenous antibiotic prophylaxis in primary total jointarthroplasty. http://www.aaos.org/about/papers/advistmt/1027.asp. Accessed 2013.

52. Muto CA, Jernigan JA, Ostrowsky BE, et al. 2003. SHEAguideline for preventing nosocomial transmission of multi-drug-resistant strains of Staphylococcus aureus and entero-coccus. Infect Control Hosp Epidemiol 24:362–386.

53. Cantoni L, Glauser MP, Bille J. 1990. Comparative efficacyof daptomycin, vancomycin, and cloxacillin for the treat-ment of Staphylococcus aureus endocarditis in rats androle of test conditions in this determination. AntimicrobAgents Chemother 34:2348–2353.

54. Patel M, Kumar RA, Stamm AM, et al. 2007. USA300genotype community-associated methicillin-resistantStaphylococcus aureus as a cause of surgical site infections.J Clin Microbiol 45:3431–3433.

55. Manian FA, Griesnauer S. 2008. Community-associatedmethicillin-resistant Staphylococcus aureus (MRSA) isreplacing traditional health care-associated MRSA strainsin surgical-site infections among inpatients. Clin Infect Dis47:434–435.

56. Recommendations for preventing the spread of vancomycinresistance. 1995. Recommendations of the Hospital Infec-tion Control Practices Advisory Committee (HICPAC).MMWR Recomm Rep 44:1–13.

57. Hiramatsu K, Aritaka N, Hanaki H, et al. 1997. Dissemina-tion in Japanese hospitals of strains of Staphylococcusaureus heterogeneously resistant to vancomycin. Lancet350:1670–1673.

58. Michel M, Gutmann L. 1997. Methicillin-resistant Staphy-lococcus aureus and vancomycin-resistant enterococci:therapeutic realities and possibilities. Lancet 349:1901–1906.

59. Fulkerson E, Valle CJ, Wise B, et al. 2006. Antibioticsusceptibility of bacteria infecting total joint arthroplastysites. J Bone Joint Surg Am 88:1231–1237.

60. UK Health protection agency. Surgical Site Infectionsurveillance service. Protocol for surveillance of surgicalsite infection. http://www.hpa.org/uk/we/HPAwebFile/HPA-web_C/1194947388966. Accessed 2013.

61. Meehan J, Jamali AA, Nguyen H. 2009. Prophylactic anti-biotics in hip and knee arthroplasty. J Bone Joint Surg Am91:2480–2490.

62. Wiesel BB, Esterhai JL. 2003. Prophylaxis of musculoskele-tal infections. In: Calhoun JH, Mader J, editors. Muscu-losketal Infections. New York: Marcel Dekker; 115–129.

63. Merrer J, Desbouchages L, Serazin V, et al. 2006. Compari-son of routine prophylaxis with vancomycin or cefazolin forfemoral neck fracture surgery: microbiological and clinicaloutcomes. Infect Control Hosp Epidemiol 27:1366–1371.





http://www.hpa.org/uk/we/HPAwebFile/HPAweb_C/1194947388966

http://www.hpa.org/uk/we/HPAwebFile/HPAweb_C/1194947388966

64. Cranny G, Elliott R, Weatherly H, et al. 2008. A systematicreview and economic model of switching from non-glycopep-tide to glycopeptide antibiotic prophylaxis for surgery.Health Technol Assess 12: iii–iv, xi–xii, 1–147.

65. Bolon MK, Morlote M, Weber SG, et al. 2004. Glycopeptidesare no more effective than beta-lactam agents for preven-tion of surgical site infection after cardiac surgery: a meta-analysis. Clin Infect Dis 38:1357–1363.

66. Mangram AJ, Horan TC, Pearson ML, et al. 1999. Guide-line for prevention of surgical site infection, 1999. HospitalInfection Control Practices Advisory Committee. InfectControl Hosp Epidemiol 20:250–278, quiz 279–280.

67. Tacconelli E, Cataldo MA, Albanese A, et al. 2008. Vanco-mycin versus cefazolin prophylaxis for cerebrospinal shuntplacement in a hospital with a high prevalence of metic-illin-resistant Staphylococcus aureus. J Hosp Infect69:337–344.

68. Finkelstein R, Rabino G, Mashiah T, et al. 2002. Vancomy-cin versus cefazolin prophylaxis for cardiac surgery in thesetting of a high prevalence of methicillin-resistant staphy-lococcal infections. J Thorac Cardiovasc Surg 123:326–332.

69. Garey KW, Lai D, Dao-Tran TK, et al. 2008. Interruptedtime series analysis of vancomycin compared to cefuroximefor surgical prophylaxis in patients undergoing cardiacsurgery. Antimicrob Agents Chemother 52:446–451.

70. Spelman D, Harrington G, Russo P, et al. 2002. Clinical,microbiological, and economic benefit of a change inantibiotic prophylaxis for cardiac surgery. Infect ControlHosp Epidemiol 23:402–404.

71. Smith EB, Wynne R, Joshi A, et al. 2012. Is it time toinclude vancomycin for routine perioperative antibioticprophylaxis in total joint arthroplasty patients? J Arthro-plasty 27:55–60.

72. Bull AL, Worth LJ, Richards MJ. 2012. Impact of vancomy-cin surgical antibiotic prophylaxis on the development ofmethicillin-sensitive staphylococcus aureus surgical siteinfections: report from Australian Surveillance Data (VIC-NISS). Ann Surg 256:1089–1092.

73. Miller LG, McKinnell JA, Vollmer ME, et al. 2011. Impactof methicillin-resistant Staphylococcus aureus prevalenceamong S. aureus isolates on surgical site infection riskafter coronary artery bypass surgery. Infect Control HospEpidemiol 32:342–350.

74. Zanetti G, Giardina R, Platt R. 2001. Intraoperativeredosing of cefazolin and risk for surgical site infection incardiac surgery. Emerg Infect Dis 7:828–831.

75. Zanetti G, Goldie SJ, Platt R. 2001. Clinical consequencesand cost of limiting use of vancomycin for perioperativeprophylaxis: example of coronary artery bypass surgery.Emerg Infect Dis 7:820–827.

76. Muralidhar B, Anwar SM, Handa AI, et al. 2006. Preva-lence of MRSA in emergency and elective patients admittedto a vascular surgical unit: implications for antibioticprophylaxis. Eur J Vasc Endovasc Surg 32:402–407.

77. Gemmell CG, Edwards DI, Fraise AP, et al. 2006. Guide-lines for the prophylaxis and treatment of methicillin-resistant Staphylococcus aureus (MRSA) infections in theUK. J Antimicrob Chemother 57:589–608.

78. Elliott RA, Weatherly HL, Hawkins NS, et al. 2010. Aneconomic model for the prevention of MRSA infections aftersurgery: non-glycopeptide or glycopeptide antibiotic pro-phylaxis? Eur J Health Econ 11:57–66.

79. Walsh EE, Greene L, Kirshner R. 2011. Sustained reduc-tion in methicillin-resistant Staphylococcus aureus woundinfections after cardiothoracic surgery. Arch Intern Med171:68–73.

80. Dhadwal K, Al-Ruzzeh S, Athanasiou T, et al. 2007.Comparison of clinical and economic outcomes of two

antibiotic prophylaxis regimens for sternal wound infectionin high-risk patients following coronary artery bypassgrafting surgery: a prospective randomised double-blindcontrolled trial. Heart 93:1126–1133.

81. Patrick S, James C, Ali A, et al. 2010. Vascular surgicalantibiotic prophylaxis study (VSAPS). Vasc EndovascularSurg 44:521–528.

82. Sewick A, Makani A, Wu C, et al. 2012. Does dualantibiotic prophylaxis better prevent surgical site infectionsin total joint arthroplasty? Clin Orthop Relat Res470:2702–2707.

83. Ritter MA, Barzilauskas CD, Faris PM, et al. 1989.Vancomycin prophylaxis and elective total joint arthro-plasty. Orthopedics 12:1333–1336.

84. Cruess RL, Bickel WS, vonKessler KL. 1975. Infections intotal hips secondary to a primary source elsewhere. ClinOrthop Relat Res 106:99–101.

85. Hall AJ. 1974. Late infection about a total knee prosthesis.Report of a case secondary to urinary tract infection. JBone Joint Surg Br 56:144–147.

86. David TS, Vrahas MS. 2000. Perioperative lower urinarytract infections and deep sepsis in patients undergoingtotal joint arthroplasty. J Am Acad Orthop Surg 8:66–74.

87. Ritter MA, Fechtman RW. 1987. Urinary tract sequelae:possible influence on joint infections following total jointreplacement. Orthopedics 10:467–469.

88. Wymenga AB, van Horn JR, Theeuwes A, et al. 1992.Perioperative factors associated with septic arthritis afterarthroplasty. Prospective multicenter study of 362 kneeand 2,651 hip operations. Acta Orthop Scand 63:665–671.

89. Glynn MK, Sheehan JM. 1984. The significance of asymp-tomatic bacteriuria in patients undergoing hip/knee arthro-plasty. Clin Orthop Relat Res 151–154.

90. Waterhouse N, Beaumont AR, Murray K, et al. 1987.Urinary retention after total hip replacement. A prospec-tive study. J Bone Joint Surg Br 69:64–66.

91. Walton JK, Robinson RG. 1982. An analysis of a malepopulation having total hip replacement with regard tourological assessment and post-operative urinary retention.Br J Urol 54:519–521.

92. Drekonja DM, Rector TS, Cutting A, et al. 2013. Urinarytract infection in male veterans: treatment patterns andoutcomes. JAMA Intern Med 173:62–68.

93. Irvine R, Johnson BL Jr, Amstutz HC. 1974. The relation-ship of genitourinary tract procedures and deep sepsis aftertotal hip replacements. Surg Gynecol Obstet 139:701–706.

94. Garibaldi RA, Burke JP, Dickman ML, et al. 1974. Factorspredisposing to bacteriuria during indwelling urethralcatheterization. N Engl J Med 291:215–219.

95. Kunin CM, McCormack RC. 1966. Prevention of catheter-induced urinary-tract infections by sterile closed drainage.N Engl J Med 274:1155–1161.

96. Jerry GJ Jr, Rand JA, Ilstrup D. 1988. Old sepsis prior tototal knee arthroplasty. Clin Orthop Relat Res 135–140.

97. Lee GC, Pagnano MW, Hanssen AD. 2002. Total kneearthroplasty after prior bone or joint sepsis about the knee.Clin Orthop Relat Res 226–231.

98. Larson AN, Hanssen AD, Cass JR. 2009. Does priorinfection alter the outcome of TKA after tibial plateaufracture? Clin Orthop Relat Res 467:1793–1799.

99. Michelson JD, Lotke PA, Steinberg ME. 1988. Urinary-bladder management after total joint-replacement surgery.N Engl J Med 319:321–326.

100. Martinez OV, Civetta JM, Anderson K, et al. 1986. Bacteri-uria in the catheterized surgical intensive care patient.Crit Care Med 14:188–191.

101. Schaeffer AJ. 1986. Catheter-associated bacteriuria. UrolClin North Am 13:735–747.



102. Skelly JM, Guyatt GH, Kalbfleisch R, et al. 1992. Manage-ment of urinary retention after surgical repair of hipfracture. CMAJ 146:1185–1189.

103. Donovan TL, Gordon RO, Nagel DA. 1976. Urinary infec-tions in total hip arthroplasty. Influences of prophylacticcephalosporins and catheterization. J Bone Joint Surg Am58:1134–1137.

104. Fitzgerald RH Jr, Nolan DR, Ilstrup DM, et al. 1977. Deepwound sepsis following total hip arthroplasty. J Bone JointSurg Am 59:847–855.

105. Surin VV, Sundholm K, Backman L. 1983. Infection aftertotal hip replacement. With special reference to a dischargefrom the wound. J Bone Joint Surg Br 65:412–418.

106. Felippe WA, Werneck GL, Santoro-Lopes G. 2007. Surgicalsite infection among women discharged with a drain in situafter breast cancer surgery. World J Surg 31:2293–2299,discussion 2300–2291.

107. Lanier ST, Wang ED, Chen JJ, et al. 2010. The effect ofacellular dermal matrix use on complication rates in tissueexpander/implant breast reconstruction. Ann Plast Surg64:674–678.

108. Sorensen AI, Sorensen TS. 1991. Bacterial growth onsuction drain tips. Prospective study of 489 clean orthope-dic operations. Acta Orthop Scand 62:451–454.

109. van den Brand IC, Castelein RM. 2001. Total joint arthro-plasty and incidence of postoperative bacteriuria with anindwelling catheter or intermittent catheterization withone-dose antibiotic prophylaxis: a prospective randomizedtrial. J Arthroplasty 16:850–855.

110. Oishi CS, Williams VJ, Hanson PB, et al. 1995. Periopera-tive bladder management after primary total hip arthro-plasty. J Arthroplasty 10:732–736.

111. Koulouvaris P, Sculco P, Finerty E, et al. 2009. Relation-ship between perioperative urinary tract infection and deepinfection after joint arthroplasty. Clin Orthop Relat Res467:1859–1867.

112. Brahmbhatt RD, Huebner M, Scow JS, et al. 2012. Nation-al practice patterns in preoperative and postoperativeantibiotic prophylaxis in breast procedures requiringdrains: survey of the American Society of Breast Surgeons.Ann Surg Oncol 19:3205–3211.

113. Phillips BT, Wang ED, Mirrer J, et al. 2011. Currentpractice among plastic surgeons of antibiotic prophylaxisand closed-suction drains in breast reconstruction: experi-ence, evidence, and implications for postoperative care.Ann Plast Surg 66:460–465.

114. Turano A. 1992. New clinical data on the prophylaxis ofinfections in abdominal, gynecologic, and urologic surgery.Multicenter Study Group. Am J Surg 164(4A Suppl):16S–20S.

115. Niederhauser U, Vogt M, Vogt P, et al. 1997. Cardiacsurgery in a high-risk group of patients: is prolongedpostoperative antibiotic prophylaxis effective? J ThoracCardiovasc Surg 114:162–168.

116. Wymenga AB, Hekster YA, Theeuwes A, et al. 1991.Antibiotic use after cefuroxime prophylaxis in hip and kneejoint replacement. Clin Pharmacol Ther 50:215–220.

117. McDonald M, Grabsch E, Marshall C, et al. 1998. Single-versus multiple-dose antimicrobial prophylaxis for majorsurgery: a systematic review. Aust N Z J Surg 68:388–396.

118. Heydemann JS, Nelson CL. 1986. Short-term preventiveantibiotics. Clin Orthop Relat Res 205:184–187.

119. Stone HH, Haney BB, Kolb LD, et al. 1979. Prophylacticand preventive antibiotic therapy: timing, duration andeconomics. Ann Surg 189:691–699.

120. Williams DN, Gustilo RB. 1984. The use of preventiveantibiotics in orthopaedic surgery. Clin Orthop Relat Res190:83–88.

121. Tang WM, Chiu KY, Ng TP, et al. 2003. Efficacy of a singledose of cefazolin as a prophylactic antibiotic in primaryarthroplasty. J Arthroplasty 18:714–718.

122. Bengston S, Knutson K, Lidgren L. 1989. Treatment ofinfected knee arthroplasty. Clin Orthop Relat Res 245:173–178.

123. Sharma D, Douglas J, Coulter C, et al. 2008. Microbiologyof infected arthroplasty: implications for empiric peri-operative antibiotics. J Orthop Surg (Hong Kong) 16:339–342.

124. Nickinson RS, Board TN, Gambhir AK, et al. 2010. Themicrobiology of the infected knee arthroplasty. Int Orthop34:505–510.

125. Peersman G, Laskin R, Davis J, et al. 2001. Infection intotal knee replacement: a retrospective review of 6489 totalknee replacements. Clin Orthop Relat Res 392:15–23.

126. Mont MA, Waldman BJ, Hungerford DS. 2000. Evaluationof preoperative cultures before second-stage reimplantationof a total knee prosthesis complicated by infection. Acomparison-group study. J Bone Joint Surg Am 82-A:1552–1557.

127. Rafiq I, Gambhir AK, Wroblewski BM, et al. 2006. Themicrobiology of infected hip arthroplasty. Int Orthop 30:532–535.

128. Al-Maiyah M, Hill D, Bajwa A, et al. 2005. Bacterialcontaminants and antibiotic prophylaxis in total hip arthro-plasty. J Bone Joint Surg Br 87:1256–1258.

129. Phillips JE, Crane TP, Noy M, et al. 2006. The incidence ofdeep prosthetic infections in a specialist orthopaedic hospi-tal: a 15-year prospective survey. J Bone Joint Surg Br88:943–948.

130. Tsukayama DT, Estrada R, Gustilo RB. 1996. Infectionafter total hip arthroplasty. A study of the treatment of onehundred and six infections. J Bone Joint Surg Am 78:512–523.

131. Moran E, Masters S, Berendt AR, et al. 2007. Guidingempirical antibiotic therapy in orthopaedics: the microbiol-ogy of prosthetic joint infection managed by debridement,irrigation and prosthesis retention. J Infect 55:1–7.

132. Azzam K, McHale K, Austin M, et al. 2009. Outcome of asecond two-stage reimplantation for periprosthetic kneeinfection. Clin Orthop Relat Res 467:1706–1714.

133. Kalra KP, Lin KK, Bozic KJ, et al. 2010. Repeat 2-stagerevision for recurrent infection of total hip arthroplasty. JArthroplasty 25:880–884.

134. Jamsen E, Stogiannidis I, Malmivaara A, et al. 2009.Outcome of prosthesis exchange for infected knee arthro-plasty: the effect of treatment approach. Acta Orthop80:67–77.

135. Kubista B, Hartzler RU, Wood CM, et al. 2012. Reinfectionafter two-stage revision for periprosthetic infection of totalknee arthroplasty. Int Orthop 36:65–71.

136. Mortazavi SM, O’Neil JT, Zmistowski B, et al. 2012. Repeat2-stage exchange for infected total hip arthroplasty: aviable option? J Arthroplasty 27:923.e1–926.e1.

137. Fletcher N, Sofianos D, Berkes MB, et al. 2007. Preventionof perioperative infection. J Bone Joint Surg Am 89:1605–1618.

138. Dellinger EP, Gross PA, Barrett TL, et al. 1994. Qualitystandard for antimicrobial prophylaxis in surgical proce-dures. The Infectious Diseases Society of America. InfectControl Hosp Epidemiol 15:182–188.

139. Scher KS. 1997. Studies on the duration of antibioticadministration for surgical prophylaxis. Am Surg 63:59–62.

140. Shapiro M, Munoz A, Tager IB, et al. 1982. Risk factors forinfection at the operative site after abdominal or vaginalhysterectomy. N Engl J Med 307:1661–1666.



141. Polk HC Jr, Trachtenberg L, Finn MP. 1980. Antibioticactivity in surgical incisions. The basis of prophylaxis inselected operations. JAMA 244:1353–1354.

142. Ohge H, Takesue Y, Yokoyama T, et al. 1999. An additionaldose of cefazolin for intraoperative prophylaxis. Surg Today29:1233–1236.

143. Morita S, Nishisho I, Nomura T, et al. 2005. The signifi-cance of the intraoperative repeated dosing of antimicro-bials for preventing surgical wound infection in colorectalsurgery. Surg Today 35:732–738.

144. Swoboda SM, Merz C, Kostuik J, et al. 1996. Doesintraoperative blood loss affect antibiotic serum and tissueconcentrations? Arch Surg 131:1165–1171, discussion1171–1172.

145. Markantonis SL, Kostopanagiotou G, Panidis D, et al.2004. Effects of blood loss and fluid volume replacement onserum and tissue gentamicin concentrations during colorec-tal surgery. Clin Ther 26:271–281.

146. Klekamp JW, DiPersio D, Haas DW. 1999. No influence oflarge volume blood loss on serum vancomycin concentra-tions during orthopedic procedures. Acta Orthop Scand70:47–50.

147. Meter JJ, Polly DW Jr, Brueckner RP, et al. 1996. Effect ofintraoperative blood loss on the serum level of cefazolin inpatients managed with total hip arthroplasty. A prospec-tive, controlled study. J Bone Joint Surg Am 78:1201–1205.

148. Polly DW Jr, Meter JJ, Brueckner R, et al. 1996. The effectof intraoperative blood loss on serum cefazolin level inpatients undergoing instrumented spinal fusion. A prospec-tive, controlled study. Spine (Phila Pa 1976) 21:2363–2367.

149. Sanford JP, Moellering RC, Gilbert DN, editors. 2000.Sanford guide to antimicrobial therapy. 30th ed; Antimicro-bial Therapy, Inc.

150. Wurtz R, Itokazu G, Rodvold K. 1997. Antimicrobial dosingin obese patients. Clin Infect Dis. 25:112–118.

151. Leader WG, Tsubaki T, Chandler MH. 1994. Creatinine-clearance estimates for predicting gentamicin pharmacoki-netic values in obese patients. Am J Hosp Pharm 51:2125–2130.

152. Janson B, Thursky K. 2012. Dosing of antibiotics in obesity.Curr Opin Infect Dis 25:634–649.

153. Begg EJ, Barclay ML, Duffull SB. 1995. A suggestedapproach to once-daily aminoglycoside dosing. Br J ClinPharmacol 39:605–609.

154. Truong J, Levkovich BJ, Padiglione AA. 2012. Simpleapproach to improving vancomycin dosing in intensivecare: a standardised loading dose results in earlier thera-peutic levels. Intern Med J 42:23–29.

155. Ho VP, Nicolau DP, Dakin GF, et al. 2012. Cefazolin dosingfor surgical prophylaxis in morbidly obese patients. SurgInfect (Larchmt) 13:33–37.

156. Rybak MJ, Lomaestro BM, Rotschafer JC, et al. 2009.Vancomycin therapeutic guidelines: a summary of consen-sus recommendations from the infectious diseases Societyof America, the American Society of Health-System Phar-macists, and the Society of Infectious Diseases Pharma-cists. Clin Infect Dis 49:325–327.

157. Traynor AM, Nafziger AN, Bertino JS Jr. 1995. Amino-glycoside dosing weight correction factors for patients ofvarious body sizes. Antimicrob Agents Chemother 39:545–548.

158. van Kralingen S, Taks M, Diepstraten J, et al. 2011.Pharmacokinetics and protein binding of cefazolin inmorbidly obese patients. Eur J Clin Pharmacol 67:985–992.

159. Edmiston CE, Krepel C, Kelly H, et al. 2004. Perioperativeantibiotic prophylaxis in the gastric bypass patient: do weachieve therapeutic levels? Surgery 136:738–747.

160. Liu C. 2011. The bundled approach to MRSA surgical siteinfection prevention: is the whole greater than the sum ofits parts?: comment on “Sustained reduction in methicillin-resistant Staphylococcus aureus wound infections aftercardiothoracic surgery”. Arch Intern Med 171:73–74.

161. Pofahl WE, Goettler CE, Ramsey KM, et al. 2009. Activesurveillance screening of MRSA and eradication of thecarrier state decreases surgical-site infections caused byMRSA. J Am Coll Surg 208:981–986, discussion 986–988.

162. Wirganowicz PZ, Eckardt JJ, Dorey FJ, et al. 1999.Etiology and results of tumor endoprosthesis revisionsurgery in 64 patients. Clin Orthop Relat Res 358:64–74.

163. Safran MR, Kody MH, Namba RS, et al. 1994. 151endoprosthetic reconstructions for patients with primarytumors involving bone. Contemp Orthop 29:15–25.

164. Malawer MM, Chou LB. 1995. Prosthetic survival andclinical results with use of large-segment replacements inthe treatment of high-grade bone sarcomas. J Bone JointSurg Am 77:1154–1165.

165. Capanna R, Morris HG, Campanacci D, et al. 1994.Modular uncemented prosthetic reconstruction after resec-tion of tumours of the distal femur. J Bone Joint Surg Br76:178–186.

166. Mittermayer F, Krepler P, Dominkus M, et al. 2001. Long-term followup of uncemented tumor endoprostheses for thelower extremity. Clin Orthop Relat Res 388:167–177.

167. Ghert M, Deheshi B, Holt G, et al. 2012. Prophylacticantibiotic regimens in tumour surgery (PARITY): protocolfor a multicentre randomised controlled study. BMJOpen 2.

168. Jansen B, Rinck M, Wolbring P, et al. 1994. In vitroevaluation of the antimicrobial efficacy and biocompatibili-ty of a silver-coated central venous catheter. J BiomaterAppl 9:55–70.

169. Karchmer TB, Giannetta ET, Muto CA, et al. 2000. Arandomized crossover study of silver-coated urinary cathe-ters in hospitalized patients. Arch Intern Med 160:3294–3298.

170. Tsuchiya H, Shirai T, Nishida H. 2013. New implanttechnology: iodine-coating for infection control. Paper pre-sented at: 26th European Musculoskeletal Oncology SocietyMeeting, Gothenburg, Sweden.

171. Tsuchiya H, Shirai T, Nishida H, et al. 2012. Innovativeantimicrobial coating of titanium implants with iodine. JOrthop Sci 17:595–604.

172. Hardes J, von Eiff C, Streitbuerger A, et al. 2010. Reduc-tion of periprosthetic infection with silver-coated megapros-theses in patients with bone sarcoma. J Surg Oncol101:389–395.

173. Gosheger G, Hardes J, Ahrens H, et al. 2004. Silver-coatedmegaendoprostheses in a rabbit model—an analysis of theinfection rate and toxicological side effects. Biomaterials25:5547–5556.

174. Buttaro MA, Gimenez MI, Greco G, et al. 2005. High activelocal levels of vancomycin without nephrotoxicity releasedfrom impacted bone allografts in 20 revision hip arthroplas-ties. Acta Orthop 76:336–340.

175. Gosheger G, Goetze C, Hardes J, et al. 2008. The influenceof the alloy of megaprostheses on infection rate. J Arthro-plasty 23:916–920.

176. Witso E, Persen L, Benum P, et al. 2004. High localconcentrations without systemic adverse effects after im-paction of netilmicin-impregnated bone. Acta Orthop Scand75:339–346.

177. Buttaro MA, Pusso R, Piccaluga F. 2005. Vancomycin-supplemented impacted bone allografts in infected hip



arthroplasty. Two-stage revision results. J Bone Joint SurgBr 87:314–319.

178. Michalak KA, Khoo PP, Yates PJ, et al. 2006. Iontophor-esed segmental allografts in revision arthroplasty forinfection. J Bone Joint Surg Br 88:1430–1437.

179. Khoo PP, Michalak KA, Yates PJ, et al. 2006. Iontophoresisof antibiotics into segmental allografts. J Bone Joint SurgBr 88:1149–1157.

180. Winkler H, Stoiber A, Kaudela K, et al. 2008. One stageuncemented revision of infected total hip replacement usingcancellous allograft bone impregnated with antibiotics. JBone Joint Surg Br 90:1580–1584.

181. Buttaro MA, Guala AJ, Comba F, et al. 2010. Incidence ofdeep infection in aseptic revision THA using vancomycin-impregnated impacted bone allograft. Hip Int 20:535–541.

182. Witso E, Persen L, Loseth K, et al. 1999. Adsorption andrelease of antibiotics from morselized cancellous bone. Invitro studies of antibiotics. Acta Orthop Scand 70:298–304.

183. Buttaro MA, Gonzalez Della Valle AM, Pineiro L, et al.2003. Incorporation of vancomycin-supplemented bone in-corporation of vancomycin-supplemented bone allografts:radiographical, histopathological and immunohistochemicalstudy in pigs. Acta Orthop Scand 74:505–513.

184. Witso E, Persen L, Loseth K, et al. 2000. Cancellous boneas an antibiotic carrier. Acta Orthop Scand 71:80–84.

185. Berbari EF, Hanssen AD, Duffy MC, et al. 1998. Riskfactors for prosthetic joint infection: case-control study.Clin Infect Dis 27:1247–1254.

186. Yang K, Yeo SJ, Lee BP, et al. 2001. Total knee arthro-plasty in diabetic patients: a study of 109 consecutive cases.J Arthroplasty 16:102–106.

187. Meding JB, Reddleman K, Keating ME, et al. 2003. Totalknee replacement in patients with diabetes mellitus. ClinOrthop Relat Res 416:208–216.

188. Pedersen AB, Mehnert F, Johnsen SP, et al. 2010. Risk ofrevision of a total hip replacement in patients with diabetesmellitus: a population-based follow up study. J Bone JointSurg Br 92:929–934.

189. Adams AL, Paxton EW, Wang JQ, et al. 2013. Surgicaloutcomes of total knee replacement according to diabetesstatus and glycemic control, 2001 to 2009. J Bone JointSurg Am 95:481–487.

190. Haverkamp D, Klinkenbijl MN, Somford MP, et al. 2011.Obesity in total hip arthroplasty—does it really matter? Ameta-analysis. Acta Orthop 82:417–422.

191. Namba RS, Paxton L, Fithian DC, et al. 2005. Obesity andperioperative morbidity in total hip and total knee arthro-plasty patients. J Arthroplasty 20:46–50.

192. Dowsey MM, Choong PF. 2008. Early outcomes andcomplications following joint arthroplasty in obese patients:a review of the published reports. ANZ J Surg 78:439–444.

193. Parvizi J, Sullivan TA, Pagnano MW, et al. 2003. Totaljoint arthroplasty in human immunodeficiency virus-posi-tive patients: an alarming rate of early failure. J Arthro-plasty 18:259–264.

194. Ragni MV, Crossett LS, Herndon JH. 1995. Postoperativeinfection following orthopaedic surgery in human immuno-deficiency virus-infected hemophiliacs with CD4 counts <

or ¼200/mm3. J Arthroplasty 10:716–721.195. Habermann B, Eberhardt C, Kurth AA. 2008. Total joint

replacement in HIV positive patients. J Infect 57:41–46.196. Wang TI, Chen CF, Chen WM, et al. 2012. Joint replace-

ment in human immunodeficiency virus-infected patients. JChin Med Assoc 75:595–599.

197. Unger AS, Kessler CM, Lewis RJ. 1995. Total kneearthroplasty in human immunodeficiency virus-infectedhemophiliacs. J Arthroplasty 10:448–452.

198. Silva M, Luck JV Jr. 2005. Long-term results of primarytotal knee replacement in patients with hemophilia. J BoneJoint Surg Am 87:85–91.

199. Rodriguez-Merchan EC. 2007. Total knee replacement inhaemophilic arthropathy. J Bone Joint Surg Br 89:186–188.

200. Shaarani SR, Collins D, O’Byrne JM. 2012. The needfor guidelines in asplenic patients undergoing totaljoint arthroplasty: a case report. Case Rep Orthop 2012:147042.

201. McCleery MA, Leach WJ, Norwood T. 2010. Rates ofinfection and revision in patients with renal diseaseundergoing total knee replacement in Scotland. J BoneJoint Surg Br 92:1535–1539.

202. Lieberman JR, Fuchs MD, Haas SB, et al. 1995. Hiparthroplasty in patients with chronic renal failure. JArthroplasty 10:191–195.

203. Sakalkale DP, Hozack WJ, Rothman RH. 1999. Total hiparthroplasty in patients on long-term renal dialysis. JArthroplasty 14:571–575.

204. Li WC, Shih CH, Ueng SW, et al. 2010. Uncemented totalhip arthroplasty in chronic hemodialysis patients. ActaOrthop 81:178–182.

205. Nagoya S, Nagao M, Takada J, et al. 2005. Efficacy ofcementless total hip arthroplasty in patients on long-termhemodialysis. J Arthroplasty 20:66–71.

206. Pour AE, Matar WY, Jafari SM, et al. 2011. Total jointarthroplasty in patients with hepatitis C. J Bone JointSurg Am 93:1448–1454.

207. Orozco F, Post ZD, Baxi O, et al. 2014. Fibrosis in hepatitisC patients predicts complications after elective total jointarthroplasty. J Arthroplasty 29:7–10.

208. Vergidis P, Lesnick TG, Kremers WK, et al. 2012. Prosthet-ic joint infection in solid organ transplant recipients: aretrospective case-control study. Transpl Infect Dis 14:380–386.

209. Tannenbaum DA, Matthews LS, Grady-Benson JC. 1997.Infection around joint replacements in patients who have arenal or liver transplantation. J Bone Joint Surg Am79:36–43.

210. Mortazavi SM, Molligan J, Austin MS, et al. 2011. Failurefollowing revision total knee arthroplasty: infection is themajor cause. Int Orthop 35:1157–1164.

211. Poss R, Thornhill TS, Ewald FC, et al. 1984. Factorsinfluencing the incidence and outcome of infection followingtotal joint arthroplasty. Clin Orthop Relat Res 182:117–126.

212. Sierra RJ, Cooney WPt, Pagnano MW, et al. 2004. Reopera-tions after 3200 revision TKAs: rates, etiology, and lessonslearned. Clin Orthop Relat Res 425:200–206.

213. Liu C, Kakis A, Nichols A, et al. 2014. Targeted Use ofVancomycin as Perioperative Prophylaxis Reduces Peri-prosthetic Joint Infection in Revision TKA. Clin OrthopRelat Res 472:227–231.

214. Cuzon G, Naas T, Demachy MC, et al. 2008. Plasmid-mediated carbapenem-hydrolyzing beta-lactamase KPC-2in Klebsiella pneumoniae isolate from Greece. AntimicrobAgents Chemother 52:796–797.

215. Nordmann P, Cuzon G, Naas T. 2009. The real threat ofKlebsiella pneumoniae carbapenemase-producing bacteria.Lancet Infect Dis 9:228–236.

216. Queenan AM, Bush K. 2007. Carbapenemases: the versa-tile beta-lactamases. Clin Microbiol Rev 20:440–458, tableof contents.

217. Schwaber MJ, Carmeli Y. 2008. Carbapenem-resistantEnterobacteriaceae: a potential threat. JAMA 300:2911–2913.



218. Bratu S, Landman D, Haag R, et al. 2005. Rapid spread ofcarbapenem-resistant Klebsiella pneumoniae in New YorkCity: a new threat to our antibiotic armamentarium. ArchIntern Med 165:1430–1435.

219. Marchaim D, Navon-Venezia S, Schwaber MJ, et al. 2008.Isolation of imipenem-resistant Enterobacter species: emer-gence of KPC-2 carbapenemase, molecular characteriza-tion, epidemiology, and outcomes. Antimicrob AgentsChemother 52:1413–1418.

220. de Smet AM, Kluytmans JA, Cooper BS, et al. 2009.Decontamination of the digestive tract and oropharynx inICU patients. N Engl J Med 360:20–31.

221. Perez F, Pultz MJ, Endimiani A, et al. 2011. Effect ofantibiotic treatment on establishment and elimination ofintestinal colonization by KPC-producing Klebsiella pneu-moniae in mice. Antimicrob Agents Chemother 55:2585–2589.



perioperative antibiotics

Documents