Analysis of clinical management of infected breast implants and of factors associated to successful breast pocket salvage in infections occurring after breast reconstruction

Analysis

of

clinical

management

of

infected

breast

implants

and

of

factors

associated

to

successful

breast

pocket

salvage

in

infections

occurring

after

breast

reconstruction

Simonetta

Franchelli

_*

_,

_Marianna

_Pesce

_,

_Ilaria

_Baldelli

_,

_Anna

_Marchese

_,

Pierluigi

Santi

_,

_Andrea

_De

_Maria

a_{SCChirurgiaPlasticaeRicostruttiva,PoliclinicoSanMartino,LargoRosannaBenzi10,Genoa16132,Italy} b_{UOCMicrobiologia,PoliclinicoSanMartino,LargoRosannaBenzi10,Genoa16132,Italy}

c_{SezionediMicrobiologia,DISC,UniversitàdiGenova,16132Genoa,Italy}

d_{ClinicaMalattieInfettive,ProgrammaInfettivologiadell’OspiteImmunocompromessoPoliclinicoSanMartino,LargoRosannaBenzi10,Genoa16132,Italy} e_{DepartmentofHealthSciences,UniversityofGenova,16132Genova,Italy}

ARTICLE INFO Articlehistory:

Received22December2017

Receivedinrevisedform9March2018 Accepted30March2018

Corresponding Editor: Eskild Petersen, Aarhus,Denmark

Keywords: Breastsurgery Breastimplantinfection Antibiotictreatment

ABSTRACT

Objectives:Considerableeffortshavebeendevotedsofartoimprovesalvageproceduresofinfected breastimplantsinabsenceofdefinedguidelinesorvalidatedclinicalprotocols.Withinacohortof prospectivelyrecruitedpatientswhounderwentbreastreconstruction,weperformeda retrospec-tivereviewofprovenimplantinfectionsinordertodescribefactorscontributingtomanagement success.

Methods: We collected data in 1293 consecutive patients who underwent two stage (expander+ prosthesis)breastreconstructionwithatleast12monthsoffollow-up.Demographicdata,timingof infection,typeofmicroorganism,intentofsalvage,fateoftheimplant,typeofantibiotictreatmentand follow-upwererecordedinaprospectivedatacollectiononclinicalrecords.

Results:Implantinfectionsoccurredin103of1293patients(8%).Amongthese,73(71%)wereproven infectionswithconfirmedmicrobiology.Implantpocketsalvagewasattemptedin43/73(59%).patients Ahigherproportionofexpanderimplantpocketsweresuccessfullysavedcomparedtoprostheticpocket (p=0,04).Gram-positivemicrorganismsrepresentedthemajorityofetiologicagents,withcoagulase negativestaphylococciprevailingoverStaphylococcusaureus.Noassociationwasobservedbetween successrateandtypeofinfectingmicroorganism.Ahigher proportionofpatientswithpreviousor intraoperativeradiotherapy orwithperioperative chemotherapyunderwent anattempt ofimplant salvage(p=0,081and0,0571trend,respectively).Nosingleantibioticregimenwassuperiortotheothers intermsofsuccessrate.Implantpocketsalvagewashigherinexpanderscomparedtoprostheses(74%vs 33%p=0,04).Highersuccessratesinimplantpocketsalvagewereevidentwhenimplantreplacement wasprecededandfollowedbyantibiotictreatmentcomparedtoinpatientantibiotictreatmentalone (100%versus57%,p=0,035).

Conclusion:Patientselectioninclinicalpracticeleadstodifferencesinpatientswithbreastimplant infectionwhoareconsideredforattemptsatimplantsalvagevs.thosewhoaretreatedwithimplant removal.Salvageofbreastimplantpocketscanbeobtainedinthemajorityofpatientswithcombined one-step implant replacementsurgeryandantibiotic treatment. Increasedefforts and protocolsto recruitpatientsintopocketsalvagemanagementareneeded.

©2018TheAuthors.PublishedbyElsevierLtdonbehalfofInternationalSocietyforInfectiousDiseases. ThisisanopenaccessarticleundertheCCBY-NC-NDlicense( http://creativecommons.org/licenses/by-nc-nd/4.0/).

Introduction

Implant-based breast reconstruction is the most common surgical option for women with breast cancer and post-mastectomy breast defects. Among a long list of possible postoperative complications (Georgiade et al., 1982; Schuster

* Corresponding authorat: Plastic and ReconstructiveSurgery Department, PoliclinicoSanMartino,LargoR.Benzi10,Genoa16132,Italy.

E-mailaddresses:simonettafranchelli@yahoo.it(S.Franchelli),

mariannap86@hotmail.com(M.Pesce),ilaria.baldelli@unige.it(I.Baldelli),

anna.marchese@unige.it(A.Marchese),plsanti@unige.it(P.Santi),

de-maria@unige.it(A.DeMaria).

https://doi.org/10.1016/j.ijid.2018.03.019

1201-9712/©2018TheAuthors.PublishedbyElsevierLtdonbehalfofInternationalSocietyforInfectiousDiseases.ThisisanopenaccessarticleundertheCCBY-NC-ND license(http://creativecommons.org/licenses/by-nc-nd/4.0/).

ContentslistsavailableatScienceDirect

International

Journal

of

Infectious

Diseases

et al., 1990; Pusic and Cordeiro, 2003), implant infection representsadevastatingeventforthepatientandadiscouraging situation for thesurgeon. Infection afterimplant-based recon-struction occurs at a mean rate of 8% (1–35%) and may vary according toclinical settingand typeof procedure(Armstrong et al., 1989; Cordeiro and McCarthy, 2006; Nahabedian et al., 2003;Spearetal.,2004;Franchellietal.,2015).

Over recent years, a consistent clinical focus targeted risk factors associated to post-implant infection, causative micro-organisms isolated frominfected implants(Brand,1993; Rieger etal.,2013;Reishetal.,2013;Momohetal.,2014;Kronowitzand Robb,2009),andprevention(Barretal.,2016).Oncebreastimplant infection is detected, there is general clinical agreement that antibiotictreatment shouldbeprovidedand, mostimportantly, thatattemptsshouldbemadetoultimatelysalvagetheimplant. Thereisaconsiderabledegreeofdisagreement,however,onthe definition of implant salvage and on standardized clinical management.Indeed,“implantsalvage”thismayincludesalvage either oftheaestheticreconstructiveresult(Rieger etal.,2013; SpearandSeruya,2010),oftheimplantpocket,oroftheoriginal implantitself(Wilkinsonetal.,1985).

Similarlywidelydivergingviewsandclinicalapproachesexist on optimal clinical management aiming at implant salvage. Standard management of breast implant infection combines implant removalwithantibiotictreatmentfollowedbydelayed positioning of a new implant weeks-months after antibiotic treatment of the previous infected pocket (Armstrong et al., 1989).Thisapproachrequiresmultiplesurgicalprocedures(e.g.,1 removal, 1 insertion) with proportionally increased associated risksandcosts.Inaddition,insomeoftheseinstances,thedelayed implant re-insertion may be technically more difficult due to pocketresorptionandtissuefibrosis(Reishetal.,2013).

Considerable effort has been devoted to improve salvage procedures minimizing the waiting time between implant removalandreimplantation.SeveralprotocolsproposedbyNorth AmericanandEuropeangroupsmayallowtissuedebridmentand implant replacement during the same surgical procedure including systemic empiric antibiotic therapy with wound drainage(Courtissetal.,1979),pre-andpostoperativeantibiotic therapywithcapsulotomywoundirrigationandimplant replace-ment (Weber andHentz,1986)orirrigationof implantpocket with saline solution combined with antibiotic therapy and implant exchange with or without capsulectomy (Chun and Schulman,2007;Laveauxetal.,2009;Princeetal.,2012;Missana etal.,2012).Thementionedseriesareuncontrolledstudiesona mixofprovenandpossibleinfections(i.e.without microbiolog-icalisolation)withoveralllimitednumberofpatientsandwith highly variable and sometimes limited follow-up (i.e., 1–6 months).

When considering actual infections, Staphylococci are the predominant cause of breast implant infection. Empiric initial treatmentisbasedonthisfactandontherelativelocalprevalence ofmethicillin-resistantstaphylococci.Gramnegative microorgan-ismsaredetectedinaminorproportionofcases,areaddressedby current guidelines for SSI (surgical site infection) prophylaxis (Berrios-Torresetal.,2017)and,whenpresentmaybeassociated with lower rates of device salvage (Spear and Seruya, 2010; Franchellietal.,2012).

In viewof the present paucity of structuredindications for managementandtreatmentandoftheneedtoimproveourclinical approachtopatientswithreconstructivebreastimplantinfection, weperformedaretrospectivereviewonbreastimplantinfections and of implant salvageprocedures at our center. Thisanalysis shows that there are discrepancies in patient selection and in outcomewhenanunstructuredapproachofantibiotictreatment andsurgeryisactive.

Patientsandmethods

Between February 1st 2009 and June 30st 2015, 1293 consecutive patients underwent two-stage (expander+implant) immediateordelayedimplantreconstructionaftermastectomyor implant replacement at the Policlinico San Martino in Genoa. Clinicaldataofallpatientsatourinstitutionarekeptonrecord. Written signed consent for data use for clinical purposes was obtainedfromallthepatients,accordingtoapprovedinstitutional policy. The present work is a retrospective elaboration of institutionalrecorddata.

Implantinfection wasdefinedaspreviouslydescribed(Franchelli etal.,2012).Briefly,criteriaforpossibleinfectionincluded:fever, minimallocaledema and inflammation; probable infectionwas defined bycellulitis, leukocytosis,systemic inflammation, echo-graphicevidenceofinflammationorperi-prostheticliquid accumu-lation without microorganism isolation. Proven infections are definedbythepresenceofpurulentdischargeand/ormicroorganism isolationinadditiontootherclinicalsigns.Implantinfectionsare furthersubdividedinto“early”(<60daysfrom surgery,DFS)and “late”(>60DFS)infections(Franchellietal.,2015).

Patientswerecensoredatapost-operativefollow-upofatleast 12months,inordertoincludelateinfections(i.e.occurring>180 DFS).

Allpatientsunderwentsurgerybythesamesurgicalteam.The sametypeofexpanderandprosthesiswasused(Allergan,Marlow International,Parkway,UnitedKingdom).Expandersand prosthe-ses had the same textured surface (Biocell: Allergan, Marlow International, Parkway, UK) and external envelope structure (Intrashiel:Allergan, Marlow International, Parkway, UK)made by three overlapped layers of silicone rubber. Expanders were positioned in a submuscular pocket after mastectomy, and progressivelyinflated;prostheses werepositionedasa replace-ment of a previous expander after its inflation or during a contralateralaugmentationmammaplasty.Antibioticprophylaxis usingCefazolin14–20mg/kg(1gfollowedby2dosesq8h)was alwaysadministered.Inpatientsallergictobeta-lactams(surgeons ASop,2018;Phillipsetal.,2016)Clindamycinwasused.

Thefollowingpatientdatawerecollected:age,tumorstageand recurrence, chemo/radio–therapy, type of implants, timing and clinicalappearanceofinfection,typeofmicroorganism,intentof pocketsalvage,fateoftheimplant(removalorsalvage),typeof antibiotictreatment,follow-upafterimplantsalvage.

Wedefinedas“perioperativechemotherapy”any chemothera-pyperformedaroundbreastreconstructionsurgeryeitherbefore oraftertheprocedure.

Standard management options for patients with proven implant infection were: (i) implant removal and antibiotic treatment,(ii)attemptofimplantpocketsalvagewithantibiotic treatment alone, (iii) combined implant replacement. Implant removalwithoutanypocketsalvageattemptisconsideredatour Center whenever at least one of the following conditions is present: implant exposure following cutaneous subcutaneous necrosis,cutaneousfistula,extensiveinflammationwithintense redness,painandedemainvolving!3breastquadrants.

Clinicalrecordanalysiswasusedtoidentifyallthosepatients whowereconsideredfor anattemptof implantpocketsalvage (“intentofsalvage”).Patientsforwhomnoprospectiveintentwas stated in the clinical record were defined as “salvage intent unknown”andwereincludedinthegroup“nointentofsalvage” (Figure1).

Successful salvage of the surgical implant was defined as retentionof theimplantfor !12monthsafter surgery.Cultures from multiple sites including wound, drainage fluid, peri-prosthetic seroma,capsular tissues and the prostheticmaterial itself, when removed, wereperformedtoidentifythe infecting

pathogens.SusceptibilitytestingwasperformedusingtheVitek2 AESautomatedsystem(bioMerièux, Marcy-l’etoile (France)and interpretedaccordingtoEUCASTstandards(EUCAST,2018).

Antibiotictreatmentwasadministeredfor10–12days.When combinedwithdebridementandimplantreplacement(Antibiotic treatment and Implant Replacement, AIR) based on available microbiological isolatesensitivity, surgery for implant replace-mentwasperformed3–4daysafteronsetofantibiotictreatment. Surgical procedures during combined salvage management includedcapsulardebridmentuptocapsulectomy,dependingon breasttissuethickness,andimplantpocketirrigationwithsaline solution.Acompleteexchangeofsurgicalinstruments/towelsand surgicalcoatswasperformedduringtheprocedurebeforeimplant replacementaccompaniedbypatientdisinfection.

To evaluate group differences, statistical analysis was per-formedemployingtwo-sidedtestsasrequired(JMP10.0Statistical software,SAS InstituteInc. Mann).Chi-square analysis, Fisher’s exacttestandMann–Whitneytestwereperformedasappropriate totestfordifferencesbetweengroups.

Results

Patientmanagementandsalvageattempts.

Implantinfectionswereobservedin103of1293patients(8%). Amongthesepatients,73(71%)hadproveninfections(PI)(Figure 1).Patientswithoutmicrobiologically confirmedinfectionwere excluded from further analysis since management and clinical outcomeinthesecasescannotbelinkedtoaspecific microorgan-ism(s) or to true infection. We next analyzed the clinical management of these 73 patients with microbiologically con-firmedPI.Implantpocketsalvageattemptswereperformedin43 of 73 patients (59%), while 30 cases (41%) underwent implant removalfollowedbyantibiotictreatmentduetoextensivecellulitis and/or overwhelming breast implant infection (Figure 1). No associations were detected between age, tumor stage, type of implantandinfectionstagingdefinedbothastimeaftersurgery and astype ofinfection basedonmicrobiologicalidentification (Table 1). A higher proportion of patients with previous or

intraoperativeradiotherapy(IORT)orperioperativechemotherapy underwentanattemptofbreastimplantpocketsalvage(p=0,081 andp=0,0571trend,respectively).

Outcomeofsalvageattemptsaccordingtopatientcharacteristicsand tomicrobiologicalisolate

Wenextstratifiedpatientsaccordingtotheoutcome(success vs.failure)ofclinicalmanagementinthosewhounderwentclinical attemptsatimplantpocketsalvage.AsshowninTable2,therewas noassociationbetweenoutcomeandage,tumorstage,relapseand radio-orchemotherapynorwithtimeofinfectionaftersurgery (earlyvslateinfections).Ahigherproportionofexpanderimplant pockets weresuccessfully savedcompared toprosthetic pocket (p=0,04). This difference was independent from the onset of infectionaftersurgerybothintermsofearlyvslateonsetwitha limitsetat60daysfromsurgery(Table2)aswellasintermsof absolutenumberofdaysfromsurgerytoinfection(p=0,161n.s.)

Among85microbiologicalisolatesobtainedfrom73patients with PI, thevast majoritywere Grampositive microorganisms (83%). As shown in Figure 2 staphylococcal infections largely prevailed overGramnegativeinfections (65/85,76,5%vs 13/85,

15,3%). Among thesestaphylococci, 49% were represented by methicillin sensitive cocci. Higher proportions of coagulase negativecoccicompared toStaphylococcusaureus,(43% vs34%, respectively) were isolated from capsular tissue and peri-prostheticfluid.

Clinicalmanagementandoutcomeofinfection

In order to verify whether successful treatment of breast implant pocket could be associated to the type of causative microorganism we next stratified the isolates by treatment outcome.

Amongstaphylococcalinfectionsatrendtowardsahigherrate of successful salvage was recorded for Coagulase Negative StaphylococciwhencomparedtoS.aureusinfectionswith18/36 vs8/29successfultreatments,respectively(p=0,0798).Inaddition amuchlowerproportionofinfectionswereduetoMRSAwhen comparedtoMSSA(3,5%vs30,6%).

Thesedataconfirmpreviousobservationsshowing Staphylo-cocci(Brand,1993;Barretal.,2016)asthemainaetiologicagentsin breast implant infections, albeit these reports had higher proportions of S. aureus. Here, indeed, a prevalence of Coagu-lase-negativestaphylococcioverS.aureusemergesandconfirms previous reports at our clinical site (Armstrong et al., 1989; Franchellietal.,2015).

Sincefactorsassociatedtotheoutcomeofaninfectionmaybe related to the host, to the microorganism or to the choice of antibiotic/clinical management, we next studied the antibiotic regimensadministeredtothepatientstogetherwiththeirsurgical management. Overall, a high number of different antibiotic regimens, according to the choice of molecule(s), dosage and duration of administration that were administered to the 43 patientsforwhomanattemptofimplantsalvagewasperformed. Eight of these patients also underwent combined surgical replacementoftheimplant(Figure1).Antibioticregimenswere groupedinto4groupsbasedonantibioticclasscompositionand areshowninTable3.Whilenoneofthese4regimenswassuperior to the others in terms of implant salvage success rate, a significantly higher proportion of success was detected when the clinical management that included systemic antibiotic administration and surgical removal (AIR) was compared to antibiotic treatment alone (8/8,100% vs 20/35, 57%respectively, p=0,035).

Discussion

Thepresentworkanalyzesthemanagementofproveninfected implant-basedbreastreconstructionsandtreatmentstrategiesina single-centerunitwithintheframeofeverydayclinicalpractice afterstratificationofoutcome(salvagevs.failure)toretainimplant pocket.

Proveninfectionsrepresentedthemajorityofalltheinfections caredfor(71%).ArelevantproportionofpatientswithPI(n30or

Table1

Demographicdataofprovenbreastimplantinfections. Proven Infections(n=73) Salvageintent Significance Tot N No N(%) Yes N(%) N" _{73 30 43} Meanage 54 52 55 ns Stage0 3 3(10) 0 ns Stage1 10 2(7) 8(19) ns Stage2 19 7(23) 12(28) ns Stage3 18 10(34) 8(19) ns Unknownstage 4 1(3) 3(7) Relapseorrecurrence 19 7(23) 12(28) ns Radiotherapy(1) 26 6(20) 22(51) P=0,0081 Chemotherapy(2) 41 21(70) 20(46) P=0,0577 Expander 61 27(90) 34(79) ns Prosthesis 12 3(10) 9(21) ns Earlyinfection 49 19(64) 30(70) ns Lateinfection 24 11(37) 13(30) ns Probableinfection 0 0 0 Proveninfection 73 30(100) 43(100)

(1)Previousorintraoperative;(2)PerioperativeChemotherapy(pre-infection);n.s.: not significant; Tot:total; (%)proportionofpatients with agivenparameter accordingtosalvageintent.

Table2

Clinicalparametersinpatientswithproveninfectionswhounderwentsalvage attemptsofimplantpocketaccordingtoinfectionoutcome.

Allinfections Success Failure Significance

N"_{43 43 28(65) 15(35)} Meanage 57 52 ns Stage0 0 0 0 Stage1 8 6(75) 2(25) ns Stage2 12 9(75) 3(25) ns Stage3 8 4(50) 4(50) ns Unknownstage 3 2(67) 1(33) Relapseorrecurrence 12 7(58) 5(42) ns Radiotherapy(1) 22 12(55) 10(45) ns Chemotherapy(2) 20 13(65) 7(35) ns Expander 34 25(74) 9(26) P:0,04 Prosthesis 9 3(33) 6(67) Earlyinfection 30 22(73) 8(27) ns Lateinfection 13 6(46) 7(54) ns

(1)Previousorintraoperativeradiotherapy;(2)PerioperativeChemotherapy (pre-infection);n.s.:notsignificant.

41%)didnotundergoanattempttorescuetheimplant,andwere managedwithimmediateimplantremovalandantibiotic treat-ment.Weobservedthatexclusionfrompocketsalvageattempts wasassociatedwithcurrentchemotherapy,whilethereversewas observedforradiotherapy.Therefore,inadditiontothelocalextent of theinfection the needto undergochemotherapyearly after breast reconstruction represents a compelling element that conditions clinical management of these patients. Accordingly, theneedforchemotherapyafterbreastimplant-based reconstruc-tionnot onlyrepresents anestablishedrisk factor forinfection (Reishetal.,2013;YiiandKhoo,2003),butalsonegativelyaffects thesubsequentmanagementoncebreastimplantinfectionoccurs. Whenconsideringthegroupofpatients whoweremanaged withanattemptatimplantpocketsalvageanoverallsuccessrateof 65%wasrecorded.Itshouldbenotedthatweincludedalsopatients withimplantskinexposureandalsoexcludedunproveninfections toavoidabiasofoverstatingtheeffectoftreatmentinsterilesites (SpearandSeruya,2010;ChunandSchulman,2007;Princeetal., 2012;Missanaetal.,2012).Therefore,theobservedoverallrateof implantpocketsalvageisencouragingcomparedtootherreports (Spear andSeruya, 2010;Bennett etal.,2011)and represents a positiveoutlookforeverydaypracticeinprovenimplantinfections. Interestinglywithinthisframe,asignificantlyincreasedsuccess rate was recorded (100% vs. 57%), when combined antibiotic treatment and implant replacement (AIR) was compared with systemic antibiotics treatment alone. This higher success rate comparedbettertootherreports(SpearandSeruya,2010;Yiiand Khoo,2003;Bennettetal.,2011; Albrightetal.,2016;Meybodi etal.,2015).

Whencomparedtomoreinvasivemanagement(e.g.removal and delayed reconstruction) requiring repeated hospitalization, AIR has the advantage of decreased hospitalization, decreased surgicalprocedures (1procedure vs. repeated procedures), and mayallowfortheprosecutionofchemotherapy.Indeedsuccessful infectedimplantsalvagewithAIRjustanticipatesthepositioning of thedefinitive prosthesis/implantwithout additional medical andeconomicburdens.

Inviewofthegreatersuccessrateinimplantpocketsalvage, which was observed for attempts performed on expanders as compared to definitive prosthetic devices and unrelated to an increasedexposuretoCTorRT,earlierreferralorself-referralor close follow-up of patients with prosthetic devices could be advised.

Inlinewithotherseries(SpearandSeruya,2010; Chunand Schulman,2007)weheredetectedawidespectrumof microbio-logical flora in infected implants including anaerobes, Gram negative rods, S. aureus and coagulase negative Staphylococci. Isolateswerenotassociatedtodifferentratesofsalvagefailure.A comparablesuccessrateoftreatmentwasobservedfor Staphylo-coccus epidermidis infections (both methicillin sensitive and resistant) compared to S. aureus. Overall, however, coagulase negativeStaphylococci representedsignificantly more common infectingagentcomparedtoS.aureusinagreementwithprevious

reports (Franchelli et al., 2015; Prince et al., 2012). Methicillin resistancewasnotafactornegativelyaffectingthepossibilityto rescuetheimplantpocketandsuccessfullytreatthepatient.Thisis incontrasttopreviousreports(Princeetal.,2012;Bennettetal., 2011),wherehoweveraveryhighnumber(33–70%)ofunproven infectionswereincluded,thuspossiblyoverestimatingfailuresin the presenceof methicillin resistant cocci.In viewof the high proportion of isolates from proven infections that were not covered by the antimicrobial spectrum of the prophylactic antibiotics,therecommendationforachangeofantibioticagent maybeindicatedaccordingtopreviousobservationsandcurrent recommendations (Berrios-Torres et al., 2017; Feldman et al., 2010).ThelimitednumberofGramnegativeinfections(18%),on the contrary, may not need additional changes in prophylactic treatment, but rather should be considered for guiding initial empirictreatmentwithcultureresultspending.

None of theantibioticregimens evaluatedhereshowed any outcomeadvantage,inlinewiththewidedispersionoftreatments andwiththedesignofthestudythatwasnotsuitedtothepurpose. Sincetherearetoourknowledgenostudiesorrecommendations onantibiotictreatmentinpatientswithbreastimplantinfection (Phillipsetal.,2016;Feldmanetal.,2010),thisidentifiestheneed forclinicalfocusinfuturework.

AccordingtothepresentobservationsofhighratesofMRSEand Coagulase-negativeStaphylococciandofpossibleGramnegative presence,empirictreatmentpendingculturesofsurgicalsamples couldbestincludepiperacillin/tazobactamandan anti-staphylo-coccalagentactiveonCoagulasenegativeStaphylococciandMRSA. The present analysis is limited by its retrospective and uncontrolleddesign.Controlledstudiesusingstandardizedclinical managementwillbeneededtoverifythepresentobservations.On theotherhand,thepresentlimitsandbiasalsoprovideauseful insight into the pitfalls of unstandardized clinicalpractice and outline the need for a more standardized approach to clinical management of breast implant pocket infections in cancer patients. On the other hand, even with these limitations, the studyhasthenoveltyofaddressingsomepointsthathavebeenso far neglected in clinical practice. There is a general lack of comprehensive analyses on the actual management of these patientsineverydaylifeinconditionswherestructured manage-ment,adviceorspecificprotocolsarenotavailable.Inthisrespect thepresentdatarepresenta first-in-kindwithseveralpointsof noveltyoverpreviousexperienceonidentificationand manage-mentofinfectedbreastimplants.Indeedthepresentstudyfarfrom representingasingle-surgeonseries,includesallthepatientsthat were sequentially seen at a single large center by a team of independent reconstructive surgeons. The study uses a careful classificationofinfection,includesonlymicrobiologicallyproven infectionsintheanalysis,andperformsforthefirsttimeadetailed analysisofeverydaymanagementdecisionsandoutcome.

Inconclusion, basedonthepresentanalysis, implantpocket salvageofprovenimplant infectionsafterbreast reconstruction representsa possibleoptionfor themajorityofpatients.Future

Table3

Comparisonbetweenimplantssavedusingantibiotictherapyplusimplantreplacementandimplantssavedusingantibiotictherapyalone.

Success Failure Significance

Antibiotic Antibioticandreplacement Antibiotic Antibioticandreplacement

blactam#inhibitoro#clindamicin 13 1 6 0 ns Glycopeptide#rifampicin/trimethoprim-sulfamethoxazole 3 1 3 0 ns

Levofloxacin 1 1 0 0 ns

Daptomicin+rifampicin 3 5 6 0 ns

Total 20 8 15 0 p=0,035

prospective work will be needed to avoid clinical bias in recruitment of eligible patients and to further improve and simplifytreatmentprotocols.

Conflictsofinterest

All authorsconcurredtotheworkand approved the manu-script.Therearenoconflictsofinterestforallauthorsactualor potential.Nooutsidefundingwasreceived.

Acknowledgements

TheauthorsareindebtedtotheotherplasticsurgeonsofPlastic and Reconstructive Unitof Policlinico SanMartino and Prof.D. Friedmanforprovidinghelpandforpatientreferral.

References

AlbrightSB,etal.One-stepsalvageofinfectedprostheticbreastreconstructions usingantibiotic-impregnatedpolymethylmethacrylateplatesandconcurrent tissueexpanderexchange.AnnPlastSurg2016;77(September(3)):280–5.

ArmstrongRW,BerkowitzRL,BoldingF.Infectionfollowingbreastreconstruction. AnnPlastSurg1989;23(October(4)):284–8.

BarrSP,etal.Infectionpreventioninbreastimplantsurgery–areviewofthe surgicalevidence,guidelinesandachecklist.EurJSurgOncol2016;42(May (5)):591–603.

BennettSP,FitoussiAD,BerryMG,CouturaudB,SalmonRJ.Managementofexposed, infectedimplant-basedbreastreconstructionandstrategiesforsalvage.JPlast ReconstrAesthetSurg2011;64(October(10)):1270–7.

Berrios-TorresSI,etal.Centersfordiseasecontrolandpreventionguidelineforthe prevention of surgical site infection, 2017. JAMA Surg 2017;152(August (8)):784–91.

Brand KG. Infection of mammary prostheses: a survey and the question of prevention.AnnPlastSurg1993;30(April(1)):289–95.

Chun JK, Schulman MR. The infected breast prosthesis after mastectomy reconstruction: successful salvage of nine implants in eight consecutive patients.PlastReconstrSurg2007;120(September(3)):581–9.

CordeiroPG,McCarthyCM.Asingle surgeon’s12-yearexperiencewithtissue expander/implant breast reconstruction:partII. Ananalysis of long-term complications,aestheticoutcomes,andpatientsatisfaction.PlastReconstrSurg 2006;18(September(4)):832–9.

CourtissEH,GoldwynRM,AnastasiGW.Thefateofbreastimplantswithinfections aroundthem.PlastReconstrSurg1979;63(June(6)):812–6.

FeldmanEM,KontoyiannisDP,SharabiSE,LeeE,KaufmanY,HellerL.Breastimplant infections: is cefazolin enough?. Plast Reconstr Surg2010;126(September (3)):779–85.

FranchelliS.Breastimplantinfectionsaftersurgicalreconstructioninpatientswith breastcancer:assessmentofriskfactorsandpathogensoverextended post-operativeobservation.SurgInfect(Larchmt)2012;13(June(3)):154–8.

FranchelliS,et al.Clinical andmicrobiological characterizationof latebreast implant infections after reconstructivebreast cancer surgery. Surg Infect (Larchmt)2015;16(October(5)):636–44.

GeorgiadeG,GeorgiadeN,McCartyJr.KS,SeiglerHF.Rationaleforimmediate reconstructionofthebreastfollowingmodifiedradicalmastectomy.AnnPlast Surg1982;8(January(1)):20–8.

KronowitzSJ,RobbGL.Radiationtherapyandbreastreconstruction:acritical reviewoftheliterature.PlastReconstrSurg2009;124(August(2)):395–408.

LaveauxC,PauchotJ,LouryJ,LeroyJ,TropetY.Acuteperiprostheticinfectionafter aestheticbreastaugmentation.Reportof threecasesofimplant“salvage”. Proposalofastandardizedprotocolofcare.AnnChirPlastEsthet2009;54 (August(4)):358–64.

MeybodiF,SedaghatN,FrenchJ,KeighleyC,MitchellD,ElderE.Implantsalvagein breastreconstructionwithsevereperi-prostheticinfection.ANZJSurg2015;87 (12):293–9.

Missana MC, Blot F, Germain M. Treatment of infected breast implant in reconstructivesurgery.AnnChirPlastEsthet2012;57(February(1)):16–24.

Momoh AO. A systematic review of complications of implant-based breast reconstructionwithprereconstructionandpostreconstructionradiotherapy. AnnSurgOncol2014;21(January(1)):118–24.

NahabedianMY, TsangarisT, Momen B,Manson PN.Infectious complications followingbreastreconstructionwithexpandersandimplants.PlastReconstr Surg2003;112(August(2)):467–76.

PhillipsBT.Areprophylacticpostoperativeantibioticsnecessaryforimmediate breastreconstruction?Resultsofaprospectiverandomizedclinicaltrial.JAm CollSurg2016;222(June(6)):1116–24.

PrinceMD.Prosthesissalvageinbreastreconstructionpatientswithperiprosthetic infectionandexposure.PlastReconstrSurg2012;129(January(1)):42–8.

PusicAL,CordeiroPG.Anacceleratedapproachtotissueexpansionforbreast reconstruction:experiencewithintraoperativeandrapidpostoperative expan-sionin370reconstructions.PlastReconstrSurg2003;111(May(6)):1871–5.

Reish RG, et al. Infection following implant-based reconstruction in 1952 consecutivebreastreconstructions:salvageratesandpredictorsofsuccess. PlastReconstrSurg2013;131(June(6)):1223–30.

RiegerUM.Bacterialbiofilmsandcapsularcontractureinpatientswithbreast implants.BrJSurg2013;100(May(6)):768–74.

surgeonsASop.Evidence-basedclinicalpracticeguideline:breastreconstruction withexpandersandimplants.2018.

SchusterRH,RotterS,BoonnW,EfronG.Theuseoftissueexpandersinimmediate breastreconstructionfollowingmastectomyforcancer.BrJPlastSurg1990;43 (July(4)):413–8.

SpearSL,SeruyaM.Managementoftheinfectedorexposedbreastprosthesis:a singlesurgeon’s 15-yearexperiencewith 69patients.PlastReconstr Surg 2010;125(April(4)):1074–84.

SpearSL,HowardMA,BoehmlerJH,DucicI,LowM,AbbruzzesseMR.Theinfectedor exposedbreastimplant:managementandtreatmentstrategies.PlastReconstr Surg2004;113(May(6)):1634–44.

TheEuropeanCommitteeonAntimicrobialSusceptibilityTesting–EUCAST.2018.

WeberJr.J,HentzRV.Salvageoftheexposedbreastimplant.AnnPlastSurg1986;16 (February(2)):106–10.

WilkinsonTS,SwartzBE,TorantoIR.Resolutionoflate-developingperiprosthetic breast infectionswithout prosthesisremoval.Aesthetic PlastSurg 1985;9 (2):79–85.

YiiNW,KhooCT.Salvageofinfectedexpanderprosthesesinbreastreconstruction. PlastReconstrSurg2003;111(March(3)):1087–92.