Clinical and Bacteriological Survey of Diabetic Foot Infections in Lisbon

Clinical

and

bacteriological

survey

of

diabetic

foot

infections

in

Lisbon

J.J.

Mendes

*

,

A.

Marques-Costa

,

C.

Vilela

,

J.

Neves

,

N.

Candeias

,

P.

Cavaco-Silva

,

J.

Melo-Cristino

a

InternalMedicineDepartment,HospitaldeSantaMarta/CentroHospitalardeLisboaCentralEPE,RuadeSantaMarta,50,1169-024Lisbon, Portugal

b_{PodiatryOutpatientClinic,Associac¸a˜oProtectoradosDiabe´ticosdePortugal,Lisbon,Portugal}

c_{ImmunologyandBacteriologyLaboratory,FaculdadedeMedicinaVeterina´riadaUniversidadeTe´cnicadeLisboa,Lisbon,Portugal} d_{SurgeryDepartment,HospitaldeSantoAnto´niodosCapuchos/CentroHospitalardeLisboaCentralEPE,Lisbon,Portugal} e_{EndocrinologyDepartment,HospitaldeCurryCabral,Lisbon,Portugal}

f

InstitutoSuperiordeCieˆnciasdaSau´de–EgasMoniz,Caparica,Portugal

g_{InstituteofMicrobiology,FaculdadedeMedicinadeLisboa,Lisbon,Portugal}

a

r

t

i

c

l

e

i

n

f

o

Articlehistory: Received31July2011 Receivedinrevisedform 16September2011 Accepted3October2011

Publishedonline21October2011 Keywords: Epidemiology Diabeticfoot Infection Microbiology Portugal

a

b

s

t

r

a

c

t

Aims:Anepidemiologicalsurveyofdiabeticfootinfections(DFIs)inLisbon,stratifyingthe bacterialprofilebasedonpatientdemographicaldata,diabeticfootcharacteristics(PEDIS classification),ulcerdurationandantibiotictherapy.

Methods:Atransversalobservationalmulticenterstudy,withclinicaldatacollectionusinga structuredquestionnaireandmicrobiologicalproducts(aspirates,biopsiesorswabs col-lectedusingtheLevinemethod)ofclinicallyinfectedfootulcersofpatientswithdiabetes mellitus(DM).

Results: Forty-ninehospitalizedandambulatorypatientswereenrolledinthisstudy,and 147microbialisolateswerecultured.Staphylococcuswasthemaingenusidentified,and methicillin-resistantStaphylococcusaureus(MRSA)waspresentin24.5%oftotalcases.Inthe clinicalsamplescollectedfrompatientsundergoingantibiotictherapy,93%oftheantibiotic regimenswereconsideredinadequatebasedontheantibioticsusceptibilitytestresults.The averagedurationofanulcerwithanyisolatedmulti-drugresistant(MDR)organismwas29 days,andprevioustreatmentwithfluoroquinoloneswasstatisticallyassociatedwith multi-drugresistance.

Conclusions: StaphylococcusaureuswasthemostcommoncauseofDFIsinourarea. Preva-lenceandprecocityofMDRorganisms,namelyMRSA,werehighandwereprobablyrelated toprevious indiscriminateantibiotic use.Clinicians should avoidfluoroquinolonesand morefrequentlyconsidertheuseofempiricalanti-MRSAtherapy.

#_{2011ElsevierIrelandLtd.Allrightsreserved.}

*Correspondingauthor.Tel.:+351934855822;fax:+351213560368. E-mailaddress:joaojoaomendes@hotmail.com(J.J.Mendes).

Abbreviations: CRTB,clinicallyrelevanttissueburden;DFI,diabeticfootinfection;DFU,diabeticfootulcer;DM,diabetesmellitus;ESBL, extended-spectrumb-lactamase;HCP,healthcareprovider;MDR,multi-drugresistant;MRSA,methicillin-resistantStaphylococcusaureus; PDR,pan-drugresistant.

ContentsavailableatSciverseScienceDirect

Diabetes

Research

and

Clinical

Practice

j o u r n a lh o m e p a g e :w w w . e l s e v i e r . c o m / l o c a t e / d i a b r e s

0168-8227/$–seefrontmatter#_{2011ElsevierIrelandLtd.Allrightsreserved.} doi:10.1016/j.diabres.2011.10.001

1.

Introduction

Diabetes mellitus (DM) isa serious health problem that is rapidlyexpandingworldwide[1]. Oneofthemorefrequent diabeticcomplicationsisdiabeticfoot,whichresultsfroma complex interaction between a number of risk factors. Neuropathy(with alterations inmotor, sensitive and auto-nomic functions) has a central role, causing ulcerations becauseof traumaor excessivepressureon deformedfeet thatlackprotectivesensitivity[2].Oncetheprotectivelayerof skin is broken, the deep tissues are exposed to bacterial colonization. Infections are facilitated by immunological deficits(especiallyinneutrophils),whicharerelatedtoDM, andtheyrapidlyprogresstothedeeptissues.PatientswithDM frequentlyrequireminorormajoramputationsofthelower limbs(15–27%),andinmorethan50%ofcases,infectionisthe preponderantfactor[2].

Staphylococcus aureus is the most prevalent isolate in diabetic foot ulcers (DFUs), together with other aerobes (includingStaphylococcusepidermidis,Streptococcusspp., Pseudo-monasaeruginosa,Enterococcusspp.andcoliformbacteria)and anaerobes [3,4]. Because of the polymicrobial nature of diabetic foot infections (DFIs), Karchmer and Gibbons [5] questioned the need for precisely defining the causative microorganismandsuggestedatreatmentstrategybasedonly ontheknowledgeofthegeneralepidemiology.Morerecently, an increasein the incidence ofmulti-drug resistant (MDR) organisms,namelymethicillin-resistantS.aureus(MRSA)and extended-spectrumb-lactamase(ESBL)-producing gram-neg-ativebacteria,isthreateningtheoutcomeofanti-infectious therapy inthe communityand inhospitalizedpatients[4]. Therefore,thecurrent guidelines[6] andexpert opinion [7] advise providers to obtain specimens for culture before initiatingempiricantibiotictherapytohelpwiththeselection ofadefinitivetherapy.

AlthoughPortugalhasoneofthehighestprevalencesof DM,lowerextremityamputations[8]andMRSAskinandsoft tissueinfections[9]inEurope,thereisvirtuallynodataonthe prevalence and characterization of DFIs. Therefore, we performed an epidemiological survey of DFIs in Lisbon, stratifyingthebacterialprofilesbasedonpatient demograph-icaldata,characteristicsofdiabeticfoot(PEDISclassification), ulcer duration and current and recent (3 months prior) antibiotictherapy.

2.

Subjects,

materials

and

methods

This transversal observational study was conducted at 4 clinicalcenters(2outpatientclinics,1generalsurgeryward and1vascularsurgeryward)inLisbonfromJanuary2010to June2010.Astructuredquestionnairewasdevelopedtorecord medical histories, examination details and investigation reports by health care providers (HCPs). Specimens were collectedfrompatientswithDMandclinicallyinfectedfoot ulcers,asadvisedbycurrentclinicalguidelines[6].ADFUwas definedasafull-thicknesswoundbelowtheankleinadiabetic patient, irrespective ofduration[10]. Infectionwas defined clinically by symptoms and signs of inflammation, as

described by the infection item on the PEDIS system [10]. Specimenswereobtainedfrompatientsbeforethefirstdoseof antibioticsorwhileunderantibiotictherapywithprogression ofinfectionsignsandclinicaldeteriorationoftheulcer.

ThisstudywasapprovedbytheFacultyofMedicineofthe University of Lisbon Research Ethics Committee and the PortugueseDataProtectionAuthority,andwritteninformed consentwasobtainedforeverypatient.

2.1. Clinicalcharacterization

Forclinicalcharacterization,9studyfactorswererecordedfor eachpatient:age,gender,DMduration(fromdiagnosis),last HbA1c value (accepted if collected in the last 3 months), hypertensionand dyslipidemia(asdefinedaccordingtothe American Diabetes Association (ADA) guidelines for the diabetic population [11]), active tobacco abuse (defined as 20packsinthepreviousyear),presenceofischemicheart disease(definedasprevioushistoryofmyocardialinfarction, coronary artery bypass graftor percutaneous transluminal coronary angioplasty) and chronic renal failure (defined as calculated glomerularfiltration rate<30mLmin 1_1.73m 2_,

permanentrenalreplacementtherapyorprevioustransplant).

2.2. Diabeticfootcharacterizations

Forcharacterizationofdiabeticfoot,weusedtheInternational WorkingGroupoftheDiabeticFootPEDISsystem[10],which classified all foot ulcers in subcategories of five main categories(perfusion,extent/size,depth/tissueloss,infection andsensation),accordingtostrictcriteria.Forthedefinitionof osteomyelitis,aminimumofapositiveprobe-to-bonetest[12] wasaccepted,butclinicianswereencouragedtosubstantiate their diagnosis with the appropriate imaging studies. The numberofpreviousulcersandpreviousminor(toeorpartof the foot) or major(above the ankle)amputations was also recorded.

2.3. Antibiotictherapy

HCPswereaskedtoregisterallcurrentandrecent(overthe previous3months)antibiotictherapies.

2.4. Collectionofsamples

All HCPs were instructed on the proper methods for the collection of culture material, and a written protocol was provided.Inthecaseofabscesswithintactintegument(and other closed lesions), the protocol suggested sampling by needle aspirationunder strictaseptic technique. Forulcers and other open wounds, biopsy specimens were required, except in situations where the HCP considered that the invasive procedure could place the patient at risk (pain induction or risk of enlarging the ulcer). In only these situations,superficialswabsampleswereaccepted,instrict accordancewiththeNationalInstituteforHealthandClinical Excellence diabetic foot guideline [6]. For either of the procedures, debridement of necrotic tissue and cleansing with simple saline before sampling was obligatory. For

biopsies, shaving or punch techniques, as previously de-scribed [13], were required.For swab sampling,HCPs were instructed on a standardized procedure [14], based on the Levine 1cm2 _{swab method, using a flocked swab (ESwab}

CollectionSystem,Copan).

2.5. Transport

Aspirates were transported in buffered isotonic agar with reductionagentmedia(Port-A-CulVial,BDBBL),andbiopsies andswabsweretransportedinmodifiedliquidAmiesmedium (ESwabPreservationSystem,Copan).Transporttothe labora-tory(MicrobiologyLaboratory,FacultyofVeterinaryMedicine, TechnicalUniversityofLisbon)within2hofcollectionwas assuredbyanon-callexpresscourier.

2.6. Processingandmicrobiologicalanalysisofwound specimens

Standardmethodsforsampleprocessingandisolationand identificationofaerobicandanaerobicbacteriawereused[15]. Biopsy sampleswere weighed to the nearest milligram in sterilePetridishesandhomogenizedinPBSusingapearljar.A 100-mLvolumeofthehomogenatewasusedforserialdilutions inPBS.Foraspiratesamples,a100-mLvolumeoftherecovered fluid was directly used for serial dilutions in PBS. Swab sampleswerevortexedwiththeswabinsidefor5s,andthena 100-mLvolumeofthesuspensionwasusedforserialdilutions inPBS.Quantificationwasperformedusingthe10-foldserial dilution method [15], and 100mL of each dilution was inoculatedontoMacConkeyagar(Merck)/ColumbiaANCagar with 5% sheep blood (BioMe´rieux) and, in duplicate, in Schaedleragarwith 5%sheepblood (BioMe´rieux).Thefirst twoplateswereincubatedunderaerobicconditionsat358C for 24–48h, and the two Schaedler plates were incubated underanaerobicconditions(AnaerocultA,Merck)for48–96h. Additionally,sampleswereinoculatedinBrainHeartInfusion Broth (Difco, BHIB) to allow recovery offastidious or low-concentrationorganisms.Isolateswereidentifiedbystandard methods [15]. In some instances, unusual strains were identified using partial 16S rRNA gene sequencing [16]. Antimicrobial susceptibility testing of the aerobic isolates wasperformedusingthestandarddiscdiffusionmethod,as recommended by the Clinical and Laboratory Standards Institute [17]. Quantitative results were expressed in CFU/ mLforneedleaspirationsamples,CFU/gforbiopsysamples andCFU/cm2_{forswabsamples.Consistentwiththestudyby}

Billetal.[18]andtheresultsofarecentsystematicreview[19], aswabcountof>105_CFU/cm2_{wasconsideredequivalenttoa}

tissuecountof>105CFU/goraneedleaspirationsampleof >105CFU/mL;allofthesevaluesareconsideredtorepresenta clinicallyrelevanttissueburden(CRTB).

2.7. Multidrugresistanceprofiles

Methicillin-resistantS.aureus(MRSA),methicillin-resistantS. epidermidis(MRSE)andothercoagulase-negativeStaphylococcus spp.(MRCN)weredefinedasstrainsphenotypicallyresistant to cefoxitin (by the disc diffusion method). Vancomycin-resistantEnterococcusspp.(VRE)weredefinedasstrainsthat

werephenotypicallyresistanttovancomycin. (ESBL)-produc-ing gram-negative strains were phenotypically confirmed using the cephalosporin/clavulanate combination disc test [20].Multi-drugresistant(MDR)P.aeruginosaandAcinetobacter baumanniistrainsweredefinedasthoseresistanttoatleast three of six antibiotics, including amikacin, gentamicin, ciprofloxacin, piperacillin,ceftazidimeand imipenem. Pan-drugresistant(PDR)P.aeruginosaandA.baumannii/calcoaceticus strainsweredefinedasthosesensitiveonlytocolistin[21].All ofthesestrains(MRSA,MRCN,VRE, [ESBL]-producing gram-negative bacteria, and MDR and PDR P. aeruginosa and A. baumannii/calcoaceticus)wereconsideredtobeMDRorganisms.

2.8. Statisticalanalyses

Qualitative variables were expressed as percentages, and quantitativevariablesareexpressedasmeansSD(standard deviation).Significanceofthestudyvariableswastestedusing Student’s t-test, the Chi-square test or Fisher’s exact test, where appropriate. A two-tailed p value of <0.05 was considered to be statistically significant. Additionally, the ulcerduration(indays)wasstratifiedbymicrobialisolateand visuallysummarizedinaboxplot,withtheboxesrepresenting thelowerandupperquartiles,theverticallinethemedian,the barstheminimumandmaximumdatapoints,andthesolid diamondsymbolthemean.

3.

Results

Atotalof49patients(meanageof62.712.7yearsanda male-to-femaleratioof6.8)wereadmittedduringthestudyperiod. Their clinical and diabetic foot characteristics,stratified in accordancewiththesamplecollectionmethod,areshownin Table1.Amongthesepatients,themeandurationofDMwas 23.012.8 years, 26.5% had HbA1c levels <58mmol/mol (<7.5%), >90% had hypertension and/or dyslipidemia, and 30.6%and10.2%hadischemicheartdiseaseandchronicrenal failure,respectively.Two-thirds ofthepatients had under-gone recentantibiotictherapy,andone-thirdwascurrently undergoingantibiotictherapy.Themajorityofthe samples camefromoutpatients(65.3%),andswabbingwasthemost commonly used method (63.3%) for sample collection. However, 92.8% of hospitalized patients and all clinically suspectedosteomyelitispatientshadsamplescollectedbyan invasivetechnique.Therewerestatisticallysignificant differ-ences in the isolation rates of microorganisms from deep tissuesamplesandsuperficialswabs,withfeweraerobesper sample, in particular gram-positive bacteria (2.31.0 vs. 1.31.2),isolatedfromswabs,buttherewasnodifferencein theisolationrateofanaerobesorMDRorganisms.

Outofthe49patientsenrolledinthisstudy,147microbial isolates(comprising 43species)werecultured,which repre-sentsanaverageof3.01.4organismspersample. System-aticresultsarepresentedinTable2.Aerobeswerepresentin 98.0%ofcases,withgram-positivebacteriacomprising66.0% ofthetotalnumberofisolates.Staphylococcus wasthemain genusidentified,withS.aureuspresentin51%ofthesamples and in94.1%ofthecaseswithaCRTB.Coagulase-negative Staphylococcusspp.werethesecondmostfrequently

encoun-teredaerobicgram-positive isolates,withS.epidermidisand StaphylococcuslugdunensiscommonlyassociatedwithaCRTB. Corynebacterium spp. and other uncommon gram-positive bacteriawerealsoidentifiedbut notinclinicallysignificant quantities.Streptococcusspp.wereinfrequently(4.1%)isolated. Gram-negative aerobes comprised 19.0% of the isolated organisms,whileP.aeruginosa,thesinglemostpredominant species,wasisolatedinonly12.2%ofcases.Proteusspp.were thenextmostfrequentlyrecoveredgram-negativebacteria, although largely (75.0%) in non-CRTB cases. A. baumannii/ calcoaceticuswereidentifiedin8.2%ofthecasesandwerethe

non-PDR speciesfoundexclusively inthe non-CRTB cases. Anaerobeswerefoundin30.6%ofpatients,with Peptostrepto-coccus spp. accounting for 55.0% of all anaerobic isolates, followedbytheBacteroidesfragilisgroup,whichaccountedfor 25%oftheseisolates,butthislastgroupwasmorefrequently identified in non-CRTB. Candida spp. were infrequently encountered,representingonly1.4%ofthetotalisolates.

MDRorganismswerepresentin38.8%ofcases,whileMRSA was found in 24.5% of patients, thereby making it the predominantlyisolatedpathogen.MRSEandother methicil-lin-resistantcoagulase-negativeStaphylococciwerealso

iden-Table1–ClinicalandmicrobiologicalcharacteristicsofDFIsstratifiedbythesamplecollectionmethod.

Total(n=49) Swabsamples(n=31) Deeptissuesamplesa_(n=18)

Hospitalization(%) 34.7 12.9 72.2

Demographicaldata

Age(years) 62.712.7 60.213.5 67.010.1

Malegender(%) 83.7 87.1 77.8

Diabetesmellitus

Controlofdiabetes(HbA1c<7%) 20.4% 16.1% 17.8%

Duration(years) 23.012.8 22.512.8 23.713.1

Co-morbidities

Hypertension(%) 93.9 96.8 88.9

Dyslipidemia(%) 95.9 93.4 100

Activetobaccoabuse(%) 38.7 32.2 50.0

Organlesions

Ischemicheartdisease(%) 30.6 35.5 22.2

Chronicrenalfailure(%) 10.2 12.9 5.6

Diabeticfootcharacterization

Numberofpreviousulcers 1.61.5 1.91.6 1.21.2

Previousamputation(%) 46.9 51.6 38.9

Major 10.2 9.7 11.1

Minor 38.8 45.2 27.8

Durationofpresentulcer(days) 30.631.9 33.425.9 25.740.5

Neuroischemic(%) 53.1 54.8 50.0 Osteomyelitis(%) 30.6 0.0 83.3 PEDIS Perfusion 1(%) 44.9 43.9 46.6 2(%) 40.8 40.7 41.0 3(%) 14.3 19.3 12.4 Extent(cm2_{) 13.356.9 1.20.6 34.391.7} Depth 1(%) 18.4 29.0 0.1 2(%) 51.0 71.0 16.6 3(%) 30.6 0.0 83.3 Infection 2(%) 61.2 87.1 16.6 3(%) 36.7 12.9 77.7 4(%) 2.0 0.0 5.4 Sensation 2(%) 100 100 100 Antibiotictherapy Previous(%) 65.3 67.7 61.2 Current(%) 30.6 23.0 43.7 Isolates Monomicrobial(%) 16.3 12.9 22.1

Totalnumber(persample) 3.01.4 3.21.3 2.71.4

Aerobes 2.51.1 2.70.9 2.31.3

Gram-positive 2.01.0 2.31.0 1.61.1

Gram-negative 0.60.6 0.50.5 0.70.7

Anaerobes 0.40.6 0.40.6 0.30.6

MDRorganisms 0.60.9 0.50.8 0.91.0

a _{Biopsies(n=14)andaspirates(n=4).}

tified but accounted for only 4.8% of the isolates. Gram-negativeMDRorganismswereidentifiedinatotalof18.9%of the patients.Ofthe isolated A.baumannii and P.aeruginosa strains,38.5%werePDR,andtheremainderwereMDR.

Althoughalongitudinalstudyusingsequential microbio-logicalsampleswasnotperformed,visuallyrepresentingthe relationshipbetweenthemicrobialisolatesandulcerduration inaboxplotgraph(Fig.1)revealedapattern:gram-positive bacteriaappearedinulcersofshortduration,whileanaerobes associatedwitheithergram-positiveor-negativeorganisms appeared in ulcers of longer duration. This finding was independent of previous or current antimicrobial therapy. The average duration of an ulcer with any isolated MDR organismwas29days.

Intheclinicalsamplescollectedfrompatientsundergoing antibiotictherapy (Table3), which correspondedmainly to hospitalizedpatientswithosteomyelitis,93%oftheantibiotic regimenswereconsideredinadequatebasedontheantibiotic susceptibilitytestresults.Quantitativeandqualitative differ-enceswerefoundinthesesamples,withfewer microorgan-isms identified (2.10.9 vs. 3.41.3); in particular, fewer gram-positive(86.7%vs.100%)andanaerobic(6.7%vs.41.2%)

bacteriawereidentified;however,therewasahigher preva-lence ofMDRorganisms(66.7%vs. 26.5%).Althoughall the clinicalvariableswereexamined,multi-drugresistancewas onlystatisticallyassociatedwithcurrentantibiotictreatment (withanyclassofantibiotics)andwithprevious fluoroquino-lonetreatment(Table4).

4.

Discussion

DFIsarecommon,complex,andcostly.Theyaccountforthe largest numberofproximatenontraumatic lowerextremity amputations [2]. This publichealth problem isparticularly importantintheunderdiagnosedandundertreateddiabetic Portuguesepopulation[8].Toourknowledge,thisisthefirst publishedepidemiological studythat reportsthe infectious microbiota and clinical characteristics of diabetic foot in patientslocatedinPortugal.Thisstudyreflectsthe clinical profilesofinpatientsandoutpatientsintheLisbonarea,but becausethesamplewasrelativelysmall,thestudypopulation was heterogeneous,andsomecontroversialmethodological issueswereutilized(notably,theuseofswabsand

quantita-Table2–DistributionoftheDFIisolates.

n % %(/patients) CRTB

Aerobes 125 85.0 98.0 63.2%

Gram-positive 97 66.0 95.9 64.9%

Staphylococcusspp. 54 36.7 79.6 66.7%

Staphylococcusaureus(MRSA) 32(17) 21.8(11.6%) 51.0(24.5%) 93.8%(94.1%) Staphylococcusepidermidis(MRSE) 7(3) 4.8(2.0%) 14.3(4.1%) 42.9%(66.7%) Othercoagulase-negative Staphylococcusspp.(MRCN)a _{15(3) 10.2(2.0%) 20.4(4.1%) 20.0%(33.3%)}

Streptococcusspp.b _{6 4.1 12.2 100%} Enterococcusspp.c_{(VRE) 13(1) 8.8(0.7%) 20.4(2.0%) 76.9%(100%)} Corynebacteriumspp.d _{12 8.2 28.6 50.0%} OtherGram-positivese _{12 8.2 22.4 41.7%} Gram-negative 28 19.0 51.0 57.1% Enterobacteriaceae 16 10.9 16.3 56.3% Escherichiacoli 1 0.7 2.0 100% Klebsiellaspp.(ESBL) 2(1) 1.4(0.7%) 4.1(2.0%) 100%(100%) Proteusspp.f _{8 5.4 16.3 25.0%} OtherEnterobacteriaceaeg _{5 3.4 4.1 80.0%}

Nonfermentingnegativebacilli 12 8.2 20.4 58.3%

MDRPseudomonasaeruginosa(PDR-PA) 7(2) 4.8(1.4%) 12.2(4.1%) 71.4%(100%) MDRAcinetobacterbaumannii/calcoaceticus(PDR-AB) 5(3) 3.4(2.0%) 8.2(6.1%) 40.0%(66.7%)

Anaerobes 20 13.6 30.6 75.0%

Peptostreptococcusspp. 11 7.5 22.4 100%

Bacteroidesfragilisgroup 5 3.4 4.1 20.0%

Otheranaerobesh _{4 2.7 4.1 75.0%}

Yeastsi _{2 1.4 4.1 –}

Inbracketsarethemulti-drugresistant(MDR)organismsofeachspecies;CRTB:clinicallyrelevanttissueburden;MRSA:methicillin-resistant Staphylococcusaureus;MRSE:methicillin-resistantStaphylococcusepidermidis;MRCN:methicillin-resistantcoagulase-negativeStaphylococcusspp. otherthanStaphylococcusepidermidis;VRE:vancomycin-resistantEnterococci;ESBL:extended-spectrumbeta-lactamasesproducing Enterobacter-iaceae;MDR:multi-drugresistant;PDR-PA/PDR-AB:pan-drug-resistantAcinetobacterbaumannii/pandrug-resistantPseudomonasaeruginosa.

a _{Staphylococcuslugdunensis(n=2)andothercoagulase-negativeStaphylococcusspp.(n=13).} b _{Streptococcusagalactiae(n=3),Streptococcusmitisgroup(n=1)andStreptococcusdysgalactiae(n=2).} c _{Enterococcusfaecalis(n=9)andEnterococcusfaecium(n=1).}

d _{Corynebacteriumamycolatum/striatum(n=9)andotherCorynebacteriumspp.(n=3).}

e _{Dermabacter hominis (n=1), Leuconostoc spp. (n=1), Arcanobacterium spp. (n=2), Arthrobacter spp. (n=1), Kocuria varians/rosea (n=2),}

Cellulomonasspp./Micrococcusspp.(n=1)andBrevibacteriumspp.(n=4).

f _{Proteusmirabilis(n=4)andProteusvulgaris(n=4).}

g _{Enterobacterspp.(n=1),Serratiamarcescens(n=2)andMorganellamorganii(n=2).}

h _{Fusobacteriumspp.(n=1),Prevotellaspp.(n=1),Eggerthellaspp.(n=1)andVeinonellaspp.(n=1).} i _{Candidaalbicans(n=1)andCandidaparapsilosis(n=1).}

tive results), care must be taken when interpreting these results.

Thebaselinecharacteristicsofthesamplepopulationare in line with those previously reported by European DFU studies[22],exceptforthehighpercentageofmalepatients and low percentage of patients with controlled DM (as evaluatedbyHbA1c).Thiscanbepartiallyexplainedbythe hypothesis of a recent study [23] that reported that male genderandpoorglycemiccontrolareindependentriskfactors forinfectionandnon-healingDFUs.Thehighprevalenceof co-morbiditiesisduetothelowcut-offsusedinthedefinitions.

Clinical guidelines [6] use infection severity and other clinicalcharacteristicsofDFUsasthebasisforselectingan appropriatetreatmentapproach,includingantibiotictherapy. Our studyusedthe PEDISclassification,andtherewereno statisticalrelationshipsbetweenthediabeticfoot character-istics, other than the duration of the ulcer and a clinical suspicionofosteomyelitis,andspecificpathogens.Wecannot becertainthatthelackofsignificantassociationswasdueonly tothesmallsamplesize,however.

Itiswelldocumentedintheliterature[3,4]thatDFIsare polymicrobialinnature.Inthepresentstudy,polymicrobial Fig.1–Aboxplotrepresentingtheulcerdurationdata(indays),stratifiedbythemicrobialisolate(theboxesrepresentthe lowerandupperquartiles,theverticallinethemedian,thebarstheminimumandmaximumdatapoints,andthesolid diamondsymbolthemean).MDROs:multi-drugresistantorganisms,GP:gram-positiveaerobes,GN:gram-negative aerobes,andA:anaerobes.

Table3–DistributionoftheDFIisolatesinrelationtocurrentantibiotictherapy.

Total (n=49)

Notunderantibiotic therapy(n=34) Underantibiotic therapy(n=15) pa Hospitalization(%) 34.7 17.6 73.3 <0.01 Isolates

Totalnumber(persample) 3.01.4 3.41.3 2.10.9 <0.01

Aerobes

Numberpresentpersample 2.51.1 2.91.0 1.91.0 <0.01

Sampleswith1(%) 98.0 100 93.3 NS

Gram-positive

Numberpresentpersample 2.01.0 2.31.0 1.51.1 0.02

Sampleswith1(%) 95.9 100 86.7 0.03

Gram-negative

Numberpresentpersample 0.60.6 0.60.6 0.40.6 NS

Sampleswith1(%) 51.0 58.8 33.3 NS

Anaerobes

Numberpresentpersample 0.40.6 0.50.6 0.10.5 NS

Sampleswith1(%) 30.6 41.2 6.7 0.01

MDRorganisms

Numberpresentpersample 0.60.9 0.40.7 1.11.0 <0.01

Sampleswith1(%) 38.8 26.5 66.7 <0.01

Antibiotictherapycoversisolatedpathogens – – 7.0%b _–

a _{Notunderantibiotictherapyvs.underantibiotictherapy.} b _{Ofthetotalofpatientscurrentundergoingantibiotictherapy.}

cultureswereobtainedfrom83.7%ofpatientswitharateof isolationof3.01.4bacteriaperpatient,independentofthe sampling method, which is similar to the results seen in previous studies. In agreement with published western studies[3,4],weisolatedpredominantlyaerobicgram-positive cocci from acute infections, while a more complex flora, includinggram-negativeandanaerobicbacteriawasobtained fromchronicwounds.

WealsofoundthatS.aureus,eitheraloneorasacomponent of a mixed infection, to be the most frequently isolated pathogen. Coagulase-negative Staphylococcus spp. were also frequentlyfound,oftenwithamethicillin-resistance pheno-type.Streptococcusspp.,whicharewell-recognizedpathogens in DFIs, were infrequently isolated. This can be partially justified by the high prevalence of present and recent antibiotictherapy.Enterococcusspp.,consideredlow-virulence commensalorganisms,exceptindiabeticandother compro-misedpatients,wereidentifiedin20.4%ofpatients,whichisin accordancewithotherstudies[3,4].

Instrictaccordance withotherwesternstudies[3,4],but unlikestudiesfromIndiaandotherAsiancountries[24],we isolatedrelativelyfewaerobicgram-negativeorganisms.

Inourstudy,thehighpercentageofP.aeruginosaandlow percentageofProteusspp.isolateswithaCRTBwasconsistent withtheviewthatthefirstspeciescancauseseveretissue damageinDMpatientsandshouldberegardedassignificant inthatpopulation,whilethelatteraremostcommonly non-pathogenic[7].

Independent of the sampling method, anaerobes were isolated inone-third of the patientsand almost always in mixedculture.Thisisincontrasttothefindingsofseveral otherstudiesthatfailedtoisolateanaerobes,possiblybecause ofsuboptimal study protocols [25]. The anaerobes isolated fromourstudyareconsistentwithotherreportedstudies[26], inwhichPeptostreptococcusspp.werethepredominantisolates.

AlthoughtheexactroleofanaerobicbacteriainDFIsisstill underdebate,ourstudyisinlinewiththeexpertopinion[7] that suggeststhatanaerobes aremorelikelytobeisolated fromlong-standinginfections.

Otherimportantfactorstoconsiderwheninterpretingthe resultsofourstudyarethatDFIisaclinicaldiagnosisandthat both the quantitative and qualitative aspects of wound microbiology are critical determinants of an infection’s course.All thepatientsenrolledinourstudyhadclinically infectedDFUs,andwebasedourconclusionsonaqualitative microbiological analysis, considering the diversity of the microorganisms and the potential for microbial synergy, andonquantitativemicrobiologicalanalysis,whichprovided a goodindication ofthemicrobial load. Assumingthatthe qualitativemicrobiologyremainsconstant,theprobabilityof woundinfection increaseswiththemicrobial load, uptoa critical level at which infection or a failure to heal is considered to be almost inevitable. In this paper, CRTB represented thequantitative aspectofwoundmicrobiology andwasusedonlyasapotentialindicatorofthe microorgan-isms’relevanceinclinicallyinfectedDFUs.

One of the main limitations of our study is that the quantitative and qualitative microbial evaluations were predominantlyperformedusingswabsamples.Whiletissue biopsiesandfluidaspiratesareconsideredthegoldstandard fordiagnosingwoundinfections[25],theseinvasivetestsare performed infrequently with small wounds and in many practicesettings,suchasoutpatientclinics,duetoconcerns over enlargingthe ulcer or inducing pain [14,25,27]. In our study, we introduced a standardized procedure that was strictlyconsistentwiththecurrentclinicalguidelines[6].Our method used quantitative aerobic and anaerobic swab cultures as an alternative methodwhen the HCP believed aninvasiveprocedurewouldplacethepatientatrisk.While thisdecisionwasbasedonthemicrobiologicalexperimental

Table4–RelationshipbetweenMDRorganismsandrecent(=3months)orcurrentantibiotictherapy.

Non-MDR(n=30) MDRa_{(n=19) p}b

Previousantibiotictherapy 63.3% 73.7% NS

Penicillins(includingassociationswithb-lactamaseinhibitors) 63.3% 79.0% NS

Cephalosporins 13.3% 26.3% NS Carbapenems 10.0% 5.3% NS Aminoglycosides 0.0% 0.0% NS Sulphamides 13.3% 15.8% NS Fluoroquinolones 23.3% 63.2% <0.01 Glycopeptides 6.7% 5.3% NS Oxazolidinones 0.0% 5.3% NS Others 3.3% 5.3% NS

Currentantibiotictherapy 16.7% 52.6% <0.01

Penicillins(includingassociationswithb-lactamaseinhibitors) 6.7% 0.0% NS

Cephalosporins 0.0% 0.0% NS Carbapenems 10.0% 15.8% NS Aminoglycosides 0.0% 5.3% NS Sulphamides 3.3% 0.0% NS Fluoroquinolones 10.0% 15.8% NS Glycopeptides 3.3% 5.3% NS Oxazolidinones 0.0% 5.3% NS Others 0.0% 5.3% NS

Coverstheisolatedpathogens 40.0% 0.0% 0.03

a _{MRSA,MRSE,MRCN,VRE,ESBL-producingnegatives,PDRPseudomonasaeruginosaandPDRAcinetobacterbaumannii/calcoaceticus.} b _{Non-MDRvs.MDR.}

and clinical evidence supporting the hypothesis that the resultsform quantitative swabs arehighly correlated with thosefrominvasiveprocedures (sensitivitiesfrom93.5%to 100% and specificities from 76.3% to 94.2% have been previouslyreported [14]), this hypothesisisnot consensual in the scientific community. Some authors have reported consistency between swab and deep tissue biopsy sample cultures [28,29], while others believe that superficial swab cultures of DFIs only complicate patient evaluation by sampling the superficial wound compartment, which may contain colonizing organisms rather than true pathogens. Thesedivergentconclusionsmaybeexplainedby different andnon-standardizedprotocols.Whileweacknowledgethata standardizedquantitativeswabsamplingprotocolmaybean imperfectanddifficult-to-implementmethodintheclinical setting,itclearlyhasmeritsintheresearchfield,atleastina settingwithahighprevalenceofthe multi-drugresistance settingsuchasinourstudy;whenproperlyinterpreted,they canprovideusefulinformation[27].

We had a surprisingly high number of swab samples (mainly from outpatient clinics) from patients with small superficialulcers.Therewerestatistically significant differ-encesbetweenthesuperficialanddeepsamples,probablydue toswab-associatedandimpossible-to-eliminatewound con-tamination by members of the endogenous microbiota (mainlygram-positiveaerobes).Thisresultmayexplainthe high prevalence of Corynebacterium spp. and other low-virulencecolonizers(e.g.,DermabacterhominisandLeuconostoc spp.),whichweremainlyculturedfromswabsamples.

Inthepresentstudy,MDRorganismswereculturedfrom 38.8% of the patients, the majority (24.5%) of which were MRSA. Most of the other international studies that have reportedasimilarlyhighpercentageofMDRorganismswere single-center,hospital-basedstudies[24].Thehighprevalence insuchstudiesmaybeexplainedbytheinstitution’suseof broadspectrumantibiotics,resultinginapathogen-selective survivaladvantage.Inourmulticenterstudy,wedidnotfind anystatisticallysignificantdifferencesbetweentheinpatients andoutpatients,andthemeandurationofulcerswithisolated MDRorganismswasshort(29days).

Wealsofoundahighpercentageofpatients(65.3%)who hadreceivedantibioticsinthepreviousthreemonthsanda statisticalassociationbetweenthe presenceofMDR organ-isms and previous fluoroquinolone therapy. This class of antibioticshasbeenwidelyusedinPortugalformanyyears [30],andothershavedescribed[31]howtheyusecorrelates with the spread of MDR organisms, particularly MRSA. Therefore, our results suggest that multi-resistance in our areaiswidespreadindiabeticpatientswithfootulcers,and fluoroquinoloneabuse(includinginadequatedosingor sub-optimal therapy duration) in the community could be a potentialcause.

We also evaluated samples from DFIpatients receiving antibiotic therapy,mainlyhospitalizedpatientswith osteo-myelitis,whohadsignsofinfectionprogressionandclinical deteriorationoftheirulcers.Microbialisolationwas signifi-cantlyinfluencedbysystemicantibiotictherapy,withfewer microorganisms (mostly anaerobic bacteria) identified but withasignificantlygreaterprevalenceofMDRorganisms.This findingmaybeexplained byselectivepressurebecausethe

majority of these patients were under broad-spectrum antibiotic therapy, mostly with carbapenems. There are surprisinglyfewpublishedclinicaltrialsonantibiotictherapy forDFIs,andtheavailabledatadonotallowcurrentguidelines to recommend any specific antibiotic regimen. In 2010, however, the PortugueseDirectorate-General ofHealth [32] publishedaclinicalguidelinesuggestingtheuseof isoxazo-lylpenicillins or clindamycinforsuperficialinfections, ami-nopenicillinswithab-lactamaseinhibitororfluoroquinolones combined withclindamycinfordeepinfections,and carba-penemsorureidopenicillinswithab-lactamaseinhibitorfor moresevereinfections.Thesameguidelinealsoconsidered thepotentialuseofcotrimoxazole,vancomycin,linezolidor tigecyclineifMRSAwassuspectedbutdidnotmentionany suspicion criteria. Although these guidelines are typically considered by HCPs, our study showed that the initial empirical antibiotictherapy coveredtheisolatedpathogens ofpatientswithclinicallydeterioratingulcersinonly7.0%of thecases.Therapeuticfailurewasrelatedtothepresenceof MDRorganisms,namelyMRSA.

Inconclusion,ourobservationalstudyprovidesaunique pictureoftheDFIpatterninourregion.Boththeprevalence and precocityofMDRorganismswerealarminglyhigh and were probably related to indiscriminate antibiotic use. Fluoroquinolones,becauseoftheirpharmacological charac-teristics,safetyandprovenclinicaleffectiveness,areamong theantimicrobialagentscurrentlyrecommendedby authori-tative DFIguidelines.However, resistancehasbeendirectly linkedtotheuseofthesecompounds,andthepresentstudy describes a statistical association that should encourage clinicians, and ultimately healthauthorities, toavoid their widespreaduse.By contrast,duetothehighprevalence of MRSAinDFIsinourarea,we suggestempirical anti-MRSA therapy followed by de-escalation to rationalize care and improveoutcomes.

Conflict

of

interest

Therearenoconflictsofinterest.

Acknowledgments

TheauthorswouldliketothankR.OliveiraandL.Pratafrom Associac¸a˜oProtectoradosDiabe´ticosdePortugal;J.Cabete,L. Mota-Capita˜o, C. Pereira-Alves, and L. Vasconcelos from Centro Hospitalarde Lisboa Central, EPE; F. Malheiro from HospitaldeCurryCabral.

r

e

f

e

r

e

n

c

e

s

[1] ShawJE,SicreeRA,ZimmetPZ.Globalestimatesofthe prevalenceofdiabetesfor2010and2030.DiabetesResClin Pract2010;87(1):4–14.

[2] RichardJL,SottoA,LavigneJP.Newinsightsindiabeticfoot infection.WorldJDiabetes2011;2(2):24–32.

[3] CitronDM,GoldsteinEJ,MerriamCV,LipskyBA,Abramson MA.Bacteriologyofmoderate-to-severediabeticfoot

infectionsandinvitroactivityofantimicrobialagents.J ClinMicrobiol2007;45(9):2819–28.

[4] TasciniC,PiaggesiA,TagliaferriE,IacopiE,FondelliS, TedeschiA,etal.Microbiologyatfirstvisitof moderate-to-severediabeticfootinfectionwithantimicrobialactivity andasurveyofquinolonemonotherapy.DiabetesResClin Pract2011.

[5] KarchmerAW,GibbonsGW.Footinfectionsindiabetes: evaluationandmanagement.CurrClinTopInfectDis 1994;14:1–22.

[6] NationalInstituteforHealthandClinicalExcellence.NICE clinicalguideline119:inpatientmanagementofdiabetic footproblems.London:NationalInstituteforHealthand ClinicalExcellence;2011.

[7] JosephWS,LipskyBA.Medicaltherapyofdiabeticfoot infections.JAmPodiatrMedAssoc2010;100(5):395–400. [8] Organizationforeconomicco-operationanddevelopment

(OECD).Healthataglance2009:OECDindicators;2009. Availablefrom:www.oecd.org/dataoecd/32/34/

43541373.pdf[lastupdateSeptember2009,lastaccessed June2011].

[9] MoetGJ,JonesRN,BiedenbachDJ,StilwellMG,FritscheTR. Contemporarycausesofskinandsofttissueinfectionsin NorthAmerica,LatinAmerica,andEurope:reportfromthe SENTRYAntimicrobialSurveillanceProgram(1998–2004). DiagnMicrobiolInfectDis2007;57(1):7–13.

[10] SchaperNC.Diabeticfootulcerclassificationsystemfor researchpurposes:aprogressreportoncriteriafor includingpatientsinresearchstudies.DiabetesMetabRes Rev2004;20(Suppl.1):S90–5.

[11] Executivesummary:standardsofmedicalcarein diabetes-2009.DiabetesCare2009;32(Suppl.1):S6–12.

[12] LaveryLA,ArmstrongDG,PetersEJ,LipskyBA. Probe-to-bonetestfordiagnosingdiabeticfootosteomyelitis:reliable orrelic?DiabetesCare2007;30(2):270–4.

[13] AlaviA,NiakosariF,SibbaldRG.Whenandhowtoperform abiopsyonachronicwound.AdvSkinWoundCare 2010;23(3):132–40.quiz141–2.

[14] BonhamPA.Swabculturesfordiagnosingwound infections:aliteraturereviewandclinicalguideline.J WoundOstomyContinenceNurs2009;36(4):389–95. [15] MurrayPR,BaronEJ,JorgensenJH,PfallerMA,YolkenRH.

Manualofclinicalmicrobiology.Washington,DC:ASM Press;2003.

[16] WooPC,LeungAS,LeungKW,YuenKY.Identificationof slidecoagulasepositive,tubecoagulasenegative Staphylococcusaureusby16SribosomalRNAgene sequencing.MolPathol2001;54(4):244–7.

[17] ClinicalandLaboratoryStandardsInstitute.Performance standardsforantimicrobialdisksusceptibility tests—M02-A10(includesM100-S20).Wayne,PA:Clinicaland LaboratoryStandardsInstitute;2010.

[18] BillTJ,RatliffCR,DonovanAM,KnoxLK,MorganRF, RodeheaverGT.Quantitativeswabcultureversustissue

biopsy:acomparisoninchronicwounds.OstomyWound Manage2001;47(1):34–7.

[19] NelsonEA,O’MearaS,CraigD,IglesiasC,GolderS,DaltonJ, etal.Aseriesofsystematicreviewstoinformadecision analysisforsamplingandtreatinginfecteddiabeticfoot ulcers.HealthTechnolAssess2006;10(12).iii–iv,ix–x,1–221. [20] PatersonDL,BonomoRA.Extended-spectrum

beta-lactamases:aclinicalupdate.ClinMicrobiolRev 2005;18(4):657–86.

[21]FalagasME,KoletsiPK,BliziotisIA.Thediversity ofdefinitionsofmultidrug-resistant(MDR)and pandrug-resistant(PDR)Acinetobacterbaumanniiand Pseudomonasaeruginosa.JMedMicrobiol2006;55(Pt 12):1619–29.

[22] PrompersL,SchaperN,ApelqvistJ,EdmondsM,JudeE, MauricioD,etal.Predictionofoutcomeinindividualswith diabeticfootulcers:focusonthedifferencesbetween individualswithandwithoutperipheralarterialdisease. TheEURODIALEStudy.Diabetologia2008;51(5):747–55. [23] ShakilS,KhanAU.Infectedfootulcersinmaleandfemale

diabeticpatients:aclinico-bioinformativestudy.AnnClin MicrobiolAntimicrob2010;9:2.

[24] ZubairM,MalikA,AhmadJ.Clinico-microbiologicalstudy andantimicrobialdrugresistanceprofileofdiabeticfoot infectionsinNorthIndia.Foot(Edinb)2011;21(1):6–14. [25] BowlerPG,DuerdenBI,ArmstrongDG.Wound

microbiologyandassociatedapproachestowound management.ClinMicrobiolRev2001;14(2):244–69. [26] NgLS,KwangLL,YeowSC,TanTY.Anaerobiccultureof

diabeticfootinfections:organismsandantimicrobial susceptibilities.AnnAcadMedSingapore2008;37(11):936–9. [27] DowG.Bacterialswabsandthechronicwound:when,

how,andwhatdotheymean?OstomyWoundManage 2003;49(5ASuppl.):8–13.

[28] PellizzerG,StrazzaboscoM,PresiS,FurlanF,LoraL, BenedettiP,etal.Deeptissuebiopsyvs.superficialswab culturemonitoringinthemicrobiologicalassessmentof limb-threateningdiabeticfootinfection.DiabetMed 2001;18(10):822–7.

[29] SlaterRA,LazarovitchT,BoldurI,RamotY,BuchsA,Weiss M,etal.Swabculturesaccuratelyidentifybacterial pathogensindiabeticfootwoundsnotinvolvingbone. DiabetMed2004;21(7):705–9.

[30] FerechM,CoenenS,Malhotra-KumarS,DvorakovaK, HendrickxE,SuetensC,etal.EuropeanSurveillanceof AntimicrobialConsumption(ESAC):outpatientquinolone useinEurope.JAntimicrobChemother2006;58(2):423–7. [31] MullerA,MaunyF,TalonD,DonnanPT,HarbarthS,

BertrandX.Effectofindividual-andgroup-levelantibiotic exposureonMRSAisolation:amultilevelanalysis.J AntimicrobChemother2006;58(4):878–81.

[32] Direcc¸a˜oGeraldaSau´de.Diabeticfoot.Nationalprogram forpreventionandcontrolofdiabetes.In:Clinical Guidelinesn8:05/PNPCD;2010.