w w w . e l s e v ie r . c o m / l o c a t e / b j i d
The
Brazilian
Journal
of
INFECTIOUS
DISEASES
Brief
communication
Early
use
of
polymyxin
B
reduces
the
mortality
of
carbapenem-resistant
Klebsiella
pneumoniae
bloodstream
infection
Qiqiang
Liang
a,
Man
Huang
a,∗,
Zhijiang
Xu
baSecondAffiliatedHospitalofZhejiangUniversity,GeneralIntensiveCareUnit,Zhejiang,China bSecondAffiliatedHospitalofZhejiangUniversity,ClinicalLaboratory,Zhejiang,China
a
r
t
i
c
l
e
i
n
f
o
Articlehistory:
Received1October2018 Accepted21December2018 Availableonline21February2019
Keywords:
carbapenem-resistantKlebsiella pneumoniaebloodstreaminfection PolymyxinB-basedcombination therapy
Drugusetiming
a
b
s
t
r
a
c
t
PolymyxinBisoneofthelastresortoptionforcarbapenem-resistantKlebsiellapneumoniae
(CRKP)bloodstreaminfectioninChina.Therefore,thetimingofadministrationofpolymyxin isfrequentlydelayed.Wecollected40casesofCRKPbloodstreaminfections(BSIs)treated withcombinationsbasedonpolymyxinBover30months.Theprimaryoutcome,30-day mortalityrate,was52.5%(21/40).EarlyadministrationofpolymyxinBismeantto adminis-terthedrugwithin48hofdiagnosingbacteremia.Delayedadministrationwasconsidered whenpolymyxinBwasadministeredafter48hofbacteremiaonset.PolymyxinBduration andtotaldosagesweresimilarinthetwogroups(11.57daysversus11.76days,p=0.919; 1306.52mgversus1247.06mg,p=0.711).Comparedwithdelayedadministration,earlyuse ofpolymyxinB-basedcombinationtherapyhadasignificantincreaseintherateof bacte-rialclearance(65.22%versus29.41%,p=0.025;OR=0.533)anddecreased30-daymortality (39.13%versus70.59%,p=0.045;OR=0.461)andoverallmortality(43.48%versus82.35%,
p=0.022;OR=0.321).
©2019SociedadeBrasileiradeInfectologia.PublishedbyElsevierEspa ˜na,S.L.U.Thisis anopenaccessarticleundertheCCBY-NC-NDlicense(http://creativecommons.org/ licenses/by-nc-nd/4.0/).
Introduction
Theincreasingprevalenceofcarbapenem-resistantKlebsiella pneumoniae(CRKP)isahealthcarecrisisinChina.Bloodstream infection (BSI) caused byCRKP is amore serious situation duetoineffectiveantibacterialsandhighmortality.1
Thera-peuticoptionsagainsttheseinfectionsarelimited; “second
∗ Correspondingauthor.
E-mailaddress:huangman@zju.edu.cn(M.Huang).
line”agentsincludecolistin,tigecycline,aminoglycosides,and carbapenems.2 The best available treatment against CRKP
infections isstill under study owingto different chemore-sistancemechanismsandlimitedeffectiveness.Ratherthan monotherapy,cliniciansareincreasinglyresortingto combi-nation therapies, such as dualcarbapenems, carbapenems combinedwithpolymyxinBortigecycline.3–5However,clinical
outcomesarenotconfirmedintheabsenceofevidence-based trials.
Regardedasoneofthelastresorts,polymyxinBhasbeen usedalternativelyinChina,especiallyfollowingthesurgeof CRKPonlysusceptibletocolistin.Thus,researchersarepaying
https://doi.org/10.1016/j.bjid.2018.12.004
1413-8670/©2019SociedadeBrasileiradeInfectologia.PublishedbyElsevierEspa ˜na,S.L.U.ThisisanopenaccessarticleundertheCC BY-NC-NDlicense(http://creativecommons.org/licenses/by-nc-nd/4.0/).
moreattention topolymyxinBoptionsas monotherapyor combinationtherapy,medicationdoses,andadverseeffects.2
Combination therapy is probablybetter than monotherapy forseriousCRKPbacteremia,butthetimingcliniciansshould implementsuchtherapyisstillunclear.2,5,6Ourretrospective
observationalstudy aimed toestimate the impact ofearly initiation ofpolymyxinBon the outcome ofpatients with CRKPBSI.
Methods
SettingA1200-bedacademictertiarycarehospitalwith40intensive carebeds,inHangzhou, China.Recordsfrompatients with positivebloodcultureforCRKPwereretrievedfromOctober 1st2015toMarch31st2018.PatientswhoreceivedpolymyxinB treatmentwereselectedfromthehospitaldatabaseand infor-mationonbasiccharacteristics,laboratoryresults,therapeutic options,andoutcomeswasabstracted.Theclinical parame-ters,originatingfromthehospitaldatabasewithoutsubjective judgments,werenearlyobjective,whichavoidedrecallbias and selectionbias.Thekey outcomewas 30-daymortality, andsecondaryoutcomesincludedbacterialclearancerateand totalmortality.Thewriting processforthisreportfollowed STROBErecommendations.
Exclusion criteria included the following: (a) age less than 18 or older than 85 years; (b) pregnancy or breast-feedingwomen;(c)end-stagerenaldisease;terminalstages of solid tumor or hematological malignancies; termi-nal pulmonary diseases; or severe autoimmune disease; (d) other refractory microbial infections in the period of BSIs, including carbapenem-resistant Acinetobacter bau-mannii (CRAB) and Pseudomonas aeruginosa (CRPA) and otherCarbapenem-resistantEnterobacteriaceae(CRE) infec-tions;methicillin-resistant Staphylococcusaureus(MRSA)and vancomycin-resistantenterococci(VRE)infections;orinvasive fungaldisease;(e)polymyxinBtreatmentsforlessthanthree daysforanyreason;(f)intracranialCRKPinfections;and(g) AcutePhysiology,AgeandChronicHealthEvaluation(APACHE II)scoreover30.
Definition
EarlyadministrationofpolymyxinBwasconsideredwhenthe drugwasadministeredwithin48hofbacteremiaonset;the remainingpatientsweregroupedasdelayedadministration.5
Bacterial clearancewas defined astwo negativeblood cul-tures, improved infection symptoms after treatment, and no rebound in one week thereafter All patients received polymyxinBatdosagesof1.5–2.5mg/kgdaily.Adultpatients receivedhalfofthedosageevery12h,withadjusteddosages in patients with renal insufficiency. The diagnostic crite-ria for acute renal injury (AKI) followed the 2012 Kidney DiseaseImprovingGlobalOutcomesguidelines.Invasive pro-ceduresincluded catheterizationsof deepveins or arteries for continuous renal replacement therapy (CRRT), pulse indicator continuous cardiac output (PICCO), and extracor-poreal membrane oxygenation (ECMO). Immunodeficiency
was determined mainlybylaboratoryindicators, historyof chemotherapy and long-term use of either glucocorticoids orimmunosuppressants.Severethrombocytopeniameanta bloodplateletcountlessthan20×109/L.Theseverityrating
ofthepatient’sacuteconditionatadmissionwasbasedon APACHEIIscores,theCharlsoncomorbidityindexes,theSOFA scores,andthePittbacteremiascores.Infectioussourceswere categorizedasdefiniteinfectionorcolonizationinothersites precedingBSIonset;otherwisetheBSIwasdeemedprimary. Septicshockdefinitionfollowedsepsis3.0.7
Bacterialdetection
Drug susceptibility test was carried out with the broth microdilutionmethodbyalaboratoryphysicianwithanalysis instruments(VITEK2AST-GN16French).Minimuminhibitory
concentration (MIC)determination and interpretationwere determined on a VITEK2 system, based on the Clinical
andLaboratoryStandardsInstitute(CLSI)8recommendations.
Carbapenem resistance meant imipenem and meropenem MIC≥4mg/L.Itisnoteworthythatresistancetotigecycline andcolistinmeantMIC>2mg/L,respectively,referringtothe EuropeanCommitteeonAntimicrobialSusceptibilityTesting (EUCAST,version8.0,2018).9
Statisticalanalysis:Datawereanalyzedusingthestatistical softwarepackageSPSS,version22(SPSS,Inc.,Chicago,USA). Categoricalvariablesinthetwogroupswerecomparedwith thePearson2orFisher’sexacttestifcellshadanexpected
fre-quencyof5orfewer,andcontinuousvariableswitht-testfor independentsamples.30-Daysurvivalfrombacteremiaonset wasplottedasKaplan–Meiercurves.Theoddsratio(OR)and 95%confidenceinterval(CI)werecalculatedtoevaluatethe strengthofassociation.Ap-valuelessthan0.05was consid-eredstatisticallysignificant.
Results
Atotalof69patientswhohadCRKPbacteremiaandreceived polymyxin B between Oct 1, 2015, and Mar 31, 2018were retrieved.Fortycaseswereultimatelyanalyzed,and29cases were excluded. Refractory mixed infections occurred in 15 cases,includingsevenCRABandfourCRPAinfections,three caseswithMRSA,threecaseswithseriousintracranialCRKP infections,andonecaseofinvasiveaspergillosis.Moredetails canbeseenintheflowchart(Fig.1).
Table 1reports thecharacteristics ofpatientswith bac-teremia duetoCRKPand thevariables associatedwiththe timingofpolymyxinBbetweenearlyanddelayed adminis-tration. The earlyuse and delayed use groups had similar characteristics and greathomogeneity.Asatraumacenter, complicatedmultipleinjuries,ascommonprimarydiseases leadingtoICUadmission,accountedfor37.5%,thesameas infectiousdiseasesinourstudy.Age,sexandsourceof infec-tion were similarin the twogroups, and types ofprimary diseaseswerenotsignificantlydifferent.Broncho-pulmonary infection was considered the source of bacteremia in 16 patientsandthegastrointestinaltractinfivecases.Primary bloodstreaminfections,withoutclearlyCRKPinfectiousor col-onizedsites,accountedforalargeproportion(42.5%).
Positive blood culture of CRKP; N=118
CRKP bacteremia; N=112;
Cases who used polymyxin B; N=69;
Cases using polymyxin B; N=40;
Early use of polymyxin B; N=23; Lately use of polymyxin B; N=17;
Excluding cases of contamination; N=6;
Excluding cases whose treatments without polymyxin B; N=43;
Excluding cases:
1. Other refractory microorganism infections, N=15;
2. Polymyxin B treatments less than 3 days, N=3;
3. Intracranial CRKP infection with polymyxin B intrathecal injection, N=3; 4. APACHEII score over 30, N=3; 5. Treatment abandoning due to non-medical factors, N=2;
6. Elderly patient older than 85; N=1; 7. Terminal stages of bronchogenic carcinoma, N=1;
8. Invasive aspergillosis in the period of BSIs, N=1
Fig.1–Flowchartofpatientswithbloodstreaminfectionsduetocarbapenem-resistantKlebsiellapneumoniaetreatedwith polymyxinB.
Intermsofclinicalcharacteristics,APACHEIIscores(18.83 versus22.06,p=0.103)andCharlsoncomorbidityindexes(2.39 versus2.52,p=0.808)onadmissionintoICUweresimilarin bothearly and delayed use groups. Likewise, similar SOFA scores(15.43versus14.47,p=0.293)andPittbacteremiascores (4.17versus3.41,p=0.679)indicatethatthepatientshadthe same clinical condition when bloodstream infections were diagnosed.Laboratoryparameters,includingwhitebloodcell count,bloodplateletcount(PLT),C-reactiveprotein(CRP), pro-calcitonin(PCT),andserumbilirubinwerealsosimilarinthe twogroups.Otherconditionsthatmayaffectprognosis,such asacutekidneyinjury(AKI), invasiveprocedures, immuno-suppression,andsepticshockwerenotsignificantlydifferent inthetwogroups,indicatinghomogeneityintheseverityof infection.
All isolateswereresistant tomeropenem andimipenem andsensitivetocolistin.Nearly27.5%ofisolateswereresistant totigecycline. Althoughweused themostsensitive antibi-otic,theoveralllethalityrateofCRKPbloodstreaminfections were stillhigh.Twenty-four(60%)patientsdied,10(43.48%) patients with early administration of polymyxinB and 14 (82.35%)withdelayedadministration(p=0.022,OR=0.321;CI: 0.110–0.940).AllpatientsacceptedpolymyxinB-based combi-nationtherapies.However,therewasnodifferencebetween the threecombinationswithtigecycline,carbapenems, and amikacin. Compared with the delayed application, early use ofpolymyxinBshowedgreater bacterialclearancerate (65.22% versus 29.41%, p=0.025; OR=0.533; CI:0.295–0.965) andreduced30-daymortality(39.13%versus70.59%,p=0.045; OR=0.461;CI:0.199–0.989)andtheoverallmortality(43.48%
Table1–CharacteristicsofpatientswithbacteremiaduetoCRKPandvariablesassociatedwithadministrationtimingof polymyxinBbetweenearlyanddelayedadministration.
Earlyadministration(n=23) Delayedadministration(n=17) p-Value(ORand95%CI)
Basicinformation Age 61.57 61.35 0.962 M/F 21/2 13/4 0.374 Primarydisease Multipleinjuries 8 7 0.68 Infections 10 5 0.364 Cerebrovasculardiseases 4 2 1 Acuteabdomen 1 3 0.294 Sourceofinfection Primary 9 8 0.616 Broncho-Pulmonary 11 5 0.24 Gastrointestinaltract 2 3 0.397 Other 1 1 1 Clinicalcharacteristics APACHEIIscores 18.83 22.06 0.103 SOFAscores 15.43 14.47 0.293
Charlsoncomorbidityindex 2.87 2.53 0.541 Pittbacteremiascore 4.17 3.41 0.679 Whitebloodcellcount(109/L) 19.95 17.9 0.572
PLT(109/L) 79.18 87.47 0.769 CRP(mg/L) 189.53 166.86 0.333 PCT(g/L) 15.46 7.23 0.208 Septicshock 14 10 0.896 Severethrombocytopenia 10 6 0.601 Hyperbilirubinemia 9 8 0.616
Acutekidneyinjury 8 8 0.433
RecentoperationHistory 13 10 0.884
Invasiveprocedures 13 11 0.601
Tracheotomy 11 11 0.289
Immunosuppression 11 6 0.428
Antimicrobialsusceptibilitytest
MeropenemMIC≥4(g/mL) 23 17 1.000
ImipenemMIC≥4(g/mL) 23 17 1.000
TigecyclineMIC>2(g/mL) 6 5 0.816
ColistinMIC>2(g/mL) 0 0 1
Therapeuticregimen(death/all)
Combinationwithtigecycline 3/7(42.86%) 4/5(80%) 0.944 Combinationwithcarbapenems 6/15(40%) 6/10(60%) 0.428 Combinationwithamikacin 0/1(0%) 2/2(100%) 1 PolymyxinBduration(days) 11.57 11.76 0.919 TotalpolymyxinBdosage(mg) 1306.52 1247.06 0.771
Therapeuticeffect
Bacterialclearancerate 15(65.22%) 5(29.41%) 0.025(OR=0.533;CI:0.295–0.965) Mortalityin30days 9(39.13%) 12(70.59%) 0.045(OR=0.461;CI:0.199–0.989) Totalmortality 10(43.48%) 14(82.35%) 0.022(OR=0.321;CI:0.110–0.940) ThelengthofICU(days) 32.00 36.88 0.434
Thelengthofhospitals(days) 44.61 39.47 0.495 PLT,bloodplatelet;CRP,C-reactiveprotein;PCT,Procalcitonin;MIC,MinimalInhibitoryConcentration.
versus 82.35%, p=0.022; OR=0.321; CI: 0.110–0.940). Fig. 2
showstheKaplan–Meiercurvesofpatientstreatedwithearly and delayed administration of polymyxin B (p=0.042). It shouldbepointedout thatpolymyxinBduration andtotal dosagesinthetwogroupsweresimilar(11.57daysversus11.76 days,p=0.919;1306.52mgversus1247.06mg,p=0.711, respec-tively).Therefore, mortalitywas notcorrelated withlonger treatmentdurationorhigherdosageofpolymyxinB.During
thetreatment,therewereneithercaseswithdetectablerenal toxicitynorneurovirulencerelatedtopolymyxinB.
Discussion
Inourstudy,preemptiveuseofpolymyxinBbeforeblood cul-tureresultshadbetterbacterialclearanceandmorefavorable
1.0 0.8 0.6 0.4 0.2 0.0 0 5 10 15 Follow-up (days) Early application Delayed application Sur viv al (%) 20 25 30
Fig.2–Kaplan–MeiercurvesofpatientstreatedwithearlyanddelayedadministrationofpolymyxinB.
outcomeinpatientswithBSIduetoCRKPthandelayeduseof thedrug.
PolymyxinBisanundisputed effectiveantibioticchoice for treating patients with CRKP bloodstream infections accordingtocultureresults.However,thelimitationsof tradi-tionalmicrobiologicaltechniquesinidentifyingbloodstream pathogenswilldelaytreatmentinitiationofseriousinfections suchasCRKPBSI.Furthermore,polymyxinBwasonlyrecently approvedforentryintotheChinesemarket,andthepriceis exceedinglyexpensive.Asaresult,cliniciansregarditasthe lastresort,preferringtigecyclineorcarbapenems,postponing theuseofpolymyxinBandunavoidablyaffectingprognosis. Gutiérrez-Gutiérrezetal.observedthatappropriatetherapy startinguptofivedaysafterinfection wasassociatedwith decreasedmortality(OR=0.44)inpatientswithBSIsdueto CPE.5
Early appropriatetherapy(starting up to twodays after infection) canalsoreduce mortality(OR=0.84).We defined early use of polymyxin Bas administering it before blood culture results, which is analogous to empirical antibiotic treatment.Incontrast,thedelayedgroupcanbeconsideredas havingreceivedtargetedantibiotictherapy.10However,there
issomedifferencebetweenearlyuseinourstudyand empir-icaltreatment, as halfofthe patients (57.5%)had primary infectionsites,suchasthebroncho-pulmonaryand gastroin-testinaltract.Oncetheinfectioussymptomsoccurred,BSIdue toCRKPpromptedearlytargetedtherapiesfirst.Nonetheless, mostpatientswithprimaryBSIshadhighriskfactors, includ-inginvasiveprocedures,immunosuppressionandhistoryof carbapenemsuse.11,12Patientsinthedelayedusegroupalso
had high riskofCRKP infections,but other antibiotic regi-menswerechoseowingtotherefusalofthepatients’relatives andtheavailabilityofalternativeregimens.Thetotal treat-mentduration anddosagesofpolymyxinBwere similarin thetwogroups.Therefore,theadvantagesofearly adminis-trationcannotbe attributedtothesefactors.Wespeculate that early application of polymyxinB can reach an effec-tivebloodconcentrationasquicklyaspossibleandtruncating
the duration of bacteremia, eventually reducing the death rate.
Theimpact ofthe underlyingconditionsinmortalityof criticalpatientspresentingwithCRKPbacteremiaarevitaland couldoutweighttheinfluenceofantibiotictherapy.Wetried our best tobalance the characteristics ofpatientstoavoid confoundingprimarydiseasesbyemployingstricteligibility criteria, such as excluding patients with terminaldisease, intracranialCRKPinfections,andhighAPACHEIIscores(>30 points).
Althoughanoptimalregimenisyettobedefined,most studieshaverecommendedpolymyxinB-basedcombinations in instead of monotherapy to reduce mortality.2,4,5,13 In a
recentrandomizedcontrolledtrial(RCT)byMicalPauletal. polymyxin B combined with a carbapenem was not more effectivethanmonotherapyincarbapenem-resistant Gram-negativebacterial infections.However,the mainbacterium wasCRABratherthanCRKP,andpneumoniaandinfections of other sites were included in this RCT, leading to great heterogeneity.14 In the study by Gutiérrez-Gutiérrez et al.
polymyxinBwasadministeredto154patients,halfofthemon monotherapyhada30-daymortalityof54%(40/74),whereas thosepatientsonpolymyxinB-containingregimenshada 30-day mortalityof 38% (28/74).5 Differentcombinations have been recommended as there are many possible treatment regimens.Carbapenem-containingregimens,asDaikosetal. reported,isamongthemosteffectivecombinations.6The
sur-vivalrate was52% withpolymyxinBwithcarbapenemsin our report, which was similar to the 50% synergy seen in vitro.15Inourstudy,polymyxinBwasfrequentlycombined
witheitheracarbapenem,ortigecycline,oraminoglycosides. Two-thirdsofpatientsreceivedthecombinationofpolymyxin Bandacarbapenem,withahighmortalityrate(61.11%). How-ever, wecannot beabletodraw any definitiveconclusions because ofthe small samplesizes and nonsignificant sta-tisticalresults.Ourstudyhassomelimitations,suchasthe impactsofunmeasuredvariablesandconfoundingfactorsdue totheobservationalnatureofthestudy.
Conclusion
WerecommendearlyadministrationofpolymyxinB-based combinationtherapybeforebloodcultureresultsforpatients withbloodstreaminfectionsduetoCRKP.Irrespectiveofdrug dosageandtreatmentduration,earlyusecaneffectively elim-inatebacteria,improving30-dayandoverallmortality.
Authors’
contributions
QiqiangLiangcollectedandanalyzedthedataandwrotethe firstdraftofthemanuscript.ZhijiangXuassistedwiththe sta-tisticalanalysesofthemicroorganisms.ManHuangreviewed andmodifiedthedraftofthemanuscript.
Funding
Nofunding.
Conflicts
of
interest
Theauthorsdeclarenoconflictsofinterest.
Acknowledgements
XinXUandTingtingGenassistedinselectingtheclinicaldata.
r
e
f
e
r
e
n
c
e
s
1. ZhangY,WangQ,YinY,etal.Epidemiologyof
Carbapenem-resistantEnterobacteriaceaeinfections:report fromChinaCRENetwork.AntimicrobAgentsChemother. 2017.
2. Rodríguez-Ba ˜noJ,Gutiérrez-GutiérrezB,MachucaI,Pascual A.Treatmentofinfectionscausedby
extended-spectrum-BetaLactamase-,AmpC-,and
carbapenemase-producingEnterobacteriaceae.ClinMicrobiol Rev.2018.
3. TumbarelloM,TrecarichiEM,DeRosaFG,etal.Infections causedbyKPC-producingKlebsiellapneumoniae:differencesin therapyandmortalityinamulticentrestudy.JAntimicrob Chemother.2015;70:2133–43.
4.JacobsDM,SafirMC,HuangD,MinhajF,ParkerA,RaoGG. Triplecombinationantibiotictherapyfor
carbapenemase-producingKlebsiellapneumoniae:asystematic review.AnnClinMicrobiolAntimicrob.2017;16:76.
5.Gutiérrez-GutiérrezB,SalamancaE,deCuetoM,etal.Effect ofappropriatecombinationtherapyonmortalityofpatients withbloodstreaminfectionsdueto
carbapenemase-producingEnterobacteriaceae(INCREMENT): aretrospectivecohortstudy.LancetInfectDis.2017;17:726–34.
6.DaikosGL,TsaousiS,TzouvelekisLS,etal.
Carbapenemase-producingKlebsiellapneumoniaebloodstream infections:loweringmortalitybyantibioticcombination schemesandtheroleofcarbapenems.AntimicrobAgents Chemother.2014;58:2322–8.
7.SingerM,DeutschmanCS,SeymourCW,etal.TheThird InternationalConsensusDefinitionsforSepsisandSeptic Shock(Sepsis-3).JAMA.2016;315:801–10.
8.InstituteCaLS.Performancestandardsforantimicrobial susceptibilitytesting.M100S28.ClinicalandLaboratory StandardsInstitute;2018.
9.GiskeC.TheEuropeanCommitteeonantimicrobial susceptibilitytesting–EUCASTbreakpointtables;2018
http://www.eucast.org/clinicalbreakpoints/
10.Martin-LoechesI,LeoneM,MadachK,MartinC,EinavS. Antibiotictherapyinthecriticallyill–expertopinionofthe IntensiveCareMedicineScientificSubcommitteeofthe EuropeanSocietyofAnaesthesiology.EurJAnaesthesiol. 2017;34:215–20.
11.XuL,SunX,MaX.Systematicreviewandmeta-analysisof mortalityofpatientsinfectedwithcarbapenem-resistant Klebsiellapneumoniae.AnnClinMicrobiolAntimicrob. 2017;16:18.
12.GiacobbeDR,DelBonoV,BruzziP,etal.Previousbloodstream infectionsduetootherpathogensaspredictorsof
carbapenem-resistantKlebsiellapneumoniaebacteraemiain colonizedpatients:resultsfromaretrospectivemulticentre study.EurJClinMicrobiolInfectDis.2017;36:663–9.
13.MachucaI,Gutierrez-GutierrezB,Gracia-AhufingerI,etal. Mortalityassociatedwithbacteremiaduetocolistin-resistant Klebsiellapneumoniaewithhigh-levelmeropenemresistance: importanceofcombinationtherapywithoutcolistinand carbapenems.AntimicrobAgentsChemother.2017;61.
14.PaulM,DaikosGL,Durante-MangoniE,etal.Colistinalone versuscolistinplusmeropenemfortreatmentofsevere infectionscausedbycarbapenem-resistantGram-negative bacteria:anopen-label,randomisedcontrolledtrial.Lancet InfectDis.2018;18:391–400.
15.ZusmanO,AvniT,LeiboviciL,etal.Systematicreviewand meta-analysisofinvitrosynergyofpolymyxinsand carbapenems.AntimicrobAgentsChemother. 2013;57:5104–11.