Considerations for the optimal management of antibiotic therapy in elderly patients

(1)

Article

Reference

Considerations for the optimal management of antibiotic therapy in elderly patients

FALCONE, Marco, et al. & ESCMID Study Group for Infections in the Elderly (ESGIE)

Abstract

Objectives: To maximise efficacy and minimise toxicity, special considerations are required for antibiotic prescription in elderly patients. This review aims to provide practical suggestions for the optimal management of antibiotic therapy in elderly patients.

FALCONE, Marco, et al . & ESCMID Study Group for Infections in the Elderly (ESGIE).

Considerations for the optimal management of antibiotic therapy in elderly patients. Journal of Global Antimicrobial Resistance , 2020, vol. 22, p. 325-333

DOI : 10.1016/j.jgar.2020.02.022 PMID : 32165285

Available at:

http://archive-ouverte.unige.ch/unige:153195

Disclaimer: layout of this document may differ from the published version.

1 / 1

(2)

Review

Considerations for the optimal management of antibiotic therapy in elderly patients

Marco Falcone

^a,

*, Mical Paul

^b

, Giusy Tiseo

^a

, Dafna Yahav

^c

, Virginie Prendki

^d

,

Lena E. Friberg

^e

, Roberto Guerri

^f

, Gaetan Gavazzi

^g

, Cristina Mussini

^h

, Marco Tinelli

ⁱ

, for the ESCMID Study Group for Infections in the Elderly (ESGIE)

aDepartmentofClinicalandExperimentalMedicine,UniversityofPisa,ViaParadisa2,Pisa56124,Italy

bDivisionofInfectiousDiseases,RambamHealthCareCenterandTheRuthandBruceRappaportFacultyofMedicine,Technion–IsraelInstituteofTechnology, Haifa,Israel

cInfectiousDiseasesUnit,RabinMedicalCenter,BeilinsonHospital,PetahTikva,Israel

dDivisionofInternalMedicineoftheAged,DepartmentofRehabilitationandGeriatrics,GenevaUniversityHospitalsandUniversityofGeneva,Hôpitaldes Trois-Chêne,Geneva,Switzerland

eDepartmentofPharmaceuticalBiosciences,UppsalaUniversity,Uppsala,Sweden

fInfectiousDiseasesDepartment,HospitaldelMar,DepartmentofMedicine,UniversitatAutònomadeBarcelona,Barcelona,Spain

gUniversityClinicsofGeriatrics,UniversityHospitalofGrenoble-Alpes,GREPIEA7408UniversityofGrenobleAlpes,Grenoble,France

hClinicofInfectiousDiseases,UniversityofModenaandReggioEmilia,Modena,Italy

iDivisionofInfectiousandTropicalDiseases,HospitalofLodi,Lodi,Italy

ARTICLE INFO

Articlehistory:

Received2February2020

Receivedinrevisedform14February2020 Accepted26February2020

Availableonline9March2020

Keywords:

Elderly Pharmacokinetics Pharmacodynamics Antibiotics Dosage Adverseevents

ABSTRACT

Objectives:Tomaximiseefﬁcacyandminimisetoxicity,specialconsiderationsarerequiredforantibiotic prescriptionin elderlypatients. This review aimstoprovide practical suggestionsfor theoptimal managementofantibiotictherapyinelderlypatients.

Methods:Thiswasanarrativereview.Aliteraturesearchofpublishedarticlesinthelast15yearson antibiotics andelderlypatientswas performedusingtheCochrane Libraryand PubMedelectronic databases.Thethreepriorityareaswereidentiﬁed:(i)pharmacokinetics/pharmacodynamics(PK/PD)for optimising dosage regimens and route of administration; (ii) antibiotic dosages in some special subpopulations; and (iii) treatmentconsiderations relatingto differentantibiotic classesand their adverseevents.

Results:CliniciansshouldunderstandthealteredPK/PDofdrugsinthispopulationowingtoco-morbid conditionsandnormalphysiologicalchangesassociatedwithageing.Thebodyofevidencejustiﬁesthe needforindividualiseddoseselection,especiallyinpatientswithimpairedrenalandliverfunction.

Cliniciansshouldbeawareofthemajordrug–druginteractionscommonlyobservedintheelderlyaswell aspotentialsideeffects.

Conclusion:Antibiotictherapyintheelderlyrequiresacomprehensiveapproach,includingstrategiesto improveappropriateantibioticprescribing,limittheiruseforuncomplicatedinfectionsandensurethe attainmentofanoptimalPK/PDtarget.Tothispurpose,furtherstudiesinvolvingtheelderlyareneededto betterunderstandthePKofantibiotics.Moreover,itisnecessarytoassesstheroletherapeuticdrug monitoringinguidingantibiotictherapyinelderlypatientsinordertoevaluateitsimpactonclinical outcome.

4.0/).

* Correspondingauthor.

E-mailaddress:marco.falcone@unipi.it(M.Falcone).

http://dx.doi.org/10.1016/j.jgar.2020.02.022

JournalofGlobalAntimicrobialResistance22(2020)325–333

ContentslistsavailableatScienceDirect

Journal of Global Antimicrobial Resistance

j o u r n a lh o m e p ag e :w w w . e l s e vi e r . c o m / l o c a t e/ j g a r

(3)

Contents

1. Introduction ... 326

2. Methods ... 326

3. Pharmacokinetics/pharmacodynamicsandrouteofadministrationofantibioticsintheelderly ... 327

3.1. β-Lactamantibiotics ... 327

3.2. Vancomycin ... 327

3.3. Linezolid ... 327

3.4. Daptomycin ... 327

3.5. Fluoroquinolones ... 329

3.6. Gentamicin ... 329

3.7. Colistin ... 329

3.8. Routeofantibioticadministration ... 329

4. Antibioticdosagesinspecialpatientpopulations ... 329

4.1. Elderlypatientswithchronickidneydisease(CKD)orhaemodialysis ... 329

4.2. Elderlycirrhoticpatients ... 330

5. Adverseeffectsofantibioticsinelderlypatients ... 330

6. Limitations ... 332

7. Conclusions ... 332

Conﬂictofinterests ... 332

Acknowledgment ... 332

References ... 332

1.Introduction

The World Health Organization (WHO) predicts that the numberofpeopleaged60 yearswillrisefrom900millionto 2billionbetween2015and2050(movingfrom12%to22%ofthe totalglobalpopulation)[1].Ageingisariskfactorfordeveloping infections. Antibiotics are among the most frequently newly- prescribeddrugsinelderlypatients,especiallyinthoseresidingin nursinghomesorlong-termcarefacilities(LTCFs)[2].Multipleco- morbidities,changesindrugpharmacokinetics(PK)andpharma- codynamics (PD), and the presence of polypharmacy with the inherentrisk ofadversedrugreactions anddrug–drugordrug– diseaseinteractionsmakethechoiceoftheoptimalantibioticvery challenging in elderly patients [3]. Appropriate antibiotic prescription,eitherintermsofdrugchoiceordosage,isofparamount importanceamongelderlypatients,butbalancingefficacy,safety, tolerabilityanddevelopmentofantimicrobialresistanceisdifficult inthispatientpopulation.Theobjectiveofthisreviewistodiscuss specialconsiderationsforantibiotictherapyinelderlypatients,in generalandforspecificantibiotics.

2.Methods

This review with recommendations for practice has been produced bya teamof experts belongingtotheESCMID Study GroupforInfectionsintheElderly (ESGIE).Thepaneldiscussed unmetneedsofantibiotictherapyinelderlypatientsandidentiﬁed the following three areas of interest: (i) PK considerations for optimisingdosageregimens and routeof administrationin the elderly; (ii) antibiotic dosages in special elderly populations (speciﬁcallypatients withrenal disease,cirrhotic patients, and patientswithalteredfatandmusclebodycomposition);and(iii) adverse events and drug interactions of antibiotics in elderly adults.

Electronic databases including the Cochrane Library and PubMed were utilised for a comprehensive search using the following combinations of keywords for the identiﬁcation of relevantstudies:[(‘elderly’OR‘LTCF’OR‘nursinghomes’OR‘aged’) AND (‘pharmacokinetics’ OR ‘pharmacodynamics’) AND ‘antibiotics’], [‘elderly’ AND ‘antibiotics’AND (‘chronic renal failure’ OR‘hemodialysis’OR‘kidneydisease’OR‘dosageadjustment’)],

Fig.1.Mainpharmacokinetic/pharmacodynamic(PK/PD)alterationsofantibioticsinelderlypatients.Ofnote,theroleofproteinbindingondrugadjustmentisnotclear.

Changesinproteinbindingdonotrequireadjustmentsindosingregimensexceptincasesofintravenousdrugswithahighextractionratio(seetextformorecomments).

326 M.Falconeetal./JournalofGlobalAntimicrobialResistance22(2020)325–333

(4)

[‘elderly’AND‘antibiotics’AND(‘cirrhosis’OR‘hepaticfailure’)], and[‘elderly’AND‘antibiotics’AND‘adverseevents’].Articleswere selectedonthebasisofthefollowingcriteria:focusedonelderly patientsandpublishedinthelast15years.

Despitethelackofuniversallyacceptedagecriteriatoclassify agegroups,weagreedtoincludeinthisreviewallstudiesfocusing onpatientsaged65years.Casereportsandmini-caseserieswere excluded.

3.Pharmacokinetics/pharmacodynamicsandrouteof administrationofantibioticsintheelderly

AvarietyofpathophysiologicalchangesmayaffectthePKand PDofantibioticsinelderlypatients(Fig.1).

PKisaffectedbychangesinbodycomposition.Sarcopeniaand malnutrition can occur in elderly patients and substantially inﬂuencethePKofadministereddrugs.Decreasedfattissueand leanbodymassmayaffectdrugdistributiondependingontheir lipophilicity.Areducedmassofadiposetissueaccumulateslower amountsoflipophilicdrugs,whereaschangedmusclemassand redistributed body water affect the distribution of hydrophilic drugs. This could lead to greater ﬂuctuations of drug plasma concentrations and higher peak concentrations in the central compartment. Since malnutrition has been associated with reducedcontentofsomehepaticcytochromes,drugmetabolism maybereducedinpatientswithcachexia[4].

Drug distribution can be affected by oedema secondary to chronicheartfailureandbyascitessecondarytocirrhosis.Changes inplasmaproteinbindingdonotusuallyinﬂuencetheclinicaldrug exposure in a patient [5]. Of consequence, no adjustments in dosingregimensareneededexceptinrarecasesof intravenous drugswithahighextractionratioandnarrowtherapeuticindex thataregivenparenterally.

Morphologicalandfunctionalchangessuchasdelayedgastric emptying,reducedsplanchnicbloodﬂowandalteredgastricpH canaffectthebioavailabilityoforallyadministereddrugs[6].

Impairmentofrenalbloodﬂow,glomerularﬁltrationrateand capacityofrenaltubularsecretionincreasetheplasmahalf-lifeof drugseliminatedbythekidneys[6].Adescriptionoftheeffectsof ageingonthemetabolismandeliminationofseveralantibioticsis showninTable1.

PDfromthehost'sperspective(i.e.thehost'sclinicalresponse tothedrug)isaffectedbyimmunesenescence,thephysicalability todealwithcertaininfections(e.g.coughinginpneumonia)and, speciﬁcally,theabilitytodealwithsevereinfectionsfunctionally andcognitively(seeFig.1).Thesameantibioticlevelsatthesame siteof theinfection causedby thesamebacteriummighthave differentclinicaleffectsinyoungerandelderlypatients.

3.1.β-Lactamantibiotics

Regardingβ-lactams,itis wellknownthatthePDindexthat optimises efficacy is the percentage of time the unbound concentrationremainsabovetheminimuminhibitoryconcentra- tionofthetargetmicro-organism(%fT_>MIC).Thereisevidencethat ageaffectsthePKofβ-lactamsandtheeffectismainlymediatedby reduced renal clearancein the elderly.An increasein systemic exposuretoceftarolineof33%wasattributedtodecreasedrenal function in elderly subjects [8]. Nearly 70% of meropenem is excretedfromthebodyintheurine,andcreatinineclearance(CLCr) issignificantlycorrelatedwithmeropenemclearance[9–11].The PK of doripenem has been also specifically studied in elderly patientswithnosocomialpneumonia:theareaundertheplasma concentration–timecurve(AUC)washigherandtheelimination half-life (t1/2) was longer in elderly patients compared with younger healthy subjects [12]. However, there might be other

factors. For example, the 24-h AUC (AUC_0–24) for ampicillin/

sulbactamwassigniﬁcantlylowerinelderlypatients[13],likelyas aconsequenceofanincreaseoftheirvolumeofdistribution(Vd)in theacutephaseofpneumonia.Dataregardingnewdrugssuchas avibactamsuggest thatthemaximumconcentration(Cmax)was lowerandthet1/2waslongerinelderlymalesubjectscompared withyoungerones[14].

Dosagesforelderlysubjectsshouldbebasedatleastonrenal function (Table 1). Speciﬁc considerations for other newer antibioticinelderlypatientshaveyettobedeveloped.

3.2.Vancomycin

Advanced age is a recognised risk factor for vancomycin- induced nephrotoxicity [15]. Vancomycin clearance correlates with CLCr because up to 90% of administered vancomycin is excretedunchangedintheurinewhenrenalfunctionisnormal.

ThePK/PDindexthatbestpredictsvancomycinefﬁcacyistheAUC/

MICratio,andanAUC/MICratioof400hasbeenproposedasan efficacytargetforvancomycintherapy[11].Thelatestpublished guidelinesemphasisetheroleoftheAUCover24htoMICbybroth microdilution (AUC/MICBMD) of 400 as the primary PK/PD predictor of vancomycin activity ifthe MIC is 1mg/L, stating that trough-only monitoring may be insufficient to guide vancomycin dosing in all patients [16]. Some studies have evaluatedtheriskofnephrotoxicityinelderlypatientswithhigh (15mg/L) rather than low (<15mg/L) average vancomycin troughs levels [17]. Bourguignon et al. built and validated a vancomycinPKmodelforpatientsaged>80yearsusingBayesian approaches.Theauthorsfoundhighinterindividualvariabilityin PKparametersinthisspecificpopulation[18].

Unfortunately,therapeuticdrugmonitoring(TDM)ofvancomy- cininparticularsettingssuchasLTCFsisdifﬁculttoperform.Based onthecurrentbestavailableevidence,dailyAUCs(assuming an MICBMDof1mg/L)shouldbemaintainedbetween400and600mgh/

Ltomaximiseefﬁcacyandminimisethelikelihoodofnephrotoxicity [19].Toimplementmodel-basedTDM,software,accountingforthe dosinghistory,should bevalidatedinthis patientpopulationto achievetheoptimalPK/PDvancomycintarget(https://link.springer.

com/article/10.1007%2Fs40262-012-0020-y).

3.3.Linezolid

The AUC0–24of linezolidappearstobecorrelated withbody weightandage,showingatendencytoincreaseasbodyweight decreasesand/orasageincreases[20].ArecentPKstudyshowed that patients treated with the conventional linezolid dose of 600mg twice daily had a 20-fold interindividual variation in antibiotic trough concentrations, with a positive correlation betweenlinezolidtroughconcentrationsandpatientage[21].In a TDM analysis, very old patients(80 years old)had concen- trationsthreetimeshighercomparedwithpatientsaged<40years [21]. Similarly, Tinelli et al. documented that elderly patients treatedwiththeconventionaldoseoflinezolidof600mgtwice daily havelinezolidtroughconcentrations exceeding theupper therapeutic threshold, set according to available literature, at 8.0mg/L[22].Baselineplateletcountandtherapydurationof10 days are the most important predictors of linezolid toxicity.

However,atthistime,thereisnoclearrecommendationofhowto adjustthelinezoliddosetoavoidoverexposureofthedruginthe elderly.

3.4.Daptomycin

AcomparisonofPKparametersofdaptomycininyoungadult (18–30years)and geriatric(75years)volunteersshowed that

M.Falconeetal./JournalofGlobalAntimicrobialResistance22(2020)325–333 327

(5)

Table 1

Effects of ageing on the metabolism and elimination of selected antibiotics.

Antibiotic PK parameter Changes in the elderly Clinical implications

Cmax AUC t1/2 Vss CL Renal

excretion Ampicillin/sulbactam

2 g q8 h[13]

33.86.5 mg/L 51.42.5 mg h/L 2.11.5 h 96.965.7 L CLtot, 602.3273.8 mL/

min

75–80% IncreasedVd

ReducedCmax

More frequent dosing of ampicillin 2 g/

sulbactam 1 g may be necessary to avoid risk of underdosing (in those with impaired renal function, the longert1/2

allows a sufﬁcientCmin).

Ceftaroline 600 mg single dose[8]

31.84.6mg/

mL

94.113.6mg h/mL 3.10.4 h 17.93.0 L CLtot, 95.713.4 mL/min CLR, 54.912.7 mL/min

88% Systemic exposure modestly higher (33%) than in healthy young subjects.

Dosages for elderly subjects should be determined based on renal status.

Meropenem[9]: 70% CLCrand age have the most signiﬁcant

impact on PK of meropenem.

Individualised PK/PD-guided dosing associated with better clinical outcomes and reduced antibiotic use compared with standard dosing.

Equation 1 Vd= 10.8(body

weight/70)^0.99

CL (L/min) = 14.6(CLCr/ 83)^0.62(age/35)^0.34

Equation 2 Vd= 14.6(body

weight/61)

CL (L/h) = 9.7(CLCr/120) Doripenem 500 mg

single dose[12]

22.4035.5mg/

mL

57.0259.9 mg h/L 1.8929.3 h 70% Longert1/2and higher AUC compared

with that of healthy subjects.

CLCrwas the most signiﬁcant covariate on doripenem CL.

In nosocomial pneumonia, PK analysis showed that 500 mg doripenem q8 h may provide a favourable antibiotic effect against bacteria with MICs up to 2mg/mL, but less is known about safety.

Vancomycin (Cminof 10–15mg/mL)[7]:

CLR, 75% Elderly patients with poor renal

function are likely to have increased AUC values and a poor prognosis.

AUC/MIC of 250–450mg h/mL is a suitable target for initial empirical treatment of MRSA pneumonia in the elderly.

Consider alternative agents in elderly patients with renal failure.

Survivors 24.58.2mg/

mL

34495.8 mg h/L 26.513.1 h 62.36.6 L 40.816.9 mL/min Non survivors 25.58.0m^g/

mL

394.7209.9 mg h/L 31.523 h 63.64.1 L 35.518.9 mL/min

Vancomycin[18]: AUC/MIC: CLR, Severe hypoALB inﬂuencest1/2of

vancomycin and treatment outcomes in elderly patients (increased % of nephrotoxicity in the severe hypoALB group).

In elderly patients, evaluation and improvement of nutritional status is essential.

Severe hypoALB 26.81.8m^g/

mL

426.343m^{g h/mL} ^33.2^{5.4 h} ^{64.01.1 L} ^33.7^{3.7 mL/min} Non-severe hypoALB 25.71.0mg/

mL

340.114.0mg h/mL 24.91.6 h 62.30.7 L 40.72.1 mL/min Ciproﬂoxacin 200 mg

q12 h[24]

1.30–4.44m^g/

mL

13.715.5 mg h/L – 78.4113.17 L CLtot, 18.394.15 L/h 50–70% Strong inﬂuence of CLCrand body weight on ciproﬂoxacin CL andVd, respectively.

CLCrand body weight should be considered for dosage optimisation of ﬂuoroquinolones in elderly.

Ciproﬂoxacin[25]: For MICs of 1 mg/L, all simulated

patients reach the efﬁcacy target.

For higher MICs, proposed regimens were inefﬁcient for patients with renal failure.

Efﬁcacy should be evaluated by observing the value of the index AUC/

MIC.

Dose reduction in elderly patients with renal impairment does not ensure optimal drug exposure against pathogens with higher MIC.

500 mg q12 h 339.8061.73 mg h/L

250 mg q12 h 206.0935.98 mg h/L

250 mg q24 h 123.2922.49 mg h/L

PK, pharmacokinetics;Cmax, maximum plasma concentration; AUC, area under the concentration–time curve;t1/2, elimination half-life;Vss, volume of distribution at steady-state; CL, clearance; q8 h, every 8 h; CLtot, total clearance;

Vd, volume of distribution;Cmin, trough concentration; CLCr, creatinine clearance; PD, pharmacodynamics; MIC, minimum inhibitory concentration; MRSA, methicillin-resistantStaphylococcus aureus; hypoALB, hypoalbuminaemia;

q12 h, every 12 h; q24 h, every 24 h.

328M.Falconeetal./JournalofGlobalAntimicrobialResistance22(2020)325–333

(6)

CmaxandVdwerenotsigniﬁcantlydifferent[23].Ofnote,theextent ofexposure (AUC0t andAUC01)of daptomycinwas higherin geriatricsubjects.TheauthorsconcludedthatchangesinthePKof daptomycininelderlypatientsareattributabletochangesinrenal function,withlowerexcretionofdaptomycinintheurine[23].

3.5.Fluoroquinolones

Antibacterialeffectivenessoffluoroquinolonesdependsonthe AUC₀₂₄/MICratio.Populationanalysisrevealedastronginfluence of CLCr and body weighton fluoroquinolone PK in the elderly patient population [24]. Dose reductions according to renal functionaregenerallywelladaptedtogeriatricpatientsifbacteria havelowMICs.However,inpatientswithmoderateorsevererenal impairmentandinfectioncausedbystrainswithMIC1mg/L,a dosagereductioncannotallowanoptimalexposure,withriskof suboptimaltreatment[25].

3.6.Gentamicin

Therearenorecentstudiesofaminoglycosidesintheelderly population.Ithasbeenreportedthatlowalbuminconcentrations are associated with a largerVd and lower Cmax of gentamicin, despitethefact thatgentamicin hasnegligible proteinbinding.

Thus,although albuminlevelsdo notdirectlyaffectgentamicin levels,theymightbeaconvenientmarkeronwhichgentamicin dosingcanbeadjusted[26].However,AUC/MICispromotedasthe bestpredictorofresponseforaminoglycosides,andclearanceof gentamicinistotallycorrelatedwithCLCr.Thus,CLCristhemost importantfactorpredictingexposuretoaminoglycosidesandtheir nephrotoxicity.Thus,TDM,doseadjustment,shorttreatmentsand avoiding other nephrotoxic drugs could beuseful strategies to avoidnephrotoxicityinelderlypatients[27].

3.7.Colistin

Colistin is an old antibiotic that has recently re-emerged becauseoftheincreaseinbacterialresistanceowingtothespread ofmultidrug-resistant(MDR) Gram-negative bacteria. However, thereissubstantialoverlapintheplasmaconcentrationsrequired foranantibacterialeffectandthosethatincreasethelikelihoodof colistin-associated nephrotoxicity. PK studies using modern methodology to separate colistimethate from colistin showed

largeinterindividualvariabilityincolistinlevels,associatedwith nephrotoxicityandprobablywithefﬁcacy[28,29].Thus,although notyetimplementedclinically inmostcentres,elderlypatients wouldbeneﬁtfrommonitoringofcolistinlevels.

3.8.Routeofantibioticadministration

Intravenous (i.v.) administration of antibiotics in elderly patients leads to quick delivery of antibiotics. However, i.v.

administration can be challenging in elderly adults because of the risk of skintears, haematomas, inadvertentextraction and phlebitis. These complications may occur more frequently in elderlypatientsreceivingantiplateletoranticoagulantagents.

Intramuscular (i.m.) injection maybe used for several antibiotics (ceftriaxone, gentamicin, penicillin) and allows their administration at home, facilitating earlier hospital discharge.

However, this method can also lead to complications such as haematomaorabscess.Thesubcutaneousrouteisfrequentlyused to administer treatments (midazolam, morphine succinate) in geriatricsettingsbecauseitistechnicallylesstimeconsumingfor nursesandispainlessandsafertoperformcomparedwithi.v.and i.m. injection. Subcutaneous administration of antibiotics is relativelycommonamong Frenchgeriatricsbut itsuseinother Europeancountriesisunknown[30].

Oraladministrationisconvenientinelderlypatientsbutmight not beeffective asmany of theantibiotics usedin critically ill patients are not available as oral formulations as theyare not absorbedfromthegastrointestinaltract.Moreover,dysphagiacan affect safeoral intake.Compliance is doubtfulin patientswith dementia.Morphologicalandfunctionalchangessuchasdelayed gastric emptying, reduced splanchnic blood ﬂow and altered gastric pH can affect the bioavailability of orally administered drugs[6].

4.Antibioticdosagesinspecialpatientpopulations

4.1.Elderlypatientswithchronickidneydisease(CKD)or haemodialysis

The balance between attainment of optimal PD targets of efﬁcacyandsafetyconcernsrepresentsachallengeforclinicians treatingpatientswithCKD.Itisestimatedthat20%ofsubjects aged >60 years are affected by advanced CKD, and in many

Table2

Proposeddoseadjustmentofantibioticsaccordingtovariousdegreesofchronickidneydisease(CKD)derivedfromstudiesspeciﬁcallyconductedinelderlypatients.

Antibiotic CKDclass Recommendeddosage(s)

Meropenem,short-termdurationinfusion[36] CLCr51mL/min 1.0gq8hfor1.0gunitdose 0.5gq8hfor0.5gunitdose CLCr26–50mL/min 1.0gq12hfor1.0gunitdose

0.5gq12hfor0.5gunitdose CLCr10–25mL/min 0.5gq12hfor1.0gunitdose 0.25gq12hfor0.5gunitdose CLCr<10mL/min 0.5gq24hfor1.0gunitdose

0.25gq24hfor0.5gunitdose

HD 0.25–0.5gq24h(additionaldoseafterHD)

Meropenem,extendedorcontinuousinfusion[38] CLCr>100mL/min Highdosageseitheradministeredoverextendedorcontinuousinfusion CLCr>50–100mL/min 24-hcontinuousinfusion:3.0gq24h

CLCr50mL/min Extended1.0gq8h CLCr<30mL/min 400mgq12–24h

Levoﬂoxacin[37] CLCr80mL/min 500mgq12h

CLCr60–79mL/min 750mgq24h CLCr40–59mL/min 500mgq24h CLCr20–39mL/min 750mgq48h CLCr<20mL/min 500mgq48h

HD 500mgq48h

CLCr,creatinineclearance;HD,haemodialysis.

(7)

Westerncountrieselderlyadultsnowrepresentthemostrapidly growing population initiating haemodialysis [31]. Dose adjust- mentofantibioticsinelderlypatientswithCKD isimportantto ensureefﬁcacyandalowriskofadverseevents.

A crucial aspect is the method for the estimation of renal functioninelderlysubjects.Anorexia,cachexiaandmalnutrition areveryfrequentlyobservedinelderlypatients,especiallythose residing in LTCFs. Loss of muscular mass directly affects the formationofcreatinineinmuscletissuesasabreakdownproduct ofcreatinephosphate.SincetheprevalenceofsarcopeniainLTCF residentsrangesbetween40%and85%,creatininelevelscouldlead toanoverestimationofrenalfunctionowingtothereducedmuscle mass[32].Thereisnoexpertaccordanceontheformula[Chronic KidneyDiseaseEpidemiologyCollaboration(CKD-EPI),Modifica- tionofDietinRenalDisease(MDRD)orCockcroft–Gault]thatbest estimates CLCr in elderly patients[33]. Available equations for estimation of glomerular filtrationrate, including CDK-EPI and newer formulae (revised Lund–Malmö, Full Age Spectrum and BerlinInitiativeStudy),appeartobecomparableinelderlypatients [34]. However, each of them may have limitations regarding accuracy.In elderlypatients, theMDRD andCKD-EPI equations significantlyoverestimateCLCr,leadingtodosecalculationerrors for many drugs, particularly in individuals with severe renal impairment[35].Optimally,CLCrshouldbedirectlymeasuredto correctlyassessrenalclearanceandtoavoiddrugoverdosingwith hydrophilicantibioticsinelderlyadults[33].

Several studies focusing on antibiotic dose adjustment in patientswith chronic renal failure and in patientsundergoing haemodialysis have been reviewed. Among these, only three studies(evaluating dose adjustments of meropenem and levo- floxacin)havebeenspecificallyconductedinelderlypatientswith variousdegreesofkidneyfailure[36–38](Table2).However,useof dosagesadjustedforrenalfunctiondoesnotdefinitelyeliminate the risk of underexposure or overexposure. In elderly patients receivinglevofloxacin,a>20%riskofunderexposurewasobserved whenusing500mgevery24h(q24h)or750mgq24hinpatients with CLCr values of 40–59mL/min/1.73m² and 60–79mL/min/

1.73m²,respectively[37].Similarly,a>10%riskofoverexposure wasobservedwhenusing500mgevery48hor500mgevery12h inpatientswithCLCrvaluesof<20mL/min/1.73m²and>80mL/

min/1.73m²,respectively[37].ThesedatasupportaTDM-guided approachtoantibioticdosageadjustmenttopreventdrug-related toxicityinelderlypatients.

4.2.Elderlycirrhoticpatients

Amongcirrhoticpatients,theChild–Pughscoreguidesdosage.

It is knownthathepatic impairment maydirectly orindirectly decrease proteinbinding, metabolism and renal elimination of antibiotics.AlterationsinthePKofcommonlyusedantibioticsthat undergohepatobiliaryclearanceinpatientswithhepaticdisease hasbeen extensivelyreviewed, and doseadjustment hasbeen proposedfordifferentantibiotics,especially thosethatundergo phaseImetabolism,havehighproteinbindingorareassociated withahighfrequencyofhepatotoxicity[39].However,thePKof antibioticsthatdonotundergohepaticmetabolism,suchasmost β-lactams,canalsobeaffectedinpatientswithliverdiseasefor severalreasons.IncreasedVdoccurringinpatientswithadvanced cirrhosis,oedema,ascitesandthird-spaceexpansioncanleadto lowerconcentrationsofhydrophilicantibiotics[40].Inaddition, ﬂuctuationsinrenalfunctionoccurringinpatientswithcirrhosis mayresult in frequently changing or unpredictable changes in antibioticelimination. On the other hand, the PK of lipophilic antibiotics(ﬂuoroquinolones,glycylcyclines,macrolides,lincosa- mides,metronidazole,oxazolidinonesandtetracyclines)areless affectedbychangesintheVdbecausetheynormallydistributeinto

tissuesandcells.However,whenintravascularﬂuiddiffusesinto tissue owing to a low oncotic pressure, which is common in patientswith impairedliver function,a redistribution of drugs fromcellscanoccur[40].Somestudiessuggestedthatadminis- teringβ-lactamsbycontinuousinfusionincirrhoticpatientscould beassociatedwithabetteroutcome[41].

However, since it is difﬁcult to predict changes in drug concentrations in cirrhotic patients, TDM represents the best methodtoachieve the mostappropriate concentrationsin this population. Unfortunately, no studies on antibiotic dosages in elderly patientswith various degree of liver failure have been conducted.

5.Adverseeffectsofantibioticsinelderlypatients

Aspolypharmacyisfrequentinelderlypatients,prescriptionof antibiotictherapycanleadtoevenmoreadverseeventsbecauseof interactions with common drugs. Common adverse events detectedinelderlypatientsaredescribedinTable3[17,42–58].

Certainadverseeventsarespeciﬁcallypertinenttotheelderly population. Pre-existing cardiovascular diseases make elderly patients more vulnerable to cardiac side effects of antibiotic.

Although not specifically designed for elderly patients, the CLARICOR trial, which included >4000 patients with a mean ageof65yearswithstablecoronaryheartdisease,showedhigher ratesofcardiovascularandall-causemortalityinpatientstreated with clarithromycin versus placebo [42]. Other studies that included elderly patients did not confirm a high risk of arrhythmias in patients treated with macrolides[43,44]. Con- versely,theuseoffluoroquinolonesrequiresawarenessaboutthe risk ofcardiac events,mainlyrepresented byQTc prolongation and risk of arrhythmia [45]. No differences in cardiac safety profilehavebeenobservedbetweenlevofloxacinand moxiflox- acin,butthelattercausesQTcprolongation[59].Severalstudies haveindicatedthatfluoroquinoloneusemaybeassociatedwith an increased risk of aortic aneurysm or dissection [46,59].No studieshavespecificallybeenconductedinelderlypatients,buta subgroupanalysisofalargecohortofpatientsshowedthatthe risk of aortic aneurysm or dissection was not higher among patients aged >65 years receiving fluoroquinolones compared withotheragegroups[46].Thus,cliniciansshouldpayattention when quinolones are prescribed in elderly patients with pre- existingvasculardiseases.

Risk of cutaneo-muscular toxicity, mainly represented by increased creatine phosphokinase (CPK) and tendon ruptures and Stevens–Johnson syndrome associated with the use of daptomycin and ﬂuoroquinolones, respectively, are higher in elderlycomparedwithyoungerpatients[45,46,59,60].

Thrombocytopenia occurs in 24% of elderlypatients treated withlinezolidandisassociatedwithbaselineplateletcount(low baseline platelet count is associated with a higher risk) and durationoftreatment[48,61].Vancomycinandaminoglycosides areamongthemostfearedantibioticsinelderlypatientsbecause ofthehighriskofnephrotoxicity.Thetwomostimportantfactors associated with aminoglycoside renal toxicity appear to be treatmentduration(3days)andconcomitantuseofnephrotoxic drugs such angiotensin-converting enzyme inhibitors or loop diuretics [27]. PK monitoring could be helpful in order to individualiseaminoglycosidedoseandintervalschemes[27].

Metabolicadverseevents, suchas increasedrisk ofhypogly- caemiainelderlypatientstreatedwithhypoglycaemicagents[56]

andelectrolytesdisorders[57],shouldbealsomonitoredinelderly patients. Although rare, neurotoxicity induced by antibiotics is often unpredictable and is potentially dangerous in elderly patients[58,62].Thehigherriskclassesincludeﬂuoroquinolones, macrolides,sulfonamides,nitrofuransandsomeβ-lactams,suchas

330 M.Falconeetal./JournalofGlobalAntimicrobialResistance22(2020)325–333

(8)

piperacillin/tazobactam, cephalosporins (particularly cefepime) andcarbapenems[62].Antibiotic-inducedneurotoxicityincludesa greatrangeofmanifestations,includingdeliriumand psychosis, andtheunderlyingmechanismsareunknown.Seizuresandnon-

convulsive status epilepticus are potentially life-threatening complications of carbapenem and cefepime therapy [63] and couldoccurinelderlypatients,especiallyifpre-existingdiseaseof thecentralnervoussystemispresent.

Table3

Antibioticadverseeventsbyorgansystem.

Toxicity Reference Typeofstudy Antibiotic Objective Patients Mean

age (years)

Results

Cardiac (prolonged QTand arrythmias)

[42] RCT Clarithromycin

vs.placebo

Todeterminewhetherclarithromycin affectsmortalityandcardiovascular morbidityinpatientswithstable coronaryheartdisease

4373 65 Highercardiovascularmortalityin clarithromycingroup

[43] Retrospective Macrolidesvs.

non-macrolides

Toevaluatethe30-dayriskofventricular arrhythmiaassociatedwithmacrolides

616,59 vs.

705,132

73.7 Nohigherriskofventriculararrhythmia inmacrolidegroup

[44] Retrospective Azithromycin Toevaluatetheassociationof azithromycinuseandcardiovascular eventsinpatientswithpneumonia

73,690 77.8 Nohighercardiovascularevents,no higherarrhythmiaincidence

[45] RCT Moxiﬂoxacin

(MFX)vs.

levoﬂoxacin (LVX)

Toassessthecardiacrhythmsafetyof MFXvs.LVXinpatientswithCAP

387vs.

195

78.1 (MFX) vs.77.5 (LVX)

8.3%ofMFX-treatedpatientsand5.1%of LVX-treatedpatientshadacardiacevent (arrhythmiaorcardiacarrest)(P=0.29) QTcprolongationinMFXgroup Vascular(aortic

aneurysm and dissection)

[46] Retrospective Fluoroquinolones Toinvestigatetheriskofaortic aneurysmordissectionamongpatients receivingﬂuoroquinoloneoramoxicillin

360,088 vs.

360,088

67.9 Fluoroquinoloneuseassociatedwithan increasedriskofaorticaneurysmor dissection,buttheriskwasnotaffected byage

Cutaneo- muscular

[47] Prospective Fluoroquinolones Toevaluateadverseeventsinpatients treatedwithquinolones

657,950 65 Highincidenceoftendonruptures duringquinolonetreatment Haematological [48] Retrospective Linezolid Toevaluateefﬁcacyandsafetyof

linezolidintheelderly

50 81 Thrombocytopenia(24%ofpatients) wasassociatedwithbaselineplatelet countanddurationoftreatment Hepatological [49] Case–control Clarithromycin,

cefuroxime, quinolones

Todeterminetheassociationbetween acuteliverinjuryandpreviousexposure toanantibioticagent

144 77 Moxiﬂoxacinandlevoﬂoxacinwere associatedwithanincreaseinriskof acuteliverinjury

Renal [50] Retrospective Gentamicin, amikacin

Toevaluatetheincidenceofkidney injuryinpatientstreatedwith aminoglycosides

278 74 Highincidenceofkidneyinjuryduring aminoglycosidetherapy

[51] Retrospective Vancomycin Todeterminewhetherhigher vancomycindosingstrategiesleadto excessiveratesofadverseeventsinthe elderly

92 77 Nephrotoxicityoccurredin32%of patients

Age>80yearsisariskfactorfor

nephrotoxicity [17] Retrospective Vancomycin Todeterminetheoverallrateof

developmentofnephrotoxicityin elderlypatientsreceivingvancomycin

124 67 Patientswithhigh(15mg/L)rather thanlow(<15mg/L)average vancomycintroughshaveelevated nephrotoxicity

[52] Prospective Aminoglycosides Toassessthesafetyofaminoglycosides inelderlypatients

249 75 Increaseof>50%increatininevalues wasrecordedin12.4%ofpatients.Renal damagecorrelatedwithahigh aminoglycosidetroughlevel(>1.1m^g/

mL) [53] Retrospective Colistin Todetermineriskfactorsforcolistin-

associatednephrotoxicityinpatients whoreceivedcolistin

129 61.7 Nephrotoxicityoccurredin48%of patients

Advancedageisariskfactorfor nephrotoxicity

[54] Retrospective Macrolides Toevaluatetheriskofacuteadverse eventsinelderlytreatedwithmacrolides andacalciumchannelblocker(CCB)

190,309 76 Co-prescribingclarithromycinwitha CCBwasassociatedwithahigherriskof acutekidneyinjurythanco-prescribing azithromycin

[55] Prospective Piperacillin/

tazobactam(TZP)

Toclarifytheefﬁcacy,safetyand pharmacokineticsofTZPinlateelderly patients

22 85 Nephrotoxicitywasobservedin18.2%of cases

CLCr<40mL/min,renalimpairmentwas ariskfactorforseverenephrotoxicity Metabolism [56] Retrospective Allantibiotics Todeterminetheriskofhypoglycaemia

inolderpatientstreatedwith sulfonylureaswhoﬁllaprescriptionfor anantimicrobialdrug

68,186 >65 Clarithromycin,levoﬂoxacin, trimethoprim/sulfamethoxazole, metronidazoleandciproﬂoxacinwere associatedwithhigherratesof hypoglycaemia

Electrolyte disorders

[57] Retrospective Allantibiotics Toevaluateincidenceofhypokalaemia inpatientstreatedforboneinfections

150 59 Olderageisassociatedwithincreased riskofhypokalaemia

Neurological [58] Retrospective Ertapenem Tocomparethecharacteristicsof ertapenem-treatedadultpatientswith andwithoutdevelopmentofseizures

165 79 Seizuresoccurredinthe1.9%ofpatients treatedwithertapenem

RCT,randomisedcontrolledtrial;CAP,community-acquiredpneumonia;CLCr,creatinineclearance.