Infectious complications associated with parenteral nutrition in intensive care unit and non-intensive care unit patients

(1)

w w w . e l s e v i e r . c o m / l o c a t e / b j i d

The

Brazilian

Journal

of

INFECTIOUS

DISEASES

Original

article

Infectious

complications

associated

with

parenteral

nutrition

in

intensive

care

unit

and

non-intensive

care

unit

patients

Pedro

Henrique

Comerlato

a,∗,1

_,

_Joel

_Stefani

b,1

_,

_Marina

_Verc¸oza

_Viana

a,c,1

_,

Luciana

Verc¸oza

Viana

a,d,1

a_Universidade_Federal_do_Rio_Grande_do_Sul,_Hospital_de_Clínicas_de_Porto_Alegre,_Faculdade_de_Medicina,_Programa_de_{Pós-Graduac¸ão} emCiênciasMédicas:Endocrinologia,PortoAlegre,RS,Brazil

b_Universidade_Federal_do_Rio_Grande_do_Sul,_Faculdade_de_Medicina,_Porto_Alegre,_RS,_Brazil

c_Universidade_Federal_do_Rio_Grande_do_Sul,_Hospital_de_Clínicas_de_Porto_Alegre,_Servic¸o_de_Medicina_Intensiva,_Porto_Alegre,_RS,_Brazil d_Universidade_Federal_do_Rio_Grande_do_Sul,_Hospital_de_Clínicas_de_Porto_Alegre,_Servic¸o_de_Nutrologia,_Porto_Alegre,_RS,_Brazil

a

r

t

i

c

l

e

i

n

f

o

Articlehistory:

Received11December2019

Accepted24February2020

Availableonline20March2020

Keywords:

Parenteralnutrition

Catheter

Centralvenouscatheter

CLABSI

Hospital-acquiredinfection

a

b

s

t

r

a

c

t

Introduction:Malnutritionisassociatedwithanincreasedriskofcomplicationsin

hospi-talizedpatients,andparenteralnutrition(PN)isusedwhenoralorenteralfeedingisnot

possible.ThisstudyaimedatanalyzingassociationsbetweenPNcharacteristicsand

infec-tiouscomplicationsinhospitalizedpatients.

Materialandmethods:This wasaretrospectivecohortstudyconductedina tertiarycare

universityhospital.DatafromconsecutiveadultpatientssubmittedtoPN(January2016

to December 2017;ICU and ward)were reviewed bymeans ofan electronic database.

Patient’sclinicalcharacteristics,PNprescriptionandcatheterinsertionproceduredatawere

extractedandanalyzed.Themainoutcomewasthedevelopmentofcentralline–associated

bloodstreaminfection(CLABSI).Thesecondaryoutcomeswereotherinfectious

complica-tionsandmortality,aswellasfactorsassociatedwithCLABSI.

Results:Weanalyzed165patientsand247cathetersusedforparenteralnutritioninfusion.

TheCLABSIratewas6.47per1000catheter-days.Intheunivariableanalysis,CLABSIwas

associatedwithlongerhospitalizationtime,longerPNtime,longercathetertime,catheter

insertionperformedbyasurgeonorasurgicalresident,andproceduresperformedoutside

theICU.Inanextendedtime-dependentCoxregression,novariablewasassociatedwith

ahigherriskofCLABSI,andadditionalPNdaysdidnotincreasetherateofCLABSI.The

overallmortalityratewas24.8%.Onlythepatients’comorbidityindexwasassociatedwith

deathinthemultivariableanalysis.

∗ _{Corresponding}_author.

E-mailaddress:pedrocomerlato@gmail.com(P.H.Comerlato).

1 _These_authors_contributed_equally_to_this_work.

https://doi.org/10.1016/j.bjid.2020.02.002

(2)

Discussion: Inourstudy,patientswhoneededPNhadanoverallCLABSIrateof6.47per1000

catheter-days.TheseoutcomeswerenotassociatedwithPNandcathetercharacteristics

studiedafteradjustmentforcathetertime.Theoverallmortalityratewas24.8%anditwas

notassociatedwithPNinmultivariableanalyses,onlywithCharlsoncomorbidityindex.

anopenaccessarticleundertheCCBY-NC-NDlicense(http://creativecommons.org/

licenses/by-nc-nd/4.0/).

Introduction

Malnutritionisassociatedwithanincreasedriskof

complica-tions,highermortalityrate,longerhospitalstays,andhigher

hospitalizationcosts.1 _Nutritional_support_is _an_alternative

to overcome this problem, and it is indicated forpatients

unabletofeedorally.2_There_are_two_available_options:_enteral

nutrition,usuallychosentopreservethepatient’s

gastroin-testinaltransit,3_and_parenteral_nutrition_(PN),_used_when_it

isimpossibletoachievepartialorfullenteralnutrition(EN)

requirements.Apragmaticmulticenterrandomizedclinical

trialevaluatedPNversusENinICUpatientsofdeveloped

coun-triesandfoundnodifferenceinbothnutritionalstrategiesin

termsofnumberoftreatedinfectiouscomplicationsor90-day

mortality.4

The infection rate related to a central venous catheter

(CVC)used for PN varies according to the deﬁnition used.

This rate can reach up to 18 infectious events per 1000

catheter-days,5,6_a_higher_rate_compared_with_central_catheter

infectionsindevicesnotusedforPN(twoinfectiousevents

per 1000 catheter-days in US intensive care units (ICUs)

and6.8infectiouseventsper1000catheter-daysin

develop-ing countries’ ICUs7_). _However, _most _central _line _infection

data come from developed countries whereresources

dif-fer(includingthetypesofPNavailableandthedeviceused

forPN nutrition) from emerging countries.8 _A_multicentric

Brazilianpublication reported 10.22 bloodstream infections

per1000catheters/day,andtheriskfactorsforinfectionwere

multiple–lumen catheters, duration of catheterization and

lengthofstayintheICU,butPNwasnotevaluatedasavariable

inthatstudy.9Indeed,anotherstudyperformedinthesame

countryshowedthatPNwasariskfactorforcentralvenous

catheterinfection.10

MoststudiesoninfectionratesofPNrefertospeciﬁc

pop-ulations, such as critically ill, cancer or trauma patients.5

Few studies evaluated different diseases, non-critically-ill

patients,catheterbundlesandphysiciansexperiencetoinsert

CVC,usingarecentlyinsertedversusanalreadyusedcatheter

for nutrition purposes. Studying this more heterogeneous

cohortmayinfermoreassociationswiththerouteofnutrition

itselfand notregardingspeciﬁcgroups.Furthermore,

char-acteristicsofthe vascularaccess correlated withincreased

oddsofinfectioninPNusersareunknown,andthe

recom-mendationsregardingthebestvascularaccesstoPNhavelow

tovery lowquality ofevidence.11 _Therefore,_the_aim_of_the

presentstudy wastoexaminemediatorsofPNandcentral

line–associatedbloodstreaminfection(CLABSI)ina

tertiary-care-levelhospital.Thesecondaryaimwastoanalyzetherate

ofothercomplicationsinpatientssubmittedtoPN.

Materials

and

methods

A retrospective cohort study was conductedinan 800-bed

tertiary-careuniversityhospitalinthesouthofBrazilthrough

reviewofelectronicmedicalrecordsofalladultinpatients

sub-mittedtoPNintheperiodofJanuary2016toDecember2017.

Patients who receivedPN forless than 72hwere excluded

from theanalyses,aswere thosewho receivedPNthrough

a long-termcatheter(LTC) due totheir out-of-hospital use

and possibilityoflackofnotiﬁcationorevenoccurrenceof

an outcome in another institution. A peripherally inserted

centralcatheter(PICC)wereusedinthehospitalduringthe

studyonlyinexperimentalsituationsandtheywerenot

ana-lyzedbecauseofthepossibilityofbiasduetodifferentiated

carerelatedtoanewtechnology/deviceinthepopulation.The

studywasapprovedbythelocalresearchethicscommittee.

The assistant physician (based on local protocol and

currentguidelines)deﬁnedthechoiceforthetotalor

supple-mentaryPN.2,3_All_of_the_prescribed_solutions_were_two-in-one

(2:1), combining glucose and aminoacids, separately from

intravenouslipidemulsion.Theavailablesolutionsand the

productsusedwereFreseniusKabi—Germany,Aminoven10%,

Lipovenos MCT 20%, and glucose 50%, with electrolytes,

vitamin K, trace elements, and addition of multivitamins.

Glycemiccontrolduringhospitalizationwasanattributionof

theattendingphysician,aswastheCVCinstallation,although

bothare standardizedprocedures.Thelocalprotocol about

carewithcentrallinesincludesqualiﬁedpersonnelanda

bun-dleforbestcareofCVC.12_According_to_our_hospital_protocol,

allphysicianswereencouragedtostartenteralororaldietand

discontinuethePNsolutionassoonaspossibleandthedevice

shouldberemoved,sinceitwouldbenolongernecessary.

Demographics characteristics, clinical data,13–15 _and

aspectsofCVCinsertionwere reviewed.Dailyrecords from

theinsertionoftheﬁrstCVCusedforPNuntildischargeor

deathwererevised.Patientswereclassiﬁedaccordingtothe

indicationforPN:totalPN,whentherewascontraindication

orintolerancetoanyamountofenteralororaldietor

supple-mentalPN,whenitwasnotpossibletoachievethenutritional

goal only withenteral or oral diet. For each patient, total

hospitalization time, totalPN time, totaltime withCVCin

use,andtimebetweentheCVCinsertionandstartofPNwere

calculated.

The main outcome was development of central

line–associated bloodstream infection (CLABSI), deﬁned

asCVCwithclinicalsigns ofinfection andnoothersource

ofbacteremia,exceptthecatheterupto48haftertheCVC’s

withdrawal, plus1) onepositiveblood cultureforaknown

(3)

16 exclusions:

1 PN started in another hospital withoutd full description

6 PNS interrupted before 72 h 2 PNS received through PICC 7 PNs received through LTC

165 medical charts included 181 medical charts reviewed

247 CVCs analyzed 1 patient needed 6 CVCs 2 patients needed 5 CVCs 4 patients needed 4 CVCs 4 patients needed 3 CVCs 49 patients needed 2 CVCs 105 patients needed 1 CVCs

Fig.1–FlowchartofIncludedandExcludedPatients.CVC, centralvenouscatheter;LTC,long-termcatheter;PICC, peripherallyinsertedcentralcatheter.

We also recorded as secondary outcomes other

infec-tions(pulmonaryinfection,abdominalinfection,bacteremia

not related to CVC, fungemia, urinary infection, operative

wound infection based on clinical diagnosis), death, and

hyperglycemia, as well as factors associated with CLABSI.

Hyperglycemiawasarbitrarilydeﬁnedasatleastfourepisodes

ofcapillaryglucose>200mg/dlduringPNinfusion;aneedfor

aregularinsulinprescriptiontoachieveglycemiccontrol;ora

descriptionofdecompensateddiabetes.

Sample size calculation, considering 18.3% cumulative

incidenceofCLABSIasfoundinastudyperformedina

sim-ilarpopulationinthesamehospital,10_was_estimated_in₂₃₁

cathetersevaluationtoidentifyfactorsassociatedwithCVC

infection,consideringapowerof95%andamarginoferrorof

5%.

Statisticanalyseswereconductedasappropriated.

Contin-uousvariableswerereportedasmeanandstandarddeviation,

medianandinterquartilerange, ornumberofpatientsand

percentages.Thedifferencesbetweenthegroupswere

ana-lyzedwithStudent’st-test, Mann-Whitney U-test,or␹2_, _as

appropriate.Generalizedestimatingequationswereusedfor

comparisonin relation to CVC (more than one device per

patient is possible). In multivariable analysis independent

variables were included in the model according to their

signiﬁcanceintheunivariateanalysis(p<0.05)ortheir

bio-logical importance. The results were expressed as hazard

ratio(HR)withtheirrespective95%conﬁdenceintervals(CI).

For analysis ofCVC infection, Cox regression adjustments

were performed for time-dependentcovariables (CVCtime

in days) until the occurrence of the patient’s ﬁrst event.

TheothercathetersinsertedaftertheoccurrenceofCLABSI

Table1–Characteristicsoftheincludedpatients (n=165). Characteristics Value Age(years) 56.3(±16.6) Malesex 92(55.8%) Weight(kg) 70.15(±16.6) BMI(kg/m2₎ _25.42_(±5.6) Surgicaladmission 132(80%) Abdominalsurgery 119(72.1%)

PNstartedintheICU 71(43%)

Hospitalizationtime(days) 43(27.5−64.5)

Charlson(comorbidityindex) 4(2−6)

SAPS3a _63.4₍_±14.5) SOFAa ₅₍₃₋₇₎ Vasoactivedrugsa ₂₁_(12.7%) PNtime(days) 15(9−25) TotalPN 125(75.7%) SupplementalPN 40(24.2%) Comorbidities DM 35(21.2%)

Coronaryarterydisease 16(9.7%)

Heartfailure 6(3.6%)

Stroke 16(9.7%)

Pulmonarydisease 20(12.1%)

Hepaticdisease 8(4.8%)

Cancer 73(44.2%)

Chronickidneydisease 13(7.9%)

PNdailyprescription Calories(kcal) 1598(±423.3) Calories(kcal/kg) 25.2(20.2−27.6) Protein(g/kg) 1.5(1.24−1.61) Glucose(g/kg) 3.08(2.52−3.52) Lipids(g/kg) 0.8(0.58−0.91) Outcomes Mortality 41(24.8%) Hyperglycemia 62(37.6%) Anyinfection 107(64.8%) Pulmonaryinfection 28(17%) Abdominalinfection 60(36.4%)

Operativewoundinfection 7(4.2%)

Urinaryinfection 9(5.5%)

BacteremianotrelatedtoCVC 7(4.2%)

CLABSI 24(14.5%)

Fungemia 12(7.3%)

Nrepresentsthenumberofpatients(andpercentage).Mean (± standarddeviation)ormedian(interquartilerange).BMI,bodymass index;ICU,intensivecareunit;SAPS3,simpliﬁedacutephysiology score3;SOFA,sequentialorganfailureassessment;PN,parenteral nutrition;DM,diabetesmellitus;CVC,centralvenouscatheter;and CLABSI,centralline–associatedbloodstreaminfection.

a _Only_collected_in_the₇₁_patients_who_started_PN_in_the_ICU.

were excluded from this analysis. The data were stored

and analyzed in the statistical programs SPSS 22.0 (IBM

SPSS Statistics for Windows, Armonk, NY) and R version

3.5.1 (Foundation for Statistical Computing, Vienna,

Aus-tria). In all analyses, a p-value <0.05 was considered as

statisticallysigniﬁcant. Thestudy was conductedin

accor-dancewithlocalregulationsandwiththecurrentguidelines

for observationalstudies.17 _All _data _were _analyzed

(4)

Table2–UnivariableanalysisforevolutiontoCLABSIathospitalization.

Variables: CLABSI(24patients) No-CLABSI(141patients) p-value

Age(years) 55.9±16.1 56.4±16.7 .77

BMI(kg/m2₎ _26.4_±_5.7 _25.2_±_5.6 _.36

Charlson(comorbidityindex) 5.5(2−6) 4(2−6) .29

Postoperative 18(75%) 113(80.1%) 1

Hospitalizationtime(days) 66(53.5−82) 38(27−59) .0001

PNtime(days) 30(11.5−43) 14(9−23) .003

DM 5(20.8%) 30(21.3%) 1

Hyperglycemia 8(33.3%) 54(38.3%) .81

PNstartedinICU 9(37.5%) 61(43.9%) .719

SupplementalPN 2(8.3%) 38(27%) 0.09

Caloriesinfused/day(kcal) 1537±402.7 1608±427 .448

Proportionofcaloriesfromglucose(%) 45(42–47.5) 45(42–48) .74

Procedureperformedbyasurgeona,b ₈₁_±_7.9%_(61−92%) ₅₆_±_3.7%_(49−64%) _.025

ProcedureperformedinICUb ₁₆_±_7.4%_(6−36%) ₃₉_±_3.6%_(32−46%) _.03

CVCtime(days)b _20.6_±_1.6_{(17.4−23.7)} _17.27_±_0.8_{(15.7−18.8)} _.034

Double-lumenb ₉₆_±_3.5%_(78−100%) ₉₅_±_1.5_(91−97%) _.741

Subclavian-siteb ₃₉_±_8.5%_(24−56%) ₃₈_±_3.3%_(32−45%) _.933

Ultrasound-guidedb ₃₇_±_8.3%_(23−54%) ₅₂_±_3.6%_(45−59%) _.123

PNinfusedinarecentlyinserted(<48h)CVCb ₈₂_±_7.5%_(63−93%) ₇₇_±_2.8%_(71−82%) _.55

BMIisbodymassindex;ICU,intensivecareunit;PN,parenteralnutrition;DM,diabetesmellitus;CVC,centralvenouscatheter;andCLABSI, centralline–associatedbloodstreaminfection.ThecellsrepresentN(%),mean±SDormedian(interquartilerange).

a _Surgeon_or_a_surgical_resident.

b _Estimated_marginal_mean_±_standard_error_and_95%_Wald_conﬁdence_interval,_through_analysis_by_GEE_(log-gamma_{distribution).}

Results

Atotal of 181medical chartsofpatients who received PN

between January 2016and December 2017 (24 consecutive

months)werereviewed.Sixteenpatientswereexcluded

leav-ing165patientsand247CVCs(Fig.1).

Descriptionofstudycohort

Table1summarizesthemaincharacteristicsoftheincluded

patients.Mostpatientsweremale,56.3±16.6yearsold,

over-weight,medianCharlsonindexwas4,andthemostfrequent

comorbiditywascancer.Meannutritionalprescription,caloric

andproteic,wasadequate.Overallmortalityratewas24.8%.

Themostprevalentoutcomewasanyinfectiouscomplication

duringPNadministration,mainlyduetoabdominalinfection.

Clinicaloutcomes

Table 2 summarizes the ﬁndings associated with CLABSI.

Therewere 28episodes ofCLABSI (11.3%of247CVCs),but

someeventsoccurredinthesamepatient.Atleastoneepisode

ofbloodstreaminfectionoccurredin24patients(14.5%of165

patients).ConsideringthetimeusedforeachCVC,theCLABSI

indexwas6.47per 1000CVC-days.Intheunivariable

anal-ysis,CLABSIwasassociatedwithlongerhospitalizationtime,

longerPNtime,longerCVCtime,catheterinsertionperformed

byasurgeonorasurgicalresident,andproceduresperformed

outsidetheICU.Noassociationwasfoundwithtotalcalories

ofPN,proportionofmacronutrients,hyperglycemia,

supple-mentalPN,useofultrasoundorcomorbiditiesatthebeginning

ofPN.Furthermore,noCLABSIoccurredinlessthan5daysof

CVCuse(medianof15days),andusingarecentlyinserted

device(withlessthan48hofuse)whenstartingPNwasnot

associatedwithalowerrateofCLABSI.Inanextended

time-dependent Coxregression,novariable wasassociatedwith

ahigherriskofCLABSIintheunivariableandmultivariable

analysis(Table3).Additionalinformationaboutthe247CVC

insertionproceduresisavailableinSupplementaryTable.

Regarding CLABSI epidemiology, coagulase-negative

staphylococciwere present in13 cases(46.4%),followedby

fungalinfections(Candida)ineightcases(28.6%)and

Staphy-lococcus aureus in two cases (7.1%). Klebsiella, Enterococcus, Pseudomonas, Enterobacter, and Escherichia were responsible

for one case of CLABSI each (3.6%). The median time for

bloodculturepositivityinCLABSIcaseswas13.9h(12−24h)

forperipheralbloodculturesand12.2h(9.9–19.8h)forblood

culturescollectedfromthePNpathway.

Overallmortalityratewas24.8%inthisstudy.Higher

Charl-sonindex,startingPNinICU,developmentofanyinfection

during PN administration and development of abdominal

infection during PN administration were related to death

(Table4).Inmultivariateanalysiswiththesevariables,onlythe

Charlsoncomorbidityindexremainedsigniﬁcantlyassociated

withmortality(HR1.175;CI1.052–1.312;p=.004).

Discussion

Inthisstudy,alargesampleofpatientssubmittedtoPNovera

two-yearperiodinauniversityhospitaloftheSouthofBrazil

wasanalyzed.Tothebestofourknowledge,thisisoneofthe

largestcohortsidentiﬁedintheliteratureanalyzingpatients

receiving PN both inthe generalward and inthe ICU

set-tings.Therateofinfectiouscomplicationsintheseindividuals

ishigh.Patientswho neededPN hadahigher incidenceof

CLABSIcomparedtopatientswithCVCandwithoutPNinthe

(5)

asso-Table3–EvolutiontoCLABSIinatime-dependentCoxregression.

HR 95%CI p-value

Univariabletime-dependent

Procedureperformedbyasurgeona _2.235 _0.82−6.07 _.11

NumberofpreviousCVCneededforPN 1.148 0.42−1.76 .7

PNtimeuntilcurrentCVC 1.002 0.94−1.05 .92

TotaltimeofPN 0.991 0.99−1.02 .15

HospitalizationtimeuntilcurrentCVC 1.001 0.97−1.02 .91

Totaltimeofhospitalization 0.995 0.99−1.01 .42

Multivariabletime-dependent

Procedureperformedbyasurgeona _2.215 _0.81−6.01 _.11

TotaltimeofPN 1.009 0.99−1.02 .16

ExtendedCoxmodelfortime-dependentcovariates,through ¨R¨survivalpackage.HR:HazardRatio;CI:95%conﬁdenceinterval.Rsquare=0.019. Concordance=0.566.Likelihoodratiotest=4.38.Waldtest4.28.Logranktest4.54p=0.1.

a _Surgeon_or_a_surgical_resident.

Table4–Univariableanalysisforevolutiontodeathathospitalization.

Variables Death(41patients) Discharge(124patients) p-value

Age(years) 60.2±17.1 55.06±16.3 .087

BMI(kg/m2₎ _24.1_±_3.9 _25.8_±₆ _.099

Charlson(comorbidityindex) 5(4−7) 3(2−6) .001

Postoperative 33(80.5%) 99(79.8%) 1

Hospitalizationtime(days) 43(29−67) 43(27−63.75) 0.76

PNtime(days) 17(9−25) 15(9−24.5) .815

DM 12(29.3%) 23(18.5%) .21

Hyperglycemia 19(46.3%) 43(34.7%) .25

PNstartedinICU 26(63.4%) 45(36.3%) .003

AnyinfectionduringPN 33(80.5%) 74(59.7%) .026

AbdominalinfectionduringPN 21(51.2%) 39(31.5%) .036

PulmonaryinfectionduringPN 10(24.4%) 18(14.5%) .22

CLABSIduringPN 5(12.2%) 19(15.3%) .8

SupplementalPN 7(17.5%) 33(26.6%) .3

Caloriesinfused/day(kcal) 1521.5±383.6 1623.4±434.1 .18

Proportionofcaloriesfromglucose(%) 45(43–48) 45(41.2–48) .79

BMIrepresentsbodymassindex;ICU,intensivecareunit;PN,parenteralnutrition;DM,diabetesmellitus;CVC,centralvenouscatheter;and

CLABSI,centralline–associatedbloodstreaminfection.ThecellsrepresentN(%),mean±SDormedian(interquartilerange).

ciatedwithCLABSIandadditionaldaysofPNdidnotincrease therateofCLABSIinmultivariateanalysesinourstudy.

Inanearlierstudyconductedinthesamehospitalalmost 20yearsearlier,10_PN_was_shown_to_be_{independently}

associ-atedwithCLABSIinmultivariateanalysis.Thatstudydiffers

fromthepresentinvestigationforincludingonlyICUpatients,

and becausemicrobiological analysesofall patients(blood

culture or catheter tip) were performed. The association

betweenPNandinfectioncouldbeduetocolonizationofthe

device.Probablyforthesamereason,atwo-foldhigherrateof

CLABSIper1000catheterdayswasidentiﬁedincomparison

withthecurrentstudy,althoughimprovementsinprocedures

andincathetercarethathavebeenestablishedovertimemay

havealsoinﬂuencedthisdifference.

Thehigh incidence ofCLABSI found in our study (6.47

per1000CVC-daysor11.7%ofallCVCs)whencomparedto

patientswithCVCandwithoutPNintheliterature18–20_is_still

withintherange(whichreaches18.8per1000CVC-days)ofthe

internationalliteratureforPN-associatedCVCinfection6_)._In_a

time-dependentCoxregression,PNtimewasnotan

indepen-dentfactorthatcouldjustifyahigherincidenceofCLABSIin

thispopulation.Itisdifﬁculttoidentifyreasonsforthis

inci-dence,sincethestudywasconductedinauniversityhospital

accreditedbytheJointCommission21_and_speciﬁc_bundles_for

CVCcareareavailableinourhospital.Nevertheless,Brazilis

anemergingcountryanddataaboutcatheterinfection,

espe-ciallyinpatientsreceivingPN,arescarce.

Dissanaikeetal.22_found_an_association_between_CLABSI

and higher total calorie infusion, which was not observed

in our cohort. Most of the patients in Dissanaike’s study

received morethan 30−40kcal/kg/day,incontrastwith the

current study, wherethe local protocol encouraged a goal

of 22−25kcal/kg/day and few patients received morethan

30kcal/kg/day, in accordance with recent guidelines.3 _We

believethatsuchﬁndingsindicatethatavoiding

hyperalimen-tationmayreducetherateofCLABSIandotherunfavorable

outcomes,as alreadydemonstrated bystudiesthat limited

totalcaloriesandcomparedparenteralandenteralnutrition

usingthesamecalorictarget.4 _One_possible_explanation_for

the high CLABSI rate in our study was the use of

two-in-onebagsseparatedfromintravenouslipidemulsionthatare

supposed tobe associatedwithanincreasedrisk of

(6)

stilllimitedandnotsufﬁcienttoendorseorrefutesuchan

association.23

ThepresentstudydidnotidentifylowerratesofCLABSI

whenanewCVCwasinstalledafterindicationofPN,thus

notjustifyingtheneedforanewdevice orreplacementof

theCVCwheninitiatingsuchtherapy.Othercatheter-related

factors,suchasthenumberoflumens,werealsonot

associ-atedwithhigherchanceofinfectioninourstudy,althoughthe

analysiswasnotrobustenoughbecauseofthelowprevalence

ofmono-lumencatheters(lessthan5%)usedinourhospital.

Therefore,itisimpossibletorefutethisassociationfoundin

theliterature,24andalthoughrecommended,thereisapaucity

ofevidenceregardingPN-dedicatedlumens.25

Ourmortalityratewashigh,andtheage-adjustedCharlson

comorbidityindexindicatesthatoursampleofpatientswas

sickerthanpopulationinotherPNstudies,probablyjustifying

thehighermortality.26,27_This_comorbidity_metrics_is_the_most

commonlystudiedprognosticmeasureofillnessburdenin

clinicalresearch28_and_is_probably_related_to_increased_rates_of

chronicdiseaseandmortality.29–31_Our_study_failed_to_identify

prolongedhospitalizationorPNtimeasindependentfactors

tojustifythisrate.Onlygreaternumberofcomorbiditieswas

associatedwithmortalityinthemultivariateanalysis.

Amongthe limitations,the studymethodologydoes not

allowforcause-effectinference,althoughitispossibleto

gen-eratehypotheses.Multivariateanalysisandlogisticregression

wereperformedtomitigatebiasand confounding.

Further-more,ourhighrateofinfectiondoesnotinvalidatetheﬁnding

thatCLABSIwasnotassociatedwithspeciﬁcsofPNor

vascu-laraccesscharacteristics.Inaddition,theretrospectivedesign

may hinder outcome recovery and related factors due to

underreportinginthemedicalrecords.Tocounteract

under-reporting,we choselaboratory resultsand the outcome of

hospitalization(death or discharge)asthe mainoutcomes.

Thestudywasnotpoweredtodetectmortalitydifferencea

priori,andthisaspectshouldbeconsideredwhenanalyzing

data.

TheexclusionofLTCandPICCfromtheanalysesisalsoa

limitation.LTCmayleadtounderreportingduetotheir

out-of-hospitaluseandpossibilityoflackofnotiﬁcationoreven

occurrenceofanoutcomeinanotherinstitution.Ithasalready

beenstatedthat PICCwereusedinthehospitalduringthe

studyonlyinexperimentalsituationsandtheywerenot

ana-lyzedbecauseofthepossibilityofbiasduetodifferentiated

careinvolvinganewtechnology.Asamitigatingfactor,less

than10ofthesedevices(sevenLTCandtwoPICC)wereused

forPNinthehospitalinthestudyperiod(3.6%ofallcatheters

usedforPN),possiblynotaffectingtheresults.

Inconclusion,patientswhoneededPNinourstudyhad

aconsiderablerateofCLABSIandotherinfectious

complica-tions.NovariablewasassociatedwithhigherriskofCLABSI

intheunivariateandmultivariateanalysesafteradjustment

forcathetertime.Themortalityratewashighandnot

asso-ciatedwithPNinmultivariateanalyses,onlywithCharlson

comorbidityindex.

Declarations

of

interest

None.

Acknowledgements

We are grateful to PhD VanessaBielefeldt Leotti for

assis-tance instatisticalanalysis. Supportforthepublicationfee

wasprovidedbyFIPE-HCPA(Fundac¸ãodeIncentivoàPesquisa

eEventos-HospitaldeClínicasdePortoAlegre).

Appendix

A. Supplementary

data

Supplementary material related to this article can be

found, in the online version, at doi:https://doi.org/10.

1016/j.bjid.2020.02.002.

r

e

f

e

r

e

n

c

e

s

1.CorreiaMI,WaitzbergDL.Theimpactofmalnutritionon morbidity,mortality,lengthofhospitalstayandcosts evaluatedthroughamultivariatemodelanalysis.ClinNutr. 2003;22:235–9.

2.McClaveSA,TaylorBE,MartindaleRG,WarrenMM,Johnson DR,BraunschweigC,etal.Guidelinesfortheprovisionand assessmentofnutritionsupporttherapyintheadultcritically illpatient:SocietyofCriticalCareMedicine(SCCM)and AmericanSocietyforParenteralandEnteralNutrition (A.S.P.E.N.).JPEN-ParenterEnter.2016;40:159–211.

3.Clinicalguidelinesfortheuseofparenteralandenteral nutritioninadultandpediatricpatients,2009.JPEN-Parenter Enter.2009;33:255–9.

4.HarveySE,ParrottF,HarrisonDA,BearDE,SegaranE,BealeR, etal.Trialoftherouteofearlynutritionalsupportincritically illadults.NewEnglJMed.2014;371:1673–84.

5.GavinNC,ButtonE,KeoghS,McMillanD,RickardC.Does parenteralnutritionincreasetheriskofcatheter-related bloodstreaminfection?Asystematicliteraturereview. JPEN-ParenterEnter.2017;41:918–28.

6.YilmazG,KoksalI,AydinK,CaylanR,SucuN,AksoyF.Risk factorsofcatheter-relatedbloodstreaminfectionsin parenteralnutritioncatheterization.JPEN-ParenterEnter. 2007;31:284–7.

7.RosenthalVD,BijieH,MakiDG,MehtaY,ApisarnthanarakA, MedeirosEA,etal.InternationalNosocomialInfection ControlConsortium(INICC)report,datasummaryof36 countries,for2004-2009.AmJInfectControl.2012;40:396–407.

8.HajjejZ,NasriM,SellamiW,GharsallahH,LabbenI,Ferjani M.Incidence,riskfactorsandmicrobiologyofcentral vascularcatheter-relatedbloodstreaminfectioninan intensivecareunit.JInfectChemother.2014;20:163–8.

9.BicudoD,BatistaR,FurtadoGH,SolaA,MedeirosEASd.Risk factorsforcatheter-relatedbloodstreaminfection:a prospectivemulticenterstudyinBrazilianintensivecare units.BraziJInfectiDis.2011;15:328–31.

10.BeghettoMG,VictorinoJ,TeixeiraL,AzevedoMJ.Parenteral nutritionasariskfactorforcentralvenouscatheter–related infection.JPEN-ParenterEnter.2005;29:367–73.

11.KovacevichDS,CorriganM,RossVM,McKeeverL,HallAM, BraunschweigC.AmericanSocietyforParenteralandEnteral NutritionGuidelinesfortheselectionandcareofcentral venousaccessdevicesforadulthomeparenteralnutrition administration.JPEN-ParenterEnter.2019;43:15–31.

12.O’GradyNP,AlexanderM,BurnsLA,DellingerEP,GarlandJ, HeardSO,etal.Guidelinesforthepreventionofintravascular catheter-relatedinfections.ClinInfectDis.2011;52:e162–93.

(7)

13.CharlsonME,PompeiP,AlesKL,MacKenzieCR.Anewmethod ofclassifyingprognosticcomorbidityinlongitudinalstudies: Developmentandvalidation.JChronDis.1987;40:373–83.

14.MoralezGM,RabelloLSCF,LisboaTC,LimaMFA,HatumRM, DeMarcoFVC,etal.ExternalvalidationofSAPS3and MPM(0)-IIIscoresin48,816patientsfrom72BrazilianICUs. AnnIntensiveCare.2017;7:53.

15.MinneL,Abu-HannaA,deJongeE.EvaluationofSOFA-based modelsforpredictingmortalityintheICU:Asystematic review.CritCare.2008;12:R161.

16.HoranTC,AndrusM,DudeckMA.CDC/NHSNsurveillance deﬁnitionofhealthcare-associatedinfectionandcriteriafor speciﬁctypesofinfectionsintheacutecaresetting.AmJ InfectControl.2008;36:309–32.

17.vonElmE,AltmanDG,EggerM,PocockSJ,GotzschePC, VandenbrouckeJP.TheStrengtheningtheReportingof ObservationalStudiesinEpidemiology(STROBE)statement: guidelinesforreportingobservationalstudies.JClin Epidemiol.2008;61:344–9.

18.ParientiJJ,MongardonN,MegarbaneB,MiraJP,KalfonP,Gros A,etal.Intravascularcomplicationsofcentralvenous catheterizationbyinsertionsite.NEnglJMed. 2015;373:1220–9.

19.LinK-Y,ChengA,ChangY-C,HungM-C,WangJ-T,Sheng W-H,etal.Centralline-associatedbloodstreaminfections amongcritically-illpatientsintheeraofbundlecare.J MicrobiolImmunol.2017;50:339–48.

20.RosenthalVD,Al-AbdelyHM,El-KholyAA,AlKhawajaSAA, LeblebiciogluH,MehtaY,etal.InternationalNosocomial InfectionControlConsortiumreport,datasummaryof50 countriesfor2010-2015:device-associatedmodule.AmJ InfectControl.2016;44:1495–504.

21.JointCommissiononAccreditationofHealthcare

Organizations&JointCommissionResources.JtCommJQual Saf.2005.

22.DissanaikeS,SheltonM,WarnerK,O’KeefeGE.Theriskfor bloodstreaminfectionsisassociatedwithincreased

parenteralcaloricintakeinpatientsreceivingparenteral nutrition.CritCare.2007;11:R114.

23.SlatteryE,RumoreMM,DouglasJS,SeresDS.3-in-1vs2-in-1 parenteralnutritioninadults:areview.NutrClinPract. 2014;29:631–5.

24.TempletonA,SchlegelM,FleischF,RettenmundG,SchobiB, HenzS,etal.Multilumencentralvenouscathetersincrease riskforcatheter-relatedbloodstreaminfection:prospective surveillancestudy.Infection.2008;36:322–7.

25.GavinNC,ButtonE,CastilloMI,Ray-BarruelG,KeoghS, McMillanDJ,etal.Doesadedicatedlumenforparenteral nutritionadministrationreducetheriskofcatheter-related bloodstreaminfections?Asystematicliteraturereview.JInfus Nurs.2018;41:122–30.

26.AdjemianD,ArendtBM,AllardJP.Assessmentofparenteral nutritionprescriptioninCanadianacutecaresettings. Nutrition.2018;49:7–12.

27.AkbarT.Catheter-relatedbloodstreaminfectionsinadults receivingparenteralnutrition:doesthetimetakentoreport bloodculturesimpactonclinicalmanagement?ClinMed. 2019;19:40.

28.deGrootV,BeckermanH,LankhorstGJ,BouterLM.Howto measurecomorbidity:acriticalreviewofavailablemethods.J ClinEpidemiol.2003;56:221–9.

29.PetersonJC,PagetSA,LachsMS,ReidMC,CharlsonME.The riskofcomorbidity.AnnRheumDis.2012;71:635–7.

30.FraccaroP,KontopantelisE,SperrinM,PeekN,MallenC, UrbanP,etal.Predictingmortalityfromchange-over-timein theCharlsonComorbidityIndex:Aretrospectivecohortstudy inadata-intensiveUKhealthsystem.Medicine.

2016;95:e4973.

31.ÖzbilginS,HanciV,ÖmürD,ÖzbilginM,TosunM,YurtluS, etal.Morbidityandmortalitypredictivityofnutritional assessmenttoolsinthepostoperativecareunit.Medicine. 2016;95:e5038.