effectiveness of intraoral chlorhexidine protocols in the...

Effectiveness of Intraoral Chlorhexidine Protocols in thePrevention of Ventilator-Associated Pneumonia: Meta-Analysis and

Systematic Review

Cristina C Villar DDS PhD, Claudio M Pannuti DDS PhD, Danielle M Nery DDS,Carlos M R Morillo DDS, Maria Jose C Carmona MD PhD, and Giuseppe A Romito DDS PhD

BACKGROUND: Ventilator-associated pneumonia (VAP) is common in critical patients and re-lated with increased morbidity and mortality. We conducted a systematic review and meta-analysis,with intention-to-treat analysis, of randomized controlled clinical trials that assessed the effective-ness of different intraoral chlorhexidine protocols for the prevention of VAP. METHODS: Searchstrategies were developed for the MEDLINE, EMBASE, and LILACS databases. MeSH terms werecombined with Boolean operators and used to search the databases. Eligible studies were random-ized controlled trials of mechanically ventilated subjects receiving oral care with chlorhexidine orstandard oral care protocols consisting of or associated with the use of a placebo or no chemicals.Pooled estimates of the relative risk and corresponding 95% CIs were calculated with randomeffects models, and heterogeneity was assessed with the Cochran Q statistic and I2. RESULTS: The13 included studies provided data on 1,640 subjects that were randomly allocated to chlorhexidine(n � 834) or control (n � 806) treatments. A preliminary analysis revealed that oral application ofchlorhexidine fails to promote a significant reduction in VAP incidence (relative risk 0.80, 95% CI0.59–1.07, I2 � 45%). However, subgroup analyses showed that chlorhexidine prevents VAP develop-ment when used at 2% concentration (relative risk 0.53, 95% CI 0.31–0.91, I2 � 0%) or 4 times/d(relative risk 0.56, 95% CI 0.38–0.81, I2 � 0%). CONCLUSIONS: We found that oral care withchlorhexidine is effective in reducing VAP incidence in the adult population if administered at 2%concentration or 4 times/d. Key words: chlorhexidine; clinical protocols; ventilator-associated pneumonia;meta-analysis; infection; critical care. [Respir Care 2016;61(9):1245–1259. © 2016 Daedalus Enterprises]

Introduction

Ventilator-associated pneumonia (VAP) is defined aspneumonia that develops �48 h after endotracheal intu-bation and initiation of mechanical ventilation.1 VAP isthe second most common nosocomial infection in ICUsand the first most common in patients receiving mechan-ical ventilation.2 The condition is associated with increases

in length of hospitalization and ICU stay, morbidity, mor-tality, and health-care costs.3,4 Despite recent advances indiagnosis and treatment of VAP, it continues to be a med-ical problem of major importance, with an attributablemortality rate between 33 and 50%.5 Thus, preventive in-terventions are needed to limit its occurrence.

The development of VAP is related to microbial colo-nization of the normally sterile lower respiratory tract bymicroorganisms commonly found in the trachea, orophar-

Drs Villar, Pannuti, Morillo, and Romito are affiliated with the Disciplineof Periodontics, School of Dentistry, and Dr Carmona is affiliated withthe Discipline of Anesthesiology, School of Medicine, University of SaoPaulo-Sao Paulo, Brazil. Dr Nery is affiliated with the Complexo Hos-pitalar Municipal de Sao Bernardo do Campo, Sao Bernardo do Campo-Sao Paulo, Brazil.

The authors have disclosed no conflicts of interest.

Correspondence: Cristina C Villar DDS PhD, Division of Periodontics,Department of Stomatology, School of Dentistry, University of Sao Paulo,Avenida Prof. Lineu Prestes, 2227, Sao Paulo-Sao Paulo, 05508-000,Brazil. E-mail: [email protected].

DOI: 10.4187/respcare.04610

RESPIRATORY CARE • SEPTEMBER 2016 VOL 61 NO 9 1245

ynx, stomach, and small or large intestines.6 Although themain route of infection leading to lower respiratory tractinfection remains unknown, the primary source of infec-tion for VAP is thought to be the oropharyngeal tract.7

Based on this, a significant number of studies have inves-tigated the effect of topical oral antiseptics in VAP pre-vention. Among these antiseptics, chlorhexidine gluconatehas attracted considerable attention, as evidenced by nu-merous randomized controlled clinical trials that have in-vestigated the effect of oral chlorhexidine use in VAPprevention.8-25

Results from the aforementioned randomized controlledtrials and meta-analyses26-30 that analyzed the effect oforal care with chlorhexidine on VAP prevention are con-flicting. Discrepant findings may have resulted from dif-ferences in study populations, diagnostic criteria for VAP,chlorhexidine concentration, and frequency of use. Meta-analyses have not reported the impact of specific protocolsof oral care with chlorhexidine on VAP prevention. More-over, previous meta-analysis mixed together outcomes re-ported on intention-to-treat and per-protocol basis. There-fore, in this paper, a systematic review and meta-analysis,with intention-to-treat analysis, of randomized controlledclinical trials was conducted to determine the effectivenessof oral decontamination with chlorhexidine and to com-pare specific protocols of oral care with chlorhexidine inVAP prevention.

Methods

Focused Question

We conducted a systematic review of the literature toassess the following focused PICO (patient or population,intervention, control or comparator, and outcome) ques-tion: In subjects endotracheally intubated and mechani-cally ventilated, does oral decontamination with chlorhexi-dine prevent the development of VAP, when comparedwith placebo or standard care or no treatment? As a secondaim, this systematic review assessed the question: Whichdose, frequency, or mode of use provides the best effect inthe prevention of VAP? This systematic review was re-ported according to the PRISMA statement guidelines.31

Eligibility Criteria

Type of Studies. Only randomized controlled trials thatreported data using an intention-to-treat approach or pro-vided enough information that per-protocol results couldbe adjusted into an intention-to-treat format were eligiblefor this review.

Study Population. The population of interest includedintubated subjects receiving mechanical ventilation.

Type of Intervention and Comparison. Oral decontam-ination protocols using chlorhexidine (test group) werecompared with standard oral care protocols consisting ofor associated with the use of (1) a placebo or (2) no treat-ment.

Outcome Measures. The primary outcome was incidenceof VAP, reported as the number/percentage of affectedsubjects.

Search Strategy

Search strategies were developed for the MEDLINE,EMBASE, and LILACS databases. MeSH terms and keywords were combined with Boolean operators and used tosearch the databases. All searches were done without lan-guage restriction, up to January 2015. The following termswere used: ([chlorhexidine OR “gluconate chlorhexidine”OR “oral decontamination” OR “oral hygiene” OR anti-septics OR “antiseptic decontamination”] AND [“ventila-tor-associated pneumonia” OR VAP OR “nosocomial pneu-monia” OR pneumonia OR intubation OR “mechanicalventilation” OR “intensive care units” OR “critical care”])AND (“clinical trial” OR RCT OR “randomized controlledtrial” OR “randomized controlled clinical trial”). Electronicsearch was complemented by manual searches of the ref-erence lists of selected full articles.

Exclusion Criteria

Reviews, in vitro and animal studies, case reports, ob-servational studies, and studies without control groups werenot included.

QUICK LOOK

Current knowledge

In recent years, many regimens of oral care with chlo-rhexidine have been used on mechanically ventilatedpatients to prevent the development of ventilator-asso-ciated pneumonia (VAP). However, results from ran-domized controlled trials and meta-analysis that ana-lyzed the effect of oral care with chlorhexidine on VAPprevention are still conflicting.

What this paper contributes to our knowledge

Results from this systematic review and meta-analysisindicate that oral care with chlorhexidine is effective inreducing VAP incidence only in the adult populationand if administered at a 2% concentration or 4 times/d.

INTRAORAL CHLORHEXIDINE FOR PREVENTION OF VAP

1246 RESPIRATORY CARE • SEPTEMBER 2016 VOL 61 NO 9

Screening Methods and Data Extraction

Two calibrated reviewers (DMN and CCV) indepen-dently screened titles and abstracts. Studies appearing tomeet the inclusion criteria or those with insufficient infor-mation in the title and abstract to make a clear decision,were selected for full manuscript evaluation, which wascarried out independently by the same 2 reviewers to de-termine study eligibility. Any disagreement was solved bydiscussion with a third reviewer (CMP). Reference lists ofprevious reviews and included studies were hand-searched.Studies that met the inclusion criteria underwent a validityassessment. Reasons for rejecting studies were recorded.Agreement between reviewers was described by kappacoefficient. Data were extracted independently by the samereviewers.

The following data were extracted and recorded: cita-tion, setting and location of the trial, characteristics ofparticipants, characteristics of the intervention (concentra-tion, dose, frequency, and type of application), samplesize, definition of VAP, and length of follow-up.

Quality Assessment and Data Synthesis

Quality assessment of the included studies was performedindependently by 2 reviewers (DMN and CCV), with dis-agreements resolved by a third adjudicator (CMP). Thefollowing 6 domains were assessed as having low risk,high risk, or unclear risk of bias, according to the Co-chrane Collaboration’s tool for assessing risk of bias.32

Then studies were categorized as follows: (1) low risk ofbias, if all domains were met; (2) unclear risk of bias, ifone or more domains were classified as having unclearrisk of bias; and (3) “high risk” of bias, if one or moredomains were not met.

Data Analysis

Analyses were performed using Review Manager 5.3software (Cochrane Information Management System).Data on the incidence of VAP was extracted as dichot-omous variables. Pooled estimates of the relative riskand the corresponding 95% CI were calculated usingrandom effects models. Subgroup statistical heteroge-neity among the studies was assessed with the CochranQ statistic and I2.

Results

The computerized search strategy yielded 211 citations,of which 53 were screened for potentially meeting theinclusion criteria (Fig. 1). Independent screening of ab-stracts led to the rejection of 30 articles (Fig. 1). The fulltext of the remaining 23 publications was obtained for

review and possible inclusion. Scanning of reference listsyielded one additional study (Fig. 1). Of the 24 publica-tions preselected, 9 articles were further excluded for rea-sons indicated in Figure 1. As a result, 13 studies pub-lished in English between September 2000 and November2012 were included in this meta-analysis. The character-istics of the final trials retained are reported in Table 1.

Subject Selection and Characteristics

The 13 included studies provided data on 1,640 subjectswho were randomly allocated to chlorhexidine (n � 834)or control (n � 806) treatments. Studies enrolled subjectsexpected to require orotracheal or nasotracheal intubationand mechanical ventilation.8-10,12,14,15,17-19,21,22,24,25 Amongthese, some studies required mechanical ventilation for atleast 48 h.10,12,22,24 Other studies only included subjectswith medical conditions suggesting an ICU stay of �48 h,15

3 d,18 or 5 d8,9 (Table 1). In 3 studies, research subjects

Fig. 1. Flow chart.



Tab

le1.

Cha

ract

eris

tics

ofth

eIn

clud

edSt

udie

s

Stud

y(Y

ear)

Subj

ect

Sour

ceIn

clus

ion

Cri

teri

aE

xclu

sion

Cri

teri

aD

iagn

ostic

Info

rmat

ion

Exp

erim

enta

lG

roup

Con

trol

Gro

upFo

llow

-Up

Gra

pet

al(2

004)

10Su

rgic

al,

trau

ma

and

neur

osci

ence

ICU

and

Em

erge

ncy

Dep

artm

ent;

Uni

ted

Stat

es

Age

�18

y,en

dotr

ache

ally

intu

bate

dan

dm

echa

nica

llyve

ntila

ted

for

48h

Ede

ntul

ous

subj

ects

CPI

S�

6C

HX

0.12

%so

lutio

n;si

ngle

full-

mou

thap

plic

atio

nde

liver

edas

20sp

rays

(n�

5)or

bysw

abbi

ng(n

�2)

atea

rly

post

-in

tuba

tion

peri

od

Stan

dard

oral

care

(n�

5)Su

bjec

tsw

ere

follo

wed

for

72h

afte

rin

tuba

tion

orun

tilex

tuba

tion

ifex

tuba

ted

befo

re72

hFo

urri

eret

al(2

005)

86

ICU

s(3

inun

iver

sity

hosp

itals

and

3in

gene

ral

hosp

itals

);Fr

ance

Age

�18

y,m

edic

alco

nditi

onsu

gges

ting

ICU

stay

�5

dan

dre

quir

ing

mec

hani

cal

vent

ilatio

nby

orot

rach

eal

orna

sotr

ache

alin

tuba

tion;

only

subj

ects

hosp

italiz

edfo

r�

48h

befo

reIC

Uad

mis

sion

wer

ein

clud

ed

Ede

ntul

ous

subj

ects

,w

itha

trac

heos

tom

ytu

be,

faci

altr

aum

a,po

st-s

urgi

cal

and

requ

irin

gsp

ecif

icor

opha

ryng

eal

care

,al

lerg

yto

CH

X

Tem

pera

ture

�38

°Cor

�36

°C;

new

infi

ltrat

eson

ches

tra

diog

raph

s,le

ukoc

ytos

is(�

10�

103/

mm

3)

orle

ukop

enia

(�3

�10

3/m

m3),

posi

tive

quan

titat

ive

cultu

reof

trac

heal

aspi

rate

(�10

⁶C

FU/m

L)

and/

orbr

onch

oalv

eola

rla

vage

flui

d(�

10⁴C

FU/m

L)

CH

X0.

2%ge

l(n

�11

4);

appl

ied

over

dent

alan

dgi

ngiv

alsu

rfac

esby

nurs

esw

eari

ngst

erile

glov

es,

afte

rm

outh

rins

ing

and

orop

hary

ngea

las

pira

tion;

gel

was

left

inpl

ace,

and

the

oral

cavi

tyw

asno

tri

nsed

afte

rap

plic

atio

n;C

HX

appl

icat

ion

star

ted

with

inth

efi

rst

24h

ofin

tuba

tion;

inte

rven

tion

was

perf

orm

edat

leas

t3

times

/dun

tilda

y28

,di

scha

rge,

orde

ath;

toot

hbr

ushi

ngw

asno

tal

low

ed

Plac

ebo

gel

(n�

114)

;ap

plie

dac

cord

ing

toth

esa

me

oral

care

prot

ocol

used

inth

eex

peri

men

tal

grou

p

Ent

ire

ICU

stay

Koe

man

etal

(200

6)12

5m

ixed

and

2su

rgic

alIC

Us;

Net

herl

ands

Age

�18

y,re

quir

ing

mec

hani

cal

vent

ilatio

nfo

r�

48h,

incl

uded

with

in24

haf

ter

intu

batio

nan

dst

art

ofm

echa

nica

lve

ntila

tion

Prea

dmis

sion

imm

unoc

ompr

omis

edst

atus

,pr

egna

ncy,

phys

ical

cond

ition

not

allo

win

gor

alap

plic

atio

nof

stud

ym

edic

atio

n

New

,pe

rsis

tent

,or

prog

ress

ive

infi

ltrat

eon

ches

tra

diog

raph

s,in

com

bina

tion

with

atle

ast

3of

4cr

iteri

a:re

ctal

tem

pera

ture

�38

°Cor

�35

.5°C

,bl

ood

leuk

ocyt

osis

(�10

�10

3/

mm

3),

and/

orle

ftsh

ift

orle

ukop

enia

(�3

�10

3/m

m3),

puru

lent

aspe

ctof

trac

heal

aspi

rate

,an

da

posi

tive

sem

iqua

ntita

tive

cultu

refr

omtr

ache

alas

pira

tes

(�10

5C

FU/m

l)af

ter

48h

ofm

echa

nica

lve

ntila

tion

CH

X2%

inV

asel

ine

petr

oleu

mje

lly(n

�12

7);

2cm

ofpa

ste

(0.5

g)pu

ton

agl

oved

fing

ertip

and

adm

inis

tere

dto

each

side

ofth

ebu

ccal

cavi

ty,

afte

rre

mov

ing

rem

nant

sof

the

prev

ious

dose

with

asa

line

moi

sten

edga

uze;

inte

rven

tion

was

perf

orm

ed4

times

/d,

until

VA

Pdi

agno

sis,

deat

h,ex

tuba

tion,

orw

ithdr

awal

;C

HX

2%/c

olis

tinin

Vas

elin

epe

trol

eum

jelly

(n�

128)

used

acco

rdin

gto

the

prot

ocol

desc

ribe

dab

ove

Vas

elin

epe

trol

eum

jelly

(n�

130)

,ap

plie

dac

cord

ing

toth

esa

me

oral

care

prot

ocol

used

inth

eex

peri

men

tal

grou

ps

Subj

ects

wer

efo

llow

edun

tilex

tuba

tion,

deat

h,de

velo

pmen

tof

pneu

mon

ia,

orw

ithdr

awof

cons

ent

Tan

tipon

get

al(2

008)

14T

ertia

ryca

re,

ICU

s,or

gene

ral

med

ical

war

ds;

Tha

iland

Age

�18

y,re

quir

ing

mec

hani

cal

vent

ilatio

nPn

eum

onia

aten

rollm

ent,

alle

rgy

toC

HX

New

,pe

rsis

tent

,or

prog

ress

ive

infi

ltrat

eon

ach

est

radi

ogra

ph,

inco

mbi

natio

nw

ithat

leas

t3

ofth

efo

llow

ing

4cr

iteri

a:bo

dyte

mpe

ratu

re�

38°C

or�

35.5

°C,

leuk

ocyt

osis

(�10

�10

3le

ukoc

ytes

/mm

3)

orle

ukop

enia

(�3

�10

3

leuk

ocyt

es/m

m3),

puru

lent

trac

heal

aspi

rate

,an

d/or

apo

sitiv

ese

miq

uant

itativ

ecu

lture

oftr

ache

alas

pira

tesa

mpl

esfo

rpa

thog

enic

bact

eria

CH

X2%

solu

tion

(n�

102)

;or

alca

reco

nsis

ted

ofto

oth

brus

hing

,su

ctio

ning

ofor

alse

cret

ions

,an

dru

bbin

gth

eor

opha

ryng

eal

muc

osa

with

15m

Lof

CH

X;

inte

rven

tion

was

perf

orm

ed4

times

/dun

tilre

mov

alof

the

endo

trac

heal

tube

Nor

mal

salin

eso

lutio

n(n

�10

5),

appl

ied

acco

rdin

gto

the

sam

eor

alca

repr

otoc

olus

edin

the

expe

rim

enta

lgr

oup

Subj

ects

wer

efo

llow

edun

tilex

tuba

tion

orde

velo

pmen

tof

pneu

mon

ia (con

tinu

ed)



Tab

le1.

Con

tinue

d

Stud

y(Y

ear)

Subj

ect

Sour

ceIn

clus

ion

Cri

teri

aE

xclu

sion

Cri

teri

aD

iagn

ostic

Info

rmat

ion

Exp

erim

enta

lG

roup

Con

trol

Gro

upFo

llow

-Up

Bel

lissi

mo-

Rod

rigu

eset

al(2

009)

15

Clin

ical

and

surg

ical

ICU

;B

razi

lE

xpec

ted

ICU

stay

�48

h,in

clud

edw

ithin

24h

afte

rIC

Uad

mis

sion

,re

gard

less

ofw

heth

erre

ceiv

ing

mec

hani

cal

vent

ilatio

n;tr

ache

otom

ized

subj

ects

wer

ein

clud

ed

CH

Xhy

pers

ensi

tivity

,pr

egna

ncy,

form

alin

dica

tion

for

CH

Xus

e,or

pres

crip

tion

ofan

othe

ror

alto

pica

lm

edic

atio

n

Acc

ordi

ngto

the

crite

ria

defi

ned

byth

eC

DC

and

NN

ISsy

stem

CH

X0.

12%

solu

tion

(tot

aln

�98

,on

mec

hani

cal

vent

ilatio

nn

�64

);af

ter

mec

hani

cal

clea

ning

ofth

em

outh

bynu

rses

,15

mL

was

appl

ied

over

all

surf

aces

ofth

eor

alca

vity

;in

terv

entio

nw

aspe

rfor

med

3tim

es/d

until

ICU

disc

harg

e

Plac

ebo

solu

tion

(tot

aln

�96

,on

mec

hani

cal

vent

ilatio

nn

�69

),ap

plie

dac

cord

ing

toth

esa

me

oral

care

prot

ocol

used

inth

eex

peri

men

tal

grou

p

Ent

ire

ICU

stay

Scan

napi

eco

etal

(200

9)17

Tra

uma

ICU

;U

nite

dSt

ates

Subj

ects

expe

cted

tobe

intu

bate

dan

dm

echa

nica

llyve

ntila

ted

with

in48

hof

ICU

adm

issi

on

Witn

esse

das

pira

tion,

conf

irm

eddi

agno

sis

ofpo

st-o

bstr

uctiv

epn

eum

onia

,hy

pers

ensi

tivity

toC

HX

,th

rom

bocy

tope

nia,

a�d

ono

tin

tuba

te�

orde

r,ag

e�

18y,

preg

nanc

y,le

gal

inca

rcer

atio

n,tr

ansf

erfr

oman

othe

rIC

U,

oral

muc

ositi

s,im

mun

osup

pres

sion

,re

adm

issi

onto

the

ICU

CPI

S�

6as

soci

ated

with

the

pres

ence

of�

104

CFU

/mL

ofa

targ

etpu

tativ

ere

spir

ator

ypa

thog

enin

bron

choa

lveo

lar

lava

gefl

uid

ora

posi

tive

pleu

ral

flui

dcu

lture

inth

eab

senc

eof

prev

ious

pleu

ral

inst

rum

enta

tion

CH

X0.

12%

alco

holic

solu

tion

(n�

58);

1ou

nce

appl

ied

usin

gan

oral

foam

appl

icat

orov

eral

lte

eth

and

intr

a-or

also

fttis

sues

and

suct

ione

daf

ter

1m

in(2

times

/d)

�to

oth

brus

hing

with

asu

ctio

nto

othb

rush

(2tim

es/d

)�

swab

bing

with

1.5%

hydr

ogen

pero

xide

solu

tion

(6tim

es/d

)�

appl

icat

ion

ofa

mou

thm

oist

uriz

er;

CH

X0.

12%

(n�

58);

oral

care

prot

ocol

iden

tical

toth

eon

ede

scri

bed

abov

e,ex

cept

that

CH

Xw

asap

plie

don

ly1

time/

d

Plac

ebo

(n�

59);

1ou

nce

appl

ied

usin

gan

oral

foam

appl

icat

orov

eral

lte

eth

and

intr

a-or

also

fttis

sues

and

suct

ione

daf

ter

1m

in(2

times

/d)

�to

oth

brus

hing

with

asu

ctio

nto

othb

rush

(2tim

es/d

)�

swab

bing

with

Pero

xam

int

solu

tion

(6tim

es/d

)�

appl

icat

ion

ofa

mou

thm

oist

uriz

erto

the

oral

muc

osa

Subj

ects

wer

efo

llow

edfo

rup

to21

dor

until

disc

harg

efr

omIC

U,

extu

batio

n,or

deat

h

Cab

ovet

al(2

010)

18Su

rgic

alIC

U;

Cro

atia

Age

�18

y,m

edic

alco

nditi

onsu

gges

ting

ICU

stay

�3

d,ev

entu

alre

quir

emen

tfo

rm

echa

nica

lve

ntila

tion

byor

otra

chea

lor

naso

trac

heal

intu

batio

n

Ede

ntul

ous

subj

ects

Tem

pera

ture

�38

°Cor

�36

°C,

infi

ltrat

eson

ches

tra

diog

raph

s,le

ukoc

ytos

is(�

10�

103/m

m3)

orle

ukop

enia

(�3

�10

3/m

m3),

posi

tive

cultu

refr

omtr

ache

alas

pira

tean

d/or

bron

choa

lveo

lar

lava

ge

CH

X0.

2%ge

l(t

otal

n�

30,

with

out

adi

agno

sis

ofpn

eum

onia

atba

selin

en

�17

);ap

plie

dov

erde

ntal

and

ging

ival

surf

aces

bynu

rses

wea

ring

ster

ilegl

oves

,af

ter

mou

thri

nsin

gw

ithbi

carb

onat

eis

oton

icse

rum

follo

wed

byor

opha

ryng

eal

aspi

ratio

n;ge

lw

asle

ftin

plac

e,an

dth

eor

alca

vity

was

not

rins

edaf

ter

appl

icat

ion;

CH

Xap

plic

atio

nst

arte

dea

rly

afte

rin

tuba

tion;

inte

rven

tion

was

perf

orm

ed3

times

/ddu

ring

ICU

stay

Plac

ebo

gel

(tot

aln

�30

,w

ithou

ta

diag

nosi

sof

pneu

mon

iaat

base

line

�23

),ap

plie

dac

cord

ing

toth

esa

me

oral

care

prot

ocol

used

inth

eex

peri

men

tal

grou

p

Unt

ilIC

Udi

scha

rge

orde

ath

(con

tinu

ed)



Tab

le1.

Con

tinue

d

Stud

y(Y

ear)

Subj

ect

Sour

ceIn

clus

ion

Cri

teri

aE

xclu

sion

Cri

teri

aD

iagn

ostic

Info

rmat

ion

Exp

erim

enta

lG

roup

Con

trol

Gro

upFo

llow

-Up

Ber

ryet

al(2

011)

19Su

rgic

al-m

edic

alIC

U;

Uni

ted

Stat

es

Age

�15

year

s,in

tuba

ted

and

able

tobe

rand

omiz

edw

ithin

12h

ofin

tuba

tion

Subj

ects

who

requ

ired

spec

ific

oral

hygi

ene

proc

edur

esin

rela

tion

tofa

cio-

max

illar

yor

dent

altr

aum

a/su

rger

y;ha

dbe

enin

the

ICU

prev

ious

lydu

ring

the

curr

ent

peri

odof

hosp

italiz

atio

n;re

ceiv

edir

radi

atio

nor

chem

othe

rapy

onad

mis

sion

toth

eIC

Uor

inth

epr

eced

ing

6w

ks;

orsu

ffer

edfr

omau

toim

mun

edi

seas

es

New

orw

orse

ning

radi

olog

ical

infi

ltrat

es,

toge

ther

with

�2

ofth

efo

llow

ing:

tem

pera

ture

�37

.5°C

or�

35°C

,w

hite

cell

coun

t�

11,0

00/m

m3

or�

4,00

0/m

m3,

chan

gein

char

acte

rist

ics

ofbr

onch

ial

secr

etio

nsfr

omm

ucoi

dto

muc

o-pu

rule

ntor

puru

lent

,in

crea

sein

frac

tion

ofin

spir

edox

ygen

orpo

sitiv

een

d-ex

pira

tory

pres

sure

requ

irem

ent

by�

20%

tom

aint

ain

oxyg

ensa

tura

tion

abov

e92

%

CH

X0.

2%aq

ueou

sso

lutio

n(n

�71

);ir

riga

tion

with

CH

X(2

times

/d),

with

seco

ndho

urly

irri

gatio

nw

ithst

erile

wat

eran

dco

mpr

ehen

sive

clea

ning

ofth

em

outh

usin

ga

soft

,pe

diat

ric

toot

hbru

sh(3

times

/d)

Con

trol

I(n

�78

):st

erile

wat

erri

nsed

seco

ndho

urly

and

com

preh

ensi

vecl

eani

ngof

the

mou

thus

ing

aso

ft,

pedi

atri

cto

othb

rush

3tim

es/d

;C

ontr

olII

(n�

76):

sodi

umbi

carb

onat

em

outh

was

hri

nsed

seco

ndho

urly

and

com

preh

ensi

vecl

eani

ngof

the

mou

thus

ing

aso

ft,

pedi

atri

cto

othb

rush

3tim

es/d

Prot

ocol

was

follo

wed

until

the

patie

ntw

asex

tuba

ted

orup

onIC

Udi

scha

rge,

trac

heot

omy,

orde

ath

Jaco

mo

etal

(201

1)21

Ter

tiary

care

PIC

U;

Bra

zil

Chi

ldre

nw

ithco

ngen

ital

hear

tdi

seas

eun

derg

oing

card

iac

surg

ery

with

orw

ithou

tca

rdio

pulm

onar

yby

pass

adm

itted

toth

ePI

CU

inth

epo

stop

erat

ive

peri

od

Subj

ects

with

apr

eope

rativ

edi

agno

sis

ofpn

eum

onia

,hy

pers

ensi

tivity

toC

HX

,co

ngen

ital

orac

quir

edim

mun

odef

icie

ncy,

intr

aope

rativ

ede

ath,

failu

reto

perf

orm

oral

hygi

ene

peri

oper

ativ

ely

Acc

ordi

ngto

the

crite

ria

defi

ned

byth

eC

DC

and

NN

ISsy

stem

CH

X0.

12%

alco

holic

solu

tion

(n�

89);

0.3

ml

ofso

lutio

n/kg

ofbo

dyw

eigh

tw

asus

edpr

eope

rativ

ely

(bef

ore

intu

batio

n),

for

30s,

asan

oral

rins

ein

child

ren

�6

yol

d;in

child

ren

�6

yol

dan

ddu

ring

orot

rach

eal

intu

batio

n,th

eso

lutio

nw

asap

plie

dto

the

oral

muc

osa,

ging

iva,

tong

ue,

and

toot

hsu

rfac

esov

era

peri

odof

30s

with

asp

atul

aw

rapp

edw

ithga

uze;

inte

rven

tion

was

perf

orm

ed2

times

/dpo

stop

erat

ivel

yun

tilPI

CU

disc

harg

eor

deat

h

Plac

ebo

solu

tion

(n�

75),

appl

ied

acco

rdin

gto

the

sam

eor

alca

repr

otoc

olus

edin

the

expe

rim

enta

lgr

oup

Subj

ects

wer

efo

llow

edun

tilPI

CU

disc

harg

eor

deat

h

Kus

ahar

aet

al(2

012)

22T

ertia

ryca

rePI

CU

;B

razi

lC

hild

ren

likel

yto

requ

ire

intu

batio

nan

dm

echa

nica

lve

ntila

tion

with

in24

hof

PIC

Uad

mis

sion

New

born

stat

us,

diag

nosi

sof

pneu

mon

iaat

adm

issi

on,

hype

rsen

sitiv

ityto

CH

X,

dura

tion

ofm

echa

nica

lve

ntila

tion

�48

h,ch

ildre

nw

ithtr

ache

osto

my

orw

hore

ceiv

edtr

ache

alin

tuba

tion

for

�24

hbe

fore

PIC

Uad

mis

sion

The

deve

lopm

ent

ofV

AP

was

quan

tifie

dus

ing

CPI

San

dco

nfir

med

byth

eal

tern

ate

pneu

mon

iacl

inic

alcr

iteri

afo

rin

fant

san

dch

ildre

n,as

defi

ned

byth

eC

DC

/Nat

iona

lH

ealth

care

Safe

tyN

etw

ork

CH

X0.

12%

gel

(n�

46);

gel

was

appl

ied

toa

toot

hbru

sh,

and

all

toot

hsu

rfac

esan

dve

ntra

lsu

rfac

eof

the

tong

uew

ere

clea

ned

and

aspi

rate

dw

itha

vacu

um;

next

,th

ege

lw

asap

plie

dby

asw

abov

eral

lgi

ngiv

alsu

rfac

es;

inte

rven

tion

was

perf

orm

ed2

times

/d,

for

upto

15d

orun

tilde

ath

orex

tuba

tion

Plac

ebo

gel

(n�

50),

appl

ied

acco

rdin

gto

the

sam

eor

alca

repr

otoc

olus

edin

the

expe

rim

enta

lgr

oup

Subj

ects

wer

efo

llow

edun

tilth

eyex

ited

the

stud

yor

wer

edi

scha

rged

from

the

hosp

ital

Ozç

aka

etal

(201

2)24

Res

pira

tory

ICU

;T

urke

yD

enta

tesu

bjec

tsex

pect

edto

bein

tuba

ted

and

mec

hani

cally

vent

ilate

dfo

r�

48h

afte

rIC

Uad

mis

sion

Witn

esse

das

pira

tion,

conf

irm

eddi

agno

sis

ofpo

st-o

bstr

uctiv

epn

eum

onia

,hy

pers

ensi

tivity

toC

HX

,th

rom

bocy

tope

nia,

a“d

ono

tin

tuba

te”

orde

r,ag

e�

18y,

preg

nanc

y,or

alm

ucos

itis,

read

mis

sion

toth

esa

me

ICU

,su

rviv

alex

pect

atio

n�

1w

kan

ded

entu

lism

The

pres

ence

of�

104

CFU

/m

Lof

ata

rget

puta

tive

resp

irat

ory

path

ogen

inm

ini-

bron

choa

lveo

lar

lava

gefl

uid

ora

posi

tive

pleu

ral

flui

dcu

lture

inth

eab

senc

eof

prev

ious

pleu

ral

inst

rum

enta

tion

CH

X0.

2%so

lutio

n(n

�32

);30

mL

appl

ied

byor

alsw

abto

all

teet

han

din

tra-

oral

soft

tissu

esan

dsu

ctio

ned

afte

r1

min

;in

terv

entio

nw

aspe

rfor

med

4tim

es/d

,fo

rup

to14

dor

until

disc

harg

efr

omIC

U,

extu

batio

n,or

deat

h

Salin

eso

lutio

n(n

�34

),ap

plie

dac

cord

ing

toth

esa

me

oral

care

prot

ocol

used

inth

eex

peri

men

tal

grou

p

Subj

ects

wer

efo

llow

edfo

rup

to14

dor

until

ICU

disc

harg

e,ex

tuba

tion,

orde

ath

(con

tinu

ed)



were children.21,22,25 All of the remaining studies includedsubjects age �15 y19 or �18 y8,9,10,12,14,18 (see Table 1).

Trials were set in various ICUs and emergency services.Most studies included subjects from clinical surgicalICUs8,9,15,18,19 and pediatric ICUs.21,22,25 Two studies in-cluded subjects admitted to trauma ICUs.10,17 Moreover,single trials were carried out in neuroscience ICU,10 mixedICU,8 emergency department,10 and general medicalwards14 (Table 1). Average sample sizes, inclusion andexclusion criteria, and VAP diagnostic methods and crite-ria varied considerably among studies (Table 1).

Oral Care With Chlorhexidine

Included studies were also quite heterogeneous in theirintervention regimens. Among them, chlorhexidine wasused at concentrations of 0.12%,10,15,17,21,22 0.2%,8,9,18,19,24

1%,25 and 2%12,14 (Table 1). Chlorhexidine was applied asoral rinse solution,10,14,15,19,21,24 gel,8,9,18,22,25 Vaseline pe-troleum jelly,12 and foam17 (Table 1). When specified bythe authors, chlorhexidine solutions were reported as aque-ous19 or alcoholic.17,21

The frequency of chlorhexidine oral application alsovaried among the studies. Chlorhexidine was used in asingle dose at intubation,10 once/d,17 twice/d,17,19,21,22 3times/d,8,9,15,18,25 or 4 times/d12,14,24 (Table 1).

Methodological Quality of the Studies

Studies’ individual risk of bias were assessed and listedin Table 2. Details related to the method of randomizationwere provided in all studies.8-10,12,14,15,17-19,21,22,24,25 Allo-cation concealment was adequately described only in 4studies.8,15,17,21 Moreover, one study25 reported that allo-cation was concealed but did not provide details of theconcealment. The remaining 8 studies did not provide anyinformation about allocation concealment.9,10,12,14,18,19,22,24

Whereas study subjects and personnel were blinded inonly 8 trials,8,12,15,17,18,21,22,25 outcome assessors wereblinded in all studies. 8-10,12,14,15,17-19,21,22,24,25

Incomplete outcome data were adequately addressed in5 studies.12,15,21,24,25 In 3 studies,9,17,18 the reasons for miss-ing data in each group were not provided. In 2 others, thedropout rate was significant higher in the chlorhexidinegroup.10,19 In Fourrier et al,9 the proportion of missingoutcomes compared with observed event risk was highenough to induce relevant bias. Finally, the reasons formissing outcomes were likely to be related to the trueoutcome in Kusahara et al.22

Sample size calculation was not described in 4 stud-ies.9,10,18,22 Moreover, in the other 4, final sample size wassmaller than the number indicated by sample size calcu-lations.8,14,19,25T

able

1.C

ontin

ued

Stud

y(Y

ear)

Subj

ect

Sour

ceIn

clus

ion

Cri

teri

aE

xclu

sion

Cri

teri

aD

iagn

ostic

Info

rmat

ion

Exp

erim

enta

lG

roup

Con

trol

Gro

upFo

llow

-Up

Seba

stia

net

al(2

012)

25PI

CU

;In

dia

Age

�3

mo

and

�15

y,re

quir

ing

orot

rach

eal

orna

sotr

ache

alin

tuba

tion

and

mec

hani

cal

vent

ilatio

n

Subj

ects

who

had

been

mec

hani

cally

vent

ilate

dfo

r�

24h

befo

rePI

CU

adm

issi

on,

with

trac

heos

tom

ies,

with

inac

cess

ible

oral

cavi

ties,

hype

rsen

sitiv

ityto

CH

X

Insu

bjec

tsw

ithno

evid

ence

ofpr

eexi

stin

gpn

eum

onia

,di

agno

sis

ofV

AP

was

mad

eba

sed

onth

ecr

iteri

aes

tabl

ishe

dby

the

CD

C;

insu

bjec

tsw

ithun

derl

ying

pneu

mon

ia,

wor

seni

ngof

clin

ical

and

radi

olog

ical

find

ings

and

chan

ges

inbr

onch

oalv

eola

rla

vage

flui

dfl

ora

wer

eus

edto

diag

nose

VA

P

CH

X1%

gel

(tot

aln

�41

,w

ithou

ta

diag

nosi

sof

pneu

mon

iaat

base

line

n�

15);

0.5

g(1

.5cm

)of

gel

was

appl

ied

over

the

bucc

alm

ucos

a,af

ter

mou

thas

pira

tion

and

clea

nsin

gw

ithsa

line-

soak

edga

uze;

inte

rven

tion

was

perf

orm

ed3

times

/d,

for

upto

21d

Plac

ebo

gel

(tot

aln

�45

,w

ithou

ta

diag

nosi

sof

pneu

mon

iaat

base

line

n�

16),

appl

ied

acco

rdin

gto

the

sam

eor

alca

repr

otoc

olus

edin

the

expe

rim

enta

lgr

oup

Subj

ects

wer

efo

llow

edfo

ra

peri

odof

21d

orun

tilho

spita

ldi

scha

rge,

whi

chev

erw

asea

rlie

r

CPI

S�

clin

ical

pulm

onar

yin

fect

ion

scor

eC

HX

�ch

lorh

exid

ine

CFU

�co

lony

-for

min

gun

itsV

AP

�ve

ntila

tor-

asso

ciat

edpn

eum

onia

CD

C�

Cen

ters

for

Dis

ease

Con

trol

and

Prev

entio

nN

NIS

�N

atio

nal

Nos

ocom

ial

Infe

ctio

nsSu

rvei

llanc

ePI

CU

�pe

diat

ric

ICU



Tab

le2.

Ris

kof

Bia

s

Stud

y(Y

ear)

Ran

dom

Sequ

ence

Gen

erat

ion

Allo

catio

nC

once

alm

ent

Blin

ding

ofPa

rtic

ipan

tsan

dPe

rson

nel

Blin

ding

ofO

utco

me

Ass

esso

rIn

com

plet

eO

utco

me

Dat

aSe

lect

ive

Rep

ortin

gO

ther

Sour

ces

ofB

ias

Ove

rall

Ris

kof

Bia

s

Four

rier

etal

(200

0)9

Low

(��su

bjec

tsw

ere

rand

omiz

edin

totw

ogr

oups

acco

rdin

gto

aco

mpu

ter-

sequ

ence

bala

nced

rand

omiz

atio

nta

ble�

)

Unc

lear

Hig

hL

owH

igh

(stu

dydo

esno

tsh

owth

ere

ason

sfo

rat

triti

onin

each

grou

p)

Low

Hig

h(n

osa

mpl

esi

zeca

lcul

atio

n)H

igh

Gra

pet

al(2

004)

10L

ow(“

Subj

ects

wer

era

ndom

ized

eith

erto

one

ofth

etr

eatm

ent

grou

ps(C

HG

bysp

ray

orsw

ab)

orto

the

cont

rol

grou

p(u

sual

care

)us

ing

abl

ock

rand

omiz

atio

nsc

hem

e”)

Unc

lear

Hig

hL

owH

igh

(attr

ition

was

high

erat

the

CH

Xgr

oups

,2

CH

Xgr

oups

wer

em

erge

ddu

eto

high

attr

ition

)

Low

Hig

h(n

osa

mpl

esi

zeca

lcul

atio

n;pr

oced

ures

not

clea

rly

desc

ribe

din

test

and

cont

rol

grou

ps)

Hig

h

Four

rier

etal

(200

5)8

Low

(“B

lock

rand

omiz

atio

nst

ratif

ied

bysi

tew

asus

ed�)

Low

(“�al

lra

ndom

izat

ion

lists

wer

ehe

ldin

seal

eden

velo

pes

inth

eph

arm

acy

depa

rtm

ents

ofth

esi

xce

nter

s”)

Low

Low

Hig

h(t

hepr

opor

tion

ofm

issi

ngou

tcom

esco

mpa

red

with

obse

rved

even

tri

sken

ough

toin

duce

clin

ical

lyre

leva

ntbi

asin

inte

rven

tion

effe

ctes

timat

e)

Low

Hig

h(f

inal

sam

ple

size

smal

ler

than

the

num

ber

defi

ned

byth

ere

sults

from

sam

ple

size

calc

ulat

ion)

Hig

h

Koe

man

etal

(200

6)12

Low

(“E

ligib

lesu

bjec

tsw

ere

rand

omly

assi

gned

toon

eof

thre

est

udy

grou

psby

aco

mpu

teri

zed

rand

omiz

atio

nsc

hedu

le.

Ran

dom

izat

ion

was

stra

tifie

dpe

rho

spita

l”)

Unc

lear

Low

Low

Low

(mis

sing

outc

ome

data

bala

nced

innu

mbe

rsac

ross

inte

rven

tion

grou

ps,

with

sim

ilar

reas

ons

for

mis

sing

data

acro

ssgr

oups

)

Low

Low

Unc

lear

Tan

tipon

get

al(2

008)

14L

ow(“

Eac

hel

igib

lesu

bjec

tw

asra

ndom

ized

toth

ech

lorh

exid

ine

grou

por

the

norm

alsa

line

grou

pby

stra

tifie

dra

ndom

izat

ion

acco

rdin

gto

the

sex

and

hosp

ital

loca

tion�

”)

Unc

lear

Hig

hL

owU

ncle

ar(i

nsuf

fici

ent

repo

rtin

gof

attr

ition

/exc

lusi

ons

tope

rmit

judg

men

t;it

isno

tcl

ear

ifal

lsu

bjec

tsco

mpl

eted

the

stud

y)

Low

Hig

h(s

ampl

esi

zew

assm

alle

rth

anth

enu

mbe

rde

fine

dby

the

sam

ple

size

calc

ulat

ion)

Hig

h

Bel

issi

mo-

Rod

rigu

eset

al(2

009)

15L

ow(�

�a

code

num

ber

was

sele

cted

from

abo

xan

dth

eco

rres

pond

ing

num

bere

dbo

ttle

was

plac

edbe

side

the

subj

ect’

sbe

d.R

ando

miz

atio

nw

asno

tst

ratif

ied”

)

Low

(“U

ntil

all

data

wer

eco

llect

ed;

only

the

phar

mac

ist

knew

whi

chco

denu

mbe

rsco

rres

pond

edto

whi

chki

ndof

solu

tion”

)

Low

Low

Low

(mis

sing

outc

ome

data

bala

nced

innu

mbe

rsac

ross

inte

rven

tion

grou

ps,

with

sim

ilar

reas

ons

for

mis

sing

data

acro

ssgr

oups

)

Low

Unc

lear

(dif

fere

nces

inag

ean

dfr

eque

ncy

ofco

mpr

omis

edm

enta

lst

atus

betw

een

grou

ps)

Unc

lear

Scan

napi

eco

etal

(200

9)17

Low

(“Su

bjec

tsw

ere

rand

omiz

edto

the

stud

yvi

aa

web

-bas

edsu

bjec

ten

rollm

ent

syst

emth

atpr

epar

eda

set

ofSu

bjec

tId

entif

icat

ion

Num

bers

(SID

)th

atid

entif

ied

indi

vidu

altr

eatm

ent

assi

gnm

ents

”)

Low

(“Se

aled

enve

lope

sco

ntai

ning

ara

ndom

num

ber

wer

ege

nera

ted

inbl

ocks

ofsi

xto

prov

ide

conc

ealm

ent

ofsu

bjec

tas

sign

men

tfr

omth

ein

vest

igat

ors“

)

Low

Low

Unc

lear

(rea

sons

for

mis

sing

data

inea

chgr

oup

wer

eno

tpr

ovid

ed)

Low

Low

Unc

lear

Cab

ovet

al(2

010)

18L

ow(“

�Su

bjec

tsw

ere

rand

omiz

edin

totw

ogr

oups

acco

rdin

gto

aco

mpu

ter-

gene

rate

dba

lanc

edra

ndom

izat

ion

tabl

e�)

Unc

lear

Low

Low

Hig

h(r

easo

nsfo

rm

issi

ngda

tain

each

grou

pw

ere

not

prov

ided

)

Low

Hig

h(n

osa

mpl

esi

zeca

lcul

atio

n,V

AP

was

not

clea

rly

defi

ned,

excl

usio

ncr

iteri

aw

ere

not

clea

rly

defi

ned)

Hig

h

(con

tinu

ed)



Tab

le2.

Con

tinue

d

Stud

y(Y

ear)

Ran

dom

Sequ

ence

Gen

erat

ion

Allo

catio

nC

once

alm

ent

Blin

ding

ofPa

rtic

ipan

tsan

dPe

rson

nel

Blin

ding

ofO

utco

me

Ass

esso

rIn

com

plet

eO

utco

me

Dat

aSe

lect

ive

Rep

ortin

gO

ther

Sour

ces

ofB

ias

Ove

rall

Ris

kof

Bia

s

Ber

ryet

al(2

011)

19L

ow(�

�Su

bjec

tsw

ere

rand

omiz

edin

toon

eof

thre

egr

oups

acco

rdin

gto

aba

lanc

edra

ndom

izat

ion

tabl

epr

epar

edby

abi

osta

tistic

ian�

)

Unc

lear

Hig

hL

owH

igh

(per

cent

age

ofm

issi

ngou

tcom

eda

taw

assm

alle

rin

num

ber

inth

epl

aceb

ogr

oup;

freq

uenc

yof

prot

ocol

brea

chw

ashi

gher

inth

etr

eatm

ent

grou

ps,

freq

uenc

yof

subj

ect

deat

hw

ashi

gher

inth

eso

dium

bica

rbon

ate

grou

p)

Low

Hig

h(n

umbe

rof

subj

ects

was

smal

ler

than

the

one

defi

ned

byth

esa

mpl

esi

zeca

lcul

atio

n)

Hig

h

Jaco

mo

etal

(201

1)21

Low

(��su

bjec

tsw

ere

rand

omiz

edto

the

expe

rim

enta

lor

the

cont

rol

grou

pby

mea

nsof

alis

tge

nera

ted

bya

com

pute

rize

dsy

stem

that

uses

ara

ndom

num

ber

gene

rato

rto

prod

uce

cust

omiz

edse

tsof

rand

omnu

mbe

rs�)

Low

(“T

hera

ndom

izat

ion

list

was

held

inth

eho

spita

lph

arm

acy,

and

all

inve

stig

ator

sw

ere

unaw

are

ofsu

bjec

ts’

assi

gnm

ents

”)

Low

Low

Low

(per

cent

age

ofm

issi

ngou

tcom

eda

taw

asba

lanc

edin

num

ber

acro

ssin

terv

entio

ngr

oups

,w

ithsi

mila

rre

ason

sfo

rm

issi

ngda

taac

ross

grou

ps)

Low

Low

Low

Kus

ahar

aet

al(2

012)

22L

ow(“

Chi

ldre

nw

ere

sequ

entia

llyra

ndom

ized

into

two

grou

psus

ing

aba

lanc

edra

ndom

izat

ion

tabl

ege

nera

ted

byth

eT

rue

Epi

stat

prog

ram

”)

Unc

lear

Low

Low

Hig

h(r

easo

nfo

rth

em

issi

ngou

tcom

eda

talik

ely

tobe

rela

ted

toth

etr

ueou

tcom

e)

Low

Hig

h(a

gedi

ffer

ence

,no

sam

ple

size

calc

ulat

ion)

Hig

h

Ozç

aka

etal

(201

2)24

Low

(�T

hera

ndom

izat

ion

prep

ared

ase

tof

subj

ect

iden

tific

atio

nnu

mbe

rs(S

IDs)

that

iden

tifie

din

divi

dual

trea

tmen

tas

sign

men

ts�)

Unc

lear

Hig

hL

owL

ow(p

erce

ntag

eof

mis

sing

outc

ome

data

was

bala

nced

innu

mbe

rac

ross

inte

rven

tion

grou

ps,

with

sim

ilar

reas

ons

for

mis

sing

data

acro

ssgr

oups

)

Low

Low

Unc

lear

Seba

stia

net

al(2

012)

25L

ow(�

The

rand

omse

quen

cew

asge

nera

ted

for

each

stra

tum

usin

gth

eST

AT

A9.

0pr

ogra

m�in

bloc

ksof

6�)

Low

(�R

ando

miz

atio

nan

dnu

mbe

ring

ofth

etu

bes

wer

edo

neby

pers

onne

lno

tin

volv

edin

the

stud

y,an

dth

eal

loca

tion

sequ

ence

rem

aine

dco

ncea

led

thro

ugh

the

entir

ele

ngth

ofth

est

udy�

)

Low

Low

Low

(per

cent

age

ofm

issi

ngou

tcom

eda

taw

asba

lanc

edin

num

ber

acro

ssin

terv

entio

ngr

oups

,w

ithsi

mila

rre

ason

sfo

rm

issi

ngda

taac

ross

grou

ps)

Low

Hig

h(n

umbe

rof

subj

ects

rand

omiz

edw

assm

alle

rth

anth

enu

mbe

rsu

gges

ted

byth

esa

mpl

esi

zeca

lcul

atio

n)

Hig

h

CH

X�

chlo

rhex

idin

eV

AP

�ve

ntila

tor-

asso

ciat

edpn

eum

onia



Effect of Oral Care With Chlorhexidineon VAP Prevention

A preliminary analysis including 1,640 pediatric andadult subjects revealed that oral application of chlorhexi-dine did not promote a significant reduction in VAP inci-dence (relative risk 0.80, 95% CI 0.59–1.07, I2 � 45%)(Fig. 2). Next, subgroup analyses were conducted to com-pare the effect of chlorhexidine in pediatric and adult pop-ulations. Similar to the results found in the overall studypopulation, oral care with chlorhexidine failed to preventVAP in the pediatric population (relative risk 1.13, 95% CI0.76–1.67, I2 � 0%) (see Fig. 2). Nonetheless, oral appli-cation of chlorhexidine promoted a trend toward a protec-tive effect in adult subjects (relative risk 0.70, 95% CI0.48–1.00, I2 � 47%) (see Fig. 2). Due to the limitednumber of studies that investigated the effect of oral carewith chlorhexidine on VAP prevention in pediatric sub-jects and the lack of effects of oral care with chlorhexidinein this study population, the following subgroup analyseswere conducted based on adult population data only.

Effect of Chlorhexidine Concentration

Subgroup analysis investigated chlorhexidine used inconcentrations of 0.12, 0.2, and 2% (Fig. 3). At the lowest

concentrations tested (0.12 and 0.2%), chlorhexidine failedto prevent VAP development (0.12% chlorhexidine: rela-tive risk 1.00, 95% CI 0.51–1.99, I2 � 54%; 0.2% chlo-rhexidine: relative risk 0.63, 95% CI 0.32–1.22, I2 � 57%).In sharp contrast, 2% chlorhexidine promoted a significantreduction in VAP incidence (relative risk 0.53, 95% CI0.31–0.91, I2 � 0%).

Effect of Chlorhexidine Frequency of Use

Subgroup analyses investigated chlorhexidine used ina single application at intubation and once, twice, 3times, or 4 times daily (Fig. 4). When used as a singleapplication dose at intubation, in the study published byGrap et al,10 chlorhexidine failed to reduce the inci-dence of VAP (relative risk 2.79, 95% CI 0.75–10.37).Likewise, chlorhexidine used at frequencies of once/d,twice/d, and 3 times/d also failed to prevent VAP de-velopment (once/d: relative risk 0.59, 95% CI 0.25–1.40; twice/d: relative risk 1.25, 95% CI 0.19 – 8.31,I2 � 65%; 3 times/d: relative risk 0.64, 95% CI 0.31–1.31, I2 � 62%). The protective effect of chlorhexidinewas only achieved when its frequency of use was in-creased to 4 times/d (relative risk 0.56, 95% CI 0.38 –0.81, I2 � 0%).

Fig. 2. Effect of oral care with chlorhexidine on ventilator-associated pneumonia prevention.



Effect of Chlorhexidine Used as Monotherapy or inCombination With Mechanical Means

In some studies, chlorhexidine was the only form of oralcare.8-10,12,18,24 On the contrary, in some others, chlorhexi-dine was associated with mechanical debridement.14,15,17,19

Therefore, an analysis was undertaken to assess the effec-tiveness of chlorhexidine used alone and in associationwith mechanical means for the prevention of VAP (Fig. 5).Used as monotherapy, chlorhexidine failed to reduce VAPincidence (relative risk 0.65, 95% CI 0.39–1.09, I2 �55%). Similarly, chlorhexidine did not promote a signifi-cant reduction in VAP incidence when used in conjunctionwith mechanical cleansing of the oral cavity (relative risk0.77, 95% CI 0.43–1.39, I2 � 42%).

Safety

Six studies reported that the oral use of chlorhexidinewas associated with no adverse effects.15,17,19,21,24,25 An-other 6 studies failed to provide information about its safe-ty.8-10,12,18,22 One study reported that mild and reversibleirritation of the oral mucosa was more common in subjectstreated with chlorhexidine 2% solution than in those treatedwith normal saline.14

Discussion

With an intention-to-treat analysis, the present meta-analysis provides the most comprehensive assessment todate of the effect of different protocols of oral care withchlorhexidine in VAP prevention in a non-cardiac surgerypopulation. According to our results, the effectiveness oforal care with chlorhexidine in VAP prevention is influ-enced by the age of the population and the concentrationand frequency of application of chlorhexidine.

The current study demonstrated that oral care with chlo-rhexidine promoted a trend toward VAP prevention inadult subjects but failed to prevent disease development innewborns and infants. There are 3 possible explanationsfor this discrepancy. First, it is plausible that the antimi-crobial effects of chlorhexidine cannot overcome the rel-ative immaturity of the immune system of newborns andinfants. The relevance of newborn respiratory innate im-munity to the pathogenesis of respiratory diseases in new-borns and infants is beginning to surface. Plasmatic levelsof complement components and other multifunctional sol-uble immune proteins are significantly lower in newbornscompared with adults.33 Moreover, existing evidence basedon animal models indicates that a post-natal impairment ofTLR2 and TLR4 expression negatively affects inflamma-tory responses following intratracheal administration of

Fig. 3. Effect of chlorhexidine concentration on ventilator-associated pneumonia prevention.



Gram-negative bacteria early in life.34 Likewise, a strongbias against T helper cell type 1 polarization of the im-mune response is also thought to make infants more sus-ceptible to microbial infections.35 Second, it is reasonableto speculate that the small oral cavity associated with therelatively large tongue in newborns and infants is likely topose technical difficulties in providing proper oral carewith chlorhexidine to these subjects. Last, the lack of chlo-rhexidine effect in the pediatric population might be sim-ply explained by the fact that none of the pediatric trialsused 2% formulations or rendered oral decontaminationwith chlorhexidine 4 times/d.

This meta-analysis demonstrated that the effectivenessof oral care with chlorhexidine on prevention of VAP isdose- and frequency-dependent. Subgroup analysis dem-

onstrated that 0.12 and 0.2% chlorhexidine failed to pro-mote a significant reduction in VAP incidence in adultsubjects. In sharp contrast, 2% chlorhexidine promoted asignificant reduction in the incidence of VAP, with a rel-ative risk of 0.53. Two previous meta-analyses showedsimilar results, with a relative risk of 0.53 for chlorhexi-dine 2%.28,36 The antibacterial activity of chlorhexidine isdose-dependent.37,38 Higher and longer lasting antimicro-bial activity has been reported for 2% chlorhexidine ascompared with less concentrated formulations,39 whichcould explain the superior results of oral care with 2%chlorhexidine in VAP prevention. Nonetheless, it is im-portant to note that data about the tolerance of 2% solu-tions were provided by only one study that reported mildand reversible irritation of the oral mucosa with the use of

Fig. 4. Effect of chlorhexidine frequency of use on ventilator-associated pneumonia prevention.



2% chlorhexidine solution.14 Moreover, 2% chlorhexidinesolutions are not available worldwide and are often onlymade for study purposes.

This study showed for the first time that oral care withchlorhexidine is only effective in reducing VAP incidencewhen provided 4 times/d. Numerous authors have demon-strated the immediate antibacterial effect of chlorhexidineand the persistence of its substantivity for up to 12–14 hafter its administration. However, the clinical relevanceof this information has been challenged, since risingevidence suggests that although chlorhexidine can befound in the oral cavity for �12 h, its antimicrobialactivity lasts only 7 h after a mouth rinse.37,38 Thus, it islikely that the effectiveness of oral care with chlorhexi-dine in VAP prevention is dependent on its persistentantimicrobial activity.

Along these lines, 2% chlorhexidine was used in only 2of the total of 10 adult population trials included in thismeta-analysis. On a patient level, this signifies that only33% of subjects receiving oral care with chlorhexidinewere treated with the 2% formulation. Likewise, chlo-rhexidine was administered 4 times/d in only 3 trials in-cluded in this meta-analysis, encompassing only 38% ofsubjects receiving oral care with chlorhexidine. As previ-ously mentioned, the current study showed that oral carewith chlorhexidine promoted only a trend toward VAPprevention in adult subjects. Thus, it is reasonable to spec-ulate that the overall effect of oral care with chlorhexidinein VAP prevention could have been stronger if more trialshad administered chlorhexidine at 2% or rendered treat-ment 4 times/d. Also, the wide variety of combinations of

chlorhexidine concentrations and dose intervals reportedin studies included in this meta-analysis have precludedthe investigation of potential interplays of chlorhexidineconcentration and frequency of use in VAP prevention. Aventilator bundle is a group of interventions related toventilator care that, when implemented together, promotessignificantly better outcomes. The VAP prevention bundleis a widely used ICU protocol that includes elevation ofthe head of the bed, daily sedation vacations and assess-ment of readiness to extubate, peptic ulcer disease prophy-laxis, deep vein thrombosis prophylaxis, and oral decon-tamination with chlorhexidine. Thus, it is also plausiblethat oral decontamination with chlorhexidine failed to pro-mote an overall significant reduction in VAP incidencebecause other bundle prevention measures had been suc-cessfully implemented and limited VAP development.

Subanalyses conducted to specifically assess the effec-tiveness of oral chlorhexidine used alone and in associa-tion with mechanical means in the prevention of VAPshowed that none of these protocols were able to reduceVAP incidence. These results, however, must be inter-preted cautiously due to the large methodological hetero-geneity across the limited number of studies included inthis subanalysis. The observation that oral chlorhexidinealone might be slightly superior due to its association withmechanical means for VAP prevention is likely to be mis-leading. First, no direct comparison was made betweenthese protocols. Second, the meta-analysis that assessedthe efficacy of chlorhexidine associated with mechanicalmeans included fewer subjects receiving 2% chlorhexidineand/or rendered treatment 4 times/d than the meta-analysis

Fig. 5. Effect of chlorhexidine used as monotherapy or in combination with mechanical means on ventilator-associated pneumoniaprevention.



of studies assessing the efficacy of chlorhexidine alone. Ofinterest, a meta-analysis of 4 low-quality trials found nodifference between oral care with chlorhexidine plus toothbrushing and oral care with chlorhexidine alone in termsof VAP prevention.28

Finally, only 2 studies included in this meta-analysisreported the periodontal conditions of enrolled subjects.17,24

Potential associations between periodontal disease and peri-odontal disease-associated micro-organisms and the devel-opment of nosocomial pneumonia have been already pro-posed.40 Thus, it is plausible to speculate that oral carewith chlorhexidine is more likely to prevent VAP devel-opment in subjects with periodontal infection.

Conclusions

We found that oral care with chlorhexidine is effectivein reducing VAP incidence in the adult population only ifchlorhexidine is administered at 2% or 4 times/d. Thesefindings, however, must be interpreted cautiously, due tothe high heterogeneity of the studies and small number oftrials that tested the safety and effectiveness of chlorhexi-dine at 2% or rendered treatment 4 times/d. Further inves-tigation of intervention protocols implementing oral chlo-rhexidine at high concentration and frequency to reduceVAP in subjects with a known periodontal status is re-quired before definitive recommendations can be made.

REFERENCES

1. Tablan OC, Anderson LJ, Besser R, Bridges C, Hajjeh R. Guidelinesfor preventing health-care-associated pneumonia, 2003: recommen-dations of CDC and the Healthcare Infection Control Practices Ad-visory Committee. MMWR Recomm Rep 2004;53(RR-3):1-36.

2. Safdar N, Crnich CJ, Maki DG. The pathogenesis of ventilator-associ-ated pneumonia: its relevance to developing effective strategies for pre-vention. Respir Care 2005;50(6):725-739; discussion 739-741.

3. Vincent JL, Rello J, Marshall J, Silva E, Anzueto A, Martin CD, etal. International study of the prevalence and outcomes of infection inintensive care units. JAMA 2009;302(21):2323-2329.

4. Melsen WG, Rovers MM, Groenwold RH, Bergmans DC, Camus C,Bauer TT, et al. Attributable mortality of ventilator-associated pneu-monia: a meta-analysis of individual patient data from randomisedprevention studies. Lancet Infect Dis 2013;13(8):665-671.

5. Kalanuria AA, Zai W, Mirski M. Ventilator-associated pneumonia inthe ICU. Critical Care 2014;18(2):208.

6. Inglis TJ. New insights into the pathogenesis of ventilator-associatedpneumonia. J Hosp Infect 1995;30(Suppl):409-413.

7. Bergmans D, Bonten M. Healthcare-associated Pneumonia (Chapter22) in Mayhall CG. Hospital Epidemiology and Infection Control.4th edition. Philadelphia, Pennsylvania: Lippincott Williams andWilkins; 2011:311-316.

8. Fourrier F, Dubois D, Pronnier P, Herbecq P, Leroy O, Desmettre T,et al. Effect of gingival and dental plaque antiseptic decontaminationon nosocomial infections acquired in the intensive care unit: a dou-ble-blind placebo-controlled multicenter study. Crit Care Med 2005;33(8):1728-1735.

9. Fourrier F, Cau-Pottier E, Boutigny H, Roussel-Delvallez M, Jour-dain M, Chopin C. Effects of dental plaque antiseptic decontamina-

tion on bacterial colonization and nosocomial infections in criticallyill patients. Intensive Care Med 2000;26(9):1239-1247.

10. Grap MJ, Munro CL, Elswick RK Jr., Sessler CN, Ward KR. Du-ration of action of a single, early oral application of chlorhexidine onoral microbial flora in mechanically ventilated patients: a pilot study.Heart Lung 2004;33(2):83-91.

11. Bopp M, Darby M, Loftin KC, Broscious S. Effects of daily oral carewith 0.12% chlorhexidine gluconate and a standard oral care proto-col on the development of nosocomial pneumonia in intubated pa-tients: a pilot study. J Dent Hyg 2006;80(3):9.

12. Koeman M, van der Ven AJ, Hak E, Joore HC, Kaasjager K, de SmetAG, et al. Oral decontamination with chlorhexidine reduces the in-cidence of ventilator-associated pneumonia. Am J Respir Crit CareMed 2006;173(12):1348-1355.

13. Koeman M, van der Ven AJ, Hak E, Joore JC, Kaasjager HA, deSmet AM, et al. Less ventilator-associated pneumonia after oraldecontamination with chlorhexidine: a randomised trial. Ned Tijd-schr Geneeskd 2008;152(13):752-759.

14. Tantipong H, Morkchareonpong C, Jaiyindee S, Thamlikitkul V. Ran-domized controlled trial and meta-analysis of oral decontamination with2% chlorhexidine solution for the prevention of ventilator-associatedpneumonia. Infect Control Hosp Epidemiol 2008;29(2):131-136.

15. Bellissimo-Rodrigues F, Bellissimo-Rodrigues WT, Viana JM, Teix-eira GC, Nicolini E, Auxiliadora-Martins M, et al. Effectiveness oforal rinse with chlorhexidine in preventing nosocomial respiratorytract infections among intensive care unit patients. Infect ControlHosp Epidemiol 2009;30(10):952-958.

16. Munro CL, Grap MJ, Jones DJ, McClish DK, Sessler CN. Chlo-rhexidine, toothbrushing, and preventing ventilator-associated pneu-monia in critically ill adults. Am J Crit Care 2009;18(5):428-437.

17. Scannapieco FA, Yu J, Raghavendran K, Vacanti A, Owens SI,Wood K, Mylotte JM. A randomized trial of chlorhexidine gluconateon oral bacterial pathogens in mechanically ventilated patients. CritCare 2009;13(4):R117.

18. Cabov T, Macan D, Husedzinovic I, Skrlin-Subic J, Bosnjak D,Sestan-Crnek S. The impact of oral health and 0.2% chlorhexidineoral gel on the prevalence of nosocomial infections in surgical in-tensive-care patients: a randomized placebo-controlled study. WienKlin Wochenschr 2010;122(13-14):397-404.

19. Berry AM, Davidson PM, Masters J, Rolls K, Ollerton R. Effects ofthree approaches to standardized oral hygiene to reduce bacterialcolonization and ventilator associated pneumonia in mechanicallyventilated patients: a randomised control trial. Int J Nurs Stud 2011;48(6):681-688.

20. Grap MJ, Munro CL, Hamilton VA, Elswick RK Jr., Sessler CN,Ward KR. Early, single chlorhexidine application reduces ventilator-associated pneumonia in trauma patients. Heart Lung 2011;40(5):e115-e122.

21. Jacomo AD, Carmona F, Matsuno AK, Manso PH, Carlotti AP.Effect of oral hygiene with 0.12% chlorhexidine gluconate on theincidence of nosocomial pneumonia in children undergoing cardiacsurgery. Infect Control Hosp Epidemiol 2011;32(6):591-596.

22. Kusahara DM, Peterlini MA, Pedreira ML. Oral care with 0.12%chlorhexidine for the prevention of ventilator-associated pneumoniain critically ill children: randomised, controlled and double blindtrial. Int J Nurs Stud 2012;49(11):1354-1363.

23. Meinberg MC, Cheade Mde F, Miranda AL, Fachini MM, Lobo SM.The use of 2% chlorhexidine gel and toothbrushing for oral hygieneof patients receiving mechanical ventilation: effects on ventilator-associated pneumonia. Rev Bras Ter Intensiva 2012;24(4):369-374.

24. Ozcaka O, Basoglu OK, Buduneli N, Tasbakan MS, Bacakoglu F,Kinane DF. Chlorhexidine decreases the risk of ventilator-associatedpneumonia in intensive care unit patients: a randomized clinical trial.J Periodontal Res 2012;47(5):584-592.



25. Sebastian MR, Lodha R, Kapil A, Kabra SK. Oral mucosal decon-tamination with chlorhexidine for the prevention of ventilator-asso-ciated pneumonia in children: a randomized, controlled trial. PediatrCrit Care Med 2012;13(5):e305-e310.

26. Labeau SO, Van de Vyver K, Brusselaers N, Vogelaers D, Blot SI.Prevention of ventilator-associated pneumonia with oral antiseptics: asystematic review and meta-analysis. Lancet Infect Dis 2011;11(11):845-854.

27. Li J, Xie D, Li A, Yue J. Oral topical decontamination for preventingventilator-associated pneumonia: a systematic review and meta-analysisof randomized controlled trials. J Hosp Infect 2013;84(4):283-293.

28. Shi Z, Xie H, Wang P, Zhang Q, Wu Y, Chen E, et al. Oral hygienecare for critically ill patients to prevent ventilator-associated pneu-monia. Cochrane Database Syst Rev 2013;(8):CD008367.

29. Klompas M, Speck K, Howell MD, Greene LR, Berenholtz SM.Reappraisal of routine oral care with chlorhexidine gluconate forpatients receiving mechanical ventilation: systematic review andmeta-analysis. JAMA Intern Med 2014;174(5):751-761.

30. Zhang TT, Tang SS, Fu LJ. The effectiveness of different concen-trations of chlorhexidine for prevention of ventilator-associated pneu-monia: a meta-analysis. J Clin Nurs 2014;23(11-12):1461-1475.

31. Liberati A, Altman DG, Tetzlaff J, Mulrow C, Gøtzsche PC,Ioannidis JP, et al. The PRISMA statement for reporting system-atic reviews and meta-analyses of studies that evaluate healthcareinterventions: explanation and elaboration. BMJ 2009;339:b2700.

32. Clarke M OA, Paulsen E, Higgins JPT, Green S. Assessing riskof bias in included studies. In: Higgins JPT and Green S, editors.Cochrane Handbook for Systematic Reviews of Interventions Ver-

sion 5.1.0. The Cochrane Collaboration; 2011. www.cochrane-handbook.org. Accessed April 13, 2016.

33. Firth MA, Shewen PE, Hodgins DC. Passive and active componentsof neonatal innate immune defenses. Anim Health Res Rev 2005;6(2):143-158.

34. Martin TR, Ruzinski JT, Wilson CB, Skerrett SJ. Effects of endo-toxin in the lungs of neonatal rats: age-dependent impairment of theinflammatory response. J Infect Dis 1995;171(1):134-144.

35. Siegrist CA. Vaccination in the neonatal period and early infancy. IntRev Immunol 2000;19(2-3):195-219.

36. Snyders O, Khondowe O, Bell J. Oral chlorhexidine in the preven-tion of ventilator-associated pneumonia in critically ill adults in theICU: a systematic review. Southern African J Crit Care 2011;27(2):48-56.

37. Addy M, Jenkins S, Newcombe R. The effect of some chlorhexidine-containing mouthrinses on salivary bacterial counts. J Clin Periodon-tol 1991;18(2):90-93.

38. Harper PR, Milsom S, Wade W, Addy M, Moran J, Newcombe RG.An approach to efficacy screening of mouthrinses: studies on a groupof French products (II): inhibition of salivary bacteria and plaque invivo. J Clin Periodontol 1995;22(9):723-727.

39. White RR, Hays GL, Janer LR. Residual antimicrobial activity aftercanal irrigation with chlorhexidine. J Endod 1997;23(4):229-231.

40. Scannapieco FA, Bush RB, Paju S. Associations between periodon-tal disease and risk for nosocomial bacterial pneumonia and chronicobstructive pulmonary disease: a systematic review. Ann Periodontol2003;8(1):54-69.



http://www.cochrane-handbook.org

http://www.cochrane-handbook.org

effectiveness of intraoral chlorhexidine protocols in the...

Documents