Upload
habao
View
213
Download
0
Embed Size (px)
Citation preview
REVIEW
The Epidemiology of Clostridium difficile Infectionin Japan: A Systematic Review
Thomas V. Riley . Tomomi Kimura
Received: December 14, 2017 / Published online: February 13, 2018� The Author(s) 2018. This article is an open access publication
Abstract: To increase understanding of the epi-demiology, risks, consequences and resourceutilization of Clostridium difficile infection (CDI)in Japan, a systematic literature review wasundertaken of relevant publications from Jan-uary 2006 to November 2017. Using the Pre-ferred Reporting Items for Systematic Reviewsand Meta-Analyses (PRISMA) guidelines andmethods, 55 articles met the criteria for fullreview. The majority (58%) of studies were froma single site, with the most recent data from2015. The incidence, reported prevalence andrecurrence rate of CDI in Japan were 0.8–4.71/10,000 patient-days, 0.3–5.5/1000 patients and3.3–27.3%, respectively, and varied according to
setting, population, CDI definition and detec-tion method. Most C. difficile isolates associatedwith CDI in Japan were toxin A?B?, with a lowlevel of C. difficile binary toxin-positive (CDT?)strains (0–6.8% reported across studies). Themost common C. difficile PCR ribotypes associ-ated with infection in Japan were smz/018, 002,052 and 369. Data regarding the impact of CDIon length of hospital stay were limited. Reportedall-cause mortality in patients with CDI rangedfrom 3.4 to 15.1% between 2007 and 2013. Twostudies assessed risk factors for CDI recurrence,identifying malignant disease, intensive careunit hospitalization and use of proton pumpinhibitors as factors increasing the risk of initialand/or recurrent CDI. No study analyzed initialCDI treatment in relation to recurrence. Morecomprehensive surveillance and coordinatedstudies are needed to map trends, understandrisk factors, and recognize the extent and impactof CDI in Japanese patients.Funding: Astellas Pharma, Inc.Plain Language Summary: Plain languagesummary available for this article.
Keywords: Clostridium difficile infection (CDI);Epidemiology; Japan; Outcomes; Ribotype
PLAIN LANGUAGE SUMMARY
Clostridium difficile (C. difficile) is a bacteriumthat often lives without causing harm in
Enhanced content To view enhanced content for thisarticle go to http://www.medengine.com/Redeem/104B4F6021959470.
Electronic supplementary material The onlineversion of this article (https://doi.org/10.1007/s40121-018-0186-1) contains supplementary material, which isavailable to authorized users.
T. V. Riley (&)Murdoch University, Murdoch, Australiae-mail: [email protected]
T. V. RileyEdith Cowan University, Joondalup, Australia
T. V. RileyPathWest Laboratory Medicine, Nedlands, Australia
T. KimuraAstellas Pharma, Inc., Tokyo, Japan
Infect Dis Ther (2018) 7:39–70
https://doi.org/10.1007/s40121-018-0186-1
people’s gut. However, when a person hasantibiotic treatment for another infection, thiscan cause an imbalance in the normal levels ofbacteria in the gut, and the C. difficile can growand replace many of the normal bacteria, caus-ing C. difficile infection (CDI). Symptomsinclude diarrhea, fever and pain. Although CDIis often mild, it can be very serious, particularlyin older people, and, if untreated, can be fatal.This review looked at studies published from2006 to 2017 to investigate patterns of CDIsickness (epidemiology) in Japan. A total of 55studies were useful for our review and showedthat, in general, CDI occurred less commonly inJapan than in Western countries. However,there was wide variation in the tests used todetect infection and the methods used toidentify specific types of C. difficile bacteriaresponsible for the infections. Because of thisvariety, there was a difference in the reliabilityof the results from the different studies, whichmade it difficult to make comparisons betweenstudies. However, there seemed to be consistentresults showing that certain types of C. difficilewere common in Japan. The studies were notable to tell us whether the types of C. difficilevaried over time. More studies that use reliablehigh-quality tests, and greater detailed analysisin Japan to map patterns of CDI over time areneeded. This would help us to understand theimportance of CDI in Japan.
INTRODUCTION
Clostridium difficile is the most common infec-tive cause of nosocomial diarrhea, implicated in20–30% of cases of antibiotic-associated diar-rhea [1, 2]. Appropriate patient care requiresrapid and accurate diagnosis to support optimalmanagement and prevent the spread of infec-tion. Furthermore, knowledge of specific riskfactors for C. difficile infection (CDI) in differentclinical settings is essential.
No national CDI surveillance system hasbeen implemented in Japan, and therefore it ischallenging to grasp the trend in epidemiologyover time using a standardized method. Areview of CDI in Asia published in 2013 foundonly a few molecular-typing studies providing
contemporary epidemiological information [3].According to a questionnaire-based survey of2537 hospitals in Japan in 2013, which hadvalid responses from 321 hospitals, CDI inci-dence varied between centers [4], and there waslittle information on the specific strains causinginfection.
There have been several important changesin CDI diagnosis and treatment in Japan. First,a new diagnostic kit detecting toxin A and Bplus ‘‘common’’ antigen (glutamate dehydro-genase; GDH) became available in April 2011.Second, oral and injectable metronidazole wereindicated for CDI in August 2012 [5] andSeptember 2014 [6], respectively, althoughunlicensed use of oral metronidazole for CDIhad occurred in Japan prior to 2012. Third, in2015, the Japanese Association for InfectiousDiseases and Japanese Society of Chemother-apy released guidelines for the treatment ofenteric infection, in which oral metronidazolewas designated as the first-line treatment forCDI [7]. Vancomycin was recommended forsevere cases and/or second and subsequentrecurrences [7].
Considering these recent changes in thediagnosis and treatment of CDI, there is agreater need to understand and update theepidemiology of CDI, the predominant strainscausing the infection, and the consequences,risks and resource utilization associated withCDI in clinical settings in Japan. This literaturereview was undertaken to summarize publishedepidemiological data on CDI in Japan fromJanuary 2006 to November 2017, to describedefinitions of CDI applied, molecular typingand diagnostic methods used, and key risk fac-tors and expected outcomes.
METHODS
The recent literature was reviewed in a system-atic fashion to identify studies and reportsrelating to the epidemiology of CDI in Japan.The Preferred Reporting Items for SystematicReviews and Meta-Analyses (PRISMA) guidelineswere used to inform search terms, and the lit-erature review process was conducted using thePRISMA Checklist and PRISMA Flow diagram.
40 Infect Dis Ther (2018) 7:39–70
Identification
Searches of MEDLINE-PubMed and EMBASE�weremade using the following primary search terms:C.difficile infection; pseudomembranous colitis; epi-demiology; Japan. Secondary search terms were asfollows: C. difficile diarrhea; C. difficile colitis;enterocolitis; toxic megacolon; hospital-acquireddiarrhea; nosocomial diarrhea; antibiotic-associ-ated diarrhea; incidence; Japan. The publicationswere limited to the English language from 1 Jan-uary 2006 to 27 November 2017.
Selection
Identified abstracts were reviewed by a singlereviewer to remove duplicates and to identifypublications meeting the pre-defined inclusionand exclusion criteria. Inclusion criteria were:Japanese patients or human samples with CDI;observational or non-randomized interventionalstudies; cross-sectional surveys; cohort studies;case–control studies; pharmacy records or claimsdatabases; electronic registers or electronic medi-cal/health records; insurance or administrativeclaims databases studies; registry studies; prospec-tiveor retrospective studies; longitudinalor follow-upstudies; andreviews.Publicationswere includedif they reported on: CDI epidemiology (incidence/prevalence); CDI risk factors; CDI definitions;diagnostic and laboratory test methods; CDIstrains; length of hospital stay (LOS); intensive careunit admission; CDI recurrence; and mortality.There was reliance on the individual publicationsto define CDI and no minimum (discriminatory)definition was used during the selection process.Exclusion criteria were: animal studies; in vitrostudies; case reports; editorials, commentaries andletters; congress abstracts; and non-English lan-guage publications. All publications that met cri-teria for the review were obtained as full articles,reassessed and reviewed.
Quality Determination and DataExtraction
A single reviewer assessed the quality of eachpaper/study according to Oxford Centre for Evi-dence-Based Medicine – Levels of Evidence [8]
(Enhanced Supplementary Material). As most ofthe captured studies did not fall strictly within agiven category, references were also assessed by thesame individual to ensure consistent application ofthe criteria across all publications.
Data from the selected studies were extractedby a single reviewer and used to populate sum-mary tables.
Compliance with Ethics Guidelines
The analysis in this article is based on previ-ously conducted studies and does not involveany new studies of human or animal subjectsperformed by either of the authors.
RESULTS
Identification of Relevant Publications
A total of 385 potential articles were identified,of which 55 were defined as relevant, after
Fig. 1 Assessment of search results to identify key papersfor review and data extraction. Asterisk did not meetinclusion criteria in relation to study population or design(see ‘‘Selection’’). Hash did not include reports of:Clostridium difficile infection (CDI) epidemiology (inci-dence/prevalence); CDI risk factors; CDI definitions;diagnostic and laboratory test methods; CDI strains;length of hospital stay; intensive care unit admission; CDIrecurrence; or mortality. Dagger one identified article wasan erratum of a previously identified study, therefore thestudy was counted only once
Infect Dis Ther (2018) 7:39–70 41
applying the pre-defined inclusion and exclu-sion criteria (Fig. 1). One article was an erratum[9] of a previously identified study [10], there-fore the study was counted only once. Theassigned grades of literature per relevant articleare summarized in the Enhanced Supplemen-tary Material.
Many papers were insufficiently specific: forexample, several papers reported on the valida-tion of novel C. difficile diagnostic assays orlaboratory testing methods in Japan, but didnot include clinical data. Others were excludedowing to a small sample size (n B 8), a focus onpre-clinical evaluation of CDI testing methods,or for reporting CDI contamination in non-pa-tient groups [11–15]. Several papers werereviews or editorials with limited relevance toJapanese CDI epidemiology [3, 16, 17].
Incidence and Prevalence
Twenty-four papers reported data relating to theincidence and prevalence of CDI, or to C. diffi-cile-related disease or diarrhea in Japanesecohorts (Table 1). Most reports were based onretrospective chart reviews (n = 16), eight wereprospective studies; either observational studiesor randomized controlled trials (RCTs), and onewas a systematic review and meta-analysis. Mostof the papers described and defined CDI interms of clinical diagnosis (‘diarrhea’) and lab-oratory findings (‘toxin positivity’). The papersdiffered in their approach to testing for CDI:some actively investigated C. difficile coloniza-tion across cohorts whilst others tested for C.difficile only in patients with clinical symptomssuggestive of CDI. Testing methods also varied.Because of these differences and the hetero-geneity in the patient populations examined inthe publications, including hematopoietic stemcell transplant (HSCT) patients, rheumatologypatients and those with Helicobacter pylori-posi-tive peptic ulcer, it was not possible to examinetrends over time in the incidence of CDI.
Very few papers reported incidence in termsof cases per 10,000 patient-days; most reportedobserved CDI ratios (as percentages), with CDIvariously defined, or prevalence within specificpatient subgroups or populations (Table 1).
Figure 2 depicts reported CDI incidence fromretrospective chart reviews in different patientcohorts, where CDI was defined based on diar-rhea and laboratory detection of fecal C. difficiletoxin.
A chart review of over 22,800 inpatients at asingle tertiary care center (of whom 851 weretested for C. difficile) reported on healthcare-fa-cility onset CDI, defined as diarrhea and a pos-itive toxin test (using C DIFF QUIK CHEKCOMPLETE�) [18]. The CDI prevalence was 5.5cases/1000 patients. The incidence of health-care-facility onset CDI was 3.11 cases/10,000patient-days, compared with 0.2 cases/10,000patient-days for community-onset CDI [18].The authors considered the CDI incidence(hospital- and community-onset) to be ratherlow, which they suggested may have beenattributable to the relatively low frequency oftesting for C. difficile and the relatively lowsensitivity of the EIA toxin detection method.Another large-scale outpatient study (n = 2193)reported the incidence of community-acquiredCDI as 1.4/100,000 patient-years (or 0.14/10,000 patient-years) [19], and a study of bothactively tested in- and outpatients provided anincidence of 0.8 cases/10,000 patient-days [20].A retrospective cohort study based on chart re-views at four tertiary care hospitals reported 160patients aged at least 14 years with hospital-onset CDI, as defined according to clinicalpractice guidelines, giving an incidence of 1.04/10,000 patient-days (or 1.61/1000 admissions)[21]. The low incidence of hospital-onset CDIcompared with that reported in studies fromWestern countries was suggested by the authorsto be a consequence of the different strainsprevalent in different regions, with outbreaks ofhospital-acquired CDI in Western countriesbeing attributed to highly virulent strains thatare not prevalent in Japan. The study did notexplore the strains responsible for the CDI epi-sodes at the four hospitals. The authors alsosuggested that the low frequency of CDI may bea consequence of the low frequency of testingfor C. difficile [21].
Few studies identified in our search describedmeasures to control the incidence of C. difficilein hospitals. Although not strictly an infectioncontrol measure, the prophylactic use of pre-
42 Infect Dis Ther (2018) 7:39–70
Table1
Incidenceandprevalence
ofC.d
ifficile,and
risk
factorsassociated
withC.d
ifficileinfectionreported
inJapanese
patientpopulations
Reference
Stud
yperiod
Stud
ydesign
Patient
popu
lation
C.diffdiagno
sis(n)
C.diffdefin
ition
Incidence
Prevalence
RiskfactorsforCDI
Akahoshi
etal.
[24]
Novem
ber2007–
May
2014
Retrospective,
single-center,
chartreview
n=
308
HSC
T(n
=102
autologous;
n=
206
allogeneic)
n=
30
Occurring
median7days(range
0–36)afterHSC
Tcond
itioning
Diarrhea(C
3loosestools/
24h)
infirst100days
postHSC
T
Positive
CD
toxina
orpositive
CD
toxinplus
GDH
b
Cum
ulative
incidence6.2%
inHSC
Tpopulation
(9.2%
allogeneic;1.0%
autologous)
–AllogeneicHSC
T,totalbody
irradiation,
stem
cellsource,
acuteleukem
ia,d
urationof
neutropenia–lin
kedwith
increasedrisk
forCDI
AllogeneicHSC
T:O
RforCDI
18.6
(95%
CI2.48–1
39)
p\
0.01;duration
ofneutropeniaC
17days
infirst30
days:ORforCDI
10.4
(95%
CI2.37–4
6p\
0.01)
Daida
etal.
[31]
July2003–
Septem
ber2012
Retrospective
case–control
study
viamedicalrecord
review
from
asinglecenter
Pediatricpatients
(aged0–
19years)
admittedto
hospitalwith
cancer
(n=
51)
Matchingcase
controlswere
selected
from
patientswithout
CDIadmittedto
hospitalwithin
2months(before/
after)
admission
for
patientswithCDI
(n=
94)
n=
51TestforC.d
ifftoxins
A/B
a
Appearanceof
symptom
sC
3days
afteradmission
51/189
=26.98%
–Multivariableanalysisof
risk
factorsforhospital-acquired
CDI:youn
gerageisarisk
factor:age0–
3yearsvs.age
4–6years,OR0.13
(95%
CI
0.03–0
.59);p=
0.008,
and
vs.age
C7years,OR0.12
(95%
CI0.03–0
.45);
p=
0.002.
Prolonged
neutropeniaisarisk
factor:
OR1.11
(95%
CI
1.03–1
.20);p=
0.008.
Use
ofC
4antibioticsin
the
60days
from
diagnosisor
referencedate:OR3.55
(95%
CI1.40–9
.04);
p=
0.008
Furuichi
etal.
[32]
August2012–
March
2013
Prospective,non-
interventional
cohortto
assess
ratesof
commun
ity-
acquired
C.d
iffcolonization
(singlecenter)
n=
346children
Health
yneonates
(n=
95)and
pediatricpatients
athospital
admission
(n=
251)
0(0%)C.d
iff(asymptom
atic)
colonization
inneonates
Pediatricpopulation
without
underlying
disease:C.d
iffcolonization
21.6%
and9%
toxinpositive
colonization;
vs.w
ithun
derlying
disease
30.8%
and23.1%
(colonizationandtoxin-
positive
colonization)
Culturedfecalsamples
positive
forC.d
ifftoxin
A/B
b
Asymptom
aticCDI
9%toxin-positive
colonization
inpediatricpatients
withno
underlying
disease
23.1%
toxin-positive
colonization
inpediatrics
with
underlying
disease
–Riskfactorsfortoxin-positive
C.d
iffcolonization:
underlying
disease(O
R4.17,
95%
CI1.15–1
5.04;
p=
0.049);age
12–2
3months(O
R4.19,
95%
CI1.52–1
1.52;
p=
0.01);tube
feeding(O
R24.28,
95%
CI4.70–1
25.34;
p\
0.001);toxin-positiveC.
diff(O
R8.29,9
5%CI
1.87–3
6.84;p=
0.005)
Infect Dis Ther (2018) 7:39–70 43
Table1
continued
Reference
Stud
yperiod
Stud
ydesign
Patient
popu
lation
C.diffdiagno
sis(n)
C.diffdefin
ition
Incidence
Prevalence
RiskfactorsforCDI
Hashimoto
etal.
[26]
Janu
ary1996–
Novem
ber2004
Retrospective
chartreview
(singlecenter)
n=
242
Living-donorliver
transplant
recipients(adult)
Diarrhea76/242;C.d
iffdiarrhea
11/242
C.d
iffdiarrhea:C
3loose
stoolson
C2
consecutivedays
and
positive
stoolcultureand
assays
fortoxinAcand
GDH
d
11/242
C.d
iffdiarrhea
=4.5%
e–
Malegend
er(O
R4.56;9
5%CI
1.02–3
3.3,
p=
0.05)and
serum
creatinine
md/
dLC
1.5(O
R16.0,95%
CI
3.85–6
8.3,
p=
0.0003)
predictedrisk
forC.d
iffdiarrhea
Intensityof
antibioticusedid
notpredictforC.d
iffdiarrhea
Hataet
al.
[62]
Novem
ber2007–
Decem
ber2012
Phase3,
multicenter,
open-labelRCT
(assessing
antibiotic
prophylaxis)
n=
579
Colorectalsurgery
(colorectalcancer
patients;elective
laparoscopic)
Rateof
C.d
iffinfection:
oral–IVprophylaxisgroup
1/289;
IVprophylaxis3/290
Positive
testforC.d
ifftoxins
inpatients
developing
enteritis/colitis/diarrhea
(assay
notdescribed)
Incidencerate
C.d
ifftoxins
inoral–IV
andIV
groups
0.3%
and1.0%
,respectively
(p=
NSbetween
oral–IVandIV
prophylaxis
groups)
5.2cases/1000
patientse
–
Hikone
etal.
[20]
August2011–
Septem
ber2013
Retrospective
chartreview
ofin-
andoutpatient
samples
(single
center)
n=
2193
samples
tested
forC.d
ifftoxin
In-andoutpatient
samples
tested
for
C.d
iff
107specim
enspositive
forC.
difftoxin;
76casesof
healthcare-facility
onsetCDI
Positive
C.d
ifftoxintestb
Incidencerate
0.8
cases/10,000
patient-days
30-day
and90-day
mortalityrates:
7.9%
and14.5%,
respectively
–RiskfactorsforrecurrentCDI:
malignant
disease(O
R7.98;
95%
CI1.22–5
2.2;
p=
0.03);historyof
ICU
hospitalization(O
R49.9;
95%
CI1.01–2
470;
p=
0.049)
Honda
etal.
[18]
Septem
ber2010–
August2012
Retrospective
chartreview
(singletertiary
care
center)
n=
22,863
adult
patients;and1537
C.d
ifftestsin
851
patients
Cases
ofCDIin
anon-outbreak
setting
126casesdiagnosedwithCDI
(86.5%
werehealthcare-
facilityonsetCDI)
Diarrheaandpositive
toxin
assayb
orpresence
ofpseudomem
branous
colitisf
Health
care-facility
onset
CDI:symptom
onset[
3days
from
admission
Com
mun
ity-onsetCDI:
symptom
onsetpriorto
orwithin3days
ofadmission
Health
care-facility
onsetCDI:3.11
cases/10,000
patient-days
Com
mun
ity-onset
CDI:0.2cases/
10,000
patient-
days
forCDI
attributableto
the
studyhospital
30-day
all-cause
mortalityin
CDI
cohort:15.1%
126/22,863
=5.5cases/
1000
patientse
–
44 Infect Dis Ther (2018) 7:39–70
Table1
continued
Reference
Stud
yperiod
Stud
ydesign
Patient
popu
lation
C.diffdiagno
sis(n)
C.diffdefin
ition
Incidence
Prevalence
RiskfactorsforCDI
Hosokaw
aet
al.
[25]
Janu
ary2007–
Decem
ber2008
Retrospective
cohort
(singlecenter)
n=
201patients
AllogeneicHSC
Tpatients(135
unrelatedcord
blood;39
unrelated
bone
marrowand
27related
peripheralblood
stem
cell)
167/201patientstested
for
C.d
iff
17/201
diagnosedC.d
iffdiarrhea
C.diff
diarrhea
in:11/135(9%)
unrelatedcord
blood
recipients;2/39
(6%)
unrelatedBMT
recipients;
4/27
(16%
)relatedPB
STrecipients
C.d
iffdiarrhea:[
3loose
stools/24
hfor2
consecutivedays
and
positive
ELISA
forC.
difftoxinA
Cum
ulative
incidenceof
C.
diffdiarrhea
9%at
post-transplant
day100
–Totalbody
irradiation
associated
withreducedrisk
ofC.d
iffdiarrhea
C.d
iffdiarrhea
was
notacause
ofanydeath;
norecurrence
ofC.d
iffdiarrhea
after
treatm
ent
Iwam
oto
etal.
[27]
Twoperiods:March
2004–F
ebruary
2006
andApril
2008–D
ecem
ber
2008
Prospective
observational
cohort(single
center)
n=
1226
Rheum
atology
inpatients
54casesof
healthcare
associated
infectionof
which
2wereC.
diffinfection(1
patientin
each
studyperiod)
Health
care-associated
infection:
developing
[3days
afteradmission
2/1226
inrheumatology
patients(0.16%
)e
––
Iwashima
etal.
[44]
April2005–
March
2008
Retrospective
cohort
studyassessing
genotypicfeatures
ofisolates
and
clinical
characteristicsof
CDI(single
center)
n=
610subm
itted
specim
ens
Patientswithstools
foun
dpositive
for
C.d
iffcultu
re(n
=106;
ofwhich
35excluded
asasym
ptom
atic
carriersand
n=
14excluded
fornon-toxigenic
strains)
71C.d
iffinfectioncases
assessed
PCRassessmentof
toxinA
andB;ribotyping
CDI:diarrhea
orcolitis
withpositive
testforC.
difftoxinBandno
other
enteropathogenic
microorganism
s
Recurrent
CDI:recurring
within2monthsof
previous
episode
Incidenceof
CDIs
withbinary
toxin-
positive
strains
5.6%
(noted
inpatientswithnon-
severe
CDI)
Prevalence\
5CDIcases/
month
–
Kaneko
etal.
[30]
Janu
ary2006–A
pril
2009
Retrospective
cohort
investigatingfor
CDIduring
active
phaseof
inflammatory
boweldisease
(singlecenter)
n=
137
Activeulcerative
colitis
55/137
(40.1%
)tested
samples
wereCDIpositive
Presence
oftoxinA
antigeng
ingutlavage
40.1%
inasample
tested
forpossible
CDI
––
Kobayashi
etal.
[21]
April2012–
Septem
ber2013
Retrospective
cohort
studybasedon
chartreview
atfour
teaching
hospitalsin
Japan
Patients
aged
C14
years
withhospital-onset
CDI
n=
160withhospital-onset
CDI
According
toSH
EA/IDSA
2010
guidelines,b
ased
onpositive
CD
toxin
EIA.bHospital-o
nset
CDI:hospitalized
for
cond
itionotherthan
CDIforC
2days
1.04
casesper10,000
patient-days;1.61
casesper1000
admissions
––
Infect Dis Ther (2018) 7:39–70 45
Table1
continued
Reference
Stud
yperiod
Stud
ydesign
Patient
popu
lation
C.diffdiagno
sis(n)
C.diffdefin
ition
Incidence
Prevalence
RiskfactorsforCDI
Kom
atsu
etal.
[22]
June
2008–
Decem
ber2013
Single-centerRCT
n=
379
Looking
atefficacyof
perioperative
synbiotics
topreventinfectious
complications
(particularly
surgicalsite
infection)
Colorectalsurgery
(laparoscopic)
0/168casesC.d
iffinfectionin
synbiotics
groupand1/194
incontrolgroup
Authorreportsuseof
synbiotics
suppressed
increasesin
potentially
pathogenicC.d
iff(detectedin
4%before
surgeryboth
groups;detected
in4%
7days
aftersurgeryin
synbioticgroupvs.1
3%control;p=
0.05
vs.d
aybefore
surgery
Gut
microbiotaassessed
byYIF-SCAN
andPC
Ranalyses
1/379colorectal
surgerypatients
(0.3%)e
––
Mizui
etal.
[37]
February
2010–
February
2011
Retrospective
studyof
risk
factorsforC.
diffdiarrhea
(singlecenter)
n=
2716
patients
givenan
injectableantibiotic
Studyalso
assessed
impact
ofprobiotics
Inpatientsgiven
antibiotics
29hadC.d
iffdiarrhea
(2687hadnon-C.diff
diarrhea)
Riskfactorsinvestigated
betweengroups
re:useof
antibioticsC
8days;enteral
nutrition;
IVhyperalim
entation;fasting,
proton
pumpinhibitorsH
2
blockers;serum
albumin
B2.9g/dL
C.d
iffdiarrhea;testsnot
defin
ed–
–RiskfactorsforC.d
iffdiarrhea
were:antibiotic
useC
8days
(OR4.071;
95%
CIC
I1.333–
12.430;
p=
0.014),IV
hyperalim
entation
(OR
3.414;
95%
CI1.469–
7.934;
p=
0.004),P
PIs(O
R3.224;
95%
CI1.421–
7.315;
p=
0.005),H
2blockers
(OR2.376;
95%
CI
1.047–
5.391;
p=
0.039)
Moriand
Aoki
[19]
Janu
ary2010–
Decem
ber2014
Retrospective
case–control,
epidem
iological,
single-centerstudy
assessingrisk
factorsforCDI
Outpatients
(1,914,011
patient-
yearsexam
ined)
CDIcases
Age-andsex-matched
controls(C.d
ifftoxin-
andcultu
re-
negative)
26patientshadcommun
ity-
acquired
CDI
Com
mun
ity-acquired
CDI:
outpatient
presentation
withdiarrhea,stool
cultu
repositive
C.d
ifftoxinassaya,h
Incidencefor
commun
ity-
acquired
CDI1.4/
100,000patient-
years
–84.6%
ofpatientswith
commun
ityacquired
CDI
hadpriorexposure
toantibiotics
Patientswithcommun
ity-
acquired
CDImorelikelyto
have
hadpriorantibiotics
(OR8.12;95%
CI2.43—
26.98)
Ogamietal.
[38]
4-year
period
(dates
notgiven)
Single-center,
retrospective
hospitalcohort
n=
463
Inpatientswith
antimicrobial
associated
diarrhea
95/463
cases(20.5%
)wereCDI
CDImanifestingas
antimicrobial-associated
diarrhea
(C3
stools/day[
48hafter-
wardadmission)and
stooltoxinpositive
(Aand/or
B)a
––
Increasedwarduseof
antimicrobialsclindamycin
(OR1.739;
95%
CI
1.050–
2.881;
p=
0.032)
andpiperacillin(O
R1.598;
95%
CI1.006–
2.539;
p=
0.047)
increasedrisk
ofCDI
46 Infect Dis Ther (2018) 7:39–70
Table1
continued
Reference
Stud
yperiod
Stud
ydesign
Patient
popu
lation
C.diffdiagno
sis(n)
C.diffdefin
ition
Incidence
Prevalence
RiskfactorsforCDI
Oshim
aet
al.
[40]
Publishedstudies
1990–2
016
System
aticreview
and
meta-analysisof
67publishedstudies
Adults
andpediatric
(B18
years)
patientsreceiving
PPIwho
developed
acute-onset
diarrhea.A
lso,
controlgroup
n=
17,217
inthetestgroup;
n=
286,018in
thecontrol
group
Recurrent
CDIoccurred
inn=
1279;n=
5459
inthe
controlgroup
Laboratoryconfi
rmationof
C.d
iffor
clinical
defin
ition.
Nofurther
detailprovided
––
PPIuseincreasedrisk
forinitial
CDIepisode(random
effects
model,overallOR2.34,95%
CI1.94–2
.82;
p\
0.00001)
Age-stratified
subgroup
analyses:significant
associations
betweenPP
Iuse
andinitialCDIepisodein
adults(O
R2.30,9
5%CI
1.89–2
.80;
p\
0.00001)
and
pediatrics
(OR3.00,9
5%CI
1.44–6
.23;
p\
0.00001)
Roughead
etal.
[36]
2008–2
013
(insurance
database)
1996–2
014
(hospitaldataset)
Retrospective
data
from
worker
insurancedatabase
andahospitalin-/
outpatient
dataset
from
asingle
center
1.2millionpatient
recordsexam
ined
andsequence
symmetry
analysis
used
toassessPP
Iuseasrisk
factor
for
CDI
n=
310patients
received
PPIsand
oralvancom
ycin
(proxy
indicator
forCDI)
––
––
Positive
associationbetween
PPIuseandCDI(adjusted
sequence
ratioforinsurance
dataset5.40;95%
CI
2.73–8
.75andforhospital
dataset3.21;95%
CI
2.12–4
.55)
Sadahiro
etal.
[63]
May
2008–
October
2011
Prospective,single-
center
RCT
comparing
oral
antibioticsand
probiotics
pre
surgeryto
prevent
infection
n=
310
Colon
cancer
Nochange
indetectionof
C.
difftoxinacrossthree
treatm
entgroups
(probiotics;antibiotics;
control(noprobiotics
orantibiotics))
Assessm
entof
C.d
ifftoxin
(AandB)in
stool
samples
byRID
ASC
REEN
Rates
ofCDI
increasedpost-
operativelyin
all
groups
(probiotic
group,
from
2.0%
to7.0%
;antibiotic
group,
5.1%
to9.1%
;control
group,
2.1%
to10.5%)
––
Sasabuchi
etal.
[28]
July2010–M
arch
2013
Retrospective
cohort
studyusingthe
Japanese
Diagnosis
Procedure
Com
bination
database
(multicenter)
n=
15,651
receiving
prophylaxis
n=
15,651
controls
Severe
sepsisand
receivingstress
ulcerprophylaxis
within2days
ofhospitaladmission;
propensity-
matched
controls
didnotreceive
prophylaxis
Inpropensity-m
atched
cohort,
215and204casesof
CDIin
thestressulcerprophylaxis
andcontrolgroups,
respectively
Not
specified,b
utIC
D-10
codesused
forother
defin
itions.C
DIcoded
as‘com
plication’
inmedicalrecordsduring
hospitalization
1.4%
instressulcer
prophylaxisgroup
and1.3%
incontrol
(p=
0.588)
––
Infect Dis Ther (2018) 7:39–70 47
Table1
continued
Reference
Stud
yperiod
Stud
ydesign
Patient
popu
lation
C.diffdiagno
sis(n)
C.diffdefin
ition
Incidence
Prevalence
RiskfactorsforCDI
Suzuki
etal.
[23]
April2010–M
arch
2012
Single-center
prospectivecohort
preandpost
intensiveinfection
controlmeasures
n=
80
Hospitalized
patients
–Based
onmedicalrecords
andhealthcare
resource
use
New
-onset
nosocomialC.
diff-associated
disease
C.d
iff-associated
diseasereduced
from
0.47
cases/
1000
inpatient
days
to0.11
(p\
0.001)
after
intensiveinfection
controlteam
interventions
––
Takahashi
etal.
[39]
Novem
ber2010–
October
2011
Multicenter
case–control
and
cohortstudy
n=
1026
CDI
n=
878controls
NationalHospital
Organization
cohort
Assessedfornewly
diagnosedCDIand
matched
controls
(noCDI)
93.9%
ofCDIcasesdeveloped
within48
hof
hospital
admission
GIsymptom
s,clinical
suspicionof
CDIand
positive
C.d
ifftoxins
a,h,i
from
stoolor
C.d
iffisolationfrom
stool
cultu
res,or
both
––
RiskfactorsforCDI
developm
ent:disruption
offeeding/parenteraland
enteralfeeding;first-and
second
-generationcephem
antibiotics(O
R1.44;95%
CI1.10–1
.87),third-and
fourth-generationcephem
antibiotics(O
R1.86;95%
CI1.48–2
.33),carbapenem
antibiotics(O
R1.87;95%
CI1.44–2
.42)
Com
orbidities
morecommon
inpatientswithCDI
Analysisof
924casesnoted11%
mortalitywithin30
days
ofCDIonset
Use
ofvancom
ycin
reduce
mortality(O
R0.43;95%
CI
0.25–0
.75)
PPIsandpenicillindidnot
increase
risk
forCDI
Watanabe
etal.
[64]
Janu
ary–June
2005
Multicenter,
retrospective
cohort
n=
294fecal
samples
subm
itted
forC.d
ifftesting
Hospitalized
patients
79/294
(5.5
cases/1000
beds
monthly)wereC.d
ifftoxin
A?
C.d
ifftoxintestc
5.5cases/1000
beds
monthly,assessed
forC.d
iffwere
foun
dto
beC.diff
toxinA?
––
48 Infect Dis Ther (2018) 7:39–70
Table1
continued
Reference
Stud
yperiod
Stud
ydesign
Patient
popu
lation
C.diffdiagno
sis(n)
C.diffdefin
ition
Incidence
Prevalence
RiskfactorsforCDI
Yasun
aga
etal.
[29]
2007–2
010
Retrospective
database
review
:analysisof
factors
affectingC.d
iff-
associated
disease
andoutcom
esof
C.d
iffdiarrhea
afterGIsurgery
Japanese
Diagnosis
Procedure
Com
bination
inpatientdatabase
(multicenter)
n=
143,652
Inpatients/G
Isurgical
patients
409casesof
C.d
iffdiarrhea
(0.28%
)
Higherratesin
colorectal
surgery(0.37%
)vs.
gastrectom
y(0.21%
)and
esophagectom
y(0.25%
)(p\
0.001)
CD
enterocoliticIC
D-10
code
Rate0.28%
Riskfactorsincluded:
olderage;higher
Charlson
comorbidity
index;
longer
pre-
operativeLOS;
non-academ
iccenter
care
In-hospitalmortality
higher
inC.d
iffdiarrhea
than
innon-C.d
iffdiarrhea
(3.4%
vs.
1.6%
:OR1.83;
95%
CI1.07–3
.13,
p=
0.027)
409/143,652
=2.8/1000
patientse
Riskfactorsincluded:o
lder
age;
higherCharlsoncomorbidity
index;longer
pre-operative
LOS;
non-academ
iccenter
care
In-hospitalmortalityhigher
inC.d
iffdiarrhea
(3.4%
vs.
1.6%
innon-C.d
iffdiarrhea:
OR1.83;95%
CI1.07–3
.13;
p=
0.027)
LOSattributableto
post-
operativeC.d
iffdiarrhea
12.4
days
(95%
CI9.7–
15.0;
p\
0.001)
BMTbone
marrowtransplantation;
CIconfi
denceinterval;GDH
glutam
atedehydrogenase;GIgastrointestinal;HSC
Thematopoieticstem
celltransplantation;
IVintravenous;LOSlength
ofstay;O
Rodds
ratio;
PPIproton
pump
inhibitor;RCTrand
omized
clinicaltrial;YIF-SCAN
YakulyIntestinalFlora-SC
AN
system
aTOX
A/B
QUIK
CHEK�
bC
DIFFQUIK
CHEK
COMPL
ETE�
cUNIQ
UIC
KdCD
CHECK
eCalculatedfrom
data
availablein
thepublicationandnotstated
inthepublication
fXPE
CT
C.D
IFFtoxinA/B
gC.d
iffToxin
Atest,O
xoid
hIm
mun
oCardCD
toxinA&B
iX/pectToxin
A/B
Infect Dis Ther (2018) 7:39–70 49
and probiotics has been suggested to maintainthe colonic microbiota and potentially reducethe development of CDI [22]. A single-centerRCT of 379 patients undergoing colorectal sur-gery evaluated the impact of perioperative syn-biotics (combination of pro- and prebiotics) onpost-surgical outcomes and fecal microbiotacomposition, finding that patients administeredsynbiotics before surgery had a lower incidenceof C. difficile in their fecal microbiota comparedwith control patients. The incidence of CDI waslow, with only one patient in the control groupdeveloping CDI, while none in the treatedgroup did so. The authors suggested a potential
role for synbiotics in suppressing overgrowth ofC. difficile after surgery [22]. The literaturesearch also identified a study reporting theimpact of infection control interventions onCDI occurrence. A medical record-based C. dif-ficile-associated disease at a single center wasreported to have an incidence of 0.47 cases/1000 inpatient days, which fell to 0.11 cases/1000 patient-days after intensive infectioncontrol intervention [23]. Infection controlmeasures included carbapenem restriction andcontinuous instruction to the ward staff oninfection control measures [23].
Some of the epidemiological reports high-lighted the incidence and prevalence of CDI inparticular patient groups (Table 1; Fig. 2). Forexample, in patients undergoing HSCT, thecumulative incidence of CDI was 6.2% for allpatients, compared with 9.2% in the allogeneicHSCT subpopulation, 1.0% in the autologousHSCT subpopulation and 9% in a cohort ofHSCT recipients who received unrelated cordblood [24, 25]. In a cohort of liver transplantpatients, CDI-associated diarrhea occurred at arate of 4.5% [26]. Among rheumatology inpa-tients, CDI was observed in 0.16% [27]; in alarge cohort of patients with sepsis, hospital-acquired CDI was observed in 1.3% of patientswithout and 1.4% with ulcer prophylaxis [28]. Aretrospective database review of over 140,000gastrointestinal (GI) surgery patients reportedCDI (ICD-10 definition) in 0.28% of the studypopulation, or a prevalence of 2.8 cases/1000patients [29]. In a smaller cohort study ofpatients with active ulcerative colitis, 40.1%tested positive for possible CDI [30]. A study ofpediatric patients with cancer who were hospi-talized at a single center reported CDI (clinicalsymptoms and positivity for toxin EIA usingTOXA/B QUIK CHEK) in 27% of the studypopulation [31]. The authors suggested that thehigh incidence of CDI compared with otherstudies of similar patient populations in non-Japanese settings may be because of the muchlonger length of stay for patients with cancer inJapan compared with other countries [31].
Among the studies that reported on possibleC. difficile colonization of patient groups was anepidemiological study of hospitalized pediatricpatients. At least 1 in 10 pediatric patients
Fig. 2 Clostridium difficile infection (defined as diarrhea/CD toxin) reported in retrospective cohorts of Japanesepatients. CRC colorectal cancer, GI gastrointestinal, HSCThematopoietic stem cell transplantation, IBD inflamma-tory bowel disease, pts patients, RA rheumatoid arthritis.Patient numbers represent those diagnosed with Clostrid-ium difficile infection
50 Infect Dis Ther (2018) 7:39–70
harbored C. difficile asymptomatically, whilefecal cytotoxin was found in 9% of otherwisehealthy children and 23.1% of children withunderlying disease [32]. However, these find-ings should be interpreted with caution as theinclusion of patients younger than 3 years oldmay increase the likelihood that the diarrheahad a cause other than C. difficile [33].
The reviewed studies included reports of bothhospital and community patients. No paperswere identified that specifically related topatients in long-term healthcare facilities andonly one paper reported on changes in infectionover time [23]. Therefore, depending on thepatient group and methods used to define andassess CDI, the incidence varied between 0.8 and4.71/10,000 patient-days, while the prevalencewas between 0.3 and 5.5 cases/1000 patients.
Risk Factors
Known risk factors for CDI include: co- andpreviously administered broad-spectrumantibiotics; age and comorbidities; poor infec-tion-control practices; GI tract surgery; andgastric-acid suppressing agents [13, 34–36].Thirteen studies in our search, including a largedatabase study representing 40% of all adult-care hospitalizations, identified risk factors forCDI in Japanese patients, which included olderage, higher comorbidity index; gastric acid-suppressing proton pump inhibitors (PPIs); anda longer pre-operative LOS before GI surgery[19, 20, 24–26, 29, 31, 32, 36–40] (Table 1).Malignant disease and intensive care unit (ICU)stay were linked with increased risk for CDIrecurrence in in- and outpatients [20]. Six arti-cles reported on the number of days spent inhospital prior to surgery or CDI diagnosis,indicating a wide variation in inpatient staybefore diagnosis [18, 20, 29, 38, 41, 42]. Theonly study to formally compare pre-operativenumber of days in hospital for patients who didand did not develop CDI found no differencebetween the two patient groups, with a median(interquartile range) of 6 (3–14) days and 5 (3–8)days, respectively (p\0.001) [29].
Patients undergoing HSCT are recognized tobe at particular risk of infection. Allogeneic
HSCT, conditioning for HSCT, acute leukemiaand prolonged neutropenia in the first 30 daysafter HSCT may all confer an increased risk forCDI as reported in a single-center study [24]. Incontrast, it was noted that among allogeneicHSCT patients, treatment with total body irra-diation may reduce post-transplant risk of CDI[25].
A study of pediatric patients reported thattube feeding was significantly associated withhigher colonization rates by toxin-positiveC. difficile [32].
Antibiotic use was a risk factor for C. difficilediarrhea in a number of studies [19, 31, 37, 38];however, in a study of liver transplant patients,the intensity of antibiotic use (measured as useof preoperative antibiotics or the number ofantibiotics used postoperatively) was not a pre-dictor for C. difficile diarrhea [26]. Among hos-pitalized pediatric patients with cancer, use of awide variety of antibiotics (the study specified atleast four different types) in the 60 days prior toCDI diagnosis was a significant risk factor forthe development of CDI [31].
Specific Strains Responsible for CDI
A review of CDI in Asia in 2013 reported thatPCR ribotypes 027 and 078 were rare, whilevariant toxin A-/toxin B? strains of ribotype017 were common. Furthermore, in Japan,common ribotypes include 014, 002 and 001,and ribotype smz/018 has been implicated inwidespread disease [3]. The review noted that avariety of typing techniques has been used inJapan, including tcdA and tcdB detection, pulsedfield gel electrophoresis (PFGE), PCR ribotypingand slpA typing. Although molecular typinghad identified toxigenic A-B? strains, theauthors did not comment on binary (CDT)toxin assessment or C. difficile surveillance inJapan [3].
In our literature search, 16 papers providedfurther details of testing methods used in Japanand described reports on the isolates and strainsassociated with CDI in Japanese cohorts(Table 2). The methods used to detect CDI and/or detect, isolate and type C. difficile includedstool culture and C DIFF QUIK CHEK
Infect Dis Ther (2018) 7:39–70 51
Table2
Summaryof
studiesdescribing
C.d
ifficilestrains,testmethods
andassayforbinary
toxinin
Japan
Reference
Stud
yperiod
Stud
ydesign
Patient
popu
lation
C.diffdefin
ition
Testmetho
dsIsolates
andstrains
Binarytoxin
Collin
set
al.[3]
Pre-2013
Narrative
review
andmeta-
analysisof
42cohortstudies
onC.d
iffin
Asian
coun
tries
Various
Various
InJapan:
tcdA
andtcdB
characterization,P
FGE,
PCRribotyping
andslpA
typing
Predom
inance
ofribotype
smz(018)in
pastdecade
Other
common
ribotypes:
014,
002,
001
Not
specifically
mentioned
inreview
ofJapanese
papers
Iwashima
etal.
[44]
April2005–
March
2008
Retrospective
cohortstudy
assessing
genotypic
features
ofisolates
and
clinical
characteristicsof
CDI(single
center)
n=
610subm
itted
specim
ens
Patientswithstools
foun
dpositive
forC.
diffcultu
re(n
=106;
ofwhich
35excluded
asasym
ptom
atic
carriersandn=
14excluded
fornon-
toxigenicstrains)
71CDIcasesassessed
CDIdefin
edas:diarrheaor
colitiswithtoxinB
positive
C.d
iffandno
otherenteropathogenic
microorganism
s
PCRassessmentof
toxins
AandBandribotyping
Isolates
A?B?CDT?:4/71
A?B?CDT-:58/71
A-B?CDT-:9/71
Ribotype
A?B?CDT?:2werej52;
1was
nc07109;
1was
km0403
A?B?CDT-:19
weresm
z;14
wereyok;
13werehr;1
2other
A-B?CDT-:6weretrf;2
werefr;1was
sgf
Nopredom
inantribotype
spreading;thedominant
typesweresm
z,yokandhr
(hr=
equivalent
toribotype
014)
Noribotypes027and078
foun
din
thestudy
Durationof
CDIlonger
inyokgroup(p\
0.05)
Incidenceof
CDIs
withbinary
toxin-
positive
strains5.6%
(noted
innon-severe
CDI)
52 Infect Dis Ther (2018) 7:39–70
Table2
continued
Reference
Stud
yperiod
Stud
ydesign
Patient
popu
lation
C.diffdefin
ition
Testmetho
dsIsolates
andstrains
Binarytoxin
Katoet
al.
[43]
February
2004–
April2004
Single-centerstudy
tovalidate
efficacyof
slpA
sequence
typing
28samples
positive
fortoxinAfrom
17patientswith
C.d
iffdiarrhea
C.d
iffdiarrhea:22/28
samples
positive
bystoolcultu
re
Not
specifically
defin
ed(see
testingmethods)
Detection
oftoxinA
using
UNIQ
UIC
K
slpAsequence
typing
Allsamples
except
hj2-2
isolatewere
A?B?CDT-
(positive
fortoxinA
andtoxinB
butnegative
forbinary
toxin)
smz-1(n
=10)andsm
z-2
(n=
6)accoun
tedfor73%
strains
Strain
patternsuggested
nosocomialinfection
Yok-1,yok-2,t25–1
,hr-1and
hj2-2identifiedin
atleast1
patienteach
Binarytoxinassessed
Katoet
al.
[49]
2003–2
007
Multicenter
study
typing
C.d
iffisolates
byslpA
sequencing
160stoolsamples
from
symptom
atic
patients
(hospitalized
witha
diagnosisof
antibiotic-associated
diarrhea
orcolitis)
Not
specifically
defin
ed(see
testingmethods)
90stoolsam
plesweretyped
ofwhich
77werepositive
bycultu
reforC.d
iff
Stoolcultu
re:PC
Rfortoxins
AandB,and
CDT
PCRribotyping
andslpA
sequence
typing
Smzsequence
type
was
dominantanddetected
bycultu
reand/or
typing
in61/99stoolsamples
positive
fortoxiccultu
reand/or
direct
slpA
sequencing
(smzin
62%;
smz-01,smz-02,smz-04)
One
isolatetype
gc8
correspond
edwithPC
Rribotype
027BI/NAP1
/027);no
PCRribotype
078
foun
d
Directtyping
from
DNA
extractedfrom
stool
samples:77/90were
positive
forC.d
iffand
typing
results
agreed
with
isolated
strain
typing
slpAsubtypes
smz-01,-02
and
-04foun
din
51/86(59%
)of
stoolsamples
that
were
tcdB
-positiveC.diff
cultu
red
andin
67%
ofstoolwhere
direct
typing
couldbe
obtained
Of87
isolates,7
5(86%
)wereA?B?
and12
(14%
)were
A-B?;3A?B?
isolates
werepositive
forPC
Rdetecting
thebinary
toxin
gene
(A?B?CDT?)
Infect Dis Ther (2018) 7:39–70 53
Table2
continued
Reference
Stud
yperiod
Stud
ydesign
Patient
popu
lation
C.diffdefin
ition
Testmetho
dsIsolates
andstrains
Binarytoxin
Kaw
ada
etal.
[47]
October
2009–January
2010
Single-centerstudy
evaluating
asinglekitfor
rapiddetection
ofGDH
and
toxinA/B
infeces(as
diagnosisof
C.
diffinfection)
60specim
ensfrom
60patientswith
antibiotic-associated
diarrhea
C.diff
cultu
rewasreference
method
C.d
iffevaluated28
inpatientsdiagnosedas
having
CDI
Evaluationof
CDIFFQUIK
CHEK
COMPL
ETE�vs.
GDH
detectionby
Immun
oCardandtoxin
A/B
detectionby
TOX
A/B
The
kithadGDH
sensitivity
100%
;specificity
93.3%;
negative
predictive
value
100%
KithadToxin
A/B
sensitivity
78.6%,specificity96.9%
comparedwithtoxigenic
cultu
re(culture
Bpositive)
The
22/23specim
ensthat
weredualpositive
for
GDH
andtoxinA/B
were
cultu
repositive
Dualn
egativesby
thekitwere
C.d
iffcultu
renegative
Not
reported
Kikkawa
etal.
[65]
Janu
ary–June
2005
Multicenter
study
lookingat
prevalence
ofA-/B?
strains
infecalsamples
subm
ittedforC.
difftests
C.d
iffisolated
in159/332specim
ens
Aspertestmethods
Culture
PCRanalysisof
toxigenic
typing
Genotypingby
PCR,
ribotyping
andPF
GE
332sample;C.d
iffisolated
from
159:
137strains
(41%
exam
ined
specim
ens
and86%
ofisolated
C.
diff)
wereA?B?;10
(3%
and6%
)wereA-B?
and
12(4%
and8%
)were
A-B-
Therefore
10(6.3%)of
159
C.d
iffstrainswereA-B?
All10
A-/B?
strainshad
identicalpatternby
PCR
ribotyping
Not
reported
54 Infect Dis Ther (2018) 7:39–70
Table2
continued
Reference
Stud
yperiod
Stud
ydesign
Patient
popu
lation
C.diffdefin
ition
Testmetho
dsIsolates
andstrains
Binarytoxin
Kobayashi
etal.
[45]
April2008–
March
2009
Single-center
retrospective
studyto
test/validatethe
3-dayrulefor
ordering
aC.
difftoxintestin
Japanese
patients
1597
stoolcultu
res
from
992patients;
880CD
toxintests
performed
in529
patients
83speciesfrom
81specim
ensconsidered
entericpathogens
Aspertestmethods
CD
toxinby
TOX
A/B
QUIK
CHEK�
Rateof
positive
stoolcultu
rein
differentpatientgroups:
14.2%
outpatients;3.6%
inpatientB
3days;1.3%
inpatientsC
4days
RespectiveCD
toxinpositive
testrates:1.9%
outpatients;
7.1%
inpatientB
3days;
8.5%
inpatientsC
4days
The
studyvalidates
the3-day
rule:the
rulecanbe
used
toestimatethepre-test
probability
ofastool
microbiologicaltest
Not
reported
Kun
ishima
etal.
[66]
February
2003–
February
2006
Single-centerstudy
ofantimicrobial
susceptibilityof
C.d
iffisolates
Studied157C.d
iffisolates
from
patientswith
diarrhea
and
probableCDI
–Antim
icrobialsensitivityof
isolates:broth
microdilution
method
todeterm
ineMIC
sof
15drugs
Foun
dno
strainsresistantto
either
metronidazoleor
vancom
ycin
Not
reported
Kuw
ata
etal.
[46]
April2012–
March
2013
Single-centerstudy
ofmolecular
epidem
iology
and
antimicrobial
sensitivityof
C.
diffisolates
C.d
iffisolates
(n=
130)
–Toxin
genotypes;MLS
Tand
eBURST
analysis
Resultscomparedwith9
strainspreviouslyanalyzed
byPC
Rribotyping
Strainsidentifiedby
CDIFF
QUIK
CHEK
COMPL
ETE�;multiplex
PCRfortoxigenictype
95toxigenicstrains(73%
),including7A-B?CDT-
and3A?B?CDT?
(23
sequence
types)
35(27%
)non-toxigenic
strains(12sequence
types)
Sequence
type
(ST)17was
mostcommon
(21.8%
)
MLS
TandeBURST
show
ed139strainsbelonged
to7
groups
andsingletons;most
A?B?CDT-
(89/91,
98%)wereclassedinto
group1
MLS
TandeBURST
suggest
mostA?B?CDT-
strains
(including
ST17,S
T2,
ST8)
may
bederivedfrom
ST28
Thisstudyreported
aprevalence
ofA-B?CDT-
(5%)
andA?B?CDT?
(2%),which
isconsidered
low
comparedwith
MLST
studiesin
China
andSpain
Infect Dis Ther (2018) 7:39–70 55
Table2
continued
Reference
Stud
yperiod
Stud
ydesign
Patient
popu
lation
C.diffdefin
ition
Testmetho
dsIsolates
andstrains
Binarytoxin
Mikam
oet
al.
[52]
May
2012–
May
2015
Phase3,
multicenter
(35
inJapan),
double-blin
dRCT,n
=93
CDIdiagnosedby
EIA
(97%
)and
stoolcultu
re(3%)
Adults
(C18
years)
prescribed
SOC
antibioticsforCDI
withplanned
duration
10–1
4days
Inpatients,
n=
86;C
65years,
n=
85
Diarrhea(C
3loosestools/
24h)
?positive
stool
testfortoxigenicC.d
iff
Cellcultu
recytotoxicity
assays,stool
cultu
rewith
toxigenicstrain
typing,
stoolcultu
rewithtoxin
detectionfrom
C.d
iffisolates
orcommercially
availableassays
(ELISA/
PCRwithC
94%
specificity)
54strainsidentifiedfrom
cultu
re.P
CRribotypes
were052(28%
),018
(19%
),002(15%
),369
(9%),159(6%),005(4%),
173(4%),012(2%),014
(2%),043(2%),056(2%),
103(2%),212(2%),235
(2%),254(2%),632(2%).
052isolated
from
11of
35sitesand018isolated
from
9of
35sites
–
Moriet
al.
[42]
12-m
onth
period
in2010
Single-center
retrospective
analysisof
stool
cultu
redatabase
tostudyextent/
reasonsfor
incorrect
diagnosisof
CDI
n=
975stoolcultu
resamples
Definitions:toxigenic
C.d
iff,C
.diff
withany
toxingene;CDI,
diarrhea
plus
atoxigenic
C.d
iffisolate
PCRassayof
toxingene
A,B
andbinary
PCRribotyping
Incidenceof
healthcare-
facilityonsetCDI(within
48h)
estimated
at1.6
cases/10,000
patient-days
The
prevalence
rate
oftoxigenicC.d
iffin
allstool
cultu
reswas
13%
(127/975)
177C.d
iffisolates
detected
ofwhich
127were
toxigenic:124(70%
)A?B?;3(1.7%)A-B?
The
mostcommon
ribotype
was
369(21.6%
),with018
(10.8%
);014/020and002
were9.9%
each
Clin
icallyim
portantisolates
such
as027and078were
notidentified
58(45.7%
)withtoxigenicC.
diffhadun
form
edstool;
incidenceof
CDIwas
1.6/
10,000
patient-days
But
ofthese58
cases,40
were
notdiagnosedin
routine
testingdueto
lack
ofclinicalsuspicion(24.1%
)or
anegative
C.d
ifftoxin
assayresult(44.8%
)
AmongA?B?,
12/177
(6.8%)were
CDT?
56 Infect Dis Ther (2018) 7:39–70
Table2
continued
Reference
Stud
yperiod
Stud
ydesign
Patient
popu
lation
C.diffdefin
ition
Testmetho
dsIsolates
andstrains
Binarytoxin
Oka
etal.
[67]
2002–2
005
Two-center
study
ofmolecular
characterization
ofC.diff
isolates
from
single,
relapseand
recurrentcases
n=
73clinicalisolates
ofC.d
iff
(n=
20isolates
from
20singleinfections;
n=
53isolates
from
20recurrentcases)
Astestmethods
PFGEandPC
Rribotyping,
andPC
Rtoxindetection
11ribotypes
Of73
strainsstudied,
67strains(91.8%
)A?B?;2
weretoxinA-,B
?[B?]
(2.7%);4(5.4%)were
A-B-
80%
ofrelapses
werecaused
bythesamestrain
asthe
firstinfection;
20%
were
dueto
adifferentstrain
–
Sawabe
etal.
[50]
Novem
ber1999–
October2004
Molecular
analysis
ofC.diff
isolates
linkedwith
diarrhea
orcolitisat
asingle
center
n=
148isolates
Astestmethods
PCRandPF
GEribotyping
Toxin
(A,B
andCDT)
determ
ined
byPC
R
26PC
Rribotypesam
ong148
isolates
Shift
from
predom
inant
ribotype
a(15/33;45%
in2000)to
ribotype
f(identicalto
smz)
(18/28;
64%
in2004)
PFGEallowed
furthersub-
classification:
fisolates
were
of4typesand11
subtypes
Onlyoneribotype
027
recovered
110/148(74%
)A?B?CDT-;
33/148
(22%
)A-B?CDT-;
5/148(3%)
A?B?CDT?
Infect Dis Ther (2018) 7:39–70 57
Table2
continued
Reference
Stud
yperiod
Stud
ydesign
Patient
popu
lation
C.diffdefin
ition
Testmetho
dsIsolates
andstrains
Binarytoxin
Senohetal.
[51]
April2011–
March
2013
fornon-
outbreak
2010
and2009
outbreak
data
Multicenter
study
toassessC.d
iffisolates
inJapan
n=
120C.d
iffisolates
during
anon-
outbreak
season;
n=
18andn=
21isolates
from
hospitalsduring
outbreaks
Astestmethods
Toxin
detectionandtyping
byPC
R120outbreak
isolates:80%
wereA?B?CDT-,
15.8%
were
A-B?CDT-;4.2%
A?B?CDT?
PCR-ribotypesm
z(A?B?CDT-)accoun
ted
for34.2%
isolates
AllA-B?CDT-
isolates
werePC
Rribotype
trf
Non-outbreakisolates:Japan
ribotypessm
z(018)and
ysmz39.2%,and
Japan
ribotype
trf15.8%
Types
smz/ysmzalso
predom
inated
inoutbreaks
5binary
toxin-positive
isolates(only1was027and
1was
078)
Alltrfisolates
wereA-B?
(new
ribotype
369)
Highratesof
resistance
toantimicrobialsobserved
inthe018isolates
See‘Isolates
and
strains’
Shim
izu
etal.
[41]
April2013–
March
2014
Studyto
evaluate
differencesin
diseaseseverity
scoreaccording
totoxigenic
cultu
retesting
andGDH/EIA
testing(single
center)
n=
334fecalsamples
from
patientswith
diarrhea
Severe
CDIdefin
edas
pseudomem
branous
colitison
endoscopy,
admission
toIC
Uor
any
twoof
age[
60years,
temperature[
38.3�C
,serum
albumin
\2.5g/dL
,white
cell
coun
t[15,000
cells/
mm
3
Simultaneousdetectionof
GHDandtoxins
A/B
byC
DIFFQUIK
CHEK
COMPL
ETE�
252GDH-negative/EIA
toxin-negative
(i.e.n
oCDI)
82GDH-positive,of
which
25wereEIA-positive
(CDI)and57
EIA
toxin-
negative
(equivocalcases)
Whentoxins
weredetected
intheinitialscreeningtest
(GDH-positive/EIA
toxin-
positive),casesweremore
severe
than
inthoseonly
identifiedaftertoxigenic
cultu
re
–
58 Infect Dis Ther (2018) 7:39–70
COMPLETE�; stool culture, PCR ribotyping andslpA sequence typing; stool culture, PCR andPFGE; and rRNA-targeted RT-qPCR and multi-plex PCR for toxin gene profiling. Reportedmethods of testing for C. difficile toxins inclu-ded: toxin A testing by Uniquick (in 2004) [43];PCR assessment for toxin A, B and CDT genes[44]; TOX A/B QUIK CHEK� [45]; and multiplexPCR [46]. Rapid detection methods evaluatedand reported in Japan include detection of GDHand toxin A/B in feces (using Immunocard� andTOX A/B, respectively) [47]; and detection ofGDH and toxin A/B simultaneously using CDIFF QUIK CHEK COMPLETE� [41]. Althoughmass spectrometry methods have also beenreported, these were not considered suitable fortyping C. difficile [48].
Our literature review (Table 2) suggests thatmost C. difficile isolates associated with infectionin Japan produce both toxins A and B. Availablestudies suggest a low prevalence of binary toxin-positive (CDT?) strains—between 0 and 6.8%reported across studies [19, 44, 46, 49–51]. A smallstudy from 2015 found that outbreak and non-outbreak isolates were predominantly smz/ysmz(by slpA typing) and that of five binary toxin-positive strains, one was ribotype 027 and one 078[51]. Individual papers appear to support theconclusion that ribotypes 027 and 078 are rare inJapan [44, 49, 50], and that smz/018, yok, 002,014and 369 (trf by slpA typing) are common[3, 42–44, 46, 49–51]. A recently published sub-study of a global RCT that isolated 54 strains oftoxigenic C. difficile (by EIA/PCR assays) from thestool cultures of93hospitalizedpatientsat35 sitesin Japan appeared to corroborate this finding [52].The most common PCR ribotypes were 052 (28%of isolates), 018 (19% of isolates), 002 (15% ofisolates) and 369 (9% of isolates), and 052 wasconsidered to be an ‘established’ strain, as it waswidely distributed across Japan. Ribotypes 027and 078 were not isolated in the substudy [52].
Recurrence
Twelve publications reported on CDI recurrence(Table 3) [18, 21, 24–26, 31, 39–42, 44], withrates in studies that included specific definitionsof recurrence ranging from 3.3% in 30 HSCT
Table2
continued
Reference
Stud
yperiod
Stud
ydesign
Patient
popu
lation
C.diffdefin
ition
Testmetho
dsIsolates
andstrains
Binarytoxin
Yuhashi
etal.
[68]
Retrospective
assessmentof
casestested
for
C.d
iffdiarrhea
(singlecenter)
n=
68
Astestmethods
EIA
testingforC.d
iffdiarrhea
(CDIFFQUIK
CHEK
COMPL
ETE�)
Patientsgroupedas
toxin-
positive
stool;toxin-
negative/toxin-positive
isolate;dualtoxinnegative
(stool
andisolate)
39toxin-positive;14
toxin-
positive
isolategroup;anda
dualtoxin-negative
stool
andisolategroupn=
15.
Allcasesconfi
rmed
tobe
GDH
positive
byEIA
Toxin-negativestool
specim
ensassociated
with
shorterdiarrhea
duration
–
A?B?CDT?
toxinA-positivetoxinB-positive,binary
toxinpositive
strain;A
?B?CDT-
toxinA-positive,toxinB-positive,binary
toxinnegative
strain;A
-B?CDT-
toxinA-negative,toxinB-positive,
binarytoxinnegative
strain;E
IAenzymeim
mun
oassays;GDH
glutam
atedehydrogenase;ICUintensivecareun
it;M
LST
multilocussequencetyping;PF
GEpulsed-fieldgelelectrophoresis;R
CTrand
omized
clinicaltrial;slpAsurface-layerproteinAencoding
gene;SO
Cstandard
ofcare
Infect Dis Ther (2018) 7:39–70 59
Table3
Incidenceof,and
risk
factorsassociated
with,
C.d
ifficileinfectionrecurrence
inJapanese
patientpopulations
Reference
Patient
popu
lation
Treatmentof
initialCDI
Definition
ofrecurrence
Recurrence
Recurrencerisk
factors
Akahoshi
etal.[24]
HSC
T(n
=102
autologous;n=
206
allogeneic);n=
30
withCDI
Oralmetronidazole(500
mg
three-times
daily,1
0–14
days)
New
episodeof
diarrhea
andpositive
toxinEIA
within365days
afterfirst
episodeof
CDI
1/30
(3.3%)within
100days
afterHSC
T
NR
Daida
etal.
[31]
Pediatricpatients(aged
0–19
years)admitted
tohospitalwithcancer
Oralmetronidazole(30mg/kg)
for[
10days
untilresolution
ofsymptom
sandneutrophil
recovery
to[
500/lL
Presence
ofCDI2weeks
afterresolution
of
prim
aryCDIsymptom
s
13/51(26%
)Statisticaltestsnot
performed.R
ecurrence
morecommon
in
youn
gerage(0–3
years;
9/13,6
9%)than
older
children(19/38,5
0%)
Hashimoto
etal.[26]
Retrospective
chartreview
(single
center)of
242living
donorliver
transplant
recipients(adults)
n=
11withC.d
iffdiarrhea
Oralvancom
ycin
(n=
8;dose
notgiven)
orconservative
managem
ent(nodetailgiven)
Nodefin
itiongiven,
but
patientsassessed
from
hospitaladmission
to
3monthsaftertransplant
3/11
(27.3%
);2/8(25%
)in
patientswho
received
vancom
ycin
NR
Hikone
etal.[20]
In-andoutpatient
samples
tested
forC.
diff(n
=2193
samples)
n=
76withhealthcare-
associated
CDI
Oralmetronidazoleor
vancom
ycin
(doses
notgiven)
formedian14
days
(range
6–52
days)
New
episodeof
CDIwithin
8weeks
from
theprevious
episode;diagnosisbased
onpresence
ofdiarrhea
andpositive
toxinEIA
14/76(18.4%
)Univariateanalysis:no
risk
factorsidentified
Multivariateanalysis:
malignant
disease(O
R
7.98;95%
CI1.22–5
2.2;
p=
0.03)andIC
U
hospitalization(O
R
49.9;95%
CI
1.01–2
470;
p=
0.049)
60 Infect Dis Ther (2018) 7:39–70
Table3
continued
Reference
Patient
popu
lation
Treatmentof
initialCDI
Definition
ofrecurrence
Recurrence
Recurrencerisk
factors
Honda
etal.[18]
CDIcasesin
anon-
outbreak
setting
n=
126withCDI
(86.5%
were
healthcare-facility
onsetCDI)
Oralmetronidazole(500
mg
three-times
daily),oral
vancom
ycin
(125
mgor
500mgfour-tim
esdaily),
combination
oral
metronidazole(500
mgthree-
times
daily)plus
vancom
ycin
(125
mgor
500mgfour-tim
es
daily),combination
oral
metronidazoleplus
rectal
vancom
ycin,com
bination
oral
andrectalvancom
ycin,o
rno
treatm
ent(stopun
necessary
antimicrobials)
New
episodeof
diarrhea
andpositive
toxinassay
within30
days
sincelast
date
ofcompleting
therapyforfirstCDI
episode
8/126(6%)
NR
Hosokaw
a
etal.[25]
AllogeneicHSC
T
patients(135
unrelatedCBT;39
unrelatedBMT
and
27relatedPB
SCT)
n=
17withC.d
iffdiarrhea
Oralmetronidazoleor
oral
vancom
ycin
(dosageand
duration
notgiven)
New
episodeof
diarrhea
andpositive
toxintest
within8weeks
after
improvem
entof
first
properlytreatedepisode
0NR
Iwashima
etal.[44]
CDIcasesin
ahospital
setting
n=
71consecutive
patientswithCDI
Noform
alregimensspecified;
3patientsreceived
treatm
ent
priorto
recurrentCDI.
Vancomycin
mentioned
(total
dose
range6–
12ggivenfor
rangeof
1–28
days)
New
CDIepisodewithin
2monthsafterrecovery
from
previous
CDI
episode
9/71
(12.7%
)NR
Infect Dis Ther (2018) 7:39–70 61
Table3
continued
Reference
Patient
popu
lation
Treatmentof
initialCDI
Definition
ofrecurrence
Recurrence
Recurrencerisk
factors
Kobayashi
etal.[21]
Patientsaged
C14
years
withhospital-onset
CDI,n=
160
Metronidazole(n
=88,5
5%),
vancom
ycin
(n=
52,3
3%),
metronidazole?
vancom
ycin
(n=
10,6
.3%),no
antimicrobial(n
=10,6
.3%).
Dosesandduration
oftherapy
notspecified
According
toSH
EA/IDSA
2010
guidelines[69]
–re-
emergenceof
CDI
symptom
s,accordingto
infectious
disease
physician
judgment,B4weeks
aftercompletionof
treatm
entforinitialCDI
episode
23/160
(14%
)Recurrencedidnotdiffer
accordingto
severe/
non-severe
CDIor
accordingto
whether
treatm
entadhered/did
notadhere
toclinical
guidelines
Moriet
al.
[42]
Stoolcultu
redatabase
n=
58caseswithCDI
Vancomycin
ormetronidazole
(dosageandduration
not
given)
Nodefin
itiongiven,
but
recurrence
recorded
within60
days
following
symptom
onset
3/58
(5.2%)
NR
Oshim
a
etal.[40]
Adults
andpediatric
(B18
years)patients
receivingPP
Iwho
developedacute-onset
diarrhea
(n=
17,217).
Also
,control
group
(n=
286,018)
Recurrent
CDI
(reportedin
9studies)
occurred
inn=
1279;
n=
5459
inthe
controlgroup
–Not
givenin
thissystem
atic
review
andmeta-analysis
ofpublishedstudies,but
basedon
recurrence,as
reported
inpublished
studies.CDIpresence
basedon
laboratory
confi
rmationof
C.d
iffor
clinicaldefin
ition
–PP
Iuseincreasedrisk
for
recurrentCDI(pooled
OR1.73,9
5%CI
1.39–2
.15;
p=
0.02)
62 Infect Dis Ther (2018) 7:39–70
Table3
continued
Reference
Patient
popu
lation
Treatmentof
initialCDI
Definition
ofrecurrence
Recurrence
Recurrencerisk
factors
Shim
izu
etal.
2015
[41]
Patientsin
ahospital
settingwithdiarrhea
(n=
334fecal
samples)
n=
28patientswith
severe
CDI
Metronidazole,vancomycin
(dosageandduration
not
given)
orno
treatm
ent
Nodefin
itiongivenand
duration
ofassessment
notdetailed
Overall:
7/28
(25%
);5/16
(31.3%
)in
patientswith
GDH-positive/EIA
toxin-positive
testand
2/12
(16.7%
)in
patients
withinitialGDH-
positive/EIA
toxin-
negative
test,b
utwho
hadconfi
rmed
positive
toxigeniccultu
re
Nodifference
in
incidenceof
recurrence
betweenthetwogroups
(p=
0.662)
Takahashi
etal.[39]
NationalHospital
Organizationcohort
Assessedfornewly
diagnosedCDI
(n=
878patients)and
matched
controls(no
CDI)
Metronidazole,vancomycin
(dosageandduration
not
given)
orno
treatm
ent
Nodefin
itiongiven,
but
recurrence
assessed
and
recorded
within30
days
ofinitialCDIepisode
34/714
(4.8%)am
ong
patientstreatedforCDI
NR
BMT
bone
marrow
transplantation,
CBT
cord
bloodtransplantation,
CIconfi
denceinterval,EIA
enzymeim
mun
oassays,
HSC
Thematopoietic
stem
cell
transplantation,
ICUintensivecare
unit,N
Rnotrecorded,O
Rodds
ratio,PB
SCTperipheralbloodstem
celltransplant,P
PIproton
pumpinhibitor,SH
EA/IDSA
SocietyforHealth
care
Epidemiology
ofAmericaandInfectious
Disease
Societyof
America
Infect Dis Ther (2018) 7:39–70 63
patients [24] to 27.3% in a cohort of 11 livertransplant recipients with CDI [26]. The timeperiod over which recurrence was defined orassessed varied widely from 14 to 365 days afterthe initial CDI episode [18, 21, 24, 25, 31, 44].One retrospective chart review of 242 livertransplant recipients, 11 of whom developedCDI, reported recurrence in 2 of 8 patients whoreceived vancomycin [26]; no analysis of riskwas made in relation to the choice of treatmentfor CDI. A retrospective cohort study based onchart reviews of hospital-onset CDI cases at fourteaching hospitals found that neither theseverity of CDI nor adherence to clinical prac-tice guidelines affected the risk of CDI recur-rence [21]. In a retrospective study, multivariateanalysis identified malignant disease (p = 0.03)and ICU hospitalization (p = 0.049) as risk fac-tors for CDI recurrence within 8 weeks of theprevious CDI episode [20]. A systematic reviewand meta-analysis of published studies reportedthat use of PPIs was significantly associated withrecurrent CDI (pooled OR 1.73, 95% CI1.39–2.15; p = 0.02) (Table 3) [40].
Healthcare Utilization
A large retrospective chart review of 22,863patients by Honda et al. reported a median LOSamong 126 CDI cases of 41.5 days (17.5 daysbefore and 18 days post-CDI diagnosis) [18]. Thesparsity of available LOS data did not allow anyanalysis of trends for CDI-related LOS over theyears examined. The adjusted attributable LOSand costs related to CDI were reported from onelarge database study of 143,652 hospitalizedpatients [29]. There were 409 cases of CDI withthe infection contributing to a LOS increase of12.4 days (95% CI 9.7–15.0; p\0.001) and a costincrease of US$6576 (95% CI 3753–9398;p\0.001) compared with control patients whodid not develop CDI. However, of note, thisstudy included patients who had undergonespecific surgical procedures and had CDI identi-fied using diagnostic codes rather than by diag-nostic tests [29], which could potentially lead toinaccurate identification of CDI.
Cases of CDI may require patient transfer tothe ICU. The study by Honda et al. found that
9.5% of 126 CDI cases needed ICU admission[18]. There were 65 (51.6%) patients who wereclassified as having severe CDI, using theseverity assessment score developed by Zar et al.[53]; three patients (2.4%) underwent CDI-re-lated colectomy/diverting loop ileostomybecause of critical illness or failure of medicaltherapy. Another retrospective chart reviewbased on stool samples from in- and outpatientsat a single hospital identified 76 patients withhospital-onset CDI, with three (3.9%) casesrequiring ICU care [20].
Mortality
All-cause mortality within 30 days ranged from3.4 to 15.1% [18, 20, 21, 29, 39]. Honda et al.reported that mortality was associated with anincreased Zar severity score [18]. Of note, theZar severity criteria will score a patient 1 point(2 points are defined as severe CDI) based onlyon age 60 years or older [53], and the medianage of patients in the Honda et al. study was78 years [18]. A multicenter retrospective cohortstudy of 160 hospitalized patients with CDIreported 30-day all-cause mortality of 13%, andfound no significant difference between themortality rate among patients with severe andnon-severe CDI, or among those whose treat-ment adhered and did not adhere to clinicalpractice guidelines [21]. At 90 days, all-causemortality among CDI cases was reported as14.5% in one retrospective chart review [20].Hosokawa et al. [25] concluded that, amongpatients who underwent unrelated cord bloodtransplantation, overall survival at 2 years wasno different for those who developed CDI thanfor those who did not (42 vs. 46%, respectively;p = 0.77). One database study demonstratedthat inpatient mortality was significantly higherin CDI patients than in those without CDI (3.4vs. 1.6%; p = 0.008) [29], although the results ofthis study should be viewed with caution owingto the reliance on recorded diagnoses of CDIfrom administrative databases (which are lesswell validated than those in prospective cohortsor registries, for example), the inclusion of onlypatients undergoing GI surgery, and the loss ofmany patients from the propensity score-
64 Infect Dis Ther (2018) 7:39–70
matched analysis. The use of vancomycin wasassociated with reduced mortality (OR 0.43;95% CI 0.25–0.75) in a multicenter, case–con-trol and cohort study of 1026 CDI patients [39].
DISCUSSION
This systematic literature review identified 55papers providing insights into the rates of CDI,patient groups affected and impact of CDI inJapan. Most of the studies were retrospectivedata reviews, and many focused on patientswith suspected CDI, and the rates and preva-lence of CDI in those groups. Fewer studiesreported overall rates of hospital- and commu-nity-acquired CDI. Nevertheless, the currentliterature suggests that hospital-onset CDI inJapan occurs at an incidence of 0.8–4.7 cases/10,000 patient-days; lower than that reported inEurope for winter 2012–2013 (country range0.7–28.7/10,000 patient bed-days), and similarto the US for hospital-onset CDI in 2013 (6.0/10,000 patient-days for laboratory-identifiedCDI and 4.4/10,000 patient-days for traditionalsurveillance-detected CDI) [54, 55]. However,direct comparisons between studies are difficultowing to differences in design, population sizeand detection methods used. The prevalence ofCDI in Japan (0.3–5.5/1000 patient admissions)was lower than that in a US Veterans HealthAdministration (VHA) report from 2014, whichrecorded that from October 2010 to June 2012,CDI prevalence at admission was between 5.3and 6.9/1000 admissions, in settings where 50%of hospitals used nucleic acid amplification tests[56]. An analysis of US VHA data for February2012 to January 2014 reported a pooled CDIadmission prevalence rate of 0.38/100 admis-sions [57].
CDI in Japan appears to have similar epi-demiology to that in South Korea, althoughdistinct from regions of South Asia [3]. There area number of factors that contribute to theunique epidemiology of CDI in Japan. First, iftesting for CDI is not conducted, this makesdata-gathering on both the rate of testing andtest results difficult. While there is a nationalsurveillance program for infectious diseases inJapan, there is no national C. difficile screening
program. Methods of testing are also animportant consideration. Indeed, datasets fromCDI surveillance programs highlight thatcountry-to-country variations in CDI incidencelargely reflect differences in surveillance meth-ods and how rigorously CDI is investigated intesting strategies [54]. This literature reviewhighlights that testing, typing and laboratorymethods for assessment and diagnosis of CDIhave evolved in Japan in recent years, with newmethods continuing to be evaluated againstestablished methods. However, some authorshave noted that many reports of CDI epidemi-ology in Japan have relied on insensitive testingmethods, such as EIA toxin tests [16]. Whilemore sensitive nucleic acid amplification testsare recommended in clinical practice guidelinesfor the detection of C. difficile [7], these are notyet subject to reimbursement in Japan and aretherefore not widely used, thus reliance on EIAremains common in clinical practice. It couldbe argued that a low detection frequency, asseen for CDI toxin tests in stool samples, may infact reflect a low disease burden, as demon-strated by Shimizu et al., who reported thatpatients with positive EIA toxin tests in stoolshad more severe CDI than those patients withnegative stool toxin tests who then had positivecultures [41]. It is recognized that CDI is under-diagnosed in many regions and countriesbecause of a combination of absence of clinicalsuspicion and suboptimal laboratory diagnos-tics [54]. This may also be the case in Japan,although no papers in this search specificallyfocused on reporting under-diagnosed or mis-sed CDI cases.
Increased LOS is both a risk factor for, and anoutcome of, CDI [29]. A notable factor in Japanis the general tendency for a longer LOS com-pared with many other countries, which mayhave an impact on, and thus affect, CDI risk andrates [58]. Incidence of CDI is typically pre-sented as per patient-days (i.e. patient-bed daysor inpatient days) and it is important tounderstand the differences in overall LOS indifferent geographic regions when comparingincidence data. National statistics from Japanshow that in 2014 mean LOS was 31.9 days forall patients and 41.7 days in those C 65 years[59], whereas the comparable 2014 statistics
Infect Dis Ther (2018) 7:39–70 65
from the USA were 4.6 days for all patients (notrestricted to those with CDI), 5.2 days forpatients aged 65–74 years and 5.3 days for thoseaged 75 years and over [60]. This suggests that iffollow up was similar between nations, roughlyeight times more patients would be included inthe US study compared with a Japanese study toyield 10,000 patient-days as the denominator.Given the lengthy LOS in Japan, as in studiesfrom other countries, rates of, and risks for, CDIin Japanese patients are increased by factorssuch as older age, presence of comorbidities,certain clinical/surgical procedures and expo-sure to drug therapies, including antibiotics [1].We found that CDI recurrence similarlydepended on a variety of factors, includingclinical circumstances and settings, rangingfrom 3.3% in HSCT patients [24] to 27.3% inliver transplant patients [26]. We found veryfew data on risk factors for CDI recurrence, suchas the choice of treatment for initial CDI. Onestudy identified malignant disease and ICUhospitalization as risk factors for CDI recurrence[20], although the broad CIs cited for both(Table 3) suggest that analysis of risk factors in alarger population is required. As previouslyestablished [1], one systematic review and meta-analysis identified in our review revealed astrong association between PPI use and the riskfor initial and recurrent CDI in both adults andchildren [40].
Most CDI cases in Japan are caused by strainsof C. difficile that produce both toxins A and B.Recent reviews of CDI in Asia, which are sup-ported by our literature review, showed thatribotype smz/18 (or 018) predominates in Japan[3] and that ribotype trf/369 is gaining promi-nence in the country [61]. Several of the studieswe reviewed confirmed a low prevalence ofribotypes 027 and 078 in Japan. It has beenpostulated that the low prevalence of thesehighly virulent strains that cause sporadic out-breaks in Western countries may account forthe low general incidence of CDI in Japan [21].Further, we found that recent papers reportingon CDI strain and ribotype increasingly inclu-ded a binary toxin assessment, although CDT?C. difficile was rare in Japan [42, 46].
There are a number of limitations to thisreview. Most reports on CDI epidemiology in
Japan included here were based on hospitalcohort data, and were often (in 32 of 55 studies)single-site reports concerning either inpatientsor patients discharged from hospital to thecommunity. We did not find papers reportingon CDI rates in long-term care facilities,although, given the prolonged LOS reported instudies in our review, some hospitals may ‘re-place’ long-term care facilities in some instan-ces. There was a large degree of heterogeneity inrelation to the definition of initial and recurrentCDI in the studies included in our review. Thishampered our ability to evaluate the epidemi-ological data in a consistent manner andreduced the potential to draw robust conclu-sions. Although the papers reviewed spannedmore than a 10-year period, there were no datadescribing local infection changes over time.However, we did find sporadic snapshots of CDIin certain patient subgroups and cohorts. Ourreview was further limited by a lack of infor-mation relating to the rate of CDI testing andno data on community-associated CDI.
Future studies and surveillance are necessaryto gather data on the numbers and types ofpatients affected by, or at risk of, CDI in Japan.Furthermore, data on the clinical impact of CDIare important for management and resourceplanning, and to ensure optimal patient careand outcomes.
CONCLUSION
The current literature offers some insights intothe evolving epidemiology of CDI in Japan, yethighlights a number of unresolved questions.Notably, heterogeneity in the CDI definitions inthe studies we reviewed limited our ability todraw robust conclusions. With the availabilityof newer diagnostic tools and release of clinicalpractice guidelines for CDI, there is a need toundertake more comprehensive and coordi-nated studies and surveillance of both CDI casesand C. difficile isolates to map current trends,review the impact of infection control mea-sures, increase knowledge of risk factors, andmore fully understand the extent and impact ofhospital-onset CDI in Japanese patient popula-tions today.
66 Infect Dis Ther (2018) 7:39–70
ACKNOWLEDGEMENTS
Funding. This analysis was initiated byAstellas Pharma, Inc. The article processingcharges were funded by Astellas Pharma, Inc.
Editorial Assistance. Editorial support tothe authors was provided by Rhian HarperOwen, Winnie McFadzean and Iona Easthopefor Cello Health MedErgy (Europe), funded byAstellas Pharma, Inc.
Authorship. All named authors meet theInternational Committee of Medical JournalEditors (ICMJE) criteria for authorship for thismanuscript, take responsibility for the integrityof the work as a whole, and have given finalapproval for the version to be published.Fidaxomicin, an Astellas product, is available insome geographical regions for the treatment ofCDI. All authors had full access to all of the datain this study and take complete responsibilityfor the integrity of the data and accuracy of thedata analysis.
Disclosures. Thomas V. Riley has receivedgrants from Cepheid, Merck, Sanofi and Otsuka.Tomomi Kimura is a full-time employee ofAstellas Pharma, Inc. (Japan).
Compliance with Ethics Guidelines. Thisarticle is based on previously conducted studiesand does not involve any new studies of humanor animal subjects performed by either of theauthors.
Data Availability. The results of the litera-ture searches analysed during the current revieware available from the corresponding author onreasonable request.
Open Access. This article is distributedunder the terms of the Creative CommonsAttribution-NonCommercial 4.0 InternationalLicense (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommer-cial use, distribution, and reproduction in anymedium, provided you give appropriate creditto the original author(s) and the source, provide
a link to the Creative Commons license, andindicate if changes were made.
REFERENCES
1. Chakra CNA, Pepin J, Sirard S, Valiquette L. Riskfactors for recurrence, complications and mortalityin Clostridium difficile infection: a systematic review.PLoS ONE. 2014;9:e107420.
2. Rupnik M, Wilcox MH, Gerding DN. Clostridiumdifficile infection: new developments in epidemiol-ogy and pathogenesis. Nat Rev Microbiol.2009;7:526–36.
3. Collins DA, Hawkey PM, Riley TV. Epidemiology ofClostridium difficile infection in Asia. AntimicrobResist Infect Control [Internet]. 2013;2:21. http://aricjournal.biomedcentral.com/articles/10.1186/2047-2994-2-21.
4. Yamagishi Y, Mikamo H. Recent epidemiology ofClostridium difficile infection in Japan. Jpn J Anti-biot. 2015;68:345–58.
5. Shionogi & Co. Antitrichomonal agent, Flagyl�—approval for additional indication for amoebicdysentery, giardiasis, anaerobic infection andinfectious enteritis [Internet]. 2012 [cited 2017 Jun7]. http://www.shionogi.co.jp/en/company/news/2012/pmrltj0000000wim-att/e_120810-2.pdf.
6. Pfizer Inc. Acquired manufacturing and marketingapproval for anaerobic bacterial infection treatmentagent ‘‘Anemetro� IV infusion solution 500 mg’’[Internet]. 2014 [cited 2017 Jun 7]. http://www.pfizer.co.jp/pfizer/company/press/2014/2014_07_04.html.
7. Onishi K, Somoda Y, Imamura A, Iwabuchi C,Okuda MNT. JAID/JSC infection treatment guide-line 2015. Intestinal infection. Jpn J Chemother.2016;64:31–65.
8. CEBM. Oxford Centre for Evidence-based Medi-cine—Levels of Evidence (March 2009) [Internet].Cent. Evidence-Based Med. 2009 [cited 2017 Feb23]. http://www.cebm.net/oxford-centre-evidence-based-medicine-levels-evidence-march-2009/.
9. Cairns MD, Preston MD, Hall CL, Gerding DN,Hawkey PM, Kato H, et al. Correction for Cairnset al., ‘‘Comparative genome analysis and globalphylogeny of the toxin variant Clostridium difficilePCR ribotype 017 reveals the evolution of twoindependent sublineages’’. J Clin Microbiol.2017;55:1971.
Infect Dis Ther (2018) 7:39–70 67
10. Cairns MD, Preston MD, Hall CL, Gerding DN,Hawkey PM, Kato H, et al. Comparative genomeanalysis and global phylogeny of the toxin variantClostridium difficile PCR ribotype 017 reveals theevolution of two independent sublineages. J ClinMicrobiol. 2017;55:865–76.
11. Nomura K, Fujimoto Y, Yamashita M, Morimoto Y,Ohshiro M, Sato K, et al. Absence of pseudomem-branes in Clostridium difficile-associated diarrhea inpatients using immunosuppression agents. Scand JGastroenterol. 2009;44:74–8.
12. Sasahara T, Ae R, Watanabe M, Kimura Y, YonekawaC, Hayashi S, et al. Contamination of healthcareworkers’ hands with bacterial spores. J Infect Che-mother. 2016;22:521–5.
13. Fujimori S. What are the effects of proton pumpinhibitors on the small intestine? World J Gas-troenterol. 2015;21:6817–9.
14. Senoh M, Kato H, Murase T, Hagiya H, Tagashira Y,Fukuda T, et al. Reverse transcription polymerasechain reaction-based method for selectivelydetecting vegetative cells of toxigenic Clostridiumdifficile. Microbiol Immunol. 2014;58:615–20.
15. Tojo M, Nagamatsu M, Hayakawa K, Mezaki K,Kirikae T, Ohmagari N. Evaluation of an automatedrapid diagnostic test for detection of Clostridiumdifficile. PLoS ONE. 2014;9:e106102.
16. Honda H, Dubberke ER. Clostridium difficile infec-tion in solid organ transplant recipients. Curr OpinInfect Dis. 2014;27:336–41.
17. Arimoto J, Horita N, Kato S. Diagnostic test accu-racy of glutamate dehydrogenase for Clostridiumdifficile: systematic review and meta-analysis. SciRep. 2016;6:29754.
18. Honda H, Yamazaki A, Sato Y, Dubberke ER. Inci-dence and mortality associated with Clostridiumdifficile infection at a Japanese tertiary care center.Anaerobe. 2014;25:5–10.
19. Mori N, Aoki Y. Clinical characteristics and riskfactors for community-acquired Clostridium difficileinfection: a retrospective, case-control study in atertiary care hospital in Japan. J Infect Chemother.2015;21:864–7.
20. Hikone M, Ainoda Y, Tago S, Fujita T, Hirai Y,Takeuchi K, et al. Risk factors for recurrent hospital-acquired Clostridium difficile infection in a Japaneseuniversity hospital. Clin Exp Gastroenterol.2015;8:191–6.
21. Kobayashi K, Sekiya N, Ainoda Y, Kurai H, ImamuraA. Adherence to clinical practice guidelines for themanagement of Clostridium difficile infection in
Japan: a multicenter retrospective study. Eur J ClinMicrobiol Infect Dis Germany. 2017;36:1947–53.
22. Komatsu S, Sakamoto E, Norimizu S, Shingu Y,Asahara T, Nomoto K, et al. Efficacy of perioperativesynbiotics treatment for the prevention of surgicalsite infection after laparoscopic colorectal surgery: arandomized controlled trial. Surg Today.2016;46:479–90.
23. Suzuki H, Senda J, Yamashita K, Tokuda Y, Kane-saka Y, Kotaki N, et al. Impact of intensive infectioncontrol team activities on the acquisition ofmethicillin-resistant Staphylococcus aureus, drug-re-sistant Pseudomonas aeruginosa and the incidence ofClostridium difficile-associated disease. J Infect Che-mother. 2013;19:1047–52.
24. Akahoshi Y, Kimura S, Nakano H, Harada N,Kameda K, Ugai T, et al. Significance of a positiveClostridium difficile toxin test after hematopoieticstem cell transplantation. Clin Transplant.2016;30:703–8.
25. Hosokawa K, Takami A, Tsuji M, Araoka H. Relativeincidences and outcomes of Clostridium difficileinfection following transplantation of unrelatedcord blood, unrelated bone marrow, and relatedperipheral blood in adult patients: a single institutestudy. Transpl Infect Dis. 2014;16:412–20.
26. Hashimoto M, Sugawara Y, Tamura S, Kaneko J,Matsui Y, Togashi J, et al. Clostridium difficile-asso-ciated diarrhea after living donor liver transplanta-tion. World J Gastroenterol. 2007;13:2072–6.
27. Iwamoto M, Kamimura T, Nagashima T, Kamata Y,Aoki Y, Onishi S, et al. Healthcare-associatedinfections in rheumatology in Japan. RheumatolInt. 2012;32:801–4.
28. Sasabuchi Y, Matsui H, Lefor AK, Fushimi K, Yasu-naga H. Risks and benefits of stress ulcer prophy-laxis for patients with severe sepsis. Crit Care Med.2016;44:e464–9.
29. Yasunaga H, Horiguchi H, Hashimoto H, Matsuda S,Fushimi K. The burden of Clostridium difficile-asso-ciated disease following digestive tract surgery inJapan. J Hosp Infect. 2012;82:175–80.
30. Kaneko T, Matsuda R, Taguri M, Inamori M, OguraA, Miyajima E, et al. Clostridium difficile infection inpatients with ulcerative colitis: investigations ofrisk factors and efficacy of antibiotics for steroidrefractory patients. Clin Res Hepatol Gastroenterol.2011;35:315–20.
31. Daida A, Yoshihara H, Inai I, Hasegawa D, Ishida Y,Urayama KY, et al. Risk factors for hospital-acquiredClostridium difficile infection among pediatric
68 Infect Dis Ther (2018) 7:39–70
patients with cancer. J Pediatr Hematol Oncol.2017;39:e167–72.
32. Furuichi M, Imajo E, Sato Y, Tanno S, Kawada M,Sato S. Characteristics of Clostridium difficile colo-nization in Japanese children. J Infect Chemother.2014;20:307–11.
33. Schutze GE, Willoughby RE. Clostridium difficileinfection in infants and children. Pediatrics.2013;131:196–200.
34. Matsumoto K, Kanazawa N, Shigemi A. Factorsaffecting treatment and recurrence of Clostridiumdifficile infections. Biol Pharm Bull.2014;37:1811–5.
35. Janarthanan S, Ditah I, Adler DG, Ehrinpreis MN.Clostridium difficile-associated diarrhea and protonpump inhibitor therapy: a meta-analysis. Am JGastroenterol. 2012;107:1001–10.
36. Roughead EE, Chan EW, Choi N-K, Griffiths J, JinX-M, Lee J, et al. Proton pump inhibitors and risk ofClostridium difficile infection: a multi-country studyusing sequence symmetry analysis. Expert OpinDrug Saf Engl. 2016;15:1589–95.
37. Mizui T, Teramachi H, Tachi T, Tamura K, Shiga H,Komada N, et al. Risk factors for Clostridium difficile-associated diarrhea and the effectiveness of prophy-lactic probiotic therapy. Pharmazie. 2013;68:706–10.
38. Ogami N, Yoshida J, Ishimaru T, Kikuchi T, Mat-subara N, Ueno T, et al. Is Clostridium difficileinfection influenced by antimicrobial use density inwards? Jpn J Antibiot. 2013;66:87–95.
39. Takahashi M, Mori N, Bito S. Multi-institution case-control and cohort study of risk factors for thedevelopment and mortality of Clostridium difficileinfections in Japan. BMJ Open. 2014;4:e005665.
40. Oshima T, Wu L, Li M, Fukui H, Watari J, Miwa H.Magnitude and direction of the associationbetween Clostridium difficile infection and protonpump inhibitors in adults and pediatric patients: asystematic review and meta-analysis. J Gastroen-terol Jpn. 2018;53:84–94.
41. Shimizu H, Mori M, Yoshimoto N. Clostridium dif-ficile infection is more severe when toxin is detectedin the stool than when detected only by a toxigenicculture. Intern Med. 2015;54:2155–9.
42. Mori N, Yoshizawa S, Saga T, Ishii Y, Murakami H,Iwata M, et al. Incorrect diagnosis of Clostridiumdifficile infection in a university hospital in Japan.J Infect Chemother. 2015;21:718–22.
43. Kato H, Kato H, Nakamura M, Iwashima Y, Naka-mura A, Ueda R. Rapid analysis of Clostridium
difficile strains recovered from hospitalized patientsby using the slpA sequence typing system. J InfectChemother. 2009;15:199–202.
44. Iwashima Y, Nakamura A, Kato HH, Kato HH,Wakimoto Y, Wakiyama N, et al. A retrospectivestudy of the epidemiology of Clostridium difficileinfection at a University Hospital in Japan: geno-typic features of the isolates and clinical character-istics of the patients. J Infect Chemother.2010;16:329–33.
45. Kobayashi M, Sako A, Ogami T, Nishimura S,Asayama N, Yada T, et al. Validation of the 3-dayrule for stool bacterial tests in Japan. Intern Med.2014;53:533–9.
46. Kuwata Y, Tanimoto S, Sawabe E. Molecular epi-demiology and antimicrobial susceptibility ofClostridium difficile isolated from a universityteaching hospital in Japan. Eur J Clin MicrobiolInfect Dis. 2015;34:763–72.
47. Kawada M, Annaka M, Kato H. Evaluation of asimultaneous detection kit for the glutamatedehydrogenase antigen and toxin A/B in feces fordiagnosis of Clostridium difficile infection. J InfectChemother. 2011;17:807–11.
48. Kiyosuke M, Kibe Y, Oho M. Comparison of twotypes of matrix-assisted laser desorption/ionizationtime-of-flight mass spectrometer for the identifica-tion and typing of Clostridium difficile. J MedMicrobiol. 2015;64:1144–50.
49. Kato H, Kato H, Ito Y, Akahane T, Izumida S,Yokoyama T, et al. Typing of Clostridium difficileisolates endemic in Japan by sequencing of slpAand its application to direct typing. J Med Micro-biol. 2010;59:556–62.
50. Sawabe E, Kato H, Osawa K, Chida T, Tojo N, Ara-kawa Y, et al. Molecular analysis of Clostridium dif-ficile at a university teaching hospital in Japan: ashift in the predominant type over a five-year per-iod. Eur J Clin Microbiol Infect Dis.2007;26:695–703.
51. Senoh M, Kato H, Fukuda T, Niikawa A, Hori Y,Hagiya H, et al. Predominance of PCR-ribotypes,018 (smz) and 369 (trf) of Clostridium difficile inJapan: a potential relationship with other globalcirculating strains? J Med Microbiol.2015;64:1226–36.
52. Mikamo H, Aoyama N, Sawata M, Fujimoto G, DorrMB, Yoshinari T. The effect of bezlotoxumab forprevention of recurrent Clostridium difficile infec-tion (CDI) in Japanese patients. J Infect Chemother.2018;24:123–9.
Infect Dis Ther (2018) 7:39–70 69
53. Zar FA, Bakkanagari SR, Moorthi KMLST, Davis MB.A comparison of vancomycin and metronidazolefor the treatment of Clostridium difficile-associateddiarrhea, stratified by disease severity. Clin InfectDis [Internet]. 2007;45:302–7. http://www.ncbi.nlm.nih.gov/pubmed/17599306.
54. Davies KA, Longshaw CM, Davis GL, Bouza E, Bar-but F, Barna Z, et al. Underdiagnosis of Clostridiumdifficile across Europe: the European, multicentre,prospective, biannual, point-prevalence study ofClostridium difficile infection in hospitalisedpatients with diarrhoea (EUCLID). Lancet InfectDis. 2014;14:1208–19.
55. Durkin MJ, Baker AW, Dicks KV, Lewis SS, Chen LF,Anderson DJ, et al. A comparison between NationalHealthcare Safety Network laboratory-identifiedevent reporting versus traditional surveillance forClostridium difficile infection. Infect Control HospEpidemiol. 2015;36:125–31.
56. Evans ME, Simbartl LA, Kralovic SM, Jain R, RoselleGA. Clostridium difficile infections in VeteransHealth Administration acute care facilities. InfectControl Hosp Epidemiol. 2014;35:1037–42.
57. Reeves JS, Evans ME, Simbartl LA, Kralovic SM,Kelly AA, Jain R, et al. Clostridium difficile infectionsin veterans health administration long-term carefacilities. Infect Control Hosp Epidemiol. [Internet].2016 [cited 2017 Feb 23];37:295–300. http://www.ncbi.nlm.nih.gov/pubmed/26686361.
58. Organisation for Economic Co-operation andDevelopment. Length of hospital stay [Internet].2016 [cited 2017 Jun 26]. https://data.oecd.org/healthcare/length-of-hospital-stay.htm.
59. Ministry of Health Labour and Welfare. Summaryof patient survey 2014. [Internet]. 2015 [cited 2017Jul 24]. http://www.mhlw.go.jp/english/database/db-hss/dl/sps_2014_03.pdf.
60. Agency for Healthcare Research and Quality(AHRQ). HCUP fast stats—trends in inpatient stays2005-2014 [Internet]. 2014 [cited 2017 Jul 24].https://www.hcup-us.ahrq.gov/faststats/NationalTrendsServlet.
61. Collins DA. Epidemiology of Clostridium difficileinfection in the Asia-Pacific region [Internet]. TheUniversity of Western Australia; 2016. http://research-repository.uwa.edu.au/files/11692621/
THESIS_DOCTOR_OF_PHILOSOPHY_COLLINS_Deirdre_Anne_2016.pdf.
62. Hata H, Yamaguchi T, Hasegawa S, Nomura A, HidaK. Oral and parenteral versus parenteral antibioticprophylaxis in elective laparoscopic colorectal sur-gery (JMTO PREV 07-01): a phase 3, multicenter,open-label. Random Trial Ann Surg.2016;263:1085–91.
63. Sadahiro S, Suzuki T, Tanaka A, Okada K, Kamata H,Ozaki T, et al. Comparison between oral antibioticsand probiotics as bowel preparation for electivecolon cancer surgery to prevent infection:prospective randomized trial. Surgery.2014;155:493–503.
64. Watanabe M, Hitomi S, Sawahata T. Nosocomialdiarrhea caused by Clostridium perfringens in theTsukuba-Tsuchiura district. Jpn J Infect Chemother.2008;14:228–31.
65. Kikkawa H, Hitomi S, Watanabe M. Prevalence oftoxin A-nonproducing/toxin-B-producing Clostrid-ium difficile in the Tsukuba-Tsuchiura district. Jpn JInfect Chemother. 2007;13:35–8.
66. Kunishima H, Chiba J, Saito M, Honda Y, Kaku M.Antimicrobial susceptibilities of Clostridium difficileisolated in Japan. J Infect Chemother.2013;19:360–2.
67. Oka K, Osaki T, Hanawa T, Kurata S, Okazaki M,Manzoku T. Molecular and microbiological char-acterization of Clostridium difficile isolates fromsingle, relapse, and reinfection cases. J Clin Micro-biol. 2012;50:915–21.
68. Yuhashi K, Yagihara Y, Misawa Y, Sato T, Saito R,Okugawa S, et al. Diagnosing Clostridium difficile-associated diarrhea using enzyme immunoassay:the clinical significance of toxin negativity in glu-tamate dehydrogenase-positive patients. InfectDrug Resist. 2016;27:93–9.
69. Cohen SH, Gerding DN, Johnson S, Kelly CP, LooVG, McDonald LC, et al. Clinical practice guideli-nes for Clostridium difficile infection in adults: 2010update by the society for healthcare epidemiologyof America (SHEA) and the infectious diseasessociety of America (IDSA). Infect Control HospEpidemiol. 2010;31:431–55.
70 Infect Dis Ther (2018) 7:39–70