Clinical and Molecular Epidemiology of Healthcare‐Associated Infections Due to Extended‐Spectrum β‐Lactamase (ESBL)–Producing Strains ofEscherichia coliandKlebsiella pneumoniaeThat

Clinical and Molecular Epidemiology of Healthcare‐Associated Infections Due to Extended‐Spectrum β‐Lactamase (ESBL)–Producing Strains of Escherichia coli and Klebsiella pneumoniaeThat Harbor Multiple ESBL Genes • Author(s): Anucha Apisarnthanarak, MD; Pattarachai Kiratisin, MD, PhD; Linda M. Mundy, MDSource: Infection Control and Hospital Epidemiology, Vol. 29, No. 11 (November 2008), pp.1026-1034Published by: The University of Chicago Press on behalf of The Society for Healthcare Epidemiologyof AmericaStable URL: http://www.jstor.org/stable/10.1086/591864 .

Accessed: 15/05/2014 08:59

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .http://www.jstor.org/page/info/about/policies/terms.jsp

.JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range ofcontent in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new formsof scholarship. For more information about JSTOR, please contact [email protected].

.

The University of Chicago Press and The Society for Healthcare Epidemiology of America are collaboratingwith JSTOR to digitize, preserve and extend access to Infection Control and Hospital Epidemiology.

http://www.jstor.org

This content downloaded from 195.78.108.144 on Thu, 15 May 2014 08:59:59 AMAll use subject to JSTOR Terms and Conditions

http://www.jstor.org/action/showPublisher?publisherCode=ucpress

http://www.jstor.org/action/showPublisher?publisherCode=shea

http://www.jstor.org/action/showPublisher?publisherCode=shea

http://www.jstor.org/stable/10.1086/591864?origin=JSTOR-pdf

http://www.jstor.org/page/info/about/policies/terms.jsp


infection control and hospital epidemiology november 2008, vol. 29, no. 11

o r i g i n a l a r t i c l e

Clinical and Molecular Epidemiology of Healthcare-AssociatedInfections Due to Extended-Spectrum b-Lactamase (ESBL)–

Producing Strains of Escherichia coli and Klebsiella pneumoniaeThat Harbor Multiple ESBL Genes

Anucha Apisarnthanarak, MD; Pattarachai Kiratisin, MD, PhD; Linda M. Mundy, MD

objectives. To characterize healthcare-associated infections due to extended-spectrum b-lactamase (ESBL)–producing strains of Esch-erichia coli and Klebsiella pneumoniae that harbor multiple ESBL genes, as opposed to a single ESBL gene.

methods. All patients with a confirmed healthcare-associated infection due to an ESBL-producing strain of E. coli or K. pneumoniaewere enrolled in the study. Molecular typing of isolates was performed, and the comparative risks and outcomes of patients were analyzed.

results. Among 71 patients with healthcare-associated infection due to an ESBL-producing strain of E. coli or K. pneumoniae, the genefor CTX-M, with or without other ESBL genes, was identified in all 51 (100%) of the patients infected with an E. coli strain and in 18(90%) of the 20 patients infected with a K. pneumoniae strain. Of these 71 patients, 17 (24%) met the definition of healthcare-associatedinfection due to an ESBL-producing strain that harbored multiple genes; in multivariate analysis, previous exposure to 3 or more classesof antibiotics (adjusted odds ratio, 4.5 [95% confidence interval, 1.7–75.2]) was the sole risk factor for healthcare-associated infection dueto an ESBL-producing strain that harbored multiple ESBL genes. Isolates recovered from patients with healthcare-associated infection dueto an ESBL-producing strain that harbored multiple ESBL genes were more resistant to various antibiotic classes, and, compared withpatients with healthcare-associated infection due to an ESBL-producing strain that harbored a single ESBL gene, they were more likely tohave ineffective initial empirical antimicrobial therapy (52% vs 94%; odds ratio, 5.1 [95% confidence interval, 1.04–14.5]).

conclusions. CTX-M ESBL is highly prevalent in Thailand. Patients with healthcare-associated infection due to an ESBL-producingstrain that harbored multiple ESBL genes were more likely to have had ineffective initial empirical antimicrobial therapy, and, given thatantibiotic selection pressure was the only associated risk, we suggest focused antimicrobial stewardship programs to limit the emergenceand spread of healthcare-associated infection due to ESBL-producing strains in this middle-income country.

Infect Control Hosp Epidemiol 2008; 29:1026-1034

From the Division of Infectious Diseases, Faculty of Medicine, Thammasart University Hospital, Pratumthani (A.A.), and the Department of Microbiology,Faculty of Medicine, Siriraj Hospital, Bangkok (P.K.), Thailand; and the Saint Louis University School of Public Health, Saint Louis, Missouri (L.M.M.).

Received April 21, 2008; accepted July 9, 2008; electronically published October 9, 2008.� 2008 by The Society for Healthcare Epidemiology of America. All rights reserved. 0899-823X/2008/2911-0005$15.00. DOI: 10.1086/591864

Infections caused by microorganims that produce extended-spectrum b-lactamase (ESBL), a plasmid-encoded b-lacta-mase, play a major role in healthcare-associated, multidrug-resistant infections worldwide.1-4 The unfavorable clinical andeconomic outcomes associated with infections due to ESBL-producing gram-negative microorganisms are of global con-cern.5-6 The emergence of additional resistance to other an-tibiotic classes among ESBL-producing strains of Escherichiacoli and Klebsiella pneumoniae has complicated and compro-mised clinical care.3,7 Studies from developed countries havereported the risk factors for, and the outcomes of, healthcare-associated infections due to ESBL-producing strains of mi-croorganisms among adults,8-13 whereas there is a paucity ofsuch data available from Asian-Pacific countries.14-17

In Thailand, antibiotic-management programs are uncom-mon, and the rates of multidrug-resistant E. coli and K. pneu-moniae infections have steadily increased over the pastdecade.18,19 No data exist for the clinical and molecular epi-demiology of patients with healthcare-associated infectiondue to ESBL-producing strains of microorganisms that harbormultiple ESBL genes. Studies that integrate information onmolecular epidemiology and clinical outcomes would provideinsight into the epidemiology, treatment, and prevention ofhealthcare-associated infections due to ESBL-producingstrains. We therefore conducted a study to describe the clinicaland molecular epidemiology of patients with healthcare-as-sociated infection due to ESBL-producing strains of E. colior K. pneumonia that harbor multiple ESBL genes, as opposedto a single ESBL gene.



hais due to esbl-producing microorganisms 1027

methods

Setting and Patients

Our study was conducted at the Thammasart University Hos-pital, a 500-bed academic medical center in Pratumthani,Thailand. Study patients were identified from the source re-cords of the hospital’s clinical microbiological laboratory,which performed bacterial cultures of all clinical specimensprocured. The records of the Division of Infection Controlwere reviewed to ensure that all eligible patients were iden-tified. All consecutive adults with culture-positive specimensof ESBL-producing strains of E. coli or K. pneumoniae whomet the definition of healthcare-associated infection due tothese microorganisms from January 1 through 30, 2006, wereeligible for the study and, if chosen, were prospectively fol-lowed for outcomes (eg, clinical response to initial antimi-crobial therapy, crude mortality, duration of hospital stay afterinfection, and hospital charges accrued after infection). Eachpatient had a single specimen submitted for molecular typing.To evaluate the risk factors for, and the outcomes of, health-care-associated infections due to ESBL-producing strains ofE. coli or K. pneumoniae that harbor multiple ESBL genes,all patients with this type of healthcare-associated infectionwere compared with patients with healthcare-associated in-fection due to ESBL-producing strains of E. coli or K. pneu-moniae that harbor a single ESBL gene. Institutional reviewboard approval was obtained prior to the initiation of ourstudy.

Definitions and Data Collection

Healthcare-associated infection was defined as (1) infectionthat occurred in a patient more than 48 hours after hospitaladmission; (2) infection that occurred in a patient less than48 hours after hospital admission, if that patient had beenhospitalized within the prior 30 days; or (3) infection thatoccurred in a patient transferred from another hospital orfrom a nursing home.17 Patients enrolled in our study werealso assessed for the criteria of infection, as delineated by theCenters for Disease Control and Prevention.20 Infection witha strain of E. coli or K. pneumoniae that harbored multipleESBL genes was defined by the recovery of an isolate thathad at least 2 ESBL genes. Empirical antimicrobial therapywas defined as inappropriate if an antimicrobial agent activeagainst the isolate (as determined by in vitro susceptibilitytesting), prescribed at the usual recommended dose, was notadministered within 48 hours after the index blood cultureresult.21,22 Oxyimino-b-lactams (cefuroxime, cefotaxime, cef-triaxone, ceftazidime, and cefepime) and aztreonam wereconsidered inappropriate regardless of their minimum inhib-itory concentration.22 Time at risk was defined as the intervalfrom admission to the index blood culture result. The failureof initial empirical antimicrobial therapy was considered tohave occurred if there was persistence of fever and presence

of leukocytosis and initial signs of infection.5 Data were ab-stracted from the inpatient medical record and included de-mographic, clinical, pharmacological, laboratory, and costvariables. Cost data were obtained from the hospital costaccounting database. Hospital costs represented the sum ofdirect and indirect costs required to provide healthcare ser-vices and medications that were related to each patient’s di-agnosis and management plan. All costs in Thai baht currencywere converted to US dollars (exchange rate, 35 baht per USdollar).

Microbiology Methods

Identification of E. coli and K. pneumoniae isolates was per-formed by use of an automated identification system (VitekSystem; bioMerieux). Susceptibilities to antimicrobial agentswere determined by use of the disk-diffusion method (cef-tazidime and cefotaxime with and without clavulanic acid),with interpretations taken from the Clinical and LaboratoryStandards Institute (CLSI) guidelines.23 ESBL production wasconfirmed by double-disk synergy testing, in accordance withCLSI guidelines, and minimum inhibitory concentrations forceftazidime, cefotaxime, ceftriaxone, and carbapenems weredetermined by use of the Etest (AB Biodisk). We consideredESBL-producing strains of E. coli or K. pneumoniae to bemultidrug resistant if resistance to more than 2 classes ofother additional antimicrobial agents (quinolones, trimeth-oprim-sulfamethoxazole, or aminoglycosides) was identified.

Molecular Typing and Pulsed-Field Gel Electrophoresis(PFGE)

All ESBL-producing strains of E. coli or K. pneumoniae werefurther typed for genes encoding b-lactamase (bla), usingpolymerase chain reaction (PCR) sequencing as previouslydescribed elsewhere.24,25 Both strands of the PCR-amplifiedbla genes were subjected to DNA sequencing, to identify se-quences coding for 7 major ESBL families (TEM, SHV, CTX-M, VEB, OXA, PER, and GES), and DNA sequence analysisby the 3100 Genetic Analyzer (Applied Biosystems). For mo-lecular typing, the chromosomal DNA of all isolates recoveredwas compared by PFGE analysis by use of a CHEF MapperXA apparatus (Bio-Rad Laboratories). DNA restriction withthe enzyme XbaI (New England BioLabs) was carried out asdescribed elsewhere.26 The PFGE patterns were compared us-ing Fingerprinting II Basic software, version 3.0 (Bio-RadLaboratories). The Dice coefficient was used for similaritymeasurements of the band patterns. Isolates were consideredto be unique strains if the band pattern similarity was lessthan 85%.26

Statistical Analysis

Data analysis was performed using SPSS, version 12 (SPSS).Categorical variables were compared using the x2 or Fisherexact test, as appropriate. Continuous variables were com-



1028 infection control and hospital epidemiology november 2008, vol. 29, no. 11

table 1. Demographic and Clinical Characteristics of and Outcomes for Patients With Healthcare-Associated In-fection Due to Extended-Spectrum b-Lactamase (ESBL)–Producing Strains of Escherichia coli or Klebsiella pneumoniae,at the Thammasart University Hospital (January 1–30, 2006)

Variable

No. (%) of patients with healthcare-associatedinfection due to ESBL-producing strain

PTotal

( )n p 71

With strainharboring

multiple genes( )n p 17

With strainharboring

single gene( )n p 54

Age, median (range), years 58 (15–89) 56 (15–80) 58 (15–89) .83Male sex 36 (51) 9 (53) 27 (50) .43Underlying disease

Diabetes 23 (32) 8 (47) 16 (30) .18Cerebrovascular accident 22 (31) 7 (41) 15 (22) .30Othera 15 (21) 5 (30) 10 (19) .34

APACHE-II score at admission, median (range) 17 (4–23) 18 (6–23) 17 (4–22) .71Time at risk, median (range), days 10 (0–27) 12 (6–27) 5 (0–14) .005Type of infection

Urinary tract 44 (62) 10 (59) 34 (63) .75Pneumonia 12 (17) 4 (24) 8 (15) .40Bloodstream 6 (8) 2 (12) 4 (7) .57Otherb 9 (13) 2 (12) 7 (13) .89

Previous exposure to x3 classes of antibioticsc 45 (63) 15 (88) 30 (67) .01Empirical antibiotic regimenc

3G or 4G cephalosporins 40 (56) 9 (53) 31 (57) .74Quinolones 12 (17) 4 (24) 8 (15) .40b-lactam and/or b-lactamase inhibitors 12 (17) 3 (18) 9 (17) .92Otherd 8 (11) 2 (12) 6 (11) .48

Inadequate antibiotic treatment 60 (85) 17 (100) 43 (80) .04Outcomes

Ineffective initial antimicrobial therapy 44 (62) 16 (94) 28 (52) .001Crude mortality 22 (31) 5 (30) 17 (31) .87Length of hospitalization, median (range), days 24 (1–48) 25 (1–48) 24 (1–40) .71Hospital costs accrued after infection,

median (range), $ 2,145 (45–3,225) 2,155 (85–3,225) 2,089 (45–2,990) .83

note. Each patient had a single specimen submitted for molecular typing (ie, a single isolate was recovered from each patient);APACHE, Acute Physiology and Chronic Health Evaluation; 3G or 4G, third- or fourth-generation.a Previous receipt of corticosteroids or immunosuppressive therapy; human immunodeficiency virus infection; or chronic liver disease.b Intra-abdominal, biliary infection and skin and soft tissue infection.c Within the previous 3 months; a patient may have received 11 antibiotic.d Penicillin, macrolides, trimethoprim-sulfamethoxazole, first- and second-generation cephalosporins, and/or aminoglycosides.

pared using the Wilcoxon rank sum test. All P values were2 tailed, and a P value of less than .05 was considered sta-tistically significant. Adjusted odds ratios (aORs) and 95%confidence intervals (CIs) were computed. To identify riskfactors for multidrug-resistant healthcare-associated infectiondue to ESBL-producing strains of E. coli or K. pneumoniae,the variables (eg, underlying diabetes, previous exposure to3 or more classes of antibiotics, and time at risk) that wereassociated with more than 10% of the patients who had thistype of infection (by use of bivariable analysis; ) orP ! .20that had a prior clinical significance (eg, Acute Physiologyand Chronic Health Evaluation II score) were entered intobackward, stepwise logistic regression models. Significant var-iables were grouped if they were closely correlated, and only1 variable per group was chosen for entry into a model. The

final model was chosen on the basis of plausibility and onthe basis of having the lowest �2 log likelihood function.

results

Study Sample

During the 6-month study period, 51 (26%) of 200 patientswith E. coli infection had a healthcare-associated infection dueto an ESBL-producing strain of E. coli, and 20 (21%) of 96patients with K. pneumoniae infection had a healthcare-asso-ciated infection due to an ESBL-producing strain of K. pneu-moniae. Each of these 71 patients had a single episode of health-care-associated infection with either the E. coli or K.pneumoniae strain. The median age was 58 years (range, 15–89 years); 35 patients (49%) were female, 23 (32%) had dia-




table 2. Antimicrobial Susceptibility Data of Healthcare-Asso-ciated Infections Due to Extended-Spectrum b-Lactamase (ESBL)–Producing Strains of Escherichia coli or Klebsiella pneumoniae, at theThammasart University Hospital (January 1–30, 2006)

Drug(s)MIC90,mg/L

No. (%) of susceptible isolates

Total( )n p 71

With singleESBL gene( )n p 54

With multipleESBL genes

( )n p 17

Ceftazidime x256 26 (37) 26 (48) 0a

Cefotaxime x256 12 (17) 12 (22) 0a

Ceftriaxone x256 8 (11) 8 (15) 0a

Imipenem 0.19 71 (100) 54 (100) 17 (100)Meropenem 0.06 71 (100) 54 (100) 17 (100)Ertapenem 0.25 71 (99) 54 (100) 17 (100)Amox-CA … 28 (40) 26 (48) 2 (12)a

Gentamicin … 25 (35) 24 (44) 1 (6)a

Amikacin … 35 (49) 29 (54) 6 (35)Ciprofloxacin … 15 (21) 14 (26) 1 (6)a

Pip-Taz … 28 (40) 26 (48) 2 (12)a

TMP-SMZ … 10 (14) 10 (19) 0 (0)a

note. Each patient had a single specimen submitted for molecular typing(ie, a single isolate was recovered from each patient); Amox-CA, amoxicillin–clavulanic acid; MIC, minimum inhibitory concentration; Pip-Taz, pipera-cillin-tazobactam; TMP-SMZ, trimethoprim-sulfametoxazole.a .P ! .05

table 3. Molecular Epidemiology of 71 Ex-tended-Spectrum b-Lactamase (ESBL)–ProducingStrains of Escherichia coli and Klebsiellapneumoniae

Microorganism,protein(s) found

No. (%)of isolates

E. coliCTX-M-14 32 (63)CTX-M-15 7 (14)CTX-M-15 and VEB-1 4 (8)CTX-M-55 8 (16)

K. pneumoniaeCTX-M-14 4 (20)CTX-M-14 and SHV-12 7 (35)CTX-M-15 2 (10)CTX-M-15 and SHV-12 2 (10)CTX-M-15 and VEB-1 2 (10)CTX-M-55 1 (5)SHV-12 and VEB-1 2 (10)

note. There were 51 E. coli isolates recovered from51 patients and 20 K. pneumoniae isolates recovered from20 patients (ie, a single isolate was recovered from eachpatient), and only ESBL genes were reported.

betes, and 45 (63%) received antibiotic(s) prior to the isolationof the ESBL-producing strain of E. coli or K. pneumoniae (Table1). Of the 51 patients with healthcare-associated infection dueto an ESBL-producing strain of E. coli, 32 (63%) had urinarytract infection, 8 (17%) had pneumonia, and 4 (8%) hadbloodstream infection. Of the 20 patients with healthcare-as-sociated infection due to an ESBL-producing strain of K. pneu-moniae, 12 (60%) had urinary tract infection, 4 (20%) hadpneumonia, and 2 (10%) had bloodstream infection. Seventeenpatients (24%) had a healthcare-associated infection due to anESBL-producing strain of E. coli or K. pneumoniae that har-bored multiple genes (Table 1).

Phenotypic and Molecular Characteristicsof the Clinical Isolates

Resistance to other antibiotics varied among the 71 isolatesof E. coli or K. pneumoniae recovered from patients (Table2). Of these isolates, 17 were recovered from patients whohad a multidrug-resistant healthcare-associated infection dueto an ESBL-producing strain of E. coli ( ) or K. pneu-n p 11moniae ( ), and none of these 17 isolates were resistantn p 6to imipenem, meropenem, or ertapenem. The 51 E. coli and20 K. pneumoniae isolates were screened for bla genes by useof PCR, with a focus on the major ESBL families (TEM, SHV,CTX-M, VEB, OXA, PER, and GES; Table 3).

Among the 51 E. coli isolates, the blaTEM gene was detectedin 37 (73%) isolates, and the blaCTX gen was detected in all51 (100%); the blaSHV gene was not detected. All of the blaTEM

genes found in the 37 E. coli isolates encoded for a non-ESBL

TEM-1 protein. Of the 51 isolates that carried a blaCTX gene,32 (63%) carried a blaCTX gene that encoded for the CTX-M-14 protein, 11 (21%) carried a blaCTX gene that encodedfor the CTX-M-15 protein, and 8 (16%) carried a blaCTX genethat encoded for the CTX-M-55 protein. An ESBL VEB-1protein was detected, together with CTX-M-15, in 4 isolates.Four (8%) of the 51 E. coli isolates had multiple ESBL genes.There was no association between the isolates’ resistance pat-terns and the specific genotypes.

Among the 20 K. pneumoniae isolates, the blaTEM gene wasdetected in 11 (55%), the blaSHV gene was detected in 18(90%), and the blaCTX gene was detected in 18 (90%) isolates.All 11 K. pneumoniae isolates carried the blaTEM gene thatencoded for a non-ESBL TEM-1 protein, and of the 18 isolatesthat carried a blaSHV gene, 4 (22%) carried a blaSHV gene thatencoded for a non-ESBL SHV-1 protein. The bla gene codingfor the non-ESBL SHV-11 protein was detected in 3 isolates;11 isolates carried a blaSHV gene that encoded for an ESBLSHV-12 protein, and 4 isolates carried a blaVEB gene thatencoded for an ESBL VEB-1 protein. Genes coding for TEM-1 and SHV-12 were detected in 2 isolates. Six isolates carrieda blaCTX gene that encoded for the CTX-M-15 protein, 11isolates carried a blaCTX gene that encoded for CTX-M-14,and 1 isolate carried a blaCTX gene that encoded for CTX-M-55. TEM-1 was identified, together with CTX-M-15, in 4isolates. Multiple ESBL genes were detected in 13 (65%) ofthe 20 K. pneumoniae isolates. Two E. coli isolates (both car-rying genes coding for CTX-M-15 and VEB-1; Figure 1) and4 K. pneumoniae isolates (3 carrying genes coding for CTX-M-14 and SHV-12, and 1 carrying genes coding for CTX-M-15 and SHV-12; Figure 2) were found to be genetically related



1030

fig

ur

e1.

Den

drog

ram

sof

gen

etic

rela

tedn

ess

ofh

ealt

hcar

e-as

soci

ated

ESB

L-pr

odu

cin

gst

rain

sof

Esc

heri

chia

coli

base

don

puls

ed-fi

eld

gel

elec

trop

hor

esis

(PFG

E)

patt

ern

s.T

he

nu

mbe

rson

the

righ

tar

en

um

eric

alpe

rcen

tage

sre

late

dto

the

sim

ilari

tyin

dex.



1031

fig

ur

e2.

Den

drog

ram

sof

gen

etic

rela

tedn

ess

ofh

ealt

hcar

e-as

soci

ated

ESB

L-pr

odu

cin

gst

rain

sof

Kle

bsie

llapn

eum

onia

eba

sed

onpu

lsed

-fiel

dge

lel

ectr

oph

ores

is(P

FGE

)pa

tter

ns.

Th

en

um

bers

onth

eri

ght

are

nu

mer

ical

perc

enta

ges

rela

ted

toth

esi

mila

rity

inde

x.




by PFGE (by use of the Dice coefficient to find a band patternsimilarity of 85% or more). Two isolates that were recoveredfrom 2 separate patients with healthcare-associated infectiondue to an ESBL-producing strain of E. coli were geneticallyrelated; these patients were admitted to the hospital in Feb-ruary 2006, to the same medicine unit and with the sameservice team. Four isolates recovered from 4 separate patientswith healthcare-associated infection due to an ESBL-produc-ing strain of K. pneumoniae were genetically related; all 4patients were admitted to the hospital in May 2006, to thesame medicine unit and with the same service team.

Assessment of Risk Factors for and Outcomes ofHealthcare-Associated Infections Due to ESBL-ProducingMicroorganisms

In multivariate analysis, previous exposure to 3 or more clas-ses of antibiotics (aOR, 4.5 [95% CI, 1.7–75.2]; ) wasP p .01the sole risk factor for healthcare-associated infection due toESBL-producing strains that harbored multiple ESBL genes.Overall, the crude mortality rate was 31% among the 71patients with healthcare-associated infection due to an ESBL-producing strain of E. coli or K. pneumoniae (ie, 22 patientsdied). Sixty (85%) of the 71 patients received inadequateantimicrobial therapy. The median delay from time of spec-imen procurement to receipt of an appropriate antibiotic reg-imen was 2 days (range, 1–6 days). There was no differencewith respect to crude mortality, length of stay, or hospitali-zation cost between healthcare-associated infection due toESBL-producing strains that harbored multiple ESBL genesand those that harbored single ESBL genes (Table 1). How-ever, healthcare-associated infections due to ESBL-producingstrains of E. coli or K. pneumoniae that harbored multiplegenes were more resistant to various classes of antibiotics(Table 2). Additionally, compared with patients with health-care-associated infections due to ESBL-producing strains ofE. coli or K. pneumoniae that harbored a single ESBL gene,the patients infected with the strains that harbored multiplegenes were more likely to have ineffective initial empiricalantimicrobial therapy (52% vs 94%; OR, 5.1 [95% CI, 1.04–14.5]; ).P p .001

discussion

Our study has 4 major findings. First, our data provide evi-dence that the CTX-M ESBL was endemic in Thailand in2006, whereas SHV and VEB were detected in a minority ofclinical isolates. Second, several isolates recovered from pa-tients with healthcare-associated infections due to ESBL-pro-ducing strains that harbored multiple ESBL genes (24%) hadhigh-level resistance to various antibiotic classes, and thesepatients were more likely to have ineffective initial empiricalantimicrobial therapy. Third, we identified prior antibioticexposure as the sole risk factor for healthcare-associated in-fections due to ESBL-producing strains that harbor multipleESBL genes, compared with those that harbor a single ESBL

gene. Fourth, the prevalence of ESBL-producing strains ofbacteria appeared to be polyclonal, because none of the majorepidemic strains were identified. Our findings confirm that,beyond its importance as a pathogen in nonhospitalized pa-tients,27-29 CTX-M ESBL-producing microorganisms play aclinically relevant role in healthcare-associated infections.

Although most ESBL-producing Enterobacteriaceae in theUnited States belong to the TEM and SHV families,30 theemergence of the CTX-M family has occurred worldwide.31-33

Although the prevalence of ESBL-producing isolates is seem-ingly high in our study, until now the prevalence of ESBLtypes in Thailand has not been thoroughly explored. In priorThai studies, SHV-12 and SHV-5 were the most prevalenttypes among Enterobacteriaceae, and although the VEB-1ESBL was identified, there was no detection of TEM or CTX-M types.34,35 We previously reported the first detection of aCTX-M ESBL in Thailand—where CTX-M-15 and a novelCTX-M-55 were detected.29,36 Unlike many CTX-M ESBLs,CTX-M-15, a member of the CTX-M-1 family, is the ESBLcapable of hydrolyzing ceftazidime because of the singleamino acid mutation D240G.37 Our current study clearlydemonstrates that the CTX-M ESBLs, mainly CTX-M-14 andCTX-M-15, have become highly endemic in Thailand. Fur-thermore, our results confirm the earlier reports that SHV-12 is widespread among K. pneumoniae isolates in thisregion.34,35

A prior study of ESBL-producing isolates recovered fromhospitalized patients detected clonally unrelated CTX-M–pro-ducing isolates, suggesting that community-associated acqui-sition had occurred.38 In contrast, isolates producing TEM andSHV enzymes typically were reported in epidemic scenarios.39,40

However, some outbreaks caused by CTX-M producing isolateshave been reported.41-43 In our study, the emergence of theCTX-M ESBL and of the genetic distinction of the majority ofthe healthcare-associated infections due to ESBL-producingstrains implicates antibiotics in having an adverse effect onendogenous flora, in the emergence and spread of healthcare-associated infections due to ESBL-producing strains rather thandue to patient-to-patient transmission of exogenous flora.44

Notably, in all clonal-related isolates, CTX-M-14 or CTX-M-15 were detected along with SHV-12 and VEB-1, thus sup-porting the recommendation by Paterson and Yu45 for en-hanced contact precautions, active surveillance in high-riskareas, and decreased oxyimino-b-lactams consumption, to con-trol epidemic TEM-SHV–producing E. coli. In contrast, thesporadic CTX-M among healthcare-associated infections dueto ESBL-producing strains may represent the emergence ofresistance in endogenous flora or its spread by undetectedhorizontal transmission.

Recognized risk factors for ESBL-producing strains of E.coli or K. pneumoniae, such as greater severity of illness, deviceutilization, antibiotic exposures, and proxy markers (eg,longer duration of hospitalization and longer length of stayin intensive care units), have been extensively reported.8-17 Instudies that evaluated risk factors specific to CTX-M–pro-




ducing strains of E. coli, diabetes mellitus, renal disease, uri-nary tract infection, gynecological surgery, and the presenceof a urinary catheter were independent risk factors, and ahigh proportion of these infections were acquired in thecommunity.40,46 However, when only patients infected with aCTX-M–producing clone were considered, the risk factorswere cardiovascular disease, urinary incontinence, the pres-ence of a urinary catheter, and polymicrobial infection.46 Useof oxyimino-b-lactams, of which cefotaxime was the mostfrequently used, was related to CTX-M–producing isolates,thus increasing the cefotaximase activity of these enzymes,whereas the use of fluoroquinolones was associated with theciprofloxacin-resistant SHV-TEM–producing isolate.40,46 Pre-vious reports suggested that healthcare-associated infectionsdue to ESBL-producing strains of E. coli or K. pneumoniaewere independent predictors of longer length of stay, highermortality, and excess hospital charges, whereas inadequateantimicrobial therapy was associated with highermortality.5,6,8,9,17 Because we compared patients who hadstrains that harbored multiple ESBL genes with patients whohad strains that harbored a single ESBL gene, no differencein clinical or economic outcomes was anticipated, and it wasnot surprising to find that ESBL-producing strains of E. coliand K. pneumoniae that harbored multiple genes had a greaterresistance to various classes of antibiotic and that initial em-pirical antimicrobial therapy was ineffective.

We recognize the study limitation of selection bias, as notedin all observational studies, and attempted to reduce such biasby including all consecutive patients identified from the sourcerecords of the hospital’s clinical microbiological laboratory. Thesmall sample size limited our capacity to detect other possiblerisk factors and outcomes of healthcare-associated infectiondue to ESBL-producing strains of E. coli or K. pneumoniae.With the detection of the gene for CTX-M with or withoutother ESBL genes in more than 95% of isolates, we were notable to evaluate risk factors specific to each ESBL gene. Lastly,because we did not perform active surveillance to detect col-onization in our study population, we may have underesti-mated the clonal spread of healthcare-associated infection dueto ESBL-producing microorganisms. Despite these limitations,to our knowledge, our investigation is the first to demonstratethat CTX-M among healthcare-associated infections is highlyendemic in Thailand. Furthermore, patients with healthcare-associated infection due to ESBL-producing strains of E. colior K. pneumoniae that harbored multiple ESBL genes had high-level resistance to various classes of antibiotics and were morelikely to have ineffective initial empirical antimicrobial therapy,indicating a need for focused antimicrobial stewardship pro-grams. Additional studies correlating the clinical and molecularepidemiology of the emergence and dynamic spread of health-care-associated infection due to ESBL-producing microorgan-isms will continue to guide treatment and prevention strategiesin this middle-income country.

acknowledgments

Financial support. A.A. was supported by the National Research Council.Potential conflicts of interest. L.M.M. is a consultant to GlaxoSmithKline

and a speaker with the Robert Michael Educational Institute. All other au-thors report no conflicts of interest relevant to this article.

Address reprint requests to Anucha Apisarnthanarak, MD, Division ofInfectious Diseases, Faculty of Medicine, Thammasart University Hospital,Pratumthani, Thailand, 12121 ([email protected]).

references

1. Rice LB, Willey SH, Papanicolaou GA, et al. Outbreak of ceftazidimeresistance caused by extended-spectrum b-lactamases at a Massachusettschronic-care facility. Antimicrob Agents Chemother 1990; 34:2200–2209.

2. Burwen DR, Banerjee SN, Gaynes RP. Ceftazidime resistance amongselected nosocomial gram-negative bacilli in the United States. The Na-tional Nosocomial Infections Surveillance System. J Infect Dis 1994; 170:1622–1625.

3. Itokazu GS, Quinn JP, Bell-Dixon C, Kahan FM, Weinstein RA. Anti-microbial resistance rates among aerobic gram-negative bacilli recoveredfrom patients in intensive care units: evaluation of a national postmar-keting surveillance program. Clin Infect Dis 1996; 23:779–784.

4. Nijssen S, Florijn A, Bonten MJM, Schmitz FJ, Verhoef J, Fluit AC. b-Lactam susceptibilities and prevalence of ESBL-producing isolates amongmore than 5000 European Enterobacteriaceae isolates. Int J AntimicrobAgents 2004; 24:585–591.

5. Lautenbach E, Patel JB, Bilker WB, Edelstein PH, Fishman NO. Ex-tended-spectrum b-lactamase–producing Escherichia coli and Klebsiellapneumoniae: risk factors for infection and impact of resistance on out-comes. Clin Infect Dis 2001; 32:1162–1171.

6. Paterson DL, Ko WC, Von Gottberg A, et al. International prospectivestudy of Klebsiella pneumoniae bacteremia: implications of extended-spectrum beta-lactamase production in nosocomial infections. Ann In-tern Med 2004; 140:26–32.

7. Hyle EP, Lipworth AD, Zaoutis TE, et al. Risk factors for increasingmultidrug resistance among extended-spectrum b-lactamase–producingEscherichia coli and Klebsiella pneumoniae. Clin Infect Dis 2005; 40:1317–1324.

8. Cordery RJ, Roberts CH, Cooper SJ, Bellinghan G, Shetty N. Evaluationof risk factors for the acquisition of bloodstream infections with ex-tended-spectrum b-lactamase-producing Escherichia coli and Klebsiellaspecies in the intensive care unit; antibiotic management and clinicaloutcome. J Hosp Infect 2008; 68:108–115.

9. Cosgrove SE. The relationship between antimicrobial resistance and pa-tient outcomes: mortality, length of hospital stay, and health care costs.Clin Infect Dis 2006; 42(Suppl 2):S82–S89.

10. Lautenbach E, Strom BL, Bilker WB, Patel JB, Edelstein PH, FishmanNO. Epidemiological investigation of fluoroquinolone resistance in in-fections due to extended-spectrum b-lactamase–producing Escherichiacoli and Klebsiella pneumoniae. Clin Infect Dis 2001; 33:1288–1294.

11. Skippen I, Shemko M, Turton J, Kaufmann ME, Palmer C, Shetty N.Epidemiology of infections caused by extended-spectrum b-lactamase-producing Escherichia coli and Klebsiella spp.: a nested case-control studyfrom a tertiary hospital in London. J Hosp Infect 2006; 64:115–123.

12. Pena C, Gudiol C, Tubau F, et al. Risk-factors for acquisition of extended-spectrum b-lactamase-producing Escherichia coli among hospitalized pa-tients. Clin Microbiol Infect 2006; 12:279–284.

13. Tumbarello M, Spanu T, Sanguinetti M, et al. Bloodstream infectionscaused by extended-spectrum-b-lactamase-producing Klebsiella pneu-moniae: risk factors, molecular epidemiology, and clinical outcome. An-timicrob Agents Chemother 2006; 50:498–504.

14. Apisarnthanarak A, Mundy LM. Prevalence, treatment, and outcome ofinfection due to extended-spectrum b-lactamase–producing microor-ganisms. Infect Control Hosp Epidemiol 2006; 27:326–327.




15. Kang CI, Kim SH, Kim DM, et al. Risk factors for and clinical outcomesof bloodstream infections caused by extended-spectrum b-lactamase-producing Klebsiella pneumoniae. Infect Control Hosp Epidemiol 2004;25:860–867.

16. Lin MF, Huang ML, Lai SH. Risk factors in the acquisition of extended-spectrum b-lactamase Klebsiella pneumoniae: a case-control study in adistrict teaching hospital in Taiwan. J Hosp Infect 2003; 53:39–45.

17. Apisarnthanarak A, Kiratisin P, Saifon P, Kitphati R, Dejsirilert S, MundyLM. Risk factors for and outcomes of healthcare-associated infection dueto extended-spectrum b-lactamase–producing Escherichia coli or Kleb-siella pneumoniae in Thailand. Infect Control Hosp Epidemiol 2007; 28:873–876.

18. Apisarnthanarak A, Danchaivijitr S, Khawcharoenporn T, et al. Effec-tiveness of education and an antibiotic-control program in a tertiarycare hospital in Thailand. Clin Infect Dis 2006; 42:768–775.

19. Dejsirilert S, Apisarnthanarak A, Kitphati R, et al. The status of anti-microbial resistance in Thailand among gram-negative pathogens inbloodstream infections: NARST data, 2000–2003 [abstract FP-A-3]. In:Program and abstracts of the 9th Western Pacific Congress on Che-motherapy and Infectious Diseases (Bangkok). Bangkok: Medinfo GD,2004:185.

20. Garner JS, Jarvis WR, Emori TG, Horan TC, Hughes JM. CDC definitionsfor nosocomial infections, 1988. Am J Infect Control 1988; 16:128–140.

21. Kollef MH. Inadequate antimicrobial treatment: an important deter-minant of outcome for hospitalized patients. Clin Infect Dis 2000; 31(Suppl 4):S131–S138.

22. Rodriguez-Bano J, Navarro MD, Romero L, et al. Bacteremia due toextended-spectrum b-lactamase–producing Escherichia coli in the CTX-M era: a new clinical challenge. Clin Infect Dis 2006; 43:1407–1414.

23. CLSI. Performance standards for antimicrobial susceptibility testing: 17thinformational supplement. CLSI document. Wayne, PA: CLSI, 2007:M100-S17.

24. Coque TM, Oliver A, Perez-Diaz JC, et al. Genes encoding TEM-4, SHV-2, and CTX-M-10 extended-spectrum ß-lactamases are carried by mul-tiple Klebsiella pneumoniae clones in a single hospital (Madrid, 1989 to2000). Antimicrob Agents Chemother 2002; 46:500–510.

25. Oliver A, Perez-Diaz JC, Coque TM, et al. Nucleotide sequence andcharacterization of a novel cefotaxime-hydrolyzing ß-lactamases (CTX-M-10). Antimicrob Agents Chemother 2001; 45:612–620.

26. Tenover, FC, Arbeit RD, Goering RV, et al. Interpreting chromosomalDNA restriction patterns produced by pulsed-field gel electrophoresis:criteria for bacterial strain typing. J Clin Microbiol 1995; 33:2233–2239.

27. Colodner R, Rock W, Chazan B, et al. Risk factors for the developmentof extended-spectrum b-lactamase-producing bacteria in nonhospitali-zed patients. Eur J Clin Microbiol Infect Dis 2004; 23:163–167.

28. Calbo E, Romanı V, Xercavins M, et al. Risk factors for community-onset urinary tract infections due to Escherichia coli harboring extended-spectrum b-lactamases. J Antimicrob Chemother 2006; 57:780–783.

29. Apisarnthanarak A, Kiratisin P, Saifon P, Kitphati R, Dejsirilert S, MundyLM. Clinical and molecular epidemiology of community-onset, ex-tended-spectrum b-lactamase-producing Escherichia coli infections inThailand: a case-case-control study. Am J Infect Control 2007; 35:606–612.

30. Paterson DL, Hujer KM, Hujer AM, et al. Extended-spectrum b-lacta-mases in Klebsiella pneumonia bloodstream isolates from seven countries:dominance and widespread prevalence of SHV- and CTX-M-type b-lactamases. Antimicrob Agents Chemother 2003; 47:3554–3560.

31. Mehrgan H, Rahbar M. Prevalence of extended-spectrum b-lactamase-producing Escherichia coli in a tertiary care hospital in Tehran, Iran. IntJ Antimicrob Agents 2008; 31:147–151.

32. Lavilla S, Gonzalez-Lopez JJ, Sabate M, et al. Prevalence of qnr genesamong extended-spectrum b-lactamase-producing enterobacterial iso-lates in Barcelona, Spain. J Antimicrob Chemother 2008; 61:291–295.

33. Wu JJ, Chen HM, Ko WC, Wu HM, Tsai SH, Yan JJ. Prevalence ofextended-spectrum b-lactamases in Proteus mirabilis in a Taiwanese uni-versity hospital, 1999 to 2005: identification of a novel CTX-M enzyme(CTX-M-66). Diagn Microbiol Infect Dis 2008; 60:169–175.

34. Chanawong A, M’Zali FH, Heritage J, Lulitanond A, Hawkey PM. SHV-12, SHV-5, SHV-2a and VEB-1 extended-spectrum b-lactamases inGram-negative bacteria isolated in a university hospital in Thailand. JAntimicrob Chemother 2001; 48:839–852.

35. Girlich D, Poirel L, Leelaporn A, et al. Molecular epidemiology of theintegron-located VEB-1 extended-spectrum b-lactamase in nosocomialenterobacterial isolates in Bangkok, Thailand. J Clin Microbiol 2001; 39:175–182.

36. Kiratisin P, Apisarnthanarak A, Saifon P, Laesripa C, Kitphati R, MundyLM. The emergence of a novel ceftazidime-resistant CTX-M extended-spectrum b-lactamase, CTX-M-55, in both community-onset and hos-pital-acquired infections in Thailand. Diagn Microbiol Infect Dis 2007;58:349–355.

37. Moubareck C, Daoud Z, Hakime NI, et al. Countrywide spread of com-munity- and hospital-acquired extended-spectrum b-lactamase (CTX-M-15)-producing Enterobacteriaceae in Lebanon. J Clin Microbiol 2005;43:3309–3313.

38. Ben-Ami R, Schwaber MJ, Navon-Venezia S, et al. Influx of extended-spectrum b-lactamase–producing Enterobacteriaceae into the hospital.Clin Infect Dis 2006; 42:925–934.

39. Safdar N, Maki DG. The commonality of risk factors for nosocomialcolonization and infection with antimicrobial-resistant Staphylococcusaureus, Enterococcus, gram-negative bacilli, Clostridium difficile, and Can-dida. Ann Intern Med 2002; 136:834–844.

40. Rodriguez-Bano J, Navarro MD, Romero L, et al. Clinical and molecularepidemiology of extended-spectrum b-lactamase–producing Escherichiacoli as a cause of nosocomial infection or colonization: implication forcontrol. Clin Infect Dis 2006; 42:37–45.

41. Moriguchi N, Itahashi Y, Tabata N, et al. Outbreak of CTX-M-3-typeextended-spectrum b-lactamase-producing Enterobacter cloacae in a pe-diatric ward. J Infect Chemother 2007; 13:263–264.

42. Lavollay M, Mamlouk K, Frank T, et al. Clonal dissemination of a CTX-M-15 b-lactamase-producing Escherichia coli strain in the Paris area,Tunis, and Bangui. Antimicrob Agents Chemother 2006; 50:2433–2438.

43. Pai H, Kim MR, Seo MR, Choi TY, Oh SH. A nosocomial outbreak ofEscherichia coli producing CTX-M-15 and OXA-30 b-lactamase. InfectControl Hosp Epidemiol 2006; 27:312–314.

44. Harris AD, Perencevich EN, Johnson JK, et al. Patient-to-patient trans-mission is important in extended-spectrum b-lactamase–producing Kleb-siella pneumoniae acquisition. Clin Infect Dis 2007; 45:1347–1350.

45. Paterson DL, Yu VL. Extended-spectrum b-lactamases: a call for im-proved detection and control. Clin Infect Dis 1999; 29:1419–1422.

46. Lavigne JP, Marchandin H, Delmas J, et al. CTX-M b-lactamase-pro-ducing Escherichia coli in French hospitals: prevalence, molecular epi-demiology, and risk factors. J Clin Microbiol 2007; 45:620–626.



Documents

Clinical and Molecular Epidemiology of Healthcare‐Associated Infections Due to Extended‐Spectrum β‐Lactamase (ESBL)–Producing Strains ofEscherichia coliandKlebsiella pneumoniaeThat