Role of Persistent Environmental Contamination And

Home , Cefdaloxime, Ceftiolene

Role of Persistent Environmental Contamination and Universal Glove and Gown Use in the Acquisition of Extended Spectrum Beta-Lactamase-Producing Enterobacteriaceae in Intensive Care Unit Patients

Item Type dissertation

Authors Ajao, Adebola Oluwakemi

Publication Date 2011

Abstract Background: Extended-spectrum beta-lactamase (ESBL)- producing Gram-negative rods (GNR) are emerging pathogens that are associated with considerable morbidity, mortality and costs among hospitalized patients. The association between persistent environ...

Keywords extended spectrum beta-lactams; patient-to-patient transmission; universal glove and gown; Cross Infection-- prevention & control; Escherichia coli; Intensive Care Units; Klebsiella pneumoniae; Universal Precautions

Download date 02/10/2021 01:13:51

Link to Item http://hdl.handle.net/10713/780

CURRICULUM VITAE

Name: Adebola Oluwakemi Ajao

Permanent Address: 211 Riverbend Lane Nashville, TN, 37221

Degree and Date to be Conferred: Ph.D., 2011

Collegiate Institutions Attended:

2006 -2011 University of Maryland, Baltimore, MD Ph.D., Molecular Epidemiology

2000-2001 Boston University, Boston, MA M.P.H Epidemiology and Biostatistics

1997-2000 Purdue University, West Lafayette, IN B.S., Microbiology

Published Manuscripts

1. Harris AD, Johnson JK, Thom KA, Morgan DJ, McGregor JC, Ajao AO, Moore AC, Comer AC, Furuno JP. Risk factors for Development of Intestinal Colonization with Imipenem-Resistant Pseudomonas aeruginosa in the Intensive Care Unit Setting. Infect Control and Hosp Epidemiol. 2011 Jul; 32(7):719-22.

2. Ajao AO, Harris AD, Roghmann MC, Johnson JK, Zhan M, McGregor JC, Furuno JP. A Systematic Review of Measurement and Adjustment for Colonization Pressure in Methicillin-resistant Staphylococcus aureus, Vancomycin-resistant Enterococci and Clostridium difficile Acquisition Studies. Infect Control and Hosp Epidemiol. 2011 May;32(5):481-9

3. Ajao AO, Robinson G, Lee MS, Ranke TD, Venezia RA, Furuno JP, Harris AD, Johnson JK. Comparison of Culture Media for Detection of Acinetobacter baumannii in Surveillance Cultures of Critically-ill Patients. Eur J Clin Microbiol Infect Dis. 2011 Apr 12

Abstracts

1. Ajao AO, Harris AD, Roghmann M, Perencevich EN, Schweizer ML, Chen WH, Furuno JP. Age as a Non-linear Predictor in the Progression from Methicillin- resistant Staphylococcus aureus (MRSA) Colonization to Infection. 49th Annual Interscience Conference on Antimicrobial Agents and Chemotherapy (ICAAC), San Francisco, CA, September 2009.

2. Furuno JP, Harris AD, Perencevich EN, Ajao AO, Schweizer ML, Andersen DA, Chan JA, Miller RR. Infection and Rapid Hospital Readmission. 49th Annual Interscience Conference on Antimicrobial Agents and Chemotherapy (ICAAC), San Francisco, CA, September 2009.

3. Ajao AO, Harris AD, Roghmann M, Johnson JK, Zhan M, McGregor JC, Furuno JP. Systematic Review of Measurement and Adjustment for Colonization Pressure in Methicillin-resistant Staphylococcus aureus (MRSA) and Vancomycin- resistant enterococci (VRE). Society for Healthcare Epidemiology of America (SHEA) Annual Meeting, Atlanta, Georgia, March 2010.

4. Roghmann M, Ajao AO, Lydecker A, Adhikari RP, Karauzum H, Chen WH, Johnson JK, Aman MJ. Lower Antibody Levels to Staphylococcus aureus Exotoxins associated with Sepsis in Hospitalized Patients with Invasive S. aureus Infections. 50th Annual Interscience Conference on Antimicrobial Agents and Chemotherapy (ICAAC), Boston MA, September 15, 2010 (oral presentation).

5. Ajao AO, Ranke T, Lee M, Robinson G, Furuno JP, Harris AD, Thom KA, Johnson JK. Detection of Multi-drug resistant Acinetobacter baumannii (MDR- AB) in Peri-Rectal Surveillance Cultures of Critically Ill Patients for Infection Control. 50th Annual Interscience Conference on Antimicrobial Agents and Chemotherapy (ICAAC), Boston MA, September 15, 2010.

6. Ajao AO, Johnson JK, Harris AD, Zhan M, McGregor JC, Thom KA, Furuno JP. Risk of acquiring Extended spectrum-β-lactamase (ESBL)-producing Klebsiella pneumoniae and Escherichia coli from prior room occupants in intensive care unit. Society for Healthcare Epidemiology of America (SHEA) Annual Meeting, Dallas, Texas, April 2011.

7. Ajao AO, Furuno JP, Harris AD, Lee M, Robinson G, Johnson JK. Prevalence and Molecular Characterization of Extended-Spectrum-β-lactamase (ESBL) and AmpC β-lactamase Genes among Klebsiella spp. and Escherichia coli Colonizing Isolates in Intensive Care Unit Patients. Society for Healthcare Epidemiology of America (SHEA) Annual Meeting, Dallas, Texas, April 2011.

8. Johnson JK, Thom KA, Harris AD, Morgan DJ, Ajao AO, Furuno JP. Carbapenem Resistance in Extended-spectrum Beta-lactamase (ESBL)-producing Klebsiella and Escherichia coli (ESBL-KE). Interscience Conference on

Antimicrobial Agents and Chemotherapy 51st Meeting. September, 2011. Chicago, IL. (Abstract # K-818)

Professional Society Memberships

2001-present American Public Health Association 2004 -present Society of Clinical Research Associates 2008-present American Society of Microbiology 2011 Society for Epidemiology Research

Honors and Awards

2009 Recipient of ASM/ICAAC Student and Post-Doctoral Fellows Travel Grant 2011 Recipient of SHEA Travel Award

Teaching Responsibilities

Spring 2008 Teaching Assistant, Biostatistical Computing (PREV 619) Fall 2008 Teaching Assistant, Introduction to Biostatistics (PREV 600)

ABSTRACT

Title of Dissertation: The Role of Persistent Environmental Contamination and Universal Glove and Gown Use in the Acquisition of Extended Spectrum Beta-Lactamase-Producing Enterobacteriaceae in Intensive Care Unit Patients.

Adebola O. Ajao, Doctor of Philosophy, 2011

Dissertation Directed by: Jon P. Furuno, Ph.D. Assistant Professor, Department of Epidemiology and Public Health

Background: Extended-spectrum β-lactamase (ESBL)-producing Gram-negative rods

(GNR) are emerging pathogens that are associated with considerable morbidity, mortality and costs among hospitalized patients. The association between persistent environmental contamination and acquisition of ESBL-GNR and the effectiveness of universal glove and gown in reducing the transmission of ESBL-GNR among Intensive care unit (ICU) patients has not been well established.

Objectives: The objectives of this dissertation are to evaluate the role of persistent environmental contamination using prior room occupant as a proxy in the acquisition of

ESBL-GNR and to evaluate the effectiveness of universal glove and gown use in reducing the incidence of ESBL-GNR in ICU patients.

Methods: A retrospective cohort study and a quasi-experimental study were conducted using patient data obtained from the central data repository and by laboratory analysis.

Peri-anal surveillance cultures were collected from all patients on ICU admission, weekly and at ICU discharge and clinical cultures were collected as medically indicated.

Inclusion criteria were ICU length of stay >48 hours and a peri-anal surveillance culture negative for ESBL-GNR on ICU admission. Multivariable logistic regression and segmented linear regression were used to analyze the data.

Results: The first study of 18,175 admissions to the University of Maryland Medical

Center (UMMC) medical ICU (MICU) and surgical ICU (SICU) between September 1,

2001 and June 30, 2009 showed that prior room occupants’ ESBL-Klebsiella and E. coli positive status is not associated with acquiring ESBL- Klebsiella and E. coli after adjusting for potential confounders (Adjusted Odds Ratio (AOR) = 1.39, 95%

Confidence Interval (CI) = 0.94 - 2.08). The second study of 6,089 admissions to the

MICU between July 1, 2005 and June 30, 2009 showed that universal glove and gown did not reduce acquisition of ESBL-GNR immediately (p =0.48) or long term (p =0.34).

Conclusions: Our study results suggest that environmental contamination may not play a significant role in the acquisition of ESBL-GNR at UMMC. Universal glove and gown was not effective at reducing the acquisition of ESBL-GNR immediately and long term at

UMMC.

The Role of Persistent Environmental Contamination and Universal Glove and Gown Use in the Acquisition of Extended Spectrum Beta-Lactamase-Producing Enterobacteriaceae in Intensive Care Unit Patients

by Adebola O. Ajao

Dissertation submitted to the faculty of the Graduate School of the

University of Maryland, Baltimore in partial fulfillment

of the requirements for the degree of

Doctor of Philosophy

2011

DEDICATION

This dissertation is dedicated to my loving husband Yomi Ajao, who supported me throughout my doctoral program. To my sweet daughter, Oluwadara Ajao, who lived with a part-time mummy for four years. To my wonderful parents, Prof. &Mrs. S.A.

Olanrewaju, my siblings and in-laws who supported me throughout this journey. I am sincerely grateful for all your love and support.

iii

Acknowledgements

I would like to express my gratitude to my primary dissertation mentor Dr. Jon P. Furuno

and all my committee members, Dr. Anthony D. Harris, Dr. Jennifer K. Johnson, Dr.

Jessina C. McGregor, Dr. Kerri A. Thom, and Dr. Min Zhan for their time, support and

guidance throughout my dissertation process.

I would like to thank the laboratory team, Gwen Robinson, Mary Lee, Tarah Ranke, Alex

Perdieu, Chrisley Pickens, Angela Stancil and Katarina Lincalis for all their hard work.

I would like to thank the data support team, Jingkun Zhu, Angela Comer and

Joseph Rosenberg for data extraction and data validation support.

I would also like to thank the entire faculty and staff of the Department of Epidemiology

and Public Health for the knowledge that they have impacted onto me. I would like to

appreciate the entire Division of Division of Genomic Epidemiology and Clinical

Outcomes for providing me with the opportunity to train as an epidemiologist and a

supportive environment to complete my dissertation work.

Table of Contents

Chapter Page

1. Introduction and Objectives……………………………………………...1 A. Introduction………………………………………………………….1 B. Research Questions, Hypotheses, and Specific Aims……………...... 3

2. Background……………………………………………………………...5 A. Overview………………………………………………………….....5 B. β-Lactams…………………………………………………………....5 C. Extended Spectrum β-Lactams…………………………………...... 9 D. History of Extended Spectrum β-Lactams……………………….....10 E. Epidemiology of ESBL-Producing Klebsiella and E. coli………….12 F. Clinical Significance of ESBL-Producing Klebsiella and E. coli in Hospitalized Patients…………………………………...... 13 G. Factors Associated with the Acquisition of ESBL-Producing Klebsiella and E. coli in Intensive Care Unit Patients……………...15 H. Role of the Environment in the Acquisition of ESBL-Producing Klebsiella and E. coli in Intensive Care Unit Patients……………...18 I. Infection Control Interventions to prevent the spread of ESBL Producing Klebsiella and E. coli in Intensive Care Unit Patients…..21 J. Summary of Research Significance………………………………...22

3. Research Design and Methods………………………………………….25 A. Role of Persistent Environmental Contamination in the Acquisition of ESBL-Producing Klebsiella and E. coli in ICU patients……………………………………………………...... 25 I. Study Site………………………………………………………...... 26 II. Study Design………………………………………………………..26 III. Description of Environmental Cleaning Procedures at UMMC...... 28 IV. Study Covariates………………………………………………...... 29 V. Sources of Data………………………………………………...... 32 VI. Laboratory Methods…………………………………………..……33 VII. Statistical Analysis…………………………………………………37

B. The Effect of Universal Glove and Gown on the Incidence of ESBL-Producing Enterobacteriaceae……………………………...39 I. Study Design…………………………………………………...…....40 II. Study Covariates……………………………………………...... 42 III. Sources of Data…………………………………………………...... 42 IV. Laboratory Methods………………………………………….....…..42 V. Statistical Analysis………………………………………………...... 42

4. Results: Dissertation Manuscript I……………………………….….…45

5. Results: Dissertation Manuscript II…………………………………….71

6. Discussion……………………………………………………………....93 A. Extended spectrum β-lactamase producing Gram-negative rods…..93 B. Prior room occupant’s ESBL-Klebsiella or E. coli positive status was not significantly associated with acquisition of ESBL-Klebsiella or E.coli……………………………………….....95 C. Universal glove and gown use was not associated with decreased acquisition of ESBL-Enterobacteriaceae in ICU patients…………97 D. Strengths and Limitations………………...………………………...99 E. Summary and Implications of the Research Questions…...……....102

7. Appendix...…………………………………………………………....104 8. References…………………………………………………………….105

List of Tables

Table I.1: List of Antibiotics in the β-Lactam Class of Antibiotics……………………...6

Table II.2: Description of Extended Spectrum β-Lacatamase Gene Families………….10

Table II.3: TEM Derived ESBLs in US Klebsiella and E. coli Isolates………………..11

Table II.4: SHV Derived ESBLs in US Klebsiella and E. coli Isolates………………...12

Table II.5: Studies of Gastrointestinal Colonization with ESBL-Producing……….…..17 Klebsiella and E. coli in Hospitalized Patient

Table II.6: Studies of Environmental Contamination with ESBL-Producing Klebsiella and E. coli among Hospitalized Patients………………….…...…20

Table III.1: Contingency Table for the Exposure and Outcome for Study 1…………...26

Table III.3: Definitions of Patient Characteristics Collected……………………….…..31

Table III.4: Calculation of the Charlson Comorbidity Index…………………………..32

Table III.5: Nucleotide Sequence of Oligonucleotides Used for PCR Amplification.....36

Table IV.1: Characteristics of Patients at Risk for Acquiring ESBL-Klebsiella

and E. coli………………………………………………………………….57

Table IV.2: Association between Acquisition of ESBL-Klebsiella and E. coli and Study Variables…………………………………………………….….60

Table IV.3: Predictors of ESBL-Klebsiella or E. coli Acquisition during ICU Stay in Multivariable Logistic Regression Model…………………………63

Table IV.4: PFGE profile for the 17 patients colonized with the same bacterial species as their prior room occupants……………………………………...64

Table V.1: Characteristics of the Study Population by Study Period…………….….....84

Table V.3 Segmented Linear Regression Model of the effect of universal glove and gown use on acquisition rate of EBSL-GNR and ESBL-Klebsiella in patients admitted to the MICU……………………..….90

vii

List of Figures

Figure II.1: Bush-Jacoby Functional Classification Scheme of β-Lactams……………...8

Figure III.1: Flow Chart of the Impact of Persistent Environmental Contamination on the Acquisition of ESBL-Producing Klebsiella and E. coli in ICU patients………………………………………………………………….....28

Figure III.2: Effect of Universal Glove and Gown Use on the incidence of ESBL-GNR Using Quasi-Experimental Study Design…………………..41

Figure IV. 1 Study Population and Exclusion Criteria …………………………….…..56

Figure IV.2: Transmission of ESBL-Klebsiella and E. coli…………………………....60

Figure V.1: Selection of Study Population……………………………………………..83

Figure V.2 Time series plot for ESBL-GNR and ESBL-Klebsiella acquisition rate July 2005 –June 2009………………………………………………....89

viii

Acronyms CCI: Charlson Comorbidity Index CDC: Centers for Disease Control and Prevention CDR: Central Data Repository CDS: Comobidity Disease Score CLSI: Clinical Laboratory Standards Institute ESBL: Extended Spectrum-β-Lactamase ESBL-KE: Extended Spectrum-β-Lactamase-Producing Klebsiella species and E.coli ESD: Environmental Service Department HAI: Hospital Acquired Infection HCW: Health Care Workers ICD-9: International Classification of Diseases, Ninth Revision ICU: Intensive Care Unit MDRO: Multi-Drug Resistant Organisms MIC: Minimum Inhibitory Concentration MICU: Medical Intensive Care Unit MRSA: Methicillin-resistant Staphylococcus aureus NHSN: National Healthcare Safety Network NNIS: National Nosocomial Infectious Disease Surveillance PFGE: Pulsed Field Gel Electrophoresis SICU: Surgical Intensive Care Unit UMMC: University of Maryland Medical Center VISA: Vancomycin-intermediate Staphylococcus aureus

1. Introduction and Objectives

A. Introduction

Extended-spectrum β-lactamase (ESBL)-producing Gram-negative rods (GNR)

are emerging pathogens that are associated with considerable morbidity, mortality and

costs among hospitalized patients in the United States and globally.1,2 ESBLs are plasmid-mediated enzymes that hydrolyze most β-lactam antibiotics making them ineffective.1 The gastrointestinal tract is the main reservoir for ESBL-GNR in humans.7-

10 Colonization with ESBL-GNR often precedes clinical infection of the same bacterial

strain.8,10 Infections caused by ESBL-GNR are difficult to identify and treat due to the limited options for empiric therapy. Thus, there is an increased likelihood that empirically

selected therapy may not provide coverage for the ESBL-GNR which subsequently increases the risk of treatment failure and poor outcomes such as death among infected patients.2,11 Patients with unidentified ESBL-GNR may serve as a reservoir and facilitate the persistent transmission of these antibiotic-resistant bacteria to other hospitalized patients.

Intensive care units (ICUs) have higher rates of antimicrobial resistance than

other areas of the hospital14 likely due to a high concentration of critically ill patients who

are less resistant to exogenous colonization, are in frequent close contacts with physicians

and nurses and a heavy use of empiric broad spectrum antibiotics that creates selective

pressure and promotes the emergence of resistant endogenous strains.14,54 The prevalence of ESBL-producing isolates in ICU patients with nosocomial infection in the United

States ranged from 5% in E. coli to 20% in K. pneumoniae in 2003 and continues to

rise.15,16 Given this increase and the limited effective antimicrobial drugs currently available or in development there is a need for more effective infection control methods to prevent the spread of ESBL-GNR among ICU patients.

Patients colonized or infected with ESBL-GNR often contaminate their immediate hospital environment. Environmental contamination with ESBL-GNR may persist despite disinfection of the hospital rooms creating a potential for cross- transmission of ESBL-GNR from environmental sources to patients12, 13 however, these studies were conducted in outbreak situations and the association between the contaminated environment and acquisition of ESBL-GNR was not established. Only one study has assessed the association between persistent environmental contamination and

acquisition of ESBL-GNR among hospitalized patients in a non-outbreak situation. This

study was limited by the small sample size and lack of molecular typing for bacterial

strain comparability. Therefore, there is need for more studies to establish the role of

contaminated environment in acquisition of ESBL-GNR in non-outbreak situations.

Under contact precaution, gowns and gloves have been recommended for contacts

with all patients colonized or infected with Methicillin-resistant Staphylococcus aureus

(MRSA), Vancomycin-resistant enterococci (VRE), Vancomycin-intermediate

Staphylococcus aureus (VISA) or Vancomycin-resistant Staphylococcus aureus

(VRSA).17 Studies that have evaluated the effectiveness of glove and gown use in

reducing colonization rate with antibiotic-resistant bacteria have been limited to VRE

studies. Some of these studies found that gown use in addition to glove use significantly

reduced VRE colonization rate18,19 while others found no effect on VRE colonization rate

in ICU patients.123 Universal glove and gown use is a form of contact precaution used for

contact with all patients and their environment regardless of the patient’s colonization or

infection status. However, universal glove and gown has not been evaluated for its’ effectiveness in reducing the transmission of ESBL-GNR among ICU patients.

ESBLs are produced by many Gram-negative rods but Klebsiella species

(Klebsiella pneumoniae and Klebsiella oxytoca) and Escherichia coli (E. coli) are the most common and clinically important Gram-negative bacteria that produce ESBLs and will henceforth be referred to as ESBL-KE.3-6 The objective of the proposed study is to

evaluate the role of persistent environmental contamination in acquisition of ESBL-KE

and to evaluate the effectiveness of universal glove and gown use in reducing the

incidence of ESBL-GNR in ICU patients.

B. Research Questions, Hypotheses, and Specific Aims

Research Question 1

Is the incidence of ESBL-KE higher in ICU patients admitted to rooms whose immediate

prior occupant was positive for ESBL-KE compared to ICU patients admitted to rooms whose immediate prior occupant was not positive for ESBL-KE?

Specific Aim 1: To determine if occupying a room whose immediate prior occupant was

positive for ESBL-KE is associated with a higher incidence of ESBL-KE in the ICU.

Hypothesis 1

Occupying a room whose immediate prior occupant was positive for ESBL-KE is

associated with a higher incidence of ESBL-KE in the ICU after controlling for patient characteristics such as age, ICU time at risk, ICU antibiotic use, co-morbid conditions and colonization pressure.

Rationale: Contamination of the hospital environment by patients with ESBL-KE has

been recognized, however the association between contaminated hospital environment

and acquisition of ESBL-KE in a non-outbreak situations has not been well established.

Understanding the role of the environment in the transmission of ESBL-KE among

hospitalized patients is necessary to optimize existing hospital infection control efforts.

Research Question 2

Does universal glove and gown use decrease the incidence rate of ESBL-producing

Enterobacteriaceae compared to contact precaution for colonized or infected ICU

patients in non-outbreak situations.

Specific Aim 2: To determine if universal glove and gown use is associated with a lower

incidence rate of ESBL-producing Enterobacteriaceae in ICU patients.

Hypothesis 2: Universal glove and gown use decreases the incidence rate of ESBL- producing Enterobacteriaceae in ICU patients compared to contact precaution for colonized or infected patients after adjusting for patient characteristics such as age, ICU antibiotic use, co-morbid conditions, and colonization pressure.

Rationale: Studies of glove and gown use has been limited to VRE and have produced

conflicting results. The effectiveness of universal glove and gown use in reducing the

transmission of ESBL-producing Enterobacteriaceae in non-outbreak situations has not

been evaluated.

II. Background

A. Overview

The objectives of this dissertation are to evaluate the role of persistent

environmental contamination (using prior room occupant ESBL-KE status as a proxy) in the acquisition of ESBL-KE and to evaluate the effectiveness of universal glove and gown use in reducing the incidence rate of ESBL-producing Enterobacteriaceae in ICU

patients. This study takes patients’ demographic and clinical characteristics into account

in understanding these associations.

B. β-Lactams

β-lactams are one of the major classes of antibiotics. Other classes of antibiotics

include aminoglycosides, glycopeptides, ketolides, lincosamides, macrolides,

oxazolidinones, quinolones, streptogramins, sulphonamides and tetracyclines. β-lactams

include penicillins, cephalosporins, carbapenems and monobactams (Table II.1).

Cephalosporins are further divided into first, second, third, fourth and fifth generation

cephalosporins.20

Table II.1: List of Antibiotics in the β-Lactam Class of Antibiotics

Antibiotic Antibiotics in this class Class β-Lactams Penicillins Narrow β-lactamase sensitive: Spectrum Benzylpenicillin, Clometocillin, Benzathine benzylpenicillin, Procaine benzylpenicillin, Azidocillin, Penamecillin, Phenoxymethylpenicillin, Propicillin, Benzathine phenoxymethylpenicillin, Pheneticillin Β-lactamase resistant: Cloxacillin Oxacillin, Meticillin, Nafcillin Extended Aminopenicillins:[Amoxicillin, Spectrum Ampicillin, Epicillin] Carboxypenicillins:[Carbenicillin, Ticarcillin, Temocillin] Ureidopenicillins, [Azlocillin, Piperacillin, Mezlocillin] Cephalosporins 1st Generation Cefazolin, Cefacetrile, Cefadroxil, Cefalexin, Cefaloglycin, Cefalonium, Cefaloridine, Cefalotin, Cefapirin, Cefatrizine,Cefazedone, Cefazaflur, Cefradine, Cefroxadine, Ceftezole 2nd Generation Cefaclor, Cefamandole, Cefminox, Cefonicid, Ceforanide, Cefotiam, Cefprozil, Cefbuperazone, Cefuroxime, Cefuzonam, Cephamycin Carbacephem 3rd Generation Cefixime, Ceftazidime, Ceftriaxone, Cefcapene, Cefdaloxime, Cefdinir, Cefditoren, Cefetamet, Cefmenoxime, Cefodizime, Cefoperazone, Cefotaxime, Cefpimizole, Cefpiramide, Cefpodoxime, Cefsulodin, Cefteram, Ceftibuten , Ceftiolene, Ceftizoxime, Oxacephem 4th Generation Cefepime, Cefluprenam, Cefozopran,Cefpirome, Cefquinome 5th Generation Ceftobiprole Carbapenems Biapenem, Doripenem, Ertapenem, Imipenem, Meropenem, Panipenem Monobactams Aztreonam, Tigemonam β- Sulbactam, Tazobactam, Clavulanic acid lactamase Inhibitors

β-lactams have a common element: a four-atom β-lactam ring in their molecular

structure. The β-lactam ring contains the antibacterial properties that inhibit bacterial cell

wall synthesis, weakening the cell wall and leading to bacterial cell lyses and death.21 β-

lactamases are enzymes produced by some bacteria that hydrolyze β-lactams making the bacteria resistant to this class of antibiotics. β-lactamases break open the β-lactam ring, deactivating the molecule's antibacterial properties. β-lactamases are produced by certain

Gram-positive bacteria such as Staphylococcus aureus but are widely distributed in

Gram-negative organisms.22 Commercially available β-lactamase inhibitors such as

clavulanic acid, sulbactam, and tazobactam can inhibit the action of some β-lactamases by binding irreversibly to these enzymes, thus enabling the bacteria to be susceptible to

β-lactams.

β-lactamases are classified according to two schemes: the Ambler molecular

classification scheme and the Bush-Jacoby functional classification scheme.22,23 Ambler’s

molecular scheme divides β-lactamases into four major classes, namely A to D based on

their protein morphology (nucleotide and amino acid sequence similarity). Classes A, C

and D enzymes contain serine in their active site. This serine is the active component in

the breaking of the β-lactam ring while class B enzymes need zinc for this action.22

The Bush-Jacoby functional classification scheme divides β-lactamases into four main groups (Groups 1-4) and several subgroups based on their phenotypic characteristics

(substrate and inhibitor functional similarities). Group 1 enzymes are cephalosporinases that are not inhibited by clavulanic acid. Group 2 enzymes are penicillinases, cephalosporinases, or both, that are inhibited by clavulanic acid. Group 3 are

penicillinases, cephalosporinases or carbapenemases that are not inhibited by clavulanic acid. Group 4 enzymes are penicillinases that are not inhibited by clavulanic acid (Figure

II.1). Group 2 is further divided into groups 2a to 2f. Group 2b enzymes are broad- spectrum β-lactamases that inactivate both penicillins and cephalosporins. Group 2b is further divided into groups 2be and 2br. Group 2be enzymes are extended-spectrum-β- lactamases capable of inactivating extended spectrum cephalosporins and monobactams.

Group 2br enzymes are penicillinases, which have diminished inhibition by clavulanic acid.23

Figure II.1: Bush-Jacoby Functional Classification Scheme of β-Lactams

BETA LACTAMS

Group 1 Group 4 CEPHALOSPORINASE PENICILLASE

Not Inhibited by clavulanic acid Not inhinited by clavulanic acid Group 2 Group 3 PENICILLASE, CEPHALOSPORINASE PENICILLASE, CEPHALOSPORINASE Inhibited by clavulanic acid and CARBAPENASE Not Inhibited by clavulanic acid

Group 2b Group 2a BROAD SPECTRUM Group 2c Group 2d Group 2e Group 2f Penicillase, PENICILLASES Cephalosporinase CARBENICILLINASE CLOXACILANASE CEPHALOSPORINASE CARBAPENAMASE Penicillins Penicillins, Penicillins, Penicillins, Cephalosporins Cephalosporins Carbenicillin Cloxacillin Monobactams Carbapenems

Group 2be Group 2br EXTENDED-SPECTRUM-β-LACTAMSES INHIBITOR-RESISTANT Penicillins, Narrow & Extended spectrum cephalosprorins, Monobactams Penicillins

C. Extended Spectrum β-lactamases (ESBLs)

Extended-spectrum β-lactamases placed in the Bush-Jacoby group 2be hydrolyze extended-spectrum cephalosporin with an oxyimino side chain including cefotaxime, ceftriaxone, ceftazidime and oxyimino monobactam but not cephamycin or carbapenems and are inhibited by clavulanic acid. ESBLs were first discovered in Germany in 1983.24

Genes that encode for ESBLs are frequently carried on plasmids that are transmissible from one organism to another. These plasmids that carry genes that encode for ESBLs also carry genes that encode resistance to other classes of antibiotics. This makes ESBL- producing GNR resistant to multiple classes of antibiotics.2,11 Most genes that encode for

ESBLs are derived from mutated β-lactamase genes; TEM-1, TEM-2 and SHV-1.1,6 The mutation alters the amino acid configuration around the active site of β-lactamases extending the spectrum of β-lactams that can be hydrolysed by these enzymes. However, several ESBLs that are not of TEM or SHV lineage have been recently described.1,25 The different families of ESBLs that have been described in the literature are outlined below

(Table II.2).

Table II.2: Description of Extended Spectrum β-lactamase Gene Families

ESBL Families Bacteria Geographic Locations TEM family Escherichia coli USA SHV family Klebsiella species Europe Proteus species Enterobacteriacae CTX-M family Escherichia coli South America Salmonella enterica Eastern Europe UK USA OXA family Pseudomonas aeruginosa Turkey Escherichia coli France PER Family (PER-1, PER-2) Pseudomonas aeruginosa Turkey Salmonella enterica South America Acinetobacter species Italy Escherichia coli France Klebsiella species Belgium Proteus mirabilis Korea Vibrio Cholera Alcaligenes faecalis Others: VEB, GES, IBC, Escherichia coli France SFO, BES-1, TLA-1 Klebsiella species Kuwait Pseudomonas aeruginosa China Enterobacter species South Africa Greece

D. History of Extended Spectrum β-lactamases in the United States

ESBL-producing bacteria was first reported in Europe five years prior to it’s’ emergence in the United States.24 The first report of ESBL producing bacteria in the

United States was of a TEM-derived ESBL obtained from an abdominal wound culture of a patient in Massachusetts, in February of 1988.26 This was followed by reports of several outbreaks of TEM-derived ESBLs namely TEM10, TEM 12 and TEM 2627-33 all coinciding with increased use of ceftazidime. Other TEM-derived ESBLs such as TEM-

6, TEM-28 and TEM-43 were described in later outbreaks.34,35 TEM-derived ESBLs have point mutations at one or more amino acid positions 104, 164, 182, and 24036 (Table

II.3). SHV-derived ESBLs appeared in the United States in 1990.37 SHV-derived ESBLs that have been identified in the United States include SHV-5, SHV-7, SHV-8, SHV-12,

SHV-29, and SHV-47.38-40 SHV-derived ESBLs present in the United States have amino

acid substitution at a total of seven positions compared to four positions for TEM-derived

ESBLs36 (Table II.4). The CTX-M-derived ESBLs are non-TEM and non-SHV-derived

ESBLs that emerged in the late 1980s, a few years after the introduction of cefotaxime.25

CTX-M-derived ESBLs are mostly found in the community and there have only been a few reports in the United States. CTX-M-derived ESBLs were first identified in 9 E. coli

isolates from a 5 state ESBL surveillance study in 2001-2002.41 This was followed by

reports of CTX-M ESBL from Salmonella enterica in 2003 and Shigella sonnei in 2004.

Both the Salmonella and Shigella isolates were obtained from stool samples.42,43 Another

report of Shigella flexneri and Shigella sonnei isolates obtained from 1999 to 2007

identified 6 isolates that carried the CTX-M-14 and CTX-M-15 ESBL genes.44 A recent

report of a CTX-M-15 producing E. coli isolated from a middle aged woman who

developed fatal urosepsis may be an indication that CTX-M-derived ESBL is spreading in the United States.124

Table II.3: TEM Derived ESBLs in US Klebsiella and E. coli isolates

TEM-Derived ESBLs Year Identified Amino acid positions and substitutions in the US 104 164 182 240 TEM-1 Glu Arg Met Glu TEM-10 1988 Glu Ser Met Lys TEM-12 1988 Glu Ser Met Glu TEM-26 1988 Lys Ser Met Glu TEM-28 1992 Glu His Met Lys TEM-6 1993 Lys His Met Glu TEM-43 1992-1996 Lys His Thr Glu

Table II.4: SHV Derived ESBLs in US Klebsiella and E. coli isolates

SHV Derived ESBLs Year Identified Amino acid positions and substitutions in the US SHV-1 8 35 43 179 195 238 240 SHV-5 1993 Ile Leu Arg Asp Thr Gly Glu SHV-7 1993 Ile Leu Arg Asp Thr Ser Lys SHV-8 1990 Phe Leu Ser Asp Thr Ser Lys SHV-12 1996 Ile Leu Arg Asn Thr Gly Glu SHV-29 NDa Ile Gln Arg Asp Thr Ser Lys SHV-46 1998 Ile Gln Ser Asp Thr Ala Glu a: No data on the year identified. Reported in 2001

E. Epidemiology of ESBL-Producing Klebsiella and E. coli

ESBL-KE have caused serious outbreaks of infections in hospitalized patients both in the United States and worldwide.8,10,27,28,30,31,45,46 ICUs have been the epicenter of most of these outbreaks; however, outbreaks due to ESBL-KE have been reported in non-

ICU inpatients and nursing home patients.30,31,47 Colonization and infection with ESBL-

KE have also been reported in outpatient communities outside of the United States.48-52

The prevalence of ESBL-KE continues to rise among hospitalized patients in the

United States. In the United States, resistance to third-generation cephalosporins specifically ceftazidime is used to screen for ESBL production in bacterial isolates, however, resistance to ceftazidime could be due to other resistance mechanisms besides

ESBLs. National Healthcare Safety Network (NHSN) data from inpatient isolates from

300 hospitals in the United States indicated that resistance to ceftazidime increased from

6.9% in 1999 to 12.3% in 2006 (p<0.0001) for K. pneumoniae isolates and from 2.0% in

1999 to 2.7% in 2006 for E.coli isolates (p<0.001).53 A cross-sectional study of ICU patients at the University of Maryland Medical Center (UMMC) also indicated that

patients colonized with ceftazidime-resistant gram-negative bacilli increased from 18.8%

in 2003 to 31.4% in 2006 (p<0.01%). This increase was also largely driven by

ceftazidime-resistant Klebsiella isolates which accounted for 6.4% of isolates in 2003 and

22.8% of isolates in 2006.55 The prevalence of ESBL-KE is highest in ICU patients due to the presence of factors that promote the emergence of resistance. The ICU consists of a high concentration of critically ill patients who are less resistant to colonization by exogenous organisms, have frequent close contacts with physicians and nurses under urgent conditions, and have a heavy use of empiric broad spectrum antibiotics that creates selective pressure and promotes the emergence of resistant endogenous strains.14,54

F. Clinical Significance of ESBL-Producing Klebsiella and E. coli in Hospitalized

Patients

The gastrointestinal tract is the main reservoir for ESBL-KE in humans.

Gastrointestinal colonization has been shown to precede most infections caused by

ESBL-KE in hospitalized patients.8,10 Pena et al. identified prior carriage of ESBL-K.

pneumoniae in the digestive tract as an independent risk factor for ESBL-K. pneumoniae

infection or colonization (OR =3.4, 95% CI = 1.1-10.4) after adjusting for age, surgery,

severity of disease, days of invasive devices, and days of prior antibiotic therapy.10

Infections caused by ESBL-KE include respiratory tract infections, urinary tract

infections, bloodstream infections, meningitis and surgical wound infections.10,16,56

Infections due to ESBL-producing GNR are difficult to treat due to their resistance to

multiple classes of antibiotics which limits the empiric therapy options for these

infections.2,11 In addition, clinically relevant ESBL-mediated resistance is not always

detected in routine susceptibility testing. Depending on which antibiotics are used to screen for ESBLs, ESBL-KE strains may appear susceptible to some extended-spectrum

cephalosporin in vitro but these antibiotics may be ineffective against these bacteria in

vivo.57 This is likely to lead to use of inappropriate empiric therapy which increases the

risk of treatment failure and death among infected patients.58-60 Kim et al. reported that

54.5% of patients with bacteremia due to ESBL producing K. pneumoniae received

inappropriate therapy before the report of culture results compared to 3.4% of patients

with bacteremia due to non-ESBL producing K. pneumoniae (p = 0.001).59 In another

study, Du et al. showed that inappropriate antibiotic therapy was associated with

treatment failure (OR = 8.3, 95% CI = 1.1-11.2) and treatment failure was the only

independent predictor of hospital mortality among bacteremic patients due to ESBL-KE

(OR = 16.4 95% CI = 3.5-74.7).58 Crude mortality has been reported to be higher in

patients with bloodstream infections due to ESBL-KE than in patients with the non-ESBL

producing strains of Klebsiella spp. or E. coli.56,59-61 Ho et al. reported that mortality was

18% in patients with bacteremia due to ESBL-producing E. coli compared to 7% in

patients with bacteremia due to non-ESBL-producing E. coli (p=0.04).62 Ariffin et al. also

reported that sepsis related mortality was significantly higher in pediatric patients

infected with ceftazidime-resistant K. pneumoniae (50%) than in patients infected with

ceftazidime-susceptible K. pneumoniae (13%) (p=0.02).60 Carbapenems are the

treatment of choice for serious ESBL-KE infections; however, carbapenem-resistant isolates have also been reported.1

G. Factors Associated with the Acquisition of ESBL-Producing Klebsiella and E. coli

in ICU patients

Factors associated with nosocomial infection due to ESBL-KE have been well

studied in both outbreak and non-outbreak situations. Adult patients are at a greater risk of acquiring a nosocomial ESBL-KE infection if they are severely ill, have previously used a third-generation cephalosporin, had a prolonged hospital stay, and had a prolonged presence of invasive devices such as a urinary catheter, central venous catheter, arterial catheter, or endotracheal tube.56,60,61,63,64 However, factors associated with ESBL-KE

colonization in hospitalized patients have not been well studied despite their importance

in infection control. As noted by Lucet et. al., factors associated with colonization may

differ from factors associated with infections and their control may require different

interventions.8 Gastrointestinal colonization with ESBL-KE often precedes clinical infections of the same bacterial strain.8,10 In addition, patients colonized with unidentified

ESBL-KE may serve as a reservoir and facilitate the persistent transmission of these

bacteria to other hospitalized patients.

Few studies have analyzed factors associated with gastrointestinal colonization

with ESBL-KE among hospitalized patients (Table II.5). These studies identified age >60

years (OR=1.79, 95% CI 1.24-2.60), piperacillin-tazobactam (OR=2.05, 95% CI: 1.36-

3.10), vancomycin (OR= 2.11, 95% CI 1.34-3.31), chronic disease score (OR=1.15; 95%

CI: 1.04-1.27)65, duration of hospitalization (OR= 1.11, 95% CI 1.02-1.21)66,

hemodialysis (OR =10.7, 95% CI: 1.4-81)67, duration of urinary catheterization (OR: 3.5;

95% CI 1.2-10.3) and duration of mechanical ventilation (OR: 4.6; 95% CI 1.1-19.3)9 to be independently associated with ESBL-KE colonization. However, none of these studies

identified patients who were negative for ESBL-KE on hospital or ICU admission and

followed them up for acquisition of ESBL-KE during their hospital or ICU stay. The only

study that followed up patients who were negative for ESBL-KE upon SICU admission

for ESBL-KE acquisition was underpowered to perform a multivariable analysis.

Nevertheless, number of gastrointestinal operations during the hospital stay, length of

stay, and nursing workload were statistically associated with acquisition of ESBL-KE at

the bivariate analysis level in this study.7 In nursing home patients, factors associated

with gastrointestinal colonization with ESBL-KE on ICU admission included poor functional status, presence of gastrostomy tube, and prior receipt of ciprofloxacin or trimethoprim-sulfamethoxazole.47

Table II.5: Studies of Gastrointestinal Colonization with ESBL-Producing Klebsiella and E. coli in Hospitalized Patients Author, Study Design Study Surveillance Independent Other factors Year Setting Population Culture risk factors adjusted for Country Sample size Method associated with in the ESBL-KE multivariable colonization analysis Harris Prospective 5,209 Peri-anal Age >60 years Cefepime, AD. 2007 cohort study patients swabs (OR= 1.79, 95% Imipenem (Non-outbreak) admitted to CI 1.24-2.60), USA the MICU piperacillin- and SICU tazobactam (OR= 2.05, 95% CI: 1.36-3.10), vancomycin (OR= 2.11, 95% CI 1.34-3.31), chronic disease score (OR= 1.15; 95% CI: 1.04- 1.27) Bisson G. Case control 59 patients Rectal swab Duration of Age, 2002 study admitted to or hospitalization Malignancy, (Non-outbreak) ICU and stool (OR= 1.11, 95% Antibiotic days USA non-ICU specimen CI 1.02-1.21). ward D’Agata Prospective 84 patients Rectal, Hemodialysis Duration of E. 1998 cohort study admitted to inguinal fold (OR =10.7, 95% prior (Non-outbreak) the SICU throat swab. CI: 1.4-81) hospitalization USA Nasogastric >=2 weeks in fluid the last year Pena C. Prospective 188 patients Rectal swab Number of days Clinical 1997 cohort study admitted to of urinary severity at (Outbreak) the ICU catheterization admission, Spain (OR= 3.5, 95% venous and CI 1.2-10.3) and arterial mechanical catheterization, ventilation (OR= parenteral 4.6, 95% CI 1.1- nutrition, 19.3) previous antibiotic therapy Soulier Prospective 59 patients Stool sample None Multivariable A. 1995 cohort study admitted to analysis was (Non-outbreak) the SICU not performed France

H. Role of the Environment in the Acquisition of ESBL-Producing Klebsiella and

E. coli in ICU Patients

Most of the data that exists on contamination of hospital environment by ESBL-

KE have been from studies performed in outbreak situations. Despite this limitation,

these studies have demonstrated that patients colonized or infected with ESBL-KE often

contaminate their immediate environmental surfaces. Epidemic strains of ESBL-

producing Klebsiella spp. have been isolated from colonized or infected patients’ bath,45

bedsheets, bowls, sinks countertops,12,13 and soap68 during outbreaks. A few studies have

even suggested that the removal of the contaminated environmental source ended the

outbreaks. For example, the end of the ceftazidime-resistant K. pneumoniae outbreak in

an obstetrics ward in Paris, France was attributed to the removal of the contaminated

ultrasound coupling gel.69 Klebsiella spp. have been known to survive from weeks to months especially in moist areas.12,70 In some outbreak studies, contamination of the hospital environmental surfaces by ESBL-producing K. pneumoniae and K. oxytoca persisted despite disinfection and cleaning of hospital rooms’ surfaces after patients’ discharge.12,13 Following an outbreak, Kac et al. performed comparative PFGE analysis

and identified closely related PFGE patterns between environmental strains and clinical

strains for ESBL-producing K. oxytoca 3 months after the patients with the identical

strains have been discharged.12 The persistent recovery of ESBL-producing Klebsiella

spp. from environmental surfaces after colonized patients have been discharged suggests

that there is potential for cross-transmission of ESBL-producing Klebsiella spp. from environmental sources.12,71 Although the contamination of hospital environment by

patients colonized or infected with ESBL-GNR has been recognized, the association

between contaminated hospital environment and acquisition of ESBL-GNR in non-

outbreak situations has not been established. Only a few studies have described isolation

of ESBL-GNR from environmental surfaces of colonized or infected patients outside of

outbreak situations (Table II.6).71-73 However, none of these studies established an

epidemiological link between patients’ and environmental ESBL-positive bacterial samples.

Colonization and infection status of prior room occupants has been used as a low

cost proxy for persistent environmental contamination and has been shown to be

associated with acquisition of MRSA, VRE, Clostridium difficille, MDR-P. aeruginosa

and Acinetobacter baumannii among ICU patients.74,109,122 However, these studies did not

perform molecular typing for bacteria strain comparison. Specifically these studies did

not quantify the proportion of acquired bacterial isolates that were genetically similar to

bacterial isolates obtained from immediate prior room occupants. Our study uses Pulsed

Field Gel Electrophoresis (PFGE) to assess the extent of genetic similarity between

ESBL-KE isolates obtained from prior room occupants and ESBL-KE isolates acquired during the ICU stay to make a stronger case for causality if an association exists.

Table II.6: Studies of Environmental Contamination with ESBL-Producing Klebsiella and E. coli among hospitalized patients Author, Study Design Population Organisms tested Study Result Year Tande Cross- Environmental and ESBL producing 1) 63% of adults and D. 2009 sectional stool samples E.coli, Klebsiella 100% of all childrens’ Study collected from 39 Pneumoniae, stool samples tested children, 44 staff Klebsiella positive for ESBL- in an orphanage in Oxytoca, entrobacteriaceae. Mali Citrobacter 2) 100% of freundii environmental samples tested positive for ESBL-entrobacteriaceae 3) Did not assess genetic similarity between colonizing and environmental strains for ESBL-KE Touati Prospective 150 Environmental ESBL producing 1) 9% environmental A. 2008 Cohort surface samples Klebsiella contamination Study collected from 2 pneumoniae and 2) Did not assess genetic hospitals Enterobacter similarity between Cloacae colonizing and environmental strains for ESBL-KE D’Agata Prospective 233 environmental ESBL producing 1) 9% environmental 1999 Cohort study samples, 284 Enterobacter contamination samples from Cloacae, 2) 2% of gown and hands and gowns Acintobacter hands contaminated of SICU nurses, baumanii, 3) 25% of patients 333 SICU patients Pseudomonas colonized aeruginosa, 4) Epidemiological link Serratia spp., was not established Klebsiella between patient and Oxytoca, environmental samples Stenotrophomonas for ESBL-KE but was spp. established for other entrobacteriaceae (E. Cloacae, P. aeruginosa, and Stenotrophomonas spp.)

I. Infection control interventions to prevent the spread of ESBL-Producing

Klebsiella spp. and E. coli in ICU Patients

Most previous studies assessing interventions to control the acquisition of ESBL-

GNR have been conducted in outbreak situations. Data from these studies are of limited value because often during outbreaks, several interventions are applied simultaneously;

thus the effectiveness of each intervention on infection control is difficult to elucidate.75

Strategies used to control previous outbreaks have included antibiotic control polices

(restriction on the use of extended-spectrum cephalosporin), surveillance culturing of patients, the environment and healthcare workers (HCWs), cohorting of colonized and

infected patients, use of contact isolation precautions for colonized and infected patients,

and increased emphasis on proper hand hygiene.28,30,31,68,76-78

Contact isolation with glove and gown use is a preventive methodology

recommended by infection control experts for patients colonized or infected with MRSA,

VRE, VISA, or VRSA.17 Studies that have evaluated the effectiveness of glove and gown

use in reducing colonization rate with antibiotic-resistant bacteria have been limited to

VRE studies. Some of these studies found that gown use in addition to glove use

significantly reduced VRE colonization rate18,19 while one study found no effect on VRE

colonization rate in ICU patients.123

At UMMC, under contact precaution, gowns and gloves are worn for contact with

all patients colonized or infected with MRSA, VRE, and other MDROs and their

environments. Universal glove and gown use is a form of contact precaution used for

contacts with all patients and their environment regardless of the patient’s colonization or

infection status. Universal glove and gown has the advantage that it may prevent

transmission of multiple antibiotic-resistant and susceptible bacteria at the same time without requiring active surveillance for each bacterial species. Universal glove and gown use was instituted at the UMMC MICU in July 2007 and at the Surgical ICU

(SICU) in February 2008 and it is currently being evaluated for its effectiveness in reducing transmission of MRSA and VRE in ICU patients. However, universal glove and gown has not been evaluated for its’ effectiveness in reducing the transmission of ESBL-

GNR among ICU patients.

Infection control monitoring and control charting of clinical cultures for ESBL-

KE at UMMC MICU and SICU from 2001 to 2005 showed no evidence of an outbreak.

However 2% of patients were colonized on ICU admission and 3% of patients acquired

ESBL-KE during their ICU stay indicating endemicity of these bacteria at UMMC.80,81

J. Summary of Research Significance

The first study evaluates the impact of persistent environmental contamination in acquisition of ESBL-KE while taking patients’ demographics and clinical characteristics into consideration in a large cohort of patients admitted to the MICU and SICU of a tertiary healthcare facility. In addition, this study quantifies the proportion of acquired

ESBL-KE isolates that are genetically similar to the ESBL-KE isolates obtained from the immediate prior room occupants. Contamination of hospital environment by patients colonized or infected with ESBL-KE has been recognized; however, the association between contaminated hospital environment and acquisition of ESBL-KE has not been well established. Understanding the role of the environment in the transmission of ESBL-

KE among hospitalized patients is necessary to optimize existing hospital infection

control efforts.

It is unclear if the current standard of environmental cleaning of rooms is

adequate for preventing the transmission of ESBL-KE in ICU patients. If this study

shows that patients admitted to rooms whose prior occupant was colonized with ESBL-

KE acquire the same bacteria strain despite environmental cleaning, then current

environmental cleaning practices may need to be re-assessed. The result of this study may

help to emphasize the importance of adequate environmental cleaning and may draw

attention to the use of environmental sampling for monitoring the level of ESBL-KE in

the hospital environment as an infection control practice.

The second study of this proposal evaluates the effectiveness of universal glove

and gown use in decreasing the incidence rate of ESBL-GNR in a large cohort of patients

admitted to the MICU of a tertiary care facility. The prevalence of ESBL-GNR continues to rise in ICU patients in the United States despite current infection control and antibiotic stewardship interventions. Yet, there is limited data on optimal infection control methods to prevent the transmission of ESBL-GNR in non-outbreak situations. The effect of universal glove and gown use on transmission of ESBL-GNR among ICU patients has not been evaluated.

If universal glove and gown use is effective in preventing the transmission of

ESBL-GNR, it may be a novel method to reduce the incidence of ESBL-GNR and its associated poor healthcare outcomes in ICU patients. Universal glove and gown has the added advantage that it may protect against multiple species of antibiotic-resistant and

susceptible bacteria without the need to actively identify patients who are colonized with each type of bacteria.

III. Research Design and Methods

A. Role of Persistent Environmental Contamination in the Acquisition of ESBL-

Producing Klebsiella and E. coli in ICU Patients

Specific Aim: To determine if occupying a room whose immediate prior occupant was

positive for ESBL-KE is associated with a higher incidence of ESBL-KE in the ICU.

Overview: This study evaluated the impact of persistent environmental contamination in

the acquisition of ESBL-KE in ICU patients. The primary exposure of interest was

ESBL-KE positive status of the immediate prior room occupant as illustrated in Table

III.1. ESBL-KE positive status was determined using peri-anal and clinical cultures. The

exposed group consisted of patients for whom the immediate prior occupant of their

hospital room was positive for ESBL-KE. The unexposed group consisted of patients for whom the immediate prior occupant of their hospital room was negative for ESBL-KE.

The primary outcome of interest was ESBL-KE acquisition during the ICU stay as

determined by peri-anal and clinical cultures. We defined patients who acquired ESBL-

KE as those who have ESBL-KE negative peri-anal culture at ICU admission but have a

subsequent ESBL-KE positive peri-anal or clinical culture on the same ICU admission.

ESBL-KE positive status of prior room occupant was used as a proxy for persistent

environmental contamination. The use of this proxy is reasonable because patients

colonized or infected with ESBL-KE often contaminate their immediate environment and

ESBL-producing Klebsiella spp. have been shown to persist on hospital environmental

surfaces despite disinfection and cleaning of the hospital room after patients’

discharge12,13,71. We determined the proportion of acquired ESBL-KE isolates that are

genetically the same or similar to the isolates obtained from the immediate prior room

occupants. Genetic similarity was defined as isolates with greater than 80% similarity in

PFGE band pattern.80,81,82

Table III.1: Contingency Table: Exposure and Outcome for Study 1

OUTCOME OF INTEREST

EXPOSURE Acquired ESBL-KE Did not acquire OF INTEREST ESBL-KE Prior room occupant positive for ESBL-KE Prior room occupant negative for ESBL-KE

I. Study Site

The study site was the UMMC MICU and SICU. UMMC is a 729-bed, urban

tertiary-care hospital in Baltimore, Maryland. The MICU was a 10-bed private room unit

that became a 29-bed private room unit in April, 2006. The MICU provides care to adult

patients with acute or potential life-threatening medical conditions. The SICU is a 19-bed

private room unit that provides care to adults with solid organ transplantation, abdominal, genitourinary, orthopedic and otolaryngologic surgery.

II. Study Design

A retrospective cohort study was performed using data from patients admitted to

the UMMC MICU and SICU from September 1, 2001 to June 30, 2009. Patients admitted

to the UMMC MICU and SICU have obtained peri-anal cultures routinely as part of an

ongoing VRE active surveillance program since 1995. Peri-anal cultures were obtained

using Staplex II swabs (Staplex, Etobiocoke, Ontario, Canada) within 48 hours of ICU

admission, weekly, and at ICU discharge. Compliance with obtaining peri-anal cultures

was approximately 90%. All peri-anal cultures were frozen in trypic soy broth containing

15% glycerol in -80°C freezer to allow for future recovery of organisms. Recovery of

organisms from these frozen swabs has been shown to be between 91% and 98% in prior

studies.83,84 All frozen peri-anal cultures were thawed and tested by the laboratory team in the research laboratory of my co-mentor, Dr. J. Kristie Johnson at University of Maryland

Baltimore (UMB) for ESBL production.

Admission peri-anal cultures were used to identify patients colonized with ESBL-

KE at ICU admission, while weekly and discharge peri-anal cultures were used to

identify patients who acquired ESBL-KE during their ICU stay. Patients who did not

receive admission peri-anal cultures were excluded from the study population as their

ESBL-KE positive status could not be established upon ICU admission. Similarly

patients that did not receive follow-up surveillance cultures were excluded from the study

population because their ESBL-KE acquisition status could not be established. Patients who were colonized with the ESBL-KE upon ICU admission and patients who stayed in the ICU for less than 48 hours were also excluded from the study population as they were at a reduced risk for acquiring ESBL-KE. The study population consisted of all patients admitted to the MICU and the SICU during the study period that had a negative ESBL-

KE peri-anal culture on ICU admission, stayed in the ICU for at least 48 hours and thus were at risk for acquiring ESBL-KE (Figure III.1). The exact timing of ESBL-KE acquisition was unknown; therefore the time interval from ICU admission to the first positive peri-anal or clinical culture was used as a proxy for time of ESBL-KE acquisition for the study population. Among the patients who acquired ESBL-KE during

their ICU stay, only the first positive peri-anal culture was included in the analysis.

Patients with multiple admissions to the ICUs during the study period were allowed to enter the cohort of at risk patients multiple times as long as they were not colonized with

ESBL-KE on the index ICU admission. At-risk patients who changed rooms during their

ICU stay accounted for more than one ICU admission as long as they were still at risk upon moving to the new room. The UMB Institutional Review Board (IRB) approved this study for the duration of the study period.

Figure III.1: Flow Chart of the Impact of Persistent Environmental Contamination on Acquisition of ESBL-Producing Klebsiella spp. and E. coli in ICU Patients.

III. Description of Environmental Cleaning Procedures at UMMC

Terminal cleaning (disinfection and cleaning of hospital room surfaces after patient discharge) was performed by staff of the Environmental Service Department

(ESD) at UMMC. ESD staff used a 12-step cleaning process in accordance with CDC

guideline to ensure adequate cleaning and disinfection of hospital room surfaces. The process included proper hand hygiene, replenishing soap and alcohol-based hand gel dispensers, emptying waste receptacles, cleaning and dusting surfaces, spot cleaning walls, cleaning and disinfecting all furniture and bathroom surfaces, and dust mopping floors. Quaternary germicidal cleaner, SaniMaster IV was used for all environmental disinfection.85

IV. Study Covariates

Patient characteristics were collected to identify potential confounding variables.

These variables included both demographic and clinical characteristics. Patient’s

demographic information included age, and sex collected at ICU admission. Patients’ clinical characteristics included length of ICU stay, length of ICU stay at risk, length of

ICU stay of prior room occupant, ICU antibiotic use, individual comorbidities, aggregate co morbid indices and colonization pressure. Total length of ICU stay, length of ICU stay at risk and length of ICU stay of prior room occupant were evaluated separately. Studies have shown that the length of ICU or hospital stay was independently associated with

ESBL-KE colonization.7,66 Information on specific antibiotics used in the ICU was

collected and evaluated separately.56 Individual antibiotic use was classified into the

major classes of antibiotics (Table III.2). Studies have shown that the previous antibiotic

treatment with certain classes of antibiotics and specific antibiotics such as

aminoglycosides, penicillin (piperacillin), cephalosporins (ceftazidime), glycopeptides

(vancomycin), sulphonamides (trimethoprim-sulphamethoxazole) and quinolones

(ciproflaxin) were associated with ESBL-KE colonization.47,65,88 Individual comorbidities

were determined using the International Classification of Diseases, Ninth Revision (ICD-

9) at ICU discharge. Aggregate comorbidity score, the Charlson Comorbidity Index

(CCI) was calculated using ICU discharge ICD-9 codes for comorbid conditions.89

Colonization pressure was defined as the average of the daily proportion of patients positive for ESBL-KE during each patient’s ICU time at risk and was calculated separately for the MICU and the SICU.90 Colonization pressure was calculated by

dividing the number of ESBL-KE positive patients by the number of patients in the unit

daily. The average of the daily colonization pressure was calculated for the period of time

that each patient was at risk for ESBL-KE. Patients contributed to colonization pressure if

they were positive for ESBL-KE on ICU admission or they acquired ESBL-KE during

their ICU stay. Patients positive for ESBL-KE were assumed to be positive from the day

of positive culture collection until discharge. The definitions of all study variables are listed below. (Table III.3)

Table III.3: Definitions of Patient Characteristics Collected Variables Definitions Demographic Information Age, Sex Total length of ICU stay Number of days from ICU admission to ICU discharge Length of ICU stay of Number of days from ICU admission to discharge for prior prior room occupant room occupant Length of time at risk 1. Number of days from ICU admission to ESBL-KE positive peri-anal or clinical culture for patients who acquired ESBL-KE. 2. Number of days from ICU admission to discharge for patients who did not acquire ESBL-KE. ICU antibiotic use 1. Individual antibiotics used during the time at risk in the ICU for the index admission was collected. 2. Individual antibiotics was classified into their respective classes namely as Aminoglycosides, β- lactams-Penicillin, 1st, 2nd, 3rd, 4th generation cephalosporin, Monobactams, Carbapenem, Glycopeptides, Ketolides, Lincosamides, Macrolides, Oxazolidinones, Quinolone, Streptogramins, Sulphonamides, Tetracyclines and Metronidazole. Individual Individual comorbid conditions present at ICU discharge as comorbid conditions determined by ICU discharge ICD-9. Examples of co morbid conditions include renal disease, AIDS, malignancy, liver disease, coronary artery disease, and diabetes. Aggregate Comorbidity Aggregate comorbidity score was calculated using ICU Charlson Comorbidity discharge ICD-9 codes for comorbid conditions (Table III.4). Colonization Pressure Colonization pressure was defined as the average of the daily proportion of patients positive for ESBL-KE during each patient’s ICU time at risk.

Table III.4: Calculation of the Charlson Comorbidity Index (Deyo et al., 1992)

Condition Assigned Weight Myocardial infarct 1 Congestive heart failure 1 Peripheral vascular disease 1 Cerebrovascular disease 1 Dementia 1 Chronic pulmonary disease 1 Connective tissue disease 1 Ulcer disease 1 Mild liver disease 1 Diabetes 1 Hemi paraplegia 2 Moderate or severe renal disease 2 Diabetes with end organ damage 2 Any tumor 2 Leukemia 2 Lymphoma 2 Moderate or severe liver disease 3 Metastatic sold tumor 6 AIDS 6

V. Sources of Data

All patient data were collected through the UMMC Central Data Repository

(CDR) and through laboratory testing. The UMMC CDR is a relational database

containing patients’ administrative, pharmacy, admission-discharge-transfer, and clinical data maintained by the Information Technology Group of the University of Maryland. A sample of the demographic, microbiology, and pharmacy data from the UMMC CDR were validated against patients’ paper medical records for this study and previous research studies and the data have a positive and negative predictive value greater than

98%.91-93 Data from the UMMC CDR has been used extensively in epidemiologic studies

of antibiotic resistance, which have been published in the peer-reviewed literature.65,94-98

Information on demographics, antibiotic use, pre-existing comorbid conditions, and aggregate comorbidity scores were obtained from the UMMC CDR. Microbiology testing of the peri-anal cultures were used to determine ESBL-KE positive status of prior room occupants and patients who acquired ESBL-KE during their ICU stay. Length of time at risk was calculated using data from the CDR and microbiology result of peri-anal and clinical cultures. Molecular testing of all ESBL-KE isolates was used to determine the extent of genetic similarity of all ESBL-KE isolates obtained from prior room occupants and from patients who acquired ESBL-KE during their ICU stay.

VI. Laboratory Methods

Detection of ESBL-KE from Peri-anal Cultures: Frozen peri-anal cultures were removed from the freezer and innoculated onto a MacConkey agar plates (Remel,

Lenexa, KS, USA) containing 1 μg/ml ceftazidime. The MacConkey with ceftazidime agar plates were incubated at 37 degrees C for 24 hours. MacConkey agar contains bile salt and crystal violet that inhibits the growth of most Gram-positive bacteria but allows growth of Gram-negative bacteria. Lactose fermenting bacteria such as Klebsiella spp. and E.coli utilizes the lactose in the medium to produce acid, which lowers the pH of the agar to below 6.8 and the bacteria appears as pink colonies. Lactose fermenting colonies that grew on MacConkey with ceftazidime agar plates were swabbed onto a TSA with

5% sheep blood agar plate (BD, Sparks, MD, USA) and incubated for 24 hours for oxidase test (BD, Sparks, MD, USA). At the same time, these colonies were placed into

5ml saline solution and dropped into the API 20E identification strips (bioMerieux;

Durham, NC), and incubated for 24 hours following API 20E instructions. Klebsiella species and E.coli were identified by a color change in the API 20E identification strips.

Antimicrobial Susceptibility Testing: Antibiotic susceptibilities of Klebsiella species and E.coli were performed on the Muller-Hinton agar plates using the Kirby-Bauer disk diffusion test method as recommended by the Clinical Laboratory Standards Institute.99

Pure colonies of Klebsiella species and E.coli were plated onto Muller-Hinton agar plates

(BBL Microbiology Systems; Cockeysville, Maryland) using the quarter-turn method.

Antibiotic disks containing containing 30 μg of ceftazidime and 30 μg cefotaxime with and without 10 μg of clavulanic acid were dropped on to the Muller-Hinton agar plates and incubated for 24 hours. An area of clearing around the antibiotic disk (zone of inhibition) indicated that the Klebsiella or E. coli was susceptible to ceftazidime or cefotaxime. A greater than 5mm difference in the zone of inhibition for either ceftazidime or cefotaxime tested in combination with clavulanic acid versus when they are tested alone were indicative of ESBL production, according to the CLSI guidelines.99

Pulsed Field Gel Electrophoresis for Molecular Typing: PFGE has been shown to

have a good discriminatory power for ESBL-KE.80,81,100 PFGE were performed by lysing

cells of ESBL-KE colonizing isolates with lysozyme (EC lysis solution) overnight at 37

degrees C. The lysed cells were deprotonated with ESP for 48 hours at 50 degrees C.

Genomic DNA were digested with restriction enzyme XbaI according to the

manufacturer’s recommendations (New England Biolabs, Beverly, MA) for 4 hours at 37

degrees C. Genomic DNA were separated in 2% agarose with 20% turbidity using a

CHEF-DR II (Bio-Rad, Richmond, CA). Electrophoresis was performed at 200 V for 22 hours, with an initial switch time of 5 seconds and the final switch time of 40 seconds.

After electrophoresis, the gel was stained with ethidium bromide and photographed under

ultraviolent light using the Gel Doc (Bio-Rad, Hercules, CA).101 Interpretation was performed using Gel Compare software (Applied Math). A dendrogram comparing all isolates was constructed using Dice coefficient and unweighted pair group method with arithmetic mean with a position tolerance of 1. Isolates with 100% similarity in PFGE band pattern was considered the same and isolates with 80% similarity in PFGE band pattern were considered similar according to Tenover criteria.102

Polymerase Chain Reaction (PCR): PCR was used to determine which β-lactamase

derived gene (TEM, SHV or CTX-M) was present in the ESBL-KE isolates obtained from both the ESBL-KE positive prior room occupants and the patients who acquired

ESBL-KE during their ICU stay. Genomic DNA was extracted from pure colonies of

ESBL-KE isolates and used as the PCR template. Oligonucleotide primers were used to

amplify the ESBL genes (Table III.5). TEM PCR reaction contained 23 µl platinum

supermix, 0.5 µl of forward primer, 0.5 µl of reverse primer, and 1 µl of DNA. PCR

amplifications for the TEM-derived gene begun with a 5-minute denaturation step at 96

degrees C, and was followed by 35 cycles of:

• denaturation at 96 degrees C for 1 minute

• annealing at 58 degrees C for 1 minute

• extension at 72 degrees C for 1 minute

• final extension at 72 degrees C for 10 minutes

SHV PCR reaction contained 14.7 µl of water, 5 µl of 5x buffer, 2 µl of Magnesium

chloride (MgCl2), 0.2 µl of dNTP, 1 µl of SHV forward primer, 1 µl of SHV reverse

primer, 0.1 µl of Go Tag polymerase and 1 µl of DNA. PCR amplifications for the SHV

gene begun with a 2-minute denaturation step at 94 degrees C, and was followed by 35

cycles of:

• denaturation at 94 degrees C for 30 seconds

• annealing at 58 degrees C for 48 seconds

• extension at 72 degrees C for 1 minute

• final extension at 72 degrees C for 5 minutes

CTX-M PCR reaction contained 13.68 µl of water, 5 µl of 5x buffer, 2 µl of MgCl2, 0.2

µl of dNTP, 1.5 µl of CTX-M forward primer, 1.5 µl of CTX-M reverse primer, 0.12 µl

of Go Tag polymerase and 1 µl of DNA. PCR amplifications for the CTM-X gene begun

with a 2-minute denaturation step at 94 degrees C, and was followed by 35 cycles of:

• denaturation at 94 degrees C for 50 seconds

• annealing at 50 degrees C for 40 seconds

• extension at 72 degrees C for 1 minute

• final extension at 72 degrees C for 5 minutes

Table III.5: Nucleotide Sequence of Oligonucleotides used for PCR amplification

Primer Nucleotide Sequence (5’ to 3’) Product size (bp)

TEM Forward: 5’-ATG AGT ATT CAA CAT TTC CG-3’ 867

Reverse: 5’-CTG ACA GTT ACC AAT GCT TA-3’

SHV Forward: 5’-GGT TAT GCG TTA TAT TCG CC-3’ 867

Reverse: 5’-TTA GCG TTG CCA GTG CTC-3’

CTX-M Forward: 5’-CGC TTT GCG ATG TCG AG-3’ 550

Reverse: 5’-ACC GCG ATA TCG TTG GT-3’

VII. Statistical Analysis

Univariate analysis was performed to evaluate the distributions of all study

variables listed in table III.3 above. The primary exposure was ESBL-KE positive status

of the immediate prior room occupant. The outcome was ESBL-KE acquired during the

ICU stay. Bivariate analysis was performed to measure the association between covariates and ESBL-KE status of prior room occupant as well as the association between covariates and ESBL-KE acquisition. Bivariate analysis used the chi-square test or Fisher’s exact test to compare categorical variables and Student’s t-test or the

Wilcoxon Rank Sum test to compare continuous variables. Stratified analysis was performed to assess effect modification using the Breslow-Day test. When effect modification was found, a sub-analysis was performed to assess the association between prior room occupant ESBL-KE status and acquisition of ESBL-KE stratified by levels of the effect modifier. All statistical tests were performed using 2-tailed test. All variables associated with ESBL-KE acquisition at P <0.1 based on bivariate analysis and any effect modifiers were entered into a multivariable analysis.

Logistic regression was used to quantify the Odds Ratio (OR) and 95%

Confidence Interval (CI) for the association between prior room occupants’ ESBL-KE

status and acquisition of ESBL-KE by the subsequent room occupant adjusting for

potential confounders. All variables associated with ESBL-KE acquisition were included

in the initial (full) multivariable logistic regression model. Variables that were not

statistically associated with the outcome (P > 0.05) were removed from the model starting

with the variable with the largest p-value. After removing each variable, we assessed

whether removing the variable altered the regression coefficient of the primary exposure by >10%. If so, the variable was included in the model as a confounder. Statistically significant effect modifiers were included in the final model if they were biologically plausible. Preliminary data from UMMC suggested that 12% -15% of patients had repeat admissions.65 Therefore patients were allowed to enter the study multiple times; we used the robust sandwich variance estimator to account for the correlated error structure that resulted from repeat admissions.103 Among the patients admitted to rooms whose prior occupants were positive for ESBL-KE, the proportion of acquired ESBL-KE isolates that were genetically identical or similar to ESBL-KE isolates obtained from the immediate prior room occupants were quantified.

B. The Effect of Universal Glove and Gown Use on the Incidence Rate of ESBL-

Producing Enterobacteriaceae in the Medical Intensive Care Unit

Specific Aim: To determine if universal glove and gown use is associated with a lower

incidence rate of ESBL-producing Enterobacteriaceae in ICU patients.

Overview: This study evaluated the effectiveness of universal glove and gown use compared to contact precaution for colonized or infected patients in decreasing the incidence rate of ESBL-producing Enterobacteriaceae in patients admitted to the UMMC

MICU. Universal glove and gown use is a form of contact precaution defined as the use

of gloves and gowns by all healthcare workers when entering all patients’ rooms, for all

patients’ contacts and contact with patients’ environment regardless of patients’ colonization or infection status. Universal glove and gown use was instituted at the

UMMC MICU in July, 1 2007 as part of a CDC-funded research and quality improvement program. Infection control policies during the period before universal glove

and gown use included MRSA nasal and VRE peri-anal screening on ICU admission,

weekly and discharge, use of contact isolation for patients colonized or infected with

MRSA and VRE, and patients infected with MDR-Gram-negative bacteria. MDR-Gram- negative bacteria were defined as all Gram-negative bacteria susceptible to two or fewer antibiotics. Peri-anal surveillance cultures were not worked up for ESBL-producing

Enterobacteriaceae and thus patients were not isolated based on colonization with ESBL- producing Enterobacteriaceae.

The exposed group consisted of patients admitted to the MICU during the period

of universal glove and gown use (July 1, 2007 to June 30, 2009). The unexposed group

consisted of patients admitted to the MICU in the period before universal glove and gown

use (July 1, 2005 to June 30, 2007) (Figure III.2). Patients were considered to belong to the period during which their ICU admission occured. The study was ESBL-producing

Enterobacteriaceae acquired during the ICU stay as determined by peri-anal surveillance

cultures and clinical cultures. We defined patients who acquired ESBL-producing

Enterobacteriaceae as those who had a negative peri-anal culture for all ESBL-producing

Enterobacteriaceae on ICU admission but had a subsequent positive peri-anal or clinical

culture for any ESBL-producing Enterobacteriaceae on the same ICU admission. ESBL-

producing Enterobacteriaceae included in this study were Klebsiella pneumonia,

Klebsiella.oxytoca, Eschericia coli, Enterobacter cloacae, Enterobacter aerogenes,

Citrobacter freundii, Proteus mirabilis, Providential stuartii, Morganella morganii and

Serratia marcescens. These ESBL-producing Enterobacteriaceae will henceforth be

referred to as ESBL-Gram-negative rods (GNR). This study was approved by the UMB

IRB for the duration of the study period.

I. Study Design

A quasi-experimental study was performed to assess acquisition rates of ESBL-

GNR during the two years before universal glove and gown use and the two years of universal glove and gown use in patients admitted to the UMMC MICU. In this study, like all quasi-experimental studies, individual study participants were not randomized to universal glove and gown use.107 A quasi-experimental design was an appropriate design for this study because it was logistically difficult to randomize patients within the MICU to universal glove and gown use without the intervention contaminating the patients randomized to the non-intervention arm.

The study population consisted of all patients admitted to the MICU during the

study period that had an ESBL negative peri-anal culture on admission to the MICU,

stayed in the MICU for at least 48 hours and thus were at risk for acquiring ESBL-GNR.

Among patients who acquired ESBL-GNR during their ICU stay, only the first positive peri-anal surveillance or clinical culture was included in the analysis. Patients with multiple admissions to the MICU during the study period were allowed to enter the cohort of at risk patients multiple times as long as they were not positive for ESBL-GNR

on the index ICU admission.

Figure III.2: Study Design for Study 2

II. Study Covariates

Patients’ demographic and clinical characteristics were collected as described in study 1

to identify potential confounding variables.

III. Sources of data

Covariates were collected through UMMC CDR and through laboratory testing as described in study 1.

IV. Laboratory Methods

Microbiology methods were performed as described in study 1.

V. Statistical Analysis

Univariate analysis was performed to evaluate the distribution of all study

variables. The primary exposure was universal glove and gown use. The outcome was

ESBL-GNR acquired during the ICU stay. Bivariate analysis was performed to measure

the association between study covariates and universal glove and gown use as well as the

association between study covariates and acquisition of ESBL-GNR. Continuous

variables were compared using Student’s t-test or the Wilcoxon Rank Sum test while categorical variables were compared using chi-square test or Fisher’s exact test as

appropriate. Stratified analysis was performed to assess effect modification using the

Breslow-Day test. The prevalence of ESBL-GNR on MICU admission was calculated by

dividing the number of patients positive for ESBL-GNR on admission to the MICU by

the total number of patients admitted to the MICU.

Unadjusted incidence rates of ESBL-GNR were calculated by dividing the number of acquired ESBL-GNR by the total number of admission days at risk during the two study periods. Admission days at risk was calculated as number of days from MICU admission to the day of first ESBL-GNR positive peri-anal or clinical culture for admissions where acquisition of ESBL-GNR occurred and number of days from MICU admission to discharge for admissions where acquisition of ESBL-GNR did not occur during the ICU stay. Prevalence of ESBL-GNR on MICU admission and ESBL-GNR acquisition rates were compared for the period of before universal glove and gown use and the period of universal glove and gown use using the chi-square test. Statistical tests were performed using 2-tailed tests. All variables found to be associated with acquisition of ESBL-GNR at P-value of <0.1 based on bivariate analysis were entered into a multivariable analysis.

Poisson regression was used to quantify the rate ratio (RR) and 95% confidence interval (CI) for the association between universal glove and gown use and acquisition of

ESBL-GNR controlling for confounding variables. All variables associated with

acquisition of ESBL-GNR were included in the initial (full) multivariable logistic

regression model. Variables that were not statistically associated with the outcome (P >

0.05) were removed from the model starting with the variable with the largest p-value.

After removing each variable, we assessed whether removing the variable altered the

regression coefficient of the primary exposure by >10%. If so, the variable was included

in the model as a confounder. Statistically significant effect modifiers were included in

the final model if they were biologically plausible. Patients were allowed to enter the

study multiple times; therefore, we used the robust sandwich variance estimator to

account for the correlated error structure that resulted from repeat admissions.103

Multivariable analysis was performed and reported separately for all ESBL-GNR, ESBL-

Klebsiella and ESBL-E. coli.

Aggregate level analysis was performed using segmented linear regression to estimate changes in the mean (immediate effect of glove and gown use on the acquisition rate of ESBL-GNR) and changes in the trend (long term effect of glove and gown use on acquisition rate of ESBL-GNR).108 Since time is a predictor in segmented regression analysis, error terms of consecutive observations may be correlated. We assessed autocorrelation by plotting the residuals against time and visually inspecting for autocorrelation. Randomly scattered residuals with no pattern indicated no autocorrelation. When autocorrelation was detected, we estimated the Durbin-Watson statistics for the autocorrelation parameters and included them in the segmented regression model to adjust for autocorrelation.107

A. Results: Risk of acquiring extended-spectrum β-lactamase (ESBL) producing Klebsiella and Escherichia coli from prior room occupants in the intensive care unit1

Abstract

Background: Environmental contamination by patients carrying extended spectrum β-

lactamase (ESBL)-producing Klebsiella spp. and Escherichia coli has been reported;

however, few studies have assessed the association between environmental contamination

and acquisition of ESBL-Klebsiella and E. coli in non-outbreak settings. Prospective

environmental culturing is both time intensive and costly, therefore, prior room

occupants’ ESBL-Klebsiella or E. coli status was used as a proxy for hospital room

contamination. This proxy has been used by prior studies to assess this association for

other antibiotic-resistant bacteria.

Objective: To determine if admission to a room previously occupied by a patient

colonized with ESBL-Klebsiella or E. coli increases the risk of acquiring ESBL-

Klebsiella or E. coli among intensive care unit (ICU) patients.

Methods: A retrospective cohort study was conducted among patients admitted to the

University of Maryland Medical Center (UMMC) medical ICU (MICU) and surgical ICU

(SICU) between September 1, 2001 and June 30, 2009. Peri-anal surveillance cultures

were collected from all patients on ICU admission, weekly and at ICU discharge and

clinical cultures were collected as medically indicated. Inclusion criteria were ICU length of stay >48 hours and having a peri-anal culture negative for ESBL-Klebsiella or E. coli

on ICU admission. Patients were classified as “exposed” if the immediate prior occupant

*Adebola O. Ajao, MPH;1 J. Kristie Johnson, PhD; 1,2 Anthony D. Harris, MD, MPH; 1 Min Zhan, PhD;1 Jessina C. McGregor, PhD; 3,4 Kerri A. Thom, MD, MS;1 Jon P. Furuno, PhD.1 1Department of Epidemiology and Public Health, 2Department of Pathology, University of Maryland School of Medicine, Baltimore, MD 3College of Pharmacy, Oregon State University, Portland, OR 4Oregon Health and Science University, Portland, OR 45

of the room to which they were admitted had a peri-anal or clinical culture positive for

ESBL-Klebsiella or E. coli. The outcome of interest was a peri-anal or clinical culture

positive for ESBL-Klebsiella or E. coli during the ICU stay after a negative peri-anal

culture for ESBL-Klebsiella or E. coli on ICU admission. Multivariable logistic

regression was used to model the association between prior room occupants’ ESBL-

Klebsiella or E. coli status and acquisition of ESBL-Klebsiella or E. coli by the

subsequent room occupant adjusting for colonization pressure and other potential

confounders. Colonization pressure was defined as average of the daily proportion of

patients positive for ESBL-Klebsiella or E. coli during each patient’s ICU time at risk.

PFGE was performed on isolates obtained from patients colonized with the same

bacterial species as their prior room occupants. Patients were determined to acquire the

same (100% similarity in PFGE pattern) or closely related (>80% similarity in PFGE

pattern) bacterial strain as their prior room occupants.

Results: There were 18,175 admissions to the MICU and SICU during the study period.

10% (1,786) of admissions did not receive admission surveillance culture and were

excluded from the study population. Of the remaining admissions, 9,371 admissions

(7,651 unique patients) stayed in the ICU for > 48 hours and were ESBL negative on ICU

admission and therefore were considered at risk for acquiring ESBL-Klebsiella or E. coli.

Two hundred and sixty-seven admissions (3%) acquired ESBL-Klebsiella or E. coli

during their ICU stay. Prior room occupant’s positive ESBL-Klebsiella or E. coli status

was significantly associated with acquisition of ESBL-Klebsiella or E. coli before adjusting for potential confounding variables (crude odds ratio (COR) = 1.88, 95% CI =

1.29 - 2.74). After adjusting for potential confounders, the association was no longer

statistically significant (Adjusted Odds Ratio (AOR) = 1.39 95% CI = 0.94 - 2.08). Other

factors significantly associated with acquisition of ESBL-Klebsiella or E. coli after controlling for the other variables in the model were ICU time at risk > 4 days (AOR =

1.67, 95% CI 1.26 - 2.23), colonization pressure (AOR = 2.18, 95% CI =1.60 - 2.97),

hemi paraplegia (AOR = 2.40, 95% CI 1.31 - 4.41), renal disease (AOR = 1.68, 95% CI

1.19 - 2.35), aminoglycosides (AOR = 1.79, 95% CI 1.29 - 2.51), anti-pseudomonal β- lactam (AOR = 2.17, 95% CI 1.43 - 3.30), imipenem (AOR = 1.59, 95% CI 1.14 - 2.21) and anti-MRSA therapy received in the ICU (AOR = 1.72, 95% CI 1.25 - 2.37).

Malignancy (AOR = 0.63, 95% CI = 0.43 - 0.92) was associated with a decreased risk of

ESBL-Klebsiella or E. coli acquisition after controlling for the other variables in the model. Of the 267 ESBL-Klebsiella or E. coli acquisitions, 32 (12%) had prior room

occupants who were positive for ESBL-Klebsiella or E. coli. Of these 32, 17 (53%) were colonized with the same bacterial species as their prior room occupants. PFGE performed on isolates from the 17 patients suggests that 3 patients (18%) acquired the same bacteria

strain as their prior room occupants while 3 patients (18%) acquired a bacteria strain that

were closely related to that of their prior room occupants.

Conclusions: Our study results suggest that environmental contamination may not play a

major role in the transmission of ESBL-Klebsiella or E. coli at UMMC.

Introduction

Extended-spectrum β-lactamase (ESBL)-producing Gram-negative rods (GNR) are emerging pathogens that are associated with considerable morbidity, mortality and costs among hospitalized patients both in the United States and globally.1,2 ESBLs are

produced by many Gram-negative rods but Klebsiella species (predominantly Klebsiella

pneumoniae) and Escherichia coli are the most common and clinically relevant ESBL-

GNR.3-6 Infections caused by ESBL-producing Klebsiella and E. coli are difficult to treat due to the limited options for empiric therapy. Thus, there is an increased likelihood that empirically selected therapy may not provide coverage for the ESBL-Klebsiella and E. coli which subsequently increases the risk of treatment failure and poor outcomes among infected patients.2,11

Patients admitted to the intensive care units (ICUs) have higher rates of ESBL- producing Klebsiella and E. coli120 likely due to a high concentration of critically ill

patients who are less resistant to exogeneous colonization and are in frequent close

contacts with physicians and nurses. In addition, ICU patients have a heavy use of

empiric broad spectrum antibiotics that creates selective pressure and promotes the

emergence of resistant endogenous strains.14,54 The prevalence of ESBL-producing

isolates in ICU patients with nosocomial infections in the United States ranged from 5%

in E.coli to 20% in K. pneumoniae in 2003 and continues to rise.15,16

Patients carrying ESBL-producing Klebsiella and E. coli often contaminate their immediate hospital environment. Previous studies conducted in outbreak situations have suggested that contamination of environmental surfaces by ESBL-Klebsiella persisted despite disinfection and cleaning of hospital rooms after patients’ discharge.12,13 Prior

room occupants’ status has been used by previous studies as a proxy for hospital room

contamination because prospective culturing of the environment is both time intensive

and costly. Previous studies have shown that admission to an ICU room previously

occupied by patients with methicillin-resistant Staphylococcus aureus (MRSA),

vancomycin resistant enterococci (VRE), Clostridium difficile, multi-drug-resistant

(MDR)-Pseudomonas aeruginosa and Acinetobacter baumannii is associated with

acquisition of these bacteria by subsequent room occupants.74,109,110,122 While the use of

prior occupant’s status can serve as a low-cost proxy variable for environmental contamination, a previous study by Nseir et al. did not demonstrate an association between prior occupant’s status and acquisition of ESBL-producing Enterobacteriaceae.

The objective of this study was to evaluate if admission to an ICU room previously

occupied by a patient with ESBL-producing Klebsiella and E.coli increases the risk of

acquiring these bacteria by subsequent room occupants.

Methods

Study Design

A retrospective cohort study was performed using data from patients admitted to

the University of Maryland Medical Center (UMMC) medical ICU (MICU) and surgical

ICU (SICU) from September 1, 2001 to June 30, 2009. During the study period, UMMC was a 729-bed, urban tertiary-care hospital in Baltimore, Maryland. The MICU was a 10-

bed private room unit that became a 29-bed private room unit in April 2006 while the

SICU is a 19-bed private room unit. Patients admitted to the UMMC MICU and SICU

had obtained peri-anal cultures routinely as part of an ongoing VRE active surveillance

program since 1995. Peri-anal cultures were obtained using Staplex II swabs (Staplex,

Etobiocoke, Ontario, Canada) within 48 hours of ICU admission, weekly and at ICU

discharge. Compliance with obtaining peri-anal cultures has been shown to be

approximately 90%.80,81 Clinical cultures were collected as medically indicated.

Study Population

The study population i.e. patients-at risk for acquiring ESBL-Klebsiella or E. coli consisted of all patients admitted to the UMMC MICU and SICU during the study period that had a negative peri-anal surveillance culture for ESBL-Klebsiella or E. coli on ICU admission and stayed in the ICU for 48 hours or longer. Patients who did not receive an admission peri-anal surveillance culture were excluded from the study population as their

ESBL status on ICU admission could not be established. Patients who were colonized with the ESBL-Klebsiella or E. coli upon ICU admission and patients who stayed in the

ICU for less than 48 hours were also excluded from the study population as they had a deceased risk for acquiring ESBL-Klebsiella or E. coli. Patients were classified as

“exposed” if the immediate prior occupant of the room to which they were admitted had a peri-anal or clinical culture positive for ESBL-Klebsiella or E. coli. The outcome of interest was a peri-anal or clinical culture positive for ESBL-Klebsiella or E. coli during the ICU stay. Among patients who became positive for ESBL-Klebsiella or E. coli during their ICU stay, only the first positive surveillance culture was included in the analysis.

Patients with multiple admissions to the ICUs during the study period were allowed to enter the cohort of at-risk patients multiple times as long as they were not positive for

ESBL-Klebsiella or E. coli at the time of the index ICU admission. At-risk patients who

changed rooms during their ICU stay accounted for more than one ICU admission as long

as they were still at-risk upon moving to the new room. All prior room occupants’ stay was included regardless of their length of stay. Terminal cleaning (disinfection and cleaning of hospital room surfaces after patient discharge) of ICU rooms was performed by the hospitals’ hospitality services using a quaternary germicidal cleaner, SaniMaster

IV (Ecolab, MN, USA) in accordance with CDC guidelines.85 Infection control policies

during the study period included MRSA nasal and VRE peri-anal screening on ICU

admission, weekly and discharge, use of contact isolation for patients colonized or

infected with MRSA and VRE, and patients infected with MDR-Gram-negative bacteria

including ESBL-GNR. MDR-Gram-negative bacteria were defined as all Gram-negative bacteria susceptible to two or fewer antibiotics. Peri-anal surveillance cultures were not worked up for ESBL real time and thus patients were not isolated based on colonization with ESBL-GNR. This study was approved by the University of Maryland, Baltimore

Institutional Review Board.

Data sources and Variable Definitions

All patient data were collected retrospectively from the UMMC Central Data

Repository (CDR) and through laboratory testing. The UMMC CDR is a relational database containing patients’ administrative, pharmacy, admission-discharge-transfer, and clinical data. A random sample of the data from the UMMC CDR was validated against patients’ paper medical records for this study and previous research studies and the CDR data had a positive and negative predictive value greater than 98%.91-93 Patient characteristics collected included data on age, sex, ICU total length of stay, ICU time at

risk, ICU length of stay of the prior room occupant, ICU antibiotic use, individual

comorbid illnesses, aggregate comorbidity score; Charlson Comorbidity Index (CCI), and

colonization pressure. ICU time at risk was defined as the number of days from ICU

admission to ESBL-Klebsiella or E. coli positive peri-anal or clinical culture collection

date for patients who acquired ESBL-Klebsiella or E. coli and number of days from ICU

admission to ICU discharge for patients who did not acquire ESBL-Klebsiella or E. coli.

ICU antibiotic exposure was defined as antibiotics ordered during the ICU time at risk.

Antibiotics ordered included β-lactams and other classes of antibiotics such as macrolides, aminoglycosides, antifungal agents, quinolone, and anti-MRSA. β-lactams ordered were penicillin, 1st generation, 2nd generation, and 3rd generation cephalosporin,

4th generation cephalosporin, and carbapenems. Anti-MRSA ordered included

vancomycin, linezolid, telavancin, tigecycline, daptomycin, quinupristin/dalfopristin, and

dalbavancin. Individual comorbidities were defined as individual comorbid illnesses

present at hospital discharge as determined by the International Classification of

Diseases, Ninth Revision (ICD-9) at ICU discharge. Aggregate comorbidity score; CCI

was calculated using the ICU discharge ICD-9 codes for comorbid conditions.89

Colonization pressure was calculated by dividing the number of patients positive for

ESBL-Klebsiella or E. coli by the number of patients in the unit daily. The average of the

daily colonization pressure was calculated for the period of time that each patient was at

risk for acquiring ESBL-Klebsiella or E. coli. 90 Since peri-anal cultures were obtained

upon ICU admission, weekly and at discharge, patients contributed to colonization

pressure if they were positive for ESBL-Klebsiella or E. coli on ICU admission or they

acquired ESBL-Klebsiella or E. coli during their ICU stay. Patients who were colonized

with ESBL-Klebsiella or E. coli upon ICU admission were assumed to be colonized throughout their ICU stay, while patients who acquired ESBL-Klebsiella or E. coli during

ICU stay were assumed to be ESBL positive from the day of positive culture collection until ICU discharge. Colonization pressure was calculated separately for the MICU and the SICU.

Microbiology and Molecular Testing

ESBL positive status of patients admitted to the ICU during the study period was determined by microbiology testing. Briefly, peri-anal surveillance cultures were inoculated onto a MacConkey agar plates (Remel, Lenexa, KS, USA) containing 1 μg/ml ceftazidime and incubated at 37 degrees C for 24 to 48 hours. Lactose fermenting colonies were identified using the API 20E identification strips or the VITEK II

(BioMeriuex Durham, NC). ESBL testing of Klebsiella species and E. coli was performed using the Kirby-Bauer disk diffusion test method using 30 μg of ceftazidime and 30 μg cefotaxime with and without 10 μg of clavulanic acid as recommended by the

Clinical and Laboratory Standards Institute.99 ESBL positive clinical cultures were

determined by the UMMC microbiology laboratory and susceptibilities were performed

following CLSI guidelines.99 Briefly, clinical cultures were tested against ceftriazone, a

third generation cephalosporin. All isolates in the intermediate resistance range (14-

25mm) underwent ESBL confirmatory testing with Kirby-Bauer disk diffusion method as

described above. Polymerase chain reaction (PCR) was used to determine which family

of β-lactamase derived gene (TEM, SHV or CTX-M) was present in the ESBL-Klebsiella

or E. coli isolates. Pulsed Field Gel Electrophoresis (PFGE) was performed to determine

the extent of genetic similarity of ESBL-Klebsiella and E. coli isolates obtained from patients who acquired ESBL-Klebsiella or E. coli during their ICU stay and their prior room occupants who have the same β-lactamase derived genes. Briefly, bacterial isolates were lysed with lysozyme overnight at 37 degrees C. The lysed cells were deprotonated with ESP for 48 hours at 50 degrees C. Bacterial genomic DNA was digested with restriction enzyme XbaI according to the manufacturer’s recommendations (New

England Biolabs, Beverly, MA) for 4 hours at 37 degrees C and separated in 2% agarose with 20% turbidity using a CHEF-DR II (Bio-Rad, Richmond, CA). Electrophoresis was performed at 200 V for 22 hours, with an initial switch time of 5 seconds and the final switch time of 40 seconds. Gel was stained with ethidium bromide and photographed under ultraviolent light using the Gel Doc (Bio-Rad, Hercules, CA). Interpretation was performed using Gel Compare software (Applied Math).101 Isolates with 100% similarity

in PFGE band pattern were considered the same and isolates with at least 80% similarity

in PFGE band pattern was considered similar according to Tenover criteria.102

Statistical analysis

Univariate analysis was performed to evaluate the distributions of all study

variables. Bivariate analysis was performed to determine the variables associated with

acquisition of ESBL-Klebsiella or E. coli and prior room occupant’s ESBL-Klebsiella or

E. coli positive status. Normally distributed continuous variables were compared using

Student’s t-test while non-normally distributed continuous variables were compared

using the Wilcoxon Rank Sum test. Categorical variables were compared using the

Pearson chi-square test or Fisher’s exact test. Fisher’s exact test was used to compare variables with cell sizes less than five. Stratified analysis was performed to assess

potential effect modification of the association between prior room occupant’s ESBL

status and acquisition of ESBL-Klebsiella or E. coli by other covariates. All variables

associated with acquisition of ESBL-Klebsiella or E. coli at p <0.1 were entered into a

multivariable analysis. Statistical test were performed using 2-tailed tests.

Logistic regression used to quantify the association between prior room

occupants’ ESBL status and acquisition of ESBL-producing Klebsiella or E. coli

adjusting for potential confounding variables. All variables associated with acquisition of

ESBL-Klebsiella or E. coli were included in the initial (full) multivariable logistic

regression model. Starting with the full model, variables that were not statistically

associated with the outcome (P-value > 0.05) were sequentially removed from the model

starting with the variable with the largest p-value. To assess confounding, we assessed if

the removal of a variable resulted in a change in the regression coefficient for prior room

occupants’ ESBL status by more than 10%. Any confounding variables were included in

the model. Since patients were allowed to enter the study multiple times; we used the robust sandwich variance estimator to account for the correlated error structure that

resulted from repeat admissions.103

Results

There were 18,175 admissions to the MICU and SICU between September 2001

and June 2009. Among them, 1,786 (10.0%) admissions did not receive admission peri-

anal surveillance culture and were excluded from the study population. Among patients

that received admission peri-anal surveillance culture, 5,597 (30.8%) admissions had ICU

stay of less than 48 hours, 786 (4.3%) admissions were colonized with ESBL-Klebsiella

or E. coli on ICU admission, and 635 (3.5%) admissions were missing the exposure

variable and were excluded from the study population. Nine thousand three hundred and

seventy-one admissions (7,651 unique patients) were at-risk for acquiring ESBL-

Klebsiella or E. coli in the MICU and SICU during the study period and were included in

the study (Figure IV.1).

Figure IV.1: Study population and exclusion criteria

Among the study population, 6,419 patients (83.9%) had one admission while

1,232 patients (16.1%) had repeat admissions. Patient characteristics are presented in

Table IV.1. The mean age of the patients was 55.7 years (standard deviation (SD) = 15.6 years), 55% were male, and 51% of the admissions were to the MICU. The median ICU length of stay was 4 days (interquartile range (IQR): 2 - 8 days), the median ICU time at

risk was 4 days (IQR: 2 - 8 days) and the median length of stay of prior room occupants was 2 days (IQR: 1 - 5 days). The median Charlson Comorbidity Index was 2 (IQR: 1 -

4), the mean ICU chronic disease score was 8.6 (SD: 4.2). The median colonization pressure was 0.07 (IQR: 0.01 - 0.15). Based upon International Classification of

Diseases, 9th revision (ICD-9) codes, 1,864 (20%) had cancer, 1,552 (17%) had diabetes,

1,213 (13%) had renal disease, and 449 (5%) were HIV positive.

Table IV.1: Characteristics of Patients at Risk for Acquiring ESBL-Klebsiella or E. coli Patient Characteristics Mean, Median, N (%) Age (mean in years, SD) 55.7 (15.6) Male Sex (n, %) 5194 (55.4) Admission to the MICU (n, %) 4794 (51.2) Prior room occupant with ESBL-Klebsiella or E.coli (n, %) 648 (6.9) Acquired ESBL-Klebsiella or E.coli during ICU stay (n, %) 267 (2.9) ICU total length of stay (median in days, IQR) 4 (2 - 8) ICU time at risk (median, IQR) 4 (2 - 8) Length of stay of prior room occupants (median in days, IQR) 2 (1 - 5) Colonization pressure (median, IQR) 0.07 (0.01 - 0.15) Co morbidities (n, %) AIDS 449 (4.8) Malignancy 1864 (19.9) Cerebrovascular 1188 (12.7) Chronic Pulmonary 1969 (21.0) Heart Failure 1467 (15.7) Dementia 21 (0.2) Diabetes 1552 (16.6) Diabetes complicated 511 (5.5) Hemi paraplegia 257 (2.7) Metastatic tumor 693 (7.4) Mild Liver 681 (7.3) Severe Liver 679 (7.3) Myocardial Infarction 729 (7.8) Peptic Ulcer 329 (3.5) Peripheral vascular 618 (6.6) Renal disease 1213 (12.9) Rheumatology 216 (2.3) Charlson comorbidity score (median, IQR) 2 (1 - 4)

ICU antibiotic use (n, %)* Any antibiotic 7972 (85.1) β-lactam 6353 (67.8) Penicillin 4552 (48.6) Anti-pseudomonal β-lactam 5339 (56.9) Anti-anaerobic β-lactam 6003 (64.1) Piperacillin/Tazobactam 3694 (39.4) Cephalosporin 3124 (33.3) 1st Generation 1344 (14.3) 2nd Generation 319 ( 3.4) 3rd Generation 1054 (11.3) Cefepime 776 ( 8.3) Antifungal 2273 (24.3) Imipenem 954 (10.2) Aminoglycosides 730 ( 7.8) Macrolides 933 ( 9.9) Quinolone 1976 (21.1) Anti-MRSA 3636 (38.8) Vancomycin 3351 (35.8) ESBL, extended-spectrum β-lactamase; ICU, intensive care unit; MICU, medical ICU; IQR, Interquartile range; SD, Standard deviation; ICU antibiotics used was defined as antibiotics were ordered during the period at risk.

The result of ESBL-Klebsiella and E. coli transmission are shown in Figure IV.2.

There were a total of 267 admissions (3%) that acquired ESBL-Klebsiella or E. coli

during their ICU stay. Of the 267 acquisitions, 236 (88%) were from surveillance cultures

(i.e. colonization) only, 20 (8%) were from surveillance and clinical cultures and 11 (4%)

were from clinical cultures only. Ninety-four (35%) of the ESBL acquisitions were E. coli, 167 (63%) were Klebsiella pneumoniae and 6 (2%) were Klebsiella oxytoca.

Six hundred and forty-eight (6.9%) of the admissions were to ICU rooms whose prior occupant was positive for ESBL-Klebsiella and E. coli. Of these 648 admissions, 32

(4.9%) acquired ESBL-Klebsiella or E. coli compared to 2.7% who acquired ESBL-

Klebsiella or E. coli among admissions to rooms not previously occupied by a patient positive for ESBL-Klebsiella or E. coli. The 2.2% excess incidence of ESBL-Klebsiella

or E. coli acquisition is what can be attributable to the exposure in the exposed population and this translated into 44% attributable risk exposed. In essence, among the patients admitted to rooms whose prior occupants were positive for ESBL-Klebsiella or E. coli,

44% of the risk of acquiring ESBL-Klebsiella or E. coli was due to being admitted to a room whose prior occupant was positive for ESBL-Klebsiella or E.coli. The crude odds ratio for ESBL-Klebsiella or E. coli acquisition was 1.88 (95% CI = 1.29 - 2.74). Among the 32 acquisitions whose prior room occupants were positive for ESBL-Klebsiella or E. coli, 17 (53%) were colonized with the same bacterial species as their prior room occupants. For these 17 patients, the attributable risk was 0% and the crude odds ratio for

ESBL-Klebsiella or E. coli acquisition was 0.97, (95% CI = 0.59 - 1.60).

Figure IV.2: Transmission of ESBL-Klebsiella and E. coli

The associations between study variables and acquisition of ESBL-Klebsiella or

E. coli are displayed in Table IV.2. We examined if the ICU patients were admitted to

(i.e. MICU versus SICU) modified the effect of the association between prior room

occupant’s ESBL-Klebsiella or E. coli status and acquisition of ESBL-Klebsiella or E.

coli but no effect modification by ICU type was detected (Breslow-Day P-value = 0.30),

therefore data for patients admitted to both the MICU and the SICU were combined and analyzed together. Admission to the MICU, ICU time at risk, colonization pressure, universal glove and gown use, presence of malignancy, cerebrovascular disease, hemi paraplegia, metastatic tumor, peripheral vascular disease, renal disease, use of penicillin, anti-pseudomonal β-lactam, anti-anaerobic β-lactam, piperacillin/tazobactam, first- generation cephalosporin, cefepime, antifungals, imipenem, aminoglycosides and anti-

MRSA therapy were associated with acquisition of ESBL-Klebsiella or E. coli at alpha of

<0.1 and were included in the full model. None of these variables were found to be confounders by the criteria that we used to assess confounding.

Table IV.2: Association between Acquisition of ESBL-Klebsiella or E.coli and Study Variables Patient Characteristics Acquired Did not acquire P-value ESBL-KE ESBL-KE (N=267) (N=9104) Prior room occupant positive for ESBL-Klebsiella or E.coli (n, %) 32 (4.9) 235 (2.7) < 0.01 Age (mean, s.d) 57.1 (15.2) 55.6 (15.6) 0.11 Male Sex (n, %) 161 (60.3) 5033 (55.3) 0.10 Admission to the MICU (n, %) 175 (65.5) 4619 (50.7) < 0.01 ICU total length of stay (median, IQR) 11 (5 - 24) 4 (2 - 8) < 0.01 ICU time at risk (median, IQR) 8 (4 - 15) 4 (2 - 8) < 0.01 Prior room occupant length of stay 3 (1 - 5) 2 (1 - 5) 0.85 (median, IQR) Colonization pressure (median, IQR) 0.13 (0.06-0.13) 0.07 (0.01-0.14) < 0.01 Universal glove and gown use (n, %) 124 (46.4) 2677 (29.4) < 0.01 Co morbidities (n, %) AIDS 10 (3.8) 339 (4.8) 0.42 Malignancy 32 (11.9) 1832 (20.1) < 0.01 Cerebrovascular 18 (6.7) 11170 (12.8) < 0.01

Chronic Pulmonary 49 (18.1) 1920 (21.1) 0.28 Heart Failure 47 (17.6) 1420 (15.6) 0.37 Dementia 1 (0.4) 20 (0.22) 0.59 Diabetes 38 (14.2) 1514 (16.6) 0.29 Diabetes complicated 10 (3.8) 501 (5.5) 0.21 Hemi paraplegia 15 (5.6) 242 (2.7) < 0.01 Metastatic tumor 10 (3.8) 683 (7.5) 0.02 Mild Liver 14 (5.2) 667 (7.3) 1.19 Severe Liver 23 (8.6) 656 (7.2) 0.38 Myocardial Infarction 20 (7.5) 709 (7.8) 0.86 Peptic Ulcer 10 (3.8) 319 (3.5) 0.83 Peripheral vascular 10 (3.8) 608 (6.7) 0.05 Renal disease 65 (24.3) 1148 (12.6) < 0.01 Rheumatology 5 (1.8) 2111 (2.3) 0.63 Charlson co morbidity score 2 (1 - 3) 2 (1- 4) 0.11 ICU chronic disease score (mean, s.d) 8.9 (3.9) 8.6 (4.2) 0.23 ICU antibiotic use (n, %)* Any antibiotic 259 (97.0) 7713 (84.2) < 0.01 β-lactam 214 (80.2) 6139 (67.4) < 0.01 Penicillin 173 (64.8) 4379 (48.1) < 0.01 Anti-pseudomonal β-lactam 232 (86.9) 5107 (56.1) < 0.01 Anti-anaerobic β-lactam 224 (83.9) 5779 (63.5) < 0.01 Piperacillin/Tazobactam 158 (59.2) 3536 (38.8) < 0.01 Cephalosporin 91 (34.1) 3033 (33.3) 0.79 1st Generation 28 (10.5) 1316 (14.9) 0.07 2nd Generation 6 (2.3) 313 ( 3.4) 0.29 3rd Generation 23 (8.6) 1031 (11.2) 0.17 Cefepime 52 (19.5) 724 ( 7.9) < 0.01 Antifungal 85 (31.8) 2188 (20.0) < 0.01 Imipenem 69 (25.8) 888 ( 9.8) < 0.01 Aminoglycosides 53 (19.9) 677 ( 7.4) < 0.01 Macrolides 33 (12.4) 900 (9.9) 0.18 Quinolone 66 (24.7) 1910 (20.9) 0.14 Anti-MRSA 189 (70.8) 3447 (37.9) < 0.01 Vancomycin 172 (64.4) 3179 (34.9) < 0.01 ESBL, extended-spectrum β-lactamase; ICU, intensive care unit; MICU, medical ICU; IQR, Interquartile range; MRSA, Methicillin-resistant Staphylococcus aureus

Since patients were allowed to enter the study multiple times; we adjusted for the correlated error structure of the data in the multivariable analysis. The result of the multivariable logistic regression adjusting for the correlated error structure is shown in

Table IV.3. After adjusting for potential confounding variables, the association between

prior room occupant’s positive ESBL-Klebsiella or E. coli status and acquisition of

ESBL-Klebsiella or E. coli was no longer statistically significant (Adjusted Odds Ratio

(AOR) = 1.39, 95% CI = 0.94 - 2.08). ICU time at risk > 4 days (AOR = 1.67, 95% CI

1.26 - 2.23), colonization pressure (AOR = 2.18, 95% CI =1.60 - 2.97), hemi paraplegia

(AOR = 2.40, 95% CI 1.31 - 4.41), renal disease (AOR = 1.68, 95% CI 1.19 - 2.35), anti-

MRSA therapy (AOR = 1.86, 95% CI 1.36 - 2.52), aminoglycoside (AOR = 1.79, 95%

CI 1.29 -2.51), anti-pseudomonal β-lactam (AOR = 2.17, 95% CI 1.43 - 3.30), and imipenem received in the ICU (AOR = 1.59, 95% CI 1.14 - 2.21) were found to be significantly associated with acquisition of ESBL-Klebsiella or E.coli after controlling for the other variables in the model. Having malignancy (AOR = 0.63, 95% CI = 0.43 -

0.92) was found to be associated with a decreased risk of acquiring ESBL-Klebsiella or

E.coli after controlling for the other variables in the model. The adjusted odds ratio for a

separate multivariable logistic regression performed for the17 patients that acquired the

same bacterial species as their prior room occupants was 0.70, 95% CI = 0.42 - 1.17. To

our knowledge, this is the first study to identify colonization pressure and hemi

paraplegia as predictors of ESBL-Klebsiella or E. coli acquisition in ICU patients.

Table IV.3: Predictors of EBSL-Klebsiella or E. coli Acquisition during ICU Stay in Multivariable Logistic Regression Model. Study Variables AOR 95% CI ESBL-KE positive prior room occupant 1.39 0.94 - 2.08 ICU length of stay at risk (> 4 days) * 1.67 1.26 - 2.23 Colonization pressure (> 7%) * 2.18 1.60 - 2.97 Hemi paraplegia 2.40 1.31 - 4.41 Renal disease 1.68 1.19 - 2.35 Malignancy 0.63 0.43 - 0.92 Anti-MRSA received in the ICU** 1.72 1.25 - 2.37 Aminoglycoside received in the ICU** 1.79 1.29 - 2.51 Anti-pseudomonal β-lactam received in the ICU** 2.17 1.43 - 3.30 Imipenem received in the ICU** 1.59 1.14 - 2.21 ESBL, extended-spectrum β-lactamase; ICU, intensive care unit; MRSA, methicillin- resistant Staphylococcus aureus; OR, odds ratio; CI, confidence interval. *These variables were dichotomized at the median because variables were not normally distributed. **Antimicrobial drug exposure was assessed antibiotic ordered during the period between ICU admission and ICU discharge for patients that did not acquire ESBL- Klebsiella or E.coli and between ICU admission and date positive surveillance culture was collected for patients that acquired ESBL-Klebsiella or E.coli.

Of the 17 patients who were colonized with the same bacterial species as their

prior room occupants, 4 were colonized with E. coli while 13 were colonized with K.

pneumoniae. PFGE performed on isolates obtained from these 17 patients showed that 3

patients (18%) acquired the same bacteria as their prior room occupants while 3 patients

(18%) acquired a bacteria strain that was closely related to that of their prior room

occupants. (Table IV.4). We were not able to detect any ESBL genes in two of the patient

pairs possibly because only TEM, SHV and CTX-M family of ESBL genes were tested

for in this study.

Table IV.4: PFGE profile for the 17 patients colonized with the same bacterial species as their prior room occupants. Patient Organism Percent Pulsed Field Gel Electrophoresis (PFGE) Pair Similarity 1. K. pneumoniae 45% 5799 5812

2. K. pneumoniae 30% 2743 2554 3. K. pneumoniae 50% 4836 4791 4. E. coli 20% 1023 1013 5. K. pneumoniae 58% 4088 3978 6. K. pneumoniae 93% 1061 1041 7. K. pneumoniae 48% 356 292 K. pneumoniae. 5799 K. pneumoniae 5812 8. E. coli 40% 1865 1817 9. E. coli 100% 2518 2336 10. K. pneumoniae 79% 2934 2793 11. E. coli 88% 3524 3496 12. K. pneumoniae 100% 4417 4857 13. K. pneumoniae 56% 6854 6047 14. K. pneumoniae 100% 2047 2245 15. K. pneumoniae 58% 356 301 16. K. pneumoniae 100% 3774 3893 301 17. K. pneumoniae 54% 4408 4431 PFGE, pulsed field gel electrophoresis

4417

Discussion

The aim of this study was to evaluate if admission to an ICU room previously

occupied by a patient positive for ESBL-Klebsiella or E. coli increases the risk of

acquiring ESBL-Klebsiella or E. coli by subsequent room occupants. Our study results suggest that prior room occupant’s ESBL-Klebsiella or E. coli positive status was not significantly associated with acquisition of ESBL-Klebsiella or E. coli at UMMC. There were multiple levels of evidence that demonstrated this lack of association. The first adjusted analysis involving the 32 patients that acquired ESBL-Klebsiella or E. coli among the exposed did not reach statistical significance. In addition only 44% of the risk of acquiring ESBL-Klebsiella or E. coli was attributable to the exposure among the exposed group. The second adjusted analysis involving the 17 patients that acquired the same bacterial species as their prior room occupants also showed a lack of statistical association. In addition, 0% of the risk of acquiring ESBL-Klebsiella or E. coli was attributable to the exposure among this exposed group. Molecular analysis further showed that only 3 out of 17 (18%) acquired a bacteria that was 100% identical to the bacteria colonizing their prior room occupant. Our study results are consistent with the report of a previous study by Nseir et. al.109 A smaller study of colonization or infection with ESBL-GNR conducted in ICU patients in France.

Possible explanation why we were not able to demonstrate an association between

prior room occupant’s ESBL-Klebsiella or E. coli positive status and acquisition of

ESBL-Klebsiella or E. coli may be that prior room occupants’ ESBL status may not be a

valid proxy for environmental contamination with ESBL-Klebsiella and E. coli due to the

following reasons. First, patients colonized or infected with ESBL-Klebsiella and E. coli

may not shed a high enough innoculum into their environment to be transmitted to

subsequent room occupants. Second, cleaning of ICU rooms using quaternary ammonium

may be efficient in eradicating ESBL-Klebsiella and E. coli from the environment.

Current literature suggests that Klebsiella or E. coli can persist in the environment for months but does not suggest that Klebsiella or E. coli persist for a shorter duration of time than other Gram-negative bacteria such as Acinetobacter spp. and Pseudomonas aeruginosa. However, there is evidence that factors such as high innoculum, humid condition and low temperature improve persistence of most Gram-negative bacteria in the environment.125 This implies that infection control interventions for ESBL-GNR may

need to focus on other modes of transmission besides transmission through the

environment. To further explore this hypothesis, future research may need to correlate

environmental culturing data with prior room occupants’ ESBL status. Parhaps, a

prospective study where hospital environmental surfaces are swabbed after patient’s

discharge and terminal cleaning to establish environmental contamination may be needed

to properly assess the impact of environmental contamination on acquisition of ESBL-

GNR. However, this type of study may be costly to perform.

Other factors associated with acquisition of ESBL-Klebsiella and E. coli in our

study were higher colonization pressure, longer ICU time at risk, renal disease, hemi

paraplegia, aminoglycosides, anti-pseudomonal β-lactam, imipenem and anti-MRSA

therapy received during the ICU time at risk. Malignancy was found to be protective of

acquiring ESBL-Klebsiella and E. coli in our study. Our study results suggest that

patients present in the ICU when ESBL-Klebsiella and E. coli colonization pressure was

above the median colonization pressure of 7% had twice the risk of acquiring ESBL-

Klebsiella or E. coli. Colonization pressure, an important infection control metric has

been previously shown to be associated with acquisition of many nosocomial pathogens

such as MRSA, VRE, and C. difficile in ICU patients.90,112-115 However, to our

knowledge; this is the first study to identify colonization pressure as a risk factor for

acquiring ESBL-Klebsiella and E. coli.

In addition to our study, several other studies have demonstrated that a longer

length of ICU stay is a risk factor for ESBL-GNR.7,60,66 Having renal disease, receipt of

aminoglycosides and anti-MRSA is also consistent with what has been previously

reported in the literature.47,65,88,119 In this study, 85% of the anti-MRSA antibiotics received was vancomycin and previous use of a glycopeptides specifically vancomycin has been previously reported as a risk factor for ESBL-Klebsiella or E. coli colonization.47,65,88 The receipt of anti-MRSA probably eliminates MRSA and other

Gram-positive bacteria in the human flora allowing MDR-Gram-negative bacteria to flourish. The presence of hemi paraplegia, receipt of imipenem and anti-pseudomonal β- lactam during the ICU time at risk as a risk factor for acquiring ESBL-Klebsiella and E. coli has not been previously reported in the literature. The presence of malignancy was found to be protective for acquiring ESBL- Klebsiella and E. coli in this study.

One limitation of this study is that prior room occupant’s ESBL status was used as

a proxy for environmental contamination. Hospital room environmental surfaces were not

cultured in this study and thus we were not able to describe the level of environmental

contamination under which acquisition of ESBL-Klebsiella and E. coli occurred.

A second limitation is that sensitivity of peri-anal cultures for detecting ESBL-

GNR compared to stool culture (gold standard) is unknown. It is possible that some

ESBL-Klebsiella and E. coli cases may have been undetected by our surveillance method

and may have led to misclassification of ESBL-Klebsiella and E. coli positivity status.

However sensitivity of peri-anal culture compared to stool culture for fluoroquinolone-

resistant E. coli was 90%.105 Therefore, we anticipate that the sensitivity of peri-anal

cultures for ESBL-Klebsiella and E. coli would be similar and the proportion of false

negatives would be low (~10%). Thus, misclassification of ESBL-Klebsiella and E. coli

positivity status by our surveillance methods would be non-differential and would have

little potential to bias the measure of effect towards a null effect.

A third limitation of this study is that not all bacterial isolates from clinical cultures obtained from patients during our study period were tested for ESBLs by the

UMMC clinical laboratory. Only patients with bacterial isolates with intermediate

resistance to third generation cephalosporin were tested for ESBLs. Therefore, it is

possible that we might have misclassified some patients with bacteria isolates from

clinical cultures whose resistance to third-generation cephalosporin was due to the ESBL

resistance mechanism as ESBL negative. Among patients with bacterial isolates from

clinical cultures that are resistant to third-generation cephalosporin, a fraction of the

resistance are due to ESBLs while a fraction are due to some other types of resistance

mechanism such as AmpCs and carbapenemases. Our data suggests that 65% of patients

with clinical cultures positive for ESBL-Klebsiella or E. coli were already colonized with

ESBL-Klebsiella or E. coli, and were captured using our surveillance method. Any

misclassification of ESBL status has the potential to bias the estimate of effect towards a

null effect. However, we anticipate that any misclassification of ESBL status that might

occur would be non-differential and the level of misclassification would be low.

In conclusion, our study results suggest that there was no significant association between prior room occupants’ ESBL-Klebsiella or E. coli positive status and acquisition of ESBL-Klebsiella or E. coli by subsequent room occupants. This implies that environmental contamination may not play a significant role in the transmission of

ESBL-GNR. Future research should aim to evaluate the mimimum innoculum of ESBL-

Klebsiella and E. coli that is needed in the environment to colonize or infect subsequent room occupants. Future research could better evaluate the impact of persistent environmental contamination on acquisition of ESBL-GNR by directly swabbing hospital environmental surfaces after patients’ discharge and terminal cleaning to establish the presence or absence of environmental contamination by ESBL-GNR. In addition, future research should employ molecular method for bacterial strain comparison to validate that it is the same bacterial strain shed into the environment by the previous room occupant that is being transmitted to the subsequent room occupant. This information is important to properly assess if the level of ICU room cleaning is sufficient to prevent transmission of ESBL-GNR from the environment to the patients. This study also identified other conditions under which patients are most likely to acquire ESBL-Klebsiella or E .coli in

the ICU. This study highlights the importance of keeping the burden of patients colonized

or infected with ESBL-GNR (colonization pressure) low in the ICU to reduce the risk of

patient-to-patient transmission of ESBL-GNR. This study also highlights the importance

of keeping the length of ICU stay short and limited use of antibiotics such as anti-MRSA

therapy, aminoglycosides, anti-pseudomonal β-lactam, and imipenem in ICU patients to reduce the risk of acquiring ESBL-GNRs.

B. Results: The Effect of Universal Glove and Gown Use on the Incidence Rate of ESBL-Producing Enterobacteriaceae2

Abstract

Background: The prevalence of extended-spectrum β-lactamase (ESBL)-producing

Enterobacteriaceae continues to rise in intensive care unit (ICU) patients in the United

States despite current infection control and antibiotic stewardship interventions.

However, the effect of universal glove and gown use on transmission of ESBL-producing

Enterobacteriaceae among ICU patients has not been evaluated.

Objective: The aim of this study was to assess the effectiveness of universal glove and gown use in reducing the transmission of ESBL-producing Enterobacteriaceae among

ICU patients in a non-outbreak setting.

Methods: A pretest-posttest quasi-experimental study was conducted among patients admitted to the University of Maryland Medical Center (UMMC) medical ICU (MICU) between July 1, 2005 and June 30, 2009. Peri-anal surveillance cultures were collected from all patients on ICU admission, weekly and at ICU discharge and clinical cultures were collected as medically indicated. Inclusion criteria were ICU length of stay >48 hours and having an ESBL-negative peri-anal culture on ICU admission. The unexposed group were patients admitted to the MICU before the period of universal glove and gown use (June 1, 2005 to June 30, 2007) while the exposed group were patients admitted to the MICU during the period of universal glove and gown use (July 1, 2007 to June 30,

2009). The outcome of interest was any ESBL-producing Enterobacteriaceae acquired

* Adebola O. Ajao, MPH;1 J. Kristie Johnson, PhD; 1,2 Anthony D. Harris, MD, MPH; 1 Min Zhan, PhD;1 Jessina C. McGregor, PhD; 3,4 Kerri A. Thom, MD, MS;1 Jon P. Furuno, PhD.1 1Department of Epidemiology and Public Health, 2Department of Pathology, University of Maryland School of Medicine, Baltimore, MD 3College of Pharmacy, Oregon State University, Portland, OR 4Oregon Health and Science University, Portland, OR

during the ICU stay. Multivariable poisson regression and segmented linear regression

were used to model the association between universal glove and gown use and

acquisition of ESBL-producing Enterobacteriaceae.

Results: There were 6,089 admissions to the MICU during the study period. Among

them, 353 (6%) admissions did not receive admission peri-anal surveillance cultures and

were excluded from the study population. 3,989 admissions (3,134 patients) stayed in the

ICU for > 48 hours, among which 3,384, 3,638, and 3,825 admissions were at risk for

acquiring any ESBL-Enterobacteriaceae, ESBL-Klebsiella and ESBL-E. coli

respectively. Unadjusted acquisition rates of any ESBL-Enterobacteriaceae, ESBL-

Klebsiella and ESBL-E. coli were 7.8 cases, 5.1 cases and 2.7 cases per 1000 patient days

at risk in the period before universal glove and gown use and 9.1 cases, 5.3 cases and 2.4

cases per 1000 patient days at risk during the period of universal glove and gown use.

Adjusted rate ratios and 95% confidence interval for acquiring any ESBL-GNR, ESBL-

Klebsiella and ESBL-E. coli in the universal glove and gown period compared to the

period before universal glove and gown were 1.13 (0.85 - 1.50), 1.01 (0.70 - 1.45), and

0.97 (0.58 - 1.63) respectively. Other variables significantly associated with acquisition

of any ESBL-Enterobacteriaceae after were older age, male sex, colonization pressure,

renal disease, hemi paraplegia, and anti-MRSA therapy received during the time at risk in

the MICU.

Conclusions: Given the unclear role of universal glove and gown as a barrier precaution to decrease the acquisition rate of ESBL-Enterobacteriaceae in the MICU, further studies using better methodology, such as cluster randomized trials may be needed to evaluate

the effectiveness of universal glove and gown use in reducing the acquisition of ESBL-

Enterobacteriaceae in ICU patients.

Introduction

Extended-spectrum β-lactamase (ESBL)-producing Enterobacteriaceae are emerging pathogens that are associated with considerable morbidity, mortality and costs among hospitalized patients globally.1,2 ESBLs are plasmid mediated enzymes that

hydrolyze most β-lactam antibiotics making these antibiotics ineffective.1 Infections caused by ESBL-producing Enterobacteriaceae are difficult to treat due to the limited options for both empiric and definitive therapy. Thus, there is an increased likelihood that empirically selected therapy may not provide coverage for the ESBL-producing

Enterobacteriaceae which subsequently increases the risk of treatment failure and poor outcomes among infected patients.2,11 Given the limited effective antimicrobial drugs currently available or in development there is a need for more effective infection control methods to prevent the spread of ESBL-producing Enterobacteriaceae among intensive care unit (ICU) patients.

Contact isolation with glove and gown use is a preventive methodology

recommended by infection control experts for patients colonized or infected with

Methicillin-resistant Staphylococcus aureus (MRSA), Vancomycin-resistant enterococcus

(VRE), Vancomycin-intermediate Staphylococcus aureus (VISA) or Vancomycin- resistant Staphylococcus aureus (VRSA).17 Studies that have evaluated the effectiveness of glove and gown use in reducing colonization rate with antibiotic-resistant bacteria have been limited to VRE studies. Some of these studies found that gown use in addition to glove use significantly reduced VRE colonization rate18,19 while others found no effect of glove and gown use on VRE colonization rate in ICU patients.123 Universal glove and

gown use is a form of contact precaution used for contacts with all patients and their

environment regardless of patients’ colonization or infection status. Universal glove and gown has the advantage that it may simultaneously prevent transmission of multiple antibiotic-resistant and susceptible bacteria without requiring active surveillance of each bacteria species.

Universal glove and gown use was instituted at the UMMC MICU in July 1, 2007 as part of the Centers for Disease Control (CDC) funded research and quality improvement program for antibiotic-resistant bacteria and it is currently being evaluated

for its effectiveness in reducing transmission of MRSA and VRE in ICU patients.

However, universal glove and gown has not been evaluated for its’ effectiveness in

reducing the transmission of ESBL-producing Enterobacteriaceae among ICU patients.

Infection control monitoring and control charting of clinical cultures for ESBL-

Klebsiella and ESBL-E.coli at the University of Maryland Medical Center (UMMC)

medical ICU (MICU) has shown no evidence of an outbreak, however 2% of patients

were colonized on ICU admission and 3% of patients acquired ESBL-Klebsiella or

ESBL-E. coli during their ICU stay indicating endemicity of these bacteria at

UMMC.80,81 The aim of this study was to evaluate the effectiveness of universal glove and gown use in reducing the transmission of ESBL-producing Enterobacteriaceae in

patients admitted to the MICU at UMMC. Henceforth ESBL-producing

Enterobacteriaceae will be referred to as ESBL-Gram-negative rods (GNR).

Methods

Study site

This study was performed using data from patients admitted to the UMMC MICU from

July 1, 2005 to June 30, 2009. UMMC is a 729-bed, urban tertiary-care hospital in

Baltimore, Maryland. The MICU was a 10-bed private room unit that became a 29-bed private room unit in April, 2006.

Selection of study population

Patients admitted to the UMMC MICU have obtained peri-anal cultures routinely

as part of an ongoing VRE active surveillance program since 1995. Peri-anal cultures

were obtained using Staplex II swabs (Staplex, Etobiocoke, Ontario, Canada) within 48

hours of ICU admission, weekly and at ICU discharge. Compliance with obtaining peri-

anal cultures has been shown to be approximately 90%.80,81 The study population

consisted of all patients admitted to the UMMC MICU during the study period who had

an ESBL-negative surveillance peri-anal culture on ICU admission and stayed in the ICU

for at least 48 hours and thus were at risk for acquiring an ESBL-GNR. Patients who did not receive admission surveillance culture were excluded from the study population as their ESBL status on admission to the MICU could not be established. Patients who were colonized with an ESBL-GNR upon admission to the MICU and patients who stayed in

the MICU for less than 48 hours were excluded from the study population as they had a reduced risk of acquiring an ESBL-GNR. Among patients who acquired an ESBL-GNR during their ICU stay, only the first positive surveillance or clinical culture was included in the analysis. Patients with multiple admissions to the MICU during the study period

were allowed to enter the cohort of at risk patients multiple times as long as they were not

positive for an ESBL-GNR on the index MICU admission.

Study design

A pretest-posttest quasi-experimental study was performed to assess acquisition rates of ESBL-GNR during two years of universal glove and gown use and two years before universal glove and gown use. The unexposed group included patients admitted to

the MICU before the period of universal glove and gown use (July 1, 2005 to June 30,

2007) while the exposed group included patients admitted to the MICU during the period

of universal glove and gown use (July 1, 2007 to June 30, 2009).

Study outcomes

The study outcomes were any ESBL-GNR, ESBL-Klebsiella, and ESBL-E. coli acquired

during the ICU stay as determined by peri-anal surveillance and clinical cultures. Patients

were considered to have acquired an ESBL-GNR in the MICU if a surveillance or clinical culture was positive for any ESBL-GNR after a documented ESBL negative surveillance culture during the index MICU admission. Prevalence of an ESBL-GNR on MICU admission was calculated by dividing the number of patients positive for an ESBL-GNR by the total number of patients admitted to the MICU during the study period. Unadjusted incidence rates for ESBL-GNR were calculated by dividing the number of ESBL-GNR acquired by the total number of admission days at risk during the study period.

Admission days at risk was calculated as the number of days from MICU admission to the first ESBL positive peri-anal surveillance or clinical culture for admissions where an

ESBL-GNR was acquired and the number of days from MICU admission to discharge for

admissions where an ESBL-GNR was not acquired during the ICU stay. Interrupted time

series defined as a sequence of values of a particular measure taken at regularly spaced

intervals over a time period was used to present the aggregated ESBL-GNR acquisition

rates every 2 month.107 ESBL-GNRs included in this study were Klebsiella pneumonia,

Klebsiella oxytoca, Eschericia coli, Enterobacter cloacae, Enterobacter aerogenes,

Citrobacter freundii, Proteus mirabilis, Providential stuartii, Morganella morganii and

Serratia marcescens. This study was approved by the University of Maryland (UMB)

Institutional Review Board (IRB).

Potential confounding variables and variable definitions

All patient data were collected retrospectively from the UMMC Central Data Repository

(CDR) and through laboratory testing. The UMMC CDR is a relational database

containing patients’ administrative, pharmacy, admission-discharge-transfer, and clinical

data. A random sample of the data from the UMMC CDR was validated against patients’

paper medical records for this study and previous research studies. Data from the CDR

was shown to have a positive and negative predictive value greater than 98%.91-93 Data collected to assess potential confounding included age, sex, total length of ICU stay, length of ICU time at risk, ICU individual antibiotic use, aggregate comorbid indices and

colonization pressure. Length of ICU stay was defined as the number of days from ICU

admission to ICU discharge. Length of ICU time at risk was defined as the number of

days from ICU admission to ESBL positive peri-anal surveillance or clinical culture for

patients who acquired an ESBL-GNR and number of days from ICU admission to ICU

discharge for patients who did not acquire an ESBL-GNR. ICU antibiotic exposure was

defined as the class of antibiotics ordered during the ICU time at risk. Antibiotics

exposure was categorized as being present or absent during the ICU time at risk. Classes

of antibiotics ordered included β-lactams, macrolides, aminoglycosides, antifungal agents, quinolone, and anti-MRSA. β-lactams collected were penicillin, 1st generation, 2nd

generation, 3rd generation and 4th generation cephalosporin, and carbapenems. Anti-

MRSA collected included vancomycin, linezolid, telavancin, tigecycline, daptomycin, quinupristin/dalfopristin, and dalbavancin. Aggregate comorbidity score, the Charlson

Comorbidity Index (CCI) was calculated using ICU discharge ICD-9 codes for comorbid conditions.89 Colonization pressure was defined as the average of the daily proportion of

patients positive for ESBL-GNR on ICU admission during each patient’s ICU time at

risk.90 Patients who were positive for ESBL-GNR on admission to the MICU were

retained to calculate colonization pressure but were excluded from the study population

for further analysis because they were no longer at risk for acquiring ESBL-GNR in the

MICU on the index admission. Patients who were colonized with ESBL-GNR upon

admission to the MICU were assumed to be colonized throughout their ICU stay.

Microbiology testing

ESBL colonization status of patients admitted to the MICU during the study period was

determined by culturing peri-anal swabs for ESBL-GNRs. Briefly, peri-anal cultures were innoculated onto a MacConkey agar plates (Remel, Lenexa, KS, USA) containing 1

μg/ml ceftazidime and incubated at 37 degrees C for 24 to 48 hours. Both lactose and non-lactose fermenting colonies were identified using the API 20E identification strips or the VITEK II (BioMerieux, Durham, NC). Susceptibility testing for the detection of

ESBL in all GNRs was performed using the Kirby-Bauer disk diffusion test method using

30 μg of ceftazidime and 30 μg cefotaxime with and without 10 μg of clavulanic acid as recommended by the Clinical Laboratory Standards Institute guideline.99 ESBL-positive clinical cultures were determined by the University of Maryland Medical Center microbiology laboratory and susceptibilities were performed following CLSI guidelines.99 Briefly, clinical cultures were tested against ceftiazone, a third generation

cephalosporin. All isolates in the intermediate resistance range (14-25mm) underwent

ESBL confirmatory testing with Kirby-Bauer disk diffusion method as described above.

Statistical analysis

Univariate analysis was performed to evaluate the distribution of all study variables. Bivariate analysis was performed to measure the association between study covariates and universal glove and gown use as well as the association between study covariates and acquisition of ESBL-GNRs. Normally distributed continuous variables

were compared using Student’s t-test while non-normally distributed continuous variables were compared using the Wilcoxon Rank Sum test. Categorical variables were compared using chi-square test or Fisher’s exact test. Poisson regression was used to

compare acquisition rates of ESBL-GNRs in the period before universal glove and gown

use and the period of universal glove and gown use. Stratified analysis was performed to

assess effect modification using the Breslow-Day test. All statistical tests were performed

using the 2-tailed test. All variables found to be associated with acquisition of ESBL-

GNR at P <0.1 by bivariate analysis were entered into a multivariable analysis. Variables

with P-value >0.05 were removed from the multivariable model sequentially starting with

the variable with the largest P-value. To assess confounding, we assessed if the removal of the variable changed the regression coefficient of universal glove and gown use by more than 10%. Since patients were allowed to enter the study multiple times; we adjusted for the correlated error structure of the data by using the robust variance estimator in the multivariable analysis.

Poisson regression was used to calculate the rate ratio (RR) and 95% CI for the association between universal glove and gown use and acquisition of ESBL-GNRs controlling for confounding variables. Multivariable analysis was performed and reported separately for all ESBL-GNRs, ESBL-Klebsiella and ESBL-E. coli due to some evidence that ESBL-Klebsiella may be more likely to be transmitted from patient-to-patient than other GNR.80,81 Aggregate level analysis was performed using segmented linear regression to estimate changes in the mean (immediate effect of glove and gown use on the acquisition rate of ESBL-GNR) and changes in the trend (long term effect of glove and gown use on the acquisition rate of ESBL-GNR).108 Aggregate level analysis was performed using segmented linear regression rather than segmented Poisson regression in order to assess and control for any autocorrelation. Autocorrelation was assessed by plotting the residuals against time and visually inspecting the plot for autocorrelation.

Durbin-Watson statistics was estimated to account for any autocorrelation.107 Aggregate level analysis was performed for all ESBL-GNRs, ESBL-Klebsiella and ESBL-E. coli.

Result

There were 6,089 admissions to the MICU during the 48-month study period. Among them, 353 (6%) admissions did not receive admission peri-anal surveillance culture and

were excluded from the study population. Among patients that received admission peri-

anal surveillance culture, 1,747 (29%) admissions had ICU stay of less than 48 hours and

were excluded from the study population. Three thousand, nine hundred and eighty-nine admissions (3,134 patients) were in the ICU for at least 48 hours. Among them, 605 admissions were colonized with an Enterobacteriaceae harboring an ESBL on admission to the MICU leaving 3,384 admissions (2,837 patients) at risk for acquiring an ESBL-

GNR, 351 admissions (255 patients) were colonized with ESBL-Klebsiella on admission

to the MICU leaving 3,638 admissions (2,982 patients) at risk for acquiring ESBL-

Klebsiella, 164 admissions (134 patients) were colonized with ESBL-E. coli on admission to the MICU leaving 3,825 admissions (3,050 patients) at risk for acquiring

ESBL-E. coli (Figure V.1).

Figure V.1: Selection of Study Population

The prevalence of ESBL-GNR, ESBL-Klebsiella and ESBL-E. coli at MICU

admission during the period before glove and gown use were 237 (13%), 131 (7.1%), 86

(4.6%) of 1859 patients. The prevalence of ESBL-GNR, ESBL-Klebsiella and ESBL-E. coli during the period of glove and gown use were 244 (11%), 157 (6.8%), 71 (3.1%) of

2276 patients. Acquisition rates of ESBL-GNR, ESBL-Klebsiella and ESBL-E. coli were

7.8 cases, 5.1 cases and 2.7 cases per 1000 patient days at risk during the period before universal glove and gown use and 9.1 cases, 5.3 cases and 2.4 cases per 1000 patient days at risk during the period of glove and gown use. Demographic and clinical characteristics of patients who received surveillance peri-anal culture on admission to the MICU and stayed in the MICU for at least 48 hours by study period are displayed in Table V.1.

Table V.1: Characteristics of the Study Population by Study Period Patients characteristics Pre-universal Universal glove P value

glove and gown and gown use

use N=2252

N=1737

Age (mean in years, SD) 54.5 (15.5) 54.5 (16.2) 0.91

Male sex (n, %) 949 (54.6%) 1227 (54.5%) 0.93

Prevalence of ESBL-GNR colonization on MICU admission (n, %) 281 (16.2) 324 (14.4) 0.12

ESBL-GNR acquired in MICU (n, %) 210 (12.1%) 320 (14.2%) 0.05

MICU length of stay (median, IQR) 4 (2 - 7) 4 (3 - 9) < 0.01

MICU time at risk (median, IQR) 4 (2 - 7) 4 (3 - 8) < 0.01

Colonization pressure (mean, SD) 0.19 (0.06) 0.20 (0.06) < 0.01

Comorbidities (n, %)

AIDS 147 (8.4) 157 (7.4) 0.22

Malignancy 276 (15.8) 400 (17.6) 0.11

Cerebrovascular Disease 221 (12.7) 274 (12.1) 0.59

Chronic Pulmonary Disease 413 (23.7) 496 (22.0) 0.19

Heart Failure 334 (19.2) 326 (14.4) < 0.01

Diabetes 273 (15.7) 419 (18.6) 0.01

Complicated Diabetes 70 (4.0) 77 (3.4 ) 0.31

Hemi paraplegia 83 (4.7) 88 (3.9) 0.17

Metastatic tumor 114 (6.5) 147 (6.5) 0.96

Mild Liver 138 (7.9 ) 126 (3.6 ) < 0.01

Severe Liver 131 (7.5) 176 (7.8) 0.74

Myocardial Infarction 119 (6.8) 167 (7.4) 0.49

Peptic Ulcer 54 (3.1) 57 (2.5) 0.27

Peripheral vascular 50 (2.8) 81 (3.6) 0.20

Renal Disease 359 (20.6) 496 (22.0) 0.30

Rheumatology 38 (2.1) 48 (2.1 ) 0.90

Charlson Co-morbidity (median, IQR) 2.0 (1 - 4) 2.0 (1 - 4) 0.54

Chronic disease score 8.3 (4.0) 7.4 (4.9) < 0.01

MICU antibiotic use (n, %)

Any antibiotic 1451 (83.5) 1909 (84.8) 0.29

β-lactam 1109 (63.9) 1490 (66.2) 0.13

Penicillin 817 (47.0) 1148 (50.9) 0.01

Anti-pseudomonal β-lactam 1141 (65.7) 1655 (73.5) < 0.01

Anti-anaerobic β-lactam 1097 (63.2) 1566 (69.5) < 0.01

Piperacillin/Tazobactam 740 (42.6) 1083 (48.1) < 0.01

Cephalosporin 495 (28.5) 667 (29.6) 0.44

1st Generation cephalosporin 133 (7.7) 143 (6.4) 0.11

2nd Generation cephalosporin 12 (0.7) 9 (0.4) 0.21

3rd Generation cephalosporin 242 (13.9) 301 (13.4) <0.61

Cefepime 171 (9.8) 294 (13.1) < 0.01

Imipenem 270 (15.5) 263 (11.7) <0.01

Anti-fungal 367 (21.1) 462 (20.5) 0.64

Aminoglycosides 113 (6.5) 170 (7.6) 0.20

Macrolides 259 (14.9) 278 (12.3) 0.02

Quinolones 418 (24.1) 516 (22.9) 0.39

Anti-MRSA therapy 949 (54.6) 1396 (61.9) < 0.01

Vancomycin 874 (50.3) 1282 (56.9) < 0.01

ESBL, extended-spectrum β-lactamase; GNR, Gram-negative rods; IQR, Interquartile range; MICU, medical ICU; MRSA, Methicillin-resistant staphylococcus aureus; SD, Standard deviation;

The rate of acquiring ESBL-GNR for each study period was assessed by Poisson regression. The unadjusted rate ratio (RR) and 95% confidence interval (CI) of acquiring

ESBL-GNR, ESBL-Klebsiella and ESBL-E. coli in the universal glove and gown period

compared to the period before universal glove and gown were 1.16 (95% CI, 0.86 - 1.54),

1.04 (95% CI, 0.73 - 1.49) and 0.89 (95% CI, 0.54 - 1.51) respectively. We examined if

the effect of universal glove and gown use on the acquisition of ESBL-GNR was different for different levels of colonization pressure but no effect modification by colonization pressure was identified for all three ESBL outcomes. Variables associated

with each of the ESBL outcomes by bivariate analysis were entered into the multivariable

model. Results of the Poisson regression are displayed in Table V.2. Adjusted rate ratios

(ARR) and their corresponding 95% CIs were 1.13 (0.85 - 1.50), 1.01 (0.70 - 1.45), and

0.97(0.58 - 1.63) for ESBL-GNR, ESBL-Klebsiella and ESBL-E. coli respectively in the

universal glove and gown period compared to the period before universal glove and gown

use. Variables found to be significantly associated with acquisition of ESBL-GNR after

controlling for confounding variables in this study were male sex (ARR = 1.53, 95% CI

=1.14 - 2.04), length of stay at risk < 4 days (ARR = 1.45, 95% CI = 1.02 - 2.08), hemi

paraplegia (ARR = 1.82, 95% CI= 1.09 - 3.06), renal disease (ARR = 1.53, 95% CI =

1.11 - 2.09), malignancy (ARR = 0.57, 95% CI = 0.36- 0.89), third generation

cephalosporin (ARR= 0.51, 95% CI = 0.32 - 0.81), antifungal agent (ARR =0.66, 95% CI

(0.49 - 0.97) and anti-MRSA therapy received in the ICU (ARR = 1.82, 95% CI = 1.12 -

2.43). Variables found to be significantly associated with acquisition of ESBL-Klebsiella

were older age (ARR = 1.01, 95% CI= 1.00 -1.03), length of time at risk (ARR= 2.33,

95% CI=1.58 - 3.45), malignancy (ARR= 0.42, 95% CI =0.23 - 0.78), chronic pulmonary

(ARR=0.58, 95% CI=0.35 - 0.95), first-generation cephalosporin (ARR = 0.16, 95% CI =

0.05 - 0.50), third-generation cephalosporin (ARR=0.42, 95% CI = 0.19 - 0.92),

antianaerobe β-lactam (ARR= 0.54, 95% CI = 0.29 - 0.98), macrolides (ARR= 0.47, 95%

CI = 0.25 - 0.90), aminoglycosides (ARR= 1.91, 95% CI = 1.22 - 2.98), ICU anti-MRSA

therapy (ARR = 2.72, 95% CI 1.82 - 4.06), and anti-psudomonal β-lactam (ARR= 3.18,

95% CI = 1.40-7.23). Variables found to be significantly associated with ESBL-E. coli

acquisition were male sex (ARR = 2.06, 95% CI =1.18 - 3.60), length of stay at risk

(ARR = 2.22, 95% CI = 1.29 - 3.85), colonization pressure (ARR = 1.85, 95% CI = 1.10 -

3.09, and hemi paraplegia (ARR = 4.67, 95% CI 2.21 - 9.81).

Table V.2: Poisson Regression Model of the Association between Universal glove and gown use and acquisition of ESBL-producing Enterobacteriacea among patients admitted to the MICU. Study Variables RR (95% CI) P-value ESBL-GNR Universal glove and gown use 1.13 (0.85 - 1.50) 0.40 Male sex 1.53 (1.14 - 2.04) <0.01 Length of stay at risk (<4 days) 1.45 (1.02 - 2.08) 0.03 Hemi paraplegia 1.82 (1.09 - 3.06) 0.01 Renal Disease 1.53 (1.11 - 2.09) <0.01 Malignancy 0.57 (0.36 - 0.89) 0.01 Third generation cephalosporin 0.51 (0.32 - 0.81) <0.01 Antifungal Agent 0.66 (0.49 - 0.97) 0.03 Anti-MRSA therapy 1.65 (1.12 - 2.43) 0.01 ESBL-Klebsiella Universal glove and gown use 1.01 (0.70 - 1.45) 0.95 Age 1.01 (1.00 -1.03) 0.03 Length of stay at risk (<4 days) 2.33 (1.58 - 3.45) <0.01 Malignancy 0.42 (0.23 - 0.78) <0.01 Chronic pulmonary 0.58 (0.35 - 0.95) 0.03 First-generation cephalosporin 0.16 (0.05 - 0.50) <0.01 Third-generation cephalosporin 0.42 (0.19 - 0.92) 0.03 Anti-anaerobe β-lactam 0.54(0.29 - 0.98) 0.04 Macrolides 0.47(0.25 - 0.90) 0.02 Aminoglycosides 1.91 (1.22 - 2.98) <0.01 Anti-MRSA therapy 2.72 (1.82 - 4.06) <0.01 Anti-pseudomonal β-lactam 3.18 (1.40-7.23) <0.01 ESBL-E. coli Universal glove and gown use 0.97 (0.58 - 1.63) 0.91 Male sex 2.06 (1.18 - 3.60) 0.01 Length of stay at risk (>4 days) 2.22 ( 1.29 - 3.85) <0.01 Colonization pressure 1.85(1.10 - 3.09) 0.02 Hemi paraplegia 4.67 (2.21- 9.81) <0.01

ESBL, extended-spectrum-beta-lactamase; ICU, intensive care unit; MRSA, methicillin-resistant staphylococcus aureus; RR, Rate ratio; CI, confidence interval. **Antimicrobial drug exposure was assessed during the ICU time at risk.

The time series plot for the acquisition rate of all ESBL-GNR, ESBL-Klebsiella and ESBL-E. coli every 2 months during the study period are displayed in Figures V.1. A positive autocorrelation with time was detected for ESBL-GNB (p=0.0120), ESBL-

Klebsiella (p=0.005) and ESBL-E. coli (p=0.0013) and adjusted for in the final

segmented linear regression. The result of the segmented regression is displayed in

Tables V.3. For ESBL-GNR, in the two years before universal glove and gown use, there was a non-statistically significant decrease in all ESBL-GNR acquisitions (p =0.35).

Universal glove and gown use did not have a significant effect on the level (p =0.48) or the trend (p =0.34) of ESBL-GNR acquisition rate. For ESBL-Klebsiella, in the two years before universal glove and gown use, there was a non-statistically significant decrease in all ESBL-Klebsiella acquisition rate (p =0.31). Universal glove and gown use did not have a significant effect on the level (p =0.56) or the trend (p =0.07) of ESBL- Klebsiella acquisition rate. For ESBL-E. coli, in the two years before universal glove and gown use, there was a non-statistically significant decrease in all ESBL- E. coli acquisition rate (p

=0.42). Universal glove and gown use did not have a significant effect on the level (p

=0.34) or the trend (p =0.19) of ESBL- E. coli acquisition rate. No seasonality effect was

detected using the Dickey-Fuller unit root test for ESBL-GNR, ESBL-Klebsiella or

ESBL-E. coli (p <=0.05).

Figure V.2 Time series plot for ESBL-GNR, ESBL-Klebsiella and ESBL-E. coli acquisition rate July 2005 - June 2009

Micu Expansion

Universal glove and gown Intervention

Jul-Aug 05

Table V.3 Segmented Linear Regression Model of the effect of universal glove and gown use on the acquisition rate of EBSL-GNR, ESBL-Klebsiella and ESBL-E. coli in patients admitted to the UMMC MICU. Coefficient Standard t-statistic P-value Error ESBL-GNR Intercept (βo) 0.009767 0.003439 2.84 0.01 Baseline Trend (β1) -0.000736 0.000772 -0.95 0.35 Level change after intervention (β2) 0.003050 0.004215 0.72 0.48 Trend change after intervention (β3) 0.000927 0.000940 0.99 0.34 Colonization pressure 0.000108 0.000211 0.51 0.62 ESBL-Klebsiella Intercept (βo) 0.005343 0.000686 7.79 < 0.01 Baseline Trend (β1) -0.000231 0.000219 -1.06 0.31 Level change after intervention (β2) -0.000619 0.001043 -0.59 0.58 Trend change after intervention (β3) 0.000473 0.000242 1.95 0.07 Colonization pressure 0.000035 0.000054 0.65 0.52 ESBL-E. coli Intercept (βo) 0.003191 0.001887 1.69 0.11 Baseline Trend (β1) -0.000340 0.000243 - 0.83 0.42 Level change after intervention (β2) -0.002212 0.002259 -0.98 0.34 Trend change after intervention (β3) 0.000702 0.000514 1.37 0.19 Colonization pressure 0.000122 0.000095 1.28 0.22

Discussion

The aim of this study was to evaluate the effectiveness of universal glove and

gown use in reducing the transmission of ESBL-producing Enterobacteriaceae in

patients admitted to the MICU at UMMC. Our study results suggest that universal glove

and gown use was not associated with a decrease in the acquisition rate of ESBL-

producing Enterobacteriaceae immediately after the intervention or long term among patients admitted to MICU during the study period. To our knowledge, this is the first study to evaluate the impact of universal glove and gown use on the acquisition of ESBL-

GNR in ICU patients. However, other studies have evaluated the effectiveness of

universal glove and gown in reducing the transmission of MRSA and VRE and have not

found this intervention to be effective.127

The strengths of this study include the use of interrupted time series; the strongest

quasi-experimental design and segmented regression for analyzing the time series data.

Segmented regression used in this study allowed us to assess both the immediate and long

term effect of universal glove and gown use on the acquisition of ESBL-GNR. Second, this study spanned a period of four years, which enabled us to statistically check that any changes in the incidence rate of ESBL-GNR were not due to periodic changes. Third, we were able to control for any effect of increasing prevalence of ESBL-GNR over time by adjusting for colonization pressure defined as the percentage of patients colonized with

ESBL-GNR on ICU admission.

There are possible explanations why we may not have detected a benefit from universal glove and gown use if a benefit truly exists. First, we were not able to consistently evaluate compliance with glove and gown use throughout the study period.

Most previous reports of compliance with glove or gown use have been low, ranging from 25% - 59%.116-118 In this study, compliance was measured between 2008 and 2009

and was estimated to be 90%. Thus we were not able to accurately assess if compliance

was sufficient throughout the study period to provide a true measure of the efficacy of

glove and gown use. Second, it is possible that there might have been changes in other practices such as a decrease in hand hygiene when universal glove and gown use was implemented. However, we were not able to control for the effect of changes in these other practices in our analysis because these data were not collected during the study period. Third, there was already a downward trend in the acquisition rate for all ESBL-

GNR before the intervention. Previous studies have shown that interventions are less

likely to be successful if there is evidence that the outcome is already changing in the

desired direction prior to implementing the intervention.126

To further explain our study results, we visually examined the time-series plot to

assess if the MICU expansion from a 10-bed unit to a 29-bed unit in April 2006 had an

impact on the acquisition rate of all ESBL-producing Enterobacteriaceae, however, we did not observe any long term increase in acquisition rate of ESBL-producing

Enterobacteriaceae after the MICU expansion.

Factors found to be associated with acquisition of ESBL-producing

Enterobacteriaceae identified in this study were older age, male sex, higher colonization

pressure, renal disease, hemi paraplegia, aminoglycosides, anti-pseudomonal β-lactam,

anti-MRSA therapy received during the time-at risk in the ICU. Renal disease, receipt of

aminoglycosides and anti-MRSA therapy is consistent with what has been previously

reported in the literature. 47,65,88,119,127

In conclusion, given the unclear role of universal glove and gown use as a barrier

precaution in reducing the acquisition of ESBL-GNR in our study, further studies using

better methodology, such as cluster randomized trials may be needed to evaluate the

effectiveness of universal glove and gown use in reducing acquisition of ESBL-GNR in

ICU patients. Future studies should ensure that compliance with the universal glove and

gown use is measured throughout the study period. In addition, other infection control

interventions occurring concurrently during the study period should be measured

throughout the study period and adjusted for in the analysis to accurately assess the

effectiveness of universal glove and gown use.

C. Discussion

A. Extended spectrum β-lactamase producing Gram-negative rods

The prevalence of ESBL-GNR among ICU patients continues to rise causing

considerable morbidity, mortality and costs among hospitalized patients in the United

States and globally.1,2,15,16 Infections caused by ESBL-GNR are difficult to treat due to the limited options for empiric therapy which increases the risk of treatment failure and death among infected patients.2,11 Environmental contamination by patients colonized or

infected with ESBL-GNR has been documented12,13 however, the association between

environmental contamination and acquisition of ESBL-GNR among hospitalized patients

in non-outbreak situations has not been well established. Given the limited effective

antimicrobial drugs currently available or in development there is a need for more effective infection control methods to prevent the spread of ESBL-GNR among ICU

patients.

Universal glove and gown use is a form of contact precaution where gloves and

gowns are used for contacts with all patients and their environment regardless of the

patient’s colonization or infection status. Universal glove and gown has the advantage

that it may prevent transmission of multiple antibiotic-resistant and susceptible bacteria at

the same time without requiring active surveillance for each bacteria species. At UMMC,

universal glove and gown use was instituted in the MICU in July, 1 2007 and is currently

being evaluated for its effectiveness in reducing transmission of MRSA and VRE in ICU

patients, however, universal glove and gown has not been evaluated for it’s effectiveness

in reducing the transmission of ESBL-GNR among ICU patients. The overall objectives

of this dissertation are to evaluate the role of persistent environmental contamination in

the acquisition of ESBL-GNR and to evaluate the effectiveness of universal glove and gown use in reducing the transmission of ESBL-GNR in ICU patients while controlling for other patient characteristics.

B. Prior room occupant’s ESBL-Klebsiella or E. coli positive status was not

significantly associated with acquisition of ESBL-Klebsiella or E. coli

Using a cohort of 18,175 admissions, our data suggested that prior room occupant’s ESBL-Klebsiella or E. coli positive status was not significantly associated with an increased risk of acquiring ESBL-Klebsiella or E. coli among ICU patients.

However, other factors such as longer time at risk, higher colonization pressure, having

renal disease and hemi paraplegia, receipt of aminoglycosides, anti-psedomonal beta-

lactam, and anti-MRSA agents in the ICU were identified as predictors of ESBL-

Klebsiella or E. coli acquisition in ICU patients.

Due to the retrospective nature of our study, prior room occupants’ ESBL positive

status was used as proxy for environmental contamination. Prior studies have shown an

association between prior room occupants’ positive status and acquisition of MRSA,

VRE, MDR-P. aeruginosa and A. baumannii among ICU patients.74,109,110,122 We may not

have been able to demonstrate an association between prior room occupant’s ESBL-

Klebsiella or E. coli positive status and acquisition of ESBL-Klebsiella or E. coli because

prior room occupants’ ESBL status may not be a valid proxy for environmental

contamination with ESBL-Klebsiella and E. coli. First, patients colonized or infected

with ESBL-Klebsiella and E. coli may not shed a high enough innoculum into their

environment to be transmitted to subsequent room occupants. Second, cleaning of ICU

rooms using quaternary ammonium may be efficient in eradicating ESBL-Klebsiella and

E. coli from the environment. Current literature suggests that Klebsiella and E. coli can

persist in the environment for months but does not suggest that Klebsiella and E. coli

persist for a shorter duration of time than other Gram-negative bacteria such as

Acinetobacter spp. and Pseudomonas aeruginosa.125 This implies that infection control

interventions for ESBL-GNR may need to focus on other modes of transmission besides

transmission through the environment. To further explore this hypothesis, future research

may need to evaluate what innoculum of ESBL-Klebsiella and E. coli is needed in the

environment to colonize or infect subsequent room occupants, correlate environmental

culturing data with prior room occupants’ ESBL status and employ molecular methods to

validate that it is the same bacterial strain that is being transmitted from the environment

to subsequent room occupants. This information is important to properly assess if the

current standard of ICU room cleaning is sufficient to prevent transmission of ESBL-

GNR from the environment to patients. This study highlights the importance of keeping

the burden of patients colonized or infected with ESBL-GNR (colonization pressure) low

in the ICU to reduce the risk of patient-to-patient transmission of ESBL-GNR. This study

also highlights the importance of keeping the length of ICU stay short and limited use of

antibiotics such as anti-MRSA therapy, aminoglycosides, anti-pseudomonal beta-lactam,

and imipenem in ICU patients to reduce the risk of acquiring ESBL-GNRs.

C. Universal glove and gown use was not associated with decreased acquisition of

ESBL-Enterobacteriaceae in ICU patients.

Using a cohort of 6,089 ICU admissions, our data suggested that universal glove and gown use is not associated with a decrease in the acquisition of all ESBL-producing

Enterobacteriaceae short term or long term. However, other factors such as older age, male sex, higher colonization pressure, having renal disease and hemi paraplegia and receipt of aminoglycosides, anti-pseudomonal β-lactam, anti-MRSA therapy specifically vancomycin prior to ESBL acquisition in the ICU were identified as risk factors for acquiring ESBL-producing Enterobacteriaceae in the MICU.

In this study, compliance was measured between 2008 and 2009 and was estimated to be

90%. Thus we were not able to accurately assess if compliance was sufficient throughout the study period to provide a true measure of the efficacy of glove and gown use. Second, it is possible that there might have been changes in other practices such as a decrease in hand hygiene when universal glove and gown use was implemented. However, we were not able to control for the effect of these other practices in our analysis because these data were not collected during the study period. Third, there was already a downward trend in the acquisition rate for all ESBL-GNRs before the intervention. Previous studies have shown that interventions are less likely to be successful if there is evidence that the outcome is already changing in the desired direction prior to implementing the intervention.126

In conclusion, given the unclear role of universal glove and gown use as a barrier precaution in reducing the acquisition of ESBL-GNR in our study, further studies using better methodology, such as cluster randomized trials may be needed to evaluate the effectiveness of universal glove and gown use in reducing acquisition of ESBL-GNR in

ICU patients. Future studies should ensure that compliance with the universal glove and gown use is measured throughout the study period. Other infection control interventions occurring concurrently during the study period should also be measured throughout the study period and adjusted for in the analysis to accurately assess the effectiveness of universal glove and gown use.

D. Strengths and Limitations

Both studies address important public health questions using a large cohort of

ICU patients. These studies are methodologically strong in that they controlled for known potential confounders to increase the internal validity of the study results. The study that evaluated the impact of persistent environmental contamination on acquisition of ESBL-

GNR was strengthened by its’ use of molecular method to quantify the extent of genetic

similarity between ESBL positive bacterial isolates obtained from prior room occupants

and ESBL positive bacterial isolates acquired during ICU stay. To our knowledge, this

was the first study to identify colonization pressure, an important infection control

variable as a risk factor for acquisition of ESBL-GNR in ICU patients. The study that

evaluated the effectiveness of universal glove and gown use in decreasing the incidence rate of ESBL-GNR in ICU patients was the first study to evaluate this intervention for

ESBL-GNR. This study included the use of interrupted time series; the strongest quasi-

experimental design and segmented regression for analyzing the time series data. The

segmented regression analysis allowed us to assess both the immediate and long term

effect of universal glove and gown use on the acquisition of ESBL-GNR. This study also

spanned a period of four years, which enabled us to statistically check that any changes in

the incidence rate of ESBL-GNR were not due to periodic changes. In addition, we were

able to control for any effect of increasing prevalence of ESBL-GNR over time by

adjusting for colonization pressure.

Both studies identified other risk factors that are associated with acquisition of

ESBL-GNR in ICU patients. Both study results may be generalizable to other ICUs with

similar patient population and infection control practices.

Our studies are limited by the following reasons. First, sensitivity of peri-anal cultures for ESBL detection compared to stool culture is unknown. It is possible that some ESBL positive cases may have been undetected by our surveillance method and may have led to misclassification of ESBL positivity status. However sensitivity of peri- anal culture compared to stool culture (gold standard) for fluoroquinolone-resistant E. coli was 90%.105 Therefore, we anticipated that the sensitivity of peri-anal cultures for

ESBLs would be similar and the proportion of false negatives would be low (~10%).

Therefore, we anticipate that the level of misclassification of ESBL positivity status by our surveillance methods would be low and non-differential and would have little potential to bias our measure of effect towards the null. Second, not all clinical cultures obtained from patients during the study period were tested for ESBLs. Only patients with bacterial isolates from clinical cultures susceptible to or with intermediate resistance to third generation cephalosporin were tested for ESBLs. Therefore, some patients who are

ESBL positive might have been misclassified as ESBL negative. However, since not all resistance to third generation cephalosporins are due to ESBLs and 68-82% of patients with ESBL positive clinical cultures are also colonized with ESBLs,8,104 we anticipated that the level of non-differential misclassification and its’ potential to bias our measure of effect towards the null would be low. Third, some bacteria isolates possess multiple resistant mechanisms. The presence of both ESBL and AMP C resistant mechanisms in bacterial isolates may mask ESBL expression and result in a negative ESBL confirmation test. This may have led to misclassification of ESBL positivity status, however, this phenomenom is rare and we expect that the extent of non-differential misclassification of

ESBL positivity status due to the presence of AMP C gene would be low.106

100

The first study is limited by using prior room occupant’s ESBL status as a proxy for environmental contamination. Hospital room environmental surfaces were not cultured in this study and thus we were not able to describe the level of environmental contamination under which acquisition of ESBL-Klebsiella and E. coli occurred. The second study was limited by the lack of data on compliance with glove and gown use and hand hygiene throughout the study period. Thus we were not able to accurately assess the true efficacy of glove and gown use by controlling for the effect of intervention compliance and hand hygiene in our data analysis.

101

E. Summary and Implications of the Research Questions

In summary, in this dissertation, we observed that 1) prior room occupant’s

ESBL-Klebsiella or E. coli positive status is not significantly associated with an

increased risk of acquiring ESBL-Klebsiella or E. coli. 2) Universal glove and gown use

is not associated with a decrease in the acquisition of ESBL-producing

Enterobacteriaceae short term or long term. In these studies, we identified other risk

factors such as older age, male sex, longer time at risk, higher colonization pressure,

having renal disease and hemi paraplegia and the receipt of aminoglycosides, anti-

pseudomonal beta-lactam, anti-MRSA in the ICU that are associated with acquisition of

ESBL-GNR in ICU patients.

Identifying these risk factors are important to identify patients that are most likely

to acquire ESBL-GNR and to understand the conditions under which patients are most

likely to acquire ESBL-GNR and in the ICU. This study highlights the importance of

keeping the burden of patients colonized or infected with ESBL-GNR (colonization

pressure) low in the ICU to reduce the risk of transmitting ESBL-GNR. This study also

highlights the importance of keeping the length of stay in the ICU short, and limited use

of antibiotics such as aminoglycosides, anti-pseudomonal beta-lactam, and anti-MRSA therapy in ICU patients to reduce the risk of acquiring ESBL-GNR.

It is still unclear if the environment plays a significant role in the transmission of

ESBL-GNR in ICU patients or if the current standard of cleaning of ICU rooms is adequate. Therefore, it is important for future research to determine what innoculum of

ESBL-GNR is needed on the ICU room surfaces to colonize or infect subsequent room

102

occupants, correlate environmental culturing data with prior room occupants’ ESBL

status and employ molecular methods to validate that it is the same bacterial strain that is

being transmitted from the environment to subsequent room occupants.

The prevalence of ESBL-GNR continues to rise in ICU patients in the United

States despite the current infection control efforts. Yet, there is limited data on optimal

infection control methods to prevent the transmission of ESBL-GNR in non-outbreak situations. Understanding the infection control methods that are effective against the transmission of ESBL-GNR is imperative to a successful infection control program aimed at reducing the incidence of colonization, infection, and poor healthcare outcomes associated with ESBL-GNR in ICU patients. Given the unclear role of universal glove and gown use as a barrier precaution in reducing the acquisition of ESBL-GNR in our study, further studies using better methodology, such as cluster randomized trials are needed to evaluate the effectiveness of universal glove and gown use in reducing the acquisition of ESBL-GNR in ICU patients. Future studies should ensure that other infection control interventions occurring concurrently during the study period and compliance with these interventions are measured throughout the study period and adjusted for in the analysis to accurately assess the effectiveness of universal glove and gown use.

103

D. Appendix

ESBL gene family present in isolates obtained from patients who acquired ESBL- Klebsiella and E. coli and their prior room occupants colonized with the same bacterial species. Patient Bacterial ESBL gene family in ESBL gene family in Percent Pair species isolates from prior room acquired isolates Similarity occupants by PFGE 1. K. pnemoniae SHV SHV, TEM, CTX 45% 2. K. pnemoniae SHV, CTX SHV 30% 3. K. pnemoniae SHV SHV 50% 4. E. coli No ESBL genes detected No ESBL genes detected 20% 5. K. pnemoniae SHV, CTX SHV 58% 6. K. pnemoniae SHV, TEM SHV, TEM 93% 7. K. pnemoniae SHV SHV, CTX 48% 8. E. coli CTX CTX 40% 9. E. coli No ESBL genes detected No ESBL genes detected 100% 10. K. pnemoniae SHV, CTX SHV, CTX 79% 11. E. coli CTX CTX 88% 12. K. pnemoniae SHV, CTX SHV, CTX 100% 13. K. pnemoniae SHV, TEM, CTX SHV, 56% 14. K. pnemoniae SHV, TEM SHV 80% 15. K. pnemoniae SHV, CTX SHV, CTX, TEM 58% 16. K. pnemoniae SHV, TEM SHV, TEM 100% 17. K. pnemoniae SHV, CTX, TEM SHV, CTX, TEM 54% ESBL, extended-spectrum-beta-lactamase; PFGR, pulsed field gel electrophoresis

104

References

1. Paterson DL, Bonomo RA. Extended-spectrum beta-lactamases: A clinical update. Clin Microbiol Rev. 2005;18:657-686.

2. Ramphal R, Ambrose PG. Extended-spectrum beta-lactamases and clinical outcomes: Current data. Clin Infect Dis. 2006;42 Suppl 4:S164-72.

3. Jacoby GA, Medeiros AA. More extended-spectrum beta-lactamases. Antimicrob Agents Chemother. 1991;35:1697-1704.

4. Podschun R, Ullmann U. Klebsiella spp. as nosocomial pathogens: Epidemiology, taxonomy, typing methods, and pathogenicity factors. Clin Microbiol Rev. 1998;11:589- 603.

5. Heritage J, M'Zali FH, Gascoyne-Binzi D, Hawkey PM. Evolution and spread of SHV extended-spectrum beta-lactamases in gram-negative bacteria. J Antimicrob Chemother. 1999;44:309-318.

6. Bradford PA. Extended-spectrum beta-lactamases in the 21st century: Characterization, epidemiology, and detection of this important resistance threat. Clin Microbiol Rev. 2001;14:933-51, table of contents.

7. Soulier A, Barbut F, Ollivier JM, Petit JC, Lienhart A. Decreased transmission of enterobacteriaceae with extended-spectrum beta-lactamases in an intensive care unit by nursing reorganization. J Hosp Infect. 1995;31:89-97.

8. Lucet JC, Chevret S, Decre D, et al. Outbreak of multiply resistant enterobacteriaceae in an intensive care unit: Epidemiology and risk factors for acquisition. Clin Infect Dis. 1996;22:430-436.

9. Pena C, Pujol M, Ricart A, et al. Risk factors for faecal carriage of klebsiella pneumoniae producing extended spectrum beta-lactamase (ESBL-KP) in the intensive care unit. J Hosp Infect. 1997;35:9-16.

10. Pena C, Pujol M, Ardanuy C, et al. Epidemiology and successful control of a large outbreak due to klebsiella pneumoniae producing extended-spectrum beta-lactamases. Antimicrob Agents Chemother. 1998;42:53-58.

11. Hogg GG, Forsyth JR, Hibberd J, McBride J. Importance of resistant klebsiella spp. in victoria. Med J Aust. 1993;158:722.

12. Kac G, Podglajen I, Vaupre S, Colardelle N, Buu-Hof A, Gutmann L. Molecular epidemiology of extended-spectrum beta-lactamase-producing enterobacteriaceae isolated from environmental and clinical specimens in a cardiac surgery intensive care unit. Infect Control Hosp Epidemiol. 2004;25:852-855.

105

13. Hobson RP, MacKenzie FM, Gould IM. An outbreak of multiply-resistant klebsiella pneumoniae in the grampian region of scotland. J Hosp Infect. 1996;33:249-262.

14. Flaherty JP, Weinstein RA. Nosocomial infection caused by antibiotic-resistant organisms in the intensive-care unit. Infect Control Hosp Epidemiol. 1996;17:236-248.

15. Hidron AI, Edwards JR, Patel J, et al. NHSN annual update: Antimicrobial-resistant pathogens associated with healthcare-associated infections: Annual summary of data reported to the national healthcare safety network at the centers for disease control and prevention, 2006-2007. Infect Control Hosp Epidemiol. 2008;29:996-1011.

16. National Nosocomial Infections Surveillance System. National nosocomial infections surveillance (NNIS) system report, data summary from january 1992 through june 2004, issued october 2004. Am J Infect Control. 2004;32:470-485.

17. Muto CA, Jernigan JA, Ostrowsky BE, Richet HM, Jarvis WR, Boyce JM, Farr BM. SHEA guideline for preventing nosocomial transmission of multidrug-resistant strains of staphylococcus aureus and enterococcus. Infect Control Hosp Epidemiol. 2003;24(5):362-86. Review.

18. Puzniak LA, Leet T, Mayfield J, Kollef M, Mundy LM. To gown or not to gown: The effect on acquisition of vancomycin-resistant enterococci. Clin Infect Dis. 2002;35:18-25.

19. Srinivasan A, Song X, Ross T, Merz W, Brower R, Perl TM. A prospective study to determine whether cover gowns in addition to gloves decrease nosocomial transmission of vancomycin-resistant enterococci in an intensive care unit. Infect Control Hosp Epidemiol. 2002;23:424-428.

20. Mosby's Drug Reference for Health Professionals. 2nd ed. Canada: Elsevier; 2010.

21. Tomasz A. The mechanism of the irreversible antimicrobial effects of penicillins: How the beta-lactam antibiotics kill and lyse bacteria. Annu Rev Microbiol. 1979;33:113- 137.

22. Ambler RP. The structure of beta-lactamases. Philos Trans R Soc Lond B Biol Sci. 1980;289:321-331.

23. Bush K, Jacoby GA, Medeiros AA. A functional classification scheme for beta- lactamases and its correlation with molecular structure. Antimicrob Agents Chemother. 1995;39:1211-1233.

24. Knothe H, Shah P, Krcmery V, Antal M, Mitsuhashi S. Transferable resistance to cefotaxime, cefoxitin, cefamandole and cefuroxime in clinical isolates of klebsiella pneumoniae and serratia marcescens. Infection. 1983;11:315-317.

106

25. Bonnet R. Growing group of extended-spectrum beta-lactamases: The CTX-M enzymes. Antimicrob Agents Chemother. 2004;48:1-14.

26. Jacoby GA, Medeiros AA, O'Brien TF, Pinto ME, Jiang H. Broad-spectrum, transmissible beta-lactamases. N Engl J Med. 1988;319:723-724.

27. Urban C, Meyer KS, Mariano N, et al. Identification of TEM-26 beta-lactamase responsible for a major outbreak of ceftazidime-resistant klebsiella pneumoniae. Antimicrob Agents Chemother. 1994;38:392-395.

28. Meyer KS, Urban C, Eagan JA, Berger BJ, Rahal JJ. Nosocomial outbreak of klebsiella infection resistant to late-generation cephalosporins. Ann Intern Med. 1993;119:353-358.

29. Rice LB, Marshall SH, Carias LL, Sutton L, Jacoby GA. Sequences of MGH-1, YOU-1, and YOU-2 extended-spectrum beta-lactamase genes. Antimicrob Agents Chemother. 1993;37:2760-2761.

30. Naumovski L, Quinn JP, Miyashiro D, et al. Outbreak of ceftazidime resistance due to a novel extended-spectrum beta-lactamase in isolates from cancer patients. Antimicrob Agents Chemother. 1992;36:1991-1996.

31. Rice LB, Willey SH, Papanicolaou GA, et al. Outbreak of ceftazidime resistance caused by extended-spectrum beta-lactamases at a massachusetts chronic-care facility. Antimicrob Agents Chemother. 1990;34:2193-2199.

32. Weber DA, Sanders CC, Bakken JS, Quinn JP. A novel chromosomal TEM derivative and alterations in outer membrane proteins together mediate selective ceftazidime resistance in escherichia coli. J Infect Dis. 1990;162:460-465.

33. Quinn JP, Miyashiro D, Sahm D, Flamm R, Bush K. Novel plasmid-mediated beta- lactamase (TEM-10) conferring selective resistance to ceftazidime and aztreonam in clinical isolates of klebsiella pneumoniae. Antimicrob Agents Chemother. 1989;33:1451- 1456.

34. Yang Y, Bhachech N, Bradford PA, Jett BD, Sahm DF, Bush K. Ceftazidime- resistant klebsiella pneumoniae and escherichia coli isolates producing TEM-10 and TEM-43 beta-lactamases from st. louis, missouri. Antimicrob Agents Chemother. 1998;42:1671-1676.

35. Bradford PA, Jacobus NV, Bhachech N, Bush K. TEM-28 from an escherichia coli clinical isolate is a member of the his-164 family of TEM-1 extended-spectrum beta- lactamases. Antimicrob Agents Chemother. 1996;40:260-262.

36. Bush K. Extended-spectrum beta-lactamases in north america, 1987-2006. Clin Microbiol Infect. 2008;14 Suppl 1:134-143.

107

37. Rasheed JK, Jay C, Metchock B, et al. Evolution of extended-spectrum beta-lactam resistance (SHV-8) in a strain of escherichia coli during multiple episodes of bacteremia. Antimicrob Agents Chemother. 1997;41:647-653.

38. Jacoby GA. Beta-lactamase nomenclature. Antimicrob Agents Chemother. 2006;50:1123-1129.

39. Bradford PA, Urban C, Jaiswal A, et al. SHV-7, a novel cefotaxime-hydrolyzing beta-lactamase, identified in escherichia coli isolates from hospitalized nursing home patients. Antimicrob Agents Chemother. 1995;39:899-905.

40. Billot-Klein D, Gutmann L, Collatz E. Nucleotide sequence of the SHV-5 beta- lactamase gene of a klebsiella pneumoniae plasmid. Antimicrob Agents Chemother. 1990;34:2439-2441.

41. Moland ES, Black JA, Hossain A, Hanson ND, Thomson KS, Pottumarthy S. Discovery of CTX-M-like extended-spectrum beta-lactamases in escherichia coli isolates from five US states. Antimicrob Agents Chemother. 2003;47:2382-2383.

42. Sjolund M, Yam J, Schwenk J, et al. Human salmonella infection yielding CTX-M beta-lactamase, united states. Emerg Infect Dis. 2008;14:1957-1959.

43. Kim S, Hu J, Gautom R, Kim J, Lee B, Boyle DS. CTX-M extended-spectrum beta- lactamases, washington state. Emerg Infect Dis. 2007;13:513-514.

44. Folster JP, Pecic G, Krueger A, et al. Identification and characterization of CTX-M- producing shigella isolates in the united states. Antimicrob Agents Chemother. 2010;54:2269-2270.

45. Eisen D, Russell EG, Tymms M, Roper EJ, Grayson ML, Turnidge J. Random amplified polymorphic DNA and plasmid analyses used in investigation of an outbreak of multiresistant klebsiella pneumoniae. J Clin Microbiol. 1995;33:713-717.

46. Brun-Buisson C, Legrand P, Rauss A, et al. Intestinal decontamination for control of nosocomial multiresistant gram-negative bacilli. study of an outbreak in an intensive care unit. Ann Intern Med. 1989;110:873-881.

47. Wiener J, Quinn JP, Bradford PA, et al. Multiple antibiotic-resistant klebsiella and escherichia coli in nursing homes. JAMA. 1999;281:517-523.

48. Brigante G, Luzzaro F, Perilli M, et al. Evolution of CTX-M-type beta-lactamases in isolates of escherichia coli infecting hospital and community patients. Int J Antimicrob Agents. 2005;25:157-162.

49. Pitout JD, Hanson ND, Church DL, Laupland KB. Population-based laboratory surveillance for escherichia coli-producing extended-spectrum beta-lactamases:

108

Importance of community isolates with blaCTX-M genes. Clin Infect Dis. 2004;38:1736- 1741.

50. Rodriguez-Bano J, Navarro MD, Romero L, et al. Epidemiology and clinical features of infections caused by extended-spectrum beta-lactamase-producing escherichia coli in nonhospitalized patients. J Clin Microbiol. 2004;42:1089-1094.

51. Colodner R, Rock W, Chazan B, et al. Risk factors for the development of extended- spectrum beta-lactamase-producing bacteria in nonhospitalized patients. Eur J Clin Microbiol Infect Dis. 2004;23:163-167.

52. Mirelis B, Navarro F, Miro E, Mesa RJ, Coll P, Prats G. Community transmission of extended-spectrum beta-lactamase. Emerg Infect Dis. 2003;9:1024-1025.

53. "Hoffman MS". Increasing trends in ceftazidime-resistant klebsiella pneumoniae and E.coli. 2010.

54. Maki DG. Control of colonization and transmission of pathogenic bacteria in the hospital. Ann Intern Med. 1978;89:777-780.

55. Thom KA, Johnson JA, Strauss SM, Furuno JP, Perencevich EN, Harris AD. Increasing prevalence of gastrointestinal colonization with ceftazidime-resistant gram- negative bacteria among intensive care unit patients. Infect Control Hosp Epidemiol. 2007;28:1240-1246.

56. Lautenbach E, Patel JB, Bilker WB, Edelstein PH, Fishman NO. Extended-spectrum beta-lactamase-producing escherichia coli and klebsiella pneumoniae: Risk factors for infection and impact of resistance on outcomes. Clin Infect Dis. 2001;32:1162-1171.

57. Paterson DL, Ko WC, Von Gottberg A, et al. Outcome of cephalosporin treatment for serious infections due to apparently susceptible organisms producing extended-spectrum beta-lactamases: Implications for the clinical microbiology laboratory. J Clin Microbiol. 2001;39:2206-2212.

58. Du B, Long Y, Liu H, et al. Extended-spectrum beta-lactamase-producing escherichia coli and klebsiella pneumoniae bloodstream infection: Risk factors and clinical outcome. Intensive Care Med. 2002;28:1718-1723.

59. Kim YK, Pai H, Lee HJ, et al. Bloodstream infections by extended-spectrum beta- lactamase-producing escherichia coli and klebsiella pneumoniae in children: Epidemiology and clinical outcome. Antimicrob Agents Chemother. 2002;46:1481-1491.

60. Ariffin H, Navaratnam P, Mohamed M, et al. Ceftazidime-resistant klebsiella pneumoniae bloodstream infection in children with febrile neutropenia. Int J Infect Dis. 2000;4:21-25.

109

61. Schiappa DA, Hayden MK, Matushek MG, et al. Ceftazidime-resistant klebsiella pneumoniae and escherichia coli bloodstream infection: A case-control and molecular epidemiologic investigation. J Infect Dis. 1996;174:529-536.

62. Ho PL, Chan WM, Tsang KW, Wong SS, Young K. Bacteremia caused by escherichia coli producing extended-spectrum beta-lactamase: A case-control study of risk factors and outcomes. Scand J Infect Dis. 2002;34:567-573.

63. Paterson DL, Ko WC, Von Gottberg A, et al. International prospective study of klebsiella pneumoniae bacteremia: Implications of extended-spectrum beta-lactamase production in nosocomial infections. Ann Intern Med. 2004;140:26-32.

64. De Champs C, Rouby D, Guelon D, et al. A case-control study of an outbreak of infections caused by klebsiella pneumoniae strains producing CTX-1 (TEM-3) beta- lactamase. J Hosp Infect. 1991;18:5-13.

65. Harris AD, McGregor JC, Johnson JA, et al. Risk factors for colonization with extended-spectrum beta-lactamase-producing bacteria and intensive care unit admission. Emerg Infect Dis. 2007;13:1144-1149.

66. Bisson G, Fishman NO, Patel JB, Edelstein PH, Lautenbach E. Extended-spectrum beta-lactamase-producing escherichia coli and klebsiella species: Risk factors for colonization and impact of antimicrobial formulary interventions on colonization prevalence. Infect Control Hosp Epidemiol. 2002;23:254-260.

67. D'Agata E, Venkataraman L, DeGirolami P, Weigel L, Samore M, Tenover F. The molecular and clinical epidemiology of enterobacteriaceae-producing extended-spectrum beta-lactamase in a tertiary care hospital. J Infect. 1998;36:279-285.

68. Szabo D, Filetoth Z, Szentandrassy J, et al. Molecular epidemiology of a cluster of cases due to klebsiella pneumoniae producing SHV-5 extended-spectrum beta-lactamase in the premature intensive care unit of a hungarian hospital. J Clin Microbiol. 1999;37:4167-4169.

69. Gaillot O, Maruejouls C, Abachin E, et al. Nosocomial outbreak of klebsiella pneumoniae producing SHV-5 extended-spectrum beta-lactamase, originating from a contaminated ultrasonography coupling gel. J Clin Microbiol. 1998;36:1357-1360.

70. Gould IM. Risk factors for acquisition of multiply drug-resistant gram-negative bacteria. Eur J Clin Microbiol Infect Dis. 1994;13 Suppl 1:S30-8.

71. D'Agata EM, Venkataraman L, DeGirolami P, et al. Colonization with broad- spectrum cephalosporin-resistant gram-negative bacilli in intensive care units during a nonoutbreak period: Prevalence, risk factors, and rate of infection. Crit Care Med. 1999;27:1090-1095.

110

72. Tande D, Jallot N, Bougoudogo F, Montagnon T, Gouriou S, Sizun J. Extended- spectrum beta-lactamase-producing enterobacteriaceae in a malian orphanage. Emerg Infect Dis. 2009;15:472-474.

73. Touati A, Brasme L, Benallaoua S, Madoux J, Gharout A, de Champs C. Enterobacter cloacae and klebsiella pneumoniae isolates producing CTX-M-15 recovered from hospital environmental surfaces from algeria. J Hosp Infect. 2008;68:183- 185.

74. Huang SS, Datta R, Platt R. Risk of acquiring antibiotic-resistant bacteria from prior room occupants. Arch Intern Med. 2006;166:1945-1951.

75. Stone SP, Cooper BS, Kibbler CC, et al. The ORION statement: Guidelines for transparent reporting of outbreak reports and intervention studies of nosocomial infection. Lancet Infect Dis. 2007;7:282-288.

76. Royle J, Halasz S, Eagles G, et al. Outbreak of extended spectrum beta lactamase producing klebsiella pneumoniae in a neonatal unit. Arch Dis Child Fetal Neonatal Ed. 1999;80:F64-8.

77. Pena C, Pujol M, Ardanuy C, et al. Epidemiology and successful control of a large outbreak due to klebsiella pneumoniae producing extended-spectrum beta-lactamases. Antimicrob Agents Chemother. 1998;42:53-58.

78. Rice LB, Carias LL, Bonomo RA, Shlaes DM. Molecular genetics of resistance to both ceftazidime and beta-lactam-beta-lactamase inhibitor combinations in klebsiella pneumoniae and in vivo response to beta-lactam therapy. J Infect Dis. 1996;173:151-158.

79. Wright MO, Hebden JN, Harris AD, et al. Aggressive control measures for resistant acinetobacter baumannii and the impact on acquisition of methicillin-resistant staphylococcus aureus and vancomycin-resistant enterococcus in a medical intensive care unit. Infect Control Hosp Epidemiol. 2004;25:167-168.

80. Harris AD, Kotetishvili M, Shurland S, et al. How important is patient-to-patient transmission in extended-spectrum beta-lactamase escherichia coli acquisition. Am J Infect Control. 2007;35:97-101.

81. Harris AD, Perencevich EN, Johnson JK, et al. Patient-to-patient transmission is important in extended-spectrum beta-lactamase-producing klebsiella pneumoniae acquisition. Clin Infect Dis. 2007;45:1347-1350.

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

111

83. Lautenbach E, Santana E, Lee A, et al. Efficient recovery of fluoroquinolone- susceptible and fluoroquinolone-resistant escherichia coli strains from frozen samples. Infect Control Hosp Epidemiol. 2008;29:367-369.

84. Green HP, Johnson JA, Furuno JP, et al. Impact of freezing on the future utility of archived surveillance culture specimens. Infect Control Hosp Epidemiol. 2007;28:886- 888.

85. Sehulster L, Chinn RY, CDC, HICPAC. Guidelines for environmental infection control in health-care facilities. recommendations of CDC and the healthcare infection control practices advisory committee (HICPAC). MMWR Recomm Rep. 2003;52:1-42.

86. Carling PC, Von Beheren S, Kim P, Woods C, Healthcare Environmental Hygiene Study Group. Intensive care unit environmental cleaning: An evaluation in sixteen hospitals using a novel assessment tool. J Hosp Infect. 2008;68:39-44.

87. Carling PC, Parry MF, Von Beheren SM, Healthcare Environmental Hygiene Study Group. Identifying opportunities to enhance environmental cleaning in 23 acute care hospitals. Infect Control Hosp Epidemiol. 2008;29:1-7.

88. Pessoa-Silva CL, Meurer Moreira B, Camara Almeida V, et al. Extended-spectrum beta-lactamase-producing klebsiella pneumoniae in a neonatal intensive care unit: Risk factors for infection and colonization. J Hosp Infect. 2003;53:198-206.

89. Deyo RA, Cherkin DC, Ciol MA. Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases. J Clin Epidemiol. 1992;45:613-619.

90. Bonten MJ, Slaughter S, Ambergen AW, et al. The role of "colonization pressure" in the spread of vancomycin-resistant enterococci: An important infection control variable. Arch Intern Med. 1998;158:1127-1132.

91. McGregor JC, Kim PW, Perencevich EN, et al. Utility of the chronic disease score and charlson comorbidity index as comorbidity measures for use in epidemiologic studies of antibiotic-resistant organisms. Am J Epidemiol. 2005;161:483-493.

92. Furuno JP, Harris AD, Wright MO, et al. Prediction rules to identify patients with methicillin-resistant staphylococcus aureus and vancomycin-resistant enterococci upon hospital admission. Am J Infect Control. 2004;32:436-440.

93. Harris AD, Nemoy L, Johnson JA, et al. Co-carriage rates of vancomycin-resistant enterococcus and extended-spectrum beta-lactamase-producing bacteria among a cohort of intensive care unit patients: Implications for an active surveillance program. Infect Control Hosp Epidemiol. 2004;25:105-108.

94. Schweizer ML, Furuno JP, Harris AD, et al. Clinical utility of infection control documentation of prior methicillin-resistant staphylococcus aureus colonization or

112

infection for optimization of empirical antibiotic therapy. Infect Control Hosp Epidemiol. 2008;29:972-974.

95. Furuno JP, Harris AD, Wright MO, et al. Value of performing active surveillance cultures on intensive care unit discharge for detection of methicillin-resistant staphylococcus aureus. Infect Control Hosp Epidemiol. 2007;28:666-670.

96. McGregor JC, Perencevich EN, Furuno JP, et al. Comorbidity risk-adjustment measures were developed and validated for studies of antibiotic-resistant infections. J Clin Epidemiol. 2006;59:1266-1273.

97. Furuno JP, McGregor JC, Harris AD, et al. Identifying groups at high risk for carriage of antibiotic-resistant bacteria. Arch Intern Med. 2006;166:580-585.

98. Furuno JP, Perencevich EN, Johnson JA, et al. Methicillin-resistant staphylococcus aureus and vancomycin-resistant enterococci co-colonization. Emerg Infect Dis. 2005;11:1539-1544.

99. Clinical and Laboratory Standards Institute. Performance standards for antimicrobial susceptibility testing; nineteenth informational supplement. 2009.

100. Minarini LA, Gales AC, Palazzo IC, Darini AL. Prevalence of community-occurring extended spectrum beta-lactamase-producing enterobacteriaceae in brazil. Curr Microbiol. 2007;54:335-341.

101. Johnson JK, Smith G, Lee MS, et al. The role of patient-to-patient transmission in the acquisition of imipenem-resistant pseudomonas aeruginosa colonization in the intensive care unit. J Infect Dis. 2009;200:900-905.

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

103. Williams RL. A note on robust variance estimation for cluster-correlated data. Biometrics. 2000;56:645-646.

104. Selden R, Lee S, Wang WL, Bennett JV, Eickhoff TC. Nosocomial klebsiella infections: Intestinal colonization as a reservoir. Ann Intern Med. 1971;74:657-664.

105. Lautenbach E, Harris AD, Perencevich EN, Nachamkin I, Tolomeo P, Metlay JP. Test characteristics of perirectal and rectal swab compared to stool sample for detection of fluoroquinolone-resistant escherichia coli in the gastrointestinal tract. Antimicrob Agents Chemother. 2005;49:798-800.

113

106. Moland ES, Hong SG, Thomson KS, Larone DH, Hanson ND. Klebsiella pneumoniae isolate producing at least eight different beta-lactamases, including AmpC and KPC beta-lactamases. Antimicrob Agents Chemother. 2007;51:800-801.

107. Wagner AK, Soumerai SB, Zhang F, Ross-Degnan D. Segmented regression analysis of interrupted time series studies in medication use research. J Clin Pharm Ther. 2002;27:299-309.

108. Shardell M, Harris AD, El-Kamary SS, Furuno JP, Miller RR, Perencevich EN. Statistical analysis and application of quasi experiments to antimicrobial resistance intervention studies. Clin Infect Dis. 2007;45:901-907.

109. Nseir,S, Blazejewski,C. Lubret,R. Wallet,F, Courcol,R.; Durocher,A. Risk of acquiring multidrug-resistant Gram-negative bacilli from prior room occupants in the intensive care unit. Clin.Microbiol.Infect. 2010.

110. Drees M, Snydman DR, Schmid CH, Barefoot L, Hansjosten K, Vue PM, Cronin M, Nasraway SA, Golan Y. Antibiotic exposure and room contamination among patients colonized with vancomycin-resistant enterococci. Infect Control Hosp Epidemiol. 2008 Aug;29(8):709-15.

111. Martínez JA, Ruthazer R, Hansjosten K, Barefoot L, Snydman DR. Role of environmental contamination as a risk factor for acquisition of vancomycin-resistant enterococci in patients treated in a medical intensive care unit. Arch Intern Med. 2003 Sep 8;163(16):1905-12.

112. Dubberke ER, Reske KA, Olsen MA, McMullen KM, Mayfield JL, McDonald LC, Fraser VJ. Evaluation of Clostridium difficile-associated disease pressure as a risk factor for C. difficile-associated disease. Arch Intern Med. 2007 May 28;167(10):1092-7.

113. Dubberke ER, Reske KA, Yan Y, Olsen MA, McDonald LC, Fraser VJ. Clostridium difficile-associated disease in a setting of endemicity: identification of novel risk factors. Clin Infect Dis. 2007 Dec 15;45(12):1543-9.

114. Lucet JC, Paoletti X, Lolom I, Paugam-Burtz C, Trouillet JL, Timsit JF, Deblangy C, Andremont A, Regnier B. Successful long-term program for controlling methicillin- resistant Staphylococcus aureus in intensive care units. Intensive Care Med. 2005 Aug;31(8):1051-7. Epub 2005 Jul 1.

115. Ajao AO, Harris AD, Roghmann MC, Johnson JK, Zhan M, McGregor JC, Furuno JP. Systematic review of measurement and adjustment for colonization pressure in studies of methicillin-resistant Staphylococcus aureus, vancomycin-resistant enterococci, and clostridium difficile acquisition. Infect Control Hosp Epidemiol. 2011 May;32(5):481-9.

114

116. Lund S, Jackson J, Leggett J, Hales L, Dworkin R, Gilbert D. Reality of glove use and handwashing in a community hospital. Am J Infect Control. 1994 Dec;22(6):352-7.

117. Stringer B, Smith JA, Scharf S, Valentine A, Walker MM. A study of the use of gloves in a large teaching hospital. Am J Infect Control. 1991 Oct;19(5):233-6

118. Shay DK, Goldmann DA, Jarvis WR. Reducing the spread of antimicrobial-resistant microorganisms. Control of vancomycin-resistant enterococci. Pediatr Clin North Am. 1995 Jun;42(3):703-16.

119. Ben-Ami R.Schwaber MJ, Navon-Venezia S, Schwartz D, Giladi M, Chmelnitsky I, Leavitt A, Carmeli Y. Influx of extended-spectrum beta-lactamase-producing enterobacteriaceae into the hospital. Clin Infect Dis. 2006 Apr 1;42(7):925-34. Epub 2006 Feb 27.

120. Tonkic,M.; Goic-Barisic,I.; Punda-Polic,V. Prevalence and antimicrobial resistance of extended-spectum beta-lactamases-producing Escherichai coli and Klebsiella pneumonia strains isolated in a university hospital in Split Croatia. Int.Microbiol., 2005, 8, 2, 119-124, Spain.

121. Hirai Y. Survival of bacteria under dry conditions; from a viewpoint of nosocomial infection. J Hosp Infect 1991; 19: 191–200.

122. Shaughnessy MK, Micielli RL, DePestel DD, Arndt J, Strachan CL, Welch KB, Chenoweth CE. Evaluation of hospital room assignment and acquisition of Clostridium difficile infection. Infect Control Hosp Epidemiol. 2011 Mar;32(3):201-6.

123. Slaughter S, Hayden MK, Nathan C, Hu TC, Rice T, Van Voorhis J, et al. A comparison of the effect of universal use of gloves and gowns with that of glove use alone on acquisition of vancomycin-resistant enterococci in a medical intensive care unit. Ann Intern Med 1996;125:448-456.

124. Owens RC Jr, Johnson JR, Stogsdill P, Yarmus L, Lolans K, Quinn J. Community Transmission in the United States of a CTX M-15-Producing Sequence Type ST131 Escherichia coli Strain Resulting in Death. J Clin Microbiol. 2011 Jul 13.

125. Kramer A, Schwebke I, Kampf G. How long do nosocomial pathogens persist on inanimate surfaces? A systematic review. BMC Infect Dis. 2006 Aug 16;6:130.

126. Davey P, Brown E, Fenelon L, Finch R, Gould I, Hartman G, Holmes A, Ramsay C, Taylor E, Wilcox M, Wiffen PJ. Interventions to improve antibiotic prescribing practices for hospital inpatients. Cochrane Database of Systematic Reviews 2005, Issue 4. Art. No.: CD003543. DOI: 10.1002/14651858.CD003543.pub2

127. Huskins WC, Huckabee CM, O'Grady NP, Murray P, Kopetskie H, Zimmer L, Walker ME, Sinkowitz-Cochran RL, Jernigan JA, Samore M, Wallace D, Goldmann DA;

115

STAR*ICU Trial Investigators. Intervention to reduce transmission of resistant bacteria in intensive care. N Engl J Med. 2011 Apr 14;364(15):1407-18.

116