Summary MULTIDRUG RESISTANT GRAM NEGATIVE ORGANISMS EPIDEMIOLOGY AND OUTCOMES IN THE ICU

Risk Factors for Nosocomial ICU acquired Multidrug-resistant Gram-negative Infection 120 Analysis 5a: SIRS patients with no GNB 121 infection/ colonization compared with ICU acquired M

i

DECLARATION

I hereby declare that the thesis is my original work and it has been written by me entirely I have duly acknowledged all sources of information which have been used in the thesis

This thesis has also not been submitted for any degree in any university

previously

Anupama Vasudevan Date: 12-04-2015

ii

ACKNOWLEDGEMENTS

Amidst heavy pressure of multifarious commitments, Prof Paul Anantharajah Tambyah, MD not only suggested this topic to probe in detail but also evinced personal interest in editing, scrutinizing and correcting my thesis and this herculean task cannot be simply acknowledged by mere words I profusely in debt my gratitude for this guidance and great help ever I would like to thank him for providing me many opportunities to hone my epidemiological and statistical skills

I would like to acknowledge and thank Prof Li Jialiang, PhD for rendering his expert advice with the statistical methods and analysis employed in this work I would like to thank Prof Lim Tow Keang and Prof Adeline Seow, thesis advisory committee members for their continuous support in my journey with this project by offering valuable suggestions and advice

The help offered by the data entry personnel, Ms.Jaminah Ali, Ms.Anisa Shah, Mr.Satish and Mr.Jegan are immensely thanked wholeheartedly I would also like to thank the timely help offered by Mr.Teo Ken Wah in proof reading

my final thesis The strength and moral support provided by Dr.Revathi Sridhar , Dr.Indumathi Venkatachalam and Dr Charumathi Sabanayagam are greatly acknowledged I would like to thank Eryanna Binte Yunus for all the administrative help rendered during this period

iii

The amount of perseverance, forbearance and moral support rendered by

my husband, Mr.Prakash Narasimhan cannot be thanked in mere words of praise and acknowledgements Last but not the least, I wish to put my sincere thanks to

my children, Akkshan & Akshara for their patience rendered in all occasions of family outings, gatherings and functions I like to thank my parents, my pillars of confidence and my mother in law for providing a calm and peaceful atmosphere while working on my project

I profusely place my sincere acknowledgements to all those who directly

or indirectly helped me in completing the project successfully

iv

TABLE OF CONTENTS

List of Publications, Awards & Conference presentations xvii

CHAPTERS

1 Introduction and Objectives of the Study 1

2 Literature Review

Drug Resistant Nosocomial Infections 13 Drug Resistant Gram Negative Bacilli Infections 14

Acinetobacter baumannii 14

Pseudomonas aeruginosa 17

Escherichia coli& Klebsiella pneumoniae 19

Multidrug-Resistant Gram Negative Bacilli 23

c1a.Multidrug-Resistant Gram-negative Bacilli 47

Multidrug-resistant Gram-negative

bacilli infection 48

Multidrug-resistant Gram-negative

bacilli bacteremia 50 Multidrug-resistant Gram-negative

bacilli site-specific Infections 53 c1b.Impact of delayed appropriate antibiotic therapy 54

Multidrug-resistant Gram-negative

bacilli infection 54 Multidrug-resistant Gram-negative

v

bacilli bacteremia 55 Multidrug-resistant Gram-negative

bacilli site-specific Infections 56 c2 Impact of initial choice of antibiotics on survival 69

Multidrug-resistant Gram-negative

Multidrug-resistant Gram-negative

e Prevention and Control of spread/incidence of

multidrug-resistant Gram-negative bacilli 82

Detailed methodology (chapter oriented) 98

4 Characteristics of Patients and Prevalence of Nosocomial

Multidrug-resistant Gram-negative Infection 110

Nosocomial susceptible Gram-negative

bacilli infection 114 Nosocomial multidrug-resistant Gram negative

bacilli infection 115

5 Risk Factors for Nosocomial ICU acquired Multidrug-resistant

Gram-negative Infection 120

Analysis 5a: SIRS patients with no GNB 121 infection/ colonization compared with

ICU acquired MDR-GNB infection Analysis 5b: SIRS patients with Nosocomial ICU acquired 125 MDR-GNB and SGNB infection

vi

6 Prediction Tool for Nosocomial ICU acquired Multidrug-resistant

Gram-negative Infection 133

Segregation into Risk categories 136

7a, 7b: Impact of Nosocomial MDR-GNB infection & effect of

inappropriate antibiotic therapy among patients with

GNB infection 148

Analysis 7a-1: MDR-GNB infection when compared with patients with no GNB Infection 148 Analysis 7a-2: MDR-GNB infection when

compared with patients with SGNB Infection 156 Analysis 7b: Impact of early inappropriate

antibiotics on mortality among patients with

7c: Impact of initial choice of antibiotics on survival among 168

SIRS patients

Analysis 7c-1: Independent Risk factors for

Analysis 7c-2: Effect of different groups of

initial antibiotics 172 Analysis 7c-3: Effect of Carbapenem on

acquisition of MDR-GNB 176

Analysis 8a: Cases compared with propensity matched

vii

controls A (uninfected controls) 188 Analysis 8b: Cases compared with propensity matched

controls B (SGNB infected controls) 190

Significance and Limitations of the study 200 Recommendations and Future directions 202

Annex

viii

SUMMARY

Antimicrobial resistance with the potential for global spread is a major public health threat Paralleling the increase of resistance amongst Gram-positive bacteria, resistant Gram-negative bacteria have increased and are of grave concern with a dwindling pipeline of new antibiotics

Intensive care units (ICU) are the epicenter of nosocomial infections and drug resistant organisms cause many of these Drug resistant nosocomial infections due to Gram-negative bacteria(GNB) result in increased morbidity and possibly increased mortality in critically ill patients Furthermore, patients with Gram-negative nosocomial infections are often treated empirically with broad spectrum antibiotics that result in “collateral damage” by increasing the risk of resistance

We conducted a prospective observational study(August’07-December’11) and all patients who had stayed for more than 24 hours at medical and surgical ICUs of National University Hospital were included Multidrug-resistant Gram-negative bacilli(MDR-GNB) was defined in accordance with the European CDC recommendations

During the study period, a total of 2949 patients were enrolled in the study 2364(80.2%) had systemic inflammatory response syndrome(SIRS) on ICU admission Amongst the patients with SIRS, 183(7.7%) patients acquired nosocomial multidrug-resistant Gram-negative bacilli infection in the hospital and

of these, 76(41.5%) were acquired in the ICU

ix

The independent risk factors for nosocomial ICU acquired MDR-GNB infection by logistic regression were identified to be presence of any Gram-negative bacilli, administration of carbapenems in the past 6 months, surgery, dialysis therapy with end stage renal disease and a stay of >5 days in the ICU

The accuracy of the prediction tool (GSDCS) constructed using these risk factors was measured by the area under the curve It was 0.77(95%CI 0.68-0.89) and 0.78(95%CI 0.69-0.89) for nosocomial multidrug-resistant Gram-negative bacilli infection and bacteremia respectively by receiver operating curve analysis when prospectively validated with a different cohort of critically ill patients(Jan-Sep 2012) When externally validated among 47 patients with Gram-negative bacilli infection from two different hospitals, the AUC was 0.7(95%CI: 0.53-0.88) for predicting nosocomial multidrug-resistant Gram-negative bacilli infection

By Cox proportional hazards survival analysis, multidrug-resistant Gram-negative bacilli infection as a time-varying covariate increased the risk of 28-day mortality while only severity of illness increased risk of in-ICU mortality The use

of initial broad spectrum antibiotics showed no survival benefit among SIRS patients but increased the risk of “collateral damage” by increasing incidence of multidrug-resistant Gram-negative bacilli during the patient’s hospital stay Those patients with MDR-GNB infection who were continued on an inappropriate antibiotic even after 72 hours of culture had a higher risk of mortality

A nested propensity matched case control study of costs associated with multidrug-resistant Gram-negative bacilli infection showed that total costs

x

associated with the hospital stay, laboratory and antibiotics increased amongst multidrug-resistant Gram-negative bacilli infected patients when compared with both uninfected controls and SGNB infected control patients

With the help of the prediction tool, the clinicians have the option to prescribe more targeted appropriate antibiotics at the bedside for critically ill patients With a 72 hour window, clinicians now have the opportunity to modify the antibiotics after the culture results are known in order to reduce bacterial selection of future resistance These strategies need to be validated in large cohorts in other settings to help combat the emerging global threat of antibiotic resistance

xi

LIST OF TABLES Chapter 2b:

2b (1): Risk factors for multidrug-resistant Gram-negative bacilli 37 2b (2): Risk Factors for multidrug-resistant

Acinetobacter baumannii 38

2b (3): Risk factors for multidrug-resistant

Pseudomonas aeruginosa 40

2b (4): Risk factors for multidrug-resistant

E.coli & Klebsiella pneumoniae 43

Chapter 2c:

2c1 (1): Mortality and multidrug-resistant

Gram-negative organisms 61 2c1 (2): Mortality and multidrug-resistant

Acinetobacter baumannii 62 2c1 (3): Mortality and multidrug-resistant

Pseudomonas aeruginosa 64

2c1 (4): Mortality and Resistant Escherichia coli &

Klebsiella pneumoniae 65

Chapter 2d:

2d (1): Costs and Resistant Gram Negative Bacteria 81

Chapter 2e:

2e (1): Prediction Tool for Resistance in Gram Negative Bacilli 91

Chapter 4:

4(2):Gram-negative positive isolates 114 4(3):Prevalence of multidrug-resistant Gram-negative bacilli 116 4(4): Nosocomial MDR-GNB and SGNB infection

(whole cohort and SIRS patients) 119

Chapter 5:

5a (1): Nosocomial ICU acquired MDR-GNB Infection and SIRS

patients with no GNB: Patient characteristics

and univariate analysis 123 5a (2): Nosocomial ICU acquired MDR-GNB Infection:

Independent risk factors- logistic regression (Comparison with SIRS

patients with no GNBInfection/Colonization) 124 5b (1): Nosocomial ICU acquired MDR-GNB

and SGNB Infection:

Patient characteristics and univariate analysis 126

xii

5b (2): Nosocomial ICU acquired MDR-GNB Infection:

Independent risk factors- logistic regression (Comparison nosocomial ICU acquired SGNB Infection) 127

Chapter 6:

6(1): Independent risk factors of ICU acquired

6(2): Sensitivity and Specificity values 136

Chapter 7:

7a-1 (1): Patient Characteristics: SIRS patients with MDR-GNB

infection and no GNB Infection/Colonization 150 7a-1 (2): Multivariable Analysis: Independent Risk factors for 152

Mortality: MDR-GNB and no GNB patients Infection/Colonization

7a-1 (3): Univariate Analysis - Effect of nosocomial MDR-GNB: 154

MDR-GNB and no GNB patients 7a-1 (4): Multivariable Analysis - Effect of Nosocomial ICU

acquired MDR-GNB Bacteremia on in-ICU mortality:

7a-1 (5): Multivariable Analysis - Effect of Nosocomial

MDR-GNB Infection & Bacteremia on 28-day mortality: MDR-GNB and noGNB patients 155 7a-2(1): Patient Characteristics: MDR-GNB and SGNB Infected

7a-2(2): Multivariable Analysis: Independent Risk factors for

Mortality: MDR-GNB and SGNB Infected Patients 158 7a-2(3): Univariate Analysis - Effect of nosocomial MDR-GNB:

MDR-GNB and SGNB Infected Patients 159 7b (1): Univariate Analysis- Effect of Inappropriate antibiotics

7b (2): Univariate Analysis- Effect of Inappropriate antibiotics

7b (3): Multivariable analysis: Effect of inappropriate

antibiotics Nosocomial MDR-GNB Infections 162 7b (4): Multivariable analysis: Effect of inappropriate

antibiotics- Nosocomial MDR-GNB Infections 162 7c-1 (1): Patient Characteristics and Univariate analysis 170 7c-1 (2): Independent risk factors for Mortality: Cox proportional

xiii

7c-2(1): Univariate Analysis- Effect of different groups

7c-2(2): Multivariable analysis: Effect of Carbapenems

7c-2(3): Multivariable analysis: Effect of Carbapenems

and Cefazolin @ 24 & 48 H- In-ICU Mortality 175 7c-2(4): Multivariable analysis: Effect of Carbapenems

and Cefazolin @ 24 & 48 H- 28-day Mortality 176 7c-3(1): Risk of MDR-GNB with initial Carbapenem use 176

Chapter 8:

8(1) : Balance check after Propensity Matched scoring 183 8a (1): Patient Characteristics: Cases and Propensity matched

Uninfected Controls (Controls A) 185 8b (1): Patient Characteristics: Cases and Propensity matched

Controls infected with SGNB(Controls B) 187 8a (2): Comparison of costs between the propensity matched

8a(3): Survivors: Comparison of costs between the propensity

8b (2): Comparison of costs between the propensity matched

8b (3): Survivors:Comparison of costs between the

propensity matched cases and controls B 192 8a (3): Comparison of length of stay between the propensity

8b (3): Comparison of costs between the propensity matched

xiv

LIST OF FIGURES Chapter 4:

4(1): Epi Curve: MDR-GNB & GNB positive clinical isolates 113 4(2): Nosocomial SGNB infection in ICU & Hospital: Sites 114 4(3): Nosocomial SGNB Infection in ICU & Hospital: Organisms 115 4(4-1): Nosocomial MDR-GNB infection by years 117 4(4-2): Prevalence of nosocomial MDR-GNB infection 117 4(5): Nosocomial MDR-GNB Infection in ICU & Hospital: Sites 118 4(7): Nosocomial MDR-GNB Infection in ICU & Hospital:

Organisms 118

Chapter 5:

Chapter 6:

6(1): Prevalence of Nosocomial ICU acquired MDR-GNB

Infection & Bacteremia - Risk categories 137 6(2): Area Under the curve- Nosocomial ICU acquired

MDR-GNB Infection: Prospective Validation 138 6(3): Prevalence of Nosocomial ICU acquired

MDR-GNB Infection Prospective Validation- Risk categories 138 6(4): Area Under the curve-Nosocomial ICU acquired

MDR-GNB Bacteremia: Prospective Validation 139 6(5): Prevalence of Nosocomial ICU acquired

MDR-GNB Bacteremia Prospective Validation-

Risk categories 139 6(6): Area Under the curve- Nosocomial ICU acquired

MDR-GNB Infection: External Validation 141 6(7): Prevalence of Nosocomial ICU acquired

MDR-GNB Infection External Validation- Risk categories 141 6(8): Area Under the curve- Nosocomial ICU acquired

MDR-GNB Infection(ICU patients): External Validation 142 6(7): Prevalence of Nosocomial ICU acquired

MDR-GNB Infection(ICU patients: External Validation-

Risk categories 142

Chapter 7:

7a-1(1): Kaplan-Meier survival curves: SIRS patients with 153

MDR-GNB infection and no GNB patients 7a-2(1): Kaplan-Meier survival curves: MDR-GNB and