Epidemiology of Clostridium difficile infection in the Asia- Pacific … · Epidemiology of...

Epidemiology of Clostridium difficile infection in the Asia-

Pacific region

Deirdre Collins, BA (Mod), MInfDis

This thesis is presented for the degree of Doctor of Philosophy

The University of Western Australia

2016

Microbiology and Immunology

School of Pathology and Laboratory Medicine

i

TABLE OF CONTENTS

CHAPTER 1. LITERATURE REVIEW .......................................................................... 1

1.1. INTRODUCTION ..................................................................................................... 2

1.1.1. Pathogenesis of CDI ................................................................................................ 2

1.1.2. Toxins of C. difficile ............................................................................................... 2

1.1.3. Epidemiology .......................................................................................................... 3

1.1.3.1. Community-associated CDI ................................................................................. 5

1.1.3.2. Animal and food sources of C. difficile ............................................................... 5

1.1.3.3. Infection control for C. difficile infection ............................................................ 6

1.2. METHODS FOR DETECTION AND TYPING OF C. DIFFICILE ........................ 6

1.2.1. Laboratory diagnosis for CDI ................................................................................. 6

1.2.2. Molecular assays for C. difficile detection .............................................................. 7

1.2.2.1. Issues with NAATs .............................................................................................. 8

1.2.3. Impact of diagnostic method on patient outcome ................................................. 10

1.3. C. DIFFICILE TYPING METHODS ...................................................................... 11

1.3.1. PCR ribotyping ..................................................................................................... 11

1.3.2. PFGE ..................................................................................................................... 11

1.3.3. MLST and other sequence-based typing methods ................................................ 12

1.4. WORLDWIDE EPIDEMIOLOGY OF C. DIFFICILE INFECTION .................... 12

1.5. EPIDEMIOLOGY OF C. DIFFICILE INFECTION IN ASIA-PACIFIC

COUNTRIES .................................................................................................................. 15

1.5.1. East Asia ............................................................................................................... 15

1.5.1.1. Japan ................................................................................................................... 15

1.5.1.2. Korea .................................................................................................................. 17

1.5.1.3. People’s Republic of China ................................................................................ 18

1.5.1.4. Taiwan ................................................................................................................ 20

1.5.1.5. Hong Kong ......................................................................................................... 21

1.5.2. Southeast Asia ....................................................................................................... 22

1.5.2.1. Philippines .......................................................................................................... 22

1.5.2.2. Thailand.............................................................................................................. 22

1.5.2.3. Malaysia ............................................................................................................. 23

1.5.2.4. Indonesia ............................................................................................................ 24

ii

1.5.2.5. Singapore ........................................................................................................... 24

1.5.2.6. Cambodia ........................................................................................................... 25

1.5.2.7. Vietnam .............................................................................................................. 25

1.5.3. South Asia ............................................................................................................. 25

1.5.3.1. India ................................................................................................................... 25

1.5.3.2. Bangladesh ......................................................................................................... 26

1.5.3.3. Pakistan .............................................................................................................. 26

1.5.4. Australia ................................................................................................................ 26

1.6. EPIDEMIOLOGY ................................................................................................... 28

1.6.1. Prevalent C. difficile RTs in Asia .......................................................................... 28

1.6.2. Infection prevention and control ........................................................................... 29

1.7. CONCLUSIONS ...................................................................................................... 31

1.8. RESEARCH AIMS .................................................................................................. 32

CHAPTER 2. MATERIALS AND METHODS ............................................................. 33

2.1. MATERIALS ........................................................................................................... 34

2.1.1. Media .................................................................................................................... 34

2.1.2. Buffers and solutions ............................................................................................ 34

2.1.3. PCR reagents ......................................................................................................... 34

2.1.4. Kits ........................................................................................................................ 34

2.2. LABORATORY METHODS .................................................................................. 35

2.2.1. Culture ................................................................................................................... 35

2.2.1.1. Pre-reduction of agar plates ............................................................................... 35

2.2.1.2. Direct culture ...................................................................................................... 35

2.2.1.3. Indirect culture ................................................................................................... 35

2.2.1.4. Toxigenic culture ............................................................................................... 35

2.2.2. Phenotypic testing of C. difficile isolates .............................................................. 36

2.2.3. DNA extraction ..................................................................................................... 36

2.2.4. Molecular characterisation of C. difficile isolates ................................................. 36

2.2.4.1. Toxin profiling ................................................................................................... 36

2.2.4.2. PCR ribotyping .................................................................................................. 36

2.3. STUDY SETTINGS ................................................................................................ 37

2.3.1. Australia: Study setting and population ................................................................ 37

2.3.2. Indonesia: study setting and population ................................................................ 38

iii

2.3.3. Malaysia: study setting and population ................................................................. 38

2.4. STUDY METHODS ................................................................................................ 38

2.4.1. Monitoring routine testing for C. difficile in WA ................................................. 38

2.4.1.1. Study background .............................................................................................. 38

2.4.2. A prospective study of C. difficile infection in two tertiary-care hospitals in Perth

......................................................................................................................................... 39

2.4.2.1. Setting and study design .................................................................................... 39

2.4.2.2. Data collection ................................................................................................... 39

2.4.2.3. C. difficile culture and ribotyping ...................................................................... 40

2.4.2.4. Statistical methods ............................................................................................. 40

2.4.3. Laboratory-identified C. difficile infection in Australian healthcare .................... 40

2.4.3.1. Study setting ....................................................................................................... 40

2.4.3.2. Data collection ................................................................................................... 41

2.4.3.3. C. difficile culture and ribotyping ...................................................................... 41

2.4.3.4. Statistical analysis .............................................................................................. 41

2.4.4. C. difficile infection in diarrhoeal emergency department patients in WA .......... 41

2.4.4.1. Study background .............................................................................................. 41

2.4.4.2. Data collection and analysis ............................................................................... 42

2.4.5. Routine identification of C. difficile in clinical samples in WA ........................... 42

2.4.6. Data collection - Indonesia.................................................................................... 43

2.4.7. Data collection - Malaysia .................................................................................... 43

2.4.7.1. In-patient study................................................................................................... 43

2.4.8. Data analysis ......................................................................................................... 43

2.5. ISOLATE COLLECTIONS ..................................................................................... 43

2.5.1. Australian C. difficile isolates ............................................................................... 44

2.5.1.1. 2012 surveillance snapshot study ....................................................................... 44

2.5.1.2. A prospective study of C. difficile infection in two tertiary-care hospitals in

Perth ................................................................................................................................ 44

2.5.1.3. Monitoring routine testing for C. difficile in WA .............................................. 44

2.5.1.4. C. difficile infection in diarrhoeal emergency department patients in WA........ 44

2.5.1.5. Asia-Pacific C. difficile isolates ......................................................................... 44

2.5.2. Other Asian isolate collections.............................................................................. 45

2.5.2.1. Singaporean isolates ........................................................................................... 45

2.5.2.2. Japanese isolates................................................................................................. 46

iv

2.5.2.3. Indonesian isolates ............................................................................................. 46

2.5.2.4. Malaysian isolates .............................................................................................. 46

2.6. DEFINITIONS ......................................................................................................... 46

2.7. ETHICAL APPROVALS ........................................................................................ 47

CHAPTER 3. DESCRIPTIVE EPIDEMIOLOGY OF C. DIFFICILE INFECTION IN

AUSTRALIA .................................................................................................................. 48

3.1. INTRODUCTION ................................................................................................... 49

3.1.1. Aims of this chapter .............................................................................................. 49

3.2. RESULTS ................................................................................................................ 50

3.2.1. Monitoring routine testing for C. difficile in WA ................................................. 50

3.2.1.1. Prevalence of C. difficile .................................................................................... 50

3.2.1.2. Sources of testing requests ................................................................................. 50

3.2.1.3. Demographic characteristics of C. difficile-positive cases ................................ 50

3.2.2. A prospective study of C. difficile infection in two tertiary-care hospitals in Perth

......................................................................................................................................... 51

3.2.2.1. Descriptive characteristics of CDI cases ............................................................ 51

3.2.2.2. Comparison of clinical signs for severe vs non-severe CDI cases..................... 52

3.2.2.3. Comparison of characteristics for CO-CDI and HO-CDI cases ........................ 52

3.2.3. Laboratory-identified C. difficile infection in Australian healthcare .................... 52

3.2.3.1. Demographic and clinical characteristics of CDI cases ..................................... 52

3.2.4. C. difficile infection in emergency department patients in WA ............................ 53

3.3. DISCUSSION .......................................................................................................... 56

3.3.1. C. difficile testing practices in Australia ............................................................... 56

3.3.2. Prevalence of C. difficile among Australian diarrhoeal patients ........................... 58

3.3.3. Increasing incidence rates of CDI in Australia ..................................................... 59

3.3.4. Characteristics of CDI patients in Australia ......................................................... 61

3.3.4.1. Advanced age ..................................................................................................... 61

3.3.4.2. Clinical signs of severe illness ........................................................................... 61

3.3.4.3. Recurrent CDI .................................................................................................... 62

3.3.4.4. Antimicrobial use ............................................................................................... 63

3.3.4.5. PPI use ................................................................................................................ 64

3.3.4.6. Hospitalisation and residence in aged care facilities.......................................... 64

3.3.4.7. Comorbidities ..................................................................................................... 65

3.3.4.8. CDI among females............................................................................................ 65

v

3.3.5. CA-CDI in Australia ............................................................................................. 66

3.3.5.1. Definition of CA-CDI ........................................................................................ 66

3.3.5.2. CO-CDI in Australia .......................................................................................... 67

3.3.5.3. Possible community sources of C. difficile in Australia .................................... 68

3.3.5.4. Strengths of the studies outlined here ................................................................ 70

3.3.5.5. Limitations of studies outlined here ................................................................... 70

3.3.5.6. Implications and recommendations.................................................................... 71

3.3.5.7. Future studies ..................................................................................................... 72

3.3.5.8. Chapter summary ............................................................................................... 72

CHAPTER 4. PREVALENCE OF C. DIFFICILE INFECTION IN INDONESIA AND

MALAYSIA .................................................................................................................... 74

4.1. INTRODUCTION ................................................................................................... 75


4.2. RESULTS ................................................................................................................ 76

4.2.1. Indonesia ............................................................................................................... 76

4.2.2. Malaysia ................................................................................................................ 77

4.2.3. Descriptive epidemiology in Indonesia ................................................................. 77

4.2.4. Descriptive epidemiology in Malaysia ................................................................. 77

4.2.4.1. Comparison of C. difficile-positive patients with C. difficile-negative patients 77

4.2.4.2. Comparison of CDI cases with non-CDI cases .................................................. 78

4.2.4.3. Comparison of C. difficile cases with toxigenic vs non-toxigenic strains ......... 79

4.3. DISCUSSION .......................................................................................................... 79

4.3.1. C. difficile testing in Malaysia and Indonesia ....................................................... 79

4.3.2. Prevalence of C. difficile in Malaysia and Indonesia ............................................ 80

4.3.3. Characteristics of CDI patients in Malaysia ......................................................... 81

4.3.3.1. Age ..................................................................................................................... 81

4.3.3.2. Recent hospitalisation ........................................................................................ 81

4.3.3.3. Extended length of stay ...................................................................................... 81

4.3.3.4. Recent nasogastric tube feeding ......................................................................... 82

4.3.3.5. Recent antimicrobial use .................................................................................... 82

4.3.3.6. PPI use ................................................................................................................ 83

4.3.3.7. Comorbidities ..................................................................................................... 83

4.3.3.8. Markers of severe CDI ....................................................................................... 83

4.3.3.9. Sex ...................................................................................................................... 83

vi

4.3.4. High prevalence of non-toxigenic C. difficile in Malaysia and Indonesia ............ 84

CHAPTER 5. MOLECULAR EPIDEMIOLOGY OF C. DIFFICILE IN ASIA

PACIFIC COUNTRIES .................................................................................................. 88

5.1. INTRODUCTION ................................................................................................... 89


5.2. RESULTS ................................................................................................................ 90

An overview of all isolates included in analyses is given in Table 5.1. ......................... 90

5.2.1. Australian C. difficile isolates ............................................................................... 90

5.2.1.1. 2012 surveillance studies ................................................................................... 90


Perth ................................................................................................................................ 91

5.2.1.3. Routine testing study .......................................................................................... 91

5.2.1.4. Emergency department study ............................................................................. 91

5.2.2. Asia-Pacific isolates .............................................................................................. 91

5.2.3. Other Asian isolate collections.............................................................................. 92

5.2.3.1. Singapore isolates .............................................................................................. 92

5.2.3.2. Japan isolates ...................................................................................................... 92

5.2.3.3. Indonesian isolates ............................................................................................. 92

5.2.3.4. Malaysian isolates .............................................................................................. 92

5.3. DISCUSSION .......................................................................................................... 92

5.3.1. C. difficile molecular epidemiology in Australia .................................................. 92

5.3.1.1. Prevalent RTs circulating in Australia ............................................................... 92

5.3.1.2. CDT-positive strains in Australia....................................................................... 94

5.3.1.3. Diversity in Australian RTs ............................................................................... 97

5.3.2. Molecular epidemiology of C. difficile in Asian countries ................................... 97

5.3.2.1. Prevalent RTs circulating in Asia ...................................................................... 97

5.3.2.2. CDT-positive strains in Asia ............................................................................ 101

5.3.3. Comparison of Australian and Asian molecular epidemiology .......................... 101

5.3.4. Strengths of the studies described here ............................................................... 102

5.3.5. Limitations of studies described here .................................................................. 102

5.3.6. Implications and recommendations..................................................................... 103

5.3.7. Future studies ...................................................................................................... 103

5.3.8. Chapter summary ................................................................................................ 104

CHAPTER 6. GENERAL DISCUSSION .................................................................... 105

vii

6.1. SUMMARY OF FINDINGS ................................................................................. 106

6.2. THE DESCRIPTIVE EPIDEMIOLOGY OF CDI IN ASIA-PACIFIC

COUNTRIES ................................................................................................................ 107

6.3. CA-CDI .................................................................................................................. 108

6.4. THE MOLECULAR EPIDEMIOLOGY OF C. DIFFICILE IN THE ASIA-

PACIFIC REGION ....................................................................................................... 109

6.5. STRENGTH AND ORIGINALITY OF STUDIES ............................................... 111

6.6. LIMITATIONS OF STUDIES .............................................................................. 112

6.7. IMPLICATIONS AND RECOMMENDATIONS FOR POLICY ........................ 112

6.8. DIRECTIONS FOR FUTURE RESEARCH ......................................................... 114

6.9. CONCLUSIONS .................................................................................................... 115

viii

SUMMARY

Clostridium difficile is an anaerobic Gram-positive spore-forming bacillus which causes

healthcare-related infection in developed countries. Highly resistant spores allow it to

persist in healthcare facilities, causing diarrhoea primarily in patients who have recently

been treated with antimicrobials that compromise the gut microbiota. The main risk

factors for C. difficile infection (CDI) are advanced age, hospitalisation and

antimicrobial use. CDI has been studied in detail in North America and Europe, where it

has caused significant outbreaks since the early 2000s. Recently, CDI has been

increasingly associated with onset in the community (CO-CDI) and association with

community-based reservoirs (CA-CDI) as opposed to exposure to healthcare facility

reservoirs including infected or colonised patients and contaminated surfaces.

Until recently, few reports existed about CDI in Australia despite evidence that

infection rates were increasing. Of great concern, particularly for Australian health

authorities, was that the epidemiology of CDI in Asia was largely unknown mainly due

to a lack of local awareness and a lack of testing. The overarching aim of this project

was to investigate the epidemiology/molecular epidemiology of CDI in the Asia-Pacific

region, focusing initially on Australia and its nearest neighbours, and then broadening

the approach to include other countries in Asia. The main objectives were:

• To examine aspects of the epidemiology of CDI in Australia.

• To describe the prevalence and epidemiology of CDI in Indonesia and Malaysia,

two of Australia’s closest neighbours.

• To provide an overview of CDI in other countries in the Asia-Pacific region.

• To describe the molecular epidemiology of C. difficile isolates from Asia-Pacific

countries.

Various studies were performed to achieve these objectives. Nationwide surveillance of

CDI was performed in Australia by collecting C. difficile isolates from cases between

October and November, 2012. A cross-sectional study of CDI patients in two major

hospitals in Western Australia (WA) was performed to investigate aspects of their

infection and treatment. The epidemiology of CA-CDI and CO-CDI were investigated

by examining the prevalence of CDI across all diarrhoeal patients whose stool

specimens were submitted to the primary enteric diagnostic laboratory in WA, and by

ix

performing a retrospective study of patients attending emergency departments with

diarrhoeal disease in six hospitals in WA. Testing for CDI was investigated in Malaysia

and Indonesia by performing toxin-specific enzyme immunoassays (EIAs) locally, with

confirmation by toxigenic culture of stool specimens performed in Perth, WA.

Molecular typing was performed on a collection of almost 1,500 C. difficile isolates

collected from Australia and other Asian countries.

The prevalence of CDI in WA was calculated as 6.4%, where 50.8% of cases were CO-

CDI, and 24.6% overall were undiagnosed by routine testing. The prevalence of CA-

CDI among Australian patients varied from 7.5% among WA hospital inpatients to 44%

among CDI cases presenting to hospital emergency departments. CA-CDI patients in

Australia were significantly younger than HCFA-CDI patients (57.0 years vs 70.0 years

nationwide, p<0.05). A risk factor for CA-CDI among emergency department patients

was previous antibiotic use (odds ratio [OR] 7.31, p<0.001).

CDI prevalence among diarrhoeal patients in Malaysia (15.4%) and Indonesia (11.7%)

was considerably higher than in WA. CDI patients in Indonesia and Malaysia appeared

to be younger (median age 50.0 years in Indonesia, 55.0 years in Malaysia) than

Australian CDI patients (median ages 60-70 years). Malaysian CDI patients exhibited

many characteristics typical of CDI patients worldwide including extended

hospitalisations (median 12.5 days in Malaysia, 15.0 days in WA), recent antibiotic use

(100.0%), particularly amoxycillin/amoxycillin clavulanate and cephalosporins, and

multiple comorbidities including haematological malignancies (20.0%), diabetes

mellitus (40.0%), chronic kidney disease (20.0%) and heart disease (20.0%).

The most common C. difficile strains in Asian countries were ribotype (RT) 017

followed by RTs 014/020, 018, 002 and 012, with wide variation between countries,

while the most common strains in Australia were RT 014/020 followed by RTs 002,

056, 054 and 070. Binary toxin-positive strains such as RTs 027 and 078 were rarely

observed in Asian and Australian isolate collections, despite being common in North

America and Europe. Non-toxigenic C. difficile strains were frequently found among

Malaysian and Indonesian isolates, and no significant differences were identified for

Malaysian patients infected with toxigenic strains compared with non-toxigenic strains.

The studies described here have expanded on our knowledge of CDI in Asia-Pacific

countries, and have provided valuable baseline data for comparison with other regions

x

of the world. These provide the first data on the prevalence of CDI in Indonesia, and the

molecular epidemiology of C. difficile in Malaysia and Indonesia. CO-CDI was

identified as a significant issue in Australia, warranting greater surveillance in the

community and further investigations of community-based sources of infection. The

molecular ecology of C. difficile in Asia differs from other regions of the world, and

highlights the need for continuous monitoring of the strains of C. difficile transmitting

within and between countries and continents. The widespread unregulated use of

antimicrobials in Asia, in humans and in agriculture, highlights the need for a “One

Health” approach to reduce the incidence of CDI in Asian countries in the future. The

findings presented here can serve to generate local recommendations for CDI

management and further studies.

xi

DECLARATION

All work presented in this thesis was performed by me, except where otherwise

indicated. All instances where work was not performed by me have been acknowledged

within the text.

_______________________________________

Deirdre Collins

xii

ACKNOWLEDGEMENTS

This PhD has been a long and convoluted process, and I would like to acknowledge and

thank everybody who has helped me along the way.

First and foremost, I would like to thank my principal supervisor, Professor Thomas

Riley. He has been a bountiful source of advice, entertainment, lifts home, many

lunches and hot chocolates, and a great support throughout this PhD. Thank you Tom

for all your encouragement and direction over the past four years.

I am particularly thankful towards our study co-investigators in Perth, namely Professor

Antonio Celenza and Associate Professor Linda Selvey. Professor Celenza was

extremely helpful in setting up the Emergency Department study, and Professor Selvey

has been a continuous source of support and advice on all aspects of the study.

I am very grateful to my study co-investigators overseas. Professor Hamimah Hassan,

Associate Professor Rina Karunakaran, Dr Mohamad Nazri Aziz, Dr Fairuz Amran,

Associate Professor Dr Ariza Adnan, Dr RamLiza RamLi, Dr Siti Asma Hassan and Dr

Yeong Yeh Lee contributed to the Malaysian prevalence studies. In Indonesia, our co-

investigators were Dr Mohammed Hussein Gasem, Dr Bambang Isbandrio, Dr I.G.

Arinton, Dr Pujo Hendriyanto and Dr I.G.A. Putra Hartana. Mr. Kentaro Ouchi

coordinated the multi-country study of CDI in the Asia-Pacific region. All of these

people were a pleasure to work with, and were very welcoming and accommodating

during study visits, and helpful throughout all aspects of the studies.

The following people performed experimental work or provided data for analysis in this

thesis. All such instances have been acknowledged in the text. I am very grateful for

their involvement and cooperation. They have all contributed to fun in the office and

laboratory, and I’ve truly enjoyed their company and friendship over the years. In

alphabetical order they are: Ms Grace Androga, Ms Suzanne Ditchburn, Ms Christine

Duncan, Dr Briony Elliott, Ms Michelle Fisher, Dr Niki Foster, Ms Noellene Foster, Ms

Stacey Hong, Ms Tanya Scheller, Mr Daniel Knight, Ms Su Chen Lim, Mr Alan

McGovern, Ms Papanin Putsathit, Ms Lauren Tracey and Ms Welma van Schalkwyk,

and the staff of the PathWest enteric diagnostic laboratory.

xiii

Last but by no means least, a special thank you to R and MC, for always making me

smile through this journey.

Some of the work presented in this thesis was funded by the following sources:

Alere Inc., the Australian Commission on Safety and Quality in Healthcare, the Department of Health Western Australia, Otsuka Pharmaceutical Co. Ltd. and Sanofi S.A.

I am grateful for an International Postgraduate Research Scholarship and an Australian Postgraduate Award provided by The University of Western Australia.

xiv

PUBLICATIONS AND PRESENTATIONS

In all the papers listed here, authorship contribution was determined according to the

National Health and Medical Research Council/Australian Vice-Chancellors’

Committee guidelines, i.e. I made a significant contribution to: a) conception and

design, or analysis and interpretation of data; b) drafting the article or revising it

critically for important intellectual content and c) final approval of the version to be

published. Where I was responsible for writing the manuscripts editorial assistance was

provided by co-authors.

Journal articles

1. Collins DA, Hawkey PM, Riley TV. Epidemiology of Clostridium difficile infection in Asia. Antimicrobial Resistance and Infection Control 2013; 2: 21.

Wrote the paper, revised it critically (90%).

2. Tan XQ, Verrall AJ, Jureen R, Riley TV, Collins DA, Lin RT, Balm MN, Chan

D, Tambyah PA. The emergence of community-onset Clostridium difficile infection in a tertiary hospital in Singapore: a cause for concern. International

Journal of Antimicrobial Agents 2014; 43: 47-51. Performed C. difficile culture and molecular typing (10%).

3. Foster NF, Collins DA, Ditchburn SL, Duncan CN, van Schalkwyk JW, Golledge CL, Keed ABR, Riley TV. Epidemiology of Clostridium difficile infection in two tertiary-care hospitals in Perth, Western Australia: a cross-sectional study. New Microbes and New Infections 2014; 2: 64-71. Performed patient data analysis and contributed to Methods and Results

sections of manuscript (25%).

4. Collins DA, Elliott B, Riley TV. Molecular methods for detecting and typing of

Clostridium difficile. Pathology 2015; 47: 211-8. Wrote all sections of the paper except discussion of molecular typing methods,

revised it critically (60%).

5. Mori N, Yoshizawa S, Saga T, Ishii Y, Murakami H, Iwata M, Collins DA, Riley TV, Tateda K. Incorrect diagnosis of Clostridium difficile infection in a university hospital in Japan. Journal of Infection and Chemotherapy 2015: 21: 718-22. Performed C. difficile culture and molecular typing (10%).

xv

6. Collins DA, Riley TV. Routine detection of Clostridium difficile in Western Australia. Anaerobe 2016: 37: 34-37. Designed experiments, performed data collection, culture and analysis, wrote

the paper, revised it critically (90%).

7. Cheng AC, Collins DA, Elliott B, Ferguson JK, Paterson DL, Thean S, Riley

TV. Laboratory-based surveillance of Clostridium difficile circulating in Australia, September - November 2010. Pathology 2016: 48: 257-260.

Wrote the paper, revised it critically (40%).

8. Collins DA, Putsathit P, Elliott B, Riley TV. Laboratory-based surveillance of

Clostridium difficile in Australia in 2012. Submitted. Performed all patient data analysis, wrote the paper, revised it critically (60%).

9. Collins DA, Selvey L, Celenza A, Riley TV. Clostridium difficile infection in

diarrhoeal emergency department patients in Western Australia. Manuscript in preparation.

Designed experiments, performed data collection and analysis, wrote the paper,

revised it critically (80%).

10. Collins DA, Riley TV, Sohn KM, Wu Y, Ouchi K, Ishii Y, Tateda K. Clostridium difficile infection in the Asia-Pacific region. Manuscript in preparation.

Performed local data collection, performed culture and molecular typing,

analysed all data, wrote the paper, revised it critically (50%).

xvi

Presentations

1. 8th International Conference on the Molecular Biology and Pathogenesis of the Clostridia (ClostPath 8), October 2013, Palm Cove, Queensland, Australia. A snapshot of Clostridium difficile strains circulating in Australia, October-November 2012. Collins DA, Putsathit P, Riley TV. Poster presentation.

2. C9-Go8 HDR Student Forum, “Sustainable Health Futures: Graduate Perspectives from China and Australia”, October 2013, Shanghai Jiao Tong University, Shanghai, China. Clostridium difficile: a major international pathogen but under-recognised in Asia. Collins DA. Oral presentation.

3. “Science Lands in Parliament”, Western Australian Parliament House, October 2013, Perth, Australia. Clostridium difficile: The most common cause of infectious diarrhoea but where is it coming from and how can we stop it? Foster NF, Knight DR, Slimings C, Collins DA, Squire M, Elliott B, Carson K, Riley TV. Poster presentation.

4. 7th International Congress of the Asia Pacific Society of Infection Control, March 2015, Taipei, Taiwan. Hypervirulent Clostridium difficile in Asia. Collins DA, Riley TV. Invited oral presentation. Abstract in Journal of Microbiology, Immunology and Infection 2015: 48, S14.

5. 5th International Clostridium difficile Symposium, May 2015, Bled, Slovenia. Monitoring and examination of Clostridium difficile testing requests in Western Australia. Collins DA, Riley TV. Poster presentation.

6. 2nd Asia-Pacific Forum on Advanced Laboratory Techniques: Molecular Epidemiology of Clostridium difficile Infections. October 2015, Hangzhou, China. Clostridium difficile infection in Singapore. Collins DA, Riley TV. Invited oral presentation.

7. Australian Society for Microbiology WA Branch “Around the Labs”: Clostridium difficile in Asia. December 2015, Perth, Western Australia. Collins DA, Riley TV. Oral presentation.

8. Faculty of Medicine and Health Sciences, Universiti Putra Malaysia and Department of Medicine, Hospital Universiti Sains Malaysia, June 2016. Clostridium difficile infection in Malaysia. Collins DA, Lee YY, Hassan SA, Hassan H, Karunakaran R, Aziz MN, Amran F, Adnan A, RamLi R, Riley TV. Oral presentation.

9. Australian Society for Microbiology Annual Scientific Meeting and Exhibition, Perth, Australia, July 2016. Clostridium difficile infection in diarrhoeal emergency department patients in Western Australia. Collins DA, Selvey L, Celenza A, Riley TV. Poster presentation.

xvii

10. 13th Biennial Congress of the Anaerobe Society of the Americas, July 2016,

Nashville, USA. Clostridium difficile infection in South-East Asia. Collins DA, Putsathit P, Gasem MH, Habibie T, Arinton IG, Handrianto P, Agung IG, Lee YY, Hassan SA, Nadiah HZ, Kiratisin P, Elliott B, Riley TV. Oral presentation.

xviii

LIST OF ABBREVIATIONS

°C degrees Celsius

µg micrograms

µL microlitres

AAD antibiotic-associated diarrhoea

ACSQH Australian Commission on Safety and Quality in Healthcare

ACT Australian Capital Territory

ADP adenosine diphosphate

Ag antigen

AH Armadale Hospital

AIDS acquired immunodeficiency syndrome

ASID Australian Society for Infectious Diseases

BA blood agar

BD Becton Dickinson

bp base-pairs

BRH Bunbury Regional Hospital

CA community-associated

CCCA cell culture cytotoxicity assay

CCFA cycloserine-cefoxitin fructose agar

CDC Centers for Disease Control

CDI Clostridium difficile infection

CDRN Clostridium difficile Ribotyping Network

CDT binary toxin

CE sequence-based capillary gel electrophoresis

CI confidence interval

CMB cooked meat broth

CO community onset

COPD chronic obstructive pulmonary disease

xix

dL decilitres

DNA deoxyribonucleic acid

dNTPs deoxynucleotide triphosphate

ED emergency department

EIA enzyme immunoassay

EPEC enteropathogenic Escherichia coli

FDA Food and Drug Administration

FTH Fremantle Hospital

GDH glutamate dehydrogenase

GDP gross domestic product

GORD gastro-oesophageal reflux disease

GP general practice/general practitioner

h hour(s)

H2RA H2 receptor agonist

HCFA healthcare facility-associated

HISWA Healthcare Infection Surveillance Western Australia

HIV human immunodeficiency virus

HO healthcare facility onset

HREC human research ethics committee

ICU intensive care unit

IDSA Infectious Diseases Society of America

IHD ischaemic heart disease

IP inpatient

IQR interquartile range

IR incidence rate

ISR intergenic spacer region

kb kilobase-pairs

km kilometres

LAMP loop-mediated isothermal amplification

xx

LOS length of stay

MIC minimum inhibitory concentration

MIC50 MIC required to inhibit growth of 50% of organism

MIC90 MIC required to inhibit growth of 90% of organism

min minute(s)

MLST multi-locus sequence typing

MLVA multi-locus variable number of tandem repeats analysis

mM millimoles

n number

N/A not applicable

NAAT nucleic acid amplification test

NPV negative predictive value

NSW New South Wales

NT Northern Territory

OGD oesophago-gastro-duodenoscopy

OP outpatient

OR odds ratio

p probability

PaLoc pathogenicity locus

PathWest PathWest Laboratory Medicine WA

PCR polymerase chain reaction

pd patient days

PFGE pulsed field gel electrophoresis

PMC pseudomembranous colitis

PPI proton pump inhibitor

PPV positive predictive value

pyo person years of observation

QEIIMC Queen Elizabeth II Medical Centre

QLD Queensland

xxi

QX QIAxcel type (local ribotyping nomenclature)

RACF residential aged care facility

rDNA ribosomal DNA

REA restriction endonuclease analysis

RFLP restriction fragment length polymorphism

RPH Royal Perth Hospital

RNA ribonucleic acid

rRNA ribosomal RNA

RT ribotype

rt real-time

SA South Australia

SCGG Sir Charles Gairdner Group

SCGH Sir Charles Gairdner Hospital

SDH Swan District Hospital

SHEA Society for Healthcare Epidemiology of America

SNP single nucleotide polymorphism

SPSS Statistical Package for the Social Sciences

ST sequence type

TAS Tasmania

TcdA toxin A/enterotoxin

TcdB toxin B/cytotoxin

UK United Kingdom

UPGMA unpaired group method of alignment

USA United States of America

UWA The University of Western Australia

VIC Victoria

w/v weight/volume

WA Western Australia

WACHS Western Australian Country Health Service

xxii

WCC white cell count

WGS whole genome sequencing

y year(s)

xxiii

LIST OF TABLES

The page number listed immediately precedes the table.

Table 1.1. Antimicrobial resistance rates and MIC values for Clostridium difficile

isolates from different countries............................................................................... 31

Table 2.1. Toxin PCR mix......................................................................................... 36

Table 2.2. Ribotyping PCR mix................................................................................. 36

Table 2.3. PCR Primers............................................................................................. 36

Table 2.4. Toxin PCR cycling conditions.................................................................. 36

Table 2.5. Ribotyping PCR cycling conditions......................................................... 36

Table 3.1. Prevalence of C. difficile among routine and non-requested groups, and

determination of overall prevalence, July 2014......................................................... 51

Table 3.2. Demographic characteristics of C. difficile-positive cases by request source

and study group.......................................................................................................... 51

Table 3.3. Descriptive characteristics of inpatients with CDI in WA (N=80), December

2011-May 2012......................................................................................................... 51

Table 3.4. Comparison of clinical signs for severe vs non-severe CDI cases.......... 51

Table 3.5. Comparison of characteristics for CO-CDI and HO-CDI cases.............. 51

Table 3.6. Demographic and clinical characteristics of Australian CDI patients in all

States and Territories except NT, September-October 2012. ................................... 52

Table 3.7. RT by acquisition mode and severity, where clinical data were

available..................................................................................................................... 52

Table 3.8. Characteristics of diarrhoeal patients on ED presentation, WA 2013-

2014………………………………………………………………........................... 54

Table 3.9. Characteristics of diarrhoeal patients admitted to hospital from ED........ 55

Table 3.10. Characteristics of CDI patients............................................................... 55

Table 3.11. Risk factor analysis for CA-CDI............................................................ 56

Table 4.1. Results of EIA and culture analysis in Indonesian inpatients................... 76

Table 4.2. Results of EIA analysis in Malaysian inpatients...................................... 77

xxiv

Table 4.3. Demographic and clinical characteristics of Indonesian inpatients with

CDI............................................................................................................................. 77

Table 4.4. Demographic and clinical characteristics of Malaysian inpatients in Kota

Bahru.......................................................................................................................... 78


Bahru, comparing CDI cases with non-CDI cases.....................................................79


Bahru with toxigenic C. difficile vs non-toxigenic C. difficile strains....................... 79

Table 5.1. Summary of all isolates collected………………………......................... 90

Table 5.2. Australian isolates collected November 2012.......................................... 90

Table 5.3. Comparison of top 10 isolates in Australian 2010 and 2012 surveillance

studies........................................................................................................................ 91

Table 5.4. Ribotypes of isolates collected from 70 CDI patients at two WA hospitals,

2011-2012………………………………………….................................................. 91

Table 5.5. WA routine testing study isolates………................................................. 91

Table 5.6. ED study isolates...................................................................................... 91

Table 5.7. Summary of participating sites and recruited participants, Asia-Pacific

study………………………………………………………………………………. 91

Table 5.8. Asia-Pacific C. difficile isolates, overall study total.................................91

Table 5.9. Singaporean isolates collected December 2011 - May 2012.................. 92

Table 5.10. Japanese isolates, collected in Tokyo, 2010........................................... 92

Table 5.11. Indonesian isolates, collected in Central Java Province, July 2014- February

2015........................................................................................................................... 92

Table 5.12. Malaysian isolates, collected in Kota Bahru and Kuala Lumpur, July 2015-

February 2016............................................................................................................ 92

xxv

LIST OF FIGURES

The page number listed immediately precedes the figure.

Figure 1.1. Reports of molecular types of C. difficile circulating in Japan.............. 17

Figure 1.2. Reports of molecular types of C. difficile circulating in Korea............... 18

Figure 1.3. Reports of molecular types of C. difficile circulating in China............... 19

Figure 1.4. Reports of molecular types of C. difficile circulating in (a) Hong Kong, (b) Singapore and (c) Thailand........................................................................................ 21

Figure 2.1. States and Territories of Australia........................................................... 37

Figure 2.2. Locations of Malaysian and Indonesian study sites................................ 38

Figure 3.1. Age distribution of all patients with C. difficile-positive specimens, July

2014........................................................................................................................... 51

Figure 3.2. Age distribution of CDI patients in SCGH and RPH, December 2011-May 2012........................................................................................................................... 51

Figure 3.3. Age distribution of CDI patients in SCGH and RPH, CO-CDI cases only, December 2011-May 2012........................................................................................ 51

Figure 3.4. Age distribution of laboratory-identified CDI cases nationwide Australia, October-November 2012........................................................................................... 53

Figure 3.5. Age distribution of laboratory-identified CDI cases nationwide Australia CO-CDI cases only, October-November 2012.......................................................... 53

Figure 3.6. Age distribution of ED-presenting CDI patients..................................... 54

Figure 5.1. Distribution of C. difficile RTs among patient types............................... 91

Figure 5.2. Distribution of C. difficile RTs in Asia Pacific countries....................... 91

Figure 5.3. Banding patterns of QX 239 and RT 018................................................ 97

1

CHAPTER 1. LITERATURE REVIEW

The following publications have arisen from this chapter:

Collins DA, Elliott B, Riley TV. Molecular methods for detecting and typing of

Clostridium difficile. Pathology 2015;47:211-8.

Collins DA, Hawkey PM, Riley TV. Epidemiology of Clostridium difficile infection in

Asia. Antimicrobial Resistance and Infection Control 2013;2:21.

2

1.1. INTRODUCTION

Clostridium difficile is an anaerobic, spore-forming bacterium first isolated from the gut

microflora of healthy neonates in the 1930s1. Thirty years passed before it was shown to

cause pseudomembranous colitis (PMC) and diarrhoea following administration of

antibiotics, often with fatal outcome2-5. Following widespread use of broad spectrum

cephalosporin antimicrobials during the 1980s and 1990s, C. difficile infection (CDI)

became a common nosocomial issue6.

Since the early 2000s, strains of C. difficile have emerged and spread throughout North

America and Europe causing major epidemics of CDI with significant morbidity and

mortality7-9. C. difficile is currently considered the primary antibiotic resistant threat to

the USA10 and recognised as the most common cause of nosocomial infection in that

country11. In other developed countries including Canada, the UK, Spain, Italy,

Australia and the Netherlands CDI is a significant cause of nosocomial infection and

places a considerable burden on healthcare systems9, 12-18. While data are limited,

C. difficile is likely to be a significant cause of illness in developing countries also.

1.1.1. Pathogenesis of CDI

C. difficile forms spores which are highly resistant to disinfectants, allowing it to persist

for long periods of time in healthcare facilities and infect patients whose normal gut

microflora are depleted by antibiotic use. When infection occurs, ingested spores

germinate in the gut and produce toxins which disrupt junctions between intestinal

epithelia and actin cytoskeleton assembly19. This can manifest clinically in a range of

symptoms including diarrhoea, colitis, severe PMC and, occasionally, fulminant colitis

with toxic megacolon and intestinal perforation. Asymptomatic carriage is commonly

seen in neonates but appears to be rare in adults20. The incubation period of C. difficile

has not been adequately described, but one study estimated that 82% of CDIs occurred

within 4 weeks of a potential transmission from a donor within a hospital21.

1.1.2. Toxins of C. difficile

The main virulence factors of C. difficile are toxin A (TcdA/enterotoxin) and toxin B

(TcdB/cytotoxin). These are encoded by the genes tcdA and tcdB on the pathogenicity

locus (PaLoc), which contains three other genes that regulate expression of tcdA and

tcdB and promote toxin release22. Toxigenic strains always produce toxin B, often

3

produce toxin A also, and cause disease. Non-toxigenic strains do not produce toxins A

and B.

Variant C. difficile strains differ from the reference strain VPI 10463 in restriction sites

and length of tcdA and tcdB, and other PaLoc regions. Many groups of variant strains

have been identified and are defined by a series of overlapping polymerase chain

reactions (PCRs) spanning the PaLoc, assigning strains to toxinotypes I-XXXI. The

best-characterised variant strains are toxin A-negative toxin B-positive (A-B+) strains23.

Currently there are six known toxinotypes of A-B+ strains; VIII, X, XVI, XVII, XXX

and XXXI24, 25. Toxinotype VIII (ribotype [RT] 017) strains are the most common A-B+

strains and are identified by a 1.8kb deletion within repetitive regions of tcdA.

Toxinotype VIII C. difficile strains have caused many outbreaks of CDI worldwide18, 26-

30.

A third toxin produced by C. difficile is binary toxin (CDT). The genes cdtA and cdtB

encode its two components, CDTa (enzymatic component) and CDTb (binding

component), and are found on the CDT locus, which is located independently of the

PaLoc31. While its role in disease is not well understood, binary toxin is known to be an

actin-specific ADP-ribosyltransferase32, 33. It appears to promote adherence to intestinal

epithelia, but additional unknown factors may be involved22. Binary toxin is not

produced by all strains of C. difficile, but has been found in new strains which are

emerging as major causes of epidemics in humans, RTs 027 and 0787, 8, 34.

1.1.3. Epidemiology

C. difficile primarily causes nosocomial infection, and is responsible for 20-25% of

cases of antimicrobial-associated diarrhoea (AAD) in hospital patients35, 36. Nosocomial

reservoirs are likely to be infected and colonised patients and contaminated healthcare

facility environments While some suggest that healthcare workers could become

colonised and transmit infections also37, others do not 37. The main risk factors for CDI

are antimicrobial exposure within the preceding 2-3 months, hospitalisation, long term

residence in a healthcare facility and advanced age. Studies have shown colonisation in

20-40% of hospitalised adults compared with 2-3% in healthy adults22. High rates of

asymptomatic colonisation in neonates have been found with prevalence ranging from

40-84%38, 39. Individuals who develop CDI while hospitalised and >48 h after

admission, are classified as having healthcare facility-associated (HCFA)-CDI 40.

4

The incidence of CDI in Europe is estimated at 7 cases/10,000 patient days (pd)41. In the

USA, the crude incidence of HCFA-CDI is 92.8 per 100,000 population, and

48.2/100,000 for community associated (CA)-CDI42, and C. difficile is estimated to

cause 14,000 deaths each year and cost $1 billion in excess medical costs per year10. In

Australia, the incidence of CDI appears to be lower than in Europe and the USA,

estimated at 4.09/10,000 pd15.

Since 2001, outbreaks of CDI in North America have been attributed to a strain of

C. difficile, referred to as RT 027, or BI by restriction endonuclease analysis (REA)

typing, or NAP1 by pulsed field gel electrophoresis (PFGE), or toxinotype III. This

strain is binary toxin-positive (CDT+), has an 18bp deletion in the tcdC gene and is

resistant to later generation fluoroquinolone antimicrobials7. RT 027 has since spread

beyond North America to Europe, Australia and several Asian countries (albeit in Asia a

fluoroquinolone-susceptible strain as opposed to the epidemic RT 027 strain)19, 43, 44,

although large outbreaks have only been seen in Europe45. The increased incidence of

infections caused by this strain has been accompanied by greater severity of disease,

with higher case-fatality rates and related morbidity43. The greatest impact has been

seen in patients >65 years of age.

C. difficile RT 078 (toxinotype V) has also contributed to the increased incidence of

CDI over the past 10 years. It was the most common RT reported in pigs and cattle in

the USA in 200746, and the third most common RT causing disease in humans in a

European hospital survey45. Like RT 027, RT 078 strains are CDT+ and produce higher

quantities of toxins A and B than other strains. RT 078 strains carry a 39-base pair

deletion in tcdC combined with a point mutation at position 184 resulting in a stop

codon20.

In Australia, another CDT+ strain, RT 244, emerged around 2010 and rapidly increased

in prevalence. It was associated with severe disease in Australia and New Zealand47, 48.

Like RT 027, RT 244 strains have a single nucleotide deletion at position 117 of the

tcdC gene (tcdC∆117)48. Unusually, RT 244 infections have rarely been reported from

other countries.

The true incidence of infections with RTs 027, 078 and 244, and other epidemic strains,

is likely to be underestimated as many countries rarely perform molecular

epidemiological analysis. Many laboratories only test for toxins A and/or B or toxigenic

5

potential, overlooking CDT+ strains. As a result, discovery of emerging strains is often

delayed, and these strains become established in healthcare environments before

infection control measures can be employed.

1.1.3.1. Community-associated CDI

CA-CDI is defined as CDI detected in individuals with no history of hospitalisation in

the preceding 12 weeks40. Recently CA-CDI has been reported with increasing

frequency15, 39, 49 and substantial proportions (22%-35%) of CA-CDI cases have no

recent history of antimicrobial use49, 50. In the absence of antimicrobial use, proton

pump inhibitors (PPIs) have been implicated as a risk factor for CDI or CA-CDI,

however debate continues over whether they play a role51, 52. Clear sources/reservoirs of

CA-CDI have yet to be identified but the ubiquity of spores in the environment suggests

diverse sources exist. Possible reservoirs for CA-CDI include soil, infants, water, pets,

production animals, meat and vegetables22, 50.

1.1.3.2. Animal and food sources of C. difficile

An increasing global demand for meat for consumption has led to significant changes in

farming practices, including administration of antibiotics to farmed animals as

prophylaxis for infections which could eventuate from the crowded conditions of

intensive farms. It is estimated that between 2010 and 2030 global consumption of

antimicrobials will increase by 67%, and in developing countries such as India and

China, antimicrobial consumption in farming could increase by up to 99%53. The

widespread use of antimicrobials contributes to increasing rates of antimicrobial

resistance among bacteria. C. difficile is currently considered the primary antimicrobial

resistant threat organism by the USA Centres for Disease Control (CDC).

Given that CDI is antimicrobial-associated, the widespread use of antimicrobials in

farming suggests that C. difficile may be highly prevalent among farmed animals. Many

studies have now shown that C. difficile colonises and can cause infection in animals

including dogs, horses, pigs, cows, chickens, elephants, and many others20, 54-56. The

prevalence of C. difficile in some studies of farmed piglets has approached 67%57-59.

Strains which are pathogenic in humans, in particular RT 078, have been isolated from

animals, suggesting interspecies transmission may occur20. A whole genome sequencing

study of RT 078 isolates in the Netherlands identified 0 single nucleotide polymorphism

6

(SNP) differences between isolates from pigs and pig farmers, indicating interspecies

spread is possibly occurring60. Several reports have detailed isolation of C. difficile

including RTs 078, 017 and 027 from retail meat products and vegetables, and food

contamination is likely to be contributing to CA-CDI61-63.

1.1.3.3. Infection control for C. difficile infection

Infection control guidelines for CDI patients in hospitals recommend isolation of

patients and implementation of contact precautions, enhanced hygiene and

environmental cleaning64. Antimicrobial surveillance is also an important measure to

reduce the incidence of CDI in hospitals. Several studies have reported decreases in CDI

incidence following reductions in prescribing of certain antimicrobial classes65, 66.

1.2. METHODS FOR DETECTION AND TYPING OF C. DIFFICILE

Rapid diagnosis of CDI is important to allow early isolation and treatment of patients,

reducing potential patient-to-patient transmission and length of hospital stay (LOS) for

those affected. C. difficile strain typing can identify outbreaks within a hospital or the

wider community. Over the past 30 years, significant advances have been made in

molecular techniques for C. difficile detection and typing.

1.2.1. Laboratory diagnosis for CDI

Diagnosis of CDI is dependent upon detecting either production of a C. difficile toxin,

or the ability to produce toxin. Detection of C. difficile alone is insufficient to diagnose

CDI due to the ubiquity of non-toxigenic strains. Moreover, due to the ubiquity of

asymptomatic C. difficile colonisation, diarrhoea (passing unformed, loose or watery

stools at least three times in 24 h) must be present in a patient to confirm a CDI

diagnosis.

Until the early 1980s, the most commonly used test for CDI laboratory diagnosis was

detection of toxin B in stool by cell-culture cytotoxicity assay (CCCA)67, and this has

remained a gold standard in diagnosis to this day. However, due to the complexity of

CCCA and the long turn-around time, rapid enzyme immunoassay (EIA) kits have

become favoured. These EIAs detect the presence of toxins by using antibodies linked

to a chromogenic marker. Initially, these kits were designed to detect toxin A, for two

reasons. First, it was wrongly assumed that C. difficile produced either both toxin A and

7

toxin B, or neither toxin. Second, toxin A was more immunogenic and it was easier to

raise antibodies against it. The discovery of A-B+ strains of C. difficile shifted attention

to EIA kits that detected both toxins, although even these still have greater avidity for

toxin A68. During the process of developing one of the early toxin A kits one

manufacturer inadvertently produced a test that detected glutamate dehydrogenase

(GDH), an enzyme produced by C. difficile and C. sordellii, as well as a few other

bacteria. Detection of GDH remains a useful screening test but detection of C. difficile

toxin is still considered as necessary for confirmation of CDI69.

Culture of C. difficile takes several days, requires anaerobic culture facilities, and does

not differentiate asymptomatic carriers from those with disease (i.e. it lacks specificity),

or toxigenic from non-toxigenic strains, but isolation of the bacterium allows a broader

array of tests to be performed, including an assessment of the organism’s toxigenic

status. So-called toxigenic culture has become the gold standard against which all new

tests should be assessed. The early low sensitivity of EIAs, averaging around 60%70, has

driven the continuing search for highly sensitive, rapid techniques for CDI laboratory

diagnosis. Currently, molecular methods have gained popularity for their rapid turn-

around time and high sensitivity. However, they are expensive and may in fact be too

sensitive for the diagnosis of disease as opposed to detecting the organism71.

1.2.2. Molecular assays for C. difficile detection

Nucleic acid amplification tests (NAATs) for C. difficile have been in use since the

early 1990s. These early PCR tests targeted various genes including tcdA and tcdB72-74.

Initially, they were relatively labour intensive, requiring manual DNA extraction and

visualisation of results using gel electrophoresis or Southern blot analysis. Some early

primers also showed cross-reactivity with other clostridia72. Throughout the 1990s,

DNA extraction kits and NAAT techniques improved, and the development of real-time

PCR (rtPCR) techniques followed. A number of assays gained US Food and Drug

Administration (FDA) approval for use within diagnostic laboratories, ushering in a

new age of rapid diagnostic techniques.

The first NAAT assay to receive FDA approval was the BD GeneOhm™ Cdiff assay in

2009. This and other early assays required a manual DNA extraction step, while most of

the newer assays are fully automated. In the automated assays a small amount of stool

sample is placed into a closed cartridge or tube in which the entire process (DNA

8

extraction, amplification reaction and product detection) takes place. These automated

assays are desirable due to their short hands-on processing time and reduced risk of

cross-contamination of samples.

The majority of the assays employ rtPCR-based reactions, while other methods are

loop-mediated isothermal DNA amplification (LAMP) and helicase-dependent

amplification with chip-based detection. Almost all of the assays require a specific

system instrument for performance of the assay, e.g. the BD MAX™ assay is performed

on the BD MAX™ system. Due to its almost universality in toxigenic strains, tcdB is

generally favoured as a target gene. Exceptions are the illumigene® assay, which targets

a highly conserved portion of the 5’ end of tcdA, and the EasyScreen™ C. difficile

Reflex Kit, which is not designed for primary screening and targets cdtA and a single

base pair deletion at position 117 within the tcdC gene (tcdC∆117) to putatively identify

RT 027 strains75. The EasyScreen™ C. difficile Reflex Kit also detects a gyrA gene

mutation which codes for fluoroquinolone resistance76. The Verigene® and Xpert®

C. difficile/Epi assays also additionally detect the CdtLoc for identification of binary

toxin-positive strains and tcdC∆117. The IMDx C. difficile assay targets a tcdB-variant

gene found in RT 017 that has caused epidemics of CDI in Europe and Asia28, 77.

Reported accuracy measurements of NAATs are generally high but can vary widely for

some assays. The positive predictive value (PPV) of the Verigene® assay was lowest at

42% compared to direct toxigenic culture in one study78, while PPVs for other assays

including BD MAX™, illumigene® and Xpert® exceeded 80%79-83. The illumigene®

assay had the lowest sensitivity of 73-95%82, 84, 85. Reported specificities and negative

predictive values (NPVs) are high, exceeding 90% for most NAATs.

1.2.2.1. Issues with NAATs

While existing data show that NAAT assays for C. difficile are analytically sensitive

and specific, questions remain over their clinical specificity. Individuals can be

colonised with toxigenic C. difficile which would be detected by NAAT. Detection of

toxin genes does not necessarily correlate with expression of toxin, nor disease.

Given the heterogeneity of C. difficile strains, some may not be detected by a particular

assay, or may be incorrectly identified. For example, some strains such as RT 244 also

carry the tcdC∆117 deletion and cdt genes48, leading them possibly to be mis-identified

9

as RT 027 by the Xpert® and Verigene® assays. The illumigene® assays fail to detect

several A- RTs, and particularly RT 033, an A-B-CDT+ strain86. Currently, almost all

assays focus on tcdB detection, but if a tcdB-variant strain was to proliferate, these

assays could fail to identify them. With constant shifts in circulating strains, validation

of assays is required in diagnostic laboratories. Surveillance must be in place to alert

clinical laboratories when a new predominant strain is identified. This information can

be used to assess assay performance in individual laboratories.

Several studies have concluded that inappropriate test ordering can also impact the

clinical specificity of tests. One study found 36% of patients for whom a C. difficile test

was ordered did not have clinically significant diarrhoea, and 20% had recently taken

laxatives87. The Infectious Diseases Society of America, Society for Healthcare

Epidemiology of America (IDSA/SHEA) and the Australasian Society for Infectious

Diseases (ASID) recommend that only unformed/diarrhoeal stool samples be tested for

C. difficile, except in the case of suspected ileus64, 88. While CDI may be suspected in

patients who have been hospitalised for >48 h40, with the increase of CA-CDI it is

important to collect history of recent antibiotic and laxative use. This will help to

determine whether a test for C. difficile is appropriate in any diarrhoeal patient.

Several American studies reported that CDI rates increased when laboratories adopted

molecular testing. Surveillance data showed increases ranging from 43% to 67% in CDI

incidence attributable to changing from toxin EIA to NAAT, which have persisted over

time.89

Some studies have measured variability in test performance depending upon the strain

type present. One study found significant differences in EIA performance for toxin A/B

and GDH were seen compared to the Xpert® C. difficile assay for non-RT 027 strains,

whereas they had equivalent performance for RT 027 strain detection. The sensitivity of

the EIA was significantly lower than that of the Xpert® assay for detecting RTs 002,

027 and 10690. In contrast, other studies have not found a discrepancy in detection based

on strain type91.

Another significant issue with NAATs is that they cost up to two to three times more

than EIAs71. The issue of cost-effectiveness has been addressed in several studies. It has

been postulated that the enhanced accuracy provided by NAATs may lead to fewer tests

for result confirmation ordered overall92. A model based on testing 1,000 patients,

10

assuming 10% prevalence of CDI and using published performance characteristics of

various test methods, was tested. The model showed that NAAT testing alone,

compared to toxin EIA alone and various two-step algorithms using GDH detection,

resulted in the greatest number of CDI patients being isolated within 24 h. The

enhanced accuracy of NAAT would lead to earlier detection of true-positive patients,

reducing the amount of time that false-negative patients (by EIA) could continue to

spread C. difficile within the hospital93.

Another study estimated that in diagnosing CDI 1 day earlier in cases otherwise missed

by EIA, NAAT alone could save US$200,000 annually by avoiding the costs of repeat

testing94. They found that the increased cost of a NAAT was justified by the earlier

detection of CDI cases, again theoretically preventing additional cases of nosocomial

CDI and reducing LOS for CDI patients. Babady et al. also found that using NAAT

alone was cost-effective considering the consequent reductions in turn-around time and

labour costs95.

Some laboratories use a multi-step algorithmic approach for diagnosis, given the high

cost of NAAT. Usually a GDH EIA screen is used initially, and only GDH-positive

specimens are further investigated by NAAT in a two-step algorithm, or in a three-step

approach a toxin EIA is also performed prior to NAAT. While an algorithmic approach

saves on costs, the prevalence of GDH-positive, non-toxigenic C. difficile in the

population may be high, leading to unnecessary secondary NAAT. In addition, as

previously mentioned, EIAs are generally less sensitive than NAAT for C. difficile

detection. Thus laboratories must monitor strains currently in circulation and adjust

their testing approach accordingly. There is no clear optimal approach and testing

guidelines vary from country to country64, 88, 96.

1.2.3. Impact of diagnostic method on patient outcome

While there are many supporters of NAATs alone for C. difficile detection, there are

still arguments for the application of toxin detection. Some believe that at least two

detection methods are required either in parallel or series for C. difficile detection.

While cell culture cytotoxicity assay and toxigenic culture are still considered reference

methods, they are not optimal for rapid detection97. One large English study found that

mortality was significantly higher in patients found positive by cytotoxin assay

compared to a cytotoxin/toxigenic culture negative group (odds ratio [OR] 1.61, 95%

11

confidence interval [CI] 1.12-2.31)69. No significant increase in mortality was apparent

when toxigenic C. difficile alone was detected. These results have been replicated,

showing that patients with a positive NAAT but no detectable toxin have less severe

and shorter-lived symptoms, lower bacterial load, are less likely to develop

complications and have lower all-cause mortality compared with those who had toxin

detected directly98-100.

1.3. C. DIFFICILE TYPING METHODS

Typing methods including PCR ribotyping, multi-locus variable number tandem repeat

analysis (MLVA), multilocus sequence typing (MLST), and PFGE are important for

epidemiological investigation101.

1.3.1. PCR ribotyping

PCR ribotyping is one of the most widely favoured methods of molecular typing. It is

based on amplification of the 16S-23S ribosomal ribonucleic acid (rRNA) intergenic

spacer region (ISR)102. Amplification of this region generates a unique fingerprint

visualised by capillary or slab gel electrophoresis of PCR products or sequencing

reactions using the PCR primers. These fingerprints can be matched with reference

fingerprints by eye, or by using software such as Bionumerics (Applied Maths, Saint-

Martens-Latem, Belgium), which is much more sensitive. This requires PCR or

sequencing facilities as well as access to reference strain fingerprints. The central

reference laboratory is currently based in the Public Health England C. difficile

Ribotyping Network (CDRN) at Leeds General Infirmary, which assigns internationally

recognised numbers to RTs103. PCR ribotyping by capillary electrophoresis also allows

faster visualisation of profiles, as well as exchange of electronic files between

laboratories that use compatible platforms allowing quicker identification of novel RTs.

1.3.2. PFGE

PFGE involves digestion of chromosomal DNA using a rare cutting endonuclease such

as smaI, and separation of the resulting DNA fragments by electrophoresis104. This

generates unique fingerprints that can be visualised and matched with reference strains

in much the same manner as RTs are matched. PFGE is favoured in the USA and

Canada, where the PulseNet network facilitates sharing of fingerprint patterns between

laboratories. PFGE is difficult to standardise due to differences between laboratory

12

equipment and methods, and many isolates are found to be non-typeable due to the

sensitivity of their DNA to degradation.

1.3.3. MLST and other sequence-based typing methods

Sequence-based typing methods are rapidly gaining favour due to their high sensitivity

and the decreasing costs of sequencing technology. MLST involves studying a set of

genes and comparing sequence variations between strains. Allelic types are assigned to

each gene to give an allelic profile. In general housekeeping genes are used, but some

virulence genes are sometimes included. The PubMLST site is the reference standard to

upload allelic profiles and assign widely recognised sequence types (STs)105, 106. This

has the benefit of being available worldwide online. In most cases STs can be equated

with RTs, however some STs correspond to multiple RTs, and some RTs can

correspond to multiple STs107.

A similar method to MLST, MLVA compares the number of tandem repeats present at

several loci and generates an allelic profile108, 109. The greater the number of loci

included, the more discriminatory the method is. Both MLST and MLVA can be

performed by PCR of the target loci and subsequent sequencing, or by whole genome

sequencing and examination of the relevant loci.

Recently whole genome sequencing (WGS) techniques have become more accessible

with a broader array of services and laboratories offering sequencing for rapidly

declining costs. The advent of WGS means it is likely that sequence-based methods will

be favoured in the future for molecular typing of C. difficile. SNP analysis of WGS data

is the most sensitive method of distinguishing between strains of the same RT or ST.

This applies to outbreak investigations and epidemiological studies110, allowing sources

of outbreaks and phylogeographic and evolutionary origins of strains to be traced107.

Currently PCR ribotyping cannot be performed in silico because the 16S-23S rRNA

operon is largely repetitive and thus difficult to reconstruct from WGS data at the

moment. This is likely to change as read lengths and quality improves.

1.4. WORLDWIDE EPIDEMIOLOGY OF C. DIFFICILE INFECTION

Following the emergence of C. difficile RT 027 in the early 2000s, surveillance for CDI

was established in North America and several European countries, giving a broader

understanding of the epidemiology of CDI in those regions. The incidence of CDI has

13

remained high and CDI currently places a heavy burden on the USA healthcare systems.

The most recent estimates of incidence were 66/100,000 person years of observation

(pyo) in persons <65 years of age, and 383-677/100,000 pyo in people aged ≥65

years111. The USA CDC estimates that CDI costs over $1 billion in excess medical costs

per year, with 250,000 infections and 14,000 deaths per year10. CDI-attributable deaths

increased by 400% between 2000 and 2007, largely due to the emergence of RT 02710.

RT 027 has persisted in North America and is still the most common strain isolated,

followed by RTs 106, 001, 002 and 014/020112, 113. Recurrent CDI is frequently reported

from North America with probabilities of 25% for a first recurrence and 38% of a

second in Canada114.

The most recent study of the incidence of CDI in the USA by Olsen et al. identified

proportions of community onset (CO)-CDI that varied widely depending on

administrative datasets examined, from 41.4% of cases in one privately insured

population dataset to 90% among younger persons in another private healthcare dataset.

The study suggested that previous estimations of CO-CDI in the USA may have been

inaccurate due to examining data that omitted many outpatient settings111.

Meanwhile the incidence of CDI in Western Europe has decreased in some countries,

most recently estimated at 7 cases/10,000 pd overall across Europe41. The UK in

particular has reported decreasing incidence rates since 200716, and significant

reductions in deaths related to CDI from 2007-2012115. A survey of testing methods and

prevalence across Europe found that suboptimal testing practices are present in many

European hospitals, leading to widespread underdiagnosis of CDI in European

countries. The study estimated up to 40,000 inpatients with CDI may be undiagnosed

every year in European hospitals41.

The most frequently reported RTs in Europe are RTs 014/020, 001, 078 and 02741, 45.

RT 027 appears to have decreased in prevalence in Western European countries over

recent years. The UK has reported large reductions in prevalence of RT 027, while an

increasingly diverse array of strains are now causing CDI16. RT 027 has persisted

elsewhere in Europe and was most frequently represented in Germany, Hungary, Poland

and Romania in a recent pan-European survey41, and is increasingly reported in other

Eastern European countries including Slovenia116 and Bosnia and Herzegovina117.

14

Reports on CDI from the Middle East are rare, however RT 027 has been reported from

Saudi Arabia118. Otherwise, a report from Lebanon identified a prevalence of 26/88

(29.5%) stools tested by toxigenic culture. The majority of isolates were A-B+,

described as toxinotype 0-like119. Local testing practices were suboptimal and facilities

for identification of strains of interest were lacking.

In South and Central America, little is known on the epidemiology of CDI, given an

absence of local testing due to poor facilities. The incidence of CDI in Brazil was

estimated at 1.8 cases/1,000 pd over 2006-2009 in intensive care unit (ICU) patients120.

Meanwhile a study in Peru in 2007 reported an overall incidence of CDI at 12.9/1,000

admissions121. The incidence of CDI among hospitalised patients in Argentina was

6.4/1,000 pd while among Mexican inpatients it was estimated at 0.4/1,000 pd122.

Another Argentinian study in a Buenos Aires hospital found that the incidence of CDI

fluctuated from 2000 to 2005, peaking at 84 cases/10,000 admissions in 2001123.

Chile has reported isolation of RT 027 as well as RTs 012 and 046124, 125. Brazil has not

reported RT 027 to date126, while RTs 014, 043, 046 and 106 have been reported from

small human and animal isolate collections127-129. The epidemic RT 027 strain of

C. difficile was reported from Costa Rica in 2010130. A variant strain of RT 012 with the

18bp tcdC deletion seen in RT 027 was also identified in Costa Rica and exhibited high

recurrence rate and disease severity131. RT 017 increased in prevalence dramatically in a

hospital in Buenos Aires from 2000 to 2005. MLVA showed a clonal relationship

among Argentinian isolates, and country-specific clonal relationships among

comparative RT 017 isolates from other countries including the Netherlands, Poland,

Japan and Canada123.

In Africa, poor diagnostic facilities in many countries leave the epidemiology of CDI

largely unknown. In Tanzania, the prevalence of C. difficile among 141 diarrhoeal

stools was 6.4% by culture, however positive samples were primarily from children and

toxigenicity was not measured132. Among human immunodeficiency virus (HIV)

patients with diarrhoea in Nigeria, the prevalence of CDI was estimated at 43% for

inpatients and 14% for outpatients133. Meanwhile in Zimbabwe, toxigenic culture of 268

diarrhoeal stool specimens identified a prevalence of 8.6% of C. difficile134. In South

Africa, C. difficile was the most frequently identified bacterial pathogen (16% of 156

diarrhoeal cases) by PCR among diarrhoeal patients in Cape Town135. The predominant

15

strain identified in a South African collection of 32 strains from 145 patients tested was

RT 017 (50%) followed by RT 001 (15.6%). RT 027 or other CDT+ strains were not

detected136.

Thus, while there are plenty of data on the epidemiology of CDI in North America and

Europe, limited data are available for the rest of the world, particularly Asia. A recent

survey found that awareness of CDI in physicians is poor in Asia, with underestimation

of its contribution to antibiotic-associated disease and recurrence rates137.

Comprehensive culture and toxin testing for C. difficile are lacking in many Asian

hospitals. As a consequence, reporting on C. difficile is less common in Asia, so data on

incidence and circulating strains are scarce.

1.5. EPIDEMIOLOGY OF C. DIFFICILE INFECTION IN ASIA-PACIFIC

COUNTRIES

Diarrhoea causes significant morbidity and mortality in Asian countries. Aetiological

studies have reported other diarrhoeal pathogens that are apparently more common than

C. difficile, such as rotavirus, Shigella spp., enteropathogenic Escherichia coli (EPEC),

Vibrio cholerae, Salmonella spp. and Campylobacter jejuni138. Thus, these pathogens

are tested for more frequently than C. difficile. AAD is not commonly reported for

Asian countries either. What data are available for Asian countries, and Australia, are

described below.

1.5.1. East Asia

1.5.1.1. Japan

In Japan, the current population is skewed towards older age; people >65 years

accounted for 20.1% of total Japanese population in 2005 and 68% of hospitalised

patients are aged ≥65 y139. Given the aged hospital population it is likely that CDI is

highly prevalent in Japan, however few reports on the incidence or prevalence of CDI in

Japan exist.

A survey in Tenri Hospital between 1976 and 1988 found that C. difficile was the most

common enteropathogen isolated (423/800) in AAD patients in a total of 3,950

inpatients140. Kobayashi also reported isolation of C. difficile from PMC patients (8/9)

and AAD patients (16/196)141, a lower proportion than that reported by Aihara et al.140.

16

One report from Sapporo described the incidence of CDI from 2010 to 2012 as

3.11/10,000 pd, with a median case age of 78 y. The majority of cases were HCFA-CDI

(93%) while CA-CDI was identified among 3.2% of 126 patients. The mortality rate

was high; 15% died within 30 days of completing treatment, however advanced age and

comorbidities were likely to be contributing factors139. Another study in Tsukuba

reported a decrease in the incidence of CDI from 0.47 to 0.11 cases/1,000 pd from 2010

to 2012. The decrease was concurrent with the introduction of an intervention on

carbapenem prescribing within the hospital, and rates of Pseudomonas aeruginosa

infection also decreased significantly142.

A report from Tokyo described an incidence of CA-CDI of 1.4/100,000 patient years.

CA-CDI was generally not severe and was not associated with death of any patients.

The only significant risk factor identified was antimicrobial exposure.143 More recently

an incidence of 0.6779 per 1,000 pd was reported from a single hospital in Yamaguchi

Prefecture144.

Molecular typing studies have provided epidemiological information on C. difficile in

Japan since the 1990s. A variety of typing techniques have been employed in Japan,

including tcdA and tcdB characterisation74, 145, 146, PFGE147, 148, PCR ribotyping148, 149

and slpA typing150. Application of molecular typing techniques has identified variant A-

B+ strains in regions across Japan23, 30, 146, 151, 152. PCR typing of tcdA on six A-B+ strains

from Japan and Indonesia identified indistinguishable repeating sequences with two

deletions (1,548 and 273 nucleotides in size)145. A later study also grouped A-B+ strains

from Japan and Indonesia, as toxinotype VIII and RT fr/017, with several subtypes

identified by PFGE28. Risk factors for infection with A-B+ strains in one hospital were

exposure to antineoplastic agents, use of nasal feeding tubes and assignment to a

particular internal medicine ward30.

Several ribotyping studies have been performed, indicating a predominance of one

particular RT, “smz”, over the past decade148, 153, 154. A 5-year study in a Tokyo hospital

followed the proliferation of RT “smz” that peaked in 2004 (64% of cases)153. RT

“smz” is recognised internationally as RT 018, and three major subtypes have been

identified within Japan by slpA typing, two of which are widespread150, 155, 156. RT 018

strains have caused CDI in recent years in Korea157, Germany158, Austria, Spain and

Slovenia45, and have been responsible for outbreaks of disease in Italy since 200717.

17

Other common RTs are 014, 002 and 001152-154 (Figure 1.1). A more recent study

outlined the emergence of another A-B+ strain, RT 369. The study analysed 120 isolates

collected in 15 hospitals from 2011-2013. RT 018 predominated (34.2%) and a strain

with a similar banding profile to RT 018, “ysmz” (5.0%), was also common in one

hospital only. RT 369 was the second most common strain (15.8%) and corresponded to

“trf” which was reported previously in Japanese literature several times152, 154(Figure

1.1), most notably causing an outbreak of CDI in Tenri Hospital in 1999/2000, where

ribotyping was not performed at the time30.

RT 027 strains have been isolated only occasionally in Japan153and RT 078 strains have

been reported rarely in humans159 but have caused disease in thoroughbred racehorses

and piglets160, 161.

1.5.1.2. Korea

The overall prevalence of CDI in Korea appears to have increased markedly over the

past decade. A retrospective review in a Seoul hospital found an increase in CDI

incidence from 1.9/10,000 admissions in 1998-1999 to 8.82/10,000 in 2006-2007162.

Moreover, a country-wide survey of 17 tertiary hospitals from 2004 to 2008 found the

incidence of CDI increased from 1.7/1,000 adult admissions to 2.7/1,000 admissions.

The rate of relapse was 8.9%, and complicated CDI was identified among 3.6%.163

Another study in Seoul reported a high incidence of 71.6/100,000 pd for HCFA-CDI

over 2009-2010157. Diagnostic methods likely have differed between sites or changed

over time, contributing to these apparent increases.

Risk factors for recurrent CDI included antibiotic therapy, anaemia, and tube feeding in

one study164, while another found PPI use alone was associated with recurrent

disease165, and another suggested that stool colonisation with vancomycin-resistant

enterococci was a risk factor which could indicate exposure to broad spectrum

antibiotics and healthcare environments166. The proportion of CA-CDI among all CDI

within a Busan hospital was 7.1%166, while 59.4% of cases of CDI presenting at the

emergency department of a Seoul hospital were CA-CDI167. Another study from

Goyang Province from 2012 to 2014 described 251 cases, 213 (84.9%) of which HCFA-

CDI, and 38 (15.1%) CA-CDI168.

Figure 1.1. Reports of molecular types of C. difficile circulating in Japan. (a) n=87, 1996- 1999148; (b) n=148, 1999-2004153; (c) n=71, 2005- 2008152;

(d) n=87, 2003-2007154; (e) n=130, 2012-2013156 (f) n=120, 2010-2013159

18

Variant A-B+ strains have been common in Korea since the 1990s169 and increased in

prevalence significantly between 2002 and 2005, with a peak of 50% in 2004169-172.

These epidemic A-B+ strains belong to RT 017 group and are widespread in Korea169,

170, 173. Subtyping of RT 017 strains revealed six different PFGE pulsotypes and 13

restriction fragment length polymorphism (RFLP)-based subtypes174. Shin et al.

reported that 72% of PMC cases between 2006 and 2010 were caused by A-B+

strains171. Meanwhile RT 018 is the most prevalent strain reported in recent years157, 175-

178, and may have originated from Japan where RT 018 has predominated179 (Figures

1.1, 1.2). Both RTs 017 and 018 were associated with recurrence of disease177, and RT

018 has shown higher rates of resistance to clindamycin and moxifloxacin than other

strains157. RT 027 was first detected in 2006180 in a hospital-associated case of PMC but

has failed to proliferate in Korea157 and RT 078 was reported as the most common

(3.1%) CDT+ strain173, while others including RT 122 have been reported in a minority

of strains175.

1.5.1.3. People’s Republic of China

Few reports on C. difficile from mainland China existed until recently. An English

language review of the Chinese language literature found two studies from Southern

China in general hospital patients with 21/183 patients (1994-1997) and 13/257 patients

(reported in 2006) with CDI (no diagnostic criteria reported)181. Another study from

Beijing identified 36 cases among 71,428 inpatients from 1998 to 2001181. Further

studies were limited to special groups of patients with malignancy and those receiving

chemotherapy or stem cell transplantation, with incidence rates varying from 1.6% to

3.5%, and an exceptionally high rate (27%) reported in stem cell transplant patients181.

In most studies a sampling frame was not reported which, in view of a very low rate of

submission of samples to laboratories in China, makes assessment of the true incidence

of CDI impossible181.

One of the most comprehensive Chinese studies of CDI was conducted between March

2007 and April 2008 at a 1,216 bed hospital in Shanghai182-184. During the study period,

42,936 patients were discharged and 587 patients had stool samples submitted for

testing by toxin assay and culture182. Overall the incidence of CDI was 17.1 per 10,000

admissions183. CDI was mild, possibly due to the younger mean age of patients (62.8

years) compared with a large European survey where 63% were ≥ 65 years45.

Figure 1.2. Reports of molecular types of C. difficile circulating in Korea. (a) n=187, 1980-2006169; (b) n=105, 1995-2008174; (c) n=337, 2006-2008173;

(d) n=194, 2009-2010157; (e) n=67, 2010-2011175; (f) n=98, 2011-2012176; (g) n=51, 2008-2011177; (h) n=1407, 2000-2009178

19

Another study from Huashan examined AAD which developed in 1.0% of 20,437

antibiotic treated hospital patients from 2012 to 2013. Toxigenic C. difficile was

isolated from 30.6% of AAD cases. Again, severe disease was relatively rare and only

one case developed PMC. Carbapenem use significantly increased risk of CDI (OR

2.31, 95% CI 1.22-4.38)184. Another study found a high prevalence of toxigenic

C. difficile at 19.2% among 416 specimens from diarrhoeal patients185.

A study of carriage rates in healthy adults identified a prevalence of 5.5%, 6.3% in

healthcare workers, and 13.4% in healthy kindergarten age children. The molecular

types carried by children differed from those carried among adults, meanwhile healthy

adults carried similar strains to diarrhoeal adults186.

Several molecular typing studies have been reported, some employing ribotyping,

others MLST or a combination of both. RTs 017, 012 and 046 are consistently among

the most commonly reported strains in Chinese isolate collections (Figure 1.3)183, 184, 186-

190. One MLST study collection included 16 Guangzhou isolates from the 1980s

comprised of the RT equivalents: 017 (9 isolates), 046 (1 isolate), 020 (5 isolates) and

ST119 (1 isolate). A study of Chinese ST37 (RT 017) isolates collected from the 1980s

until 2012 revealed low diversity among strains, indicating RT 017/ST37 has persisted

for many decades on mainland China191. In recurrent CDI cases the most common RTs

were RT 020 followed by 001, 002 and 023. Among the 62 cases studied, 32 recurred

with the same strain isolated, otherwise patients were reinfected with a new strain192.

RT 027 was reported from Beijing for the first time in two patients, detected in 2012

and 2013. Both cases were outpatients. While the strains demonstrated high

fluoroquinolone resistance, their MLVA profiles differed from the international

epidemic RT 027 strain193. Another paper reported RT 027 identified in 2012 in a 44

year old female Crohn’s disease patient with recurrent diarrhoea that had lasted for 3

years. She had never been abroad but her medical history included long term

hospitalisation. The strain’s susceptibility to fluoroquinolones was not reported on, and

it is unclear whether the strain was related to the epidemic RT 027 strain.194

RT 078 was also recently reported for the first time from China. Eight isolates were

described, four of which were from patients and four from environmental surfaces

within the two hospitals where patients were affected195.

Figure 1.3. Reports of molecular types of C. difficile circulating in China. (a) n=75 isolates collected in 2008 183; (b) n=110, 2008-2009197; (c) n=104,

1980-2012191, (d) n=21, 2009-2010190; (e) n=94, 2010-2013189; (f) n=181, 2009-2011198; (g) n=63, 2012-2013185; (h) n=62, 2008-2011193; (i) n=33,

2012-2013187; (j) n=94, 2012-2014194

20

1.5.1.4. Taiwan

A similar situation to Japan and China is found in Taiwan where testing is rare so the

true incidence of CDI has not been determined. The number of CDI cases dramatically

increased by 5-6 times in patients ≥65 years between 2003 and 2007 according to one

study199. Increasing awareness and testing probably contributed to this rise as another

report by Chan et al. demonstrated. They found the proportion of positive cultures

remained constant at ~10% while testing increased from 2002 to 2009 as awareness of

CDI increased200. The incidence at a northern Taiwan hospital was 0.45/1,000 pd

overall, and 7.9/1,000 pd in ICU201. The prevalence of C. difficile among diarrhoeal

emergency department patients in Taipei from 2006 to 2009 was 2.3%. The age of

patients was widely distributed, and CDI was observed as occurring more frequently in

the warm, wet spring and summer months202. Isolates were predominantly A-B+ (75%).

Another retrospective study reviewed patients from 2007 to 2013; 225 cases of CDI

occurred among 241,391 inpatients. The majority (68%) were aged >65 y203.

Colonisation at admission was high in a study in Tainan Hospital, where 20.0% of

enrolled patients were colonised, including 13.2% with toxigenic C. difficile.

Colonisation with toxigenic C. difficile occurred after average 60.6 days in hospital.

Colonised patients were at higher risk for CDI than those with no colonisation and those

colonised with non-toxigenic strains204

RT 017 strains of C. difficile were found (6%) in a sample of 142 Taiwanese isolates

where 57 RTs were identified205, however other internationally recognised RTs have

rarely been reported, apart from a few CDT+ strains.

Sporadic reports of RT 027 exist, the earliest from a long term hospitalised 83 year old

male in August 2013206. Other RT 027 cases were identified in 2014, in a 75 year old

male207 and a 69 year old female with no travel history in Jan 2014, whose symptoms

commenced 7 days following admission. The strain was moxifloxacin susceptible, so it

is unlikely to be the epidemic RT 027 strain208.

Other CDT+ strains of C. difficile detected in Taiwan include RT 126, which was

isolated from three cases which clustered between January and June 2012 and were

indistinguishable by repetitive sequence-based PCR. One patient developed PMC and

21

recurrent CDI209. RTs 126 and 127 were also the most frequent RTs identified in pigs in

Taiwan (28% and 43% respectively), while RT 078 was present in 18%57.

1.5.1.5. Hong Kong

CDI has been better recognised and studied for longer in Hong Kong than in mainland

China and Taiwan. The earliest report was from 1991-3 when 20 patients had

C. difficile cultured as part of a survey of the faecal flora in 120 post bone marrow

transplant patients210. A survey at Queen Mary Hospital, Hong Kong in 1996-7

reported 100/3,112 (3.2%) patients with diarrhoea positive by culture for C. difficile211.

A more recent survey at the same hospital using CCCA between September and

December 2008 detected 37/723 (5.1%) positive samples212. The number of patients

diagnosed in 2003 was 82 compared to 66 in 2008, suggesting the overall number of

cases was constant. There are also studies in special groups of patients, most notably 12

cases in the elderly receiving enteral feeding213 and in 38 patients being treated for

tuberculosis214. Testing for CDI increased following the isolation of RT 027 among one

of 12 hyper-toxigenic (detectable cytotoxin at 100-fold dilution) isolates identified in a

3-month study in a university hospital in 2008212. The crude incidence of CDI was

calculated as 22-23/100,000 population from 2011-2013215.

There are few data on RTs of C. difficile circulating in Hong Kong, the most

comprehensive coming from a retrospective study of 345 isolates from 307 patients in

2009 from a healthcare region spanning five hospitals. Unusually 70% were of a

pattern not represented by 23 of the most internationally common RTs, with a further

11.6% being non-typeable (Figure 1.4). RT 002 represented 9.4% of strains, presumed

to be causing cross-infection as the incidence between 2004 and 2008 was 0.53/1,000

admissions which rose in 2009 to 0.95/1,000 admissions. An elevated frequency of

sporulation (20.2%) was found in 35 of these RT 002 isolates compared to 3.7% for 56

randomly selected isolates of other RTs (Figure 1.4)216. RT 017 strains were also found

at a prevalence of 0.7%216. More recently, RT 002 was again the most prevalent strain

identified across three hospitals form 2011-2013, and was associated with increased

morbidity and mortality215. RT 027 was first detected in 2008 in stool of a patient on

steroids with no history of travel in the previous 1.5 years, and as yet there have been no

further reports of 027 from Hong Kong212.

Figure 1.4. Reports of molecular types of C. difficile circulating in (a, b) Hong Kong, (c) Singapore and (d) Thailand. (a) n=345, 2009 216; (b) n=100,

2011-2013218; (c) n=164, 2013219; (d) n=53, 2006-2008220

22

Another study across four hospitals reported that the incidence of CDI increased by

17% from 2008 to 2012. Introduction of infection control measures including contact

precautions and decontamination resulted in an immediate significant reduction by 47%

in 2010, then further declines of -19.4% and -19.8% over 2011 to 2012. The proportion

of RT 002 did not change significantly over time217. Similarly, an outbreak of CDI in a

male rehabilitation ward of a public hospital was contained when infection control

measures were implemented. The majority of cases were caused by RT 002221.

1.5.2. Southeast Asia

1.5.2.1. Philippines

The underestimation of the role of C. difficile in enteric disease in Asia was

demonstrated recently by a study in the Philippines. Historically, patients with colitis

were diagnosed with amoebic colitis, presumed to be caused by Entamoeba histolytica.

However, 43.6% of colitis cases were positive for C. difficile when testing by EIA for

toxin A/B was introduced222. Given that metronidazole is the standard treatment for

both CDI and amoebic colitis, CDI would be masked if testing was not carried out. This

study contrasted a previous report where stools of 37 of 2,135 patients admitted during

2004 who developed nosocomial diarrhoea were tested by EIA for toxin A/B.

Surprisingly, no C. difficile toxin was detected223.

1.5.2.2. Thailand

C. difficile has been a known cause of AAD in Thailand since the early 1990s224-228. In

the earliest report, faecal cytotoxin was detected in 52.5% of 206 diarrhoeal patients

while culture gave low recovery rates (4.8%) over a 26-month period. Cytotoxin was

detected in the stools of 61% of antibiotic-treated patients, and 51% of non-antibiotic-

treated patients224. In contrast, Thamlikitkul et al. reported low detection of toxin A by

EIA in clindamycin-treated (10/140 patients) and patients treated with beta-lactams

(10/140)227. The difference in prevalence found in both studies was probably due to the

use of different toxin assays.

The prevalence of C. difficile in AAD patients was reported as 18.64% by culture and

44-46% by PCR for tcdA and tcdB on the same stool samples, showing the need for

both culture and testing for presence of toxin226. Similarly, a recent study in Bangkok

found that combining toxin EIAs with direct tcdB PCR on stools increased positive

23

diagnoses two-fold, compared with EIAs alone229. Similarly, increased detection of

C. difficile by 20% was also reported when a two-step algorithm was adopted at

Chulalongkorn Hospital in Bangkok230.

Wongwanich et al. found a higher prevalence of CDI in 201 HIV-positive patients

(58.8%) than in 271 HIV-negative patients (36.5%), and six PFGE subtypes were

found231. Again toxin A EIA alone gave lower prevalence than when combined with

culture. Another study on acquired immunodeficiency syndrome (AIDS) patients where

only toxin A EIA was performed reported a lower prevalence of CDI (16/102

patients)232. The discrepancies between studies where lower prevalence is reported

when toxin A assay alone is used suggest it is likely that A-B+ strains are widespread,

and this has been confirmed by their high prevalence where molecular analysis has been

performed220.

Wongwanich et al. reported on a collection of 77 C. difficile isolates from 44

asymptomatic children and infants, and 33 diarrhoeal adult patients. In this study, tcdA

PCR-negative isolates predominated, numbering 57 overall and 18/33 adult diarrhoeal

isolates. Fourteen PFGE pulsotypes and eight subtypes were found225. Meanwhile the

first report on RTs circulating in Thailand revealed a predominance of RTs 017 and

014/020 (Figure 1.4). The mean age of patients was 64 years, and 62.3% were female.

No CDT+ strains were identified. Interestingly, MLST identified these RT 017 strains as

ST45, differing from other regions of Asia where RT 017 was identified as ST37220.

1.5.2.3. Malaysia

Reports on CDI in Malaysia are rare. Seven cases of CDI were reported in 1988

identified over a 3-year period in Kuala Lumpur233. Another study used gas-liquid

chromatography to detect C. difficile from stools; 5/54 samples tested were positive for

C. difficile234. A case report was published in 1997 describing CDI in a patient with

Henoch-Schonlein purpura235. In the most informative study of CDI in Malaysia to date,

toxin A/B assays on 175 stool samples from inpatients with AAD were performed in a

tertiary hospital in Kota Bharu in north-eastern Malaysia. Twenty-four (13.7%) were

positive for toxin, with the majority of infected patients aged >50 years236. In another

study of 159 patients with type 2 diabetes mellitus from the same region, 14 were

positive by toxin A/B EIA. Patients with CDI were younger (mean 53.8 y p=0.02) and

used PPIs more frequently (p=0.01) than non-CDI patients237. Meanwhile another

24

prevalence study performed in Kuala Lumpur identified a prevalence of 6.1% using

toxin A/B EIA among 147 diarrhoeal patients238. No typing data or other molecular

analysis has been reported on Malaysian C. difficile isolates.

1.5.2.4. Indonesia

Like Malaysia, reports on CDI in Indonesia are uncommon. An aetiological study of

diarrhoea in Indonesian children identified C. difficile in 1.3% of stool samples tested.

Toxin A EIA was performed but no culture or toxin B assays were carried out, so the

true prevalence of C. difficile may have been greater138. Toxin A-B+ strains of

C. difficile have been isolated from asymptomatic children <2 years old in Surabaya28,

145, 146. A molecular study of a number of isolates from South-East Asia included eight

isolates from Indonesia, five of which were toxinotype VIII and RT 017, and grouped

with international RT 017 epidemic strains. Two were A+B+ toxinotype 0, and one A-B+

isolate was CDT+, toxinotype XVI28. Otherwise, a cross-sectional study performed in a

rural general hospital from 2003 to 2014 in hepatocellular carcinoma patients identified

4/16 (25%) toxin A/B positive patients239. The small sample population and number of

cases make it difficult to draw any conclusions about the prevalence of C. difficile in

Indonesia.

1.5.2.5. Singapore

A 50-month aetiological study (1985 to 1989) of diarrhoea in 4,508 patients at the

National University Hospital found C. difficile in 35 of only 365 cases where C. difficile

culture was requested. The incidence of CDI at Singapore General Hospital was

3.2/1,000 admissions, with cases more common in males and patients >50 years.

Culture of stool samples identified cases (43%) which were negative by toxin A/B EIA.

CDI was three times more common in Malay patients than in patients of Indian

heritage240. An increase in incidence of CDI was observed in a 1,200 bed general

hospital from 2001 to 2006. The incidence rate rose from 1.49 cases/10,000 pd to

6.64/10,000 pd, with a concurrent increase in positive toxin assays on stool241. A

broader surveillance study in three hospitals found that CDI incidence decreased from

0.52/1,000 pd in 2006 to 0.3/1,000 pd in 2008, as testing increased over the same time

period242.

25

CDT+ strains of C. difficile have been reported in Singapore, including RT 027 strains,

which have caused sporadic hospital-acquired disease243. A collection of 118 C. difficile

isolates in one study contained several unnamed RT groups, one of which was most

likely RT 017 (14 A-B+ isolates)240. Meanwhile the most comprehensive molecular

typing study found C. difficile RT 053 was most common, then RTs 043, 017 and 020

(Figure 1.4)219.

1.5.2.6. Cambodia

The only report of CDI in Cambodia was in 2/40 (5%) of human immunodeficiency

virus (HIV)-positive patients with chronic diarrhoea, identified by toxin A EIA244.

1.5.2.7. Vietnam

The only report on C. difficile from Vietnam involved an evaluative study of a

diagnostic assay involving a multiplex PCR detecting seven diarrhoeal pathogens

including target genes of Shigella spp., Campylobacter spp., rotavirus, norovirus and

adenovirus as well as tcdA and tcdB of C. difficile, which had not been included in

diagnostic assay previously. C. difficile was detected in 45/479 (9%) specimens tested,

387 of which were from children <15 years old. The cost of the assay was high,

meaning it was unlikely to be introduced for routine diagnostic use245.

1.5.3. South Asia

1.5.3.1. India

C. difficile has been recognised as an aetiological agent of colitis, PMC, and hospital-

associated diarrhoea in India since the 1980s246-248. An early report found 21/93 (22.5%)

AAD cases were positive by culture and toxin assay in Nehru Hospital in 1983-1984247.

In Calcutta, C. difficile was isolated in 38/341 (11.1%) hospitalised patients with acute

diarrhoea over 1 year249. A hospital in Delhi reported CDI in 26/156 (16.7%) diarrhoeal

hospitalised patients, detected by culture and toxin A EIA250. Infection control measures

subsequently introduced at this hospital apparently reduced incidence of CDI by more

than 50% over 5 years251. A retrospective review by Ingle et al. in a Mumbai hospital

found 17/99 (17%) patients between 2006 and 2008 were diagnosed with CDI by toxin

A/B EIA252. One report found a prevalence of non-toxigenic C. difficile of 12.6%

among 79 hospitalised patients, five of whom subsequently developed diarrhoea with

26

positive culture and toxin assay. The study group also detected widespread

contamination of surfaces on beds (51%) and hands of hospital workers (62.5%)253.

Several reports exist of acute diarrhoea in hospitalised children (7-11%) caused by

C. difficile254-256. CDI was detected by toxin A/B EIA in 5/50 HIV-positive patients

with diarrhoea257. In a study in Mumbai from 2009 to 2010, 150 patients with acute

diarrhoea were tested by toxin A/B EIA and 12 (8%) were positive. Two patients

(1.3%) only were defined as CA-CDI258.

While several prevalence studies have been reported, the molecular epidemiology of

C. difficile strains in India is not well understood. One study from Chandigarh described

174 isolates, 68 (39.1%) of which were A+B+ while a further 53 (30.5%) carried one

toxin gene, and 16 of which carried cdtA or cdtB259. Another study in 138 children

found a prevalence of 15.2% by culture, and a prevalence of 10.9% by toxin A/B EIA,

but only four isolates were toxigenic260. Meanwhile, a study of C. difficile among pet

dogs found a prevalence of 13.7%; of 10 toxigenic isolates RTs 012, 014 and 046 were

most frequent261. It is likely that similar RTs circulate in people in India also.

1.5.3.2. Bangladesh

In the 1980s, Akhtar reported testing by CCCA directly on 361 stool samples of patients

with gastroenteritis, diarrhoea, and AAD. All samples were negative for C. difficile.

However, culture identified five (two toxigenic and three non-toxigenic) isolates of

C. difficile262. In the 1990s, an aetiological study found that 13/814 (1.6%) children

admitted to hospital with diarrhoea were infected with C. difficile (diagnosed by

CCCA). Seven of the cases were concurrently infected with another diarrhoeal agent263.

No more recent reports are available for Bangladesh.

1.5.3.3. Pakistan

The only mention of C. difficile in literature from Pakistan is in a study of AAD patients

in Karachi. From 2002 to 2009, 191 AAD patients were monitored and 29.2% were

culture-positive for C. difficile, however no further analyses were performed to

determine toxigenicity264.

1.5.4. Australia

27

Despite being a significant cause of morbidity and mortality, few studies on CDI in

Australia existed until recently. An early study in 1983 found C. difficile by culture or

toxin B detection in 26 (14.5%) of 179 hospitalised patients over 6 months265. Over the

following 10 years the incidence of CDI in a Western Australia (WA) hospital increased

significantly from 22 cases/100,000 pd to 50/100,000 pd266. This occurred over the

same period as a substantial increase in cephalosporin use. From 1993 to 1998 the

incidence appeared to stay relatively stable (2-3 cases per 1,000 discharges) then a

significant decrease was detected in 1999 (2.09 to 0.87 per 1,000 discharges) and into

2000. The decrease was attributed to introduction of a change in prescribing policy for

third generation cephalosporin use65, 267.

Following the outbreaks of CDI across North America and Europe caused by RT 027,

surveillance of CDI was made mandatory in Australia to determine whether the strain

had entered the country. Three separate known introductions of the fluoroquinolone-

resistant epidemic RT 027 strain of C. difficile into Australia have occurred, once in

2008268 and twice in 2010269, none of these resulted in outbreaks of a scale mirroring

those in the Northern Hemisphere269. The failure of RT 027 to become established in

Australia may be partly attributable to controls on fluoroquinolone prescribing among

humans270. In addition, fluoroquinolone use among animals is not permitted271.

However, a marked increase in rates of CDI in Australia occurred early in the 2nd

decade of the 2000s, from 3.25/10,000 pd in 2011 to 4.03/10,000 pd in 201215. The

observed increase was unlikely to have been the result of changes in diagnostic

practices, although many Australian laboratories had recently adopted molecular

diagnostic methods15. Interestingly, the greatest increase was seen in rates of CA-CDI,

which comprised 26% of all cases from 2011 to 201215. Similar increases in overall

incidence and in CA-CDI were reported from all States of Australia at this time272, 273.

One partial explanation for the increase in CDI in Australia was the emergence of a

previously unseen strain of C. difficile, RT 244, as a cause of severe CA-CDI. RT 244

strains produce CDT like RT 027, however, are fluoroquinolone susceptible48. A small

case-control study of RT 244 infections conducted in Melbourne from June 2011 to

May 2012 found a 30-day mortality of 42%48. Unusually, a similar outbreak occurred in

New Zealand at the same time, where RT 244 was reported to cause severe CA-CDI

with increased mortality47. Despite the small number of apparent CA-CDI cases caused

28

by RT 244, its recent appearance and proliferation simultaneously in two countries

suggests importation from overseas, most likely in food.

Several animal studies performed in Australia have identified high prevalence of

C. difficile among piglets, sheep and lambs, and veal calves. Many of the RTs of

C. difficile identified have been found frequently in human studies elsewhere, including

RTs 014, 056 and 01254-56, 274. However, remarkably, the most commonly reported

strain among Northern Hemisphere animals, RT 078, has not been identified in

Australian animals.

1.6. EPIDEMIOLOGY

1.6.1. Prevalent C. difficile RTs in Asia

Ribotyping data are available for China, Japan, Singapore, Hong Kong, Taiwan, and

Korea. Overall, the most prevalent RTs in Asia appear to be 017, 018, 014, 002, and

001. While RTs 027 and 078 have caused major outbreaks in North America and

Europe, they have been reported only to have caused sporadic cases or small clusters of

CDI in Asia, in Singapore, Hong Kong, Korea, and Japan153, 180, 205, 212, 216, 243, 275. RT

078 has only been reported from Korea and China to date173, 183. Another CDT+ strain,

RT 130, was recently reported from Korea157.

Meanwhile, RT 017, A-B+, toxinotype VIII strains are widespread in Asia, having

caused epidemics worldwide (Figures 1.1-1.4). In China and Korea RT 017 is the most

common RT of C. difficile in circulation, and this strain is prevalent in Japan, Taiwan,

and Hong Kong also173, 174, 182, 197. Exposure to antineoplastic agents, use of nasal

feeding tubes, and care in a particular hospital ward were associated with infection with

RT 017 strains in one hospital in Japan30. RT 017 strains have persisted in China and

Taiwan while they appear to have declined in Korea (Figure 1.2). The emergence of

C. difficile RT 369, another A-B+ strain in Japan, is also of interest and warrants close

monitoring. ST81 (A-B+) is frequently reported among MLST studies in Asia also.

Corresponding RTs of this ST have not been reported. It is possible that RT 369 and

ST81 correspond to each other, but this cannot be confirmed from the literature

currently.

In Japan, the smz/RT 018 strain of C. difficile appears to have persisted as the most

prevalent RT for over a decade (Figure 1.1). In addition, smz/018 was the most

29

prevalent strain of C. difficile isolated in a tertiary hospital in Seoul between September

2008 and January 2010179, suggesting possible spread from Japan to Korea.

Interestingly, RT 018 has caused outbreaks of CDI in Italy in 2007/2008 and was the

fourth most prevalent RT in Europe in the most recent reports17, 45. It is not clear

whether smz/018 is prevalent in other Asian countries, as comparative typing with

reference smz or 018 strains may not have been performed.

RTs 017 and 018 have caused widespread disease in Asia and across the world. Unlike

the other major epidemic RT 027 and 078 strains, they do not produce binary toxin, and

RT 018 does not appear to possess variant toxin genes17. Some other virulence factors

may have contributed to their spread. The resistance of RT 018 isolates to clindamycin

and fluoroquinolones could have allowed them to be selected148, 179, 216. Another

virulence factor may be high sporulation rates. The epidemic RT 002 isolates in Hong

Kong sporulated at a higher rate than other isolates, potentially allowing them to persist

in the hospital environment and cause outbreaks216.

1.6.2. Infection prevention and control

Few reports of infection control in Asian hospitals exist. A hospital in India introduced

control measures including disinfection of surfaces, rapid detection of C. difficile by

toxin assays, isolation of patients, controls on prescription of antibiotics and education

of staff members. The prevalence of CDI (initially 15% among cases of nosocomial

diarrhoea) was reduced by 50% while the number of tests requested increased as health

workers became more aware of CDI251. A hospital-wide computerised antimicrobial

stewardship scheme was introduced in a hospital in Taiwan. While the incidence of

some antibiotic resistant organisms decreased, the isolation rate of C. difficile remained

constant at 10%200, indicating that other infection control measures besides

antimicrobial stewardship would be required to control CDI in hospitals.

In Korea, a reduction in incidence from 4.7 to 1.53/1,000 pd was reported in an ICU

following ICU-only implementation of infection control procedures including

education, isolation, hand hygiene, contact precautions and environmental

disinfection276. Meanwhile the overall incidence of CDI in the hospital increased from

0.93 to 1.17/1,000 pd. Similarly in Hong Kong, the incidence of CDI decreased

significantly by 47% following implementation of infection control precautions217. In

30

Japan, introduction of restrictions on carbapenem use in one hospital resulted in a

significant decrease of CDI from 0.47 to 0.11 cases/1,000 pd66.

Asia is going through a period of rapid demographic change. With its dense, growing

population, infection control is a pertinent issue. As C. difficile is now the major cause

of nosocomial infection in North America and Europe, control measures could be

applied in Asia to prevent the same situation there. A number of issues exist which

could contribute to the spread of CDI in Asia.

As wealth and the aged population are increasing, more people have access to hospital

care and enter aged care facilities. It is likely that CDI incidence could increase as these

high-risk populations increase in size. For example, modelling of the future age

structure of the Chinese population suggests that there will be a much larger population

at risk for CDI. Using census data from 2005 (population 1.3 billion) when only 100

million individuals were ≥ 65 y, by 2026 there will be 200 million individuals ≥ 63 y277.

Antibiotic use in most Asian countries is poorly regulated. A review of Southeast Asian

countries found that 47% of pneumonia cases do not receive an appropriate antibiotic,

54% of diarrhoea cases are unnecessarily treated with antibiotics, and 40% of antibiotics

are prescribed at inadequate doses278. It is estimated that more than half of the 210,000

tons of antibiotics produced in China in 2007 were used in human medicine279. In many

cases inappropriate antibiotics are prescribed without any laboratory testing. Studies in

India have found the most commonly prescribed antibiotics for cough and respiratory

disease are fluoroquinolones, a known risk factor for CDI278. In addition, antibiotics are

freely available over the counter without prescription in most Asian countries, leading

to misuse in the community279, 280. Moreover, Asian countries show some of the highest

rates of antibiotic consumption in the world. A study found that India and China were

the greatest consumers of antibiotics in 2010, and Hong Kong, Malaysia, Singapore and

South Korea all featured among the top eight antibiotic consumers per person281.

Given the free use of antibiotics by the general public it would be feasible that CA-CDI

could be common in Asia. Studies in Asian countries have neglected to address the

issue of CA-CDI, apart from two studies in Korea which found conflicting proportions

of 7% and 59% of all CDI surveyed being CA-CDI. It would be appropriate to monitor

CA-CDI more closely in Asia in the future, as it is increasingly recognised elsewhere in

the world282.

31

Despite widespread antibiotic use few studies in Asia have measured antimicrobial

susceptibility of clinical C. difficile isolates. High rates of resistance to moxifloxacin,

ciprofloxacin and clindamycin have been found in isolates from Korea, Japan, Northern

Taiwan and China (Table 1.1). Heteroresistance to metronidazole has been reported

from China283, and one study reported resistance among 17.5% of strains284, warranting

close monitoring. Resistance to vancomycin was reported in one strain in Thailand220.

Production and consumption of meat products is also increasing in Asia292. Intensive

farming of poultry, seafood and swine is already in place, and increasing with

worldwide and local demand293. The risk of C. difficile colonisation and/or infection in

animals would most likely increase with intensive farming practices including crowding

of animals and prophylactic antibiotic use. Thus contamination of meat products and

animal-human transmission could occur. To date, few reports have been made of

C. difficile in the environment or animals in Asia. RT 078 was reported in thoroughbred

racehorses in Japan160 and RTs 126, 127 and 078 in pigs in Taiwan57. RTs 014, 012 and

046 were reported in pet dogs in India261.

1.7. CONCLUSIONS

CDI is not widely recognised in Asia so consequently the extent of the disease is not

known. Although relatively few studies on C. difficile have been performed in Asia,

what work that has been done demonstrates that CDI is possibly a significant cause of

nosocomial disease in Asian countries. It does appear that awareness is increasing in the

region, and testing and surveillance are on the rise. However, routine testing is required

to inform on the prevalence of CDI throughout the region, and this stage has not yet

been reached.

In addition, the molecular types of strains in circulation should be taken into account

when selecting an appropriate testing regimen for countries across Asia. Based on in

silico MLST data, phylogenetic analysis has identified six distinct clades of C. difficile

designated 1, 2, 3, 4, 5, and C-I circulating in the world. Certain clades appear to

localise to particular continents, and clade 4 appears to be associated with Asia or South

East Asia. Clade 4 largely consists of non-toxigenic strains, as well as A-B+ strains,

dominated by RT 017107. Another A-B+ strain, RT 369, has also been reported in

significant numbers from Japan. The widespread prevalence of the RT 017 group of A-

B+ strains in Asian countries shows that assays for toxin B or the tcdB genes are

Table 1.1. Antimicrobial resistance rates and MIC values for Clostridium difficile isolates from different countries

Reference Isolates

(n)

[Resistance rate] (%), MIC50 (mg/L), MIC90 (mg/L)

Erythromycin Clindamycin Tetracycline Moxifloxacin Ciprofloxacin Piperacillin/

tazobactam

Metronidazole Vancomycin

China

183 75 [76], 128, 128 [66.7], 128, 128 [41.3], 4, 64 [45.3], 4, 128 [100], 64, 128 [0], <16/4,

16/4 [0], 0.25, 0.25 [0], 1, 2

197 110 [85.3], 128, 128 [88.1], 128, 128 [62.7], 16, 32 [61.8], 16, 128 [100], 64, 128 [0], 8/4, 16/4 [0]a, 0.125, 0.25 [0], 0.5, 1

285 60 [73.3], 16, 64 [41.7], 8, >16 [30] [100]

187 226 [99.6], 16, 16 [0], 0.25, 0.5 [0], 0.25, 0.5

286 9 [100], 256, 256 [100], 256, 256 [22.2], 1, 32 [0], 0.064, 0.125 [0], 0.38, 0.5

193 62 [87.1], -,>128 [16.1], -, 32 [0], -,0.5 [0], -,0.5

284 57 [100], 95, 267 [63.1], 2.8, 37 [7.0], 4.2,

26.0 [17.5], 0.8, 2.56 [0], 0.4, 1.5

189 94 [57.9] [34.0] [32.9] [100] [0] [0]

Hong Kong

216 35 [100] 0.5, 0.75 0.75, 1.5

Japan

287 73 [87.7], >256, >256 [87.7], >256, >256 [100], >32, >32 [0], 0.19, 0.25 [0], 2, 4

153 72 [12]b [100] [0] [0]

156 130 [59], 8, >256 [99], 16, >32 [0], 0.25, 0.5 [0], 1,2

288 157 >32, >32 32, >32 8, 16 0.25, 0.5 0.5, 1

Korea

Table 1.1 (continued)

173 120 [50], 128, >128 [42], 2, 16 [0], 8, 16 [0], 1,4 [0], 0.5, 1

157 111 [82] [83]

179 131 [67.9] [82] [0] [0]

289 123 [75] [85]

290 11 [0], 3,6 [0], 1,2 [0], 8/4, 16/4 [0], 1,2 [0], 0.5, 1

Singapore

240 68 [63], 8, >512 [0], 0.5, 1 [0], 1, 1

Taiwan

291 113 [46], 4, >256 [16] [0]

Thailand

220 53 [45.2], 1.5, >32 >32, >32 [0], 0.06, 0.125 [1.8], 0.5,

0.75

a18 isolates showed heteroresistance to metronidazole, b subset of 25 isolates tested

32

preferable to toxin A assays for diagnosis of CDI. The more virulent epidemic strains

027 and 078 do not appear to have become established in Asia, while RT 017 strains

have caused epidemics, not only in Asia but elsewhere18, 77.

Widespread unregulated antibiotic use and inappropriate prescribing in South East

Asian countries suggests that CDI could be widespread in those regions where

surveillance is currently lacking. Asia may be facing a “perfect storm” as heavy usage

of antibiotics combined with an ageing increasingly hospitalised population. Increasing

laboratory capacity in the region as well as improving surveillance should be seen as

essential in preventing unnecessary morbidity and mortality.

1.8. RESEARCH AIMS

The overarching aim of this PhD project was to describe the epidemiology of CDI in the

Asia-Pacific region. Australia is increasingly being seen as part of Asia, and with a rise

in the movement of people and produce between Asia and Australia, there is a need to

understand CDI in both regions, in particular the molecular epidemiology of strains of

C. difficile causing infection.

With this in mind, specific objectives were to:

• To examine aspects of the epidemiology of CDI in Australia.

• To describe the prevalence and epidemiology of CDI in Indonesia and Malaysia,

two of Australia’s closest neighbours.

• To provide an overview of CDI in other countries in the Asia-Pacific region.

• To describe the molecular epidemiology of C. difficile isolates from Asia-Pacific

countries.

33

CHAPTER 2. MATERIALS AND METHODS

34

2.1. MATERIALS

2.1.1. Media

All agar plates and broths other than ChromID™ were prepared by PathWest

Laboratory Medicine Excel Media, Mt Claremont, Western Australia (WA). These

included:

Cycloserine-cefoxitin-fructose agar containing 0.1% sodium taurocholate (CCFA),

Blood agar (BA) and,

Robertson’s Cooked Meat Broth (CMB) containing 5mg/L gentamicin, 250mg/L

cycloserine and 8mg/L cefoxitin.

ChromID™ C. difficile agar (bioMérieux, Marcy l’Etoile, France) was purchased pre-

prepared by bioMérieux.

2.1.2. Buffers and solutions

Ultra Pure H2O (Fisher Biotec, Perth, WA, Australia).

Chelex-100 solution, 5% (w/v): 50g Chelex-100 resin suspended in 1000mL Ultra Pure

water.

2.1.3. PCR reagents

AmpliTaq Gold® DNA Polymerase (Applied Biosystems, Foster City, CA, USA)

Magnesium Chloride (Applied Biosystems)

Buffer II (Applied Biosystems)

dNTPs (Fisher Biotec)

Primers (GeneWorks, Thebarton, SA, Australia)

Ultra Pure H2O

2.1.4. Kits

MinElute® PCR Purification kit (QIAGEN, Venlo, Limburg, the Netherlands)

35

2.2. LABORATORY METHODS

2.2.1. Culture

2.2.1.1. Pre-reduction of agar plates

CCFA and BA plates were pre-reduced at 35°C in an anaerobic chamber (Don Whitley

Scientific Ltd, Shipley, West Yorkshire, United Kingdom), in an atmosphere containing

80% nitrogen, 10% hydrogen and 10% carbon dioxide at 75% relative humidity for ≥2 h

before use for culture.

2.2.1.2. Direct culture

A swab or 10µL of stool was streaked for single colonies on ChromID® 294 or on pre-

reduced CCFA (CCFA plates were used routinely in the laboratory until 2014, then

replaced with ChromID®). Plates were incubated at 35°C for 24 (for ChromID® plates)

or 48 h (ChromID® and CCFA). C. difficile appeared as yellow, ground-glass, flat,

round colonies with irregular filamentous or smooth edges on CCFA. On ChromID®,

presumptive C. difficile colonies appeared as grey or black ground-glass, flat colonies

with smooth or filamentous edges.

2.2.1.3. Indirect culture

Stool specimens from Malaysia and Indonesia were enriched in CMBs. A swab of

specimen was inoculated into a CMB and incubated aerobically at 35°C for ≥48 h.

Enriched cultures were ethanol-shocked by adding 1mL of post-incubation CMB to

1mL of absolute ethanol. The mixture was incubated at room temperature for ≥1 h, then

10µL was inoculated onto CCFA or ChromID® and incubated as described in section

2.2.1.2.

Colonies with the morphological appearance of C. difficile were subcultured onto pre-

reduced blood agar (BA) and incubated anaerobically for 48 h at 35°C.

2.2.1.4. Toxigenic culture

Culture was performed as described above, then DNA was extracted from C. difficile

isolates on BA and tcdB polymerase chain reaction (PCR, as described in section

2.2.4.1) was performed to confirm the presence of toxigenic C. difficile.

36

2.2.2. Phenotypic testing of C. difficile isolates

C. difficile colonies on BA were identified by characteristic colony morphology (flat,

ground-glass colonies with smooth or filamentous edges), “horse manure” smell and

chartreuse fluorescence under ultraviolet light. C. difficile isolates were further

confirmed using the proline aminopeptidase test (DIATABS; Rosco Diagnostica).

2.2.3. DNA extraction

Bacteria were collected from a pure 48 h culture on BA using a 1µL inoculation loop,

and emulsified in 100µL of 5% w/v Chelex®-100. The solution was vortexed, then

heated to 100°C on a heat block for 12 min, then spun in a centrifuge at 14,000×g for 12

min at 4°C. The resulting supernatant was pipetted into a sterile Eppendorf tube and

stored at -20°C for future use.

2.2.4. Molecular characterisation of C. difficile isolates

2.2.4.1. Toxin profiling

All PCR reactions were performed according to the conditions outlined from Tables 2.1

to 2.4 with a positive control (reference C. difficile strain R6, ribotype [RT] 027) and a

negative (water only) control. Isolates were characterised for toxigenic properties using

PCR reactions for the tcdA, tcdB, cdtA and cdtB genes74, 295. Isolates were categorised as

tcdA+ if they yielded a 250bp product from primers NK3-NK2 (Table 2.3) and a

1,300bp product from primers NK11-NK974. Isolates that did not produce a 250bp

product from NK3-NK2 were categorised as tcdA-. Otherwise isolates were considered

positive for the target toxin gene if they yielded a product from the necessary primers.

The reaction mixes and cycling conditions are detailed in Tables 2.1 and 2.4. Toxin

PCR products were visualised using gel capillary electrophoresis on the QIAxcel

Advanced system (QIAGEN, Venlo, Limburg, the Netherlands).

2.2.4.2. PCR ribotyping

PCR ribotyping was performed according to the methods of Stubbs et al.296 The primers

used were complementary to the 3’ end of the 16S rRNA gene and the 5’ end of the 23S

rRNA gene (Table 2.3). The PCR amplified the intergenic spacer region (ISR). The

reaction mix and cycling conditions are detailed in Tables 2.2 and 2.5.

Table 2.1. Toxin PCR mix.

Reagents Final concentration

UltraPure H2O

PE Buffer 2 1×

MgCl2 2mM

dNTPs 0.2mM

PE Taq Gold 5U/tube

Forward primer 0.2µM

Reverse primer 0.2µM

BSA 0.01%

Template DNA

Total 20µl

Table 2.2. Ribotyping PCR mix.

Reagents Final concentration

UltraPure H2O

PE Buffer 2 1×

MgCl2 4mM

dNTPs 0.4mM

PE Taq Gold 3.75U/tube

CD16S Primer 0.4µM

CD23S Primer 0.4µM

BSA 0.02%

Total 40 µl

Table 2.3. PCR primers used throughout.

Gene Primers Sequence (5’ to 3’) Reference

Ribotyping CD16S CTGGGGTGAAGTCGTAACAAGG Stubbs et al., 1999

CD23S GCGCCCTTTGTAGCTTGACC Stubbs et al., 1999

tcdA NK2 CCCAATAGAAGATTCAATATTAAGCTT Kato et al., 1991

NK3 GGAAGAAAAGAACTTCTGGCTCACTCAGGT Kato et al., 1991

tcdA rep NK9 CCACCAGCTGCAGCCATA Kato et al., 1991

NK11 TGATGCTAATAATGAATCTAAAATGGTAAC Kato et al., 1991

tcdB NK104 GTGTAGCAATGAAAGTCCAAGTTTACGC Kato et al ., 1991

NK105 CACTTAGCTCTTTGATTGCTGCACCT Kato et al., 1991

cdtA cdtApos TGAACCTGGAAAAGGTGATG Stubbs et al., 2000

cdtArev AGGATTATTTACTGGACCATTTG Stubbs et al., 2000

cdtB cdtBpos CTTAATGCAAGTAAATACTGAG Stubbs et al., 2000

cdtBrev AACGGATCTCTTGCTTCAGTC Stubbs et al., 2000

Table 2.4. Toxin PCR cycling conditions.

Temperature (°C) Time Number of cycles

95 10 min 1

94 30 sec

45 55 30 sec

72 1 min

72 7 min 1

4 Hold

Table 2.5. Ribotyping PCR cycling conditions.

Temperature (°C) Time (minutes) Number of cycles

95 10 1

94 1

25 55 1

72 2

72 7 1

4 Hold

37

Following the PCR reaction, the products were purified using the MinElute® PCR

Purification kit (QIAGEN, Venlo, Limburg, the Netherlands). The purified PCR

products were visualised by capillary gel electrophoresis on the QIAxcel Advanced

system (QIAGEN).

Cluster analysis of PCR ribotyping densitometric curves or band profiles was performed

using Dice or Pearson coefficient similarity with relationships represented in unpaired

group method of alignment (UPGMA) or neighbour-joining dendrograms within

Bionumerics v 7.5 (Applied Maths, Saint-Martens-Latem, Belgium). Ribotypes were

identified by comparison of the band profiles with a reference library. The reference

library included strains from the European Centre for Disease Prevention and Control

and the most prevalent ribotypes circulating in Australia (B. Elliott, T. V. Riley et al.,

unpublished data). Isolates which did not correspond to internationally recognised

ribotypes in the library collection were assigned with internal nomenclature prefixed

with “QX” (e.g. QX 001). All PCR ribotyping was performed in the C. difficile

reference laboratory in PathWest at Queen Elizabeth II Medical Centre (QEIIMC).

2.3. STUDY SETTINGS

2.3.1. Australia: Study setting and population

Australia covers a land area of 7.7 km2, with a range of climates from the tropical

northern region, to inland deserts, to warm temperate coastal regions. The country is

divided into two Territories and six States (Figure 2.1). It is sparsely populated by 21.5

million people, 64% of which live in major cities297. The studies described here were

primarily conducted in WA, which spans an area of 2.5 million km2 with a population

of 2.2 million people, mainly concentrated in the Perth Metropolitan Region. The

following sites were included in studies: Armadale Hospital (AH, 265 beds), Fremantle

Hospital (FTH, 450 beds), Royal Perth Hospital (RPH, 855 beds), Sir Charles Gairdner

Hospital (SCGH, 600 beds), Swan District Hospital (SDH, 200 beds), all serving the

Perth Metropolitan region, And Bunbury Regional Hospital (BRH, 100 beds), serving

the South West region of WA.

PathWest Laboratory Medicine (WA) (PathWest) is the single public sector pathology

service for the State of WA. The PathWest enteric laboratory receives and processes

faecal specimens collected in public and private hospitals, and community health

Figure 2.1. States and Territories of Australia.

38

centres, as well as general practitioners throughout the State, processing up to 30,000

specimens per year.

2.3.2. Indonesia: study setting and population

The Indonesian prevalence study was conducted in Central Java Province of Indonesia.

The province population is 32.7 million, with a high density of 1,000 persons/km2 and a

tropical climate. The average gross domestic product (GDP) per capita is $3,475

(www.jatengprov.go.id).

The study was conducted in four hospitals (Figure 2.2): Dr Kariadi Hospital, Semarang

(900 beds), Prof Margono Sukaryo Hospital, Purwokerto (600 beds), Kartini Hospital,

Jepara (400 beds), Ketileng Hospital, Semarang (200 beds).

2.3.3. Malaysia: study setting and population

The Malaysian prevalence study was conducted in Kota Bahru, Kelantan Province and

in Kuala Lumpur, Selangor State. The largest ethnic group in Malaysia is Malay (50%),

followed by Chinese (24%), indigenous (11%), Indian (7%) and others. Kelantan is

relatively isolated and rural compared with other regions of Malaysia. The population of

Kota Bahru is 1.5 million with a density of 97/km2. Kuala Lumpur is the most populous

city in Malaysia with 1.7 million people in the metropolitan area, and 7.5 million in the

greater Kuala Lumpur region (www.statistics.gov.my).

The average GDP per capita in Kelantan Province is US$2,600, which is lower

compared with other provinces (www.kelantan.gov.my). Kuala Lumpur’s average GDP

per capita is considerably higher at US$37,500 (www.statistics.gov.my).

The prevalence study was conducted in: Hospital Universiti Sains Malaysia, Kota Bahru

(700 beds), Hospital Universiti Kebangsaan Malaysia, Kuala Lumpur (870 beds), and

Hospital Sungai Buloh, Kuala Lumpur (620 beds) (Figure 2.2).

2.4. STUDY METHODS

2.4.1. Monitoring routine testing for C. difficile in WA

2.4.1.1. Study background

Figure 2.2. Locations of Malaysian and Indonesian study sites.

39

The enteric laboratory at PathWest was monitored for this study. All faecal specimens

submitted for testing to the diagnostic laboratory were surveyed for the month of July

2014. Routine testing for C. difficile was performed on diarrhoeal specimens meeting

the following criteria: a specimen received from any adult hospital inpatient (IP), any

hospital emergency department (ED) patient, residents of long term care facilities or

patients aged >65 years attending outpatient (OP) clinics. In other cases C. difficile

testing was specifically requested by the physician ordering the test.

Diarrhoeal specimens where C. difficile infection (CDI) testing was specifically

requested or routinely performed (routine group)298 were processed as normal by the

staff of the PathWest enteric laboratory. Routine testing was performed using the BD

MAX™ Cdiff assay, as well as tests for bacterial, parasitic and viral causes of diarrhoea

and gastroenteritis. C. difficile-positive routine specimens were subsequently cultured

on ChromID™ (by A McGovern). Diarrhoeal specimens received where CDI was not

requested or routinely performed (non-requested group) were examined by toxigenic

culture. All routine isolates of C. difficile were PCR ribotyped as part of routine

surveillance activities (by A McGovern, P Putsathit, S Hong and G Androga), and non-

requested C. difficile isolates were ribotyped separately.

The overall prevalence of tcdB-positive C. difficile was calculated. Patients were

grouped according to whether they were hospital IPs, attending OP clinics or a

community general practice (GP). Basic demographic details were available for testing

requests. The overall prevalence and % females were compared with the equivalent

proportions identified by routine testing by χ2. Median ages were calculated and

compared using Mann Whitney U test. Statistical calculations were performed using

GraphPad InStat 3.

2.4.2. A prospective study of C. difficile infection in two tertiary-care hospitals in

Perth

2.4.2.1. Setting and study design

The observational cross-sectional study was conducted at SCGH and RPH. Diarrhoeal

faecal samples sent to the PathWest enteric laboratory for C. difficile testing were

monitored from December 2011 to May 2012.

2.4.2.2. Data collection

40

Patients were eligible for inclusion in the study if they experienced three or more

episodes of diarrhoea within a 24 h period and had a C. difficile-positive laboratory test

result. Patient specimens were screened using glutamate dehydrogenase (GDH) enzyme

immunoassay (EIA), followed by confirmation with the BD GeneOhm™ Cdiff assay.

Patients provided informed consent to be included in the study. Patients were not

approached if they were considered unable to consent due to severity of their current

condition. Medical record reviews and patient interviews with a standardised

questionnaire were performed to collect data regarding previous medication including

antibiotics, their hospitalisation and outcomes. Patient recruitment and data collection

were coordinated and performed by N Foster, C Duncan, W Van Schalkwyk, T Scheller

and S Ditchburn, and the data were provided for analysis.

2.4.2.3. C. difficile culture and ribotyping

C. difficile culture and PCR ribotyping were performed in the C. difficile reference

laboratory (by M Fisher and P Putsathit), and the data were provided for analysis.

2.4.2.4. Statistical methods

Data were collated in a Microsoft Office Excel 2010 database and statistical analyses

were performed using InStat 3.0. Cases were classified as having severe CDI,

community onset (CO)-CDI, and according to location of onset and acquisition.

Prevalences of characteristics associated with severe CDI and with hospital onset (HO)-

CDI and CO-CDI were calculated. Prevalence ratios and their 95% CIs were calculated.

Fisher’s exact and χ2 analyses were performed to compare prevalence ratios.

2.4.3. Laboratory-identified C. difficile infection in Australian healthcare

2.4.3.1. Study setting

The prospective study was conducted within all Australian States and Territories except

the Northern Territory (NT) from October – November 2012 (Figure 2.1). Isolates were

received from one public hospital in the Australian Capital Territory (ACT), from four

public and one private hospital and one private pathology laboratory in New South

Wales (NSW), one public hospital and one private laboratory in Queensland (QLD),

one private and one public laboratory in South Australia (SA), one public hospital in

41

Tasmania (TAS), two public hospitals and a private laboratory in Victoria (VIC), and

one public and two private laboratories in Western Australia (WA), providing coverage

of most healthcare facilities in participating regions. Most participating laboratories

provided diagnostic services to public hospitals. Participating laboratories routinely

performed culture for C. difficile or cultured positive specimens for inclusion in the

isolate collection. Isolates from duplicate specimens taken within 7 days were excluded.

2.4.3.2. Data collection

CDI was diagnosed with either tcdB PCR or positive toxin EIA. Basic demographic and

clinical data were collected by local staff in participating sites. Collection of data and

isolates was coordinated by TV Riley and de-identified data were provided for analysis.

Cases were classified according to location of onset and acquisition. Cases were also

classified as having severe CDI or complicated severe CDI. Median ages and

proportions of characteristics were calculated. Characteristics were compared by the χ2

test.

2.4.3.3. C. difficile culture and ribotyping

Culture and PCR ribotyping were performed in the C. difficile reference laboratory (by

P Putsathit), and the data were provided for analysis.

2.4.3.4. Statistical analysis

Data were collated in a Microsoft Office Excel database and statistical analyses were

performed using InStat 3.0. Median ages were compared using the Mann Whitney U

test and proportions were compared using Fisher’s exact test. Severity was compared

with RT by the χ2 test. Univariate odds ratios (ORs) were calculated for infection with

individual RTs for outcomes of severe CDI, complicated CDI, or community associated

(CA)-CDI.

2.4.4. C. difficile infection in diarrhoeal emergency department patients in WA

2.4.4.1. Study background

42

Six hospitals in WA were involved in the study from July 2013 to June 2014. These

hospitals were SCGH, RPH, SDH, AH, FTH and BRH. During the study period all

patients ≥18 years of age presenting at EDs with diarrhoeal symptoms were asked to

provide a stool sample for testing. Specimens were processed routinely by staff in the

PathWest enteric laboratory. Testing requests were monitored for requests submitted

from EDs of participating hospitals over the study period to identify patients for

inclusion in the study.

2.4.4.2. Data collection and analysis

Patient details and laboratory data for specimens which were positive for a bacterial

diarrhoeal pathogen were collected from the PathWest Ultra database. Medical record

reviews were performed collecting information about basic demographics, recent

medication use including antibiotics and clinical findings during their presentation. If

patients were admitted further information on treatments and outcomes was collected.

Culture and PCR ribotyping were performed as part of routine surveillance (by S Hong

and P Putsathit), and the data were provided for analysis. Patient laboratory data were

collected in a Microsoft Excel 2010 spreadsheet, and medical chart review information

was collected in EpiInfo 7. IBM SPSS Statistics version 22.0 was used to merge de-

identified patient laboratory results with corresponding chart review data, and for all

statistical analysis. Proportions were compared using the χ2 test. Univariate odds ratios

were calculated, and a multivariate model using a backwards stepwise logistic

regression was used to identify risk factors for CA-CDI. Assistance with this work was

kindly provided by A/Professor Linda Selvey, Curtin University. Variables were

included in the multivariate analysis if univariate analysis suggested an association

(p<0.2) or if an association was identified from the literature.

2.4.5. Routine identification of C. difficile in clinical samples in WA

This work was performed by the enteric laboratory at PathWest.

Diarrhoeal patient specimens were screened using a GDH EIA (C. DIFF CHEK™-60;

Alere, Sinnamon Park, Qld, Australia)299. GDH-positive specimens were further

confirmed as containing toxigenic C. difficile using the BD GeneOhm™ assay to detect

43

tcdB (BD, Franklin Lakes, NJ, USA)300. All specimens that were GDH-negative were

regarded as negative for C. difficile.

From April 2014, the BD MAX™ Cdiff assay (BD, Franklin Lakes, NJ)301 replaced the

previous testing algorithm as the sole diagnostic test for C. difficile in PathWest. The

BD MAX™ Cdiff assay detects tcdB.

2.4.6. Data collection - Indonesia

Diarrhoeal stool samples were collected in Indonesian sites from July 2014 to February

2015. Patients were recruited locally and all samples were tested locally at all sites with

QUIK CHECK COMPLETE™. All GDH-positive stool specimens were frozen and

sent to the PathWest C. difficile reference laboratory on commercial transport swabs in

Cary-Blair medium (Medical Wire & Equipment Co. Ltd., England). De-identified

basic demographic and clinical information was provided for analysis.

2.4.7. Data collection - Malaysia

2.4.7.1. In-patient study

Samples in the Malaysian study sites were collected from July 2015 to February 2016.

Study participants were aged 18 to 80 years with symptoms of antibiotic-associated

diarrhoea (AAD). All samples were tested locally with QUIK CHECK COMPLETE™.

De-identified patient demographic and clinical information was provided for analysis.

2.4.8. Data analysis

A CDI case was defined as any patient where a toxigenic strain and/or C. difficile toxin

was detected. Median ages and length of stay (LOS) were calculated with interquartile

ranges (IQRs) and compared using Mann Whitney U test. For Indonesian inpatients,

characteristics of HO-CDI and CO-CDI were compared. For Malaysian inpatients, the

sensitivity and specificity of the EIA test were calculated using toxigenic culture as the

reference standard. Characteristics of Malaysian inpatients with any C. difficile detected

and C. difficile-negative patients were compared using χ2. Characteristics of CDI cases

were compared with all other patients, and with those infected with non-toxigenic

strains by χ2.

2.5. ISOLATE COLLECTIONS

44

Isolates were collected from 12 countries overall, from several separate studies. For all

isolate collections outlined below, PCR ribotyping and toxin PCRs were performed

(Section 2.2.6).

2.5.1. Australian C. difficile isolates

2.5.1.1. 2012 surveillance snapshot study

Isolates for the 2012 national snapshot study described in section 2.4.3 were collected at

the PathWest C. difficile reference laboratory. PCR ribotyping and toxin PCRs were

performed (by P Putsathit) on 542 isolates. The proportions of the top 10 RTs in this

collection were compared to a similar collection from 2010302 using the χ2 test to

determine any significant differences.


Perth

Isolates were collected for 70 patients included in the cross-sectional study of CDI

patients at two WA hospitals described in section 2.4.2. PCR ribotyping and toxin PCRs

were performed by M Fisher and P Putsathit.

2.5.1.3. Monitoring routine testing for C. difficile in WA

Isolates from the study of routine diagnostics of CDI in WA (section 2.4.1) were

collected, totalling 57.

2.5.1.4. C. difficile infection in diarrhoeal emergency department patients in WA

Isolates were collected for the study of C. difficile in emergency department patients

(section 2.4.4). In total 66 isolates were collected.

2.5.1.5. Asia-Pacific C. difficile isolates

A descriptive CDI study was conducted at 48 hospital sites across 13 countries;

Australia, China, Hong Kong, India, Indonesia, Japan, Malaysia, the Philippines,

Singapore, South Korea, Taiwan, Thailand and Vietnam. Cases were recruited from 01

July 2013 to 20 January 2014 with a maximum target of 100 isolates per country. A

CDI case was defined as a patient with at least three episodes of diarrhoea (unformed

45

stool assuming its container’s shape) within a 24 h period, with detection of free toxin

in the stool by EIA or cell culture cytotoxicity assay (CCCA), detection of toxigenic C.

difficile by direct real-time-PCR (RT-PCR) for tcdB or toxigenic culture, or

colonoscopic findings of PMC. Patients with diarrhoea caused by bacteria other than C.

difficile were excluded.

In all countries involved except Australia, China, Indonesia and India, C. difficile assay-

positive specimens were forwarded to LSI Medience Corporation in Tokyo where

culture was performed, using direct plating on ChromID® and enrichment in

cycloserine-cefoxitin fructose broth supplemented with 0.1% sodium taurocholate and

incubated for 7 days, followed by subculture on ChromID®. Isolates were sent to the

PathWest C. difficile reference laboratory.

Australian C. difficile assay-positive specimens were routinely collected by the

PathWest C. difficile reference laboratory. Specimens were cultured directly and

indirectly on ChromID™, then DNA was extracted from BA subcultures and PCR

ribotyping was performed (by A McGovern, S Hong and P Putsathit).

In China, C. difficile assay-positive specimens were cultured in the Chinese Centre for

Disease Control and Prevention, then frozen and sent to the PathWest C. difficile

reference laboratory. Due to Indian government restrictions on sharing of bacterial

isolates, no Indian isolates were collected for the study.

In total, 414 isolates, 15 of which were isolated by enrichment only, were recovered

from 600 recruited cases and forwarded to the PathWest C. difficile reference

laboratory. Isolates were cultured on BA for DNA extraction and molecular

characterisation.

2.5.2. Other Asian isolate collections

2.5.2.1. Singaporean isolates

Isolates were collected for a Singaporean study of the descriptive epidemiology of

CDI303. In total 70 EIA-positive stool specimens were collected and sent to the

PathWest C. difficile reference laboratory on transport swabs in Amies charcoal

medium (Medical Wire & Equipment Co. Ltd., England). The swabs were cultured

46

directly on pre-reduced CCFA and indirectly in CMBs before culture on CCFA. DNA

was extracted from BA cultures for molecular analysis.

2.5.2.2. Japanese isolates

A study of CDI was conducted in the Toho University School of Medicine, a 972-bed

tertiary teaching hospital in Tokyo, Japan. A retrospective study of stool culture was

conducted over 12 months in 2010304. C. difficile was detected by PCR for tcdB.

Patients <2 years of age were excluded from the analysis. A total of 118 isolates of

C. difficile were sent to the PathWest C. difficile reference laboratory for molecular

analysis. The isolates were cultured on BA for DNA extraction and molecular analysis.

2.5.2.3. Indonesian isolates

Molecular typing was performed on 74 isolates collected for the Indonesian prevalence

study described in Section 2.4.6.

2.5.2.4. Malaysian isolates

Isolates were collected for the Malaysian prevalence study described in Section 2.4.7. A

collection of 82 isolates was typed.

2.6. DEFINITIONS

A diarrhoeal faecal specimen was defined as loose or water, assuming the shape of its

container.

CA-CDI was defined as infection identified within 48 h of admission with no hospital

admissions in the previous 12 weeks, or healthcare facility associated (HCFA)-CDI if

discharged from hospital or resident in a residential aged care facility (RACF) in the

previous 4 weeks, otherwise cases were defined as “indeterminate”40. Cases were also

classified by location of onset of infection, i.e. onset in a healthcare facility (HO-CDI)

or community onset (CO-CDI)305. Recurrent CDI was defined as an episode of CDI that

occurred within 8 weeks after a previous, resolved episode305.

Severity of infection was classified according to the Infectious Disease Society

(IDSA)’s Strategies to Prevent Healthcare-Associated Infections in Acute Care

47

Hospitals guidelines. Severe disease was defined as white cell count (WCC) ≥15,000

cells/µL or peak serum creatinine level 1.5 times greater than the premorbid level, or if

not recorded, 1.5 times greater than the normal range (0.5-1.5 mg/dL for men, 0.6-

1.2mg/dL for women). Complicated severe disease was defined as hypotension, shock,

ileus or megacolon.88

2.7. ETHICAL APPROVALS

All the studies described above received ethical approval from local hospital-specific

Human Research Ethics Committees (HRECs), as well as reciprocal approvals from

HRECs including the UWA HREC, the Curtin University HREC and the WA Country

Health Service HREC (Reference numbers 2012-185, 52/2013, AR/13/44, CTBU-588,

RA/4/1/6130, 2013:03, 2014-036, 2011-133 and RA-11/036).

48

CHAPTER 3. DESCRIPTIVE EPIDEMIOLOGY OF

C. DIFFICILE INFECTION IN AUSTRALIA

Work presented in this chapter has been published:

Collins DA, Riley TV. Routine detection of Clostridium difficile in Western Australia.

Anaerobe 2016;37:34-7.

Foster NF, Collins DA, Ditchburn SL, Duncan CN, van Schalkwyk JW, Golledge CL,

Keed ABR, Riley TV. Epidemiology of Clostridium difficile infection in two tertiary-

care hospitals in Perth, Western Australia: a cross-sectional study. New Microbes and

New Infections 2014;2:64-71.

49

3.1. INTRODUCTION

As discussed previously, C. difficile infection (CDI) has not been studied in detail to

date in Australia. The incidence of hospital-identified CDI increased significantly from

3.2/10,000 patient days (pd) in 2011 to 4.0/10,000 pd in 2012. The apparent increase in

infections could be partially attributable to increased testing or adoption of molecular

diagnostics, but studies have also indicated that the greatest proportion of the increased

national incidence was attributable to community associated (CA)-CDI15, 273.

Testing for CDI is not mandatory in diarrhoeal cases in Australia. Many outpatient

clinics and community medical practices do not necessarily consider CDI in diarrhoeal

patients, despite the recent reported increases in CA-CDI in Australia and overseas15, 306.

Furthermore, CDI is not a notifiable disease in Australia, however, the Australian

Commission on Safety and Quality in Healthcare (ACSQH) mandated hospital

surveillance from 2010 in all States and Territories to monitor CDI in all public

hospitals and all private hospitals that also admitted public patients. In Western

Australia (WA), culture and molecular typing of all hospital-identified C. difficile was

performed as a research project within the Healthcare Infection Surveillance WA

(HISWA) program from 2011 to 2015, and this program is currently continuing

routinely within PathWest. Hospital-identified cases are defined as CDI in a patient

either admitted to hospital or attending emergency or outpatient (OP) departments.

The recent increases in CDI overall and in CA-CDI in particular in Australia required

further investigation. This chapter describes four studies examining the epidemiology of

CDI in Australia. The first study examined testing requests for CDI in WA to determine

its prevalence among diarrhoeal patients. The second study was a descriptive analysis of

a cross-sectional study of CDI patients in two hospitals in WA. The third study

examined the descriptive epidemiology of CDI patients across Australia. The fourth

study monitored CDI in patients presenting to hospital emergency departments (EDs)

with CDI in WA.

3.1.1. Aims of this chapter

The overall aim of this chapter was to better understand the epidemiology of CDI in

Australia. The objectives were:

• To examine C. difficile testing practices in Australia.

50

• To determine whether CDI is underdiagnosed in Australia.

• To determine the prevalence of C. difficile among diarrhoeal patients in

Australia.

• To identify CA-CDI patients in Australia.

• To describe disease characteristics of CDI patients in Australia.

3.2. RESULTS

3.2.1. Monitoring routine testing for C. difficile in WA

3.2.1.1. Prevalence of C. difficile

Altogether 1,010 non-duplicate diarrhoeal specimens were submitted for testing during

the study period. Of these, 678 were processed routinely, identifying 49 (7.2%) positive

specimens (Table 3.1). Toxigenic culture was performed on the remaining 332

specimens resulting in 16 (4.8%) positive samples. A second pathogen was identified in

seven non-requested cases, namely Salmonella spp. in four, Campylobacter jejuni in

two, and rotavirus in one.

The overall prevalence of C. difficile among all specimens was calculated as 6.4%

(Table 3.1). The prevalence in general practitioner (GP) specimens was considerably

higher (15.4%) in the routine group than in the non-requested group (1.0%). The overall

prevalence among inpatients (IPs) was 7.2%, and 6.9% and 4.6% in outpatients (OP)

and GP specimens, respectively. The proportion of missed detections, i.e. non-requested

positive specimens among all CDI cases was 24.6%.

3.2.1.2. Sources of testing requests

IP specimens (417, 61.5%) were most common among the routine group, while the

majority of tests were performed on GP specimens in the non-requested group (208,

62.7%). IP CDI cases made up 59.2% of all routine cases and 49.2% of cases overall.

GP cases accounted for 18.5% of all cases overall, while OP cases made up the

remaining 32.3% (Table 3.1), giving a prevalence of community-onset (CO)-CDI of

50.8% among all CDI cases.

3.2.1.3. Demographic characteristics of C. difficile-positive cases

51

The age distribution of all cases is shown in Figure 3.1. In all cases the median age was

calculated as 60.0 y (interquartile range [IQR] 1.5-75.7). The median age of the non-

requested group was 0.8 y (IQR 0.6-1.3), and 72.7 y (IQR 54.8-76.3) in the routine

group (Table 3.2). Females accounted for 64.6% of cases overall.

3.2.2. A prospective study of C. difficile infection in two tertiary-care hospitals in

Perth

3.2.2.1. Descriptive characteristics of CDI cases

A total of 80 patients were recruited from 177 eligible patients with the main reason for

exclusion being inability to consent (58 patients, 32.7%). Six eligible patients declined

to participate in the study, giving a response rate of 93.6%.

The characteristics of the recruited patients are shown in Table 3.3. The age

distributions of all cases and of CO-CDI cases are shown in Figures 3.2 and 3.3,

respectively. The median age of the patients was 60.5 y (IQR 49.0-71.0) and 51.3%

were male. Patients had a median number of 5.5 diarrhoeal episodes in the previous 24

h. The median length of stay (LOS) was 15.0 days (IQR 5.4-40.0). The majority of

patients were Caucasian, with 3.8% Aboriginal and 1.3% Asian ethnicity.

Overall, three or more comorbidities were present in 26.3% of patients; the most

common were gastro-oesophageal reflux disease (GORD, 46.3%), hypertension

(42.5%), chronic pulmonary disease (36.3%), chronic renal disease (36%), cancer

(35%), immune deficiency (35%) and rheumatological disease (31.3%). Overall, 17.5%

of patients had a history of previous CDI and 13.8% of patients had undergone a

colonoscopy, sigmoidoscopy or oesophago-gastro-duodenoscopy (OGD) during the

current admission.

The majority of patients (62.5%) reported antibiotic use within the preceding 3 months

(Table 3.3). The most commonly used antibiotics were amoxycillin/amoxycillin

clavulanate (12.5%) and cephalosporins (11.3%), and 46.5% of patients (excluding

those with CO-CDI) had received antibiotics during the current admission prior to their

CDI diagnosis. The most common antibiotics received for these patients were:

amoxycillin clavulanate (14.0%), piperacillin tazobactam (7.0%), and ciprofloxacin

(7.0%).

Table 3.1. Prevalence of C. difficile among routine and non-requested groups, and determination of overall prevalence, WA July 2014.

Routine Non-requested Overall

Source N Prevalence (%) % of all cases N Prevalence (%) % of all cases Prevalence (%) % of all cases

IP 417 7.0 59.2 29 10.3 18.8 7.2 49.2

OP 209 5.7 24.5 95 9.5 56.2 6.9 32.3

GP 52 15.4 16.3 208 1.9 25.0 4.6* 18.5

Total 678 7.2 100.0 332 4.8 100.0 6.4 100.0

*p<0.05

IP inpatient; OP outpatient; GP general practice

Table 3.2. Demographic characteristics of C. difficile-positive cases by request source and study group.

Group Routine (N=49) Non-requested (N=16) Overall

Source Median age (IQR) % Female Median age (IQR) % Female Median age (IQR) % Female

IP 73.4 (58.0-76.2) 75.9 1.0 (0.8-1.0) 66.6 70.8 (54.2-75.8) 75.0

OP 57.9 (20.7-77.5) 50.0 0.7 (0.6-1.2) 55.6 1.9 (0.7-64.5) 52.4

GP 74.2(60.5-76.1) 62.5 1.2 (0.8-2.3) 100.0 58.8 (1.5-75.7) 58.3

Total 72.7 (54.8-76.3) 67.3 0.8 (0.6-1.3) 55.6 60.0 (1.5-75.7)* 64.6

*p<0.05

IP inpatient; OP outpatient; GP general practice

Table 3.3. Descriptive characteristics of inpatients with CDI in WA (N=80), December

2011-May 2012.

Characteristic n (%)

Median age (years) [IQR] 60.5 [49.0-71.0]

Male 41 (51.3)

Median number of stools in past 24 hours [IQR] 5.5 [4.0-10.0]

Patients with multiple comorbidities (≥3) 21 (26.3)

Recurrent CDI 8 (10.0)

History of CDI 14 (17.5)

Median number of admissions within last 3 months [IQR] 1 [0-2.3]

Median LOS (days) [IQR] 15 [5.4-40.0]

Ethnicity

Caucasian 76 (95.0)

Aboriginal 3 (3.8)

Asian 1 (1.3)

Medical history

GORD 37 (46.3)

Hypertension 34 (42.5)

Chronic pulmonary disease 32 (40.0)

Chronic renal disease 29 (36.3)

Cancer 28 (35.0)

Immune deficiency 28 (35.0)

Rheumatologic disease 25 (31.3)

Chronic liver disease 22 (27.5)

Heart disease 19 (23.8)

Gastritis 16 (20.0)

Type 2 diabetes mellitus 13 (16.3)

Colitis 10 (12.5)

Gastric ulcer 8 (10.0)

Connective tissue disorder 8 (10.0)

Irritable bowel disease 8 (10.0)

Diverticulitis 7 (8.8)

Cerebral vascular disease 2 (2.5)

Type 1 diabetes mellitus 1 (1.3)

HIV 0

Previous antibiotic use in the past 3 months 50 (62.5)

Amoxycillin/amoxycillin clavulanate 10 (12.5)

Fluoroquinolones 1 (1.3)

Metronidazole 6 (7.5)

Trimethoprim sulfamethoxazole 2 (2.5)

Clindamycin 4 (5.0)

Cephalosporins 9 (11.3)

Vancomycin 4 (5.0)

Piperacillin/tazobactam 5 (6.3)

Others 27 (33.8)


Median duration of prior antibiotic use (days)[IQR] 14.0 [8.0-49.0]

Antibiotic use this admission (prior to onset of diarrhoea) 20 (46.5)

Piperacillin/tazobactam 3 (7.0)

Ceftriaxone 1 (2.3)

Cephazolin 2 (4.7)

Amoxycillin/ amoxycillin clavulanate 6 (14.0)

Meropenem 2 (4.7)

Vancomycin 1 (2.3)

Ciprofloxacin 3 (7.0)

Metronidazole 1 (2.3)

Disease characteristics

Received enteral feeding 1 (1.3)

Underwent OGD/colonoscopy/sigmoidoscopy 11 (13.8)

Colonoscopy 3 (3.8)

Sigmoidoscopy 2 (2.5)

Hospital onset 43 (53.8)

Community onset 37 (46.3)

Healthcare facility associated 66 (82.5)

Healthcare associated, hospital onset 43 (53.8)

Healthcare associated, community onset 23 (28.8)

Community associated 6 (7.5)

Indeterminate 8 (10.0)

Severe CDI 16 (20.0)

Complicated severe CDI 16 (20.0)

Death 3 (3.8)

Treatment given

Metronidazole only 47 (58.8)

Vancomycin only 5 (6.3)

Both 21 (26.3)

IQR interquartile range; CDI C. difficile infection; GORD gastro-oesophageal reflux disease; LOS length of stay; HIV human immunodeficiency virus; OGD oesophago-gastro-duodenoscopy; HCFA

Table 3.4. Comparison of clinical signs for severe vs non-severe CDI cases.

Clinical Sign

All CDI (N=80)

n (%)

Non Severe CDI (N=16)

n (%)

Severe CDI (N=64)

n (%)

Prevalence ratio (95% CI)

Dehydration 53 (66.3) 44 (68.8) 9 (56.3) 1.12 (0.87-1.45)

Fever ≥ 38°C 25 (31.3) 20 (31.3) 3 (18.8) 1.13 (0.91-1.39)

Hypotension 14 (17.5) 13 (20.3) 1 (6.3) 1.20 (0.99-1.46)

Ileus 2 (2.5) 0 2 (12.6) 0*

Loss of appetite 68 (85.0) 56 (87.5) 12 (75.0) 1.24 (0.82-1.87)

Malaise 49 (61.3) 39 (60.9) 10 (62.5) 0.99 (0.79-1.23)

Megacolon 1 (1.3) 0 1 (6.3) 0

Nausea 35 (43.8) 27 (42.2) 8 (50.0) 0.94 (0.75-1.18)

Weight loss 48 (60.0) 42 (65.6) 6 (37.5) 1.27 (0.98-1.65)

CDI C. difficile infection

*p<0.05

Table 3.5. Comparison of characteristics for CO-CDI and HO-CDI cases.

Characteristic CO-CDI (N=37)

n (%)

HO-CDI (N=43)

n (%) Prevalence ratio (95% CI)

Age ≥ 65 years 15 (40.5) 17 (39.5) 1.03 (0.60-1.76)

Male 21 (56.8) 20 (46.5) 1.22 (0.80-1.87)

≥2 chronic disease 15 (40.5) 25 (58.1) 0.70 (0.48-1.11)

History of CDI 8 (21.6) 6 (14.0) 1.55 (0.60-4.06)

Severe CDI 6 (16.2) 10 (23.3) 0.69 (0.39-1.50)

Previous antibiotic use in previous 3

months 25 (67.6) 25 (58.1) 1.16 (0.83-1.63)

Antibiotic use this admission (prior to

onset of diarrhoea) 16 (43.2) 26 (60.5) 0.72 (0.46-1.11)

Readmission 23 (62.2) 21 (48.8) 1.27 (0.86-1.89)

Length of stay ≥ 20 days 8 (21.6) 26 (60.5) 0.36 (0.19-0.69)

CDI C. difficile infection; CO community onset; HO hospital onset

Figure 3.1. Age distribution of all patients with C. difficile-positive specimens, July

2014.

Figure 3.2. Age distribution of CDI patients in SCGH and RPH, December 2011-May

2012.

Figure 3.3. Age distribution of CDI patients in SCGH and RPH, CO-CDI cases only,

December 2011-May 2012.

52

Patients were predominantly treated with metronidazole (58.8%); 26.3% received a

combination of both (Table 3.3). Only five patients (6.3%) received vancomycin alone.

A total of 16 patients (20.0%) in the study had severe CDI, two of whom (2.5%) died

within 30 days. The 90-day crude mortality rate was 16.3%. A comparison of clinical

signs of CDI in patients with severe versus non-severe disease is given in Table 3.4.

Among the 80 CDI cases 55% had a previous hospital admission for any reason within

the last 4 weeks. The median number of admissions in the last 3 months was 1 (IQR 0–

2.3). Hospital onset (HO)-CDI was identified in 53.8% of patients. Of the CO-CDI

cases (46.3%), 23 (62.2%) were healthcare facility–associated (HCFA)-CDI, 6 (16.2%)

were community-associated (CA)-CDI, and 8 (21.6%) were of indeterminate

association.

3.2.2.2. Comparison of clinical signs for severe vs non-severe CDI cases

The prevalence of clinical signs associated with severe and non-severe CDI cases was

compared in Table 3.4. No significant differences were noted other than ileus among

severe CDI patients (12.6% vs 0, p<0.05). Other characteristics including fever,

megacolon, ileus and dehydration were distributed evenly among cases.

3.2.2.3. Comparison of characteristics for CO-CDI and HO-CDI cases

Characteristics of HO-CDI and CO-CDI cases were compared in Table 3.5. There were

no significant differences in prevalence characteristics examined other than extended

LOS (≥20 days). Patients with CO-CDI stayed for significantly less time than patients

with HO-CDI (prevalence ratio 0.36, 95% CI 0.19-0.69). CA-CDI patients were

significantly younger (median 57.0 y) compared to HCFA-CDI patients (70.0 y,

p<0.05). CA-CDI was significantly more common among patients aged 30-49 y (22.0%

vs 10.9%, p<0.05) while HCFA-CDI was more common among patients aged ≥65 y

(60.9% vs 39.0%, p<0.01).

3.2.3. Laboratory-identified C. difficile infection in Australian healthcare

3.2.3.1. Demographic and clinical characteristics of CDI cases

In total 542 isolates were collected from participating laboratories. Clinical and

demographic data were collected for 338 cases (Table 3.6). The age distributions for all

Table 3.6. Demographic and clinical characteristics of Australian CDI patients in all

States and Territories except Northern Territory, September-October 2012.

All cases (N=338) HCFA-CDI (N=174) CA-CDI (N=59) p

Female 58.9% 58.0% 55.9% 0.88

Median age (years) 66.0 [40.3-78.0] 70.0 [51.5-80.0] 57.0 [31.0-76.5] 0.02

Age group

0-14 y 9.2% 6.9% 5.3% 0.07

15-29 y 7.1% 6.3% 6.8% 1.0

30-49 y 14.8% 10.9% 22.0% 0.04

50-64 y 16.9% 14.9% 16.9% 0.68

≥≥≥≥65 y 52.1% 60.9% 39.0% 0.004

Recurrence 8.6% 10.3% 6.8% 0.61

Severe CDI 18.1% 21.8% 17.0% 0.46

Complicated CDI 2.1% 3.5% 1.7% 0.68

HCFA-CDI healthcare facility associated C. difficile infection; CA-CDI community

associated CDI

Table 3.7. Ribotype by acquisition mode and severity, where clinical data were available.

Ribotype

n (%)

HCFA CA Indeterminate Unknown Severe CDI Complicated CDI

001/271 (n=5) 3 (60.0) − − 2 (40.0) 1 (20.0) −

002 (n=28) 13 (46.4) 4 (14.3) 1 (3.6) 10 (35.7) 3 (10.7) −

010 (n=7) 5 (71.5) − − 2 (28.5) 2 (28.5) −

014/020 (n=76) 40 (52.6) 12 (15.8) 5 (6.6) 19 (25.0) 13 (17.1) 1 (1.3)

015/193 (n=11) 6 (54.5) 1 (9.1) − 4 (36.4) 2 (18.2) −

017 (n=9) 5 (55.6) 1 (11.1) 1 (11.1) 2 (22.2) 2 (22.2) −

053 (n=8) 4 (50.0) 3 (37.5) − 1 (12.5) 4 (50.0) 1 (12.5)

054 (n=17) 8 (47.1) 5 (29.4) − 4 (23.5) 2 (11.8) −

056 (n=24) 12 (50.0) 5 (20.8) 1 (4.2) 6 (25.0) 3 (12.5) −

070 (n=12) 4 (33.3) 2 (16.7) 1 (8.3) 5 (41.7) 1 (8.3) −

078 (n=4) 1 (25.0) 1 (25.0) 1 (25.0) 1 (25.0) 2 (50.0) −

103 (n=4) 3 (75.0) − − 1 (25.0) − 1 (25.0)

126 (n=4) 4 (100.0) − − − 2 (50.0) 1 (25.0)

244 (n=9) 5 (55.6) 2 (22.2) 1 (11.1) 1 (11.1) 1 (11.1) 1 (11.1)

251 (n=5) 1 (20.0) 1 (20.0) 2 (40.0) 1 (20.0) 1 (20.0) −

Other (n=115) 60 (52.2) 22 (19.1) 14 (12.2) 19 (16.5) 22 (19.1) 2 (1.7)

Total (n=338) 174 (51.5) 59 (17.5) 27 (7.9) 78 (23.1) 61 (18.1) 7 (2.1)

HCFA healthcare facility associated; CA community associated; CDI C. difficile infecton

53

cases, and of CO-CDI cases only, are shown in Figures 3.4 and 3.5. The median age of

patients was 66.0 y (IQR 40.3-78.0), and was significantly lower in CA-CDI patients

(57.0 years) compared to HCFA-CDI patients (70.0 years, p=0.02). CA-CDI was

significantly more common among patients aged 30-49 y (22.0% vs 10.9%, p<0.05)

while HCFA-CDI was significantly more common among patients aged ≥65 y (60.9%

vs 39.0%, p<0.01).

CO-CDI was identified in 116 (34.3%) cases, onset in a hospital or residential aged care

facility (RACF) >48 h following admission in 114 (33.7%) and 11 (3.3%) cases

respectively, or in a hospital or RACF <48 h after admission in 36 (10.7%) cases. The

majority of cases were classed as HCFA-CDI (174, 51.5%; Table 3.6); among which

137 (40.5%) were HCFA-HO-CDI and 37 (10.9%) were HCFA-CO-CDI. Fifty-nine

cases (17.5%) could be defined as CA-CDI, 45 (13.3%) were CA-CO-CDI while

healthcare onset (CA-HO-CDI) occurred in 14 (4.1%). A further 27 (7.9%) cases were

classed as indeterminate, while insufficient data were available to determine the location

of onset or acquisition in the remaining 78 (23.1%) cases.

Severe CDI was present in 61(18.1%) cases, and complicated disease in seven (2.1%)

cases. One patient underwent a colectomy, and one patient was admitted to the (ICU).

Ileus and megacolon were not recorded for any cases. No significant difference was

observed for severe or complicated CDI when CA-CDI and HCFA-CDI cases were

compared (Table 3.6). Recurrent disease was reported in 29 (8.6%) cases, with no

significant difference observed for CA-CDI compared to HCFA-CDI.

Ribotypes (RTs) were compared with disease characteristics in Table 3.7. No specific

RTs were significantly associated with HCFA or CA-CDI, or more severe disease in the

survey, however severe disease was most frequently associated with the binary toxin-

positive (CDT+) RTs 078 and 126 (50.0% each, Table 3.7). By univariate odds ratio

(OR) calculation, only RT 053 was associated with complicated CDI (OR 9.7, 95% CI

1.06-88.82, p=0.04). Otherwise no RT was associated with severe or complicated CDI,

or with CA-CDI.

3.2.4. C. difficile infection in emergency department patients in WA

From July 2013-June 2014 patient details were collected for 463 patients who presented

at emergency departments (EDs) with diarrhoeal symptoms and who provided a

Figure 3.4. Age distribution of laboratory-identified CDI cases nationwide Australia,

October-November 2012.

Figure 3.5. Age distribution of laboratory-identified CDI cases nationwide Australia

CO-CDI cases only, October-November 2012.

54

diarrhoeal stool sample for laboratory testing. A bacterial pathogen (C. difficile or other

bacteria) was identified in 154 of these patients, 77 with C. difficile and 77 with another

pathogen. Data were collected for 154 patients; 77 with CDI and 77 infected with

another pathogen including Campylobacter spp. (46), Salmonella spp. (12) and

Aeromonas spp. (4).

Characteristics of patients presenting with bacterial diarrhoea at EDs are presented in

Table 3.8. CDI patients were more commonly female (66.2%, p<0.01). The age

distribution of CDI cases (Figure 3.6) differed from non-CDI cases (Table 3.8). The

majority of non-CDI cases were aged <35 y (48.1%), while the majority of CDI cases

were aged ≥ 65 y (48.1%). Age groups were distributed more evenly across non-CDI

cases, while variation was seen in CDI cases, where 9.1% were aged 35-49 y while all

other age groups were represented by ≥19%. Residents of long-term care facilities

(LTCFs) were present only among CDI cases (7.8%), and significantly more CDI cases

(22.1%) arrived by ambulance than non-CDI cases (3.9%, p<0.01). HCFA-CDI was

identified in 49.4% of CDI cases while CA-CDI was present in 44.2%. Another 2.6%

were of indeterminate association (Table 3.8) while details of previous hospitalisations

could not be obtained for three CDI cases.

Abdominal pain was more frequently seen among non-CDI cases (76.6% vs 40.3%,

p<0.01), while blood in stool was reported equally across both groups (9.1%). Fever

(temperature ≥38°C) was more frequent among non-CDI cases (37.7% vs 19.9%).

Elevated white cell count (WCC) was more frequent among CDI cases (33.8% vs 8.0%,

p<0.01), as was elevated serum creatinine (10.8% vs 4.1% in non-CDI cases, Table

3.8).

The most common comorbidity among patients was hypertension at 35.1% among CDI

cases and 20.8% among non-CDI cases (p<0.05). Solid tumours were more common

among C. difficile cases (22.1% vs 5.2%, p<0.01). GORD was also frequent among all

patients (14.3% in CDI cases, 15.6% in non-CDI cases, Table 3.8).

Previous medication was reported more frequently in CDI cases (94.8% vs 67.5%,

p<0.01). The most commonly used medication was antibiotics (45.5% vs 11.7%,

p<0.01), followed by proton pump inhibitors (PPIs) (40.3% vs 24.7%, p<0.05). Where

the class of antibiotic was recorded, amoxycillin was most frequently used (17.8%

Table 3.8. Characteristics of diarrhoeal patients on ED presentation, WA 2013-2014.

Characteristic CDI

(N=77)

Other pathogen

(N=77)

Total

(N=154)

Female 51 (66.2) 34 (44.2)** 85 (55.2) Median age (years) [IQR] 63.7 [39.5-77.8] 54.3 [34.1-70.8] 55.7 [34.6-72.9] Age group 18-34 y 18 (23.4) 37 (48.1) 55 (35.7) 35 - 49 y 7 (9.1) 12 (15.6) 19 (12.3) 50 - 64 y 15 (19.5) 13 (16.9) 28 (18.2)

≥ 65 y 37 (48.1) 15 (19.5) 52 (33.8)

ED presentation Ambulance arrival 17 (22.1) 3 (3.9)** 20 (13.0) Long term care facility resident 6 (7.8) 0* 6 (3.9) Recurrent CDI (episode in past 8 weeks)

8 (10.4) N/A

HA-CDI 38 (49.4) N/A CA-CDI 34 (44.2) N/A Indeterminate 2 (2.6) N/A History of CDI 11 (14.3) 2 (2.6)** 13 (8.4) Abdominal pain 31 (40.3) 59 (76.6)** 90 (58.4) Hypertension 8 (10.4) 2 (2.6) 10 (6.5) Hypotension 4 (5.2) 1 (1.3) 5 (3.2)

Fever (≥38°C) 23 (19.9) 29 (37.7) 52 (33.8)

Elevated white cell count 25 (33.8) 6 (8.0)** 31 (20.8) Elevated serum creatinine 8 (10.8) 3 (4.1) 11 (7.5) Blood in stool 7 (9.1) 7 (9.1) 14 (9.1) Comorbidities Inflammatory bowel disease 2 (2.6) 1 (1.3) 3 (1.9) Hypertension 27 (35.1) 16 (20.8)* 43 (27.9) Diabetes 12 (15.6) 6 (7.8) 18 (11.7) Solid tumour 17 (22.1) 4 (5.2)** 21 (13.6) Leukaemia 4 (5.2) 1 (1.3) 5 (3.2) Lymphoma 1 (1.3) 0 1 (0.6) GORD 11 (14.3) 12 (15.6) 23 (14.9) Myocardial infarction 6 (7.8) 1 (1.3) 7 (4.5) Peptic ulcer 3 (3.9) 4 (5.2) 7 (4.5) COPD 5 (6.5) 1 (1.3) 6 (3.9) Liver disease 4 (5.2) 1 (1.3) 5 (3.2) Renal disease 9 (11.7) 1 (1.3) 10 (6.5) Cerebrovascular disease 6 (7.8) 0* 6 (3.9) Congestive heart failure 6 (7.8) 0* 6 (3.9) Connective tissue disease 2 (2.6) 1 (1.3) 3 (1.9) Peripheral vascular disease 2 (2.6) 0 2 (1.3) Previous medication Steroid 11 (14.3) 2 (2.6)** 13 (8.4) Proton pump inhibitor 31 (40.3) 19 (24.7)* 50 (32.5) H2 receptor agonist 1 (1.3) 2 (2.6) 3 (1.9) Probiotic 3 (3.9) 2 (2.6) 5 (3.2) Chemotherapy 8 (10.4) 2 (2.6) 10 (6.5)


Antibiotic 35 (45.5) 9 (11.7)** 44 (28.6) Cephalosporin 8 (11.0) 2 (2.7)* 10 (6.5) Amoxycillin/amoxycillin clavulanate

13 (17.8) 0** 13 (8.8)

Metronidazole 4 (5.5) 1 (1.3) 5 (3.4) Piperacillin/tazobactam 7 (9.5) 0** 7 (4.7) Vancomycin 5 (6.8) 0** 5 (3.4) Trimethoprim/sulfamethoxazole 4 (5.5) 0* 4 (2.7) Fluoroquinolone 5 (6.8) 4 (5.3) 9 (6.1) Laxative 5 (6.5) 2 (2.6) 7 (4.5) Treatment received in ED Pain relief 22 (28.6) 40 (51.9)** 62 (40.3) Antiemetic/antidiarrhoeal 20 (26.0) 42 (54.5)** 62 (40.3) Antibiotic 22 (28.6) 11 (14.3)* 33 (21.4) Cephalosporin 3 (3.9) 0 3 (1.9) Amoxycillin/amoxicillin clavulanate

3 (3.9) 1 (1.3) 4 (2.6)

Metronidazole 9 (11.7) 4 (5.2) 13 (8.4) Vancomycin 3 (3.9) 0 3 (1.9) Fluoroquinolone 3 (3.9) 1 (1.3) 4 (2.6) Admitted 54 (70.1) 29 (37.7)** 83 (53.9) Severe CDI in ED, not admitted 5 (6.5) N/A

Proportions were compared by χ2. *p<0.05 **p<0.01

ED emergency department; IQR interquartile range; CDI C. difficile infection; HA healthcare associated; CA community associated; GORD gastro-oesophageal reflux disease; COPD chronic obstructive pulmonary disease

Figure 3.6. Age distribution of ED-presenting CDI patients.

55

among CDI cases only, p<0.01), followed by cephalosporins (11.0% vs 2.7% in non-

CDI cases, p<0.05). Piperacillin tazobactam, vancomycin and

trimethoprim/sulfamethoxazole were all used by CDI cases only. Steroids and laxatives

were also more frequently used by CDI cases (14.3% vs 2.6%, p<0.01, 6.5% vs 2.6%

respectively, Table 3.8).

CDI cases were admitted for further treatment significantly more frequently than non-

CDI cases (68.8% vs 37.7%, p<0.01, Table 3.8). Among admitted patients, the median

length of stay was significantly longer in CDI cases (8.0 days vs 3.0 days, p<0.01, Table

3.9). Elevated WCC during admission was more frequent among CDI cases (43.1% vs

17.9%, p<0.05). CDI patients more frequently had solid tumours (31.4% vs 10.7%,

p<0.05). Previous antibiotic use was more frequent among CDI cases (41.2% vs 14.3%,

p<0.05), particularly amoxycillin/amoxycillin clavulanate (18.0% vs 0%, p<0.05, Table

3.9).

Severe CDI was present in 26 (51.0%) admitted CDI patients. Six patients (12.0%)

developed pseudomembranous colitis (PMC), one (2.0%) developed ileus, and three

(5.9%) underwent colonoscopy or sigmoidoscopy during their admission. One patient

(2.0%) underwent appendectomy (two non-CDI cases also) and one CDI patient was

admitted to ICU and subsequently died. Overall six patients (8.0%) were deceased after

30 days, all of whom were CDI cases (Table 3.9). No RT was associated with severe or

complicated CDI, or with CA-CDI or HCFA-CDI.

A significantly larger proportion of CA-CDI cases than HCFA-CDI cases were aged 18-

35 y (35.3% vs 7.9%, p<0.01). No CA-CDI cases were deceased after 30 days,

compared to 21.4% of HCFA-CDI cases. Hypertension was significantly more frequent

among HCFA-CDI cases compared to CA-CDI cases (47.4% vs 23.5%, p<0.05). PPI

use was more frequent among HCFA-CDI cases than CA-CDI cases (57.9% vs 23.5%,

p<0.01), as was use of piperacillin tazobactam (Table 3.10).

By univariate analysis, male gender and age 18-35 y were negatively associated with

CA-CDI (OR 0.38 and 0.34 respectively, p<0.05). History of CDI, use of steroids and

antibiotics were positively associated with CA-CDI (ORs 6.47, 6.47 and 5.29

respectively, p<0.05). Following multivariate analysis, age 18-34 y and 35-49 y were

negatively associated with CA-CDI (OR 0.28 and 0.15 respectively, p<0.05). Liver

disease (OR 19.05, p<0.05) and previous antibiotic use (OR 7.31, p<0.001) were

Table 3.9. Characteristics of diarrhoeal patients admitted to hospital from ED.

Characteristic CDI

(N=54)

Other pathogen

(N=29)

Total

(N=83)

Female 33 (44.6) 14 (48.3) 47 (57.3) Age group 18-34 y 5 (6.8) 12 (41.4) 17 (20.7) 35 - 49 y 3 (5.7) 5 (17.2) 8 (9.8) 50 - 64 y 14 (26.4) 4 (13.8) 18 (22.0)

≥ 65 y 31 (58.5) 8 (27.6) 39 (47.6)

Median length of stay (days) [IQR] 8.0 [3.0-13.0]

3.0 [1.25-3.0]** 4.0 [2.0-9.0]

Ambulance arrival 15 (28.3) 3 (10.3) 18 (22.0) Long term care facility 6 (11.3) 0 6 (7.3) Recurrent CDI (episode in past 8 weeks)

6 (11.1) N/A

History of CDI 8 (15.1) 1 (3.4) 9 (11.0) Deceased after 30 days 6 (12.2) 0 6 (8.0)

Fever during admission (≥38°C) 27 (52.9) 12 (42.9) 39 (49.4)

Elevated white cell count 22 (43.1) 5 (17.9)* 27 (34.2) Elevated serum creatinine 10 (20.0) 2 (7.7) 12 (15.8) Severe CDI 26 (51.0) N/A Pseudomembranous colitis 6 (12.0) N/A Ileus 1 (2.0) N/A Underwent colonoscopy/sigmoidoscopy

3 (5.9) 5 (17.9) 8 (10.1)

Underwent appendicectomy 1 (2.0) 2 (8.7) 3 (4.2) Admission to intensive care unit 1 (2.0) 0 1 (1.2) Comorbidities Inflammatory bowel disease 1 (2.0) 1 (3.6) 2 (2.5) Hypertension 23 (45.1) 7 (25.0) 30 (38.0) Diabetes 10 (19.6) 2 (7.1) 12 (15.2) Solid tumour 16 (31.4) 3 (10.7)* 19 (24.1) Leukaemia 3 (5.9) 0 3 (3.8) Lymphoma 1 (2.0) 0 1 (1.3) GORD 9 (17.6) 6 (21.4) 15 (19.0) Myocardial infarction 5 (9.8) 1 (3.6) 6 (7.6) Peptic ulcer 2 (3.9) 3 (10.7) 5 (6.3) COPD 5 (9.8) 1 (3.6) 6 (7.6) Liver disease 4 (7.8) 0 4 (5.1) Renal disease 9 (17.6) 1 (3.6) 10 (12.7) Cerebrovascular disease 6 (11.8) 0 6 (7.6) Congestive heart failure 3 (5.9) 0 3 (3.8) Connective tissue disease 2 (3.9) 1 (3.6) 3 (3.8) Peripheral vascular disease 2 (3.9) 0 2 (2.5) Previous medication Steroid 9 (17.6) 1 (3.6) 10 (12.7) Proton pump inhibitor 27 (52.9) 9 (32.1) 36 (45.6) H2 receptor agonist 1 (2.0) 1 (3.6) 2 (2.5) Probiotic 3 (5.7) 1 (3.6) 4 (5.1) Chemotherapy 8 (15.7) 2 (7.1) 10 (12.7)


Antibiotic 21 (41.2) 4 (14.3)* 25 (31.6) Cephalosporin 5 (10.0) 2 (7.4) 7 (9.1) Amoxycillin/amoxycillin clavulanate 9 (18.0) 0* 9 (11.7) Metronidazole 3 (6.0) 0 3 (3.9) Piperacillin/tazobactam 5 (10.0) 0 5 (6.5) Vancomycin 3 (5.9) 0 3 (3.8) Trimethoprim/sulfamethoxazole 3 (6.0) 0 3 (3.9) Fluoroquinolone 5 (10.0) 2 (7.4) 7 (9.1) Laxative 3 (5.9) 2 (7.1) 5 (6.3)

Denominators vary: admission records were not available for four patients due to transfer to

another hospital

Proportions were compared by χ2, medians by Mann Whitney U test. *p<0.05 **p<0.01

ED emergency department; IQR interquartile range; CDI C. difficile infection; HA healthcare associated; CA community associated; GORD gastro-oesophageal reflux disease; COPD chronic obstructive pulmonary disease

Table 3.10. Characteristics of CDI patients.

Characteristic HCFA-CDI (N=38) CA-CDI (N=34)

Female 24 (63.2) 23 (67.6) Age group 18-34 y 3 (7.9) 12 (35.3)** 35 - 49 y 3 (7.9) 4 (11.8) 50 - 64 y 11 (28.9) 4 (11.8)

≥ 65 y 21 (55.3) 14 (41.2)

Ambulance arrival 9 (23.7) 7 (20.6) History of CDI 6 (15.8) 5 (14.7) Deceased after 30 days 6 (21.4) 0* Comorbidities Inflammatory bowel disease 0 2 (5.9) Hypertension 18 (47.4) 8 (23.5)* Diabetes 8 (21.1) 4 (11.8) Solid tumour 13 (34.2) 3 (8.8) Leukaemia 3 (7.9) 1 (2.9) Lymphoma 1 (2.6) 0 GORD 8 (21.1) 3 (8.8) Myocardial infarction 5 (13.2) 1 (2.9) Peptic ulcer 2 (5.3) 1 (2.9) COPD 4 (10.5) 1 (2.9) Liver disease 1 (2.6) 3 (8.8) Renal disease 6 (15.8) 3 (8.8) Cerebrovascular disease 5 (13.2) 1 (2.9) Congestive heart failure 5 (13.2) 1 (2.9) Connective tissue disease 1 (2.6) 1 (2.9) Peripheral vascular disease 0 2 (5.9) Previous medication Steroid 6 (15.8) 5 (14.7) Proton pump inhibitor 22 (57.9) 8 (23.5)** H2 receptor agonist 1 (2.6) 0 Probiotic 3 (7.9) 0 Chemotherapy 4 (10.5) 3 (8.8) Antibiotic 21 (55.3) 14 (41.2) Cephalosporins 3 (8.3) 5 (15.6) Amoxycillin/amoxycillin clavulanate

9 (25.0) 4 (12.5)

Metronidazole 3 (8.3) 1 (3.1) Piperacillin/tazobactam 7 (18.9) 0** Vancomycin 4 (10.8) 1 (3.1) Trimethoprim/sulfamethoxazole 2 (5.6) 2 (6.3) Fluoroquinolone 4 (11.1) 1 (3.1) Laxative 2 (5.3) 3 (8.8) Ribotype 014/020 17 (44.7) 11 (32.4) 002 1 (2.6) 4 (11.8) 001 1 (2.6) 3 (4.2) 244 1 (2.6) 2 (5.9) 056 1 (2.6) 2 (5.9)

CDI C. difficile infection; HCFA healthcare facility associated; CA community associated; GORD gastro-oesophageal reflux disease; COPD chronic obstructive pulmonary disease; PPI proton pump inhibitor; H2RA H2 receptor agonist

56

positively associated with CA-CDI (Table 3.11), however the number of cases with

liver disease was small (n=3) so this finding may not be significant.

Two patients with HCFA-CDI had visited Thailand to undergo elective cosmetic

surgery in the previous month. One of those patients was infected with RT 017, the

most common strain in Asia overall307, and in Thailand220. The other patient was

infected with QX 369, a RT pattern which matched a single isolate from Thailand (from

an isolate collection of P Putsathit)220. It appears that both patients were infected with

C. difficile during their overseas hospital stay and became ill just prior to or following

their return to Australia. Another patient infected with RT 017 experienced severe CDI,

was admitted to ICU, and died due to a combination of comorbidities and CDI. They

had no history of recent travel and had HCFA-CDI.

3.3. DISCUSSION

3.3.1. C. difficile testing practices in Australia

Identification of the ability to produce toxin, generally by polymerase chain reaction

(PCR), is the most common method of C. difficile detection in Australia. Over the

course of this PhD project, the testing practice in PathWest was either an initial

glutamate dehydrogenase (GDH) enzyme immunoassay (EIA) screen followed by tcdB

PCR using BD GeneOhm™, or tcdB PCR alone using BD MAX™, introduced in April

2014. The nationwide study described in section 3.2.3 included the major pathology

laboratories for Australia in all States and Territories except the Northern Territory

(NT). All isolates included in the study were identified by tcdB PCR or by toxin EIA.

Details of which test type was used by the providing laboratory were not provided with

isolates, however tcdB PCR remains likely the most common diagnostic method used in

Australia currently. While highly sensitive, PCR may lead to overdiagnosis of CDI by

identifying cases of colonisation and may have contributed to the reports of increases in

incidence across the country69, 308.

C. difficile testing requests within WA were examined in section 3.2.1. The criteria

within PathWest for testing for C. difficile were a diarrhoeal specimen received from

any adult hospital inpatient IP, any hospital ED patient, residents of LTCFs or patients

aged >65 y attending OP clinics. Otherwise C. difficile testing would be performed only

Table 3.11. Risk factor analysis for CA-CDI.

Characteristic Univariate OR

(95% CI)

p Multivariate

OR

p

Male 0.38 (0.16-0.88) 0.03 Age group 18-34 y 0.34 (0.13-0.92) 0.03 0.28 0.02 35 - 49 y 0.36 (0.09-1.37) 0.13 0.15 0.02 50 - 64 y 0.33 (0.09-1.26) 0.10 0.21 0.05

≥ 65 y Ref Ref

History of CDI 6.47 (1.19-35.21) 0.03 Comorbidities Inflammatory bowel disease 4.75 (0.42-54.27) 0.21 Hypertension 1.17 (0.45-3.08) 0.75 Diabetes 1.58 (0.42-6.00) 0.50 Solid tumour 1.77 (0.37-8.36) 0.47 Leukaemia 2.30 (0.14-37.94) 0.56 Lymphoma GORD 0.52 (0.14-1.99) 0.34 Myocardial infarction 2.30 (0.14-37.94) 0.56 Peptic ulcer 0.55 (0.06-5.14) 0.60 COPD 2.30 (0.14-37.94) 0.56 Liver disease 7.36 (0.74-73.46) 0.09 19.05 0.03 Renal disease 7.36 (0.74-73.46) 0.09 Cerebrovascular disease Congestive heart failure Connective tissue disease 2.30 (0.14-37.94) 0.56 Peripheral vascular disease Previous medication Steroid 6.47 (1.19-35.21) 0.03 Proton pump inhibitor 0.94 (0.36-2.42) 0.90 H2 receptor agonist Probiotic Chemotherapy 3.63 (0.58-22.79) 0.17 Antibiotic 5.29 (1.99-14.02) 0.001 7.31 <0.001 Laxative 3.63 (0.58-22.79) 0.17

CDI C. difficile infection; OR odds ratio; CI confidence interval; GORD gastro-oesophageal reflux disease; COPD chronic obstructive pulmonary disease

57

if specifically requested by the treating physician. As shown in Table 3.1, GPs rarely

requested C. difficile testing (52/678, 7.7%). This would suggest that GPs rarely

considered C. difficile in cases of diarrhoea within the community. However, it is

interesting to note that the prevalence of C. difficile was highest among GP-requested

specimens (15.4%), indicating that GPs often correctly suspected CDI among patients,

probably by collecting a history of recent hospitalisation or antimicrobial use.

A minority of specimens from hospital IPs (29), from paediatric IPs, were not routinely

tested, compared to 417 routinely tested IP specimens. Paediatric IP specimens were

only routinely tested if specifically requested. The identification of paediatric specimens

that were positive for C. difficile, where no other diarrhoeal pathogen could be

identified, raises questions about whether the current testing practice is adequate in the

case of children.

A previous study in WA in 2011-2012 found a high prevalence of C. difficile (45%)

among diarrhoeal paediatric patients309. Furthermore, increasing rates of CDI among

paediatric IPs have been reported from the USA 310-312. High rates of C. difficile

colonisation among infants and a high proportion of cases of co-infection with another

diarrhoeal pathogen are seen in paediatric patients 313. It has been previously suggested

that the likelihood of isolating C. difficile from a stool sample may be increased by co-

infection with other pathogens as the altered gut flora (as a result of the co-infection)

may allow C. difficile to proliferate, and could plausibly be the case in the 10 patients

identified with co-infection.

Neonates frequently carry C. difficile but appear to be protected against CDI due to

absence of the receptor for C. difficile toxin, which is produced as the infant grows and

a more mature microflora appears; thus the risk of infection increases with age314. A

recent study from the USA315 reported that children diagnosed with CA-CDI by PCR

often lacked traditional risk factors and could not be confirmed as having CDI using

different testing methods, suggesting overdiagnosis of CDI in children with community

onset diarrhoea. This may explain the recent reports of increases in incidence of CDI in

children in the USA. However, in the case of paediatric patients with diarrhoea, with

the classic risk factors of history of recent hospitalisation or antimicrobial use, toxigenic

C. difficile could plausibly be the cause of infection. In these cases, C. difficile testing

may be appropriate.

58

It is striking to note that 50.8% of C. difficile-positive specimens overall came from

community patients (attending OP clinics and GPs), and could be classed as CO-CDI.

Among the 33 CO-CDI cases, 13 were not identified by routine testing. These data

suggest that C. difficile may be circulating at high levels in the community in Australia.

In light of this, community health practices, OP clinics and GPs should be made aware

of the recent high incidence of CA-CDI and remain vigilant for CDI among community

patients316. CO-CDI and CA-CDI in Australia are discussed in further detail below.

Australian testing practices for C. difficile appear to be appropriate, however,

overdiagnosis may be occurring due to favouring of PCR as a diagnostic assay. GPs

appear to identify patients at risk for CDI well, however, ongoing vigilance is required

in the light of increasing CA-CDI rates and the high proportion of C. difficile identified

among community diarrhoeal patient specimens. Lastly, criteria for testing paediatric

patients may need to be re-evaluated, however appear to be largely appropriate.

3.3.2. Prevalence of C. difficile among Australian diarrhoeal patients

The prevalence study described in section 3.2.1 identified that C. difficile was the most

commonly identified pathogen in diarrhoeal specimens during the study period (July

2014). The prevalence of C. difficile identified by routine testing was 7.2% (Table 3.1).

In an earlier study, the prevalence of toxigenic C. difficile among routinely tested

specimens in PathWest was estimated at 7.2% also80. That study was performed from

December 2013 to January 2014 and suggests the prevalence remained constant over

time. It is also a good comparison of the summer and winter seasons, showing no

difference in isolation rates, a change from other early studies that showed distinct

seasonal peaks in summer 2011/12 and summer 2012/1315, 305 .

By screening all diarrhoeal specimens received by PathWest, the number of C. difficile-

positive specimens over July 2014 increased from 49 to 65, but the overall prevalence

of C. difficile identified (6.4%) was lower than that identified by routine testing (Table

3.1). The overall prevalence was similar to that found in an equivalent study performed

in Spain in 2013 (6.0%)317.

Point-prevalence studies in Spain in 2008 and 2013 found that underdiagnosis of CDI

occurred in 66% and 51%, respectively, of diarrhoeal patients317, 318. A multi-country

survey across Europe from September 2011 to August 2013 reported that 23% of CDI

59

cases went undiagnosed41. In the current prevalence study, 24.6% of cases were

undiagnosed, lower than the proportion of Spanish cases, but similar to Europe overall.

While the results of the prevalence study were similar to those found in European

surveys, the demography of cases of undiagnosed CDI has not been described

previously. As discussed above, there is continuing controversy over whether

C. difficile causes infection in children. In the present study, undiagnosed C. difficile-

positive specimens were identified in children only, and 32.3% of all cases overall were

aged <20 y. In prevalence surveys both across Europe and in the Netherlands, 18% of

routinely identified CDI cases were also aged <20 y41, 319, but the age characteristics of

undiagnosed cases were not described, so accurate comparisons cannot be drawn. It

appears that current testing practices in WA are appropriate, given that only paediatric

cases, which may not be true CDI, were undetected by routine testing.

3.3.3. Increasing incidence rates of CDI in Australia

The recommended format for accurate reporting of CDI incidence rates is the number of

new cases per 10,000 pd40. The earliest estimation of CDI incidence in Australia was

reported from a 10 year study at Sir Charles Gairdner Hospital (SCGH) between 1983

and 1992266. The incidence was reported as increasing from 3/10,000 pd to 7/10,000 pd

in 1988 before stabilising again. A later publication showed a significant decrease in

incidence of CDI at SCGH when restrictions were placed on prescribing third-

generation cephalosporins within the hospital in 1998267. At that point the rates declined

to <2/10,000 pd, demonstrating the importance of exposure to cephalosporins as a risk

factor for CDI. After that, conserved practices in cephalosporin prescribing appeared to

influence the stabilised CDI incidence in Australia through the 2000s.

In a nationwide surveillance study in 2010, the crude incidence of CDI in Australia was

calculated at 23.8/100,000 population302. As Victoria (VIC) and the NT, and some

laboratories, were not participating in the study this was probably an under-estimation

of the true incidence. In the USA, a study estimated the national incidence for 2011 at

51.9/100,000 population for CA-CDI and 95.3/100,000 for HCFA-CDI, much greater

figures than those estimated for Australia42. A recent study of all hospital discharges in

the USA between 2005 and 2009 estimated a national incidence of 77.8/10,000 pd, a far

greater rate than that identified in Australia320. The USA, however, has had a long-

standing issue with control of CDI and the Centers for Disease Control (CDC) currently

60

consider C. difficile to be the greatest antibiotic resistant threat to public health10. In

Canada, rates were estimated at 35.6/100,000 in 1991 and increased to 156.3/100,000 in

2003, at the height of the emergence of RT 02712. In Germany, the incidence per

100,000 population was also reported as increasing from 1.7-3.8/100,000 in 2003 to

14.8/100,000 in 2006321, a significant rise but lower rates than estimated for Australia

for 2010. These crude rates give an overall idea that the Australian CDI incidence at that

time was similar to other regions.

The first national report on the incidence of CDI in Australia was published in 2012

when Slimings et al. estimated that the Australian incidence of CDI increased

significantly from 3.2/10,000 pd in 2011 to 4.0/10,000 pd in 201215.

Although few, other published studies have supported the view that there has been an

increase in CDI in Australia, including the prospective study described in section 3.2.2

where it was estimated that for Royal Perth Hospital (RPH) the average incidence was

6.8/10,000 pd, fluctuating from 5.1 to 8.1/10,000 pd305. At SCGH the average incidence

rate was 8.0/10,000 pd, fluctuating from a low of 3.9/10,000 pd in December 2011 to a

high of 16.3/10,000 pd in January 2012, then dropping again to baseline rates for

March, April and May 2012305. This apparent peak in incidence in SCGH did not

correlate with an obvious increase in isolation of particular RTs or in obvious outbreaks

on particular wards, possibly due to low numbers at a hospital. The peak incidence did

however coincide with the emergence of RT 244 strains causing infections Australia-

wide (discussed in more detail in Chapter 5) and anecdotally more RT 244 infections

appear to have occurred on the east coast of Australia. The peak incidence exceeded that

reported for Quebec Province during November 2004 to April 2005, when the incidence

was still high due to the ongoing CDI epidemic.

Another report from Victoria after introduction of routine surveillance of CDI estimated

the incidence at 2.2/10,000 from October 2010 to March 2011. This rate was lower than

that reported nationally and for SCGH but preceded the apparent increase reported after

late 2011272.

Increasing incidence may reflect a true increase in CDI cases, but also may be

confounded by changes in testing practices. Alterations to criteria for testing specimens

for C. difficile may lead to more positive identifications of disease. Also, introduction

into Australia of tcdB PCR for detection of C. difficile, and its subsequent popularity in

61

the late 2000s, would also have led to increased detection of C. difficile due to the

higher sensitivity, as discussed in Chapter 1. Slimings et al argued that while these

factors may have somewhat influenced the increases seen, a true increase in incidence

was apparent15.

3.3.4. Characteristics of CDI patients in Australia

CDI patients in Australia exhibit many characteristics in common with patients from

elsewhere in the world. These characteristics include advanced age, clinical signs of

severe illness, recent hospitalisation, recent antimicrobial use and multiple

comorbidities.

3.3.4.1. Advanced age

All of the studies described in this chapter identified CDI primarily among patients of

advanced age (see Figures 3.1 to 3.6). The median age overall of patients with

C. difficile-positive specimens in section 3.2.1 was 70.8 y (IQR 54.2-75.8, Table 3.2).

Patients in SCGH and RPH in the cross-sectional study in section 3.2.2 had a median

age of 60.5 y (IQR 49.0-71.0, Table 3.3). Nationwide, the median age calculated for the

snapshot study in section 3.2.3 was 66.0 y (IQR 40.3-78.0, Table 3.6) and among the

CDI cases identified in EDs, the median age was 63.7 y (IQR 39.5-77.8, Table 3.8). It is

interesting to note the difference in median age for Australian CA-CDI cases compared

to HCFA-CDI identified in Table 3.6 (57.0 y vs 70.0 y, p<0.05). In particular, patients

aged 30-49 y were significantly more common among CA-CDI cases compared to

HCFA-CDI cases (22.0% vs 10.9%, Table 3.6). In the same study, HCFA-CDI cases

were significantly more frequently aged ≥65 y (60.9% vs 39.0%, p<0.01, Table

3.6).These results agree with reports from elsewhere in the world, where generally

patients with CDI are of advanced age20, 322, but CA-CDI cases are increasingly being

identified among younger people49. This is not surprising since older people are more

likely to be hospitalised and treated for an array of comorbidities that are increasingly

common with advanced age.

3.3.4.2. Clinical signs of severe illness

Severe CDI was identified in studies described here, in various proportions in different

patient populations. The same definition of severe CDI (white cell count [WCC]

>15,000 cells/µL, serum creatinine >1.5 times the normal range) was used across all

62

studies to allow accurate comparison. Severe CDI was described in 20.0% of patients in

SCGH and RPH (Table 3.3). In the nationwide snapshot study, severe CDI was

identified in 18.1% of cases. These findings are similar, and show the cross sectional

study probably reflected the general epidemiology of CDI patients across Australia.

Other clinical signs exhibited in severe CDI cases were ileus (12.6%), loss of appetite

(75.0%), malaise (62.5%), nausea (50.0%) and dehydration (56.3%, Table 3.4).

Interestingly, fever was more commonly seen in non-severe CDI cases (31.3%) in this

study due to the definition of severe CDI used, but other guidelines consider fever a

sign of severity64.

Meanwhile in the ED patient study, 51.0% of patients who were admitted had severe

symptoms (Table 3.9). A small proportion of patients exhibited symptoms of severe

CDI when they presented at the ED but were not admitted for further treatment (5

patients, 6.5%, Table 3.8). The high proportion of severe CDI among patients admitted

via the ED with CDI is surprising. Only a small proportion of these patients (2

patients/27 with severe CDI and admitted, 7.4%) had recurrent CDI, and a small

proportion again (4 patients, 14.8%) had a history of CDI. Meanwhile 14.8% of

admitted patients with severe symptoms had recurrent CDI, and 14.8% had a history of

CDI.

It is difficult to compare severity in studies from other parts of the world and Australia,

because a standard definition of severity is not generally used. Other studies describe

hypotension and shock, fever, admission to ICU, colectomy and other factors as signs of

severity64, 96. For these analyses the standard definition given by the IDSA/SHEA

guidelines were used88.

3.3.4.3. Recurrent CDI

Recurrent CDI was identified in 10.0% of CDI patients in SCGH and RPH (Table 3.3).

Among the cases in the nationwide study, 8.6% had recurrent CDI (Table 3.6). For ED

patients with CDI, 10.4% had recurrent CDI. These numbers were reasonably constant

over the three studies, suggesting a recurrence rate of around 10% in Australia. This is

lower than that reported for European CDI patients (16%)45. Other reports from

elsewhere in the world have estimated higher rates of recurrence ranging from 15-25%,

since the late 1990s323, 324, however, during the RT 027 outbreaks in Canada even higher

recurrence rates, where the 60-day probability of recurrence increased significantly

63

from 20.8% in 1991-2002 to 47.2% in 2003-2004325, were described. There are also

some differences in the way recurrence is defined. Some studies define recurrence as

any episode of CDI occurring in the previous 90 days326. For studies described here,

recurrence was defined as an episode occurring in the 8 weeks prior to current

diagnosis, as recommended by McDonald et al.40 Elsewhere, recurrent CDI has been

associated with increased mortality327, however, this has not been explored in Australia

to date.

Recurrence can occur due to relapse of infection with the same strain or reinfection with

a new strain of C. difficile while the gut microbiota are still compromised. Recurrence

most likely occurs due to poor recovery of gut microbiota following treatment for CDI,

and/or a poor immune response due to advanced age or severity of comorbidities. It

appears that recurrence rates in Australia may be lower than those reported elsewhere,

however, varying definitions in other reports mean accurate comparisons are difficult to

draw.

Few studies of risk factors for CDI have been performed in Australia. Leonard et al.

conducted a small case-control study of diarrhoeal patients in RPH from 2009-2010328

and found that antibiotic use was a significant risk factor. In another study conducted in

Tasmania in 1997, antimicrobial use, severe comorbidity, renal disease, exposure to

antineoplastic agents and nasogastric feeding were all associated with CDI329. These

results mirror characteristics seen elsewhere in the world20, 45, 322 and in Asia (Chapter

5), and the studies described here continue to support these findings.

3.3.4.4. Antimicrobial use

The primary risk factor for CDI, recent antimicrobial use, was observed in the studies

described here. Among SCGH and RPH patients, antibiotics were used in the previous 3

months by 62.5% of patients and by 52.5% prior to onset of CDI during the current

admission (Table 3.3). Among ED patients with CDI, antibiotics were used in the

previous 8 weeks by 45.5% (Table 3.8). These figures seem lower than those reported

overseas where recent antibiotic use was reported among 79% of European CDI

patients45. A higher proportion of recent antibiotic use in the previous 8 weeks was

found in Asian CDI patients (77.6%, Table 4.9). A US study reported antibiotic use in

the previous 90 days among 87% of all CDI patients in Minnesota from 1991 to 200549.

64

Amoxycillin (12.5%) and cephalosporins (11.3%) were the most frequently reported

antibiotics used by CDI patients in SCGH and RPH (Table 3.3). Similarly, among ED

patients, amoxycillin (17.8%) and cephalosporins (11.0%) were the most frequently

used antibiotics in the previous 8 weeks. These are broad-spectrum antibiotics, which

are commonly implicated in increasing risk for CDI, particularly cephalosporins330, 331.

Restrictions in cephalosporin use were associated with a reduction in CDI incidence in

the late 1990s in Western Australia267. The findings reported in this thesis require

further investigation to determine the exact circumstances where cephalosporins were

prescribed, particularly the possibility that the antibiotics were given as an outpatient. In

addition, it is obvious that continuing controls on cephalosporin prescriptions are

required to help control CDI infection rates in Australia.

3.3.4.5. PPI use

PPIs were used significantly more frequently by CDI patients presenting in ED

compared with other diarrhoeal patients (40.3% vs 24.7%, p<0.01, Table 3.8). A

number of studies have reported PPI use as a risk factor for CDI51, 328, however debate

continues over their role, if any, in contributing to CDI52, particularly being associated

with the risk of recurrent disease165, 332. PPIs are among the top 10 most commonly used

drugs in the world333 and are frequently used by people of advanced age with

underlying comorbidities, who are also at higher risk of CDI. Whether the drugs

themselves or the underlying conditions which warrant their use contribute to CDI was

unclear until recently. Imhann et al. reported alteration of the gut microbiome in PPI

users, which suggest a role in contributing to CDI is more plausible333. Moreover, the

lack of association of use of H2RAs, another type of gastric acid suppressants, with CDI

(Table 3.8) suggests a lack of association of gastric acid disorders with CDI, and further

supports a possible association of PPIs with CDI.

3.3.4.6. Hospitalisation and residence in aged care facilities

Hospital inpatients most frequently had CDI in the studies outlined in this chapter. The

prevalence of C. difficile among IPs was 7.2%, making up 49% of all cases (Table 3.1).

HCFA-CDI was identified in 82.5% of patients in SCGH and RPH (Table 3.3) while

58.0% of all 338 cases investigated across Australia were HCFA-CDI, with a large

proportion coming from RACFs. The higher rates present at SCGH and RPH were to be

expected given they are both tertiary referral hospitals dealing with patients with an

65

array of conditions with multiple comorbidities. From the time that CDI was first

described as an entity the majority of cases worldwide have been reported as HCFA-

CDI322, 334.

3.3.4.7. Comorbidities

Comorbidities were common among CDI patients in Australia. The most common

comorbidities among SCGH and RPH CDI patients were GORD (46.3%), hypertension

(42.5%), chronic pulmonary disease (40.0%) and chronic renal disease (36.3%). Cancer

(35.0%) and immune deficiencies (35.0%) were also frequent (Table 3.3).

Comorbidities were less frequent among ED CDI cases as this was an inherently

healthier group of patients. Among CDI cases at ED, the most common comorbidity

was hypertension (35.1%) followed by solid tumour (22.1%), diabetes mellitus (15.6%)

and GORD (14.3%, Table 3.8). Again, comorbidities are frequently reported in CDI

patients worldwide, and the number and severity of comorbidities contributes to

outcomes of disease including complications, recurrence or death20.

3.3.4.8. CDI among females

Overall in the prevalence study, C. difficile-positive specimens were more often from

females (64.6%, Table 3.2). For the nationwide study of CDI cases, females made up

the majority of cases (58.9%, Table 3.6). Females were also significantly more common

among CDI patients at EDs compared with other diarrhoeal patients (66.2% vs 44.2%,

p<0.01, Table 3.8). However, males outnumbered females for the study of SCGH and

RPH cases (52.3%, Table 3.3), a slightly unusual finding. CDI has often been reported

previously as occurring more commonly in females42, 45. Why females may appear to

experience CDI more frequently has not been explored. It could be due to more frequent

visits to medical practitioners for issues such as urinary tract infections that are known

to occur more often in women, and thus women receive antimicrobial treatment more

frequently than men. It could be due to the fact that females tend to live longer than

males, so a greater proportion of older patients are more likely to be female or that

females are more likely to seek medical care. Exposure to young children, who are

frequent carriers of C. difficile, has been implicated previously as a risk factor for CA-

CDI335, and women traditionally have spent a greater proportion of time in close contact

with infants and young children than men due to breastfeeding and early stage

motherhood.

66

Evidence also shows that vertebrate females’ gut microbiota differ from males, even

when they share the same diet336. A study of gut microbiomes of the hunter-gatherer

Hadza people in Tanzania noted a difference in microbiome of females compared to

males, however some differences in diet and divisions of labour between sexes may

have contributed to the difference337. While sex-dependent differences in gut microbiota

have not been explored in depth, this could explain the apparent increased risk among

women for CDI. Further investigation is required to explore this possibility.

3.3.5. CA-CDI in Australia

CA-CDI is increasingly being described internationally42, 49. CA-CDI was identified in

17.5% of cases in the nationwide snapshot study (Table 3.6), lower than that reported by

Slimings et al., where surveillance data from the majority of States and Territories were

aggregated for analysis15. Among the CDI cases in SCGH and RPH, CA-CDI was

identified among 7.5% of patients, while community onset occurred for 28.8% (Table

3.3). Overall 50.8% of cases in the prevalence study were CO-CDI (Table 3.1).

Understandably, a higher proportion of CDI in patients presenting to EDs was CA-CDI

- 44.2% (Table 3.8) possibly indicating that patients are using hospital EDs as surrogate

GPs. CA-CDI is clearly prevalent among community patients and greater vigilance is

required among GPs and community health practices. In addition, hospital EDs need to

be aware that CA-CDI is becoming a bigger issue for them, and that they should modify

their protocols accordingly.

CA-CDI has been investigated in WA previously. In the 1990s, C. difficile was the

second most commonly detected pathogen after Campylobacter spp. among patients

with diarrhoea of community onset and acquisition338. Victorian surveillance data from

October 2010 to March 2011 identified similar proportions of CA-CDI (16%) and CO-

CDI (22%)272. In Tasmania, the rate of CA-CDI increased from 10/100,000 population

(95% CI 7.5-13.2) in 2010 to 17/100,000 in 2011 (14-21.5)273. Slimings et al. reported

that the increase in CDI nationwide in Australia was attributable to a significant

increase in CA-CDI, which was 26% overall15. More data are accumulating to show that

CA-CDI is on the increase, and CO-CDI makes up a substantial proportion of CDI in

Australia.

3.3.5.1. Definition of CA-CDI

67

One problem to emerge when examining data in the literature on CA-CDI is that the

definition of CA-CDI used may be inadequate. The incubation period of C. difficile has

not been examined, but probably may range from a few days to weeks. Patients who

develop diarrhoea after 48 h of hospitalisation are considered as having HCFA-CDI,

whereas a minimum of 12 weeks without hospitalisation is required for CA-CDI. In all

likelihood, many “indeterminate” cases could probably be classed as CA-CDI, and

many cases of CDI in patients occurring within 1-2 weeks of discharge from hospital

could have been acquired in the community. Even where hospitalisation has occurred in

the previous 4 weeks, hospital association could still be unlikely in some cases.

Moreover, HCFA-CDI is applied to cases who frequently visit outpatient clinics for

chemotherapy and other treatment, however they could also have been exposed to

C. difficile in the community. No RTs in the studies described here were significantly

associated with CA-CDI, however CO-CDI may be a better indicator for future studies.

A study in the UK of C. difficile strains over time produced some noteworthy data on

transmission within and outside hospitals. For cases defined as HCFA-CDI, and patients

who were infected with C. difficile of the same RT within 2 weeks of each other, the

hypothesis was that they would be infected with very closely related, if not

indistinguishable strains. However, when whole genome sequencing (WGS) analysis

was performed, a far greater diversity among strains was identified than expected110.

This indicated a much more diverse array of reservoirs of infection than previously

imagined. It is more likely that some of the patients in the study were exposed to

reservoirs in the community prior to hospitalisation rather than within the hospital

environment. These data further support a re-evaluation of definitions of HCFA-CDI

and CA-CDI.

3.3.5.2. CO-CDI in Australia

CO-CDI was identified frequently in the studies outlined in this chapter. In the

prevalence study, 50.8% of all C. difficile-positive specimens were from community

patients. Among routinely identified C. difficile cases, 40.8% were community patients

(Table 3.1) and the prospective study of SCGH and RPH patients identified 46.3% CO-

CDI cases. In the nationwide snapshot study 41.9% of cases were CO-CDI. These data

indicate that almost half of all CDI in Australia has onset in the community. A similar

68

proportion was reported from the USA recently, where over half of CDI cases in a

national surveillance study were CO-CDI42.

The age distribution of CDI patients presenting to the ED, all of whom are considered

CO-CDI, differed from the age distribution seen for CDI patients in general. A bi-modal

distribution was found for ED-presenting CDI patients (Figure 3.6), while in general the

age distribution for CDI patients in Australia is skewed towards older age, as

demonstrated by Figures 3.1, 3.2 and 3.4. When distributions were examined for CO-

CDI patients only in the SCGH and RPH study (Figure 3.3), a bi-modal distribution was

not as evident. Meanwhile for the nationwide study patients aged 40-45 y were the least

frequent age group, but a clear mode of distribution was not as evident (Figure 3.6).

Older age groups were more common, and a peak in children <10 y was likely to be

confounded by some colonised children. It appears that among the ED patients with

CDI, younger patients aged 19-34 y were also presenting with CDI more commonly

than other middle-aged patients (Figure 3.6).

The studies showed that severe CDI does occur in the community in WA. A Canadian

study of ED-presenting CDI patients found that older patients and patients

comorbidities or with recurrent disease were more likely to experience severe

symptoms, extended hospital stay, colectomy or ICU admission, or death339. A study in

Taipei found that CDI patients in an ED had a wide age range and cases more often

occurred in warm springs and summer seasons202. In the study outlined here, cases also

peaked in the summer months.

Five ED patients with severe CDI and a further 18 with non-severe CDI were

discharged home prior to receiving the result of their diagnostic test. Hospital inpatients

with CDI are routinely isolated and contact precautions are initiated upon a positive

diagnosis. Discharged, untreated patients could potentially further spread strains among

the community before receiving treatment. Rapid admission, isolation and treatment of

patients could reduce the burden of CA-CDI.

3.3.5.3. Possible community sources of C. difficile in Australia

69

Given the high proportion of CO-CDI and CA-CDI cases found, community sources of

C. difficile need to be identified. One possible source is food animals and their meat

products. C. difficile has been detected among Australian production mammals

including pigs274, horses340, sheep54 and cattle56. Many studies overseas have also

identified C. difficile in meat products and fresh vegetables62, 341-343. Recently C. difficile

was isolated for the first time from Australian vegetables (SC Lim, personal

communication). While there may be some contamination of Australian meat products,

the highest rates of colonisation/infection with C. difficile in animals occur in very

young animals that (apart from neonatal dairy calves in Australia) are not destined for

slaughter55, 56. Thus contamination of vegetables grown in soil containing animal

manure is one more likely explanation. There may of course be others.

In the Northern Hemisphere, the predominant animal C. difficile strain is RT 078,

frequently identified in pigs and cattle344. RT 078 has been identified in increasing rates

of CA-CDI in Europe8, 345 suggesting either a direct zoonosis or, more likely, a transfer

to the food supply. In Australia, RT 078 has not been identified in animals, and rarely

causes disease in humans. In cases of CDI caused by RT 078 in Australia, a potential

source of infection could be imported food products. Further investigations are required

to explore possible intercontinental transmission of C. difficile via imported and

exported foodstuffs.

A potential source of food-borne CDI within Australia is veal sourced from neonatal

dairy calves, since a prevalence of 56% C. difficile was identified in veal calves at

slaughter56. Some C. difficile strains isolated from cattle (RTs 126, 127, 033 and 056)

have also caused human CDI in Australia recently56. There is further discussion of

animal and human strains of C. difficile in Australia in Chapter 5.

Another potential source of transmission within the community is recently discharged

hospital patients. As demonstrated in the study of ED patients, a number of patients

were allowed to return home, while exhibiting clinical signs of severe CDI. These

patients could shed spores in their environment which could infect others at risk for

CDI. The high rates of C. difficile isolation among children in the studies described here

also indicate that children may spread C. difficile among the community. The higher

rates of CDI among younger adults found in the ED patients could be explained by

young parents being exposed to C. difficile spores shed by their infant children. Young

70

children have been implicated as possible sources of C. difficile transmission

previously335, however clear risk factors for exposure to children have not been

described frequently.

3.3.5.4. Strengths of the studies outlined here

The prevalence study provided an accurate estimate of the overall prevalence of

C. difficile in diarrhoeal patients in WA, and identified what patient populations could

be targeted to increase testing. No similar studies had previously been performed in

Australia. The results can be used to indicate to GPs the requirement for vigilance for

C. difficile among community patients, and among paediatric diarrhoeal patients.

The cross-sectional study mostly mirrored the results of the nationwide snapshot study,

so most of the findings could be applied to Australian population in general. However

some characteristics differed, like CA-CDI and CO-CDI. This is likely due to the

patient populations of SCGH and RPH, tertiary hospitals dealing with more severely ill

people with issues like malignancies, chronic kidney diseases and other comorbidities.

The nationwide snapshot was important for providing an overview of the whole country

at one point in time. It was performed close to when a peak in CDI was seen in 2011.

While it did not provide much detail, it gave an important idea of the basic

demographics and identified severe and complicated cases of CDI. It also provided

baseline data on HCFA-CDI, CA-CDI, and CO-CDI and generated a large collection of

isolates for further investigation.

3.3.5.5. Limitations of studies outlined here

All studies diagnosed CDI by positive tcdB PCR using BD GeneOhm or BD MAX or

toxigenic culture. These molecular assays may overdiagnose CDI, as they may identify

colonised patients. However all patients included in analyses had clinical diarrhoea,

which in the light of their C. difficile assay result, are likely to have true CDI.

The prospective study in SCGH and RPH was limited by the requirement to recruit

patients to provide consent for interview and medical record review. This meant that a

substantial number of CDI patients were not approached due to the severity of their

illness. That led to probable bias in the results, showing less severe disease overall. The

71

actual proportions of CA-CDI and HCFA-CDI may also have differed substantially if

all CDI patients over the study period had been included in the study.

The ED study had a number of limitations. Patients who presented at EDs may have

differed from other community patients with diarrhoea. They may have been

experiencing more severe symptoms than patients who would visit their GP or not seek

medical assistance, which could have biased the results. Patients may have been

accessing EDs for free medical care or out-of-hours care.

Another limitation of the ED study was the requirement for a stool sample from ED

patients. All patients were asked to provide a stool sample during the study period.

Many patients could not or chose not to provide a sample during their visit. Those

experiencing more severe symptoms may have been more likely to provide a sample

during their visit, which may have introduced a bias in our results regarding severity.

Patients who were recently hospitalised may have been more likely to seek care than

those who assumed they had eaten contaminated food or had a mild case of self-limiting

diarrhoea.

Despite limitations the ED study highlighted that patients in the community are

becoming ill with CDI caused by strains which can cause severe disease. The sources of

these strains need to be investigated in further detail.

3.3.5.6. Implications and recommendations

The findings in this chapter show that CDI is frequently occurring in community

patients. This requires greater vigilance for possible CDI cases in physicians practising

in outpatient clinics, EDs, and general practice clinics. Testing methods appear to be

generally adequate to detect CDI patients in Australia, however a more rapid test would

allow identification of community CDI cases on the spot, allowing action to be taken

more quickly to admit the patient for further treatment, to administer appropriate

antibiotics, and to advise them on how to minimise spreading spores in their

surroundings.

The fluctuations and increases in CDI rates over recent years have highlighted a need

for ongoing surveillance of CDI in Australia to measure trends and identify outbreaks.

Currently, mandatory reporting of CDI cases is required by hospitals for the Australian

Commission on Safety and Quality in Healthcare. This system must continue with up to

72

date reporting of rates to healthcare services to ensure swift action is taken to control

outbreaks if and when they appear.

3.3.5.7. Future studies

Future studies of CDI in Australia should continue to monitor overall trends and the

molecular epidemiology of C. difficile to identify outbreaks early and allow control

measures to be introduced quickly. In particular, the epidemiology of CO-CDI and CA-

CDI in Australia require further investigation.

Epidemiological studies of patients attending GP and OP clinics would fill in some

“missing data” regarding community-based patients with CDI. Collection of data on

recent travel and diet, exposure to children and recently hospitalised patients could

generate hypotheses for sources of CA-CDI. These hypotheses could be further tested in

case-control studies to identify risk factors for CA-CDI in Australia.

Studies of foodstuffs including meat and vegetables, both imported and locally

produced, for isolation of C. difficile, could pinpoint plausible sources of CDI in the

community. Surveys to detect C. difficile in other environmental sources including

gardening materials, water bodies and vegetable markets could also identify potential

sources of infection. C. difficile isolates arising from vegetables grown in Australia

were analysed with WGS and compared with some human isolates of the same RT from

Australia and Asia (discussed further in Chapter 5). Further studies of this nature could

confirm some community sources of infection. Furthermore, WGS studies could be

applied to Australian C. difficile isolates from healthcare institutions to determine

transmission within or outside hospitals, as previously demonstrated by Eyre et al110.

These studies could provide interesting insights to the sources and transmission of

C. difficile within Australia.

3.3.5.8. Chapter summary

The major findings of the chapter can be summarised as follows:

• C. difficile testing in Australia appeared to be adequate to detect adult CDI

patients, however the use of NAAT may overdiagnose CDI.

• The overall prevalence of C. difficile among Australian diarrhoeal patients was

6.4%.

73

• A significant proportion of C. difficile-positive specimens (24.6%) were

“undiagnosed”, although they were from paediatric patients where consideration

of CDI is controversial.

• The prevalence of CA-CDI among Australian CDI patients varied from 7.5% to

44.2% depending on patient populations studied.

• Australian CDI patients exhibit characteristics in common with CDI cases

elsewhere in the world - recent nasogastric feeding, advanced age, recent

antimicrobial exposure.

74

CHAPTER 4. PREVALENCE OF C. DIFFICILE

INFECTION IN INDONESIA AND MALAYSIA

75

4.1. INTRODUCTION

Limited surveillance for CDI has been performed in Asia, particularly southeast Asia307,

despite diarrhoea being a significant cause of morbidity and mortality in Asian

countries. A survey published in 2012 found that awareness of C. difficile infection CDI

in physicians in Asia is poor, with underestimation of its contribution to antibiotic-

associated disease and recurrence of diarrhoea137. A study from the Philippines reported

misdiagnosis of CDI as amoebiasis in colitis patients222.

Another important issue relates to the methods employed for laboratory diagnosis of

CDI. Some studies have reported using enzyme immunoassays (EIAs) for toxin A to

diagnose CDI in Asian countries138, 227, 232, 250. However, the predominant strain of

C. difficile circulating in Asia at the time was ribotype (RT) 017 which is A-B+307,

meaning that testing for toxin A would result in false negatives and under-reporting.

Finally, often other diarrhoeal pathogens are more commonly tested for, and limited

laboratory resources mean C. difficile testing is often completely lacking in the Asian

region, leading to an underestimation of its prevalence137.

There have been no published studies on the prevalence of C. difficile in Indonesia and

few from Malaysia. One study among children with acute diarrhoea in Jakarta described

C. difficile among 1.3% of 154 specimens tested by toxin A EIA138. Given that an

inappropriate test was used, it is likely that the prevalence of C. difficile was higher than

identified in this study.

A prevalence study was reported in 2013 from Kota Bharu, in the Kelantan Province of

Malaysia, where the prevalence of CDI was described as 13.7% among 175 inpatients

with diarrhoea, 105 of whom had antibiotic-associated diarrhoea (AAD). They

diagnosed CDI using a toxin A/B EIA. The majority of CDI was identified among

females (54.2%) and patients were mainly of advanced age236. CDI patients with AAD

had used antibiotics within the previous 6 weeks more frequently than other AAD

patients. One other study, also in Kota Bharu, described CDI among type 2 diabetes

mellitus patients. CDI was identified by toxin A/B EIA in 13/159 patients, and

carbapenem and metronidazole use were associated with disease237. Interestingly, 12/13

(92%) patients tested positive for toxin A, and only one was negative for toxin A,

indicating a lower prevalence of A-B+ strains among this study group than in general in

Malaysia.

76

It is likely that C. difficile is highly prevalent in patients in Indonesia and Malaysia.

Reports of inappropriate antibiotic prescribing346 and free access to antibiotics without

prescriptions suggest C. difficile and other antimicrobial-resistant organisms may be

widespread there.

The lack of information on CDI in a population that appears to be at high risk for CDI

warranted further investigation. This chapter describes two studies performed in

Malaysia and Indonesia. In Indonesia, four hospitals established testing to estimate the

prevalence of C. difficile among diarrhoeal patients and, in Malaysia, a similar

prevalence study was performed in three hospitals.


The aims of this chapter were:

• To improve laboratory capacity to test for C. difficile in diarrhoeal patients in

hospitals in Malaysia and Indonesia by introducing a rapid easy-to-use lateral

flow EIA.

• To estimate the prevalence of CDI among diarrhoeal patients in Malaysia and

Indonesia.

• To describe disease characteristics and treatment of CDI patients in Malaysia.

To describe outcomes for CDI patients in Malaysia.

4.2. RESULTS

4.2.1. Indonesia

In total, 340 samples were tested in Indonesia, of which 23 (6.8%) were identified as

glutamate dehydrogenase (GDH)-positive, toxin-positive C. difficile using QUIK

CHEK COMPLETE™ (Table 4.1). A further 47 (13.8%) specimens were GDH-

positive/toxin-negative (Table 4.1).

C. difficile was not recovered by culture from four specimens, three of which were

toxin-positive. Eight specimens yielded two unique isolates, resulting in isolation of 74

unique C. difficile strains, six of which were recovered by enrichment only. Among

these, 37 unique strains were tcdB-positive, two of which were isolated from the same

specimen. The remaining seven specimens, each of which yielded two isolates, each

Table 4.1. Results of EIA and culture analysis in Indonesian inpatients.

EIA result Culture result

Sitea n (%)

1

(N=148)

2

(N=98)

3

(N=49)

4

(N=45)

Total

(N=340)

GDH positive/Toxin positive

Culture positive [tcdB positive] 13 (8.8) [13 (8.8)] 2 (2.0) [2 (2.0)] 3 (6.1) [3 (6.1) 2 (4.4) [2 (4.4)] 20 (5.9) [20 (5.9)]

Culture negative 1 (0.7) 2 (2.0) 0 0 3 (0.9)

GDH positive/Toxin negative

Culture positive [tcdB positive] 21 (14.2) [12 (8.1)] 10 (10.2) [0] 11 (22.4) [3 (6.1)] 4 (8.9) [1 (2.2)] 46 (13.5) [16 (4.7)]

Culture negative 1 (0.7) 0 0 0 1 (0.3)

GDH negative 112 (75.7) 84 (85.7) 35 (71.4) 39 (86.7) 271 (79.5)

CDI prevalence* 26 (17.6) 4 (4.1) 6 (12.2) 3 (6.7) 39 (11.5)

a Sites: 1 Kariadi Hospital, 2 Kartini Hospital, 3 Margono Hospital, 4 Ketileng Hospital

* CDI was defined as any sample toxin-positive by EIA, or culture positive/tcdB positive.

EIA enzyme immunoassay; GDH glutamate dehydrogenase; CDI C. difficile infection

77

contained one toxigenic and one non-toxigenic strain. The overall prevalence of CDI

was 11.5%.

4.2.2. Malaysia

In total 286 diarrhoeal samples were collected from inpatients across all sites, of which

21 (7.3%) were GDH- and toxin-positive, and a further 53 (18.5%) were GDH-positive

and toxin-negative (Table 4.2). Fifteen GDH-negative samples were positive by culture.

Culture yielded 82 isolates from 77 (26.9%) positive specimens; five specimens yielded

two distinct isolates. Isolates were recovered by enrichment only from 11 specimens.

C. difficile was not recovered by culture from 12 GDH-positive specimens, only one of

which was toxin-positive by EIA. tcdB was detected in 41 isolates giving a prevalence

of 50.0% for toxigenic strains among all isolates. The overall prevalence of CDI was

15.4%, varying from 6.7% in Hospital Sungai Buloh to 21.2% in Hospital Universiti

Kebangsaan Malaysia. The sensitivity and specificity of QUIK CHEK COMPLETE

were 44% and 99%, respectively (Table 4.2).

4.2.3. Descriptive epidemiology in Indonesia

Basic data were provided for 70 GDH-positive patients only. Among 39 CDI cases,

53.8% were female and the median age was 50.0 y (interquartile range [IQR] 37.0-

58.0). Abdominal pain was present in 38.5%, blood in stool in 46.2% and fever in

48.7%. Characteristics did not differ significantly between hospital onset (HO)-CDI

patients (28, 71.8%) and community onset (CO)-CDI patients (11, 28.2%) apart from

abdominal pain which was more frequent among CO-CDI patients (63.6% vs 28.6%,

p<0.05, Table 4.3).

4.2.4. Descriptive epidemiology in Malaysia

Demographic and clinical data were received for 76 patients in Hospital Universiti Sains

Malaysia collected from July - September 2015 only.

4.2.4.1. Comparison of C. difficile-positive patients with C. difficile-negative

patients

Malaysian inpatients with any C. difficile strain identified either by positive EIA or

positive culture had significantly longer length of stay compared to other inpatients

Table 4.2. Results of EIA and culture analysis in Malaysian inpatients.

EIA result Culture result

Sitea n (%)

Total

(N=286)

1

(N=142)

2

(N=99)

3

(N=45)

GDH positive/Toxin positive Culture positive [tcdB positive] 9 (6.3) [7 (4.9)] 11 (11.1) [11 (11.1)] 0 20 (7.0) [18 (6.3)]

Culture negative 1 (0.7) 0 0 1 (0.3)

GDH positive/Toxin negative

Culture positive [tcdB positive] 21 (14.8) [5 (3.5)] 15 (15.1) [9 (9.1)] 6 (13.3) [3 (6.7)] 42 (14.7) [17 (5.9)]

Culture negative 5 (3.5) 5 (5.1) 1 (2.2) 11 (3.8)

GDH negative

Culture positive [tcdB positive] 11 (7.7) [5 (3.5)] 4 (4.0) [1 (1.0)] 0 15 (5.2) [6 (2.1)]

Culture negative 95 (66.9) 64 (64.6) 38 (84.4) 197 (68.9)

CDI prevalence* 20 (14.1) 21 (21.2) 3 (6.7) 44 (15.4)

Sensitivity [95% CI] 0.41 [0.19-0.67] 0.53 [0.30-0.74] 0.00 [0.00-0.19] 0.44 [0.29-0.60]

Specificity [95% CI] 0.98 [0.93-0.99] 1.00 [0.94-1.00] 1.00 [0.90-1.00] 0.99 [0.96-0.99]

aSites: 1 Hospital Universiti Sains Malaysia, 2 Hospital Universiti Kebangsaan Malaysia, 3 Hospital Sungai Buloh

* CDI was defined as any sample toxin-positive by EIA, or culture positive/tcdB positive.

EIA enzyme immunoassay; GDH glutamate dehydrogenase; CDI C. difficile infection; CI confidence interval

Table 4.3. Demographic and clinical characteristics of Indonesian inpatients with CDI.

Characteristic n (%) p Total

HO-CDI

(N=28)

CO-CDI

(N=11)

(N=39)

Female 15 (53.6) 6 (54.5) 0.96 21 (53.8)

Median age [IQR] 50.0 [41.3-57.3] 40.0 [29.0-65.0] 0.44 50.0 [37.0-58.0]

Abdominal pain 8 (28.6) 7 (63.6) 0.04 15 (38.5)

Bloody diarrhoea 11 (39.3) 7 (63.6) 0.17 18 (46.2)

Fever 12 (42.9) 7 (63.6) 0.24 19 (48.7)

IQR interquartile range; HO hospital onset; CDI C. difficile infection; CO community onset

78

(11.0 vs 5.0 days, p<0.01). C. difficile-positive patients were of a higher median age

than other patients (58.0 vs 35.0 years, p=0.05). A high proportion (87.0%) had a recent

admission while fewer C. difficile-negative patients had (26.4%, p<0.01). Less than half

(43.5%) of C. difficile-positive patients cases were female. The majority (95.7%) of

C. difficile-positive patients had recently used antibiotics, most frequently penicillin

(82.6%), followed by cephalosporins (52.2%). Meanwhile, significantly fewer

C. difficile-negative patients had used antibiotics (30.2%, p<0.001), including

penicillins (17.0%, p<0.001) and cephalosporins (18.9%, p<0.01). Proton pump

inhibitors (PPIs) were also significantly more frequently used among C. difficile-

positive patients (65.2%) compared to C. difficile-negative patients (28.3%, p<0.01,

Table 4.4).

4.2.4.2. Comparison of CDI cases with non-CDI cases

Characteristics of CDI cases compared with non-CDI cases are shown in Table 4.6. CDI

cases had significantly longer length of stay (LOS) compared to other inpatients (12.5

vs 6.0 days, p<0.01) and 40.0% of cases were female. A higher proportion of CDI cases

(80.0%) had a recent admission than non-CDI cases (39.4%, p<0.05). All (100.0%) CDI

cases had recently used antibiotics. The most frequently used antibiotic among CDI

cases was penicillin (80.0%), followed by cephalosporins (70.0%). Fewer non-CDI

cases had used antibiotics (42.4%, p<0.01), including penicillins (30.3%, p<0.01) and

cephalosporins (22.7%, p<0.01). Other antibiotics which were more frequently used

among CDI cases were macrolides (20.0% vs 3.0%, p<0.01) and metronidazole (20.0%

vs 3.0%, p<0.05). PPIs were also more commonly used among CDI cases (70.0%)

compared to non-CDI cases (34.8%, p<0.01).

Nasogastric feeding in the past 30 days was more frequent among CDI cases (40.0%)

than non-CDI cases (3.8%, p<0.01). Diabetes was the most common comorbidity

among CDI cases (40.0%). No individual comorbidities were significantly more

frequent among CDI cases. Having two or more comorbidities was more frequent

among CDI cases (60.0% vs 25.8%, p<0.05). Abdominal pain (40.0% vs 18.2%),

elevated white cell count (WCC, 20.0% vs 6.1%) and elevated serum creatinine (20.0%

vs 13.6%) were more common among non-CDI cases compared to CDI cases, but fever

was more frequent among CDI cases (90.0% vs 50.0%, p<0.05). Outcome of death

(30.0% vs 6.1%, p<0.05) was significantly more frequent among CDI cases. Meanwhile


Bahru.

Characteristic Any C. difficile detected n (%) Total

Yes (N=23) No (N=53) p (N=76)

Median age (years [IQR]) 58.0 [35.0-63.0] 35.0 [27.5-

59.5]

0.05 46.0 [28.0-61.0]

Female 10 (43.5) 28 (52.8) 0.45 38 (50.0)

Median length of stay (days

[IQR])

11.0 [8.0-18.0] 5.0 [5.0-7.0] <0.001 6.5 [5.0-11.0]

Previous admission 20 (87.0) 14 (26.4) <0.001 34 (44.7)

Recent medication

Any antibiotic use 22 (95.7) 16 (30.2) <0.001 38 (50.0)

Cephalosporin 12 (52.2) 10 (18.9) 0.003 22 (28.9)

Penicillin 19 (82.6) 9 (17.0) <0.001 28 (36.8)

Macrolide 3 (13.0) 1 (1.9) 0.05 4 (5.3)

Carbapenem 7 (30.4) 4 (7.5) 0.01 11 (14.5)

Aminoglycoside 0 1 (1.9) 0.51 1 (1.3)

Vancomycin 1 (4.3) 0 0.13 1 (1.3)

Metronidazole 2 (8.7) 1 (1.9) 0.16 3 (3.9)

Fluoroquinolone 2 (8.7) 1 (1.9) 0.16 3 (3.9)

Proton pump inhibitor 15 (65.2) 15 (28.3) 0.002 30 (39.5)

Chemotherapy 5 (21.7) 8 (15.1) 0.48 13 (17.1)

Medical history

Gastrointestinal illness 0 0

Nasogastric tube 6 (26.1) 2 (3.8) 0.004 8 (10.5)

Diabetes mellitus 8 (34.8) 14 (26.4) 0.46 22 (18.9)

Solid tumour 1 (4.3) 2 (3.8) 0.91 3 (3.9)

Haematological malignancy 6 (26.1) 6 (11.3) 0.11 12 (15.8)

Ischaemic heart disease 8 (34.8) 8 (15.1) 0.05 16 (21.1)

Chronic kidney disease 6 (26.1) 5 (9.4) 0.06 11 (14.5)

Cerebrovascular disease 1 (4.3) 2 (3.8) 0.91 3 (3.9)

COPD 2 (8.7) 1 (1.9) 0.16 3 (3.9)



Diarrhoea 16 (69.6) 47 (88.7) 0.04 63 (82.9)

Abdominal pain 8 (34.8) 8 (15.1) 0.05 16 (21.1)

Fever 17 (73.9) 25 (47.2) 0.03 42 (55.3)

Elevated white cell count 4 (17.4) 2 (3.8) 0.04 6 (7.9)

Elevated serum creatinine 7 (30.4) 4 (7.5) 0.009 11 (14.5)

Blood in stool 2 (8.7) 6 (11.3) 0.73 8 (10.5)

Outcome

Sepsis 3 (13.0) 0 0.03 3 (3.9)

Death 5 (21.7) 2 (3.8) 0.03 7 (9.2)

Discharge well 15 (65.2) 51 (96.2) <0.001 66 (86.8)

IQR interquartile range; COPD chronic obstructive pulmonary disease

79

a discharge after resolution of illness was significantly more frequent among non-CDI

cases (90.9% vs 60.0%, p<0.01, Table 4.5).

4.2.4.3. Comparison of C. difficile cases with toxigenic vs non-toxigenic strains

To determine whether patients with non-toxigenic strains exhibited different

characteristics than CDI cases with toxigenic strains, a comparison for these groups was

made. No significant differences were observed for various characteristics. The median

LOS for patients with toxigenic strain infection was close to significantly longer than

for patients with non-toxigenic strains (16.0 vs 10.0 days, p=0.06). Meanwhile other

characteristics did not differ between patients (Table 4.6).

4.3. DISCUSSION

4.3.1. C. difficile testing in Malaysia and Indonesia

The hospitals involved in the studies described here did not previously perform routine

testing for C. difficile. Testing among diarrhoeal patients identified a high prevalence of

CDI. This demonstrates that C. difficile has probably been ubiquitous among Indonesian

and Malaysian hospitals for a long time, so testing is necessary in these hospitals to

identify CDI patients.

One of the aims of this project was to introduce a simple test for CDI that could be used

in a developing country with minimal training. It was important to select an appropriate

assay for C. difficile testing in Malaysia and Indonesia. RT 017 is the predominant

strain in neighbouring countries Singapore219 and Thailand220, and this strain was

identified as the most prevalent strain in Asia in the multi-country study of CDI

(Chapter 5). In addition, a high prevalence of non-toxigenic strains was also identified

in these studies. These factors require consideration when selecting a suitable diagnostic

assay for local use. Of the current diagnostic assays available, an EIA for GDH and

toxin A/B is a viable option to introduce for C. difficile testing in these countries, given

their low cost and requirement for few laboratory facilities.

The QUIK CHEK COMPLETE EIA had a low sensitivity of 44% and high specificity

of 99% in Malaysia for detection of C. difficile. While the low sensitivity of the test is

not ideal, a polymerase chain reaction (PCR) assay for tcdB may be too costly and

require equipment which is currently not available in some participating hospitals.


Bahru, comparing CDI cases with non-CDI cases.

Characteristic n (%)

CDI (N=10) No CDI (N=66) p

Median age (years [IQR]) 55.0 [34.3-64.8] 45.5 [28.0-61.0] 0.35

Female 4 (40.0) 34 (51.5) 0.50

Median length of stay (days [IQR]) 12.5 [9.5-18.8] 6.0 [5.0-10.0] 0.001

Previous admission 8 (80.0) 26 (39.4) 0.02

Recent medication

Any antibiotic use 10 (100.0) 28 (42.4) 0.001

Cephalosporin 7 (70.0) 15 (22.7) 0.002

Penicillin 8 (80.0) 20 (30.3) 0.002

Macrolide 2 (20.0) 2 (3.0) 0.03

Carbapenem 3 (30.0) 8 (12.1) 0.13

Aminoglycoside 0 1 (1.5) 0.70

Vancomycin 0 1 (1.5) 0.70

Metronidazole 2 (20.0) 1 (1.5) 0.01

Fluoroquinolone 0 3 (4.5) 0.49

Proton pump inhibitor 7 (70.0) 23 (34.8) 0.03

Chemotherapy 1 (10.0) 12 (18.2) 0.52

Medical history

Gastrointestinal illness 0 0

Nasogastric tube 4 (40.0) 4 (6.1) 0.001

Diabetes mellitus 4 (40.0) 18 (27.3) 0.41

Solid tumour 1 (10.0) 2 (3.0) 0.29

Haematological malignancy 2 (20.0) 10 (15.2) 0.70

Ischaemic heart disease 2 (20.0) 14 (21.2) 0.93

Chronic kidney disease 2 (20.0) 9 (13.6) 0.59

Cerebrovascular disease 1 (10.0) 2 (3.0) 0.29

Chronic pulmonary disease 1 (10.0) 2 (3.0) 0.29

≥2 comorbidities 6 (60.0) 17 (25.8) 0.03



Diarrhoea 7 (70.0) 56 (84.8) 0.25

Abdominal pain 4 (40.0) 12 (18.2) 0.12

Fever 9 (90.0) 33 (50.0) 0.02

Elevated white cell count 2 (20.0) 4 (6.1) 0.13

Elevated serum creatinine 2 (20.0) 9 (13.6) 0.59

Severe CDI 4 (40.0) N/A

Blood in stool 1 (10.0) 7 (10.6) 0.95

Outcome

Sepsis 1 (10.0) 2 (3.0) 0.29

Death 3 (30.0) 4 (6.1) 0.02

Discharge well 6 (60.0) 60 (90.9) 0.007

IQR interquartile range; CDI C. difficile infection


Bahru with toxigenic C. difficile vs non-toxigenic C. difficile strains.

Characteristic n (%) Total

Toxigenic strain

(N=7)

Non-toxigenic

strain (N=12)

p

Median age (years [IQR]) 52.0 [35.0-63.0] 60.0 [52.3-66.8] 0.31 58.0 [35.0-

63.0]

Female 3 (42.9) 5 (41.7) 0.96 8 (42.1)

Median length of stay (days

[IQR])

16.0 [11.0-21.0] 10.0 [7.3-14.0] 0.06

Previous admission 6 (85.7) 11 (91.7) 0.68 17 (89.5)

Recent medication

Any antibiotic use 7 (100.0) 11 (91.7) 0.43 18 (94.7)

Cephalosporin 6 (85.7) 5 (41.7) 0.06 11 (57.9)

Penicillin 6 (85.7) 9 (75.0) 0.58 15 (78.9)

Macrolide 0 (0.0) 2 (16.7) 0.25 2 (10.5)

Carbapenem 2 (28.6) 2 (16.7) 0.54 4 (21.1)

Aminoglycoside 0 0 0

Vancomycin 0 1 (8.3) 0.43 1 (5.3)

Metronidazole 1 (14.3) 1 (8.3) 0.68 2 (10.5)

Fluoroquinolone 0 1 (8.3) 0.43 1 (5.3)

Proton pump inhibitor 6 (85.7) 6 (50.0) 0.12 12 (63.2)

Chemotherapy 1 (14.3) 3 (25.0) 0.58 4 (21.1)

Medical history

GI illness 0 0

Nasogastric tube 3 (42.9) 2 (16.7) 0.21 5 (26.3)

Diabetes mellitus 4 (57.1) 3 (25.0) 0.16 7 (36.8)

Solid tumour 1 (14.3) 0 0.18 1 (5.3)

Haematological malignancy 2 (28.6) 3 (25.0) 0.87 5 (26.3)

Ischaemic heart disease 2 (28.6) 5 (41.7) 0.57 7 (36.8)

Chronic kidney disease 2 (28.6) 3 (25.0) 0.87 5 (26.3)

Cerebrovascular disease 1 (14.3) 0 0.18 1 (5.3)

Chronic pulmonary disease 1 (14.3) 1 (8.3) 0.68 2 (10.5)



Diarrhoea 6 (85.7) 8 (66.7) 0.36 14 (73.7)

Abdominal pain 3 (42.9) 4 (33.3) 0.68 7 (36.8)

Fever 6 (85.7) 7 (58.3) 0.22 13 (68.4)

Elevated white cell count 0 3 (25.0) 0.15 3 (15.8)

Elevated serum creatinine 2 (28.6) 3 (25.0) 0.87 5 (26.3)

Blood in stool 1 (14.3) 1 (8.3) 0.68 2 (10.5)

Outcome

Sepsis 0 2 (16.7) 0.51 2 (10.5)

Death 2 (28.6) 1 (8.3) 0.52 3 (15.8)

Discharge well 5 (71.4) 9 (75.0) 0.99 14 (73.7)

IQR interquartile range

80

Culture for diagnostic purposes or otherwise is also not possible due to the lack of

anaerobic culture facilities. Moreover, the high prevalence of non-toxigenic strains

means culture could lead to overdiagnosis unless further testing was undertaken such as

tcdB PCR being performed on isolates.

The high specificity of QUIK CHEK COMPLETE could be advantageous to rule in

patients with CDI. The EIA identified toxin in less than half of the patients from whom

C. difficile was isolated. GDH was identified in most CDI patients, but the high

prevalence of non-toxigenic strains means GDH-positive results must be treated with

caution. Differential diagnosis of CDI could be made by taking into account results of

other assays and the history of the patient, e.g. recent penicillin or cephalosporin use

and advanced age. Given the current high prevalence of CDI among diarrhoeal patients

in these hospitals, and the limited resources available locally for testing, the continued

use of a low-cost EIA identifying both GDH and toxin would appear to be a beneficial

practice, but could underdiagnose CDI among patients.

4.3.2. Prevalence of C. difficile in Malaysia and Indonesia

The prevalence of C. difficile was 11.5% in Indonesia and 15.4% in Malaysia (p=0.15).

While the prevalence was higher in Malaysia, the prevalence in Indonesia may have

been underestimated, since GDH-negative samples were not cultured. A number of

these may have been tcdB-positive. The prevalence of C. difficile in Malaysia and

Indonesia was high compared with studies from Singapore (7-11% toxin-positive)241

and Australia (7.2%)80, and the Western Australia (WA) prevalence study described in

chapter 3 (6.4%-7.2%). A study in China reported a similar high prevalence of toxigenic

C. difficile (19.2%) among diarrhoeal inpatients186. Meanwhile in Hong Kong, the

prevalence of C. difficile among diarrhoeal specimens was 5.1% from 2007-2008, using

cell culture cytotoxicity assay (CCCA)212, and 13.3% in 2009 by CCCA and culture216.

Previously in Malaysia, in a study at Hospital Universiti Sains Malaysia, the prevalence

of C. difficile toxin was 13.7% using toxin A/B EIA among 175 diarrhoeal inpatients. In

our study, the overall prevalence at that hospital was 14.1%, but only 7.0% were

positive by toxin EIA. This indicates that the actual prevalence of CDI could have been

higher than that identified in the 2013 study.

The high prevalence of C. difficile in Indonesia and Malaysia is cause for concern. They

are developing countries with high rates of antibiotic usage and mis-usage347, 348, and a

81

high burden of diarrhoeal diseases349, 350. Limited resources mean C. difficile testing is

not currently adequate to detect cases, and anaerobic culture facilities for further

examination including molecular analysis are rare.

4.3.3. Characteristics of CDI patients in Malaysia

The characteristics of patients with CDI in Malaysia were similar to those described

elsewhere in the world.

4.3.3.1. Age

Older median age was not associated with CDI (Table 4.5), but was more common

among C. difficile-positive patients compared to C. difficile-negative patients (median

58.0 y vs 35.0 y, Table 4.4). This is interesting to note, as advanced age is considered an

important risk factor, in addition to antimicrobial use, for CDI 322. The fact that older

aged patients were more likely to be positive for any C. difficile strain (Table 4.4)

indicates that non-toxigenic strains are transmitting in the same manner as toxigenic

strains, possibly within or outside the hospital environment, and that older patients are

the most likely to be colonised or infected with C. difficile in general. Toxigenic strains

caused disease in any age group, since there was no difference in age noted for either

toxigenic vs non-toxigenic carrying cases, or for CDI cases compared with non-CDI

cases.

4.3.3.2. Recent hospitalisation

A previous hospital admission, a typical risk factor for CDI seen worldwide20, was

significantly more common among CDI patients (80.0% vs 39.4%, p< 0.05, Table 4.5)

and among patients carrying any C. difficile strain (87.0% vs 26.4%, p<0.01, Table 4.4)

than in C. difficile-negative patients. Again, this suggests that strains of C. difficile were

transmitted in the same manner, and most C. difficile-positive patients probably were

exposed to C. difficile during their previous hospital admission.

4.3.3.3. Extended length of stay

The median LOS for CDI patients was significantly greater than for non-CDI patients

(12.5 days vs 6.0 days, p<0.01, Table 4.5). For all C. difficile-positive patients, the LOS

was significantly greater (11.0 days vs 5.0 days, p<0.01, Table 4.4), and was also

greater for patients with toxigenic strains compared with patients with non-toxigenic

82

strains (16.0 days vs 10.0 days, p=0.06, Table 4.6). This is typically seen among CDI

patients worldwide20.

4.3.3.4. Recent nasogastric tube feeding

Previous nasogastric tube feeding was also a common precursor for CDI (40.0% vs

6.1%, p<0.01, Table 4.5) and for infection with any C. difficile strain (26.1% vs 26.4%,

p<0.01, Table 4.4). Nasogastric feeding is a known risk factor for CDI351. It is unclear

whether patients who receive tube feeding are at higher risk for CDI anyway due to the

severity and complications of their illness. It has been postulated that the tube actually

serves to transmit C. difficile spores directly to the gut, contaminated by healthcare

workers. Nasogastric feeding is often delivered straight to the colon to improve

tolerance by patients, bypassing gastric secretions which may kill some spores352.

4.3.3.5. Recent antimicrobial use

Previous use of antibiotics was also significantly more common among CDI cases and

C. difficile-positive patients. Penicillins were most commonly used by CDI cases

(80.0% vs 30.3%, p<0.01, Table 4.5), and C. difficile-positive patients (82.6% vs

17.0%, p<0.01, Table 4.4) and are typically associated with risk of CDI330.

Cephalosporins were also significantly associated with CDI (70.0% vs 22.7%, p<0.01,

Table 4.5) and C. difficile positivity (52.2% vs 18.9%, p<001, Table 4.4) in this study.

Cephalosporins are broad-spectrum antibiotics that have been identified as an important

risk factor for CDI330. Reduction in use of cephalosporins in Sir Charles Gairdner

Hospital in WA in the 1990s resulted in a significant decrease in the incidence of CDI,

suggesting cephalosporin use was a driving factor in increasing the incidence of CDI267.

Cephalosporins and penicillin derivatives amoxycillin/amoxycillin clavulanate were

also significantly associated with CDI in the descriptive studies of CDI patients in

Australia (Chapter 3).

Other antibiotics that were also associated with CDI were macrolides (20.0% vs 3.0%,

p<0.05, Table 4.5) and metronidazole (20.0% vs 1.5%, p<0.05, Table 4.5).

Metronidazole use was not significantly associated with C. difficile positivity in general

(Table 4.4), despite the fact that theoretically it would deplete any C. difficile strains or

similar bacteria occupying the same ecological niche of the gut microflora. This finding

may have been due to the small number of CDI cases studied here.

83

4.3.3.6. PPI use

PPI use was also significantly more common among CDI cases (70.0% vs 34.8%,

p<0.05, Table 4.5) and C. difficile-positive patients (65.2% vs 28.3%, p<0.01, Table

4.3) than C. difficile-negative patients. This finding mirrors others reported elsewhere51,

328. Similar results were found for CDI patients presenting at emergency departments

(EDs) outlined in Chapter 3. As discussed in section 3.3.4.5, the role of PPIs as a risk

factor for CDI is somewhat disputed52. PPIs are one of the most commonly used drugs

in the world, implying that PPIs could be identified as commonly associated with any

disease affecting older people with underlying health issues. A recent report of gut

microflora alteration in PPI users supports a possible link to risk of CDI however333.

4.3.3.7. Comorbidities

No individual comorbidities were significantly associated with CDI cases compared

with non-CDI cases (Table 4.5). However, patients with two or more comorbidities

more frequently had CDI than those that did not (60.0% vs 25.8%, p<0.05, Table 4.5).

Death was significantly more common among CDI cases (30.0% vs 6.1%, p<0.05,

Table 4.5). These data also are similar to reports from elsewhere in the world, indicating

that the descriptive epidemiology of CDI in Malaysia is similar in general to other parts

of the world.

4.3.3.8. Markers of severe CDI

Markers of severe infection occurred in 40.0% of CDI patients in Malaysia, which is

higher than that observed in the studies in Australia (18.1% to 20.0%) outlined in

Chapter 3. The proportion of patients experiencing elevated WCC was (20.0%, Table

4.4) was lower than that reported in Europe (29%), but the proportion with elevated

serum creatinine (20.0%) was higher than that reported in Europe (8%)45. The measure

in Europe was more accurate however, having a baseline reading prior to symptom

onset, while here elevated serum creatinine was estimated as 1.5 times the upper value

of the “normal range” (0.5-1.5 mg/dL for men, 0.6-1.2mg/dL for women). Fever was

less commonly reported among the Indonesian CDI patients compared with Malaysian

CDI patients (48.7% vs 90.0%, Tables 4.3, 4.5).

4.3.3.9. Sex

84

The majority of CDI cases in the Indonesian study were female (53.6%, Table 4.3).

Females were less commonly included in the Malaysian study (43.5%, Table 4.4) which

differs from the studies in Australia (chapter 3) and findings elsewhere in the world42, 45.

It is unclear whether this is a true finding, but it is perhaps more likely to be due to

cultural differences in seeking medical care, or in providing consent to take part in the

study.

4.3.4. High prevalence of non-toxigenic C. difficile in Malaysia and Indonesia

Half (50.0%) of all isolates were non-toxigenic in both the Indonesian and Malaysian

studies. Patients were sometimes co-infected with a toxigenic and a non-toxigenic

strain, while others had a non-toxigenic strain only. This high frequency of non-

toxigenic strains is of interest. Details on whether the patients were positive for another

pathogen, or C. difficile only, were not available but a comparison of characteristics for

all patients with CDI, with toxigenic strains only, and with non-toxigenic strains was

undertaken. The only difference which came close to statistical significance was that

patients with a toxigenic strain had a longer LOS compared to patients with non-

toxigenic strains (median 16.0 vs 10.0 days, p=0.06, Table 4.6).

The role of non-toxigenic C. difficile is unclear. Recent studies have explored the

possibility of treating patients with recent CDI with a non-toxigenic strain of C. difficile,

to occupy the ecological niche left in the gut microflora following treatment with

metronidazole or vancomycin, to prevent recurrent infection353. Whether non-toxigenic

strains in SE Asia act protectively against CDI when other pathogens are present, or

actually contribute to diarrhoeal symptoms somehow, is unknown. No significant

differences in disease characteristics for patients with toxigenic vs non-toxigenic strains

were noted in the Malaysian study (Table 4.6). The same associations for CDI were

noted for C. difficile positivity in general, apart from age which was significantly higher

for C. difficile-positive patients vs C. difficile-negative patients, but not in CDI patients

compared to non-CDI patients (Tables 4.4, 4.5). This suggests that non-toxigenic strains

were possibly protective against CDI in older patients, but were transmitted between

patients in the same manner as toxigenic strains. Larger sample sizes and collection of

further data including co-infections and recurrence rates could provide some

informative insight into the role of non-toxigenic strains in preventing disease.

4.3.5. Strengths of these studies

85

This was the first study of its kind to be performed in Indonesia, and greatly expanded

on previous knowledge of the prevalence of C. difficile in Malaysia. These studies have

provided baseline data on the prevalence and types of C. difficile circulating in Malaysia

and Indonesia. They have also informed further on the suitability of testing practices in

SE Asia. These studies are also the first to describe the epidemiology of CDI patients in

Indonesia, and expand on few previous descriptive studies reported from Malaysia.

These studies thus provide an important reference for future studies in the region.

4.3.6. Limitations of these studies

The requirement to obtain patient consent may have biased the results of both

prevalence studies. It is unknown whether patients were not included in these studies

due to severity of illness or if they chose not to participate. If these patients had been

included this could have changed the epidemiological measurements. However the

descriptive characteristics of disease described here mirror those seen elsewhere in the

world. It is possible that more severe disease and comorbidities would have been

recorded if consent was not required for patient inclusion in the studies.

Only EIA-positive faecal samples were received from Indonesia, so the true prevalence

of CDI may have been higher than determined here. Also it was not possible to

determine the sensitivity and specificity of the test in the Indonesian sites. Another

limitation in Indonesia was that few data were collected, on EIA-positive patients only,

so the descriptive epidemiology of CDI in Indonesia remains largely unknown.

We were unable to define healthcare facility associated (HCFA)-CDI and community

associated (CA)-CDI with the information gathered for either study. Future surveillance

in the region should include collection of hospitalisation information to inform on this.

Another limitation was that demographic and clinical data were only collected for a

subset of cases in Malaysia. The number of CDI cases among that subset was low, so

comparison of some disease characteristics may have been inaccurate.

4.3.7. Implications and recommendations

While the testing method introduced performed well, the discovery of a high prevalence

of non-toxigenic strains in the patient population suggests that a GDH toxin A/B EIA

may not be an ideal test for use in this region. Detection of toxin or ability to produce

86

toxin is usually considered vital for diagnosis of CDI. However, the significance of

these non-toxigenic strains has not been determined and the GDH component was

highly sensitive. An in-house PCR could be developed for low cost to identify toxigenic

C. difficile among diarrhoeal patients. This would depend on the laboratory having PCR

facilities however. Otherwise if surveillance continued, either toxigenic culture, toxin

assay or some form of PCR to detect tcdB would be required to provide accurate

estimates of the prevalence and incidence rates of CDI in SE Asia.

The high prevalence of toxigenic C. difficile identified signifies the need for

surveillance to be performed in Malaysia and Indonesia. This could be difficult due to

limited resources available in these countries. However, regular point prevalence studies

could be a valuable alternative to ongoing surveillance to gather meaningful data and to

lower the costs and effort involved.

The high prevalence of C. difficile found also indicates that infection control measures

could be introduced to attempt to minimise transmission of C. difficile among hospital

patients. Such measures could include antimicrobial stewardship programs, which could

also serve to control other antimicrobial resistant pathogens possibly circulating in the

region. An antimicrobial stewardship program introduced in Singapore General

Hospital was reported where CDI rates were not reduced354. Nonetheless, a stewardship

program could be beneficial to the hospitals described here. Another control measure

could be enforcement of contact precaution measures. In the case of developing country

health facilities, patient isolation may not always be viable, but strict contact

precautions and enhanced cleaning could serve to minimise contamination of the

surrounding environment.

4.3.8. Future studies

Surveillance or point prevalence studies, as mentioned above, would be the most

important next step to monitor C. difficile in Indonesia and Malaysia. Such studies

would provide information over time of changes in incidence or prevalence of CDI.

Monitoring the molecular types circulating in the region would show any shifts in types

circulating, or if new types were emerging. Such studies could also collect information

on previous hospitalisations of patients to describe the proportions of CA-CDI and

HCFA-CDI in the region, since these data were missing from the studies described here.

87

Any infection control measures introduced could be evaluated to measure their

effectiveness in reducing CDI incidence rates and transmission between patients.

Further investigation of non-toxigenic C. difficile strains and their role in SE Asia

would be very interesting. It is unclear why non-toxigenic strains are so prevalent

among hospital inpatients in Malaysia and Indonesia, and potentially elsewhere in SE

Asia. International studies have not reported similar rates, but it is unclear whether this

is because non-toxigenic strains were not detected or they were not considered as

publication-worthy results, or whether they are truly more common in Indonesia and

Malaysia. As mentioned above, whether non-toxigenic strains could be protective

against infection or against severe or complicated disease is not known, however no

significant differences were noted for patients infected with toxigenic strains compared

with non-toxigenic strains (Table 4.5).

4.3.9. Chapter summary

• C. difficile testing was introduced in three hospitals in Malaysia and four

hospitals in Indonesia.

• The prevalence of C. difficile among diarrhoeal patients in Malaysia was 15.4%

and 11.7% in Indonesia, not significantly different.

• CDI patients in Malaysia exhibited many characteristics typical of CDI patients

worldwide - extended length of stay, recent antibiotic use, recent nasogastric

feeding and multiple comorbidities.

• Severe CDI was described in 40.0% of Malaysian CDI patients.

• Non-toxigenic strains were commonly detected, and made up half of all isolates.

• CDI patients were less likely to be discharged well and more likely to have an

outcome of death than other diarrhoeal patients.

• No significant differences were found for Malaysian patients infected with

toxigenic strains compared with non-toxigenic strains.

88

CHAPTER 5. MOLECULAR EPIDEMIOLOGY OF

C. DIFFICILE IN ASIA PACIFIC COUNTRIES

The following publications have arisen from this chapter:

Cheng AC, Collins DA, Elliott B, Ferguson JK, Paterson DL, Thean S, Riley TV.

Laboratory-based surveillance of Clostridium difficile circulating in Australia,

September - November 2010. Pathology. 2016:48:257-260.

Collins DA, Riley TV. Routine detection of Clostridium difficile in Western Australia.

Anaerobe. 2016:37:34-7

Mori N, Yoshizawa S, Saga T, Ishii Y, Murakami H, Iwata M, Collins DA, Riley TV,

Tateda K. Incorrect diagnosis of Clostridium difficile infection in a university hospital

in Japan. Journal of Infection and Chemotherapy. 2015:21:718-22.

89

Foster NF, Collins DA, Ditchburn SL, Duncan CN, van Schalkwyk JW, Golledge CL,

Keed ABR, Riley TV. Epidemiology of Clostridium difficile infection in two tertiary-

care hospitals in Perth, Western Australia: a cross-sectional study. New Microbes and

New Infections. 2014;2,64-71.

Tan XQ, Verrall AJ, Jureen R, Riley TV, Collins DA, Lin RT, Balm MN, Chan D,

Tambyah PA. The emergence of community-onset Clostridium difficile infection in a

tertiary hospital in Singapore: a cause for concern. International Journal of

Antimicrobial Agents. 2014;43:47-51.

5.1. INTRODUCTION

Data on the molecular types of C. difficile circulating in Asia Pacific countries were rare

prior to the commencement of this PhD. The lack of adequate testing in many countries,

and of anaerobic culture facilities, led to few studies on the molecular ecology of

C. difficile being performed for most Asian countries. Molecular epidemiological

studies had only been published previously for isolate collections from China, Hong

Kong, Japan and Korea148, 152-154, 157, 169, 173, 174, 183, 197, 216. Few of these studies provided

detailed information on ribotypes (RTs) of strains, due to a lack of reference strains. In

particular, many Japanese studies used a local ribotyping nomenclature which could not

be extrapolated to internationally recognised types without sharing of reference strains

with other laboratories159.

Another form of typing, slpA typing, was also favoured in Japan for a number of years,

and only some slpA types were related to internationally recognised or local RTs150.

Tandem repeat sequence typing of C. difficile isolates has also been reported from

China285. Multi-locus sequence typing (MLST) has been performed occasionally in Asia

and studies have been reported from China186, 191, 192, Thailand220 and Japan156. MLST

types have been published with corresponding RTs188 which has allowed extraction of

some data on strains in circulation from these reports.

Common strains of C. difficile identified in Asia previously were RTs 017 and 018

(Table 1.1). C. difficile RT 017, an A-B+ strain, has been the most widely reported RT in

90

Asian in previous studies. RT 018 was historically called “smz” in Japanese studies and

has also been reported from Korea previously. Few other RTs have been reported,

besides 014, 002, 001 and 046, with many strains reported as unknown (Figures 1.1-

1.4).

The binary toxin-positive (CDT+) strains 027 and 078 were rarely reported in Asia. RT

027 has been reported sporadically from Hong Kong212, Japan275, Taiwan206, 208,

mainland China194, 195, Korea180, 355, Singapore243, however, some of these RT 027

strains were fluoroquinolone susceptible and therefore not “epidemic” strains.

Meanwhile reports of RT 078 in Asia have been limited to animals which could be

traced back to Europe160, 161.


The aims of this chapter were:

• To collect C. difficile isolates from across the Asia-Pacific region and to

perform molecular typing.

• To produce a broad overview of the molecular ecology of C. difficile in the

Asia-Pacific region.

• To compare the molecular epidemiology of C. difficile in Australia with that of

other Asia-Pacific countries.

5.2. RESULTS

An overview of all isolates included in analyses is given in Table 5.1.

5.2.1. Australian C. difficile isolates

5.2.1.1. 2012 surveillance studies

In the 2012 collection, 80 RTs were identified, the 10 most common comprising 65.5%

of the whole collection (Table 5.2). The two most common RTs were RTs 014/020

(25.5%) and 002 (10.5%), followed by RTs 056 (5.9%), 070 (4.2%) and 054 (4.2%).

Emerging RTs which were not previously seen in the top 10 or increased in prevalence

Table 5.1. Summary of all isolates collected.

Country Study Year Number of isolates

Australia Nationwide survey 2012 542

Cross-sectional WA study 2012 70

WA routine testing study 2014 61

WA ED study 2013/2014 66

Various Asia-Pacific study 2014/2015 414

Singapore Isolate collection 2011/2012 70

Japan Isolate collection 2010 118

Indonesia Prevalence study 2014/2015 74

Malaysia Prevalence study 2015/2016 82

Total 1497

Table 5.2. Australian isolates collected November 2012.

RT State/Territory n (%) Australia

NSW WA VIC SA QLD TAS/ACT n (%)

014/020 37(18.2) 34(31.5) 16(21.9) 30(39.5) 14(26.4) 7(24.1) 138(25.5)

002 29(14.3) 11(10.2) 3(4.1) 4(5.3) 5(9.4) 5(17.2) 57(10.5)

056 12(5.9) 8(7.4) 4(5.5) 2(2.6) 4(7.5) 2(6.9) 32(5.9)

070 5(2.5) 3(2.8) 2(2.7) 5(6.6) 6(11.3) 2(6.9) 23(4.2)

054 9(4.4) 4(3.7) 7(9.6) 3(3.9) 0 0 23(4.2)

015 7(3.4) 3(2.8) 5(6.8) 2(2.6) 4(7.5) 1(3.4) 22(4.1)

017 6(3.0) 4(3.7) 3(4.1) 4(5.3) 1(1.9) 0 18(3.3)

012 9(4.4) 4(3.7) 0 2(2.6) 0 0 15(2.8)

010 7(3.4) 4(3.7) 0 1(1.3) 2(3.8) 0 14(2.6)

244 5(2.5) 4(3.7) 0 0 4(7.5) 0 13(2.4)

043 2(1.0) 0 1(1.4) 4(5.3) 1(1.9) 2(6.9) 10(1.8)

018 2(1.0) 1(0.9) 3(4.1) 0 2(3.8) 2(6.9) 10(1.8)

046 2(1.0) 3(2.8) 1(1.4) 3(3.9) 0 1(3.4) 10(1.8)

027 8(3.9) 0 1(1.4) 0 0 0 9(1.7)

QX 077 4(2.0) 1(0.9) 2(2.7) 0 0 1(3.4) 8(1.5)

078 5(2.5) 0 2(2.7) 0 0 0 7(1.3)

137 1(0.5) 3(2.8) 0 1(1.3) 2(3.8) 0 7(1.3)

001 2(1.0) 1(0.9) 0 2(2.6) 2(3.8) 0 7(1.3)

QX 013 1(0.5) 0 4(5.5) 0 1(1.9) 0 6(1.1)

QX 412 3(1.5) 0 2(2.7) 1(1.3) 0 0 6(1.1)

251 2(1.0) 1(0.9) 1(1.4) 1(1.3) 1(1.9) 0 6(1.1)

QX 150 2(1.0) 2(1.9) 2(2.7) 0 0 6(1.1)

QX 076 1(0.5) 1(0.9) 1(1.4) 2(2.6) 0 0 5(0.9)

103 1(0.5) 0 2(2.7) 1(1.3) 1(1.9) 0 5(0.9)

126 2(1.0) 2(1.9) 0 1(1.3) 0 5(0.9)

005 2(1.0) 1(0.9) 1(1.4) 0 1(1.9) 0 5(0.9)

053 2(1.0) 1(0.9) 0 0 0 0 3(0.6)

Others 35 (17.2) 12 (11.1) 10 (13.7) 7 (9.2) 2 (3.8) 6 (20.7) 72 (13.3)

Total 203 (100) 108 (100) 73 (100) 76 (100) 53 (100) 29 (100) 542 (100)

RT ribotype; NSW New South Wales; WA Western Australia; VIC Victoria; SA South Australia; QLD Queensland; TAS Tasmania; ACT Australian Capital Territory

91

since the 2010 survey were RTs 015 (4.1%, p<0.05), 017 (3.3%, p<0.01), 012 (2.8%,

p<0.05) and 244 (2.4%, p<0.05, Table 5.3).


Perth

The most common RT among the 70 isolates collected was 014/020 (30.0%) followed

by 056 (5.7%) and 070 (5.7%). RT 244 was among the six most common RTs (4.3%) at

equal prevalence to 002 (4.3%, Table 5.4).

5.2.1.3. Routine testing study

RTs 014/020 (37.7%; 34.1% in routine group, 43.8% in non-requested), 002 (8.2%),

056 (4.9%) were most common in the routine diagnostics study isolates (Table 5.5).

Other strains in the top 10 included RTs 018 (3.3%), 005 (3.3%), 076 (3.3%). RTs 002,

056 and 018 were identified in routinely tested patients only, and in addition RT 018

was only found in general practice (GP) patients. RTs 103 and 247 were found in

outpatients (OPs) only (Figure 5.1).

5.2.1.4. Emergency department study

RT 014/020 made up almost half (45.5%) of the emergency department (ED) study

isolate collection. Other common strains included RT 002 (7.6%), 244 (4.5%), 001

(4.5%), 056 (4.5%, Table 5.6).

5.2.2. Asia-Pacific isolates

This was the first systematic study of C. difficile molecular epidemiology in the Asia-

Pacific region. Participating sites are summarised in Table 5.7. Overall, 65 different

RTs were identified, 34 of which were represented by single isolates. A large proportion

(19.9%) of isolates was A-B+CDT- strains, mostly RT 017 (16.6%) and RT 369 (4.1%).

Non-toxigenic strains made up 5.8% of the isolate collection. The most common RTs

were RTs 017 (16.6%), 014/020 (11.1%), 018 (9.9%) and 002 (9.2%, Table 5.8).

The most frequent strains isolated by country are shown in Figure 5.2. In Australia, RT

014/020 dominated and it was also one of the most common RTs in Thailand, Singapore

and Vietnam. RT 017 was most dominant in Thailand at >50% of strains, and was also

Table 5.3. Comparison of top 10 isolates in Australian 2010 and 2012 surveillance studies.

2010 2012

Position Ribotype n (%) Ribotype n (%)

1 014/020 99(30.0) 014/020 138(25.5)

2 002 39(11.8) 002 57(10.5)

3 054 14(4.2) 056 32(5.9)

4 056 13(3.9) 070 23(4.2)

5 070 12(3.6) 054 23(4.2)

6 005 11(3.3) 015 22 (4.1)*

7 012 9(2.7) 017 18(3.3)**

8 018 9(2.7) 012 15(2.8)

9 046 8(2.4) 010 14(2.6)

10 103 8(2.4) 244 13(2.4)*

Other 105(31.8) 187(34.5)

NT 3(0.9)

Total 330(100) 542 (100)

*p<0.05 **p<0.01

NT Isolate could not be ribotyped

Table 5.4. Ribotypes of isolates collected from 70 CDI patients at two WA hospitals,

2011-2012.

Ribotype Toxin profile n (%)

014/020 A+B+CDT- 21 (30.0)

056 A+B+CDT- 4 (5.7)

070 A+B+CDT- 4 (5.7)

002 A+B+CDT- 3 (4.3)

244 A+B+CDT+ 3 (4.3)

QX 033 A+B+CDT- 3 (4.3)

010 A-B-CDT- 2 (2.9)

017 A-B+CDT- 2 (2.9)

018 A+B+CDT- 2 (2.9)

046 A+B+CDT- 2 (2.9)

251 A+B+CDT+ 2 (2.9)

QX 075 A+B+CDT- 2 (2.9)

QX 077 A+B+CDT- 2 (2.9)

Others Various 18 (25.7)

Total 70 (100)

Table 5.5. WA routine testing study C. difficile isolates.

n (%)

Ribotype Toxin profile Routine group Non-requested group Total (%)

014/020 A+B+CDT- 16 (35.6) 7 (43.8) 23 (37.7)

002 A+B+CDT- 4 (8.9) 1 (6.3) 5 (8.2)

056 A+B+CDT- 3 (6.7) 0 3 (4.9)

018 A+B+CDT- 2 (4.4) 0 2 (3.3)

005 A+B+CDT- 1 (2.2) 1 (6.3) 2 (3.3)

076 A+B+CDT- 1 (2.2) 1 (6.3) 2 (3.3)

Others Various 18 (40.0) 6 (37.5) 24 (39.3)

Total 45 (100) 16 (100) 61 (100)

Table 5.6. ED Study isolates.


014/020 A+B+CDT- 30 (45.5)

002 A+B+CDT- 5 (7.6)

244 A+B+CDT+ 3 (4.5)

001 A+B+CDT- 3 (4.5)

056 A+B+CDT- 3 (4.5)

005 A+B+CDT- 2 (3.0)

012 A+B+CDT- 2 (3.0)

015 A+B+CDT- 2 (3.0)

017 A-B+CDT- 2 (3.0)

081 A+B+CDT- 2 (3.0)

QX 013 A+B+CDT- 2 (3.0)

010 A-B-CDT- 1 (1.5)

018 A+B+CDT- 1 (1.5)

054 A+B+CDT- 1 (1.5)

070 A+B+CDT- 1 (1.5)

103 A+B+CDT- 1 (1.5)

247 A+B+CDT- 1 (1.5)

QX 068 A+B+CDT- 1 (1.5)

QX 086 A+B+CDT- 1 (1.5)

QX 224 A-B-CDT- 1 (1.5)

QX 399 A+B+CDT+ 1 (1.5)

Total 66 (100)

Table 5.7. Summary of participating sites and recruited participants, Asia-Pacific study.

Country Participating sites Recruited

participants per

site n (%)

Total recruited

participants per

country n (%)

Number of

isolates

collected

Australia Sir Charles Gairdner

Hospital

12 (2.0) 60 (10.0) 50

Fremantle Hospital 37 (6.2)

Royal Perth Hospital 11 (1.8)

China Huashan Hospital FDU 22 (3.7) 64 (10.7) 44

Shanghai East Hospital 2 (0.3)

Ruijin Hospital SJTU 6 (1.0)

FMUU Xijing Hospital 6 (1.0)

Shanghai Renji Hospital 10 (1.7)

Sir Run Run Shaw Hospital 6 (1.0)

Peking University First

Hospital

9 (1.5)

SMU Nanfang Hospital 3 (0.5)

Hong

Kong

Princess Margaret Hospital 5 (0.5) 36 (6.0) 6

Prince of Wales Hospital 38 (4.7)

Queen Mary Hospital 3 (0.5)

India PDHN Hospital 2 (0.3) 2 (0.3) 0

Indonesia RSUP Dr Kariadi Hospital 7 (1.2) 7 (1.2) 7

Japan Toho University Omori

Medical Centre

38 (6.3) 48 (8.0) 46

Toho University Ohashi

Medical Centre

5 (0.8)

Toho University Sakura

Medical Centre

5 (0.8)

Malaysia Hospital Sungai Buloh 1 (0.2) 2 (0.3) 1

Hospital Queen Elizabeth 1 (0.2)

Philippines National Kidney Transplant

Hospital

6 (1.0) 9 (1.5) 9

Davao Doctors Hospital 1 (0.2)

The Medical City 2 (0.3)

Singapore National University

Hospital

40 (6.7) 66 (11.0) 22

Changi General Hospital 10 (1.7)


Singapore General

Hospital

16 (2.7)

South Korea Yonsei University

Severance Hospital

22 (3.7) 99 (16.5) 89

Chungnam National

University Hospital

66 (11.0)

Inje University Haeundae

Paik Hospital

11 (1.8)

Taiwan E-Da Hospital 8 (1.3) 99 (16.5) 88

National Taiwan

University Hospital

48 (8.0)

Taipei Medical University

Hospital

5 (0.8)

Veterans General Hospital

KH

6 (1.0)

National Cheng Kung

University Hospital

32 (5.3)

Thailand Ramathibodi Hospital 25 (4.2) 44 (7.3) 38

KCMH Department of

Gastroenterology

Bangkok

1 (0.2)

Songklanagarind Hospital 18 (3.0)

Vietnam National HTD Vietnam 2 (0.3) 64 (10.7) 14

Pediatric Hospital No 1 62 (10.3)

Total 600 414

Table 5.8. Asia-Pacific C. difficile isolates, overall study total.


017 A-B+CDT- 69 (16.6)

014/020 A+B+CDT- 46 (11.1)

018 A+B+CDT- 41 (9.9)

002 A+B+CDT- 38 (9.2)

012 A+B+CDT- 20 (4.8)

369 A-B+CDT- 17 (4.1)

QX 239 A+B+CDT- 15 (3.6)

QX 032 A+B+CDT- 15 (3.6)

001 A+B+CDT- 13 (3.1)

046 A+B+CDT- 12 (2.9)

106 A+B+CDT- 12 (2.9)

056 A+B+CDT- 7 (1.7)

070 A+B+CDT- 6 (1.4)

009 A-B-CDT- 4 (1.0)

043 A+B+CDT- 5 (1.2)

QX 320 A+B+CDT- 5 (1.2)

103 A+B+CDT- 4 (1.0)

QX 029 A+B+CDT- 4 (1.0)

QX 026 A+B+CDT- 4 (1.0)

QX 076 A+B+CDT- 4 (1.0)

027 A+B+CDT+ 3 (0.7)

QX 020 A+B+CDT- 3 (0.7)

QX 337 A+B+CDT+ 3 (0.7)

QX 449 A+B+CDT+ 3 (0.7)


Total 414 (100)

Figure 5.1. Distribution of C. difficile RTs among patient types.

Figure 5.2. Distribution of C. difficile RTs in Asia-Pacific countries.

92

the most common strain in China/Hong Kong and in the southeast Asian countries

Singapore, Indonesia, Malaysia, Philippines and Vietnam. RT 018 dominated in Korea

and the similar strain, QX 239 was most common in Japan. RT 002 dominated in

Taiwan and was also common in Japan. RT 369 was also among the most common

strains in Japan, and was also third most common in China and Hong Kong.

5.2.3. Other Asian isolate collections

5.2.3.1. Singapore isolates

In the Singaporean isolate collection, 18 RTs were identified. The most common was

RT 053 (22.9%), followed by RTs 014/020 (18.6%), 012 (17.1%), 046 (8.6%, Table

5.9).

5.2.3.2. Japan isolates

Among the 118 Japanese isolates, 31 RTs were represented. RT 369 was most common

(22.9%), followed by RTs 014/020 (11.0%), 018 (10.2%), 002 (10.2%), 127 (5.1%) and

QX 029 (5.1%, Table 5.10).

5.2.3.3. Indonesian isolates

More than 20 RTs were identified in the Indonesian isolate collection. The most

common RT was 017 (24.3%, Table 5.20). A high proportion of strains were non-

toxigenic (50.0%). Other toxigenic strains which were common were 053 (4.1%), QX

134 (4.1%, A-B+CDT-), and 014/020, 103, and 043 (all 2.7%, Table 5.11).

5.2.3.4. Malaysian isolates

Among 82 isolates, RTs 017 and 043 were most common (20.7% and 13.4%), followed

by non-toxigenic strain QX 002 (9.8%) and RTs 053, QX 026 and 039 (4.9% each), RT

009 (2.4%) and singleton strains of various toxin profiles (Table 5.12). As with

Indonesia, a high proportion of strains were non-toxigenic (50.0%).

5.3. DISCUSSION

5.3.1. C. difficile molecular epidemiology in Australia

5.3.1.1. Prevalent RTs circulating in Australia

Table 5.9. Singaporean isolates collected December 2011 - May 2012.


053 A+B+CDT- 16 (22.9)

014/020 A+B+CDT- 13 (18.6)

012 A+B+CDT- 12 (17.1)

046 A+B+CDT- 6 (8.6)

002 A+B+CDT- 3 (4.3)

017 A-B+CDT- 3 (4.3)

043 A+B+CDT- 3 (4.3)

087 A+B+CDT- 2 (2.9)

010 A-B-CDT- 1 (1.4)

056 A+B+CDT- 1 (1.4)

075 A+B+CDT- 1 (1.4)

106 A+B+CDT- 1 (1.4)

QX 002 A-B-CDT- 1 (1.4)

QX 026 A+B+CDT- 1 (1.4)

QX 029 A+B+CDT- 1 (1.4)

QX 087 A+B+CDT- 1 (1.4)

QX 153 A+B+CDT- 1 (1.4)

QX 161 A+B+CDT- 1 (1.4)

QX 354 A+B+CDT- 1 (1.4)

Unknown A+B+CDT- 1 (1.4)

Total 70 (100)

Table 5.10. Japanese isolates, collected in Tokyo, 2010.


369 A-B+CDT- 27 (22.9)

002 A+B+CDT- 12 (10.2)

014/020 A+B+CDT- 12 (10.2)

QX 239 A+B+CDT- 9 (7.6)

127 A+B+CDT+ 6 (5.1)

018 A+B+CDT- 5 (4.2)

QX 029 A+B+CDT- 5 (4.2)

046 A+B+CDT- 4 (3.4)

012 A+B+CDT- 3 (2.5)

017 A-B+CDT- 3 (2.5)

001 A+B+CDT- 2 (1.7)

015 A+B+CDT- 2 (1.7)

056 A+B+CDT- 2 (1.7)

131 A+B+CDT- 2 (1.7)

QX 013 A+B+CDT- 2 (1.7)

QX 026 A+B+CDT- 2 (1.7)

QX 119 A+B+CDT- 2 (1.7)

QX 289 A+B+CDT- 2 (1.7)


Total 118 (100)

Table 5.11. Indonesian isolates, collected in Central Java Province, July 2014- February 2015.


017 A-B+CDT- 18 (24.3)

QX 224 A-B-CDT- 7 (9.5)

QX 238 A-B-CDT- 6 (8.1)

QX 108 A-B-CDT- 6 (8.1)

053 A+B+CDT- 3 (4.1)

QX 134 A-B+CDT- 3 (4.1)

014/020 A+B+CDT- 2 (2.7)

103 A+B+CDT- 2 (2.7)

043 A+B+CDT- 2 (2.7)

QX 076 A+B+CDT- 2 (2.7)

QX 083 A-B-CDT- 2 (2.7)


Total 74 (100)

Table 5.12. Malaysian isolates, collected in Kota Bahru and Kuala Lumpur, July 2015-February 2016.


017 A-B+CDT- 17 (20.7)

043 A+B+CDT- 11 (13.4)

QX 002 A-B-CDT- 8 (9.8)

053 A+B+CDT- 4 (4.9)

QX 026 A+B+CDT- 4 (4.9)

039 A-B-CDT- 4 (4.9)

009 A-B-CDT- 2 (2.4)


Total 82 (100)

93

The most common RT in Australia has consistently been RT 014/020 for at least 5 years

(Tables 5.2-5.6). RT 014 is also common in European and increasingly common in

North American isolate collections 45, 113, 356. Few data are available for other regions,

but given it was also found at high frequency in the Asia-Pacific isolates represented

here, it is probably the most common strain of C. difficile in the world at present and

therefore should be studied intensively to try and determine why it has become such a

successful strain.

As part of a study that evaluated a new antimicrobial treatment for C. difficile infection

(CDI), Cheknis et al. typed 24 Australian C. difficile isolates recovered in the mid-

2000s by restriction endonuclease analysis (REA) and found that Australian strains did

not correspond well to American and European strains357. However, several RT 014

strains were identified by their corresponding REA pattern. Genomic analysis of RT

014 strains within Australia has shown a number of sub-lineages exist (D Knight,

personal communication), suggesting that RT 014 strains across the world probably

exhibit a great degree of diversity. This has been further supported by the study of Eyre

et al. which found high diversity among various RTs isolated from hospital inpatients

with CDI in Oxfordshire, UK110.

RT 002 was in the top three strains among all the Australian collections described here,

and was the fourth most common RT among the Asia-Pacific isolate collection (Tables

5.2-5.8). RT 002 appears to be less prevalent in Europe and the USA than in Australia

and Asia45, 113. A previous study in Hong Kong reported that 002 isolates showed

increased sporulation rates compared to other strains216. This may contribute to its

ubiquity in the Asia-Pacific region.

RT 056 was another common strain identified in Australian collections (Tables 5.2-5.6).

In particular RT 056 appeared to increase in prevalence significantly between 2010 and

2012 (Table 5.4). It was also prevalent in a European isolate collection45, where it was

associated with complicated disease outcome. A systematic study of disease severity vs

RT has not yet been performed in Australia, although crude analysis in the studies

outlined in Chapter 3 did not identify any RTs significantly associated with severe or

complicated CDI. A survey of Australian veal calves at slaughter found C. difficile RT

056 was commonly isolated from the gastrointestinal contents of these calves56. This

94

finding highlights a need for further study of Australian foodstuffs and comparison of

human and animal strains to determine whether RT 056 is entering the food-chain and

therefore a possible cause of foodborne infection.

The detection of RT 018 strains of C. difficile in Australia is also notable. RT 018

appeared in the top 12 most common strains in every Australian isolate collection

described here (Tables 5.2-5.8) RT 018 is highly prevalent in Asia, particularly in

Japan307. It appeared in Korea in the early 2000s157 and also emerged in Italy at around

the same time and caused widespread severe disease and a number of outbreaks17. A

recent study of Italian C. difficile strains showed RT 018 was the most prevalent RT

overall and was highly transmissible, accounting for 95.7% of secondary CDI cases358.

This apparently high transmissibility may be related to enhanced virulence, and

surveillance for this strain in Australia appears warranted to guard against increasing RT

018 infection rates in the future.

RT 017 was the most common A-B+ strain in Australia; however, it was not seen in the

high proportions seen elsewhere in other Asia-Pacific countries (Tables 5.2-5.12).

Given the high frequency of travel between Australia and its close neighbours in Asia,

the appearance of RT 017 is expected in Australia. Two RT 017 isolates were identified

among the ED patients (Table 5.6). One of those patients had visited Thailand 2 weeks

previously for elective plastic surgery. It is likely this hospital visit was the location of

their exposure to C. difficile. While the appearance of RT 017 may reflect travel from

affected regions, other avenues for introduction, such as food or animal imports, should

not be ignored as these strains have caused significant outbreaks in the Asian region.

5.3.1.2. CDT-positive strains in Australia

Australia has seemingly escaped the epidemic RT 027 and 078 strains of C. difficile that

have been linked to widespread outbreaks across North America and Europe 7, 359. RT

027 and 078 were only detected in relatively low frequencies in the 2010 and 2012

surveys (Table 5.2, 5.3). However, several other CDT+ strains, in particular RTs 244

and 251, appeared in Australia in recent years (Tables 5.3, 5.4).

95

RT 027 was detected in Australia for the first time in 2009 268. The most significant RT

027 outbreak involved at least seven patients at a Victorian hospital in early 2010269. In

the 2010 Australian Commission on Safety and Quality in Healthcare survey RT 027

strains were isolated only from New South Wales (NSW), all from Sydney (Table 5.2),

while Victorian laboratories did not participate in the survey. In the 2012 survey, nine

RT 027 isolates were recovered, eight of which were from NSW and one from VIC,

introduced by an overseas traveller (Table 5.3). The NSW isolates were mainly from

north Sydney suggesting a low-level endemicity may have been established in the area.

Otherwise, C. difficile RT 027 was not detected in any other Australian strain collection

described here. The apparent rarity of RT 027 strains in Australia, and their inability to

establish in the country, is likely to be due to its resistance to fluoroquinolones360.

Australian policies on fluoroquinolones are conservative, restricting widespread use270.

This may have helped to prevent local proliferation of RT 027 due to lack of selective

pressure.

Seven RT 078 (0.9%) isolates were identified in the 2012 survey, while only one was

detected in the 2010 survey (Tables 5.2 and 5.3). The increase in prevalence could

indicate a source in Australia, or (more likely) repeated importation from overseas.

C. difficile RT 078 is commonly found in production animals in the Northern

Hemisphere46, however to date it has not been detected in any Australian production

animals 54-56. While infections in Australian humans are increasing, the apparent lack of

RT 078 in animals in Australia suggests the source probably lies overseas. The

prevalence of RT 078 strains causing infections has also been increasing in Europe in

recent years45, indicating the possibility of importation to Australia in international

travellers. Alternatively, imported food products of animal origin from Europe could

plausibly be a source of RT 078 in Australia, or imported animals such as racehorses160.

C. difficile RT 244, another CDT+ strain, was first detected in Australia (one isolate in

Queensland) in 2010302. In the cross-sectional study in Western Australia (section

3.2.2), the prevalence of RT 244 was at 4.3%, placing it in the top five most common

RTs in circulation305. The prevalence of RT 244 was lower at 2.4% in the 2012 survey;

however its increase since the 2010 survey was considered statistically significant

(Table 5.4). RT 244 was among the five most common strains in the ED study (Table

96

5.6), further strengthening an association with community onset (CO)-CDI361. In

addition, in the 2012 nationwide survey, 11/13 RT 244 strains were isolated in private

laboratories or community hospitals, predominantly in NSW and Queensland. Two of

these could be classified as causing community associated (CA)-CDI, where clinical

data were available. These findings suggest that RT 244 was introduced to Australia

around 2010, and was transmitted from a source/reservoir in the community. The most

likely explanation would be imported food which was distributed around the country.

RT 244 strains are fluoroquinolone susceptible48 and were associated with severe CDI

in a case-control study in Melbourne. In that study the 30-day mortality was calculated

at 42%48. A corresponding study in New Zealand had similar findings, where 244

strains caused severe CA-CDI with increased mortality47. In the 2012 nationwide

survey, clinical data were available for nine RT 244 isolates; one caused severe CDI

where elevated leukocyte count was observed. The apparent propensity of RT 244

strains to cause severe CDI and its association with CO-CDI mean it requires vigilance

and further investigation of its possible origins.

C. difficile RT 251 is also CDT+, and also appeared in Australia recently. It was first

isolated in 2010302. In the 2012 survey 1/6 isolates could be classified as CA-CDI

where clinical data were collected. RT 251 also appears to have the propensity to cause

severe disease. In late 2015, a 32 year old immunocompetent patient in Victoria

developed CA-CDI caused by RT 251, and quickly progressed to colectomy followed

by death in the intensive care unit (ICU) (unpublished report). Further studies of RT 251

and its possible origins are required. It is interesting to note that neither RTs 244 nor

251 were detected in any Asian isolate collection described here. This suggests that the

source of infection is localised to Australia, however, both these RTs belong to clade 2

of C. difficile, the same clade as RT 027. Therefore, it is highly probable that both RTs

244 and 251 originated from North America like RT 027.

Among other CDT+ C. difficile strains, seven RT 126/127 isolates and one RT 033 were

detected in the 2012 survey (Table 5.3). RTs 236 and 127 were also isolated in Taiwan

and other Asian isolate collections. RT 033 is particularly notable since its toxin profile

is A-B-CDT+. It evades most molecular detection platforms due to its unusual toxin

profile86, meaning it may be more prevalent in the population than it appears to be. A

97

case report was recently published outlining severe infection with RT 033 in a

paediatric patient with ulcerative colitis362. RTs 126, 127 and 033 have also been

detected in Australian cattle and pigs56, 274, indicating possible transmission between

humans and animals by zoonotic transfer or contamination of animal food products.

5.3.1.3. Diversity in Australian RTs

It is interesting to note that the Australian isolate collections exhibited great diversity in

RTs compared to other country collections. The low population density and large area

of Australia could contribute to this. It is likely that a wide array of sources of infection

exists in Australia, indicated by the high number of isolates of single RTs recorded.

Differences in testing practices could also play a role in the apparent greater diversity of

RTs in Australia compared to other Asian countries. For example, ED patients in

Western Australia (WA) are routinely tested for C. difficile when presenting with

diarrhoea, and awareness of C. difficile among GPs and community health practices in

Australia may be greater than in other regions of Asia. In other Asian countries, poorer

awareness of risk factors among physicians could play a role in missed identification of

C. difficile in CO-CDI patients or other patients lacking classical risk factors137. Missed

diagnosis could also be likely to occur in other Asian countries where social disparities

are widespread. CDI rates were previously shown to be higher in certain racial groups in

the USA, however closer analysis revealed this was confounded by lack of access to

medical care in some racial groups of lower socioeconomic status363. A similar situation

could exist in Asia, however data do not yet exist to examine this.

5.3.2. Molecular epidemiology of C. difficile in Asian countries

5.3.2.1. Prevalent RTs circulating in Asia

RT 017 was the most common strain of C. difficile circulating in the Asia-Pacific

collection of isolates (Table 5.8). It was individually the most common strain circulating

in the Chinese, Thailand and Vietnam subset collections (Figure 5.3). It was also highly

prevalent in the Taiwanese subset (Figure 5.3). Previous typing studies of A−B+

C. difficile isolates from multiple countries have shown a high degree of similarity

among strains 23, 364. To date RT 017 has dominated international typing studies of A−B+

strains. The results presented here show it is perhaps more common on mainland Asia,

Figure 5.3. Banding patterns of QX 239 and RT 018.

98

from China through to Malaysia and Indonesia, than on the islands of Japan.

Interestingly, it has also been highly prevalent in Korea. While it appears to be endemic

to Asia, RT 017 has caused outbreaks of CDI in Canada, the Netherlands and Ireland 18,

77, 365. More recently, it was reported as the most common RT in South Africa making

up 50% of a collection of C. difficile isolates, the first report on the molecular

epidemiology of C. difficile in the region136.

Another A-B+ strain, RT 369, was also highly prevalent in the Asia-Pacific collection.

This strain was found in the Japanese isolate collection from 2010, where it was the

most common strain (Table 5.10). It appears to have emerged over recent years in

Japan, where historically it was referred to in literature using local nomenclature as

“trf”152, 159. It appears that RT 369 caused outbreaks of CDI in hospitals in 2000 and

2001, while ribotyping was not performed at the time30, 159, 366. The first detection of the

emergence of RT 369 to become one of the most prevalent strains in Japan was reported

in a study conducted on isolates collected from outbreak and non-outbreak situations

from 2009-2013. This study detected RT 369 in an outbreak setting in a hospital in

2009159.

While RT 369 was the third most common isolate in the Japan subset (Table 5.10), it

was among the top six strains in the Chinese subset (Figure 5.2). Single isolates of RT

369 were also detected in the Korean and Taiwan subsets. These findings suggest RT

369 is localised to northern Asia. RT 369 was not detected in any Australian isolate

collection described here.

RT 018 was also highly prevalent, being the most common strain detected in the Korean

subset of isolates (Figure 5.2). This corresponds to earlier studies in Korea, where RT

018 appears to have replaced RT 017 as the predominant strain since 2009 (Figure

1.2)157. RT 018 has been present in Japan since the earliest typing studies, where it was

referred to as type “smz”148, 152-154. Interestingly, RT 018 belongs to clade 1 C. difficile

107 which is likely to have its home in Europe suggesting this strain has been introduced

into possibly Japan initially from Europe and then into Korea.

A similar RT, QX 239, displayed a very similar banding profile to RT 018, indicating it

may be a sub-lineage of RT 018 (Figure 5.3). While RT 018 was highly prevalent in the

99

2010 Japanese isolate collection (10.2%, Table 5.9), QX 239 was rare (one isolate). In

the Japanese subset of Asia-Pacific isolates, however, QX 239 dominated (Figure 5.2),

indicating it may have arisen from RT 018 in Japan and replaced it. QX 239 was the

sole strain isolated in two participating sites, suggesting possible outbreaks were

occurring. QX 239 was not detected in the Korean subset of isolates. Otherwise the

third site showed greater diversity. If QX 239 and RT 018 were combined to one group,

they would make up the most prevalent RT in the Asia-Pacific collection (Table 5.8).

A RT with a similar banding profile to RT 018, designated as RT 356, was also reported

from Italy recently356. A strain, ysmz, described as having a similar banding profile to

RT 018, was also reported in the 2009-2013 Japanese survey159. The authors did not

confirm the identity of ysmz but it may correspond to RT 356/QX 239, given its similar

description. RT ysmz was localised to one hospital out of 15 surveyed in that study, and

was detected in the 2009 outbreak setting. Otherwise RT 018 was also prevalent in the

other 14 hospitals surveyed. RT 356 showed high antibiotic resistance in a European

survey, warranting close investigation and vigilance for further spread worldwide356.

QX 239 was associated with recurrent CDI in the Asia-Pacific study, which suggests it

also has enhanced antimicrobial resistance, further supporting the notion that they may

be the same strain.

RT 002 was highly prevalent in Taiwan (22.7%, Figure 5.2), and in the Japanese

collections (Figure 5.2, Table 5.10). It was previously described as the most common

strain in Hong Kong216. Too few isolates (n=6) were collected from Hong Kong in this

study to determine whether RT 002 was still the most common strain. Despite their

proximity, RT 002 was not seen in the Chinese or Korean isolate collections.

Meanwhile RT 018 and QX 239 were highly prevalent in the Korean and Japanese

collections, but were not detected in the Chinese and Taiwanese isolate collections.

The Singaporean isolate collections showed a different distribution of strain types

compared to its neighbouring south eastern Asian countries (Vietnam, Thailand,

Malaysia, Indonesia). The most common types were RTs 014/020, 053, 012 and 046

(Table 5.9). RT 053 was not as prevalent in any other isolate collection, nor reported

internationally to be as common as it appears to be in Singapore. This suggests that it is

possibly endemic to Singapore. Singapore’s multi-national population and busy

100

international travel hub may contribute to the diverse molecular epidemiology of

C. difficile isolated in Singapore

C. difficile RT 012 was among the top five most prevalent strains in Asia (Table 5.7)

and also was a top 10 strain in the Australian isolate collections. Interestingly, RT 012

has been isolated from vegetables (Lim et al, unpublished data) and pigs (Knight et al.,

unpublished) in WA where it has been increasing in prevalence over the last few years.

In general, the Japanese isolate collections showed low strain diversity. In particular, the

top three RTs in the Asia-Pacific study subset made up 82.9% of the isolate collection.

In Japanese hospitals, patients tend to have longer stays, where they complete courses of

medication, particularly antimicrobials, before discharge. This could contribute to the

low diversity in RTs seen. A UK study showed that even within a hospital patient

population, the diversity among isolates of the same RT was higher than expected,

showing that sources of infection were more diverse than previously thought110. Where

patient-patient transmission was expected to occur frequently, the diversity within strain

types showed that this did not occur as frequently as previously thought.

C. difficile RT 014/020 was also highly prevalent in the Asian isolate collections, being

the third most common strain in the Asia-Pacific collection (Table 5.6) and the 2010

Japanese collection (Table 5.10), and the second most common strain in the

Singaporean collection (Table 5.9). This provides further evidence that RT 014/020 is

an important RT throughout the world and more work on this strain should be

encouraged in the future.

RT 017 was the most prevalent strain in Thailand, Vietnam and Malaysia. Interestingly,

RT 017 was not identified in the Philippines subset, despite its geographical proximity

to countries where it is highly prevalent. This was an interesting finding, indicating the

Philippines’ molecular ecology of C. difficile may differ somewhat from mainland Asia,

however the low number of isolates collected from the Philippines (n=7) means

meaningful observations cannot be drawn.

For two patients in the ED study of CDI patients in WA (section 3.2.4), acquisition of

C. difficile most likely occurred in Thailand, due to their hospitalisation there. In each

101

case their RT (RT 017 and QX 369) suggested an origin in Thailand. Both these cases

are likely to have been healthcare facility associated (HCFA)-CDI and this raises

concerns over the risks of “medical tourism”. Increasing reports of antimicrobial

resistance in Asia and proliferation of antimicrobial resistance genes are further

increasing the risks as a result of medical tourism.

5.3.2.2. CDT-positive strains in Asia

Overall, CDT+ strains of C. difficile were rare in the Asian isolate collections (3.6%,

Table 5.8). RT 127 was isolated six times in the 2010 Japanese collection. Further

investigation revealed all the affected patients were in the same hospital ward and their

hospital stays overlapped, suggesting that most likely patient-patient transmission had

occurred. Otherwise RT 127 was only isolated once in the Taiwan subset of isolates.

RT 027 was only identified in the Philippines subset of Asia-Pacific isolates (two

isolates) and the China subset (one isolate). While the total number of Philippines

isolates was only seven, they made up 28.6% the collection. As no pre-existing

information on the molecular epidemiology of C. difficile in the Philippines existed, it is

unclear whether RT 027 causes a significant amount of CDI in the country, or why it

would be present there but not elsewhere in Asia-Pacific countries.

RT 078 was isolated once from Taiwan and once from Korea. As previously discussed,

RT 078 is predominantly found in production animals in Europe although it has been

isolated from animals in the Asia-Pacific region previously. In Japan, RT 078 was

isolated from production pigs, and genetic analysis revealed close relationship to

European pig strains of RT 078161. It was suggested that RT 078 was introduced by

importation of European pigs for meat production. In addition, RT 078 has been

reported in racehorses in Japan160. Again, it is likely that RT 078 was introduced by

importation of racehorses from Europe.

5.3.3. Comparison of Australian and Asian molecular epidemiology

The molecular epidemiology varied widely between Australia and other Asian

countries, and between Asian countries. Australia showed a greater predominance of its

primary strain, RT 014/020. RT 014/020 was found in up to 45% of isolates in the

102

routine testing study (Table 5.5). Nationwide, it varied from 25.5% to 30.0% overall of

all strains in Australia (Tables 5.2, 5.3). Otherwise the molecular types of C. difficile in

Australia were highly variable, with many singleton strains. The Singaporean isolate

collection from the Asia-Pacific study showed very similar results, where RT 014/020

made up 30.4% of strains, otherwise many RTs were represented. The previous

Singapore collection from 2011 was dominated by RT 053 (22.9%, Table 5.9)

suggesting RT 014/020 may have risen in prominence since the first study was

performed in 2011/12, or there may have been some differences in participating

hospitals. Japan was predominated by QX 239 (Figure 5.2). Korea was predominated by

RT 018 (Figure 5.2).

5.3.4. Strengths of the studies described here

This is a large collection of C. difficile isolates with 789 Australian strains and 708

Asian strains described in this chapter (Table 5.1). No previous studies have

amalgamated a collection of Asian isolates of this size. No nationwide studies on

Australian molecular epidemiology of C. difficile had been reported either. These

studies have contributed to a better international understanding of the molecular

epidemiology of C. difficile, and have greatly enhanced our knowledge of C. difficile in

the Asia-Pacific region.

5.3.5. Limitations of studies described here

For the Australian studies, no isolates were received from the Northern Territory (NT).

Therefore, whether the molecular ecology of C. difficile in the NT differs from

elsewhere in Australia remains unknown. Some variability between States and

Territories was apparent from the nationwide surveillance studies so it would be

plausible that the molecular epidemiology of C. difficile could differ in the NT.

However, the population of the NT is quite small (approximately 300,000) and so the

impact of any differences would be minimal.

Some strains given internal nomenclature (QX types) were able to be identified with

UK RT numbers over the course of this PhD, but many were not. Thus they cannot be

compared with RTs reported in other publications. Recently, CE-ribotyping was

established within PathWest Laboratory Medicine WA which allows easy sharing of

ribotyping files with the C. difficile ribotyping network (CDRN) in Leeds University. In

103

the future it will be easier and quicker to resolve QX types of interest with the

internationally recognised RTs.

One limitation of the data here is the high variation in numbers between countries.

Many isolates were obtained from Japan, Korea, Taiwan and Indonesia, while few were

collected from Philippines, Hong Kong, Vietnam. Also the number of sites varied for

countries, which may not have given a sample of isolates which was representative of

the country as a whole.

There may also have been substantial differences in physician awareness and criteria for

testing between sites and countries in Asia. As reported by Mavros et al., physician

awareness of CDI in some countries is limited, particularly in Asia137, which could lead

to minimal testing in some regions, thereby reducing the numbers of isolates available

for analysis. This could lead to under-estimation of the prevalence of some RTs in

particular areas in Asia.

5.3.6. Implications and recommendations

The studies here identified a number of possible routes of transmission of C. difficile

between countries. RT 078 in animals in Asia most likely originated in Europe. RT 017

strains identified in Australia and elsewhere originated in Asia. Medical tourists were

diagnosed with CDI in Australia, infected with strains most likely acquired during their

elective hospital stays in Asian countries. RT 012 strains were identified which were

closely related, despite isolation in multiple Asia-Pacific countries. Ongoing

surveillance of C. difficile in both humans and animals across Asia is required to track

further within-country and international transmission of disease.

Adoption of a standardised protocol for ribotyping, such as that recommended by

Fawley et al.103, would facilitate easy comparison of strains between countries in Asia

and Australia. Whole genome sequencing (WGS) of further isolates of the same RT

identified in multiple countries would help delineate the origins and modes of

transmission of C. difficile strains. Increasing numbers of studies from China are

reporting typing using MLST, which would be a more accessible method for some

laboratories to employ for molecular characterisation of strains and easy comparison

with other studies.

5.3.7. Future studies

104

The prevalence of C. difficile has still not been described for many Asian countries.

Furthermore, incidence rates have not been reported except for Korea and Singapore. A

large-scale surveillance study would be highly informative. A point prevalence study

such as that performed by the EUCLID group in Europe41 would be the most convenient

way to obtain a prevalence rate for many countries and to collect isolates for typing. If

performed regularly, it would be a valuable source of information on the changing

epidemiology of CDI in the region.

Further molecular studies on C. difficile across Asia and Australia will be welcome.

Much is yet to be determined regarding the dynamics of transmission of C. difficile.

Future studies should maintain regular systematic typing of C. difficile isolates from

Asian countries and Australia, by ribotyping, MLST or WGS. WGS studies will become

more affordable and achievable, and will be most useful to determine origins of

outbreaks, clusters of cases and movement of strains between countries and over time.

For example, comparison of RT 018, QX 239, RT 356 and ysmz by WGS would

provide valuable information about how closely related the strains are and their origins.

The high prevalence of A-B+ strains in Asia suggests that they are endemic to Asia,

however, further genomic studies of A-B+ strains from a range of countries are required

to document their phylogenetic evolution.

5.3.8. Chapter summary

• A large collection of almost 1700 isolates from across the Asia-Pacific region

was amassed and typed.

• A high degree of variability in C. difficile strains was observed across the Asia-

Pacific region.

• The most common strain in Asian countries was RT 017 followed by RTs

014/020, 018, 002 and 012, with wide variation between countries.

• In comparison, the most common strain in Australia was RT 014/020 (16-45%

of all strains) followed by RTs 002, 056, 054 and 070.

• CDT+ strains were rarely observed in Asian and Australian isolate collections.

• Animal studies in Australia have identified RTs which are also common among

Australian patients, most frequently RTs 014, 056.

105

CHAPTER 6. GENERAL DISCUSSION

106

C. difficile continues to cause widespread disease, and is a cause of considerable

morbidity and mortality worldwide. Despite mounting research on its epidemiology and

increasing recommendations on infection prevention and control, the incidence rates of

C. difficile infection (CDI) continue to increase in many parts of the world. This may be

due to increased awareness worldwide leading to increasing testing, and adoption of

testing assays with enhanced specificity, or possibly due to adopting tests that are highly

sensitive leading to over-reporting. Nevertheless, C. difficile appears to be one of the

most common nosocomial pathogens in the world, and is increasingly recognised in

community-associated disease also. The mounting evidence of high prevalence among

animals, particularly production animals, and the resulting environmental contamination

show that CDI requires a multi-faceted “One Health” approach to reduce incidence

rates.

The aims of this thesis were to describe disease characteristics, treatment and outcomes

of CDI patients in the Asia-Pacific region, to determine the prevalence of CDI in

Indonesia and Malaysia, to examine the epidemiology of CDI in Australia, and to

describe the molecular epidemiology of C. difficile isolates from Asia-Pacific countries.

These aims were achieved, raising new questions about the epidemiology of CDI in the

Asia-Pacific region. The key findings are discussed below.

6.1. SUMMARY OF FINDINGS

• The prevalence of CDI among diarrhoeal patients in Malaysia was high at 15.4%

and 11.7% in Indonesia, while it was lower in Australia at 6.4%.

• CDI patients in Asia-Pacific countries exhibited many characteristics typical of

CDI patients worldwide - advanced age, extended length of stay (LOS), recent

antibiotic use, multiple comorbidities, and received similar treatment.

• Non-toxigenic strains were commonly isolated in Malaysia and Indonesian

prevalence studies, and made up half of all isolates in those studies, and no

significant differences were found for Malaysian patients infected with toxigenic

strains compared with non-toxigenic strains.

• The prevalence of community associated (CA)-CDI among Australian patients

varied from 7.5% to 44% depending on patient populations studied.

• The most common C. difficile strains in Asian countries were ribotype (RT 017

followed by RTs 014/020, 018, 002 and 012, with wide variation between

107

countries, while the most common strains in Australia were RT 014/020

followed by RTs 002, 056, 054 and 070.

• Binary toxin-positive (CDT+) strains were rarely observed in Asian and

Australian isolate collections, despite being common in other continents.

6.2. THE DESCRIPTIVE EPIDEMIOLOGY OF CDI IN ASIA-PACIFIC

COUNTRIES

The epidemiology of CDI in Australia mirrored that reported elsewhere in Europe and

North America45, 367. Patients were of advanced age, the majority of patients were

female, and patients experienced extended LOS (Tables 3.2-3.6). The increasing

incidence rates of CDI in Australia in recent years warrant continued monitoring15.

Meanwhile the nationwide incidence of CDI in Asian countries remains largely

unknown. The prevalence of CDI among diarrhoeal inpatients in Indonesia (11.1%,

Table 4.1) and Malaysia (15.4%, Table 4.2) was high, around double that found in

Australian inpatients (7.2%, Table 4.1). This indicates that CDI is widespread in these

countries, and many physicians may be largely unaware of the problem.

Asian CDI patients showed similar characteristics to Australian CDI patients. Older age,

recent antibiotic use and recent hospitalisations were widespread. The use of antibiotics

is widespread and uncontrolled in many Asian countries, and antibiotics are freely

available in many countries without prescriptions278, 348, 368. Australia was also reported

as one of the heaviest consumers of antibiotics in the world in 2010281. In general

markers of severe infection were also lower than those identified in studies elsewhere in

the world. Some possible explanations for enhanced virulence in strains elsewhere are

increased germination, sporulation, toxin production and epithelial adherence369.

Mortality rates were low in Australia and other Asian countries (0.8%-3.8%). Deaths

due to CDI have been reported more frequently in North America and Europe10, 13, 45. It

is unclear whether this is due to population differences, differences in the prevalence of

circulating strains, presence of CDT+ strains, or other unidentified factors. A high

prevalence of non-toxigenic strains was identified in the Malaysian and Indonesian

studies. It is likely that non-toxigenic strains are also highly prevalent in neighbouring

countries in Asia. Therapeutic use of toxigenic strains to prevent recurrence has shown

positive results recently. It is possible that the high prevalence of colonisation with non-

toxigenic strains in Asia could provide a protective effect against recurrent and severe

108

CDI and may reduce mortality overall. Further studies are required to explore some of

these possibilities.

6.3. CA-CDI

CA-CDI is reported with increasing frequency worldwide15, 282, suggesting a possible

recent increase in community-based sources of infections. Meanwhile antibiotic

consumption is increasing rapidly worldwide, in both humans and farmed animals53, 281.

Australia, China, India and Singapore were identified among the heaviest consumers of

antibiotics in the world in 2010281. China and India were also among the largest

consumers of antimicrobials in livestock in 201053.

Modelling has shown that CDI incidence rates show seasonal fluctuations, peaking

following the winter influenza season, where antibiotic prescribing rates are highest370.

It is plausible that C. difficile may be propagated among livestock and transmitted via

food, animal contact and spread of manure as fertiliser on crops. Meanwhile community

populations are consuming greater quantities of antibiotics, and other possible risk

drugs including proton pump inhibitors (PPIs). These factors combined would increase

the probability of an individual having a disrupted gut flora and coming into contact

with C. difficile in their daily life. It is plausible that the global increases reported in

CA-CDI may be attributable in some part to such a theory. Vigilance is required among

general practices (GPs) and community-based health practices to identify CDI patients

in the community, and to restrict unnecessary antimicrobial prescriptions.

The emergence of C. difficile strains within the community causing severe disease in

Australia adds to the urgency for research into CDI within the community. A high

prevalence of community onset (CO)-CDI was identified among CDI patients in

Western Australia (WA, 50.8%, Table 3.1) and CA-CDI was identified in 44.2% of

patients presenting in EDs of hospitals in WA (Table 3.8). In the study of emergency

department (ED) patients, age and antibiotic use were significantly associated with CA-

CDI (Table 3.11). The current study and studies in Europe have shown that C. difficile

testing is often neglected among community patients and younger patients, leading to

probable widespread underdiagnosis of CDI371.

In the studies presented here, CA-CDI patients were consistently younger than

healthcare facility associated (HCFA)-CDI patients. Younger patients may not

109

necessarily be tested for CDI by physicians unaware of the increasing rates of CA-CDI.

Moreover, the evidence of diverse arrays of strains causing infections in hospital

patients shows that the epidemiology of transmission of C. difficile is more complex

than previously thought110. Therefore the prevalence of CA-CDI is probably higher than

previously estimated.

6.4. THE MOLECULAR EPIDEMIOLOGY OF C. DIFFICILE IN THE ASIA-

PACIFIC REGION

A large collection of strains was described using molecular techniques in the present

studies. This revealed high diversity in strains across Asia-Pacific countries, with

considerable variation between different regions.

Diversity in strains varied between countries studied in the present study (Chapter 5),

and in previous studies outlined in Chapter 1. Collections of isolates from Australia,

Hong Kong and Singapore showed high diversity in types. These countries have mixed

populations with higher proportions of individuals born overseas than other Asia-Pacific

countries. This could be contributing to the high variation in types identified.

Meanwhile Thailand and China generally showed lower diversity in strains, while the

Japanese collections showed low diversity in general. In Japan, many hospitals require

patients to remain hospitalised until all courses of antibiotics are completed. This

practice may contribute to reducing transmission of C. difficile within communities and

hospitals, and thus lower the overall diversity in molecular types in circulation. It may

also contribute to clusters of cases when infection prevention and control fails.

A number of interesting emerging strains were identified.

RT 018 appears to be widespread in north-eastern Asian countries, particularly in Korea

and Japan. RT 018 is also highly prevalent in Italy. It is interesting that the same RT has

seemingly been established in separate regions of the world. As of now it is unclear how

closely related the strains are. Multi-locus sequence typing (MLST) and single

nucleotide polymorphism (SNP) analysis could determine whether Italian and Korean

and Japanese RT 018 strains are of the same lineage, or if they arose from a common

predecessor. Meanwhile, QX 239 gave a RT pattern similar to the RT pattern of RT

018. This strain was highly prevalent in one hospital in Japan and was significantly

associated with recurrent infection (Figure 5.2). Whether this was due to resistance to

110

metronidazole or vancomycin was not determined and further study is required to

determine the susceptibility of RT 018 and QX 239 to antimicrobials and to determine

their relationship to each other.

C. difficile RT 017 has persisted for decades in Asia192, and was the most frequently

isolated strain in Asian countries. RT 017 was more localised to mainland Asian

countries, and was rarely reported from Japan. It is possible that several lineages of RT

017 exist. RT 017 is consistently reported as sequence type (ST)37 in the literature,

apart from one identification as ST45 in Thailand220. ST45 was reported by Stabler et

al. as a Clade 2 strain, grouping with A+B+ strains, while ST37 is consistently placed in

Clade 4 with other A-B+ strains188, 372. Whether the Thai ST45/RT017 strains are a

distinct lineage, or they were mis-identified as ST45 is unclear. Further genomic studies

would help to clarify the lineages of RT017 in circulation in Asia today.

C. difficile RT 369, another A-B+ strain, was identified as one of the most common

strains in Japan and China in these studies. It was also reported in Japan recently but has

not been described in detail to date. Currently its ST has not been reported.

Interestingly, another A-B+ strain, ST81 has been reported several times in other Asian

studies. It is possible that this is the same strain however MLST is required to confirm

this. If it was confirmed as ST81 further explanations on its emergence and origins

could be derived.

CDT+ C. difficile strains have remained rare in Asian countries, despite the outbreaks of

RTs 027 and 078 in North America and Europe, and of RT 244 in Australia. RT 027

strains have been reported sporadically from Asian countries, the majority of which are

fluoroquinolone-susceptible and belong to a separate lineage from the epidemic strain44.

Despite introduction of the epidemic strain into Australia, RT 027 failed to become

established locally. It is thought that conservative use of fluoroquinolones in Australia

compared to North America and Europe probably contributed to the failure of RT 027

to cause outbreaks on the scale seen in the Northern Hemisphere270. However,

fluoroquinolones are widely used in Asian countries and establishment of the epidemic

strain has not occurred. This is most likely due to a lack of introduction of the epidemic

strain to the region, but the ubiquity in Asia of other strains which are seen frequently

worldwide, such as RTs 014, 020, 001 and 002, indicates a frequent inter-continental

111

transmission of C. difficile strains. Furthermore, RT 017 has been transmitted from Asia

to other regions of the world and caused epidemics.

Why CDT+ strains do not appear to be causing disease in Asian countries is an

interesting question which requires further investigation. CDT+ strains RTs 078, 126

and 127 have been identified among Asian animals which were imported from Europe57,

160, but have only appeared sporadically in humans in Asia to date. A high prevalence of

non-toxigenic C. difficile strains as described in the present studies in Indonesia and

Malaysia could be contributing to prevention of transmission of other strains, including

CDT+ strains. Further exploration of the role of non-toxigenic strains is required to

advance our understanding of the epidemiology of CDI in the Asia-Pacific region.

6.5. STRENGTH AND ORIGINALITY OF STUDIES

Many of the studies presented here were performed in regions of the world where

investigations of C. difficile had not previously been undertaken. Few data existed for

Malaysia and none for Indonesia regarding prevalence, and these are the first

descriptions of the molecular ecology of C. difficile in these countries. No multi-country

survey of CDI had been performed in Asia previously. The study was performed with a

standardised protocol and collected valuable baseline descriptive information on CDI in

Asia.

These studies also accumulated a large collection of C. difficile strains from across the

Asia-Pacific region and have greatly added to the sparse information on the distribution

of types across Asia-Pacific countries. Some countries included did not have any

preceding published molecular data for C. difficile, including Australia, Vietnam, the

Philippines, Malaysia and Indonesia. This isolate collection can be used in many further

studies for genomic analysis to identify sources of transmission, phylogeographic

relationships and evolutionary lineages of strains. The isolates can also be used to

determine antimicrobial susceptibility and provide baseline data for future studies.

The epidemiology of CDI in Australia in particular was not studied in detail prior to the

studies presented here. Nationwide descriptive epidemiological data had not been

collected previously. The molecular epidemiology of Australian C. difficile strains had

not been published prior to these studies. The epidemiology of CA-CDI and CO-CDI

112

was also not well understood previously. These studies have provided important

information on the epidemiology of CDI in Australia.

6.6. LIMITATIONS OF STUDIES

Some isolate collections for particular countries including the Philippines, Vietnam and

Hong Kong were small, so strong conclusions about the molecular types of C. difficile

circulating in these countries could not be drawn. Isolates from India were not received

and so the molecular epidemiology of Indian C. difficile strains remains unknown.

Further studies in these countries are required to enhance our understanding of the

molecular epidemiology of CDI in these regions. However, testing in most of these

countries is likely to be inadequate to identify C. difficile, and anaerobic culture

facilities are lacking in many of these countries. Thus, improving capacity to undertake

investigations into CDI in Asia should be a priority.

The change in standard culture procedures within the laboratory from using cycloserine-

cefoxitin fructose agar (CCFA) to ChromID® may have affected isolation rates.

However, due to the large quantities of isolates amassed, the difference expected would

probably not be significant.

Antimicrobial susceptibility testing was not performed, due to a lack of funding to

perform the analysis, and because it was beyond the scope of the current studies. The

isolate collection amassed here can be employed to perform antimicrobial testing in the

future, however.

6.7. IMPLICATIONS AND RECOMMENDATIONS FOR POLICY

Specific guidelines for physicians in Asia-Pacific countries regarding CDI have not

been published to date, apart from Australasian guidelines64. Currently, one of the

primary issues regarding CDI in Asia is establishment of adequate testing, but this is

required to be within the economic means of the respective country. Many countries

included in the studies presented here are developing nations with poor facilities for

diagnostic laboratories in hospitals. Currently, enzyme immunoassay (EIA) tests are

most likely to be used due to their low cost and simplicity to perform, however their

poor sensitivity means some cases will be undiagnosed. This was demonstrated by the

low sensitivity identified in the Malaysian prevalence study (44%, Table 4.2). The high

specificity of EIAs is however beneficial for differential diagnoses to be made in

113

context with patient history, other test results and their current signs and symptoms.

Otherwise, lower cost in-house polymerase chain reaction (PCR) assays could be

performed in laboratories with the requisite facilities. However, development of more

sensitive toxin EIAs appears to be a more viable option for introduction for use in Asian

countries.

Another highly pertinent goal in Asia-Pacific countries, and one which is achievable, is

implementation of infection prevention and control procedures including introduction of

contact precautions, disinfection of patient rooms and enhanced hand hygiene. Such

control procedures have been shown to be highly effective in reducing CDI incidence in

Asian hospitals to date. Education of hospital staff on awareness of C. difficile and

introduction of contact precaution procedures could greatly reduce further transmission

of C. difficile in Asia-Pacific countries. Antimicrobial stewardship programs would also

significantly reduce CDI incidence within hospitals in Asia-Pacific countries.

In Australia, enhanced vigilance for CDI is required among community-based

physicians. CA-CDI is frequent among community patients so testing must be ordered

to identify and treat CA-CDI patients rapidly. Meanwhile, antimicrobial prescribing in

community patients must be restricted to a minimum, both in Australia and in Asian

countries, to reduce evolutionary pressures which favour proliferation of antimicrobial

resistant organisms.

Meanwhile, a “One Health” approach is required across Asia-Pacific countries to reduce

the prevalence of C. difficile among humans, animals and the environment. Education

on C. difficile is required for Asia-Pacific agriculturalists, food producers and

environmental scientists to try to reduce the transmission of C. difficile within the

community. Further studies of C. difficile in animals, food and environment are required

to determine whether the prevalence is increasing among livestock and the environment

as a result of increasing antimicrobial consumption.

Lastly, introduction of a vaccination for C. difficile would be a highly effective way to

reduce the incidence of C. difficile among hospital patients and community, in Asia-

Pacific countries and worldwide. Currently several candidate vaccines are in

development, some of which are undergoing clinical trials373-375. Since the use of

antimicrobials in livestock is like to be propagating C. difficile, introduction of

114

vaccination of livestock could also contribute to reducing overall prevalence of

C. difficile.

6.8. DIRECTIONS FOR FUTURE RESEARCH

These studies have opened up many possibilities for future research.

Continued study of C. difficile in Asia-Pacific countries is required to monitor outbreaks

and emergence of novel strains or intercontinental transmission of strains. A point

prevalence study of CDI in Asia would be highly informative and would be more

achievable than ongoing surveillance in countries lacking the requisite means. Studies

similar to that performed in Europe by Davies et al.41 would be ideal to perform in Asia

to give an overview of the prevalence of C. difficile at a specific point in time across

many countries. Molecular typing and antimicrobial susceptibility analysis could also be

performed if specimens were collected and processed in a central laboratory with the

required facilities. If point prevalence studies were performed intermittently over

several years they could provide insight about whether education of medical staff,

implementation of infection control precautions or introduction of a vaccination would

reduce the prevalence of C. difficile over time.

Many further studies could be performed using the C. difficile isolates collected in

studies outlined here. Antimicrobial susceptibility testing would be useful to determine

resistance patterns of C. difficile across the Asia-Pacific region. To date, high rates of

resistance to fluoroquinolones in particular have been reported (Table 1.1). It is

important to monitor antimicrobial resistance to determine whether resistance arises to

the drugs currently used for CDI treatment, metronidazole and vancomycin. Two

strains, QX 239 and QX 032, were both significantly associated with recurrent CDI in

the present study, indicating possible resistance to metronidazole or vancomycin. It is

important to determine whether they indeed are more resistant to antibiotics than other

strains.

Genomic analysis of the isolate collection would allow an array of analyses including

MLST, multi-locus variable number of tandem repeats analysis (MLVA) and SNP

analysis to be performed. These would be valuable to determine whether outbreaks of

certain lineages, in particular RTs 017, 369 and 018, are occurring, inter-country and

115

inter-continental transfer of strains and the evolutionary origins of strains. Such

knowledge could be applied to reduce further transmission of strains in the future.

Studies to identify C. difficile in Asian animals and food studies could provide valuable

information on what strains are present in farmed animals and the food chain, and

whether these are closely related to clinical isolates. This would elucidate where

individuals with CA-CDI are being exposed to C. difficile, and would allow policies to

be set in place to reduce transmission of C. difficile within communities and the

environment.

Lastly, the identification of a high prevalence of non-toxigenic strains in Indonesia and

Malaysia warrants further investigation. It is plausible that non-toxigenic strains are

prevalent across many neighbouring Asian countries, but they are rarely reported in the

literature because most laboratory diagnostic tests do not detect them. Epidemiological

analyses could be applied to determine whether their presence was contributing to the

apparent lower mortality and severity of CDI in Asian patients described in the studies

presented here. Since recent clinical trials have shown that administration of non-

toxigenic strains to CDI patients undergoing treatment significantly reduced the

incidence of recurrent CDI353, it is plausible that the non-toxigenic strains among

Malaysian and Indonesian patients could be inducing a similar effect.

6.9. CONCLUSIONS

C. difficile is a ubiquitous multi-drug resistant nosocomial pathogen. It is currently

considered a major threat to hospital patients worldwide, and is increasingly identified

among community members experiencing diarrhoea. It is essential to monitor the

prevalence and circulation of C. difficile strains worldwide. The studies presented here

have provided information on a previously largely under-studied region, Asia-Pacific

countries. The studies identified that C. difficile is highly prevalent in hospital patients

across Asia-Pacific countries, and Asia-Pacific patients bear many characteristics in

common with CDI patients elsewhere in the world, i.e. advanced age, hospitalisation,

and recent antimicrobial use. The molecular types of C. difficile circulating in the region

are diverse and differ to elsewhere in the world, which may contribute to lower rates of

severity and mortality seen in Asia-Pacific CDI patients compared with elsewhere.

These studies have identified that C. difficile requires vigilance among health

116

professionals across the Asia-Pacific region, and have provided many important

baseline data for further studies to expand upon.

117

REFERENCES

1. Hall I, O'Toole E. Intestinal flora in newborn infants. Am J Dis Child. 1935; 49: 390-402. 2. Larson HE, Price AB, Honour P, et al. Clostridium difficile and the aetiology of pseudomembranous colitis. Lancet. 1978; 1: 1063-6. 3. George WL, Sutter VL, Goldstein EJ, et al. Aetiology of antimicrobial-agent- associated colitis. Lancet. 1978; 1: 802-3. 4. Bartlett JG, Moon N, Chang TW, et al. Role of Clostridium difficile in antibiotic-associated pseudomembranous colitis. Gastroenterology. 1978; 75: 778-82. 5. Tedesco FJ, Barton RW, Alpers DH. Clindamycin-associated colitis. A prospective study. Ann Intern Med. 1974; 81: 429-33. 6. Gerding DN, Johnson S, Peterson LR, et al. Clostridium difficile-associated diarrhea and colitis. Infect Control Hosp Epidemiol. 1995; 16: 459-77. 7. McDonald LC, Killgore GE, Thompson A, et al. An epidemic, toxin gene- variant strain of Clostridium difficile. N Engl J Med. 2005; 353: 2433-41. 8. Goorhuis A, Bakker D, Corver J, et al. Emergence of Clostridium difficile infection due to a new hypervirulent strain, polymerase chain reaction ribotype 078. Clin Infect Dis. 2008; 47: 1162-70. 9. Kuijper E, Barbut F, Brazier JS, et al. Update of Clostridium difficile infection due to PCR ribotype 027 in Europe, 2008. Euro Surveill. 2008; 13: 18942. 10. Centers for Disease Control and Prevention (CDC). Antibiotic resistance threats in the United States, 2013. Atlanta: CDC, 2013. 11. Miller BA, Chen LF, Sexton DJ, et al. Comparison of the burdens of hospital- onset, healthcare facility-associated Clostridium difficile infection and of healthcare-associated infection due to methicillin-resistant Staphylococcus

aureus in community hospitals. Infect Control Hosp Epidemiol. 2011; 32: 387- 90. 12. Pepin J, Valiquette L, Alary ME, et al. Clostridium difficile-associated diarrhea in a region of Quebec from 1991 to 2003: a changing pattern of disease severity. CMAJ. 2004; 171: 466-72. 13. Loo VG, Poirier L, Miller MA, et al. A predominantly clonal multi-institutional outbreak of Clostridium difficile-associated diarrhea with high morbidity and mortality. N Engl J Med. 2005; 353: 2442-9. 14. Bouza E. Consequences of Clostridium difficile infection: understanding the healthcare burden. Clin Microbiol Infect. 2012; 18 Suppl 6: 5-12. 15. Slimings C, Armstrong P, Beckingham WD, et al. Increasing incidence of Clostridium difficile infection, Australia, 2011-2012. Med J Aust. 2014; 200: 272-6. 16. Public Health England. Clostridium difficile Ribotyping Network (CDRN) for England and Northern Ireland 2011–2013 report. The National Archives

[online]. 2014; http://webarchive.nationalarchives.gov.uk/20140714084352/http://www.hpa.org .uk/webw/HPAweb&HPAwebStandard/HPAweb_C/1317140658987. 17. Spigaglia P, Barbanti F, Dionisi AM, et al. Clostridium difficile isolates resistant to fluoroquinolones in Italy: emergence of PCR ribotype 018. J Clin Microbiol. 2010; 48: 2892-6.

118

18. Kuijper EJ, de Weerdt J, Kato H, et al. Nosocomial outbreak of Clostridium

difficile-associated diarrhoea due to a clindamycin-resistant enterotoxin A- negative strain. Eur J Clin Microbiol Infect Dis. 2001; 20: 528-34. 19. Rupnik M, Wilcox MH, Gerding DN. Clostridium difficile infection: new developments in epidemiology and pathogenesis. Nat Rev Microbiol. 2009; 7: 526-36. 20. Freeman J, Bauer MP, Baines SD, et al. The changing epidemiology of Clostridium difficile infections. Clin Microbiol Rev. 2010; 23: 529-49. 21. Walker AS, Eyre DW, Wyllie DH, et al. Characterisation of Clostridium difficile

hospital ward-based transmission using extensive epidemiological data and molecular typing. PLoS Med. 2012; 9: e1001172. 22. Heinlen L, Ballard JD. Clostridium difficile infection. Am J Med Sci. 2010; 340: 247-52. 23. van den Berg RJ, Claas EC, Oyib DH, et al. Characterization of toxin A- negative, toxin B-positive Clostridium difficile isolates from outbreaks in different countries by amplified fragment length polymorphism and PCR ribotyping. J Clin Microbiol. 2004; 42: 1035-41. 24. Rupnik M. Heterogeneity of large clostridial toxins: importance of Clostridium

difficile toxinotypes. FEMS Microbiol Rev. 2008; 32: 541-55. 25. Elliott B, Squire MM, Thean S, et al. New types of toxin A-negative, toxin B- positive strains among clinical isolates of Clostridium difficile in Australia. J

Med Microbiol. 2011; 60: 1108-11. 26. Akerlund T, Alefjord I, Dohnhammar U, et al. Geographical clustering of cases of infection with moxifloxacin-resistant Clostridium difficile PCR-ribotypes 012, 017 and 046 in Sweden, 2008 and 2009. Euro Surveill. 2011; 16: 19813. 27. Arvand M, Hauri AM, Zaiss NH, et al. Clostridium difficile ribotypes 001, 017, and 027 are associated with lethal C. difficile infection in Hesse, Germany. Euro

Surveill. 2009; 14. 28. Rupnik M, Kato N, Grabnar M, et al. New types of toxin A-negative, toxin B- positive strains among Clostridium difficile isolates from Asia. J Clin Microbiol. 2003; 41: 1118-25. 29. al-Barrak A, Embil J, Dyck B, et al. An outbreak of toxin A negative, toxin B positive Clostridium difficile-associated diarrhea in a Canadian tertiary-care hospital. Can Commun Dis Rep. 1999; 25: 65-9. 30. Komatsu M, Kato H, Aihara M, et al. High frequency of antibiotic-associated diarrhea due to toxin A-negative, toxin B-positive Clostridium difficile in a hospital in Japan and risk factors for infection. Eur J Clin Microbiol Infect Dis. 2003; 22: 525-9. 31. Rupnik M, Grabnar M, Geric B. Binary toxin producing Clostridium difficile strains. Anaerobe. 2003; 9: 289-94. 32. Popoff MR, Rubin EJ, Gill DM, et al. Actin-specific ADP-ribosyltransferase produced by a Clostridium difficile strain. Infect Immun. 1988; 56: 2299-306. 33. Perelle S, Gibert M, Bourlioux P, et al. Production of a complete binary toxin (actin-specific ADP-ribosyltransferase) by Clostridium difficile CD196. Infect

Immun. 1997; 65: 1402-7. 34. Goorhuis A, Debast SB, van Leengoed LA, et al. Clostridium difficile PCR ribotype 078: an emerging strain in humans and in pigs? J Clin Microbiol. 2008; 46: 1157; Author reply 58. 35. Bartlett JG. Clinical practice. Antibiotic-associated diarrhea. N Engl J Med. 2002; 346: 334-9.

119

36. Ackermann G, Thomalla S, Ackermann F, et al. Prevalence and characteristics of bacteria and host factors in an outbreak situation of antibiotic-associated diarrhoea. J Med Microbiol. 2005; 54: 149-53. 37. Friedman ND, Pollard J, Stupart D, et al. Prevalence of Clostridium difficile

colonization among healthcare workers. BMC Infect Dis. 2013; 13: 459. 38. Matsuki S, Ozaki E, Shozu M, et al. Colonization by Clostridium difficile of neonates in a hospital, and infants and children in three day-care facilities of Kanazawa, Japan. Int Microbiol. 2005; 8: 43-8. 39. Hensgens MP, Keessen EC, Squire MM, et al. Clostridium difficile infection in the community: a zoonotic disease? Clin Microbiol Infect. 2012; 18: 635-45. 40. McDonald LC, Coignard B, Dubberke E, et al. Recommendations for surveillance of Clostridium difficile-associated disease. Infect Control Hosp

Epidemiol. 2007; 28: 140-5. 41. Davies KA, Longshaw CM, Davis GL, et al. Underdiagnosis of Clostridium

difficile across Europe: the European, multicentre, prospective, biannual, point- prevalence study of Clostridium difficile infection in hospitalised patients with diarrhoea (EUCLID). Lancet Infect Dis. 2014; 14: 1208-19. 42. Lessa FC, Mu Y, Bamberg WM, et al. Burden of Clostridium difficile infection in the United States. N Engl J Med. 2015; 372: 825-34. 43. Clements AC, Magalhaes RJ, Tatem AJ, et al. Clostridium difficile PCR ribotype 027: assessing the risks of further worldwide spread. Lancet Infect Dis. 2010; 10: 395-404. 44. He M, Miyajima F, Roberts P, et al. Emergence and global spread of epidemic healthcare-associated Clostridium difficile. Nat Genet. 2013; 45: 109-13. 45. Bauer MP, Notermans DW, van Benthem BH, et al. Clostridium difficile infection in Europe: a hospital-based survey. Lancet. 2011; 377: 63-73. 46. Keel K, Brazier JS, Post KW, et al. Prevalence of PCR ribotypes among Clostridium difficile isolates from pigs, calves, and other species. J Clin

Microbiol. 2007; 45: 1963-4. 47. De Almeida MN, Heffernan H, Dervan A, et al. Severe Clostridium difficile infection in New Zealand associated with an emerging strain, PCR-ribotype 244. N Z Med J. 2013; 126: 9-14. 48. Lim SK, Stuart RL, Mackin KE, et al. Emergence of a ribotype 244 strain of Clostridium difficile associated with severe disease and related to the epidemic ribotype 027 strain. Clin Infect Dis. 2014; 58: 1723-30. 49. Khanna S, Pardi DS, Aronson SL, et al. The epidemiology of community- acquired Clostridium difficile infection: a population-based study. Am J

Gastroenterol. 2012; 107: 89-95. 50. Chitnis AS, Holzbauer SM, Belflower RM, et al. Epidemiology of community- associated Clostridium difficile infection, 2009 through 2011. JAMA Intern Med. 2013; 173: 1359-67. 51. Biswal S. Proton pump inhibitors and risk for Clostridium difficile associated diarrhea. Biomed J. 2014; 37: 178-83. 52. Naggie S, Miller BA, Zuzak KB, et al. A case-control study of community- associated Clostridium difficile infection: no role for proton pump inhibitors. Am J Med. 2011; 124: 276 e1-7. 53. Van Boeckel TP, Brower C, Gilbert M, et al. Global trends in antimicrobial use in food animals. Proc Natl Acad Sci U S A. 2015; 112: 5649-54.

120

54. Knight DR, Riley TV. Prevalence of gastrointestinal Clostridium difficile

carriage in Australian sheep and lambs. Appl Environ Microbiol. 2013; 79: 5689-92. 55. Knight DR, Squire MM, Riley TV. Laboratory detection of Clostridium difficile in piglets in Australia. J Clin Microbiol. 2014; 52: 3856-62. 56. Knight DR, Thean S, Putsathit P, et al. Cross-sectional study reveals high prevalence of Clostridium difficile non-PCR ribotype 078 strains in Australian veal calves at slaughter. Appl Environ Microbiol. 2013; 79: 2630-5. 57. Wu YC, Lee JJ, Tsai BY, et al. Potentially hypervirulent Clostridium difficile PCR ribotype 078 lineage isolates in pigs and possible implications for humans in Taiwan. Int J Med Microbiol. 2016; 306: 115-22. 58. Goldova J, Malinova A, Indra A, et al. Clostridium difficile in piglets in the Czech Republic. Folia Microbiol (Praha). 2012; 57: 159-61. 59. Noren T, Johansson K, Unemo M. Clostridium difficile PCR ribotype 046 is common among neonatal pigs and humans in Sweden. Clin Microbiol Infect. 2014; 20: O2-6. 60. Knetsch CW, Connor TR, Mutreja A, et al. Whole genome sequencing reveals potential spread of Clostridium difficile between humans and farm animals in the Netherlands, 2002 to 2011. Euro Surveill. 2014; 19: 20954. 61. Rodriguez-Palacios A, Stampfli HR, Duffield T, et al. Clostridium difficile PCR ribotypes in calves, Canada. Emerg Infect Dis. 2006; 12: 1730-6. 62. Metcalf DS, Costa MC, Dew WM, et al. Clostridium difficile in vegetables, Canada. Lett Appl Microbiol. 2010; 51: 600-2. 63. Songer JG, Trinh HT, Killgore GE, et al. Clostridium difficile in retail meat products, USA, 2007. Emerg Infect Dis. 2009; 15: 819-21. 64. Trubiano JA, Cheng AC, Korman TM, et al. Australasian Society of Infectious Diseases updated guidelines for the management of Clostridium difficile infection in adults and children in Australia and New Zealand. Intern Med J. 2016; 46: 479-93. 65. Thomas C, Riley TV. Restriction of third generation cephalosporin use reduces the incidence of Clostridium difficile-associated diarrhoea in hospitalised patients. Commun Dis Intell Q Rep. 2003; 27 Suppl: S28-31. 66. Suzuki H, Senda J, Yamashita K, et al. Impact of intensive infection control team activities on the acquisition of methicillin-resistant Staphylococcus aureus, drug-resistant Pseudomonas aeruginosa and the incidence of Clostridium

difficile-associated disease. J Infect Chemother. 2013; 19: 1047-52. 67. Bowman RA, Riley TV. Laboratory diagnosis of Clostridium difficile-associated diarrhoea. Eur J Clin Microbiol Infect Dis. 1988; 7: 476-84. 68. Wilkins TD, Lyerly DM. Clostridium difficile testing: after 20 years, still challenging. J Clin Microbiol. 2003; 41: 531-4. 69. Planche TD, Davies KA, Coen PG, et al. Differences in outcome according to Clostridium difficile testing method: a prospective multicentre diagnostic validation study of C difficile infection. Lancet Infect Dis. 2013; 13: 936-45. 70. Brecher SM, Novak-Weekley SM, Nagy E. Laboratory diagnosis of Clostridium

difficile infections: there is light at the end of the colon. Clin Infect Dis. 2013; 57: 1175-81. 71. Burnham CA, Carroll KC. Diagnosis of Clostridium difficile infection: an ongoing conundrum for clinicians and for clinical laboratories. Clin Microbiol

Rev. 2013; 26: 604-30.

121

72. Wren B, Clayton C, Tabaqchali S. Rapid identification of toxigenic Clostridium

difficile by polymerase chain reaction. Lancet. 1990; 335: 423. 73. Gumerlock PH, Tang YJ, Weiss JB, et al. Specific detection of toxigenic strains of Clostridium difficile in stool specimens. J Clin Microbiol. 1993; 31: 507-11. 74. Kato N, Ou CY, Kato H, et al. Identification of toxigenic Clostridium difficile by the polymerase chain reaction. J Clin Microbiol. 1991; 29: 33-7. 75. MacCannell DR, Louie TJ, Gregson DB, et al. Molecular analysis of Clostridium difficile PCR ribotype 027 isolates from Eastern and Western Canada. J Clin Microbiol. 2006; 44: 2147-52. 76. Drudy D, Kyne L, O'Mahony R, et al. gyrA mutations in fluoroquinolone- resistant Clostridium difficile PCR-027. Emerg Infect Dis. 2007; 13: 504-5. 77. Drudy D, Harnedy N, Fanning S, et al. Emergence and control of fluoroquinolone-resistant, toxin A-negative, toxin B-positive Clostridium

difficile. Infect Control Hosp Epidemiol. 2007; 28: 932-40. 78. Carroll KC, Buchan BW, Tan S, et al. Multicenter evaluation of the Verigene Clostridium difficile nucleic acid assay. J Clin Microbiol. 2013; 51: 4120-5. 79. Dalpke AH, Hofko M, Zorn M, et al. Evaluation of the fully automated BD MAX Cdiff and Xpert C. difficile assays for direct detection of Clostridium

difficile in stool specimens. J Clin Microbiol. 2013; 51: 1906-8. 80. Putsathit P, Morgan J, Bradford D, et al. Evaluation of the BD Max Cdiff assay for the detection of toxigenic Clostridium difficile in human stool specimens. Pathology. 2015; 47: 165-8. 81. Buchan BW, Mackey TL, Daly JA, et al. Multicenter clinical evaluation of the Portrait toxigenic C. difficile assay for detection of toxigenic Clostridium

difficile strains in clinical stool specimens. J Clin Microbiol. 2012; 50: 3932-6. 82. Walkty A, Lagace-Wiens PR, Manickam K, et al. Evaluation of an algorithmic approach in comparison with the illumigene assay for laboratory diagnosis of Clostridium difficile infection. J Clin Microbiol. 2013; 51: 1152-7. 83. Gilbreath JJ, Verma P, Abbott AN, et al. Comparison of the Verigene

Clostridium difficile, Simplexa C. difficile Universal Direct, BD MAX Cdiff, and Xpert C. difficile assays for the detection of toxigenic C. difficile. Diagn

Microbiol Infect Dis. 2014; 80: 13-8. 84. McElgunn CJ, Pereira CR, Parham NJ, et al. A low complexity rapid molecular method for detection of Clostridium difficile in stool. PLoS ONE. 2014; 9: e83808. 85. Hong G, Park KS, Ki CS, et al. Evaluation of the illumigene C. difficile assay for toxigenic Clostridium difficile detection: a prospective study of 302 consecutive clinical fecal samples. Diagn Microbiol Infect Dis. 2014; 80: 177-80. 86. Androga GO, McGovern AM, Elliott B, et al. Evaluation of the Cepheid Xpert C. difficile/Epi and meridian bioscience illumigene C. difficile assays for detecting Clostridium difficile ribotype 033 strains. J Clin Microbiol. 2015; 53: 973-5. 87. Dubberke ER, Han Z, Bobo L, et al. Impact of clinical symptoms on interpretation of diagnostic assays for Clostridium difficile infections. J Clin

Microbiol. 2011; 49: 2887-93. 88. Cohen SH, Gerding DN, Johnson S, et al. Clinical practice guidelines for Clostridium difficile infection in adults: 2010 update by the Society for Healthcare Epidemiology of America (SHEA) and the Infectious Diseases Society of America (IDSA). Infect Control Hosp Epidemiol. 2010; 31: 431-55.

122

89. Gould CV, Edwards JR, Cohen J, et al. Effect of nucleic acid amplification testing on population-based incidence rates of Clostridium difficile infection. Clin Infect Dis. 2013; 57: 1304-7. 90. Tenover FC, Novak-Weekley S, Woods CW, et al. Impact of strain type on detection of toxigenic Clostridium difficile: comparison of molecular diagnostic and enzyme immunoassay approaches. J Clin Microbiol. 2010; 48: 3719-24. 91. Goldenberg SD, Gumban M, Hall A, et al. Lack of effect of strain type on detection of toxigenic Clostridium difficile by glutamate dehydrogenase and polymerase chain reaction. Diagn Microbiol Infect Dis. 2011; 70: 417-9. 92. Peterson LR, Robicsek A. Does my patient have Clostridium difficile infection? Ann Intern Med. 2009; 151: 176-9. 93. Tenover FC, Baron EJ, Peterson LR, et al. Laboratory diagnosis of Clostridium

difficile infection can molecular amplification methods move us out of uncertainty? J Mol Diagn. 2011; 13: 573-82. 94. Larson AM, Fung AM, Fang FC. Evaluation of tcdB real-time PCR in a three- step diagnostic algorithm for detection of toxigenic Clostridium difficile. J Clin

Microbiol. 2010; 48: 124-30. 95. Babady NE, Stiles J, Ruggiero P, et al. Evaluation of the Cepheid Xpert Clostridium difficile Epi assay for diagnosis of Clostridium difficile infection and typing of the NAP1 strain at a cancer hospital. J Clin Microbiol. 2010; 48: 4519-24. 96. Crobach MJ, Dekkers OM, Wilcox MH, et al. European Society of Clinical Microbiology and Infectious Diseases (ESCMID): data review and recommendations for diagnosing Clostridium difficile-infection (CDI). Clin

Microbiol Infect. 2009; 15: 1053-66. 97. Wilcox MH, Planche T, Fang FC, et al. What is the current role of algorithmic approaches for diagnosis of Clostridium difficile infection? J Clin Microbiol. 2010; 48: 4347-53. 98. Baker I, Leeming JP, Reynolds R, et al. Clinical relevance of a positive molecular test in the diagnosis of Clostridium difficile infection. J Hosp Infect. 2013; 84: 311-5. 99. Polage CR, Chin DL, Leslie JL, et al. Outcomes in patients tested for Clostridium difficile toxins. Diagn Microbiol Infect Dis. 2012; 74: 369-73. 100. Longtin Y, Trottier S, Brochu G, et al. Impact of the type of diagnostic assay on Clostridium difficile infection and complication rates in a mandatory reporting program. Clin Infect Dis. 2013; 56: 67-73. 101. Janezic S, Rupnik M. Molecular typing methods for Clostridium difficile : Pulsed-field gel electrophoresis and PCR ribotyping. In: Mullany P, Roberts AP, editors. Methods Mol Biol: Humana Press, 2010; 55-65. 102. O'Neill G, Ogunsola FT, JS B, et al. Modification of a PCR ribotyping method for application as a routine typing scheme for Clostridium difficile. Anaerobe. 1996; 2: 205-09. 103. Fawley WN, Knetsch CW, MacCannell DR, et al. Development and validation of an internationally-standardized, high-resolution capillary gel-based electrophoresis PCR-ribotyping protocol for Clostridium difficile. PLoS ONE. 2015; 10: e0118150. 104. Alonso R, Martin A, Pelaez T, et al. An improved protocol for pulsed-field gel electrophoresis typing of Clostridium difficile. J Med Microbiol. 2005; 54: 155- 7.

123

105. Lemee L, Dhalluin A, Pestel-Caron M, et al. Multilocus sequence typing analysis of human and animal Clostridium difficile isolates of various toxigenic types. J Clin Microbiol. 2004; 42: 2609-17. 106. Griffiths D, Fawley W, Kachrimanidou M, et al. Multilocus sequence typing of Clostridium difficile. J Clin Microbiol. 2010; 48: 770-8. 107. Dingle KE, Elliott B, Robinson E, et al. Evolutionary history of the Clostridium

difficile pathogenicity locus. Genome Biol Evol. 2014; 6: 36-52. 108. Marsh JW, O'Leary MM, Shutt KA, et al. Multilocus variable-number tandem- repeat analysis for investigation of Clostridium difficile transmission in hospitals. J Clin Microbiol. 2006; 44: 2558-66. 109. Fawley WN, Wilcox MH. An enhanced DNA fingerprinting service to investigate potential Clostridium difficile infection case clusters sharing the same PCR ribotype. J Clin Microbiol. 2011; 49: 4333-7. 110. Eyre DW, Cule ML, Wilson DJ, et al. Diverse sources of C. difficile infection identified on whole-genome sequencing. N Engl J Med. 2013; 369: 1195-205. 111. Olsen MA, Young-Xu Y, Stwalley D, et al. The burden of Clostridium difficile infection: estimates of the incidence of CDI from U.S. Administrative databases. BMC Infect Dis. 2016; 16: 177. 112. Tenover FC, Akerlund T, Gerding DN, et al. Comparison of strain typing results for Clostridium difficile isolates from North America. J Clin Microbiol. 2011; 49: 1831-7. 113. Tickler IA, Goering RV, Whitmore JD, et al. Strain types and antimicrobial resistance patterns of Clostridium difficile isolates from the United States, 2011 to 2013. Antimicrob Agents Chemother. 2014; 58: 4214-8. 114. Sheitoyan-Pesant C, Abou Chakra CN, Pepin J, et al. Clinical and healthcare burden of multiple recurrences of Clostridium difficile infection. Clin Infect Dis. 2016; 62: 574-80. 115. Office for National Statistics. Deaths involving Clostridium difficile: England and Wales: 2012. Available at: http://wwwonsgovuk/peoplepopulationandcommunity/birthsdeathsandmarriag es/deaths/bulletins/deathsinvolvingclostridiumdifficileenglandandwales/2013- 08-22. 2012; Accessed 27 April 2016. 116. Janezic S, Strumbelj I, Rupnik M. Use of modified PCR ribotyping for direct detection of Clostridium difficile ribotypes in stool samples. J Clin Microbiol. 2011; 49: 3024-5. 117. Ahmetagic S, Salkic N, Ahmetagic A, et al. Clostridium difficile infection in hospitalized patients at university clinical center Tuzla, Bosnia and Herzegovina: a 4 year experience. Mater Sociomed. 2013; 25: 153-7. 118. Alzahrani N, Johani SA. Emergence of a highly resistant Clostridium difficile strain (NAP/BI/027) in a tertiary care center in Saudi Arabia. Ann Saudi Med. 2013; 33: 198-9. 119. Moukhaiber R, Araj GF, Kissoyan KA, et al. Prevalence of Clostridium difficile toxinotypes in infected patients at a tertiary care center in Lebanon. J Infect Dev

Ctries. 2015; 9: 732-5. 120. Balassiano IT, Dos Santos-Filho J, de Oliveira MP, et al. An outbreak case of Clostridium difficile-associated diarrhea among elderly inpatients of an intensive care unit of a tertiary hospital in Rio de Janeiro, Brazil. Diagn Microbiol Infect

Dis. 2010; 68: 449-55.

124

121. Garcia C, Samalvides F, Vidal M, et al. Epidemiology of Clostridium difficile- associated diarrhea in a Peruvian tertiary care hospital. Am J Trop Med Hyg. 2007; 77: 802-5. 122. Lopardo G, Morfin-Otero R, Moran V, II, et al. Epidemiology of Clostridium

difficile: a hospital-based descriptive study in Argentina and Mexico. Braz J

Infect Dis. 2015; 19: 8-14. 123. Goorhuis A, Legaria MC, Van Den Berg RJ, et al. Application of multiple-locus variable-number tandem-repeat analysis to determine clonal spread of toxin A- negative Clostridium difficile in a general hospital in Buenos Aires, Argentina. Clin Microbiol Infect. 2009; 15: 1080-86. 124. Plaza-Garrido A, Barra-Carrasco J, Macias JH, et al. Predominance of Clostridium difficile ribotypes 012, 027 and 046 in a university hospital in Chile, 2012. Epidemiol Infect. 2016; 144: 976-9. 125. Hernández-Rocha C, Barra-Carrasco J, Pizarro-Guajardo M, et al. Epidemic Clostridium difficile ribotype 027 in Chile. Emerg Infect Dis. 2012; 18: 1370-72. 126. Monteiro Ade A, Pires RN, Persson S, et al. A search for Clostridium difficile ribotypes 027 and 078 in Brazil. Braz J Infect Dis. 2014; 18: 672-4. 127. Balassiano IT, Yates EA, Domingues RMCP, et al. Clostridium difficile: a problem of concern in developed countries and still a mystery in Latin America. J Med Microbiol. 2012; 61: 169-79. 128. Secco DA, Balassiano IT, Boente RF, et al. Clostridium difficile infection among immunocompromised patients in Rio de Janeiro, Brazil and detection of moxifloxacin resistance in a ribotype 014 strain. Anaerobe. 2014; 28: 85-9. 129. Silva RO, Rupnik M, Diniz AN, et al. Clostridium difficile ribotypes in humans and animals in Brazil. Mem Inst Oswaldo Cruz. 2015; 110: 1062-5. 130. Quesada-Gomez C, Rodriguez C, Gamboa-Coronado Mdel M, et al. Emergence of Clostridium difficile NAP1 in Latin America. J Clin Microbiol. 2010; 48: 669-70. 131. Quesada-Gómez C, López-Ureña D, Acuña-Amador L, et al. Emergence of an outbreak-associated Clostridium difficile variant with increased virulence. J Clin

Microbiol. 2015; 53: 1216-26. 132. Seugendo M, Mshana SE, Hokororo A, et al. Clostridium difficile infections among adults and children in Mwanza/Tanzania: is it an underappreciated pathogen among immunocompromised patients in sub-Saharan Africa? New

Microbes New Infect. 2015; 8: 99-102. 133. Onwueme K, Fadairo Y, Idoko L, et al. High prevalence of toxinogenic

Clostridium difficile in Nigerian adult HIV patients. Trans R Soc Trop Med Hyg. 2011; 105: 667-9. 134. Simango C, Uladi S. Detection of Clostridium difficile diarrhoea in Harare, Zimbabwe. Trans R Soc Trop Med Hyg. 2014; 108: 354-7. 135. Kullin B, Meggersee R, D'Alton J, et al. Prevalence of gastrointestinal pathogenic bacteria in patients with diarrhoea attending Groote Schuur Hospital, Cape Town, South Africa. S Afr Med J. 2015; 105: 121-5. 136. Rajabally N, Kullin B, Ebrahim K, et al. A comparison of Clostridium difficile

diagnostic methods for identification of local strains in a South African centre. J

Med Microbiol. 2016; 65: 320-27. 137. Mavros MN, Alexiou VG, Vardakas KZ, et al. Underestimation of Clostridium

difficile infection among clinicians: an international survey. Eur J Clin

Microbiol Infect Dis. 2012; 31: 2439-44.

125

138. Oyofo BA, Subekti D, Tjaniadi P, et al. Enteropathogens associated with acute diarrhea in community and hospital patients in Jakarta, Indonesia. FEMS

Immunol Med Microbiol. 2002; 34: 139-46. 139. Honda H, Yamazaki A, Sato Y, et al. Incidence and mortality associated with

Clostridium difficile infection at a Japanese tertiary care center. Anaerobe. 2014; 25: 5-10. 140. Aihara M, Sakai M, Iwasaki M, et al. [Isolation of enteropathogenic microorganism from patients with infection of the digestive tract during 1976 to 1988 in Tenri Hospital]. Kansenshogaku Zasshi. 1991; 65: 864-74. 141. Kobayashi T. [Studies on Clostridium difficile and antimicrobial associated diarrhea or colitis]. Jpn J Antibiot. 1983; 36: 464-76. 142. Shimizu H, Mori M, Yoshimoto N. Clostridium difficile infection is more severe when toxin is detected in the stool than when detected only by a toxigenic culture. Intern Med. 2015; 54: 2155-9. 143. Mori N, Aoki Y. Clinical characteristics and risk factors for community- acquired Clostridium difficile infection: A retrospective, case-control study in a tertiary care hospital in Japan. J Infect Chemother. 2015; 21: 864-7. 144. Yoshida J, Kikuchi T, Asano I, et al. Clostridium difficile infection relationship to ward-monthly antimicrobial use density and days of treatment: a three-year study. Japanese Journal of Infection Prevention and Control. 2016; 31: 92-99. 145. Kato H, Kato N, Katow S, et al. Deletions in the repeating sequences of the toxin A gene of toxin A-negative, toxin B-positive Clostridium difficile strains. FEMS Microbiol Lett. 1999; 175: 197-203. 146. Kato H, Kato N, Watanabe K, et al. Identification of toxin A-negative, toxin B- positive Clostridium difficile by PCR. J Clin Microbiol. 1998; 36: 2178-82. 147. Kato H, Kato N, Watanabe K, et al. Application of typing by pulsed-field gel electrophoresis to the study of Clostridium difficile in a neonatal intensive care unit. J Clin Microbiol. 1994; 32: 2067-70. 148. Kato H, Kato N, Watanabe K, et al. Analysis of Clostridium difficile isolates from nosocomial outbreaks at three hospitals in diverse areas of Japan. J Clin

Microbiol. 2001; 39: 1391-5. 149. Kato H, Kita H, Karasawa T, et al. Colonisation and transmission of Clostridium

difficile in healthy individuals examined by PCR ribotyping and pulsed-field gel electrophoresis. J Med Microbiol. 2001; 50: 720-7. 150. Kato H, Yokoyama T, Arakawa Y. Typing by sequencing the slpA gene of Clostridium difficile strains causing multiple outbreaks in Japan. J Med

Microbiol. 2005; 54: 167-71. 151. Kikkawa H, Hitomi S, Watanabe M. Prevalence of toxin A-nonproducing/toxin- B-producing Clostridium difficile in the Tsukuba-Tsuchiura district, Japan. J

Infect Chemother. 2007; 13: 35-8. 152. Iwashima Y, Nakamura A, Kato H, et al. A retrospective study of the epidemiology of Clostridium difficile infection at a University Hospital in Japan: genotypic features of the isolates and clinical characteristics of the patients. J

Infect Chemother. 2010; 16: 329-33. 153. Sawabe E, Kato H, Osawa K, et al. Molecular analysis of Clostridium difficile at a university teaching hospital in Japan: a shift in the predominant type over a five-year period. Eur J Clin Microbiol Infect Dis. 2007; 26: 695-703. 154. Kato H, Kato H, Ito Y, et al. Typing of Clostridium difficile isolates endemic in Japan by sequencing of slpA and its application to direct typing. J Med

Microbiol. 2010; 59: 556-62.

126

155. Kato H, Kato H, Nakamura M, et al. Rapid analysis of Clostridium difficile

strains recovered from hospitalized patients by using the slpA sequence typing system. J Infect Chemother. 2009; 15: 199-202. 156. Kuwata Y, Tanimoto S, Sawabe E, et al. Molecular epidemiology and antimicrobial susceptibility of Clostridium difficile isolated from a university teaching hospital in Japan. Eur J Clin Microbiol Infect Dis. 2015; 34: 763-72. 157. Kim J, Kang JO, Kim H, et al. Epidemiology of Clostridium difficile infections in a tertiary-care hospital in Korea. Clin Microbiol Infect. 2013; 19: 521-7. 158. Joost I, Speck K, Herrmann M, et al. Characterisation of Clostridium difficile

isolates by slpA and tcdC gene sequencing. Int J Antimicrob Agents. 2009; 33 Suppl 1: S13-8. 159. Senoh M, Kato H, Fukuda T, et al. Predominance of PCR-ribotypes, 018 (smz) and 369 (trf) of Clostridium difficile in Japan: a potential relationship with other global circulating strains? J Med Microbiol. 2015; 64: 1226-36. 160. Niwa H, Kato H, Hobo S, et al. Postoperative Clostridium difficile infection with PCR ribotype 078 strain identified at necropsy in five Thoroughbred racehorses. Vet Rec. 2013; 173: 607. 161. Usui M, Nanbu Y, Oka K, et al. Genetic relatedness between Japanese and European isolates of Clostridium difficile originating from piglets and their risk associated with human health. Front Microbiol. 2014; 5: 513. 162. Lee YJ, Choi MG, Lim CH, et al. [Change of Clostridium difficile colitis during recent 10 years in Korea]. Korean J Gastroenterol. 2010; 55: 169-74. 163. Kim YS, Han DS, Kim YH, et al. Incidence and clinical features of Clostridium

difficile infection in Korea: a nationwide study. Epidemiol Infect. 2013; 141: 189-94. 164. Ryu HS, Kim YS, Seo GS, et al. Risk factors for recurrent Clostridium difficile infection. Intest Res. 2012; 10: 176-82. 165. Kim YG, Graham DY, Jang BI. Proton pump inhibitor use and recurrent Clostridium difficile-associated disease: a case-control analysis matched by propensity score. J Clin Gastroenterol. 2012; 46: 397-400. 166. Choi HK, Kim KH, Lee SH, et al. Risk factors for recurrence of Clostridium

difficile infection: effect of vancomycin-resistant Enterococci colonization. J

Korean Med Sci. 2011; 26: 859-64. 167. Shin BM, Moon SJ, Kim YS, et al. Characterization of cases of Clostridium

difficile infection (CDI) presenting at an emergency room: molecular and clinical features differentiate community-onset hospital-associated and community-associated CDI in a tertiary care hospital. J Clin Microbiol. 2011; 49: 2161-5. 168. Kim YA, Rim JH, Choi MH, et al. Increase of Clostridium difficile in community; another worrisome burden for public health. Ann Clin Microbiol. 2016; 19: 7-12. 169. Kim H, Riley TV, Kim M, et al. Increasing prevalence of toxin A-negative, toxin B-positive isolates of Clostridium difficile in Korea: impact on laboratory diagnosis. J Clin Microbiol. 2008; 46: 1116-7. 170. Shin BM, Kuak EY, Yoo HM, et al. Multicentre study of the prevalence of toxigenic Clostridium difficile in Korea: results of a retrospective study 2000- 2005. J Med Microbiol. 2008; 57: 697-701. 171. Shin BM, Kuak EY, Yoo SJ, et al. Emerging toxin A-B+ variant strain of Clostridium difficile responsible for pseudomembranous colitis at a tertiary care hospital in Korea. Diagn Microbiol Infect Dis. 2008; 60: 333-7.

127

172. Shin BM, Lee EJ, Kuak EY, et al. Comparison of VIDAS CDAB and CDA immunoassay for the detection of Clostridium difficile in a tcdA- tcdB+ C.

difficile prevalent area. Anaerobe. 2009; 15: 266-9. 173. Kim H, Jeong SH, Roh KH, et al. Investigation of toxin gene diversity, molecular epidemiology, and antimicrobial resistance of Clostridium difficile

isolated from 12 hospitals in South Korea. Korean J Lab Med. 2010; 30: 491-7. 174. Kim SJ, Kim H, Seo Y, et al. Molecular characterization of toxin A-negative, toxin B-positive variant strains of Clostridium difficile isolated in Korea. Diagn

Microbiol Infect Dis. 2010; 67: 198-201. 175. Cho SY, Nam Y, Soh YS, et al. To the Editor: Molecular epidemiology of toxigenic Clostridium difficile isolates in Korea. Epidemiol Infect. 2014; 142: 1651-2. 176. Han SH, Kim H, Lee K, et al. Epidemiology and clinical features of toxigenic culture-confirmed hospital-onset Clostridium difficile infection: a multicentre prospective study in tertiary hospitals of South Korea. J Med Microbiol. 2014; 63: 1542-51. 177. Kim J, Seo MR, Kang JO, et al. Clinical characteristics of relapses and re- infections in Clostridium difficile infection. Clin Microbiol Infect. 2014; 20: 1198-204. 178. Lee JH, Lee Y, Lee K, et al. The changes of PCR ribotype and antimicrobial resistance of Clostridium difficile in a tertiary care hospital over 10 years. J Med

Microbiol. 2014; 63: 819-23. 179. Kim J, Kang JO, Pai H, et al. Association between PCR ribotypes and antimicrobial susceptibility among Clostridium difficile isolates from healthcare- associated infections in South Korea. Int J Antimicrob Agents. 2012; 40: 24-9. 180. Kim H, Lee Y, Moon HW, et al. Emergence of Clostridium difficile ribotype 027 in Korea. Korean J Lab Med. 2011; 31: 191-6. 181. Jin K, Wang S, Huang Z, et al. Clostridium difficile infections in China. J

Biomed Res. 2010; 24: 411-6. 182. Huang H, Wu S, Wang M, et al. Molecular and clinical characteristics of Clostridium difficile infection in a University Hospital in Shanghai, China. Clin

Infect Dis. 2008; 47: 1606-8. 183. Huang H, Fang H, Weintraub A, et al. Distinct ribotypes and rates of antimicrobial drug resistance in Clostridium difficile from Shanghai and Stockholm. Clin Microbiol Infect. 2009; 15: 1170-3. 184. Huang H, Wu S, Wang M, et al. Clostridium difficile infections in a Shanghai hospital: antimicrobial resistance, toxin profiles and ribotypes. Int J Antimicrob

Agents. 2009; 33: 339-42. 185. Zhou FF, Wu S, Klena JD, et al. Clinical characteristics of Clostridium difficile infection in hospitalized patients with antibiotic-associated diarrhea in a university hospital in China. Eur J Clin Microbiol Infect Dis. 2014; 33: 1773-9. 186. Cheng JW, Xiao M, Kudinha T, et al. The role of glutamate dehydrogenase (GDH) testing assay in the diagnosis of Clostridium difficile infections: a high sensitive screening test and an essential step in the proposed laboratory diagnosis workflow for developing countries like China. PLoS ONE. 2015; 10: e0144604. 187. Tian TT, Zhao JH, Yang J, et al. Molecular characterization of Clostridium

difficile isolates from human subjects and the environment. PLoS ONE. 2016; 11: e0151964.

128

188. Dingle KE, Griffiths D, Didelot X, et al. Clinical Clostridium difficile: clonality and pathogenicity locus diversity. PLoS ONE. 2011; 6: e19993. 189. Dong D, Peng Y, Zhang L, et al. Clinical and microbiological characterization of Clostridium difficile infection in a tertiary care hospital in Shanghai, China. Chin Med J (Engl). 2014; 127: 1601-7. 190. Hawkey PM, Marriott C, Liu WE, et al. Molecular epidemiology of Clostridium

difficile infection in a major Chinese hospital: an underrecognized problem in Asia? J Clin Microbiol. 2013; 51: 3308-13. 191. Yan Q, Zhang J, Chen C, et al. Multilocus sequence typing (MLST) analysis of 104 Clostridium difficile strains isolated from China. Epidemiol Infect. 2013; 141: 195-9. 192. Du P, Cao B, Wang J, et al. Sequence variation in tcdA and tcdB of Clostridium

difficile: ST37 with truncated tcdA is a potential epidemic strain in China. J Clin

Microbiol. 2014; 52: 3264-70. 193. Chen Y, Rashid MU, Huang H, et al. Molecular characteristics of Clostridium

difficile strains from patients with a first recurrence more than 8 weeks after the primary infection. J Microbiol Immunol Infect. 2015. 194. Cheng JW, Xiao M, Kudinha T, et al. The first two Clostridium difficile ribotype 027/ST1 isolates identified in Beijing, China-an emerging problem or a neglected threat? Sci Rep. 2016; 6: 18834. 195. Wang P, Zhou Y, Wang Z, et al. Identification of Clostridium difficile ribotype 027 for the first time in Mainland China. Infect Control Hosp Epidemiol. 2014; 35: 95-8. 196. Jin H, Ni K, Wei L, et al. Identification of Clostridium difficile RT078 from patients and environmental surfaces in Zhejiang Province, China. Infect Control

Hosp Epidemiol. 2016; 37: 745-6. 197. Huang H, Weintraub A, Fang H, et al. Antimicrobial susceptibility and heteroresistance in Chinese Clostridium difficile strains. Anaerobe. 2010; 16: 633-5. 198. Chen YB, Gu SL, Wei ZQ, et al. Molecular epidemiology of Clostridium

difficile in a tertiary hospital of China. J Med Microbiol. 2014; 63: 562-9. 199. Chang SC, Su CH, Chou WH, et al. Epidemiology of nosocomial Clostridium

difficile disease in Taiwan. Taiwan Epidemiology Bulletin. 2009; 25: 163-77. 200. Chan YY, Lin TY, Huang CT, et al. Implementation and outcomes of a hospital- wide computerised antimicrobial stewardship programme in a large medical centre in Taiwan. Int J Antimicrob Agents. 2011; 38: 486-92. 201. Lee YC, Wang JT, Chen AC, et al. Changing incidence and clinical manifestations of Clostridium difficile-associated diarrhea detected by combination of glutamate dehydrogenase and toxin assay in Northern Taiwan. J

Microbiol Immunol Infect. 2012; 45: 287-95. 202. Ji D, Huang IH, Lai CC, et al. Prevalence and characterization of enterotoxigenic Bacteroides fragilis and toxigenic Clostridium difficile in a Taipei emergency department. J Microbiol Immunol Infect. 2014. 203. Lee HC, Kim KO, Jeong YH, et al. Clinical outcomes in hospitalized patients with Clostridium difficile infection by age group. Korean J Gastroenterol. 2016; 67: 81-6. 204. Hung YP, Lin HJ, Wu TC, et al. Risk factors of fecal toxigenic or non-toxigenic Clostridium difficile colonization: impact of Toll-like receptor polymorphisms and prior antibiotic exposure. PLoS ONE. 2013; 8: e69577.

129

205. Wei HL, Kao CW, Wei SH, et al. Comparison of PCR ribotyping and multilocus variable-number tandem-repeat analysis (MLVA) for improved detection of Clostridium difficile. BMC Microbiol. 2011; 11: 217. 206. Liao TL, Lin CF, Chiou CS, et al. Clostridium difficile PCR ribotype 027 emerges in Taiwan. Jpn J Infect Dis. 2015; 68: 338-40. 207. Lai MJ, Chiueh TS, Huang ZY, et al. The first Clostridium difficile ribotype 027 strain isolated in Taiwan. J Formos Med Assoc. 2016; 115: 210-2. 208. Hung YP, Cia CT, Tsai BY, et al. The first case of severe Clostridium difficile ribotype 027 infection in Taiwan. J Infect. 2015; 70: 98-101. 209. Hung YP, Lin HJ, Tsai BY, et al. Clostridium difficile ribotype 126 in southern Taiwan: a cluster of three symptomatic cases. Anaerobe. 2014; 30: 188-92. 210. Yuen KY, Woo PC, Liang RH, et al. Clinical significance of alimentary tract microbes in bone marrow transplant recipients. Diagn Microbiol Infect Dis. 1998; 30: 75-81. 211. Wong SS, Ho PL, Woo PC, et al. Bacteremia caused by Staphylococci with inducible vancomycin heteroresistance. Clin Infect Dis. 1999; 29: 760-7. 212. Cheng VC, Yam WC, Chan JF, et al. Clostridium difficile ribotype 027 arrives in Hong Kong. Int J Antimicrob Agents. 2009; 34: 492-3. 213. Lee JS, Auyeung TW. A comparison of two feeding methods in the alleviation of diarrhoea in older tube-fed patients: a randomised controlled trial. Age

Ageing. 2003; 32: 388-93. 214. Chang KC, Leung CC, Yew WW, et al. Analyses of fluoroquinolones and Clostridium difficile-associated diarrhoea in tuberculosis patients. Int J Tuberc

Lung Dis. 2009; 13: 341-6. 215. Wong SH, Ip M, Hawkey PM, et al. High morbidity and mortality of Clostridium difficile infection and its associations with ribotype 002 in Hong Kong. J Infect. 2016; 73: 115-22. 216. Cheng VC, Yam WC, Lam OT, et al. Clostridium difficile isolates with increased sporulation: emergence of PCR ribotype 002 in Hong Kong. Eur J

Clin Microbiol Infect Dis. 2011; 30: 1371-81. 217. Cheng VC, Chau PH, So SY, et al. Containment of Clostridium difficile

infection without reduction in antimicrobial use in Hong Kong. Eur J Clin

Microbiol Infect Dis. 2015; 34: 1381-6. 218. Wong SH, Ip M, Hawkey PM, et al. High morbidity and mortality of Clostridium difficile infection and its associations with ribotype 002 in Hong Kong. J Infect. 2016. 219. Chiang D, Ng S, La MV, et al. Performance assessment of the BD MAX Cdiff assay in comparison to Xpert C. difficile assay in a setting with very low prevalence of toxigenic Clostridium difficile PCR ribotype 027. Anaerobe. 2014; 30: 156-8. 220. Ngamskulrungroj P, Sanmee S, Putsathit P, et al. Molecular epidemiology of Clostridium difficile infection in a large teaching hospital in Thailand. PLoS

ONE. 2015; 10: e0127026. 221. Lam TS, Yuk MT, Tsang NC, et al. Clostridium difficile infection outbreak in a male rehabilitation ward, Hong Kong Special Administrative Region (China), 2011. Western Pac Surveill Response J. 2012; 3: 59-60. 222. Warren CA, Labio E, Destura R, et al. Clostridium difficile and Entamoeba

histolytica infections in patients with colitis in the Philippines. Trans R Soc Trop

Med Hyg. 2012; 106: 424-8.

130

223. Gutierrez AA, Alcantara CS, Alejandria M. The incidence of Clostridium

difficile-associated diarrhea in the adult intensive care units of the UP-Philippine General Hospital: a case-control study. Int J Antimicrob Agents. 2005; 26: S102- S02. 224. Wongwanich S, Ramsiri S, Vanasin B, et al. Clostridium difficile associated disease in Thailand. Southeast Asian J Trop Med Public Health. 1990; 21: 367- 72. 225. Wongwanich S, Pongpech P, Dhiraputra C, et al. Characteristics of Clostridium

difficile strains isolated from asymptomatic individuals and from diarrheal patients. Clin Microbiol Infect. 2001; 7: 438-41. 226. Wongwanich S, Rugdeekha S, Pongpech P, et al. Detection of Clostridium

difficile toxin A and B genes from stool samples of Thai diarrheal patients by polymerase chain reaction technique. J Med Assoc Thai. 2003; 86: 970-5. 227. ThamLikitkul V, Danpakdi K, Chokloikaew S. Incidence of diarrhea and Clostridium difficile toxin in stools from hospitalized patients receiving clindamycin, beta-lactams, or nonantibiotic medications. J Clin Gastroenterol. 1996; 22: 161-3. 228. Siristonpun Y, Tanphaichitra D. Value of Clostridium difficile antigen test in immunocompromised hosts in the tropics. J Chemother. 1991; 3 Suppl 1: 73-5. 229. Chotiprasitsakul D, Janvilisri T, Kiertiburanakul S, et al. A superior test for diagnosis of Clostridium difficile-associated diarrhea in resource-limited settings. Jpn J Infect Dis. 2012; 65: 326-9. 230. Issarachaikull R, Khantipong M, Sawatpanich A, et al. Prospective evaluation of a novel two-step protocol for screening of Clostridium difficile infection in hospitalized adult patients. Southeast Asian J Trop Med Public Health. 2015; 46: 1037-48. 231. Wongwanich S, Ramsiri S, Kusum M, et al. Clostridium difficile infections in HIV-positive patients. Southeast Asian J Trop Med Public Health. 2000; 31: 537-9. 232. Waywa D, Kongkriengdaj S, Chaidatch S, et al. Protozoan enteric infection in AIDS related diarrhea in Thailand. Southeast Asian J Trop Med Public Health. 2001; 32 Suppl 2: 151-5. 233. Parasakthi N, Puthucheary SD, Goh KL, et al. Clostridium difficile associated diarrhoea: a report of seven cases. Singapore Med J. 1988; 29: 504-7. 234. Hassan H, Cheng SC. Detection of Clostridium difficile from an enhancement broth by gas-liquid chromatography. Malays J Pathol. 1991; 13: 29-31. 235. Boey CC, Ramanujam TM, Looi LM. Clostridium difficile-related necrotizing pseudomembranous enteritis in association with Henoch-Schonlein purpura. J

Pediatr Gastroenterol Nutr. 1997; 24: 426-9. 236. Hassan SA, Othman N, Idris FM, et al. Prevalence of Clostridium difficile toxin in diarhoeal stool samples of patients from a tertiary hospital in North Eastern Penisular Malaysia. Med J Malaysia. 2012; 67: 402-5. 237. Hassan SA, Rahman RA, Huda N, et al. Hospital-acquired Clostridium difficile

infection among patients with type 2 diabetes mellitus in acute medical wards. J

R Coll Physicians Edinb. 2013; 43: 103-7. 238. Syuhada N, Azimatun N, Alfizah H, et al. Clostridium difficile infection: Clinico-epidemiological perspective. Med & Health. 2013; 8: 64-72. 239. Kurniawan A, Lugito NPH, Yanto TA, et al. Clostridium difficile infection in elderly hepatocellular carcinoma patients in general hospital, Karawaci, Tangerang, Banten, Indonesia. J Geriatr Oncol. 2014; 5: S17.

131

240. Koh TH, Tan AL, Tan ML, et al. Epidemiology of Clostridium difficile infection in a large teaching hospital in Singapore. Pathology. 2007; 39: 438-42. 241. Lim PL, Barkham TM, Ling LM, et al. Increasing incidence of Clostridium

difficile-associated disease, Singapore. Emerg Infect Dis. 2008; 14: 1487-9. 242. Hsu LY, Tan TY, Koh TH, et al. Decline in Clostridium difficile-associated disease rates in Singapore public hospitals, 2006 to 2008. BMC Res Notes. 2011; 4: 77. 243. Lim PL, Ling ML, Lee HY, et al. Isolation of the first three cases of Clostridium

difficile polymerase chain reaction ribotype 027 in Singapore. Singapore Med J. 2011; 52: 361-4. 244. Chhin S, Harwell JI, Bell JD, et al. Etiology of chronic diarrhea in antiretroviral- naive patients with HIV infection admitted to Norodom Sihanouk Hospital, Phnom Penh, Cambodia. Clin Infect Dis. 2006; 43: 925-32. 245. Duong VT, Phat VV, Tuyen HT, et al. Evaluation of Luminex xTAG gastrointestinal pathogen panel assay for detection of multiple diarrheal pathogens in fecal samples in Vietnam. J Clin Microbiol. 2016; 54: 1094-100. 246. Gupta U, Yadav RN. Clostridium difficile in hospital patients. Indian J Med Res. 1985; 82: 398-401. 247. Ayyagari A, Sharma P, Venkateswarlu, et al. Prevalence of Clostridium difficile in pseudomembranous and antibiotic-associated colitis in north India. J

Diarrhoeal Dis Res. 1986; 4: 157-60. 248. Bhattacharya MK, Niyogi SK, Rasaily R, et al. Clinical manifestation of Clostridium difficile enteritis in Calcutta. J Assoc Physicians India. 1991; 39: 683-4. 249. Niyogi SK, Bhattacharya SK, Dutta P, et al. Prevalence of Clostridium difficile in hospitalised patients with acute diarrhoea in Calcutta. J Diarrhoeal Dis Res. 1991; 9: 16-9. 250. Dhawan B, Chaudhry R, Sharma N. Incidence of Clostridium difficile infection: a prospective study in an Indian hospital. J Hosp Infect. 1999; 43: 275-80. 251. Chaudhry R, Joshy L, Kumar L, et al. Changing pattern of Clostridium difficile

associated diarrhoea in a tertiary care hospital: a 5 year retrospective study. Indian J Med Res. 2008; 127: 377-82. 252. Ingle M, Deshmukh A, Desai D, et al. Prevalence and clinical course of Clostridium difficile infection in a tertiary-care hospital: a retrospective analysis. Indian J Gastroenterol. 2011; 30: 89-93. 253. Vaishnavi C, Singh M. Preliminary investigation of environmental prevalence of Clostridium difficile affecting inpatients in a north Indian hospital. Indian J Med

Microbiol. 2012; 30: 89-92. 254. Niyogi SK, Dutta P, Dutta D, et al. Clostridium difficile and its cytotoxin in hospitalized children with acute diarrhea. Indian Pediatr. 1991; 28: 1129-32. 255. Gogate A, De A, Nanivadekar R, et al. Diagnostic role of stool culture & toxin detection in antibiotic associated diarrhoea due to Clostridium difficile in children. Indian J Med Res. 2005; 122: 518-24. 256. Chandra BK, Singh G, Taneja N, et al. Diarrhoeagenic Escherichia coli as a predominant cause of paediatric nosocomial diarrhoea in India. J Med

Microbiol. 2012; 61: 830-6. 257. Uppal B, Kashyap B, Bhalla P. Enteric pathogens in HIV/ AIDS from a tertiary care hospital. Indian J Community Med. 2009; 34: 237-42.

132

258. Ingle M, Deshmukh A, Desai D, et al. Clostridium difficile as a cause of acute diarrhea: a prospective study in a tertiary care center. Indian J Gastroenterol. 2013; 32: 179-83. 259. Vaishnavi C, Singh M, Mahmood S, et al. Prevalence and molecular types of Clostridium difficile isolates from faecal specimens of patients in a tertiary care centre. J Med Microbiol. 2015; 64: 1297-304. 260. Justin S, Antony B, Shenoy KV, et al. Prevalence of Clostridium difficile among paediatric patients in a tertiary care hospital, coastal Karnataka, South India. J

Clin Diagn Res. 2015; 9: DC04-7. 261. Hussain I, Sharma RK, Borah P, et al. Isolation and characterization of Clostridium difficile from pet dogs in Assam, India. Anaerobe. 2015; 36: 9-13. 262. Akhtar SQ. Isolation of Clostridium difficile from diarrhoea patients in Bangladesh. J Trop Med Hyg. 1987; 90: 189-92. 263. Albert MJ, Faruque AS, Faruque SM, et al. Case-control study of enteropathogens associated with childhood diarrhea in Dhaka, Bangladesh. J

Clin Microbiol. 1999; 37: 3458-64. 264. Haider Naqvi SA, Chaudhry FF. Clostridium difficile postantibiotic diarrhoea diagnosis. J Coll Physicians Surg Pak. 2012; 22: 640-3. 265. Riley TV, Bowman RA, Carroll SM. Diarrhoea associated with Clostridium

difficile in a hospital population. Med J Aust. 1983; 1: 166-9. 266. Riley TV, O'Neill GL, Bowman RA, et al. Clostridium difficile-associated diarrhoea: epidemiological data from Western Australia. Epidemiol Infect. 1994; 113: 13-20. 267. Thomas C, Stevenson M, Williamson DJ, et al. Clostridium difficile-associated diarrhea: epidemiological data from Western Australia associated with a modified antibiotic policy. Clin Infect Dis. 2002; 35: 1457-62. 268. Riley TV, Thean S, Hool G, et al. First Australian isolation of epidemic Clostridium difficile PCR ribotype 027. Med J Aust. 2009; 190: 706-8. 269. Richards M, Knox J, Elliott B, et al. Severe infection with Clostridium difficile

PCR ribotype 027 acquired in Melbourne, Australia. Med J Aust. 2011; 194: 369-71. 270. Cheng AC, Turnidge J, Collignon P, et al. Control of fluoroquinolone resistance through successful regulation, Australia. Emerg Infect Dis. 2012; 18: 1453-60. 271. Collignon P. Fluoroquinolone use in food animals. Emerg Infect Dis. 2005; 11: 1789-92. 272. Bull AL, Worth LJ, Richards MJ. Implementation of standardised surveillance for Clostridium difficile infections in Australia: initial report from the Victorian Healthcare Associated Infection Surveillance System. Intern Med J. 2012; 42: 715-8. 273. Mitchell BG, Wilson F, McGregor A. An increase in community onset Clostridium difficile infection: a population-based study, Tasmania, Australia. Healthcare Infection. 2012; 17: 127-32. 274. Knight DR, Squire MM, Riley TV. Nationwide surveillance study of Clostridium difficile in Australian neonatal pigs shows high prevalence and heterogeneity of PCR ribotypes. Appl Environ Microbiol. 2015; 81: 119-23. 275. Kato H, Ito Y, van den Berg RJ, et al. First isolation of Clostridium difficile 027 in Japan. Euro Surveill. 2007; 12: E070111 3. 276. You E, Song H, Cho J, et al. Reduction in the incidence of hospital-acquired Clostridium difficile infection through infection control interventions other than the restriction of antimicrobial use. Int J Infect Dis. 2014; 22: 9-10.

133

277. Chen W, Liu L. Future population trends in China: 2005-2050. The Centre of

Policy Studies (COPS). Monash University, Australia 2009. 278. Holloway KA. Promoting the rational use of antibiotics. Regional Health

Forum; WHO South East Asia Region. 2011; 15: 122-30. 279. Hvistendahl M. Public health. China takes aim at rampant antibiotic resistance. Science. 2012; 336: 795. 280. Hadi U, van den Broek P, Kolopaking EP, et al. Cross-sectional study of availability and pharmaceutical quality of antibiotics requested with or without prescription (Over The Counter) in Surabaya, Indonesia. BMC Infect Dis. 2010; 10: 203. 281. Van Boeckel TP, Gandra S, Ashok A, et al. Global antibiotic consumption 2000 to 2010: an analysis of national pharmaceutical sales data. Lancet Infect Dis. 2014; 14: 742-50. 282. Khanna S, Pardi DS. "Community-acquired Clostridium difficile infection: an emerging entity". Clin Infect Dis. 2012; 55: 1741-2. 283. Huang H, Weintraub A, Fang H, et al. Antimicrobial resistance in Clostridium

difficile. Int J Antimicrob Agents. 2009; 34: 516-22. 284. Zhu S, Zhang H, Zhang X, et al. Investigation of toxin gene diversity and antimicrobial resistance of Clostridium difficile strains. Biomed Rep. 2014; 2: 743-48. 285. Dong D, Zhang L, Chen X, et al. Antimicrobial susceptibility and resistance mechanisms of clinical Clostridium difficile from a Chinese tertiary hospital. Int

J Antimicrob Agents. 2013; 41: 80-4. 286. Lidan C, Linhai L, Yang L, et al. Molecular characterization and antimicrobial susceptibility of tcdA-negative Clostridium difficile isolates from Guangzhou, China. Diagn Microbiol Infect Dis. 2016; 84: 361-5. 287. Oka K, Osaki T, Hanawa T, et al. Molecular and microbiological characterization of Clostridium difficile isolates from single, relapse, and reinfection cases. J Clin Microbiol. 2012; 50: 915-21. 288. Kunishima H, Chiba J, Saito M, et al. Antimicrobial susceptibilities of Clostridium difficile isolated in Japan. J Infect Chemother. 2013; 19: 360-2. 289. Kim J, Pai H, Seo M-r, et al. Clinical and microbiologic characteristics of tcdA- negative variant Clostridium difficile infections. BMC Infect Dis. 2012; 12: 109. 290. Kim J, Seo MR, Kang JO, et al. Clinical and microbiologic characteristics of Clostridium difficile infection caused by binary toxin producing strains in Korea. Infect Chemother. 2013; 45: 175-83. 291. Lin YC, Huang YT, Tsai PJ, et al. Antimicrobial susceptibilities and molecular epidemiology of clinical isolates of Clostridium difficile in Taiwan. Antimicrob

Agents Chemother. 2011; 55: 1701-5. 292. Nam KC, Jo C, Lee M. Meat products and consumption culture in the East. Meat Sci. 2010; 86: 95-102. 293. Coker RJ, Hunter BM, Rudge JW, et al. Emerging infectious diseases in southeast Asia: regional challenges to control. Lancet. 2011; 377: 599-609. 294. Carson KC, Boseiwaqa LV, Thean SK, et al. Isolation of Clostridium difficile from faecal specimens--a comparison of ChromID C. difficile agar and cycloserine-cefoxitin-fructose agar. J Med Microbiol. 2013; 62: 1423-7. 295. Stubbs S, Rupnik M, Gibert M, et al. Production of actin-specific ADP- ribosyltransferase (binary toxin) by strains of Clostridium difficile. FEMS

Microbiol Lett. 2000; 186: 307-12.

134

296. Stubbs SLJ, Brazier JS, O’Neill GL, et al. PCR targeted to the 16S-23S rRNA gene intergenic spacer region of Clostridium difficile and construction of a library consisting of 116 different PCR ribotypes. J Clin Microbiol. 1999; 37: 461-63. 297. Australian Bureau of Statistics. 2011 Census of Population and Housing. Canberra 2011. 298. Bowman RA, Riley TV. Routine culturing for Clostridium difficile? Pathology. 1984; 16: 240-2. 299. Zheng L, Keller SF, Lyerly DM, et al. Multicenter evaluation of a new screening test that detects Clostridium difficile in fecal specimens. J Clin Microbiol. 2004; 42: 3837-40. 300. Terhes G, Urban E, Soki J, et al. Comparison of a rapid molecular method, the BD GeneOhm Cdiff assay, to the most frequently used laboratory tests for detection of toxin-producing Clostridium difficile in diarrheal feces. J Clin

Microbiol. 2009; 47: 3478-81. 301. Le Guern R, Herwegh S, Grandbastien B, et al. Evaluation of a new molecular test, the BD Max Cdiff, for detection of toxigenic Clostridium difficile in fecal samples. J Clin Microbiol. 2012; 50: 3089-90. 302. Cheng AC, Collins DA, Elliott B, et al. Laboratory-based surveillance of Clostridium difficile circulating in Australia, September - November 2010. Pathology. 2016; 48: 257-60. 303. Tan XQ, Verrall AJ, Jureen R, et al. The emergence of community-onset Clostridium difficile infection in a tertiary hospital in Singapore: a cause for concern. Int J Antimicrob Agents. 2014; 43: 47-51. 304. Mori N, Yoshizawa S, Saga T, et al. Incorrect diagnosis of Clostridium difficile infection in a university hospital in Japan. J Infect Chemother. 2015; 21: 718-22. 305. Foster NF, Collins DA, Ditchburn SL, et al. Epidemiology of Clostridium

difficile infection in two tertiary-care hospitals in Perth, Western Australia: a cross-sectional study. New Microbes New Infect. 2014; 2: 64-71. 306. Khanna S, Pardi DS, Aronson SL, et al. Outcomes in community-acquired Clostridium difficile infection. Aliment Pharmacol Ther. 2012; 35: 613-8. 307. Collins DA, Hawkey PM, Riley TV. Epidemiology of Clostridium difficile infection in Asia. Antimicrob Resist Infect Control. 2013; 2: 21. 308. Polage CR, Gyorke CE, Kennedy MA, et al. Overdiagnosis of Clostridium

difficile infection in the molecular test era. JAMA Intern Med. 2015; 175: 1792- 801. 309. Hart J, Putsathit P, Knight DR, et al. Clostridium difficile infection diagnosis in a paediatric population: comparison of methodologies. Eur J Clin Microbiol

Infect Dis. 2014; 33: 1555-64. 310. Zilberberg MD, Tillotson GS, McDonald C. Clostridium difficile infections among hospitalized children, United States, 1997-2006. Emerg Infect Dis. 2010; 16: 604-9. 311. Kim J, Smathers SA, Prasad P, et al. Epidemiological features of Clostridium

difficile-associated disease among inpatients at children's hospitals in the United States, 2001-2006. Pediatrics. 2008; 122: 1266-70. 312. Nylund CM, Goudie A, Garza JM, et al. Clostridium difficile infection in hospitalized children in the United States. Arch Pediatr Adolesc Med. 2011; 165: 451-7.

135

313. de Graaf H, Pai S, Burns DA, et al. Co-infection as a confounder for the role of Clostridium difficile infection in children with diarrhoea: a summary of the literature. Eur J Clin Microbiol Infect Dis. 2015; 34: 1281-7. 314. Eglow R, Pothoulakis C, Itzkowitz S, et al. Diminished Clostridium difficile

toxin A sensitivity in newborn rabbit ileum is associated with decreased toxin A receptor. J Clin Invest. 1992; 90: 822-9. 315. Kociolek LK, Patel SJ, Zheng X, et al. Clinical and microbiologic assessment of cases of pediatric community-associated Clostridium difficile infection reveals opportunities for improved testing decisions. Pediatr Infect Dis J. 2016; 35: 157- 61. 316. Tracey L, Kirke A, Armstrong P, et al. From the hospital to the home - the rise and rise of Clostridium difficile infection. Aust Fam Physician. 2015; 44: 712-7. 317. Alcala L, Reigadas E, Marin M, et al. Impact of clinical awareness and diagnostic tests on the underdiagnosis of Clostridium difficile infection. Eur J

Clin Microbiol Infect Dis. 2015; 34: 1515-25. 318. Alcala L, Martin A, Marin M, et al. The undiagnosed cases of Clostridium

difficile infection in a whole nation: where is the problem? Clin Microbiol Infect. 2012; 18: E204-13. 319. Bauer MP, Veenendaal D, Verhoef L, et al. Clinical and microbiological characteristics of community-onset Clostridium difficile infection in The Netherlands. Clin Microbiol Infect. 2009; 15: 1087-92. 320. Khanna S, Gupta A, Baddour LM, et al. Epidemiology, outcomes, and predictors of mortality in hospitalized adults with Clostridium difficile infection. Intern Emerg Med. 2016; 11: 657-65. 321. Burckhardt F, Friedrich A, Beier D, et al. Clostridium difficile surveillance trends, Saxony, Germany. Emerg Infect Dis. 2008; 14: 691-2. 322. Leffler DA, Lamont JT. Clostridium difficile infection. N Engl J Med. 2015; 372: 1539-48. 323. Nicholson MR, Thomsen IP, Slaughter JC, et al. Novel risk factors for recurrent Clostridium difficile infection in children. J Pediatr Gastroenterol Nutr. 2015; 60: 18-22. 324. Kelly CP, LaMont JT. Clostridium difficile-more difficult than ever. N Engl J

Med. 2008; 359: 1932-40. 325. Pepin J, Alary ME, Valiquette L, et al. Increasing risk of relapse after treatment of Clostridium difficile colitis in Quebec, Canada. Clin Infect Dis. 2005; 40: 1591-7. 326. Figueroa I, Johnson S, Sambol SP, et al. Relapse versus reinfection: recurrent Clostridium difficile infection following treatment with fidaxomicin or vancomycin. Clin Infect Dis. 2012; 55 Suppl 2: S104-9. 327. Olsen MA, Yan Y, Reske KA, et al. Recurrent Clostridium difficile infection is associated with increased mortality. Clin Microbiol Infect. 2015; 21: 164-70. 328. Leonard AD, Ho KM, Flexman J. Proton pump inhibitors and diarrhoea related to Clostridium difficile infection in hospitalised patients: a case-control study. Intern Med J. 2012; 42: 591-4. 329. Halim HA, Peterson GM, Friesen WT, et al. Case-controlled review of Clostridium difficile-associated diarrhoea in southern Tasmania. J Clin Pharm

Ther. 1997; 22: 391-7. 330. Slimings C, Riley TV. Antibiotics and hospital-acquired Clostridium difficile

infection: update of systematic review and meta-analysis. J Antimicrob

Chemother. 2014; 69: 881-91.

136

331. Stevens V, Dumyati G, Fine LS, et al. Cumulative antibiotic exposures over time and the risk of Clostridium difficile infection. Clin Infect Dis. 2011; 53: 42- 8. 332. CadLe RM, Mansouri MD, Logan N, et al. Association of proton-pump inhibitors with outcomes in Clostridium difficile colitis. Am J Health Syst

Pharm. 2007; 64: 2359-63. 333. Imhann F, Bonder MJ, Vich Vila A, et al. Proton pump inhibitors affect the gut microbiome. Gut. 2015; 65: 740-8. 334. Dial S, Delaney JAC, Schneider V, et al. Proton pump inhibitor use and risk of community-acquired Clostridium difficile-associated disease defined by prescription for oral vancomycin therapy. CMAJ. 2006; 175: 745-48. 335. Wilcox MH, Mooney L, Bendall R, et al. A case-control study of community- associated Clostridium difficile infection. J Antimicrob Chemother. 2008; 62: 388-96. 336. Bolnick DI, Snowberg LK, Hirsch PE, et al. Individual diet has sex-dependent effects on vertebrate gut microbiota. Nat Commun. 2014; 5: 4500. 337. Schnorr SL, Candela M, Rampelli S, et al. Gut microbiome of the Hadza hunter- gatherers. Nat Commun. 2014; 5: 3654. 338. Riley TV, Cooper M, Bell B, et al. Community-acquired Clostridium difficile- associated diarrhea. Clin Infect Dis. 1995; 20 Suppl 2: S263-5. 339. Andrews CN, Raboud J, Kassen BO, et al. Clostridium difficile-associated diarrhea: predictors of severity in patients presenting to the emergency department. Can J Gastroenterol. 2003; 17: 369-73. 340. Thean S, Elliott B, Riley TV. Clostridium difficile in horses in Australia--a preliminary study. J Med Microbiol. 2011; 60: 1188-92. 341. Rodriguez-Palacios A, Borgmann S, Kline TR, et al. Clostridium difficile in foods and animals: history and measures to reduce exposure. Anim Health Res

Rev. 2013; 14: 11-29. 342. de Boer E, Zwartkruis-Nahuis A, Heuvelink AE, et al. Prevalence of Clostridium difficile in retailed meat in the Netherlands. Int J Food Microbiol. 2011; 144: 561-4. 343. Weese JS, Avery BP, Rousseau J, et al. Detection and enumeration of Clostridium difficile spores in retail beef and pork. Appl Environ Microbiol. 2009; 75: 5009-11. 344. Squire MM, Riley TV. Clostridium difficile infection in humans and piglets: a 'One Health' opportunity. Curr Top Microbiol Immunol. 2013; 365: 299-314. 345. Burns K, Morris-Downes M, Fawley WN, et al. Infection due to C. difficile

ribotype 078: first report of cases in the Republic of Ireland. J Hosp Infect. 2010; 75: 287-91. 346. Hadi U, Duerink DO, Lestari ES, et al. Audit of antibiotic prescribing in two governmental teaching hospitals in Indonesia. Clin Microbiol Infect. 2008; 14: 698-707. 347. Hadi U, Keuter M, van Asten H, et al. Optimizing antibiotic usage in adults admitted with fever by a multifaceted intervention in an Indonesian governmental hospital. Trop Med Int Health. 2008; 13: 888-99. 348. Hadi U, Duerink DO, Lestari ES, et al. Survey of antibiotic use of individuals visiting public healthcare facilities in Indonesia. Int J Infect Dis. 2008; 12: 622- 9.

137

349. Gurpreet K, Tee GH, Amal NM, et al. Incidence and determinants of acute diarrhoea in Malaysia: a population-based study. J Health Popul Nutr. 2011; 29: 103-12. 350. Komarulzaman A, Smits J, de Jong E. Clean water, sanitation and diarrhoea in Indonesia: Effects of household and community factors. Glob Public Health. 2016; 13: 1-15. 351. Bliss DZ, Johnson S, Savik K, et al. Acquisition of Clostridium difficile and Clostridium difficile-associated diarrhea in hospitalized patients receiving tube feeding. Ann Intern Med. 1998; 129: 1012-9. 352. O'Keefe SJ. Tube feeding, the microbiota, and Clostridium difficile infection. World J Gastroenterol. 2010; 16: 139-42. 353. Gerding DN, Meyer T, Lee C, et al. Administration of spores of nontoxigenic Clostridium difficile strain M3 for prevention of recurrent C. difficile infection: a randomized clinical trial. JAMA. 2015; 313: 1719-27. 354. Liew YX, Lee W, Tay D, et al. Prospective audit and feedback in antimicrobial stewardship: is there value in early reviewing within 48 h of antibiotic prescription? Int J Antimicrob Agents. 2015; 45: 168-73. 355. Tae CH, Jung SA, Song HJ, et al. The first case of antibiotic-associated colitis by Clostridium difficile PCR ribotype 027 in Korea. J Korean Med Sci. 2009; 24: 520-4. 356. Freeman J, Vernon J, Morris K, et al. Pan-European longitudinal surveillance of antibiotic resistance among prevalent Clostridium difficile ribotypes. Clin

Microbiol Infect. 2015; 21: 248 e9-48 e16. 357. Cheknis AK, Sambol SP, Davidson DM, et al. Distribution of Clostridium

difficile strains from a North American, European and Australian trial of treatment for C. difficile infections: 2005-2007. Anaerobe. 2009; 15: 230-3. 358. Baldan R, Trovato A, Bianchini V, et al. Clostridium difficile PCR ribotype 018, a successful epidemic genotype. J Clin Microbiol. 2015; 53: 2575-80. 359. Brazier JS, Patel B, Pearson A. Distribution of Clostridium difficile PCR ribotype 027 in British hospitals. Euro Surveill. 2007; 12: E070426 2. 360. van Steenbergen J, Debast S, van Kregten E, et al. Isolation of Clostridium

difficile ribotype 027, toxinotype III in the Netherlands after increase in C.

difficile-associated diarrhoea. Euro Surveill. 2005; 10: E050714 1. 361. Eyre DW, Tracey L, Elliott B, et al. Emergence and spread of predominantly community-onset Clostridium difficile PCR ribotype 244 infection in Australia, 2010 to 2012. Euro Surveill. 2015; 20: 21059. 362. Androga GO, Hart J, Foster NF, et al. Infection with toxin A-negative, toxin B- negative, binary toxin-positive Clostridium difficile in a young patient with ulcerative colitis. J Clin Microbiol. 2015; 53: 3702-4. 363. Mao EJ, Kelly CR, Machan JT. Racial differences in Clostridium difficile

infection rates are attributable to disparities in health care access. Antimicrob

Agents Chemother. 2015; 59: 6283-7. 364. Johnson S, Sambol SP, Brazier JS, et al. International typing study of toxin A- negative, toxin B-positive Clostridium difficile variants. J Clin Microbiol. 2003; 41: 1543-7. 365. Alfa MJ, Kabani A, Lyerly D, et al. Characterization of a toxin A-negative, toxin B-positive strain of Clostridium difficile responsible for a nosocomial outbreak of Clostridium difficile-associated diarrhea. J Clin Microbiol. 2000; 38: 2706-14.

138

366. Sato H, Kato H, Koiwai K, et al. [A nosocomial outbreak of diarrhea caused by toxin A-negative, toxin B-positive Clostridium difficile in a cancer center hospital]. Kansenshogaku Zasshi. 2004; 78: 312-9. 367. Lessa FC, Gould CV, McDonald LC. Current status of Clostridium difficile

infection epidemiology. Clin Infect Dis. 2012; 55: S65-S70. 368. Hsu LY, Kwa AL, Lye DC, et al. Reducing antimicrobial resistance through appropriate antibiotic usage in Singapore. Singapore Med J. 2008; 49: 749-55. 369. Martin JS, Monaghan TM, Wilcox MH. Clostridium difficile infection: epidemiology, diagnosis and understanding transmission. Nat Rev Gastroenterol

Hepatol. 2016; 13: 206-16. 370. Furuya-Kanamori L, McKenzie SJ, Yakob L, et al. Clostridium difficile

infection seasonality: Patterns across hemispheres and continents - A systematic review. PLoS ONE. 2015; 10: e0120730. 371. Reigadas E, Alcala L, Marin M, et al. Missed diagnosis of Clostridium difficile infection; a prospective evaluation of unselected stool samples. J Infect. 2015; 70: 264-72. 372. Stabler RA, Dawson LF, Valiente E, et al. Macro and micro diversity of Clostridium difficile isolates from diverse sources and geographical locations. PLoS ONE. 2012; 7: e31559. 373. Sheldon E, Kitchin N, Peng Y, et al. A phase 1, placebo-controlled, randomized study of the safety, tolerability, and immunogenicity of a Clostridium difficile

vaccine administered with or without aluminum hydroxide in healthy adults. Vaccine. 2016; 34: 2082-91. 374. de Bruyn G, Saleh J, Workman D, et al. Defining the optimal formulation and schedule of a candidate toxoid vaccine against Clostridium difficile infection: A randomized Phase 2 clinical trial. Vaccine. 2016; 34: 2170-8. 375. Karczewski J, Bodmer JL, Cook JC, et al. Design and purification of subunit vaccines for prevention of Clostridium difficile infection. Methods Mol Biol. 2016; 1403: 385-96.

139

APPENDIX

PUBLICATIONS

REVIEW Open Access

Epidemiology of Clostridium difficile infection inAsiaDeirdre A Collins1*, Peter M Hawkey2,3 and Thomas V Riley1,4

Abstract

While Clostridium difficile infection (CDI) has come to prominence as major epidemics have occurred in NorthAmerica and Europe over the recent decade, awareness and surveillance of CDI in Asia have remained poor.Limited studies performed throughout Asia indicate that CDI is also a significant nosocomial pathogen in thisregion, but the true prevalence of CDI remains unknown. A lack of regulated antibiotic use in many Asian countriessuggests that the prevalence of CDI may be comparatively high. Molecular studies indicate that ribotypes 027 and078, which have caused significant outbreaks in other regions of the world, are rare in Asia. However, variant toxinA-negative/toxin B-positive strains of ribotype 017 have caused epidemics across several Asian countries. Ribotypesmz/018 has caused widespread disease across Japan over the last decade and more recently emerged in Korea.This review summarises current knowledge on CDI in Asian countries.

Keywords: Clostridium Difficile, Clostridium Infections In Humans, Epidemiology

IntroductionClostridium difficile causes infection ranging from milddiarrhoea to pseudomembranous colitis (PMC), prima-rily in older age patients who have been exposed to anti-biotics. Epidemics of C. difficile infection (CDI) haveoccurred in North America and Europe over recent de-cades and the epidemiology of CDI in these regions iswell-documented. These epidemics have demonstratedthe need for surveillance of the international movementof C. difficile strains [1]. Circulating strains in Asia, as inother regions, have the potential to spread internatio-nally, warranting close monitoring of the prevalence andmolecular epidemiology of CDI in the region. Indeed,it is likely that the variant toxin A-negative/ toxinB-positive (A-B+) ribotype 017 C. difficile strain origi-nated in Asia. One particular clindamycin-resistantribotype 017 strain of apparent clonal origin has domi-nated international typing studies of A-B+ strains and hasbeen the cause of epidemics in Canada, the Netherlandsand Ireland [2-4]. Unfortunately, limited data are availableon CDI in Asia. A recent survey found that awareness ofCDI in physicians is poor in Asia, with underestimation of

its contribution to antibiotic-associated disease and re-currence rates [5]. In addition, comprehensive cultureand toxin testing for C. difficile are lacking in manyAsian hospitals. As a consequence, reports of C. difficileare rare in Asia, so data on prevalence and circulatingstrains are limited. What reports are available on Asiancountries are described here.

Literature search and selection strategyPubMed searches were performed for publications priorto 1 May 2013 with the term “Clostridium difficile” com-bined with specific Asian country names. A search for“Clostridium difficile” was also performed on theWanfang and KoreaMed databases. Manual searches ofcited references of these articles were performed andrelevant English language articles and abstracts were in-cluded for analysis.

East AsiaJapanNo English language reports were found on the preva-lence of CDI in Japan. However, molecular typing stu-dies have provided some epidemiological informationsince the 1990s. A variety of typing techniques have beenemployed in Japan, including tcdA and tcdB character-isation [6,7], pulsed field gel electrophoresis (PFGE) [8],

* Correspondence: [email protected] and Immunology, School of Pathology and LaboratoryMedicine, the University of Western Australia, Perth, AustraliaFull list of author information is available at the end of the article

© 2013 Collins et al.; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the CreativeCommons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, andreproduction in any medium, provided the original work is properly cited.

Collins et al. Antimicrobial Resistance and Infection Control 2013, 2:21http://www.aricjournal.com/content/2/1/21

mailto:[email protected]

http://creativecommons.org/licenses/by/2.0

polymerase chain reaction (PCR) ribotyping [8,9] andslpA typing [10]. Application of molecular typing tech-niques has identified A-B+ strains in regions across Japan[6,11-14]. PCR typing of tcdA on six A-B+ strains fromJapan and Indonesia identified indistinguishable repea-ting sequences with two deletions (1,548 and 273 nucle-otides in size) [7]. A later study also grouped A-B+ strainsfrom Japan and Indonesia, as toxinotype VIII and ribotypefr/017, with several subtypes identified by PFGE [15].Several ribotyping studies have been performed, indi-

cating predominance of ribotype “smz” over the pastdecade [8,16,17]. A 5-year study in a Tokyo hospitalfollowed the proliferation of ribotype “smz”, peaking in2004 (64% of cases) [16]. Ribotype “smz” is recognisedinternationally as ribotype 018 (personal observation),and three major subtypes have been identified withinJapan by slpA typing, two of which are widespread[10,18]. Ribotype 018 strains have caused CDI recently inKorea [19], Austria, Spain and Slovenia [20], and havebeen responsible for outbreaks of disease in Italy since2007 [21]. Other common ribotypes were 014, 002 and001 [14,16,17]. Ribotype 027, which has caused wide-spread epidemics in North America and Europe, has beenreported only occasionally in Japan [16].

KoreaA country-wide survey of 17 tertiary hospitals in Koreafrom 2004 to 2008 found that the incidence of CDIincreased from 1.7/1,000 adult admissions to 2.7/1,000admissions [22]. Diagnostic methods were not reported,and may have differed between sites or over time, con-tributing to these apparent increases.Risk factors for recurrent CDI included antibiotic the-

rapy, anaemia, and tube feeding in one study [23], whileanother found proton pump inhibitor (PPI) use alonewas associated with recurrent disease [24]. One studyfound that the proportion of community-acquired CDI(CA-CDI) among all CDI cases within a Busan hospitalwas 7.1% [25], while another reported that 59.4% ofcases of CDI presenting at the emergency department ofa Seoul hospital were community-acquired [26].Variant A-B+ strains have been common in Korea since

the 1990s [27] and increased in prevalence among allstrains significantly between 2002 and 2005, peaking at50% in 2004 [27-30]. These epidemic A-B+ strainsbelonged to ribotype group 017 and were widespread inKorea [27,28,31]. Subtyping of 017 strains revealed six dif-ferent PFGE pulsotypes and 13 restriction fragment lengthpolymorphism-based subtypes [32]. Shin et al. reportedthat 72% of PMC cases between 2006 and 2010 werecaused by A-B+ strains [29]. Ribotype 018 was the mostprevalent strain isolated in Seoul from September 2008 toJanuary 2010 [33]. Ribotype 027 was detected in 2006 [34]in a hospital-associated case of PMC but has failed to

proliferate in Korea [19]. Type 078 was reported as themost common (3.1%) binary toxin-positive strain [31].

ChinaThe lack of reports on CDI from mainland China con-trasts with the relatively large number of reports fromneighbouring Korea and Japan. An English language re-view of the Chinese language literature found two stud-ies from Southern China in general hospital patientswith 21/183 patients (1994–1997) and 13/257 patients(reported in 2006) with CDI (no diagnostic criteriareported). Another study from Beijing identified 36 casesamong 71,428 in-patients from 1998 to 2001 [35]. Fur-ther studies were limited to special groups of patientswith malignancy and those receiving chemotherapy orstem cell transplantation, with incidence rates varyingfrom 1.6% to 3.5%, with an exceptionally high rate (27%)reported in the stem cell transplant patient study [35].In most studies a sampling frame was not reportedwhich, in view of the low rate of testing in China, makesassessment of the true incidence of CDI impossible [35].The most comprehensive Chinese study of CDI was

conducted between March 2007 and April 2008 at a1,216 bed hospital in Shanghai [36-38]. During the studyperiod, 42,936 patients were discharged and 587 patientshad stool samples submitted for testing by toxin assayand culture [36]. Overall the incidence of CDI was 17.1per 10,000 admissions [37]. CDI was mild, possibly dueto the younger mean age of patients (62.8 years) com-pared with a large European survey where 63% were ≥65years [20]. Fifty-six isolates from this study were aggre-gated with further unspecified isolates to create a collec-tion of 75 [38]. The most common ribotypes were 017,012 and 046 (Figure 1).In the absence of ribotyping studies other than that of

Huang et al., a recent report of MLST typing on 69 C.difficile strains mainly from Beijing provides some fur-ther insight on the molecular epidemiology of CDI inChina [41]. The equivalent ribotypes of the most com-mon sequence types (ST37, ST35, ST54) found were 017(23%), 046 (23%) and 012 (17%), respectively [42]. Thecollection included 16 Guangzhou isolates from the1980s comprised of the ribotype equivalents: 017 (9 iso-lates), 046 (1 isolate), 020 (5 isolates) and ST119 (1 iso-late). This suggests that ribotypes 017, 046 and perhaps012 are the most common in mainland China, a patternthat differs to other world regions.

TaiwanA similar situation to Japan and China is found inTaiwan, where testing is rare so the true incidence ofCDI cannot be defined. The number of CDI cases in-creased by 5–6 times in patients ≥65 years between 2003and 2007 according to one study [43]. Increasing

Collins et al. Antimicrobial Resistance and Infection Control 2013, 2:21 Page 2 of 9http://www.aricjournal.com/content/2/1/21

awareness and testing probably contributed to the dra-matic rise, as another report by Chan et al. demon-strates. They found the proportion of positive culturesremained constant at ~10% while testing increased from2002 to 2009 as awareness of CDI increased [44]. Theincidence rate found at a northern Taiwan hospital was0.45/1,000 patient days overall, and 7.9/1,000 patientdays in ICU wards [45]. Ribotype 017 strains were found(6%) in a sample of 142 Taiwanese isolates where 57ribotypes were identified [46], however other inter-nationally recognised ribotypes of Taiwanese isolateshave not been reported. While binary toxin-positive iso-lates were detected, ribotypes 027 and 078 have not beenreported from Taiwan [47].

Hong KongCDI has been better recognised and studied for longerin Hong Kong than in mainland China and Taiwan. Asurvey at Queen Mary Hospital, Hong Kong in 1996–7reported 100/3,112 patients with diarrhoea positive byculture for C. difficile [48]. A more recent survey at the

same hospital using a tissue culture cytotoxin assay be-tween September and December 2008 detected 37/723positive samples [49]. The number of patients diagnosedin 2003 was 82 compared to 66 in 2008, suggesting theoverall number of cases was constant. Testing for CDIincreased following the isolation of ribotype 027 amongone of 12 hyper-toxigenic (detectable cytotoxin at 100-fold dilution) isolates identified in a 3-month study in auniversity hospital in 2008 [49].There are little data on ribotyping in Hong Kong, the

most comprehensive being a retrospective study of 345isolates from 307 patients in 2009 from a healthcare re-gion spanning five hospitals. Unusually, 70% were of apattern not represented by 23 of the most internationallycommon ribotypes, with a further 11.6% being non-typable (Figure 1) [40]. Ribotype 002 represented 9.4% ofstrains, presumed to be causing cross-infection as theincidence between 2004 and 2008 was 0.53/1,000 admis-sions rising in 2009 to 0.95/1,000 admissions. An ele-vated frequency of sporulation (20.2%) was found in 35of these ribotype 002 isolates compared to 3.7% for 56

Figure 1 Ribotype distributions for studies in China (A,B), Hong Kong (C), Japan (D-G) and Korea (H-K). A: n = 75 isolates collected in2008; 33 ribotypes identified [37]. B: n = 110, December 2008-May 2009; 16 ribotypes [39]. C: n = 345, 2009; 106 ribotypes [40]. D: n = 87, March1996-November 1999; 12 ribotypes [8]. E: n = 148, November 1999-October 2004; 26 ribotypes [16]. F: n = 71, April 2005-March 2008; 20 ribotypes[14]. G: n = 87, 2003–2007; 18 ribotypes [17]. H: n = 187, 1980–2006; 39 ribotypes [27]. I: n = 105, 1995–2008; 11 ribotypes [32]. J: n = 337 toxigenicisolates, 2006–2008; 50 ribotypes [31]. K: n = 194, 2009–2010; 54 ribotypes [19].


randomly selected isolates of other ribotypes [40].Ribotype 017 strains were also found at a prevalence of0.7% [40]. Ribotype 027 was first detected in 2008 instool of a patient on steroids with no history of travel inthe previous 1.5 years and, as yet, there have been nofurther reports of 027 from Hong Kong [49].

Southeast AsiaPhilippinesThe underestimation of the role of C. difficile in entericdisease in Asia was demonstrated recently by a study inthe Philippines. Historically, patients with colitis were di-agnosed with amoebic colitis, presumed to be caused byEntamoeba histolytica. However, 43.6% of colitis caseswere positive for C. difficile when testing by enzyme im-munoassay (EIA) for toxin A/B was introduced [50].Given that metronidazole is the standard treatment forboth CDI and amoebic colitis, CDI would be masked iftesting was not carried out.

ThailandIn the earliest report of CDI in Thailand in 1990, faecalcytotoxin was detected in 52.5% of 206 diarrhoeal pa-tients while culture gave low recovery rates (4.8%) over a26-month period. Cytotoxin was detected in stool of61% of antibiotic-treated patients, and 51% of non-antibiotic-treated patients [51]. In contrast, Thamlikitkulet al. reported low detection of toxin A by EIA inclindamycin-treated (10/140 patients) and patientstreated with beta-lactams (10/140) [52]. The differencein prevalence found in both studies was probably due tothe use of different toxin assays.The prevalence of C. difficile in antibiotic-associated

diarrhoea (AAD) patients was reported as 18.64% by cul-ture and 44-46% by PCR for tcdA and tcdB on the samestool samples, showing the need for both culture andtesting for presence of toxin [53]. Similarly, a recentstudy in Bangkok found that combining toxin EIAs withdirect tcdB PCR on stool increased positive diagnosestwo-fold, compared with toxin EIAs alone [54].Wongwanich et al. found a greater prevalence of CDI

in 201 HIV-positive patients (58.8%) than in 271 HIV-negative patients (36.5%), and six PFGE subtypes werefound [55]. Again, toxin A EIA alone gave lower preva-lence than when combined with culture. Another studyon AIDS patients where only toxin A EIA was perfor-med reported a lower prevalence of CDI (16/102 pa-tients) [56]. The discrepancies between studies wherelower prevalence is reported when a toxin A assay alonewas used suggest it is likely that toxin A-negative strainswere widespread.Wongwanich et al. reported on a collection of 77 C.

difficile isolates from 44 asymptomatic children and in-fants, and 33 diarrhoeal adult patients. In this study,

tcdA PCR-negative isolates predominated, numbering 57overall and 18/33 adult diarrhoeal isolates. FourteenPFGE pulsotypes and eight subtypes were found [57].No data on circulating ribotypes have been reported.

MalaysiaReports of CDI in Malaysia are rare. In the most inform-ative study in Malaysia to date, toxin A/B assays on 175stool samples from inpatients with AAD were performedin a tertiary hospital in north-eastern Malaysia. Twenty-four (13.7%) were positive for toxin, with the majority ofinfected patients aged >50 years [58]. No ribotyping orother molecular analysis has been reported on MalaysianC. difficile isolates.

IndonesiaLike Malaysia, reports on CDI in Indonesia are uncom-mon. An aetiology study of diarrhoea in Indonesian chil-dren identified C. difficile in 1.3% of stool samplestested. Toxin A EIA only was performed so the trueprevalence of C. difficile may have been greater [59]. Theonly molecular study included eight isolates fromIndonesia, five of which were toxinotype VIII and ribotype017, and grouped with international 017 epidemic strains.Two were A+B+ toxinotype 0, and one A-B+ isolate wasbinary toxin-positive, toxinotype XVI [15].

SingaporeA 50-month aetiological study (1985 to 1989) of diarrhoeain 4,508 patients at the National University Hospital foundC. difficile in 35 of only 365 cases where C. difficile culturewas requested. The incidence rate of CDI in SingaporeGeneral Hospital was 3.2/1,000 admissions, more com-mon in males and patients >50 years. Culture of stoolsamples identified cases (43%) which were negative bytoxin A/B EIA. CDI was three times more common inMalay patients than in patients of Indian heritage [60]. Anincrease in incidence of CDI was observed in a 1,200 bedgeneral hospital from 2001 to 2006. The incidence raterose from 1.49 cases per 10,000 patient days to 6.64 per10,000 patient days, with a concurrent increase in positivetoxin assays on stool [61]. A broader surveillance study inthree hospitals found that CDI incidence decreased from0.52/1,000 patient days in 2006 to 0.3/1,000 patient daysin 2008, as testing increased over the same time period[62]. Binary toxin-positive strains have been reported, in-cluding 027 strains, which have caused sporadic hospital-acquired disease [63].

South AsiaIndiaAn early report found 21/93 AAD cases were positivefor C. difficile by culture and toxin assay in NehruHospital in 1983–1984 [64]. In Calcutta, C. difficile was


isolated in 38/341 hospitalised patients with acute diar-rhoea over 1 year [65]. A hospital in Delhi reported CDIin 26/156 diarrhoeal hospitalised patients, detected byculture and toxin A EIA [66]. Infection control measuressubsequently introduced at this hospital apparently re-duced incidence of CDI by more than 50% over 5 years[67]. A retrospective review by Ingle et al. in a Mumbaihospital found 17/99 patients between 2006 and 2008were diagnosed with CDI by toxin A/B EIA [68]. Themost recent report found a prevalence of non-toxigenicC. difficile of 12.6% among 79 hospitalised patients, fiveof whom subsequently developed diarrhoea with positiveculture and toxin assay. The study group also detectedwidespread contamination of surfaces on beds (51%) andhands of hospital workers (62.5%) [69]. Several reportsexist of acute diarrhoea in hospitalised children (7-11%)caused by C. difficile [70-72]. The molecular epidemiologyof C. difficile strains in India is not currently known.

BangladeshIn the 1990s, an aetiological study found that 13/814children admitted to hospital with diarrhoea wereinfected with C. difficile (diagnosed by cell cytotoxinassay). Seven of the cases were concurrently infectedwith another diarrhoeal agent [73]. Recent reports arenot available for Bangladesh.

EpidemiologyPrevalent C. difficile ribotypes in AsiaRibotyping data with internationally recognised nomen-clature are available for China, Japan, Singapore, HongKong, Taiwan, and Korea. Overall, the most prevalentribotypes in Asia appear to be 017, 018, 014, 002, and001. While ribotypes 027 and 078 have caused outbreaksin North America and Europe, they are reported only tohave caused sporadic cases of CDI in Asia so far, inSingapore, Hong Kong, Korea, and Japan [16,34,40,46,49,63,74]. Ribotype 078 has only been reported fromKorea and China to date [31,37]. Another binary toxin-positive strain, ribotype 130, was recently reported fromKorea [19].Meanwhile, ribotype 017, A-B+, toxinotype VIII strains

are widespread in Asia, and have caused epidemicsworldwide (Figure 1). In China and Korea 017 is themost common ribotype in circulation, and is prevalentin Japan, Taiwan, and Hong Kong also [31,32,36,39]. Ex-posure to antineoplastic agents, use of nasal feedingtubes, and care in a particular hospital ward were associ-ated with infection with 017 strains in one hospital inJapan [11]. Ribotype 017 strains have persisted in Chinaand Taiwan while they appear to have declined in Korea(Figure 1).In Japan, smz/018 appears to have persisted as the

most common ribotype for over a decade (Figure 1).

Ribotype smz/018 was the most prevalent strain isolatedin a tertiary hospital in Seoul between September 2008and January 2010 [33], indicating spread from Japan toKorea. Ribotype 018 caused outbreaks of CDI in Italy in2007/2008 and is the fourth most prevalent ribotype inEurope at present [20,21]. It is not clear whether smz/018 is prevalent in other Asian countries, as comparativetyping with a reference smz or 018 strain may not havebeen performed.Ribotypes 017 and 018 have caused widespread disease

in Asia and across the world. Unlike the other majorepidemic strains 027 and 078, they do not produce bin-ary toxin, and ribotype 018 does not appear to possessvariant toxin genes [21]. Some other virulence factorsmay contribute to their spread. The resistance of ribotype018 isolates to clindamycin and fluoroquinolones couldcontribute to their enhanced virulence [8,33,40]. Anothervirulence factor is high sporulation rate. The epidemicribotype 002 isolates in Hong Kong sporulated at a higherrate than other isolates, allowing them to persist in thehospital environment and cause outbreaks of disease [40].

Prevention and controlTwo reports of infection control in Asian hospitals werefound. A hospital in India introduced control measuresincluding disinfection of surfaces, rapid detection of C.difficile by toxin assays, isolation of patients, controls onprescription of antibiotics and education of staff mem-bers. The incidence of CDI (initially 15% among cases ofnosocomial diarrhoea) was reduced by 50% while thenumber of tests requested increased as health workersbecame more aware of CDI [67]. A hospital-widecomputerised antimicrobial stewardship scheme was in-troduced in a hospital in Taiwan. While the incidence ofsome antibiotic resistant organisms decreased, the isola-tion rate of C. difficile remained constant at 10% [44],indicating that other infection control measures besidesantimicrobial stewardship would be required to controlCDI in hospitals.Asia is going through a period of rapid demographic

change. With its dense, growing population, infectioncontrol is a pertinent issue. As C. difficile now causes themajority of nosocomial disease in North America and Eur-ope, control measures could be applied in Asia to preventthe same situation there. A number of issues exist whichcould contribute to the spread of CDI in Asia.As wealth and the aged population are increasing,

more people have access to hospital care and enter agedcare facilities. It is likely that CDI incidence could in-crease as these high-risk populations increase in size.For example, modelling of the future age structure ofthe Chinese population suggests that there will be a lar-ger population at risk for CDI. Using census data from2005 (population 1.3 billion) when only 100 million


individuals were ≥65 years old, by 2026 there will be 200million individuals ≥65 years [75].Antibiotic use in most Asian countries is poorly regu-

lated. A review of Southeast Asian countries found that47% of pneumonia cases do not receive an appropriateantibiotic, 54% of diarrhoea cases are unnecessarilytreated with antibiotics, and 40% of antibiotics are pre-scribed in under-dose [76]. In many cases inappropriateantibiotics are prescribed without any laboratory test.Studies in India have found the most commonly pre-scribed antibiotics for cough and respiratory disease arefluoroquinolones, a known risk factor for CDI [76]. Inaddition, antibiotics are freely available without prescrip-tion in most Asian countries, leading to misuse in thecommunity.Given the free use of antibiotics by the general public it

would be plausible that CA-CDI could be common in Asia.Studies in Asian countries have neglected to address theissue of CA-CDI, apart from two studies in Korea whichfound conflicting proportions of 7% and 59% of all CDIsurveyed being community-acquired. It would be appropri-ate to monitor CA-CDI more closely in Asia in the future.

Despite widespread antibiotic use few studies in Asiahave measured antimicrobial susceptibility of clinical C.difficile isolates. High resistance rates to moxifloxacin,and clindamycin have been found in isolates from Korea,Japan, Northern Taiwan and China (Table 1). Heterore-sistance to metronidazole has been reported from China,warranting close monitoring (Table 1).Production and consumption of meat products is also

increasing in Asia [80]. Intensive farming of poultry,seafood and swine is already in place, and increasingwith worldwide demand [81]. The risk of C. difficilecolonisation and/or infection in animals would mostlikely increase with intensive farming practices includingcrowding of animals and prophylactic antibiotic use.Thus contamination of food products and animal-human transmission could occur. To date, no reportshave been made of C. difficile in the environment oranimals, apart from five cases of fulminant colitiscaused by ribotype 078 in thoroughbred racehorses inJapan [82], and the isolation of C. difficile from 2/ 250(0.8%) swine faecal samples from 25 pig farms, also inJapan [83].

Table 1 Antimicrobial resistance rates and MIC values for Clostridium difficile isolates from different countries

(Reference) Numberof isolates

[Resistance rate] (%), MIC50 (mg/L), MIC90 (mg/L)

Erythromycin Clindamycin Tetracycline Moxifloxacin Ciprofloxacin Piperacillin/tazobactam

Metronidazole Vancomycin

China

[37] 75 [76], 128, 128 [66.7], 128,128

[41.3], 4, 64 [45.3], 4, 128 [100], 64, 128 [0], <16/4,16/4

[0], 0.25, 0.25 [0], 1, 2

[39] 110 [85.3], 128,128

[88.1], 128,128

[62.7], 16, 32 [61.8], 16, 128 [100], 64, 128 [0], 8/4, 16/4 [0]a, 0.125, 0.25 [0], 0.5, 1

Hong Kong

[40] 35 [100] 0.5, 0.75 0.75, 1.5

Japan

[77] 73 [87.7], >256,>256

[87.7], >256,>256

[100], >32,>32

[0], 0.19, 0.25 [0], 2, 4

[16] 72 [12]b [100] [0] [0]

Korea

[31] 120 [50], 128,>128

[42], 2, 16 [0], 8, 16 [0], 1,4 [0], 0.5, 1

[19] 111 [82] [83]

[33] 131 [67.9] [82] [0] [0]

[78] 123 [75] [85]

Singapore

[60] 68 [63], 8, >512 [0], 0.5, 1 [0], 1, 1

Taiwan

[79] 60 [73.3], 16, 64 [41.7], 8, >16 [30] [100]

[47] 113 [46], 4, >256 [16] [0]a 18 isolates showed heteroresistance to metronidazole, bsubset of 25 isolates tested


ReviewAccording to the existing evidence, CDI occurs at simi-lar rates in Asia as in other continents where CDI ismore commonly recognised and researched. The mo-lecular epidemiology of C. difficile strains in Asia indi-cates a persisting predominance of variant A-B+ ribotype017 strains and ribotype 018 strains (Figure 1). Binarytoxin-positive strains have rarely been isolated to datedespite the proliferation of ribotypes 078 and 027 inEurope and North America. Favouring toxin A EIAs fordiagnostic methods is not optimal for the Asian regiondue to the predominance of A-B+ strains. Broader sur-veillance monitoring CA-CDI and C. difficile in animalswill enhance our understanding of the epidemiology ofCDI in the region.

ConclusionsCDI is not widely recognised in Asia so in consequencethe extent of the disease is not known. Although rela-tively few studies on C. difficile have been performed inAsia, what work has been done demonstrates that CDIis a significant cause of nosocomial disease in Asiancountries. It appears that awareness is increasing andtesting and surveillance are on the rise. Routine testingis required to inform on the prevalence of CDI through-out the region. The widespread prevalence of the 017group of A-B+ strains in Asian countries shows thatassays for toxin B or the tcdB genes are preferable totoxin A assays for diagnosis of CDI. The more virulentepidemic strains 027 and 078 do not appear to havebecome established in Asia, while ribotype 017 andsmz/018 strains have caused epidemics.Widespread unregulated antibiotic use and inappropri-

ate prescribing in SE Asian countries indicates that CDIcould be widespread in those regions where surveillanceis currently lacking. Asia may be facing a “perfect storm”as heavy usage of antibiotics combines with an ageingincreasingly hospitalised population. Increasing labora-tory capacity in the region as well as improving sur-veillance should be seen as essential in preventingunnecessary morbidity and mortality.

AbbreviationsCDI: Clostridium difficile infection; PMC: Pseudomembranous colitis;PFGE: Pulsed-field gel electrophoresis; PCR: Polymerase chain reaction;PPI: Proton pump inhibitor; CA-CDI: Community-acquired Clostridium difficileinfection; EIA: Enzyme immunoassay; AAD: Antibiotic-associated diarrhoea.

Competing interestsThe authors declare that they have no competing interest.

Authors’ contributionsDAC performed literature searches, analysed and interpreted data anddrafted the manuscript. PMH planned and designed the review, interpreteddata, drafted and revised the manuscript. TVR planned and designed thereview and revised the manuscript. All authors read and approved the finalmanuscript.

AcknowledgementsD Collins is a recipient of an International Postgraduate Research Scholarshipfor study in Australia, and an Australian Postgraduate Award.

Author details1Microbiology and Immunology, School of Pathology and LaboratoryMedicine, the University of Western Australia, Perth, Australia. 2West MidlandsPublic Health Laboratory, Health Protection Agency, Heart of England NHSFoundation Trust, Birmingham, UK. 3School of Immunity and Infection,University of Birmingham, Birmingham, UK. 4Division of Microbiology andInfectious Diseases, PathWest Laboratory Medicine (WA), Perth, Australia.

Received: 7 June 2013 Accepted: 7 June 2013Published: 1 July 2013

References1. Clements AC, Magalhaes RJ, Tatem AJ, Paterson DL, Riley TV: Clostridium

difficile PCR ribotype 027: assessing the risks of further worldwidespread. Lancet Infect Dis 2010, 10(6):395–404.

2. Al-Barrak A, Embil J, Dyck B, Olekson K, Nicoll D, Alfa M, Kabani A:An outbreak of toxin A negative, toxin B positive Clostridium difficile-associated diarrhea in a Canadian tertiary-care hospital. Can Commun DisRep 1999, 25(7):65–69.

3. Kuijper E, Weerdt J, Kato H, Kato N, Dam A, Vorm E, Weel J, Rheenen C,Dankert J: Nosocomial outbreak of Clostridium difficile-associateddiarrhoea due to a clindamycin-resistant enterotoxin A-negative strain.Eur J Clin Microbiol Infect Dis 2001, 20(8):528–534.

4. Drudy D, Harnedy N, Fanning S, Hannan M, Kyne L: Emergence andcontrol of fluoroquinolone-resistant, toxin A-negative, toxin B-positiveClostridium difficile. Infect Control Hosp Epidemiol 2007, 28(8):932–940.

5. Mavros MN, Alexiou VG, Vardakas KZ, Tsokali K, Sardi TA, Falagas ME:Underestimation of Clostridium difficile infection among clinicians: aninternational survey. Eur J Clin Microbiol Infect Dis 2012, 31(9):2439–2444.

6. Kato H, Kato N, Watanabe K, Iwai N, Nakamura H, Yamamoto T, Suzuki K,Kim S-M, Chong Y, Wasito EB: Identification of toxin A-negative, toxin B-positive Clostridium difficile by PCR. J Clin Microbiol 1998, 36(8):2178–2182.

7. Kato H, Kato N, Katow S, Maegawa T, Nakamura S, Lyerly DM: Deletions in therepeating sequences of the toxin A gene of toxin A-negative, toxin B-positive Clostridium difficile strains. FEMS Microbiol Lett 1999, 175(2):197–203.

8. Kato H, Kato N, Watanabe K, Yamamoto T, Suzuki K, Ishigo S, Kunihiro S,Nakamura I, Killgore GE, Nakamura S: Analysis of Clostridium difficileisolates from nosocomial outbreaks at three hospitals in diverse areas ofJapan. J Clin Microbiol 2001, 39(4):1391–1395.

9. Kato H, Kita H, Karasawa T, Maegawa T, Koino Y, Takakuwa H, Saikai T,Kobayashi K, Yamagishi T, Nakamura S: Colonisation and transmission ofClostridium difficile in healthy individuals examined by PCR ribotypingand pulsed-field gel electrophoresis. J Med Microbiol 2001, 50(8):720–727.

10. Kato H, Yokoyama T, Arakawa Y: Typing by sequencing the slpA gene ofClostridium difficile strains causing multiple outbreaks in Japan. J MedMicrobiol 2005, 54(Pt 2):167–171.

11. Komatsu M, Kato H, Aihara M, Shimakawa K, Iwasaki M, Nagasaka Y, FukudaS, Matsuo S, Arakawa Y, Watanabe M, et al: High frequency of antibiotic-associated diarrhea due to toxin A-negative, toxin B-positive Clostridiumdifficile in a hospital in Japan and risk factors for infection. Eur J ClinMicrobiol Infect Dis 2003, 22(9):525–529.

12. van den Berg RJ, Claas EC, Oyib DH, Klaassen CH, Dijkshoorn L, Brazier JS,Kuijper EJ: Characterization of toxin A-negative, toxin B-positiveClostridium difficile isolates from outbreaks in different countries byamplified fragment length polymorphism and PCR ribotyping. J ClinMicrobiol 2004, 42(3):1035–1041.

13. Kikkawa H, Hitomi S, Watanabe M: Prevalence of toxin A-nonproducing/toxin-B-producing Clostridium difficile in the Tsukuba-Tsuchiura district.Japan. J Infect Chemother 2007, 13(1):35–38.

14. Iwashima Y, Nakamura A, Kato H, Wakimoto Y, Wakiyama N, Kaji C, Ueda R:A retrospective study of the epidemiology of Clostridium difficileinfection at a University Hospital in Japan: genotypic features of theisolates and clinical characteristics of the patients. J Infect Chemother2010, 16(5):329–333.

15. Rupnik M, Kato N, Grabnar M, Kato H: New types of toxin A-negative, toxinB-positive strains among Clostridium difficile isolates from Asia. J ClinMicrobiol 2003, 41(3):1118–1125.


16. Sawabe E, Kato H, Osawa K, Chida T, Tojo N, Arakawa Y, Okamura N:Molecular analysis of Clostridium difficile at a university teaching hospitalin Japan: a shift in the predominant type over a five-year period.Eur J Clin Microbiol Infect Dis 2007, 26(10):695–703.

17. Kato H, Kato H, Ito Y, Akahane T, Izumida S, Yokoyama T, Kaji C, Arakawa Y:Typing of Clostridium difficile isolates endemic in Japan by sequencing of slpAand its application to direct typing. J Med Microbiol 2010, 59(Pt 5):556–562.

18. Kato H, Kato H, Nakamura M, Iwashima Y, Nakamura A, Ueda R: Rapidanalysis of Clostridium difficile strains recovered from hospitalizedpatients by using the slpA sequence typing system. J Infect Chemother2009, 15(3):199–202.

19. Kim J, Kang JO, Kim H, Seo MR, Choi TY, Pai H, Kuijper EJ, Sanders I, FawleyW: Epidemiology of Clostridium difficile infections in a tertiary-carehospital in Korea. Clin Microbiol Infect 2013, 19(6):521–527.

20. Bauer MP, Notermans DW, van Benthem BHB, Brazier JS, Wilcox MH, RupnikM, Monnet DL, van Dissel JT, Kuijper EJ: Clostridium difficile infection inEurope: a hospital-based survey. Lancet 2011, 377(9759):63–73.

21. Spigaglia P, Barbanti F, Dionisi AM, Mastrantonio P: Clostridiumdifficileisolates resistant to fluoroquinolones in Italy: emergence of PCRribotype 018. J Clin Microbiol 2010, 48(8):2892–2896.

22. Kim YS, Han DS, Kim YH, Kim WH, Kim JS, Kim HS, Park YS, Song HJ, Shin SJ,Yang SK, et al: Incidence and clinical features of Clostridium difficile infectionin Korea: a nationwide study. Epidemiol Infect 2013, 141(1):189–194.

23. Ryu HS, Kim YS, Seo GS, Lee YM, Choi SC: Risk factors for recurrentClostridium difficile infection. Intest Res 2012, 10(2):176–182.

24. Kim YG, Graham DY, Jang BI: Proton pump inhibitor use and recurrentClostridium difficile-associated disease: a case–control analysis matchedby propensity score. J Clin Gastroenterol 2012, 46(5):397–400.

25. Choi HK, Kim KH, Lee SH, Lee SJ: Risk factors for recurrence of Clostridiumdifficile infection: Effect of vancomycin-resistant Enterococci colonization.J Korean Med Sci 2011, 26(7):859–864.

26. Shin BM, Moon SJ, Kim YS, Shin WC, Yoo HM: Characterization of cases ofClostridium difficile infection (CDI) presenting at an emergency room:Molecular and clinical features differentiate community-onset hospital-associated and community-associated CDI in a tertiary care hospital.J Clin Microbiol 2011, 49(6):2161–2165.

27. Kim H, Riley TV, Kim M, Kim CK, Yong D, Lee K, Chong Y, Park JW:Increasing prevalence of toxin A-negative, toxin B-positive isolates ofClostridium difficile in Korea: impact on laboratory diagnosis. J ClinMicrobiol 2008, 46(3):1116–1117.

28. Shin BM, Kuak EY, Yoo HM, Kim EC, Lee K, Kang JO, Whang DH, Shin JH:Multicentre study of the prevalence of toxigenic Clostridium difficile inKorea: results of a retrospective study 2000–2005. J Med Microbiol 2008,57(Pt 6):697–701.

29. Shin BM, Kuak EY, Yoo SJ, Shin WC, Yoo HM: Emerging toxin A-B + variantstrain of Clostridium difficile responsible for pseudomembranous colitisat a tertiary care hospital in Korea. Diagn Microbiol Infect Dis 2008,60(4):333–337.

30. Shin BM, Lee EJ, Kuak EY, Yoo SJ: Comparison of VIDAS CDAB and CDAimmunoassay for the detection of Clostridium difficile in a tcdA- tcdB + C.difficile prevalent area. Anaerobe 2009, 15(6):266–269.

31. Kim H, Jeong SH, Roh KH, Hong SG, Kim JW, Shin MG, Kim MN, Shin HB, UhY, Lee H, et al: Investigation of toxin gene diversity, molecularepidemiology, and antimicrobial resistance of Clostridium difficile isolatedfrom 12 hospitals in South Korea. Korean J Lab Med 2010, 30(5):491–497.

32. Kim SJ, Kim H, Seo Y, Yong D, Jeong SH, Chong Y, Lee K: Molecularcharacterization of toxin A-negative, toxin B-positive variant strains ofClostridium difficile isolated in Korea. Diagn Microbiol Infect Dis 2010,67(2):198–201.

33. Kim J, Kang JO, Pai H, Choi TY: Association between PCR ribotypes andantimicrobial susceptibility among Clostridium difficile isolates fromhealthcare-associated infections in South Korea. Int J Antimicrob Agents2012, 40(1):24–29.

34. Kim H, Lee Y, Moon HW, Lim CS, Lee K, Chong Y: Emergence of Clostridiumdifficile ribotype 027 in Korea. Korean J Lab Med 2011, 31(3):191–196.

35. Jin K, Wang S, Huang Z, Lu S: Clostridium difficile infections in China.Journal of Biomedical Research 2010, 24(6):411–416.

36. Huang H, Wu S, Wang M, Zhang Y, Fang H, Palmgren AC, Weintraub A,Nord CE: Molecular and clinical characteristics of Clostridium difficileinfection in a University Hospital in Shanghai. China. Clin Infect Dis 2008,47(12):1606–1608.

37. Huang H, Fang H, Weintraub A, Nord CE: Distinct ribotypes and rates ofantimicrobial drug resistance in Clostridium difficile from Shanghai andStockholm. Clin Microbiol Infect 2009, 15(12):1170–1173.

38. Huang H, Wu S, Wang M, Zhang Y, Fang H, Palmgren AC, Weintraub A,Nord CE: Clostridium difficile infections in a Shanghai hospital:antimicrobial resistance, toxin profiles and ribotypes. Int J AntimicrobAgents 2009, 33(4):339–342.

39. Huang H, Weintraub A, Fang H, Wu S, Zhang Y, Nord CE: Antimicrobialsusceptibility and heteroresistance in Chinese Clostridium difficile strains.Anaerobe 2010, 16(6):633–635.

40. Cheng V, Yam W, Lam O, Tsang J, Tse E, Siu G, Chan J, Tse H, To K, Tai J,et al: Clostridium difficile isolates with increased sporulation: emergenceof PCR ribotype 002 in Hong Kong. Eur J Clin Microbiol Infect Dis 2011,30(11):1371–1381.

41. Yan Q, Zhang J, Chen C, Zhou H, Du P, Cui Z, Cen R, Liu L, Li W, Cao B, et al:Multilocus sequence typing (MLST) analysis of 104 Clostridium difficilestrains isolated from China. Epidemiol Infect 2013, 141(1):195–199.

42. Dingle KE, Griffiths D, Didelot X, Evans J, Vaughan A, Kachrimanidou M,Stoesser N, Jolley KA, Golubchik T, Harding RM, et al: Clinical Clostridiumdifficile: Clonality and pathogenicity locus diversity. PLoS One 2011,6(5):e19993.

43. Chang SC, Su CH, Chou WH, Hwang KP, Chuang YC, Chou MY, Leu HS,Wang FD, Wang JT, Wang LS, et al: Epidemiology of nosocomialClostridium difficile disease in Taiwan. Taiwan Epidemiology Bulletin 2009,25(3):163–177.

44. Chan YY, Lin TY, Huang CT, Deng ST, Wu TL, Leu HS, Chiu CH:Implementation and outcomes of a hospital-wide computerisedantimicrobial stewardship programme in a large medical centre inTaiwan. Int J Antimicrob Agents 2011, 38(6):486–492.

45. Lee YC, Wang JT, Chen AC, Sheng WH, Chang SC, Chen YC: Changingincidence and clinical manifestations of Clostridium difficile-associateddiarrhea detected by combination of glutamate dehydrogenase and toxinassay in Northern Taiwan. J Microbiol Immunol Infect 2012, 45(4):287–295.

46. Wei HL, Kao CW, Wei SH, Tzen JT, Chiou CS: Comparison of PCR ribotypingand multilocus variable-number tandem-repeat analysis (MLVA) forimproved detection of Clostridium difficile. BMC Microbiol 2011, 11(1):217.

47. Lin YC, Huang YT, Tsai PJ, Lee TF, Lee NY, Liao CH, Lin SY, Ko WC, Hsueh PR:Antimicrobial susceptibilities and molecular epidemiology of clinicalisolates of Clostridium difficile in Taiwan. Antimicrob Agents Chemother2011, 55(4):1701–1705.

48. Wong SSY, Ho PL, Woo PCY, Yuen KY: Bacteremia caused byStaphylococci with inducible vancomycin heteroresistance. Clin Infect Dis1999, 29(4):760–767.

49. Cheng VC, Yam WC, Chan JF, To KK, Ho PL, Yuen KY: Clostridium difficileribotype 027 arrives in Hong Kong. Int J Antimicrob Agents 2009,34(5):492–493.

50. Warren CA, Labio E, Destura R, Sevilleja JE, Jamias JD, Daez ML: Clostridiumdifficile and Entamoeba histolytica infections in patients with colitis inthe Philippines. Trans R Soc Trop Med Hyg 2012, 106(7):424–428.

51. Wongwanich S, Ramsiri S, Vanasin B, Khowsaphit P, Tantipatayangkul P,Phan-urai R: Clostridium difficile associated disease in Thailand. SoutheastAsian J Trop Med Public Health 1990, 21(3):367–372.

52. Thamlikitkul V, Danpakdi K, Chokloikaew S: Incidence of diarrhea andClostridium difficile toxin in stools from hospitalized patients receivingclindamycin, beta-lactams, or nonantibiotic medications. J ClinGastroenterol 1996, 22(2):161–163.

53. Wongwanich S, Rugdeekha S, Pongpech P, Dhiraputra C: Detection ofClostridium difficile toxin A and B genes from stool samples of Thaidiarrheal patients by polymerase chain reaction technique. JMed AssocThai 2003, 86(10):970–975.

54. Chotiprasitsakul D, Janvilisri T, Kiertiburanakul S, Watcharananun S,Chankhamhaengdecha S, Hadpanus P, Malathum K: A superior test fordiagnosis of Clostridium difficile-associated diarrhea in resource-limitedsettings. Jpn J Infect Dis 2012, 65(4):326–329.

55. Wongwanich S, Ramsiri S, Kusum M, Warachit P: Clostridium difficileinfections in HIV-positive patients. Southeast Asian J Trop Med PublicHealth 2000, 31(3):537–539.

56. Waywa D, Kongkriengdaj S, Chaidatch S, Tiengrim S, Kowadisaiburana B,Chaikachonpat S, Suwanagool S, Chaiprasert A, Curry A, Bailey W, et al:Protozoan enteric infection in AIDS related diarrhea in Thailand.Southeast Asian J Trop Med Public Health 2001, 32(Suppl 2):151–155.


57. Wongwanich S, Pongpech P, Dhiraputra C, Huttayananont S,Sawanpanyalert P: Characteristics of Clostridium difficile strains isolatedfrom asymptomatic individuals and from diarrheal patients. Clin MicrobiolInfect 2001, 7(8):438–441.

58. Hassan SA, Othman N, Idris FM, Abdul Rahman Z, Maning N, Abdul RahmanR, Tiong CG: Prevalence of Clostridium difficile toxin in diarhoeal stoolsamples of patients from a tertiary hospital in North Eastern PeninsularMalaysia. Med J Malaysia 2012, 67(4):402–405.

59. Oyofo BA, Subekti D, Tjaniadi P, Machpud N, Komalarini S, Setiawan B,Simanjuntak C, Punjabi N, Corwin AL, Wasfy M, et al: Enteropathogensassociated with acute diarrhea in community and hospital patients inJakarta. Indonesia. FEMS Immunol Med Microbiol 2002, 34(2):139–146.

60. Koh TH, Tan AL, Tan ML, Wang G, Song KP: Epidemiology of Clostridiumdifficile infection in a large teaching hospital in Singapore. Pathology(Phila) 2007, 39(4):438–442.

61. Lim PL, Barkham TM, Ling LM, Dimatatac F, Alfred T, Ang B: Increasingincidence of Clostridium difficile-associated disease, Singapore.Emerg Infect Dis 2008, 14(9):1487–1489.

62. Hsu LY, Tan TY, Koh TH, Kwa AL, Krishnan P, Tee NW, Jureen R: Decline inClostridium difficile-associated disease rates in Singapore public hospitals,2006 to 2008. BMC Res Notes 2011, 4(1):77.

63. Lim PL, Ling ML, Lee HY, Koh TH, Tan AL, Kuijper EJ, Goh SS, Low BS, AngLP, Harmanus C, et al: Isolation of the first three cases of Clostridiumdifficile polymerase chain reaction ribotype 027 in Singapore.Singapore Med J 2011, 52(5):361–364.

64. Ayyagari A, Sharma P, Venkateswarlu, Mehta S, Agarwal KC: Prevalence ofClostridium difficile in pseudomembranous and antibiotic-associatedcolitis in north India. J Diarrhoeal Dis Res 1986, 4(3):157–160.

65. Niyogi SK, Bhattacharya SK, Dutta P, Naik TN, De SP, Sen D, Saha MR, DattaD, Nair GB, Mitra U, et al: Prevalence of Clostridium difficile in hospitalisedpatients with acute diarrhoea in Calcutta. J Diarrhoeal Dis Res 1991,9(1):16–19.

66. Dhawan B, Chaudhry R, Sharma N: Incidence of Clostridium difficileinfection: a prospective study in an Indian hospital. J Hosp Infect 1999,43(4):275–280.

67. Chaudhry R, Joshy L, Kumar L, Dhawan B: Changing pattern of Clostridiumdifficile associated diarrhoea in a tertiary care hospital: a 5 yearretrospective study. Indian J Med Res 2008, 127(4):377–382.

68. Ingle M, Deshmukh A, Desai D, Abraham P, Joshi A, Rodrigues C,Mankeshwar R: Prevalence and clinical course of Clostridium difficileinfection in a tertiary-care hospital: a retrospective analysis. Indian JGastroenterol 2011, 30(2):89–93.

69. Vaishnavi C, Singh M: Preliminary investigation of environmentalprevalence of Clostridium difficile affecting inpatients in a north Indianhospital. Indian J Med Microbiol 2012, 30(1):89–92.

70. Niyogi SK, Dutta P, Dutta D, Mitra U, Sikdar S: Clostridium difficile and itscytotoxin in hospitalized children with acute diarrhea. Indian Pediatr 1991,28(10):1129–1132.

71. Gogate A, De A, Nanivadekar R, Mathur M, Saraswathi K, Jog A, Kulkarni MV:Diagnostic role of stool culture & toxin detection in antibiotic associateddiarrhoea due to Clostridium difficile in children. Indian J Med Res 2005,122(6):518–524.

72. Chandra BK, Singh G, Taneja N, Pahil S, Singhi S, Sharma M: DiarrhoeagenicEscherichia coli as a predominant cause of paediatric nosocomialdiarrhoea in India. J Med Microbiol 2012, 61(Pt 6):830–836.

73. Albert MJ, Faruque AS, Faruque SM, Sack RB, Mahalanabis D: Case–controlstudy of enteropathogens associated with childhood diarrhea in Dhaka.Bangladesh. J Clin Microbiol 1999, 37(11):3458–3464.

74. Kato H, Ito Y, van den Berg RJ, Kuijper EJ, Arakawa Y: First isolation ofClostridium difficile 027 in Japan. Euro Surveill 2007, 12(1), E070111.070113.

75. Chen W, Liu L: Future population trends in China: 2005–2050. In TheCentre of Policy Studies (COPS). Monash University: Australia; 2009.

76. Holloway KA: Promoting the rational use of antibiotics. Regional HealthForum. WHO South East Asia Region 2011, 15(1):122–130.

77. Oka K, Osaki T, Hanawa T, Kurata S, Okazaki M, Manzoku T, Takahashi M,Tanaka M, Taguchi H, Watanabe T, et al: Molecular and microbiologicalcharacterization of Clostridium difficile isolates from single, relapse, andreinfection cases. J Clin Microbiol 2012, 50(3):915–921.

78. Kim J, Pai H, Kang J, Seo M-r: Clinical and microbiologic characteristics oftcdA-negative variant Clostridium difficile infections. BMC Infect Dis 2012,12(1):109.

79. Dong D, Zhang L, Chen X, Jiang C, Yu B, Wang X, Peng Y: Antimicrobialsusceptibility and resistance mechanisms of clinical Clostridium difficilefrom a Chinese tertiary hospital. Int J Antimicrob Agents 2013, 41(1):80–84.

80. Nam K-C, Jo C, Lee M: Meat products and consumption culture in theEast. Meat Sci 2010, 86(1):95–102.

81. Coker RJ, Hunter BM, Rudge JW, Liverani M, Hanvoravongchai P: Emerginginfectious diseases in southeast Asia: regional challenges to control.Lancet 2011, 377(9765):599–609.

82. Niwa H, Kato H, Senoh M, Hobo S, Kinoshita Y, Hariu K, Oku K, Kuroda T,Nakai K: Isolation of PCR ribotype 078 Clostridium difficile from five casesof fulminant colitis in thoroughbred racehorses. In 4th InternationalClostridium difficile Symposium. Slovenia: Bled; 2012.

83. Asai T, Usui M, Hiki M, Kawanishi M, Nagai H, Sasaki Y: Clostridium difficileisolated from the fecal contents of s wine in Japan. J Vet Med Sci 2013,75(4):539–541.

doi:10.1186/2047-2994-2-21Cite this article as: Collins et al.: Epidemiology of Clostridium difficileinfection in Asia. Antimicrobial Resistance and Infection Control 2013 2:21.

Submit your next manuscript to BioMed Centraland take full advantage of:

• Convenient online submission

• Thorough peer review

• No space constraints or color figure charges

• Immediate publication on acceptance

• Inclusion in PubMed, CAS, Scopus and Google Scholar

• Research which is freely available for redistribution

Submit your manuscript at www.biomedcentral.com/submit


International Journal of Antimicrobial Agents 43 (2014) 47– 51

Contents lists available at ScienceDirect

International Journal of Antimicrobial Agents

jou rn al hom epage : ht tp : / /www.e lsev ier .com/ locate / i jant imicag

The emergence of community-onset Clostridium difficile infection in atertiary hospital in Singapore: A cause for concern

X.Q. Tana, A.J. Verrall a, R. Jureenb, T.V. Rileyc, D.A. Collinsc, R.T. Linb, M.N. Balmb,D. Chanb, P.A. Tambyaha,∗

a Division of Infectious Diseases, University Medicine Cluster, National University Health System, Singaporeb Department of Laboratory Medicine, National University Health System, Singaporec Microbiology & Immunology School of Pathology & Laboratory Medicine, The University of Western Australia and Division of Microbiology, Nedlands,Perth, WA, Australia

a r t i c l e i n f o

Article history:Received 27 August 2013Accepted 17 September 2013

Keywords:Clostridium difficileRibotypingInfection control

a b s t r a c t

Increasing rates of Clostridium difficile infection (CDI) among those without traditional risk factors havebeen reported mainly in Europe and North America. Here we describe the epidemiology, clinical fea-tures and ribotypes of CDI at National University Hospital (NUH), a 1000-bed tertiary care hospital inSingapore, from December 2011 to May 2012. All laboratory-confirmed CDI cases ≥21 years old whogave informed consent were included. Clinical data were collected prospectively and participants under-went an interviewer-administered questionnaire. Cases were classified by healthcare facility exposureand severity according to the SHEA guidelines. Included cases were also subjected to PCR and wereclassified by ribotype. In total, 66 patients participated in the study, of which 33 (50.0%) were healthcare-facility-associated hospital onset (HCFA-HO). Of the 33 community-onset (CO) cases, 14 (42.4%) wereHCFA-CO, 10 (30.3%) were indeterminate and 9 (27.3%) were community-associated (CA). Of the CAcases, a majority (90.9%) had prior exposure to a healthcare facility within the last 12 weeks. Clinicalcharacteristics, exposures and outcomes were not different between HO-CDI and CO-CDI. Diagnosis wasdelayed in CO-CDI compared with HO-CDI (4 days vs. 1 day; P = 0.014). There was no difference in distri-bution of ribotypes between CO-CDI and HO-CDI, with 053 being most prevalent in both groups. CO-CDIincreasingly contributes to the burden of CDI in NUH. This may reflect a trend in other parts of Asia.Healthcare professionals should be aware of the possible role of outpatient healthcare environments toCDI risk and thus extend control measures to outpatient settings.

© 2013 Elsevier B.V. and the International Society of Chemotherapy. All rights reserved.

1. Introduction

Clostridium difficile is an important nosocomial pathogen glob-ally. In North America, investigators have drawn attention toincreasing rates of C. difficile infection (CDI) among those with-out traditional risk factors such as exposure to healthcare facilitiesor extensive broad-spectrum antibiotics [1]. These community-associated (CA) cases are often not identified by conventionalsurveillance and infection control activities. In North America, theyhave been linked to the 027 ribotype [2], which has also beendescribed in a small cluster in Singapore [3]. To better definethe epidemiology of CA-CDI [4], a prospective, hospital-based,cross-sectional study of CDI cases was undertaken to describe the

∗ Corresponding author at: 1E Kent Ridge Road, NUHS Tower Block, Level 10,Singapore 119228, Singapore. Tel.: +65 6772 4388; fax: +65 6778 3967.

E-mail addresses: [email protected], [email protected] (P.A. Tambyah).

prevalence, risk factors, clinical features and ribotypes of CDI at atertiary care hospital in Singapore.

2. Methods

2.1. Site

The study was conducted in National University Hospital(NUH), a 1000-bed tertiary care hospital in Singapore, with activehaematology, oncology, neurosurgery, cardiothoracic surgery andtransplant services covering both children and adults. Over thestudy period, NUH recorded 152 377 patient-days and 25 544 dis-charges, or a caseload of 304 754 patient-days/year and 51 088discharges/year.

2.2. Case finding

All laboratory-confirmed CDI cases who were ≥21 years oldat time of hospitalisation and who gave informed consent were

0924-8579/$ – see front matter © 2013 Elsevier B.V. and the International Society of Chemotherapy. All rights reserved.http://dx.doi.org/10.1016/j.ijantimicag.2013.09.011

dx.doi.org/10.1016/j.ijantimicag.2013.09.011

http://www.sciencedirect.com/science/journal/09248579

http://www.elsevier.com/locate/ijantimicag

http://crossmark.crossref.org/dialog/?doi=10.1016/j.ijantimicag.2013.09.011&domain=pdf



dx.doi.org/10.1016/j.ijantimicag.2013.09.011

48 X.Q. Tan et al. / International Journal of Antimicrobial Agents 43 (2014) 47– 51

included in the study over the 6-month period from December 2011to May 2012. Clinical information was gathered prospectively fromthe case notes, clinical charts, medical records, laboratory recordsand the patient or legally responsible representative in order tocomplete a standardised questionnaire. Data collected were thentranscribed into an electronic database. Assessment of cause ofdeath and clinical evidence of severity were performed by a sin-gle experienced infectious diseases physician based on pre-definedcriteria.

2.3. Other data

Denominator data such as total number of stool samples sentto the laboratory and number of patients at risk were derived fromthe computerised laboratory and hospital databases, respectively.

2.4. Definitions

Cases of CDI were defined as inpatients with three or moreunformed or loose stools conforming to the shape of a container(diarrhoea) within a 24-h period and positive for C. difficile toxin Aor B, or glutamate dehydrogenase (GDH)-positive and PCR-positivefor toxin-negative cases. Cases were classified by healthcare facil-ity exposure according to the Society for Healthcare Epidemiologyof America (SHEA) guidelines [5]. Healthcare-facility-associated(HCFA) cases included hospital-onset (HCFA-HO) cases whosesymptom onset was in hospital and >48 h after admission, andcommunity-onset (HCFA-CO) cases whose symptoms began within4 weeks of discharge from hospital. Community-associated (CA)cases were those who had not been discharged from any hospi-tal in the preceding 12 weeks with symptom onset within 48 h ofadmission. Cases with symptom onset 4–12 weeks after a hospitaladmission were classified as indeterminate.

The SHEA definitions of severe and severe complicated CDI wereused and thus severe CDI was defined when a patient had one ormore of the following: leukocytosis with a white blood cell count of≥15 000 cells/�L; a serum creatinine level ≥1.5 times higher thanthe pre-morbid level (or if pre-morbid level unknown, 1.5 timesabove the normal range) [5]. Severe complicated cases were thosewith hypotension or shock, ileus or megacolon contributed by CDI.

2.5. Statistical methods

Analysis was conducted using SPSS software v.20 (IBM Corp.,Armonk, NY). Statistical tests were two-sided, and P-values of <0.05were considered to indicate statistical significance.

2.6. Laboratory methods

Testing of stool for C. difficile was performed according to thestandard laboratory procedure in NUH using the C DIFF QUIK CHEKCOMPLETE® test (TECHLAB® Inc., Blacksburg, VA), a combinationlateral flow assay that tests for GDH as well as toxins A and B. Alltoxin-positive or GDH-positive samples on the lateral flow assaywere subjected to PCR using the Xpert® C. difficile test (Cepheid,Sunnyvale, CA) and were ribotyped by PCR [6]. Toxin-positive casesand those that were both GDH-positive and PCR-positive wereincluded. Crude bacterial template DNA for toxin profiling was pre-pared by re-suspension of cells in a 5% (w/v) solution of Chelex-100resin (Sigma–Aldrich, Castle Hill, NSW, Australia). PCR ribotypingwas performed as previously described by Stubbs et al. [7]. PCRribotyping reaction products were concentrated using a QIAGENMinElute® PCR Purification Kit (Ambion Inc., Austin, TX) and runon a QIAxcel® capillary electrophoresis platform (Ambion Inc.).PCR ribotypes were identified by comparison of banding patterns

Fig. 1. Participant recruitment. GDH, glutamate dehydrogenase.

with a reference library, which consisted of a collection of 15 refer-ence strains from the European Centre for Disease Prevention andControl (ECDC), a collection of the most prevalent PCR ribotypescurrently circulating in Australia (Elliott et al., unpublished data)and a selection of binary toxin-positive strains. Dendrogram andcluster analysis of PCR ribotyping band patterns were performedusing the Dice coefficient within BioNumerics software packagev.6.5 (Applied Maths, Sint-Martens-Latem, Belgium). Isolates thatcould not be identified with the available reference library weredesignated with internal nomenclature.

2.7. Ethical approval

Participants or their legally authorised representatives gaveinformed consent to be included in the study. Ethical approval forthe study was granted by the Domain Specific Review Board ofSingapore’s National Healthcare Group.

3. Results

During the study period, 1642 stool samples were sent for C.difficile testing. In total, 158 instances of CDI were identified in stoolspecimens submitted to the clinical laboratory. Of the 158 cases ofCDI identified, 69 were unable to be assessed for eligibility owingto discharge prior to identification by study co-ordinators, death ora language barrier. Of the 89 cases assessed for eligibility, 13 caseswere ineligible as they did not meet the minimum age requirement,4 cases were ineligible as they had not passed three stools in the24 h prior to C. difficile testing and 3 cases were ineligible as theywere recurrent cases. Of the eligible patients, only three declinedto participate in the study, giving a response rate of 95.7%. Thus, 66cases were recruited for the study (Fig. 1).

X.Q. Tan et al. / International Journal of Antimicrobial Agents 43 (2014) 47– 51 49

Table 1Characteristics of Clostridium difficile infection (CDI) cases (n = 66).

Characteristic n (%)a

Age (years) (mean ± S.D.) 62.3 ± 16.1Male 36 (54.5)Significant past medical history 63 (95.5)

Hypertension 48 (72.7)Heart disease 33 (50.0)Type 2 diabetes mellitus 25 (37.9)Chronic renal disease 23 (34.8)Chronic pulmonary disease 18 (27.3)Cancer 16 (24.2)Cerebrovascular disease 11 (16.7)Gastritis 4 (6.1)

Patients with multiple co-morbidities (≥3) 44 (66.7)Any history of previous CDI 8 (12.1)Previous antibiotic use in the past 3 months 33 (50.0)Antibiotic use this admission (prior to onset of diarrhoea) 36 (54.5)Received enteral feeding 8 (12.1)Underwent OGD/colonoscopy/sigmoidoscopy 15 (22.7)

OGD 11 (16.7)Colonoscopy 5 (7.7)Sigmoidoscopy 4 (6.1)

No. of stools in the last 24 h [median (IQR)] 5 (4–7)Time elapsed between diarrhoea and positive C. difficile

testing (days) (median (IQR)3 (1–7)

Severe CDI 27 (40.9)Death 10 (15.2)Previous admission within last 4 weeks 28 (42.4)No. of admissions within the last 3 months [median (IQR)] 1 (0–1)LoS (days from onset of diarrhoea to discharge) [median

(IQR)]20 (11–37)

Treatment givenMetronidazole only 45 (68.2)Vancomycin only 4 (6.1)Both (sequentially) 14 (21.2)None 3 (4.5)

S.D., standard deviation; OGD, oesophagogastroduodenoscopy; IQR, interquartilerange; LoS, length of stay.

a Data are n (%) unless otherwise stated.

3.1. Incidence

The incidence of laboratory-confirmed CDI in this hospital was10.7/10 000 patient-days or 6.38/1000 discharges. The medianlength of stay (LoS) for patients with CDI was 23.5 days; com-paratively, the overall median LoS is 6.1 days for all patients inNUH.

3.2. Demographics and risk factors

The mean age of studied patients was 62.3 years, and 36(54.5%) were male (Table 1). Medical co-morbidities were com-mon among the cases, with 48 (72.7%) having hypertension, 33(50.0%) heart disease and 25 (37.9%) type 2 diabetes mellitus. Eightcases (12.1%) previously had CDI, and 33 (50.0%) had reportedprevious antibiotic use in the preceding 3 months. Twenty-sevenpatients (40.9%) had severe CDI by the SHEA criteria, of which 16/66(24.2%) were severe complicated. Ten cases (15.2%) recruited tothe study died. Of the 33 patients (50.0%) who used antibiotics inthe previous 3 months, 12 (36.4%) had used amoxicillin/clavulanicacid (AMC), 8 (24.2%) had used fluoroquinolones, 5 (15.2%) hadused trimethoprim/sulfamethoxazole (SXT), 2 (6.1%) had usedmetronidazole and 6 (18.2%) had used other antibiotics. Thirty-sixparticipants (54.5%) had antibiotic use during the admission priorto onset of diarrhoea. Of these, 14 (38.9%) had used third-generationcephalosporins (10 ceftriaxone and 4 ceftazidime), 11 (30.6%) hadused piperacillin/tazobactam and 9 (25.0%) each had used AMC,carbapenems (6 meropenem and 3 imipenem), vancomycin (7intravenous and 2 oral) and metronidazole (5 intravenous and

4 oral) during the index admission prior to the onset of diar-rhoea. SXT, fluoroquinolones and clindamycin were used by fewerparticipants; seven (19.4%), six (16.7%) and one (2.8%) patients,respectively. Three patients used other antibiotics not listed above.

3.3. Hospital-onset vs. community-onset Clostridium difficileinfection

Using the SHEA definitions, 33 cases (50.0%) were HCFA-HO and33 cases (50.0%) were CO. Of the 33 CO cases, 14 (42.4%) were HCFA-CO, 10 (30.3%) were indeterminate and 9 (27.3%) were CA. Of the CAcases, a majority (88.9%) had prior exposure to a healthcare facilitywithin the last 12 weeks (within the past 12 weeks, six had visiteda specialist outpatient clinic, two had visited a primary care clinicand only one had not visited any healthcare facility).

Table 2 compares clinical features, exposures and outcomes forCO and HO cases. There was no significant difference in the meanage or sex according to site of onset. Of the nine CA cases, eight(88.9%) reported no recent exposure to antibiotics. The mediantime from onset of diarrhoeal symptoms to the diagnostic specimenbeing sent was significantly longer for CO cases than for HO cases (4days vs. 1 day; P = 0.014). HO cases had a significantly longer medianLoS from the onset of diarrhoea (14 days vs. 22 days; P < 0.01).

3.4. Ribotypes

Ribotypes were available for 61 isolates (92.4%). Of these, 42corresponded to known ribotypes, the predominant strains being053 and 012 (Table 3). Ribotype 053 was the most frequent bothin CO and HO cases (17.9% and 24.2%, respectively). The prevalenceof different ribotypes did not differ between CO-CDI and HO-CDIcases.

4. Discussion

Clostridium difficile has been recognised as an emergingpathogen, especially in Europe and North America. This study sup-ports the high incidence of CDI in Singapore. The rate of 10.7/10 000patient-days or 6.38/1000 discharges is consistent with the increas-ing prevalence observed by others [8]. Whereas Lim et al. reportedan incidence of 1.49–6.64/10 000 patient-days between 2001 and2006 [9], Molton et al. estimated an increase from 4.21/10 000patient-days to 12.08/10 000 patient-days from 2010 to 2012 [10].The decrease suggested in the study by Hsu et al. from 2006 to2008 appears to be an outlier [11]. Factors underlying this increasemay be increasing LoS (from 4.8 days to 5.6 days) and highernumbers of cancer patients (from 21 231 patient-days to 28 838patient-days between 2006 and 2011). Concurrently, use of broad-spectrum antibiotics has increased [12]. Given the attributable costper episode of CDI of US$2470–3669 [13], CDI is clearly a diseasewith great impact in Southeast Asia and there is an urgent need forimproved control in the hospital and beyond.

A high proportion of CA-CDI amongst the CO-CDI was observedin this study, markedly different from previous studies in Singapore[14] and elsewhere in Asia [15], but similar to reports from Sweden(22–28%) [16,17] and the USA (18–20%) [2].

In contrast to Dumyati et al. [2], most CA cases in the currentstudy did not recall prior exposure to antibiotics, but all except onehad actually been treated in an outpatient clinic in the preceding 3months. This apparent absence of traditional risk factors is consis-tent with the findings of Wilcox et al. [18] and may reflect genuinecommunity transmission or definitions that do not capture expo-sure to outpatient healthcare environments. This hypothesis issupported by our finding that similar ribotypes caused the CO andHO cases. Considering these non-traditional types of exposures

50 X.Q. Tan et al. / International Journal of Antimicrobial Agents 43 (2014) 47– 51

Table 2Association of clinical characteristics, exposures and outcomes with community-onset (CO) vs. hospital-onset (HO) Clostridium difficile infection (CDI).

Variable n (%) RR (95% CI)a P-valuea

CO-CDI (n = 33) HO-CDI (n = 33)

Age ≤62 yearsb

Yes 13 (44.8) 16 (55.2) 0.8 (0.5–1.4) 0.46No 20 (54.1) 17 (45.9)Mean (S.D.) 62.9 (18.0) 61.8 (14.2)

SexMale 16 (44.4) 20 (55.6) 0.8 (0.5–1.3) 0.46Female 17 (56.7) 13 (43.3)

History of CDIYes 6 (75.0) 2 (25.0) 1.6 (1.0–2.6) 0.26No 27 (46.6) 31 (53.4)

ICU admissionYes 0 (0.0) 3 (100) 0.5 (0.4–0.6) 0.08No 33 (52.4) 30 (47.6)

DeathYes 4 (40.0) 6 (60.0) 0.8 (0.4–1.7) 0.50No 29 (51.8) 27 (48.2)

Severe CDIYes 13 (48.1) 14 (51.9) 0.9 (0.6–1.5) 0.80No 20 (51.3) 19 (48.7)

Enteral feedingYes 4 (50.0) 4 (50.0) 1.0 (0.5–2.1) 1.00No 29 (50.0) 29 (50.0)

Time elapsed between diarrhoea onset and C. difficile testing ≤3 days c

Yes 14 (37.8) 23 (62.2) 0.58 (0.4–0.9) 0.03No 19 (65.5) 10 (34.5)

Treatment with vancomycinYes 8 (44.4) 10 (55.6) 0.85 (0.45–1.43) 0.58No 25 (52.1) 23 (47.9)

RR, relative risk; CI, confidence interval; S.D., standard deviation; ICU, intensive care unit.a By Fisher’s exact test, comparing proportions. RR measures risk of CO-CDI given the variable.b Age was converted to a categorical variable; 62 years was the mean age and was chosen as the cut-off.c Time elapsed was converted to a categorical variable; 3 days was chosen based on the median number of days elapsed between onset of diarrhoea and testing for CDI,

as the data were highly skewed.

may be more important as more healthcare activities shift tooutpatient settings in Asia as in Europe and North America.

Time from onset of diarrhoea to diagnosis was prolonged in theCO group, perhaps because the lack of traditional risk factors inthe CO group may make physicians less likely to entertain CDI asa cause of diarrhoea. Otherwise, most of the clinical characteristicsand disease severity of CO patients were similar to HO cases. Thelonger LoS for HO patients likely reflects the role of HO-CDI as acomplication of long hospitalisations, not severity per se, as diseaseseverity rates were similar in the two groups.

The severity of CDI was similar both for HO and CO cases. Part ofthis might be because the SHEA definition of severity was used [5],which is different from the McDonald surveillance definition [19].However, it is not clear that there was any real difference in clinicalfeatures between HO and CO cases in any case.

A substantial proportion of the isolates were of novel ribotypesthat have not been previously defined (31.1%), demonstrating thediversity of strains in Singapore. This diversity may be attributableto the high use of imported food from widespread areas toSingapore.

The majority of patients with CDI have 053 or 012 strains;it is not surprising that the most prevalent ribotypes in HO-CDIand CO-CDI patients were similar given the movement of patientsbetween hospitals and the community. Moreover, most patientshad reported recent contact with a healthcare facility.

In previous reports, the 027 ribotype was implicated in a limitednumber of CO-CDI in Singapore [3] and the USA [2]. It is possiblethat we might have missed sporadic cases of 027 CDI but there were

no sustained outbreaks to our knowledge during the study period.The ribotypes most prevalent in Singapore are different from thosereported in other studies in Asia [20–22], North America and Europe[20,23]. In studies in Europe, ribotypes 078, 014 and 172 were mostcommonly found in CO-CDI cases in The Netherlands [24], whilstribotypes 001 and 027 were most common in the UK [25].

A limitation of this study was the number of patients not suc-cessfully recruited before discharge and not invited to participatein the study. It is possible that the patients not invited had a shorterLoS than those recruited because they were less ill. However, thereis no reason to suspect that this would lead to over-representationof CO-CDI.

Table 3Ribotypes by community-onset (CO) and hospital-onset (HO) Clostridium difficileinfection (CDI).

Ribotype n (%) P-valuea

CO-CDI (n = 28) HO-CDI (n = 33) Total (n = 61)

002 1 (3.6) 0 (0.0) 1 (1.6)012 4 (14.3) 7 (21.2) 11 (18.0) 0.53014 2 (7.1) 4 (12.1) 6 (9.8) 0.68017 2 (7.1) 1 (3.0) 3 (4.9) 0.59020 2 (7.1) 3 (9.1) 5 (8.2) 1.00053 5 (17.9) 8 (24.2) 13 (21.3) 0.76064-like 1 (3.6) 0 (0.0) 1 (1.6)087 2 (7.1) 0 (0.0) 2 (3.3)Unknown 9 (32.1) 10 (30.3) 19 (31.1) 1.00

a By Fisher’s exact test, comparing proportions.

X.Q. Tan et al. / International Journal of Antimicrobial Agents 43 (2014) 47– 51 51

5. Conclusion

CO-CDI and CA-CDI are an important often unrecognised con-tributor to the increasing burden of CDI in NUH and possiblyother institutions in Asia. Physicians, infection control teams andresearchers should be aware of the possible contribution of outpa-tient healthcare environments to CDI risk and thus extend infectioncontrol and antibiotic stewardship efforts to outpatient care to con-trol the transmission of C. difficile.

Acknowledgments

The authors would like to thank: Serene Tan, Bu Duo, Ange-line Koh and Tan Teck Seng from the Investigational Medicine Unit,National University Health System (NUHS) for providing assistancewith conduct of the study as well as collection and entry of data;Thomas Riley, Deirdre Collins and their team in the University ofWestern Australia for performing the ribotyping of the samples;and Donald Chiang and team in NUHS for performing laboratorydiagnostics for the identification of C. difficile.

Funding: This work was supported by Sanofi Pasteur through anunrestricted educational grant.

Competing interests: PAT has previously received research sup-port from Sanofi-Pasteur, GSK, Inviragen and Fabentech; he has alsoreceived honoraria from Novartis, MSD and Astra-Zeneca. All otherauthors declare no competing interests.

Ethical approval: Ethical approval was given by the NationalHealthcare Group (NHG) Research & Development Office, NHGDomain Specific Review Board [DSRB ref. no. 2011/02001].

References

[1] Centers for Disease Control and Prevention (CDC). Severe Clostridium difficile-associated disease in populations previously at low risk—four states, 2005.MMWR Morb Mortal Wkly Rep 2005;54:1201–5.

[2] Dumyati G, Stevens V, Hannett GE, Thompson AD, Long C, MacCannell D, et al.Community-associated Clostridium difficile infections, Monroe County, NewYork, USA. Emerg Infect Dis 2012;18:392–400.

[3] Lim PL, Ling ML, Lee HY, Koh TH, Tan AL, Kuijper EJ, et al. Isolation of the firstthree cases of Clostridium difficile polymerase chain reaction ribotype 027 inSingapore. Singapore Med J 2011;52:361–4.

[4] Freeman J, Bauer MP, Baines SD, Corver J, Fawley WN, Goorhuis B, et al. Thechanging epidemiology of Clostridium difficile infections. Clin Microbiol Rev2010;23:529–49.

[5] Cohen SH, Gerding DN, Johnson S, Kelly CP, Loo VG, McDonald LC, et al.Clinical practice guidelines for Clostridium difficile infection in adults: 2010update by the Society for Healthcare Epidemiology of America (SHEA) and theInfectious Diseases Society of America (IDSA). Infect Control Hosp Epidemiol2010;31:431–55.

[6] Elliott B, Squire MM, Thean S, Chang BJ, Brazier JS, Rupnik M, et al. New types oftoxin A-negative, toxin B-positive strains among clinical isolates of Clostridiumdifficile in Australia. J Med Microbiol 2011;60:1108–11.

[7] Stubbs SL, Brazier JS, O’Neill GL, Duerden BI. PCR targeted to the 16S–23SrRNA gene intergenic spacer region of Clostridium difficile and construction ofa library consisting of 116 different PCR ribotypes. J Clin Microbiol 1999;37:461–3.

[8] Miller MA, Hyland M, Ofner-Agostini M, Gourdeau M, Ishak M, Canadian Hos-pital Epidemiology Committee. Canadian Nosocomial Infection SurveillanceProgram, morbidity, mortality, and healthcare burden of nosocomial Clostrid-ium difficile-associated diarrhea in Canadian hospitals. Infect Control HospEpidemiol 2002;23:137–40.

[9] Lim PL, Barkham TM, Ling LM, Dimatatac F, Alfred T, Ang B. Increasing inci-dence of Clostridium difficile-associated disease, Singapore. Emerg Infect Dis2008;14:1487–9.

[10] Molton J, Balm M, Alenton L, Salmon S, Tan XQ, Tambyah PA, et al. Clostrid-ium difficile infection in Singapore; a matter of urgency. In: 13th Asia-PacificCongress of Clinical Microbiology and Infection (APCCMI). 2012.

[11] Hsu LY, Tan TY, Koh TH, Kwa AL, Krishnan P, Tee NW, et al. Decline in Clostridiumdifficile-associated disease rates in Singapore public hospitals, 2006 to 2008.BMC Res Notes 2011;4:77.

[12] Liew YX, Krishnan P, Yeo CL, Tan TY, Lee SY, Lim WP, et al. Surveillance ofbroad-spectrum antibiotic prescription in Singaporean hospitals: a 5-year lon-gitudinal study. PLoS ONE 2011;6:e28751.

[13] Dubberke ER, Butler AM, Reske KA, Agniel D, Olsen MA, D’Angelo G, et al.Attributable outcomes of endemic Clostridium difficile-associated disease innonsurgical patients. Emerg Infect Dis 2008;14:1031–8.

[14] Chan M, Lim PL, Chow A, Win MK, Barkham TM. Surveillance for Clostridiumdifficile infection: ICD-9 coding has poor sensitivity compared to laboratorydiagnosis in hospital patients, Singapore. PLoS ONE 2011;6:e15603.

[15] Bishara J, Peled N, Pitlik S, Samra Z. Mortality of patients with antibiotic-associated diarrhoea: the impact of Clostridium difficile. J Hosp Infect2008;68:308–14.

[16] Norén T, Akerlund T, Bäck E, Sjöberg L, Persson I, Alriksson I, et al. Molecu-lar epidemiology of hospital-associated and community-acquired Clostridiumdifficile infection in a Swedish county. J Clin Microbiol 2004;42:3635–43.

[17] Karlström O, Fryklund B, Tullus K, Burman LG. A prospective nationwide studyof Clostridium difficile-associated diarrhea in Sweden. The Swedish C. difficileStudy Group. Clin Infect Dis 1998;26:141–5.

[18] Wilcox MH, Mooney L, Bendall R, Settle CD, Fawley WN. A case-control study ofcommunity-associated Clostridium difficile infection. J Antimicrob Chemother2008;62:388–96.

[19] McDonald LC, Coignard B, Dubberke E, Song X, Horan T, Kutty PK. Ad HocClostridium difficile Surveillance Working Group. Recommendations for surveil-lance of Clostridium difficile-associated disease. Infect Control Hosp Epidemiol2007;28:140–5.

[20] Huang H, Fang H, Weintraub A, Nord CE. Distinct ribotypes and rates of antimi-crobial drug resistance in Clostridium difficile from Shanghai and Stockholm.Clin Microbiol Infect 2009;15:1170–3.

[21] Sawabe E, Kato H, Osawa K, Chida T, Tojo N, Arakawa Y, et al. Molecular anal-ysis of Clostridium difficile at a university teaching hospital in Japan: a shift inthe predominant type over a five-year period. Eur J Clin Microbiol Infect Dis2007;26:695–703.

[22] Jamal W, Rotimi VO, Brazier J, Duerden BI. Analysis of prevalence, risk factorsand molecular epidemiology of Clostridium difficile infection in Kuwait over a3-year period. Anaerobe 2010;16:560–5.

[23] Vanek J, Hill K, Collins J, Berrington A, Perry J, Inns T, et al. Epidemiologicalsurvey of Clostridium difficile ribotypes in the North East of England during an18-month period. J Hosp Infect 2012;81:209–12.

[24] Bauer MP, Veenendaal D, Verhoef L, Bloembergen P, van Dissel JT, Kuijper EJ.Clinical and microbiological characteristics of community-onset Clostridiumdifficile infection in The Netherlands. Clin Microbiol Infect 2009;15:1087–92.

[25] Bignardi GE, Settle C. Different ribotypes in community-acquired Clostridiumdifficile. J Hosp Infect 2008;70:96–8.

http://refhub.elsevier.com/S0924-8579(13)00343-9/sbref0005


















































































































































































































































































































































































































































































































































































































































































































































































































































Epidemiology of Clostridium difficile infection in two tertiary-care hospitals

in Perth, Western Australia: a cross-sectional study

N. F. Foster1, D. A. Collins1, S. L. Ditchburn2, C. N. Duncan2, J. W. van Schalkwyk2, C. L. Golledge3, A. B. R. Keed4 and

T. V. Riley1,3

1) School of Pathology and Laboratory Medicine, The University of Western Australia, 2) Sir Charles Gairdner Hospital, 3) Division of Microbiology and Infectious

Diseases, PathWest Laboratory Medicine WA, Queen Elizabeth II Medical Centre and 4) Department of Infectious Diseases and Microbiology, Royal Perth

Hospital, Perth, Western Australia, Australia

Abstract

The epidemiology of Clostridium difficile infection (CDI) has changed over time and between countries. It is therefore essential to monitor

the characteristics of patients at risk of infection and the circulating strains to recognize local and global trends, and improve patient

management. From December 2011 to May 2012 we conducted a prospective, observational epidemiological study of patients with

laboratory-confirmed CDI at two tertiary teaching hospitals in Perth, Western Australia to determine CDI incidence and risk factors in an

Australian setting. The incidence of CDI varied from 5.2 to 8.1 cases/10 000 occupied bed days (OBDs) at one hospital and from 3.9 to

16.3/10 000 OBDs at the second hospital. In total, 80 patients with laboratory-confirmed CDI met eligibility criteria and consented to be in

the study. More than half (53.8%) had hospital-onset disease, 28.8% had community-onset and healthcare facility-associated disease and 7.5%

were community-associated infections according to the definitions used. Severe CDI was observed in 40.0% of these cases but the 30-day

mortality rate for all cases was only 2.5%. Besides a shorter length of stay among cases of community-onset CDI, no characteristics were

identified that were significantly associated with community-onset or severe CDI. From 70 isolates, 34 different ribotypes were identified.

The predominant ribotypes were 014 (24.3%), 020 (5.7%), 056 (5.7%) and 070 (5.7%). Whereas this study suggests that the characteristics

of CDI cases in Australia are not markedly different from those in other developed countries, the increase in CDI rate observed emphasizes

the importance of surveillance.

Keywords: Community-associated Clostridium difficile infection, healthcare facility-associated Clostridium difficile infection, molecular

epidemiology, ribotype 244

Original Submission: 27 November 2013; Revised Submission: 20 February 2014; Accepted: 5 March 2014

Article published online: 1 April 2014

New Microbes and New Infections 2014; 2: 64–71

Corresponding author: N. F. Foster, School of Pathology and

Laboratory Medicine (M504), The University of Western Australia, 35

Stirling Highway, Crawley, WA 6009, Australia

E-mail: [email protected]

Introduction

Clostridium difficile is an important nosocomial pathogen and the

most frequently diagnosed cause of infectious hospital-acquired

diarrhoea. Clostridium difficile infection (CDI) has a wide clinical

spectrum, varying from asymptomatic carriage, to mild diar-

rhoea, to pseudomembranous colitis [1]. Clostridium difficile

produces two main toxins, toxin A and toxin B, which belong to

the large clostridial toxin family. The genes for toxins A and B,

tcdA and tcdB, respectively, are located on the chromosome in a

19.6-kb pathogenicity locus togetherwith three accessory genes

[2]. Some strains also produce an actin-specific ADP-ribosyl-

transferase known as binary toxin (CDT) [3], the importance of

which is still not clear. Althoughmost clinically important strains

produce both toxins, toxin A-negative, toxin B-positive (A� B+)

strains have been widely reported [4]. In recent years, an

increase in the frequency and severity of CDI has been

ª 2014 The Authors. New Microbes and New Infections published by John Wiley & Sons Ltd on behalf of the European Society of Clinical Microbiology and Infectious Disease.

This is an open access article under the terms of the Creative Commons Attribution License, which permits use,

distribution and reproduction in any medium, provided the original work is properly cited.

ORIGINAL ARTICLE 10.1002/nmi2.43

associated with the emergence of a binary toxin-producing

strain of C. difficile NAP1/027 (North American PFGE type 1,

UK PCR ribotype 027) [5]. This fluoroquinolone-resistant strain

has been linked to epidemics in North America and Europe.

Major risk factors for clinically apparent CDI include

antimicrobial therapy, hospitalization, residence in a long-term

care facility, older age (≥65 years), and increased length of

hospital stay [1]. Little is known about the epidemiology of

CDI in Australia. Given this lack of information, we undertook

a study of CDI in two large teaching hospitals in Perth,

Western Australia. The primary objectives were to estimate

the incidence of CDI cases for hospitalized adult patients and

to describe the profile of patients with the laboratory-con-

firmed infection. The prevalence of circulating ribotypes was

also determined to understand the molecular epidemiology of

CDI in Western Australia.

Materials and Methods

Setting and study design

This was a prospective, observational, epidemiological study

conducted at the Sir Charles Gairdner Hospital (SCGH), a

600-bed tertiary teaching hospital, and the Royal Perth

Hospital (RPH), an 855-bed tertiary teaching hospital, both

located in Perth, Western Australia. Stool samples sent to the

laboratory for C. difficile testing were monitored from Decem-

ber 2011 to May 2012. All adult patients who submitted loose

stool samples for C. difficile testing over this period were

considered for the study. A minimum target of 35 patients with

laboratory-confirmed CDI per Western Australian hospital

was set. The study was approved by the Sir Charles Gairdner

Group (SCGG) and RPH Human Research Ethics Committees

(SCGG Ref. 2011-133 and RPH Ref. RA-11/036). In Western

Australia, next-of-kin/guardian/carer consent on behalf of the

patient is not acceptable so patients who could not provide

consent were not approached about the study.

Definitions and collection of data

Recently recommended definitions were applied in this study

[6,7]. Patients were classified as having ‘laboratory-confirmed

CDI’ if they experienced the passage of three or more

unformed or loose stools conforming to the shape of a

container (diarrhoea) within a 24-h period and had a C. diffi-

cile-positive laboratory test result. The laboratory testing

method was Becton Dickinson GeneOhmTM, which detects the

toxin B gene (BD, Franklin Lakes, NJ, USA). Specimens from

SCGH patients were screened using a glutamate dehydroge-

nase immunoassay (C. DIFF CHEKTM-60; Alere, Sinnamon Park,

Qld, Australia) before GeneOhm analysis. All specimens that

were glutamate dehydrogenase-negative were regarded as

negative for C. difficile. The consenting patient’s hospital record

was reviewed to confirm the CDI episode. Other data

regarding patient location at onset, risk factors and outcomes

were abstracted from the medical record and by interviewing

the patient using a standardized questionnaire. The number of

occupied bed days (OBDs) was extracted from the respective

hospital administrative databases. Patients fulfilling the case

definition of a laboratory-confirmed CDI were further classi-

fied as having: (1) ‘community-onset CDI’ (when they expe-

rienced diarrhoea in the community or 48 h or less after

admission to the hospital) or ‘hospital-onset CDI’ (when the

onset of diarrhoea was more than 48 h after admission to a

hospital); and (2) ‘healthcare facility (HCF) -associated CDI’

(when they had been discharged from a healthcare facility

within the previous 4 weeks from the onset of CDI symp-

toms), ‘community-associated CDI’ (when they had not been

discharged from a healthcare facility within the previous

12 weeks from the onset of CDI symptoms) or ‘indeterminate

CDI’ (when they had been discharged from a healthcare facility

within the previous 4–12 weeks from the onset of CDI

symptoms). Severe CDI in a patient was defined as one or

more of the following: leucocytosis with a white blood cell

count of >15 000 cells/lL, a serum creatinine level >1.5 times

the pre-morbid level (or if the pre-morbid level was unknown,

1.5 times above the normal range), evidence of severe colitis

(abdominal or radiological signs), temperature >38.5°C,

admission to an intensive care unit for complications associ-

ated with CDI, surgery (e.g. colectomy) for treatment of toxic

megacolon, perforation or refractory colitis due to CDI, death

within 30 days due to or related to CDI.

Clostridium difficile culture and ribotyping

The presence of viable C. difficile was confirmed by culture.

Direct culture of the specimen was performed on CCFA

(cycloserine-cefoxitin-fructose agar containing 0.1% taurocho-

late; PathWest Laboratory Medicine Excel Media [8]). Broth

enrichment in Robertson’s cooked meat medium containing

5 mg/L of gentamicin, 250 mg/L of cycloserine and 8 mg/L of

cefoxitin was performed concurrently and an ethanol-shocked

aliquot was subcultured on CCFA [9,10]. The identity of

putative C. difficile colonies was confirmed by morphology,

odour and chartreuse fluorescence on blood agar, and

L-proline aminopeptidase DIATABSTM (Rosco Diagnostica,

Taastrup, Denmark) reaction.

The PCR ribotyping was performed as previously

described [11]; PCR ribotyping reaction products were

concentrated using the Qiagen MinElute PCR Purification

kit (QIAGEN, Venlo, Limburg, the Netherlands) and resolved

on the QIAxcel capillary electrophoresis platform (QIAGEN).

ª 2014 The Authors. New Microbes and New Infections published by John Wiley & Sons Ltd on behalf of the European Society of Clinical Microbiology and Infectious Disease., NMNI, 2, 64–71

NMNI Foster et al. C. difficile in Perth, Australia 65

Cluster analysis of PCR ribotyping band profiles was performed

using the Dice similarity coefficient with relationships repre-

sented in a UPGMA dendrogram within BIONUMERICSTM software

package v.6.5 (Applied Maths, Saint-Martens-Latem, Belgium).

Ribotypes were identified by comparison of the band profiles

with our reference library, which consisted of a collection of 50

UK ribotypes that included 15 reference strains from the

European Centre for Disease Prevention and Control and the

most prevalent PCR ribotypes currently circulating in Australia

(B. Elliott, T. V. Riley et al., unpublished data). Isolates that could

not be identified with the available reference library were

designated with internal (QX) nomenclature.

Statistical methods

GRAPHPAD INSTAT V3.06 was used to calculate prevalence ratios

and their 95% confidence intervals, and to perform Fisher’s

exact and chi-square analyses. A value of p ≤0.05 was

considered statistically significant.

Results

A total of 2170 specimens from RPH and 2248 specimens from

SCGH were tested for C. difficile over the study period

(Fig. 1). At RPH, the proportion of tests that were

PCR-positive varied from a low of 6.3% in March to a high

of 9.0% in May, with an average of 8.1% positive. At SCGH, the

proportion of PCR-positive tests varied from a high of 13.6% in

February to a low of 7.1% in April, with an average of 10.1%

positive. The difference in proportion of tests positive

between RPH and SCGH was statistically significant in

February 2012 (chi-squared p <0.0389).

The incidence of CDI was estimated from the number of

new patients with C. difficile testing requested who had stool

specimens that were positive for C. difficile. At RPH and

SCGH throughout the study period this reflected the

proportion of specimens tested that were positive for

C. difficile (Fig. 1). There was little variation at RPH during

the nearly 6-month study period with rates varying from 5.2

to 8.1 cases/10 000 OBDs. However, at SCGH, the rates

fluctuated from a low of 3.9/10 000 OBDs in December

2011 to a high of 16.3/10 000 OBDs in January 2012 before

dropping to rates similar to RPH for March, April and May

2012. The average rate at RPH for the study period was

6.8 cases/10 000 OBDs while the average rate at SCGH was

8.0 cases/10 000 OBDs.

From 331 PCR-positive stool specimens from patients at

RPH and SCGH, 80 cases were recruited for further study

(Fig. 2). Initially 154 were excluded because the stools were

not loose or watery (n = 86) or they were repeat samples

from previously PCR-positive patients (68). From the remain-

ing 177 patients for possible inclusion in the study, a total of 88

patients were recruited, with the main reason for exclusion

being inability to consent (58 patients, 32.7%). Of the 88

patients recruited, eight were withdrawn, the main reason

being that they did not meet the definition of CDI (five

patients). Of eligible patients, only six declined to participate in

the study, giving a response rate of 93.6%.

The characteristics of the 80 patients with CDI are shown in

Table 1. The median age of the patients was 60.5 years of age

and just over half (51.3%) were male. Overall, 26.3% of people

with CDI had three or more co-morbidities, with the most

common being: oesophageal reflux (46.3%), hypertension

0

2

4

6

8

10

12

14

16

18

0

2

4

6

8

10

12

14

16

Case

s per

10,

000

OBD

% sp

ecim

ens P

CR-p

osi�

ve

MonthRPH % specimens posi�ve (n=2170) SCGH % specimens posi�ve (n=2248)

RPH Rate/10,000 OBD SCGH Rate/10,000 OBD

MayAprMarFebJanDec

FIG. 1. Rates of Clostridium difficile infection (CDI) and proportion of

specimens positive for C. difficile at Royal Perth Hospital and Sir

Charles Gairdner Hospital during the study period, by month. FIG. 2. Recruitment strategy.


66 New Microbes and New Infections, Volume 2 Number 3, May 2014 NMNI

(42.5%), chronic pulmonary disease (36.3%), chronic renal

disease (36%), cancer (35%), immune deficiency (35%) and

rheumatological disease (31.3%). A history of CDI was noted

in 17.5% of patients and 13.8% of patients had undergone a

colonoscopy, sigmoidoscopy or oesophagogastroduodenos-

copy during the admission. The median length of stay for

patients who acquired CDI was 15 days compared with a

median length of stay of 6.0 and 6.1 days for patients at RPH

and SCGH, respectively, during the same period.

Most patients (62.5%) recalled taking antibiotics within the

preceding 3 months (Table 1). The most commonly consumed

antibiotics preceding CDI were amoxicillin / amoxicillin

clavulanate (12.5%) and cephalosporins (11.3%). Just over half

(52.5%) of the CDI patients had consumed antibiotics during

the admission when CDI was diagnosed. The most common

antibiotics received for these patients were: piperacillin

tazobactam (8.8% of patients in the study had previous

exposure), cephalosporins (6.3%), amoxicillin clavulanate

(6.3%) and meropenem (6.3%). Vancomycin, commonly used

in the treatment of CDI, was also implicated in four patients

(5%) before the onset of diarrhoea.

Metronidazole predominated as the treatment of choice for

CDI with 58.8% of patients receiving solely metronidazole and

26.3% receiving both metronidazole and vancomycin (Table 1).

Interestingly, only five patients (6.3%) received vancomycin

alone despite a number of severe cases.

A total of 32 patients (40.0%) in the study had severe CDI

as defined, however, only two patients recruited into the

study (2.5%) died within 30 days. The 90-day crude mortality

rate was 16.3%; six of the 13 deaths were among patients

with severe CDI. A comparison of clinical signs of CDI in

patients with severe versus non-severe disease is given in

Table 2.

Of the 80 CDI cases 55% were readmissions, defined as

having a previous hospital admission for any reason within

the last 4 weeks. The median number of admissions in the

last 3 months was 1 (interquartile range, 0–2.3). More than

half the CDI cases (53.8%) were hospital-onset and 32.6% of

these met the definition of severe CDI. Of the commu-

nity-onset cases (46.3%), 23 (62.2%) were HCF-associated, 6

(16.2%) were community-associated, and 8 (21.6%) were of

indeterminate nature as defined. A comparison between

hospital and community-onset CDI is given in Table 3. There

were no statistically significant differences in risk factors or

outcomes examined other than length of stay where patients

TABLE 1. Characteristics of Clostridium difficile infection

cases

CharacteristicNumber ofcases (%) (n = 80)

Age (years)Median 60.5Interquartile range 49–71

Male 41 (51.3)EthnicityCaucasian 76 (95.0)Aboriginal 3 (4.8)Asian 1 (1.3)

Significant past medical historyOesophageal reflux 37 (46.3)Hypertension 34 (42.5)Chronic pulmonary disease 32 (40.0)Chronic renal disease 29 (36.3)Cancer 28 (35.0)Immune deficiency 28 (35.0)Rheumatological disease 25 (31.3)Chronic liver disease 22 (27.5)Heart disease 19 (23.8)Gastritis 16 (20.0)Type 2 diabetes mellitus 13 (16.3)Colitis 10 (12.5)Gastric ulcer 8 (10.0)Connective tissue disorder 8 (10.0)Irritable bowel disease 8 (10.0)Diverticulitis 7 (8.8)Cerebrovascular disease 2 (2.5)Type 1 diabetes 1 (1.3)

History of CDI 14 (17.5)Previous antibiotic use in the past 3 months 50 (62.5)Amoxicillin/amoxicillin clavulanate 10 (12.5)Cephalosporins 9 (11.3)Metronidazole 6 (7.5)Piperacillin tazobactam 5 (6.3)Vancomycin 4 (5.0)Clindamycin 4 (5.0)Trimethoprim sulfamethoxazole 2 (2.5)Fluoroquinolones 1 (1.3)Others 27 (33.8)

Duration of previous antibiotic use (days)Median 14Interquartile range 8–49

Antibiotic use this admission (before specimen collection) 42 (52.5)Piperacillin tazobactam 7 (8.8)Amoxicillin clavulanate 5 (6.3)Meropenem 5 (6.3)Vancomycin 4 (5.0)Fluoroquinolones 4 (5.0)Ciprofloxacin 3 (3.8)Moxifloxacin 1 (1.3)

Cephazolin 3 (3.8)Metronidazole 1 (1.3)Fluconazole 1 (1.3)Ceftriaxone 1 (1.3)Cephalexin 1 (1.3)

Received enteral feeding 1 (1.3)Underwent OGD/colonoscopy/sigmoidoscopy 11 (13.8)OGD 7 (8.8)Colonoscopy 3 (3.8)Sigmoidoscopy 2 (2.5)Number of stools in the last 24 hMedian 5.5Interquartile range 4–10

LOS (days)Median 15Interquartile range 7.5–40

CDI, Clostridium difficile infection; OGD, oesophagogastroduodenoscopy; LOS,length of stay.

TABLE 2. Clinical signs of Clostridium difficile infection

Clinical sign All CDI (%)

Non-severeCDI (%)(n = 48)

SevereCDI (%)(n = 32)

Prevalenceratio (95% CI)

Dehydration 53 (66.3) 31 (64.6) 22 (68.8) 1.07 (0.78–1.46)Hypotension 14 (17.5) 7 (14.6) 7 (21.9) 1.50 (0.58–3.87)Ileus 2 (2.5) 0 2 (6.3)Loss of appetite 68 (85.0) 42 (87.5) 26 (81.3) 0.93 (0.76–1.13)Malaise 49 (61.3) 27 (56.3) 22 (68.8) 1.22 (0.87–1.72)Megacolon 1 (1.3) 0 1 (3.1)Nausea 35 (43.8) 18 (37.5) 17 (53.1) 1.42 (0.87–2.31)Weight loss 48 (60.0) 29 (60.4) 19 (59.4) 0.98 (0.68–1.42)

CDI, Clostridium difficile infection.



with community-onset CDI stayed for significantly less time

than patients with hospital-onset CDI. Interestingly, severe

CDI was more common among community-onset cases but

this did not reach statistical significance (Fisher’s exact

p 0.173).

Of the 80 cases of laboratory-confirmed CDI, C. difficile was

isolated from 70, 34 at RPH and 36 at SCGH. The ribotypes of

these 70 isolates are shown in Fig. 3. Of the 32 ribotypes

present, the most common was ribotype 014 accounting for

one-quarter (24.3%) of isolates. The next most common were

ribotypes 020, 056 and 070 each at 5.7%, followed by ribotypes

002, 244 and QX 033 (103-like), all at 4.3%. Eighteen (25.7%)

ribotypes were unique, of which 11 could not be assigned a UK

ribotype from our reference collection.

Discussion

Little is known about the epidemiology of CDI in Australia. A

10-year study at SCGH in Western Australia between 1983

and 1992, showed that the incidence of CDI increased from

~3 cases per 10 000 OBDs in 1983 to nearly 7 cases per

10 000 OBDs in 1988 before stabilizing [12]. A significant

decrease in the incidence of CDI at SCGH was observed

10 years later in 1998 when the use of third-generation

cephalosporins was restricted within the hospital, and rates

declined to <2/10 000 OBDs [13]. Recently, routine surveil-

lance of CDI commenced in Victorian public hospitals and a

rate of CDI of 2.2/10 000 OBDs was reported for October

2010 to March 2011 [14]. The rates seen in the current study

were similar to those seen at SCGH in the late 1980s and

exceeded those recently reported in Victoria. The average

rate at RPH for the study period was 6.8 cases/10 000 OBDs

while the average rate at SCGH was 8.0 cases/10 000 OBDs

(Table 1). These rates are similar to those reported for the

Province of Ontario (8.1/10 000 patient days) in Canada in a

study performed from November 2004 until April 2005 but

less than the rates reported for Quebec Province in the same

study (13/10 000 patient days) [15]. It is interesting to note

however that the rate at SCGH in January 2012 (16.3/

10 000 OBDs) exceeded this latter figure. This very high rate

could not be explained by any obvious clustering of cases in

certain wards or by an excess of particular strains, and

requires further investigation. Differences between rates at

RPH and SCGH were also reflected in the higher proportion

of samples that were positive at SCGH (Fig. 1), so may be due

to a difference in thresholds for patient testing. Molton et al.

[16] reported a similar increase in Singapore. CDI rates

increased three-fold, from 4.2 per 10 000 patient days to 12.1

per 10 000 patient days from March 2010 to April 2012.

Community-acquired CDI has been investigated in Western

Australia previously. In a study of diarrhoeal disease with

community acquisition and onset, C. difficile was the second

most commonly detected pathogen after Campylobacter spe-

cies [17]. In the present study, 28.8% of the CDI cases were

community-onset, HCF-associated, 7.5% were commu-

nity-associated and 10.0% were of indeterminate nature as

defined. A similar breakdown of cases was reported from the

Victorian public hospital surveillance system where the same

criteria were used for the time/place of onset of cases [14].

This is likely to be an under-representation of commu-

nity-acquired CDI. The definitions used for this study [6] mean

that any patient who develops CDI within 4 weeks of

discharge from an HCF is classified as having HCF-associated,

community-onset disease, and between 4 and 12 weeks the

TABLE 3. A comparison of hospital-onset and commu-

nity-onset Clostridium difficile infection

Risk factor

Hospital-onsetCDI (%)(n = 43)

Community-onsetCDI (%)(n = 37)

Prevalenceratio (95% CI)

Age ≥65 years 17 (39.5) 15 (40.5) 1.03 (0.6–1.76)Male 20 (46.5) 21 (56.8) 1.22 (0.8–1.87)Two ormore chronicdiseases

25 (58.1) 15 (40.5) 0.70 (0.48–1.11)

History of CDI 6 (14.0) 8 (21.6) 1.55 (0.6–4.06)Severe CDI 14 (32.6) 18 (48.6) 1.49 (0.87–2.57)Previous antibioticuse in the past3 months

25 (58.1) 25 (67.6) 1.16 (0.83–1.63)

Antibiotic use thisadmission (beforespecimencollection)

26 (60.5) 16 (43.2) 0.72 (0.46–1.11)

Readmission 21 (48.8) 23 (62.2) 1.27 (0.86–1.89)LOS ≥20 days 26 (60.5) 8 (21.6) 0.36 (0.19–0.69)

CDI, Clostridium difficile infection; LOS, length of stay.

0

2

4

6

8

10

12

14

16

18

20

No.

of i

sola

tes

Ribotype

FIG. 3. Ribotypes of Clostridium difficile isolated from study partici-

pants at Royal Perth Hospital and Sir Charles Gairdner Hospital during

the study period. Unique ribotypes (*) included 001, 015, 054, 087,

103, 237, 297 and 014/020-group, and local ribotypes QX 001, 013,

024, 050, 081, 097, 103, 113, 161 and 227.



attribution was indeterminate. However, many of these cases

will be true community-acquired infection given that there is

now good evidence that the rates of community-acquired CDI

around the world are increasing [18,19]. It was noteworthy

that 19% of cases were haematology/oncology patients who,

because of their frequent visits to healthcare facilities, are

always classified as having HCF-associated CDI (data not

shown). However, many of these patients are likely to have

acquired their infection in the community.

In Australia, CDI has been driven by exposure to cepha-

losporins [13]. Not surprisingly, in the present study, previous

exposure to broad-spectrum antibiotics (piperacillin tazobac-

tam, amoxicillin clavulanate, cephalosporins and carbapenems)

was again associated with CDI. This association is similar to

other studies worldwide where cephalosporins as well as

other broad-spectrum antibiotics are the most frequently

implicated antibiotics [13,20,21].

Antibiotic use in the preceding 3 months was the only

significant risk factor identified by Leonard et al. [22] in one of

only two Australian studies assessing risk factors for CDI. In

this small case–control study of diarrhoeal patients at RPH

from 2009–2010, the crude mortality rate (12%) was similar to

the 90-day mortality rate observed in the present study. In

Tasmania in 1997, severe underlying disease, renal impairment,

exposure to antineoplastic agents, and the use of total

parenteral nutrition or nasogastric feeding, all well-known risk

factors for CDI [20], also increased the risk of developing CDI

[23]. The crude mortality rate was 17.2%, and factors

associated with a poor prognosis were older age, severe

underlying disease, renal impairment and failure to treat with

metronidazole or vancomycin. It would appear that little has

changed in Australia in relation to risk factors.

Although over a third of cases were considered severe, the

30-day mortality rate in the present study (2.5%) was low

compared with the 30-day attributable mortalities in studies

from North America: 6.9% overall in Quebec, Canada, during a

ribotype 027 outbreak [24]. This may reflect the lack of

ribotype 027. However, as noted in Fig. 2, a third of patients

assessed were not capable of consenting to inclusion in the

study. In many instances this was because they were too ill and

therefore it is possible that the mortality rate could be higher.

As part of a study that evaluated a new treatment for CDI,

Cheknis et al. [25] typed 24 Australian C. difficile isolates

recovered in the mid-2000s by restriction endonuclease

analysis and found that two-thirds of them were uncommon

types (compared with isolates from North America and

Europe). A quarter of the isolates belonged to restriction

endonuclease analysis group Y, a group that corresponds to

PCR ribotypes 014 and 020. PCR ribotyping is a widely used

and highly discriminatory method for investigating intra-species

variation among C. difficile isolates. In the present study, 32

different ribotypes were identified among 70 isolates from

consecutive patients, emphasizing the diversity of strains in

Australia. Ribotype 014 was again the most commonly

detected at a prevalence of 24.3%. This ribotype seems well

adapted to surviving in the hospital environment as it was

common in the UK before the emergence of the ribotype 027

epidemic strain [11] and is still common in Europe (~10%)

[26,27]. The study by Stubbs et al. [11] from the UK Anaerobe

Reference Unit of 2030 C. difficile strains distinguished 116

types while a study of 330 isolates in a Swedish county

distinguished 53 types [28].

Besides ribotype 014 (which is often grouped with 020 due to

difficulty distinguishing the two), the most prevalent ribotypes in

Europe appear to be 001 (~9%), 078 (~8%), 018 (~6%) and 106

(~5%). A review of ribotyping results across Asia suggested that

017, 018, 014, 002 and 001 were the most prevalent in the

region [29]. In North America, ribotype 027 predominates

(~29%); other common ribotypes include 014 or 020 (~12%),

002 (~5%), 053 (~5%) and 106 (~5%) [30,31]. With such a high

prevalence of 014 and 020 in the present study, other ribotypes

were generally less common than elsewhere with the exception

of ribotypes 056 and 070 (both 6%). Ribotype 056 has been

associated with more complicated infections [27].

Recently, C. difficile ribotype 027 was isolated for the first

time in Australia from a patient who had returned from the USA

[32] and, early in 2010, the first cluster of ribotype 027 cases

was detected in Melbourne, Victoria [33]. Interestingly, C. dif-

ficile ribotype 027 does not appear to have established in

Australia and none was detected in the present study. It is

possible that antibiotic prescribing practices in Australia have

not favoured the emergence of this epidemic strain as the use of

later fluoroquinolones is quite restricted. It is also possible that

the geographic isolation of Australia is also responsible for the

delayed appearance of ribotype 027. However, another poten-

tially epidemic ribotype that produces binary toxin, ribotype

244, was detected at a prevalence of 4.3%. Although this

appears low, it was the equal third most common ribotype

found, and one that was not in Australia 2 years ago. Preliminary

data from other States of Australia, and New Zealand, suggest

that ribotype 244 infection may be associated with more severe

disease [34]. Also of interest, ribotype 237 (A� B+ CDT+), a

recently described strain from Australian livestock [35], was

identified in one patient. Ribotype 078, the predominant

livestock strain, which is associated with community-acquired

infection in the northern hemisphere [19], was not seen. This

strain was absent in recent Australian livestock surveys [36,37].

This study has highlighted a significant increase in the rate of

CDI in Western Australia. Surveillance data from other

Australian States suggest that similar increases have occurred



Australia-wide, and that this may be at least partly due to

community-acquired CDI [38,39]. The contribution of com-

munity-acquired CDI to overall rates may also be masked by

the current definitions used. The emergence of at least one

new virulent strain of C. difficile is a concern, and this needs to

be monitored. Risk factors for acquiring C. difficile in hospital

appear to be unchanged; however, risk factors for community

acquisition require further investigation.

Acknowledgements

This study was funded by an unrestricted educational grant

from Sanofi Pasteur for which the authors are grateful.

Conflict of Interest

None declared.

References

1. Elliott B, Chang BJ, Golledge CL, Riley TV. Clostridium difficile-associated

diarrhoea. Intern Med J 2007; 37: 561–568.

2. Hammond GA, Johnson JL. The toxigenic element of Clostridium difficile

strain VPI 10463. Microb Pathog 1995; 19: 203–213.

3. Perelle S, Gibert M, Bourlioux P, Corthier G, Popoff MR. Production of

a complete binary toxin (actin-specific ADP-ribosyltransferase) by

Clostridium difficile CD196. Infect Immun 1997; 65: 1402–1407.

4. Drudy D, Fanning S, Kyne L. Toxin A-negative, toxin B-positive

Clostridium difficile. Int J Infect Dis 2007; 11: 5–10.

5. Kuijper EJ, Coignard B, Tull P. Emergence of Clostridium difficile-asso-

ciated disease in North America and Europe. Clin Microbiol Infect 2006;

12(suppl 6): 2–18.

6. McDonald LC, Coignard B, Dubberke E, Song X, Horan T, Kutty PK.

Recommendations for surveillance of Clostridium difficile-associated

disease. Infect Control Hosp Epidemiol 2007; 28: 140–145.

7. Cohen SH, Gerding DN, Johnson S et al. Clinical practice guidelines for

Clostridium difficile infection in adults: 2010 update by the society for

healthcare epidemiology of America (SHEA) and the infectious diseases

society of America (IDSA). Infect Control Hosp Epidemiol 2010; 31: 431–

455.

8. Foster NF, Riley TV. Improved recovery of Clostridium difficile spores

with the incorporation of synthetic taurocholate in cycloserine-cefox-

itin-fructose agar (CCFA). Pathology 2012; 44: 354–356.

9. Bowman RA, Riley TV. Laboratory diagnosis of Clostridium difficile-asso-

ciated diarrhoea. Eur J Clin Microbiol Infect Dis 1988; 7: 476–484.

10. Carroll SM, Bowman RA, Riley TV. A selective broth for Clostridium

difficile. Pathology 1983; 15: 165–167.

11. Stubbs SL, Brazier JS, O’Neill GL, Duerden BI. PCR targeted to the

16S-23S rRNA gene intergenic spacer region of Clostridium difficile and

construction of a library consisting of 116 different PCR ribotypes. J

Clin Microbiol 1999; 37: 461–463.

12. Riley TV, O’Neill GL, Bowman RA, Golledge CL. Clostridium diffi-

cile-associated diarrhoea: epidemiological data from Western Australia.

Epidemiol Infect 1994; 113: 13–20.

13. Thomas C, Stevenson M, Williamson DJ, Riley TV. Clostridium

difficile-associated diarrhea: epidemiological data from Western Aus-

tralia associated with a modified antibiotic policy. Clin Infect Dis 2002;

35: 1457–1462.

14. Bull AL, Worth LJ, Richards MJ. Implementation of standardised

surveillance for Clostridium difficile infections in Australia: initial report

from the Victorian Healthcare Associated Infection Surveillance

System. Intern Med J 2012; 42: 715–718.

15. Gravel D, Miller M, Simor A et al. Health care-associated Clostridium

difficile infection in adults admitted to acute care hospitals in Canada: a

Canadian Nosocomial Infection Surveillance Program Study. Clin Infect

Dis 2009; 48: 568–576.

16. Molton J, Balm M, Alen L, Salmon S. Clostridium difficile infection in

Singapore; a matter of urgency. 13th Asia-Pacific Congress of Clinical

Microbiology and Infection. Beijing, 2012.

17. Riley TV, Cooper M, Bell B, Golledge CL. Community-acquired

Clostridium difficile-associated diarrhea. Clin Infect Dis 1995; 20(suppl

2): S263–S265.

18. Khanna S, Pardi DS, Aronson SL et al. The epidemiology of commu-

nity-acquired Clostridium difficile infection: a population-based study. Am

J Gastroenterol 2012; 107: 89–95.

19. Hensgens MP, Keessen EC, Squire MM et al. Clostridium difficile

infection in the community: a zoonotic disease? Clin Microbiol Infect

2012; 18: 635–645.

20. Bartlett JG. Narrative review: the new epidemic of Clostridium

difficile-associated enteric disease. Ann Intern Med 2006; 145: 758–764.

21. Stevens V, Dumyati G, Fine LS, Fisher SG, van Wijngaarden E.

Cumulative antibiotic exposures over time and the risk of Clostridium

difficile infection. Clin Infect Dis 2011; 53: 42–48.

22. Leonard AD, Ho KM, Flexman J. Proton pump inhibitors and diarrhoea

related to Clostridium difficile infection in hospitalised patients: a

case-control study. Intern Med J 2012; 42: 591–594.

23. Halim HA, Peterson GM, Friesen WT, Ott AK. Case-controlled review

of Clostridium difficile-associated diarrhoea in southern Tasmania. J Clin

Pharm Ther 1997; 22: 391–397.

24. Loo VG, Poirier L, Miller MA et al. A predominantly clonal multi-insti-

tutional outbreak of Clostridium difficile-associated diarrhea with high

morbidity and mortality. N Engl J Med 2005; 353: 2442–2449.

25. Cheknis AK, Sambol SP, Davidson DM et al. Distribution of Clostridium

difficile strains from a North American, European and Australian trial of

treatment for C. difficile infections: 2005–2007. Anaerobe 2009; 15: 230–

233.

26. Barbut F, Mastrantonio P, Delmee M, Brazier J, Kuijper E, Poxton I.

Prospective study of Clostridium difficile infections in Europe with

phenotypic and genotypic characterisation of the isolates. Clin Microbiol

Infect 2007; 13: 1048–1057.

27. Bauer MP, Notermans DW, van Benthem BH et al. Clostridium

difficile infection in Europe: a hospital-based survey. Lancet 2011; 377:

63–73.

28. Noren T, Akerlund T, Back E et al. Molecular epidemiology of

hospital-associated and community-acquired Clostridium difficile infec-

tion in a Swedish county. J Clin Microbiol 2004; 42: 3635–3643.

29. Collins DA, Hawkey PM, Riley TV. Epidemiology of Clostridium difficile

infection in Asia. Antimicrob Resist Infect Control 2013; 2: 21.

30. Tenover FC, Novak-Weekley S, Woods CW et al. Impact of strain

type on detection of toxigenic Clostridium difficile: comparison of

molecular diagnostic and enzyme immunoassay approaches. J Clin

Microbiol 2010; 48: 3719–3724.

31. Waslawski S, Lo ES, Ewing SA et al. Clostridium difficile ribotype diversity

at six health care institutions in the United States. J Clin Microbiol 2013;

51: 1938–1941.

32. Riley TV, Thean S, Hool G, Golledge CL. First Australian isolation of

epidemic Clostridium difficile PCR ribotype 027. Med J Aust 2009; 190:

706–708.



33. Richards M, Knox J, Elliott B et al. Severe infection with Clostridium

difficile PCR ribotype 027 acquired in Melbourne, Australia. Med J Aust

2011; 194: 369–371.

34. De Almeida M, Heffernan H, Dervan A et al. Severe Clostridium difficile

infection in New Zealand associated with an emerging strain,

PCR-ribotype 244. N Z Med J 2013; 126: 9–14.

35. Squire MM, Carter GP, Mackin KE et al. Novel molecular type of

Clostridium difficile in neonatal pigs, Western Australia. Emerg Infect Dis

2013; 19: 790–792.

36. Knight DR, Riley TV. Prevalence of Clostridium difficile gastrointestinal

carriage in Australian sheep and lambs. Appl Environ Microbiol 2013; 79:

5689–5692.

37. Knight DR, Thean S, Putsathit P, Fenwick S, Riley TV. Cross-sectional

study reveals high prevalence of Clostridium difficile non-PCR ribotype

078 strains in australian veal calves at slaughter. Appl Environ Microbiol

2013; 79: 2630–2635.

38. Mitchell BG, Wilson F, McGregor A. An increase in community onset

Clostridium difficile infection: a population-based study, Tasmania,

Australia. Healthc Infect 2012; 17: 127–132.

39. Slimings C, Armstrong P, Beckingham W et al. Increasing incidence of

Clostridium difficile infection, Australia, 2011–2012. Med J Aust 2014;

200: 272–276.



Copyright © Royal College of pathologists of Australasia. Unauthorized reproduction of this article is prohibited.

Molecular methods for detecting and typing of Clostridium difficile

DEIRDRE A. COLLINS1, BRIONY ELLIOTT

1AND THOMAS V. RILEY

1,2

1School of Pathology and Laboratory Medicine, University of Western Australia, and 2Department of Microbiology,PathWest Laboratory Medicine (WA), Perth, Australia

Summary

Since the early 2000s, Clostridium difficile has emerged as amajor international pathogen. Recently, strains of C. difficile incirculation appear to be changing, with greater diversity,leading to challenges for diagnostics and surveillance. Cur-rently molecular diagnostic methods are favoured for theirhigh sensitivity and rapid processing times; however, a num-ber of issues still exist with molecular tests, in particular highcost, low clinical specificity and failure to detect some variantC. difficile strains. Molecular typing methods are used todetermine the continually evolving epidemiology of C. difficileinfection. Typing methods including PCR ribotyping andpulsed field gel electrophoresis are currently popular inEurope and North America, respectively, while high-through-put next-generation sequencing is likely to become morewidely used in years to come. This review discusses currentmolecular detection and typing techniques for C. difficile.

Key words: Clostridium difficile, diagnostics, molecular testing, PCR,

ribotyping, sequencing, surveillance, typing.

Received 13 November 2014, revised 27 January, accepted 28 January 2015

INTRODUCTION

Following widespread use of broad spectrum cephalosporinantimicrobials during the 1980s and 1990s, Clostridium diffi-cile infection (CDI) became one of the most common causes ofnosocomial diarrhoea in the world.1 Diarrhoea is a commoncomplaint among hospital patients and can result from under-lying disease or as a side effect of treatment with various drugs,particularly antimicrobials. Clostridium difficile causes20–25% of cases of antimicrobial-associated diarrhoea inhospital patients.2,3 In recent years, community-acquiredCDI has been increasing.4,5 Rapid diagnosis of CDI is desirableto allow early isolation and treatment of patients, reducingpotential patient-to-patient transmission and length of hospitalstay for those affected. In addition, C. difficile strain typing canidentify outbreaks within a hospital or the wider community.Over the past 30 years, significant advances have been made inmolecular techniques for C. difficile detection and typing.

C. difficile pathogenicity

The main virulence factors of C. difficile are toxin A (TcdA/enterotoxin) and toxin B (TcdB/cytotoxin). These are encodedby the genes tcdA and tcdB on the pathogenicity locus (PaLoc),which contains three other genes that regulate expressionof tcdA and tcdB and promote toxin release (Fig. 1).6 Toxigenicstrains always produce toxin B, often produce toxin A also,

and cause disease. Non-toxigenic strains do not produce toxinA or B.6

Variant C. difficile strains differ from the reference strain(VPI 10463) in restriction sites and length of tcdA and tcdB, andother PaLoc regions. Many groups of variant strains have beenidentified and can be defined by a series of overlappingpolymerase chain reactions (PCRs) spanning the PaLoc, allow-ing strains to be assigned to different toxinotypes I-XXXI.7

A third toxin produced by some strains of C. difficile isbinary toxin (C. difficile transferase; CDT). The genes cdtA andcdtB encode its two components, CDTa (enzymatic com-ponent) and CDTb (binding component), and are found onthe CDT locus (CdtLoc), which is carried independently of thePaLoc (Fig. 1).8 While the role of binary toxin in disease is notwell understood, binary toxin is produced by several strainssuch as NAP1/BI/027 which emerged as a major epidemicstrain in the early 2000s.9–11

C. DIFFICILE DETECTION METHODS

Until the early 1980s the most commonly used test for CDIlaboratory diagnosis was detection of toxin B in stool by cell-culture cytotoxicity,12 however, due to its complexity and longturnaround time, rapid enzyme immunoassay (EIA) kitsbecame favoured. Initially, these kits were designed to detecttoxin A, for two reasons. First, it was wrongly assumed thatC. difficile produced either both toxin A and toxin B, or neithertoxin and, second, toxin A was far more immunogenic and itwas easier to raise antibodies against it. The discovery of toxinA negative strains of C. difficile shifted attention to EIA kitsthat detected both toxins, although even these still have greateravidity for toxin A.13 During the process of developing one ofthe early toxin A kits, one manufacturer inadvertently produceda test that detected glutamate dehydrogenase (GDH), anenzyme produced by C. difficile, C. botulinum and C. spor-ogenes, as well as a few anaerobic cocci.14 These EIAs detectthe presence of toxins or GDH using antibodies linked to achromogenic marker. Detection of GDH remains a usefulscreening test because of its high sensitivity but detection ofC. difficile toxin is still considered as necessary for confir-mation of CDI.15

Culture of C. difficile takes several days and does notdifferentiate asymptomatic carriers from those with disease(i.e., it lacks specificity), or toxigenic from non-toxigenicstrains. However, isolation of the bacterium allows abroader array of tests to be performed, including determinationof the organism’s toxigenic status. So-called toxigenicculture has become the gold standard against which all newtests should be assessed. The low sensitivity of EIAs, averagingaround 60%,16 has driven the continuing search for highly

Pathology (April 2015) 47(3), pp. 211–218

M O L E C U L A R D I A G N O S T I C S I N M I C R O B I O L O G Y

Print ISSN 0031-3025/Online ISSN 1465-3931 # 2015 Royal College of Pathologists of Australasia

DOI: 10.1097/PAT.0000000000000238


sensitive, rapid techniques for CDI laboratory diagnosis.Currently, molecular methods have gained popularity fortheir rapid turnaround time and high sensitivity. However,they are expensive and may in fact be too sensitive forthe diagnosis of disease as opposed to detecting theorganism.17

Molecular assays for C. difficile detection

Nucleic acid amplification tests (NAATs) for C. difficile havebeen in use since the early 1990s. These early PCR teststargeted various genes including tcdA and tcdB.18–20 Theywere relatively labour intensive, requiring manual DNA extrac-tion and visualisation of results using gel electrophoresis orSouthern blot analysis. Some early primers also showed cross-reactivity with other clostridia.19 Throughout the 1990s, DNAextraction kits and NAAT techniques improved, and the devel-opment of real-time PCR (rtPCR) techniques followed. Anumber of assays gained US Food and Drug Administration(FDA) approval for use within diagnostic laboratories, usheringin a new age of rapid diagnostic techniques.

The currently approved molecular assays for C. difficiledetection are summarised in Table 1. The first NAAT to receiveFDA approval was the BD GeneOhm Cdiff assay in 2009. Thisand other early assays required a manual DNA extraction step,while most of the newer assays are fully automated. In theautomated assays a small amount of stool sample is placed intoa closed cartridge or tube where the entire process (DNAextraction, amplification reaction and product detection) takesplace. These automated assays are desirable due to their shorthands-on processing time and reduced risk of cross-contami-nation of samples.

The majority of the assays employ rtPCR-based reactions,some with detection of products using molecular beacons (BD

Positiveregulator

Positiveregulator

Enzymaticcomponent

Bindingcomponent

tcdR tcdB tcdE tcdA tcdC

cdtR cdtA cdtB

Toxin BHolin-like

protein Toxin ANegativeregulator

A

B

Fig. 1 The PaLoc (A) and the CdtLoc (B). The PaLoc encodes five genes. tcdAand tcdB encode toxins A and B respectively. tcdC is a polymorphic locus withmany known mutations, which encodes TcdC. The TcdC protein is believed tonegatively regulate production of toxins A and B. tcdE encodes a protein thoughtto play a role in toxin release from the cell. The CdtLoc is located downstream ofthe PaLoc. The cdtA and cdtB genes encode the proteins which comprise theenzymatic and binding subunits of binary toxin.

Table 1 Current approved molecular assays for C. difficile detection

Assay name (Manufacturer) Target Reaction type Extraction type System TAT (min)

ICEPlex C. difficile Kit (PrimeraDx) tcdB rtPCR Manual ICEPlex System <480IMDx C. difficile for Abbott m2000

(Intelligent Medical Devices)tcdA, tcdB,

tcdB-variantMultiplex rtPCR Automated Abbott m2000 System

BD Diagnostics BD Max Cdiff Assay(GeneOhm Sciences Canada)

tcdB rtPCR Automated BD Max System <180

Quidel Molecular Direct C. difficileAssay (Quidel Corporation)

tcdA, tcdB Multiplex rtPCR Automated QuantStudio Dx, AppliedBiosystems 7500 FastDx, Cepheid SmartCycler II

<70

Verigene C. difficile Nucleic AcidTest (Nanosphere)

tcdA, tcdB,tcdCD117, cdt

Multiplex rtPCR Automated Verigene System 240

Portrait Toxigenic C. difficile Assay(Great Basin Scientific)

tcdB Helicase-dependentmultiplex amplification

Automated Portrait Dx Analyzer <90

Simplexa C. difficile UniversalDirect Assay (Focus Diagnostics)

tcdB rtPCR Automated 3M Integrated Cycler <90

Xpert C. difficile/Epi (Cepheid) tcdB,tcdCD117, cdt

Multiplex rtPCR Automated GeneXpert Instrument <60

Illumigene C. difficile DNAAmplification Assay (MeridianBioscience)

tcdA LAMP Manual Illumipro-10 <60

Xpert C. difficile (Cepheid) tcdB, tcdCD117,cdt

Multiplex rtPCR Automated GeneXpert Instrument 45

ProGastro Cd Assay (Prodesse) tcdB rtPCR Manual Cepheid SmartCycler II <180BD GeneOhm C. diff Assay (BD

Diagnostics/GeneOhm Sciences)tcdB rtPCR Manual Cepheid SmartCycler II 120

EasyScreen C. difficile DetectionKit (Genetic Signatures)

tcdA, tcdB rtPCR Manual Roche LightCycler 480, BioradCFX96, Agilent (Stratagene)MX3000, Qiagen Rotor-GeneQ, Cepheid SmartCycler

<180

EasyScreen C. difficile Reflex Kit(Genetic Signatures)

tcdCD117, cdtA,gyrA mutation

rtPCR Manual Roche LightCycler 480, BioradCFX96, Agilent (Stratagene)MX3000, Qiagen Rotor-GeneQ, Cepheid SmartCycler

<180

Seeplex Diarrhea ACE (Seegene) tcdB Multiplex PCR Manual Calipher LabChip Dx, MCE-202MultiNA, Agilent 2200TapeStation

<60

Faecal Pathogens C (16Plex)(AusDiagnostics)

tcdB Multiplex PCR Manual High-Plex <180

rtPCR, real-time polymerase chain reaction; TAT, turnaround time.

212 COLLINS et al. Pathology (2015), 47(3), April


Max, BD GeneOhm), fluorescent probe-primers (Simplexa) orgold particle probe capture with silver signal amplification onan array (Verigene). The Illumigene assay differs from theothers in that it employs loop-mediated isothermal DNAamplification (LAMP). The Portrait Toxigenic C. difficileAssay utilises helicase-dependent amplification, with chip-based detection. Almost all of the assays require a specificsystem instrument for performance of the assay, e.g., the BDMax assay is performed on the BD Max system. The exceptionsare the Quidel Molecular Direct C. difficile assay, and theEasyScreen C. difficile assays, which can both be performed onseveral different platforms. Two multiplex PCRs identify tcdBof C. difficile in addition to marker genes of several otherdiarrhoeal pathogens. The Seeplex Diarrhea ACE Detectionsystem identifies four viruses and/or 10 bacteria in one reaction,while the Faecal Pathogens C kit detects five viruses and 10bacteria in a multiplex tandem PCR (Table 1).

Due to its ubiquity in toxigenic strains, tcdB is favoured as atarget gene. Exceptions are the Illumigene assay, which targetsa highly conserved portion of the 5’ end of tcdA, and theEasyScreen C. difficile Reflex Kit, which targets cdtA and a117 bp deletion within the tcdC gene (tcdCD117) and identifiesputative ribotype (RT) 027 strains (Table 1).21 The EasyScreenC. difficile Reflex Kit also detects a gyrA gene mutation whichcodes for fluoroquinolone resistance.22 The Verigene and XpertC. difficile/Epi assays additionally detect the CdtLoc for identi-fication of binary toxin-positive strains and tcdCD117. TheIMDx C. difficile assay also targets a tcdB-variant gene found inRT 017, a TcdA–TcdBþ strain that has caused epidemics ofCDI in Europe and Asia.8,23,24

Reported accuracy measurements of NAATs are generallyhigh but can vary widely for some assays (Table 2). Thepositive predictive value (PPV) of the Verigene assay was42% compared to direct toxigenic culture, while PPVs for otherassays including BD Max, Illumigene and Xpert exceeded80%. The Illumigene assay had the lowest sensitivity of73–95%. Reported specificities and negative predictive values(NPVs) are high, exceeding 90% for most NAATs.

Issues with NAATs

While existing data show that NAATs for C. difficile areanalytically sensitive and specific (Table 2), questions remainover their clinical specificity. Individuals can be colonised withC. difficile which would be detected by a NAAT. Detection oftoxin genes does not necessarily correlate with expression oftoxin, nor disease.

Given the heterogeneity of C. difficile strains, some may not bedetected by a particular assay, or may be incorrectly identified.

For example, some strains such as RT 244 also carry thetcdCD117 deletion and cdt genes,25 leading them to be possiblymis-identified as RT 027 by the Xpert and Verigene assays.26 TheIllumigene assays fail to detect several tcdA-negative ribotypes,particularly RT 033.27 Currently, almost all assays focus on tcdBdetection, but if a tcdB-variant strain was to proliferate, theseassays could fail to identify them. With shifts in circulatingstrains, validation of assays is required in diagnostic laboratories.Surveillance must be in place to alert clinical laboratories when anew predominant strain is identified. This information can beused to assess assay performance in individual laboratories.

Several studies have concluded that inappropriate test order-ing can also impact the clinical specificity of tests. One studyfound 36% of patients for whom a C. difficile test was ordereddid not have clinically significant diarrhoea, and 20% hadrecently taken laxatives.28 The Infectious Diseases Societyof America/Society for Healthcare Epidemiology of Americaand the Australasian Society for Infectious Diseases recom-mend that only unformed/diarrhoeal stool samples be tested forC. difficile, except in the case of suspected ileus.29,30 While CDImay be suspected in patients who have been hospitalised for>48 h,31 with the increase of community-acquired disease it isimportant to collect history of recent antibiotic and laxative use.This will help to determine whether a test for C. difficile isappropriate in any diarrhoeal patient.

Several American studies reported that CDI rates increasedwhen laboratories adopted molecular testing. Surveillance datashowed increases ranging from 43% to 67% in CDI incidenceattributable to changing from toxin EIA to NAAT.32 After aperiod of increase these rates declined, suggesting introductionof more specific case definitions, or perhaps decreased trans-mission.33

Some studies have measured variability in test performancedepending upon the strain type present. One study foundsignificant differences in EIA performance for toxin A/Band GDH compared to the Xpert C. difficile assay for non-RT 027 strains, whereas they had equivalent performance forRT 027 strain detection. The sensitivity of the EIA wassignificantly lower than that of the Xpert assay for detectingRTs 002, 027 and 106.34 In contrast, other studies have notfound a discrepancy in detection based on strain type.35

Another significant issue with NAATs is that they cost up totwo to three times more than EIAs.17 The issue of cost-effec-tiveness has been addressed in several studies. It has beenpostulated that the enhanced accuracy provided by NAATs maylead to fewer tests for result confirmation ordered overall.36 Amodel based on testing 1000 patients, assuming 10% preva-lence of CDI and using published performance characteristicsof various test methods, was tested. The model showed that

Table 2 Reported accuracy measurements of nucleic acid amplification test assays, where available, compared to direct toxigenic culture

Assay Sensitivity (%) Specificity (%) PPV (%) NPV (%) Reference(s)

BD Max Cdiff Assay 94–98 98–99 89–98 99 82–85

Quidel Molecular Direct C. difficile Assay 82–96 97–100 86,87

Verigene C. difficile Nucleic Acid Test 95–99 88–99 42 99 88–90

Portrait Toxigenic C. difficile Assay 98 93 91

Simplexa C. difficile Universal Direct Assay 87–98 100 87,88

Illumigene C. difficile Assay 73–95 95–100 83–100 84–100 92–105

Xpert C. difficile 96–100 92–99 83–92 98–100 82,88,92,102,103,106–108

ProGastro Cd Assay 89–100 93–100 83–100 95–100 107,109

BD GeneOhm C. diff Assay 82–97 96–100 87–100 97–100 83,92,94,102,104,107–112

NPV, negative predictive value; PPV, positive predictive value.

DETECTION AND TYPING OF C. DIFFICILE 213


NAAT testing alone, compared to toxin EIA alone and varioustwo-step algorithms using GDH detection, resulted in thegreatest number of CDI patients being isolated within 24 h.The enhanced accuracy of NAAT would lead to earlier detec-tion of true-positive patients, reducing the amount of time thatfalse-negative patients (by EIA) could continue to spread C.difficile within the hospital.37

Another study estimated that in diagnosing CDI one dayearlier in cases otherwise missed by EIA, NAAT alone couldsave US$200,000 annually by avoiding the costs of repeattesting.38 They found that the increased cost of a NAAT wasjustified by the earlier detection of CDI cases, again theoreticallypreventing additional cases of nosocomial CDI and reducinglength of stay for CDI patients. Babady et al. also found that usingNAATs alone was cost-effective considering the consequentreductions in turnaround time and labour costs.39

Some laboratories use a multi-step algorithmic approach fordiagnosis, given the high cost of NAATs. Usually a GDH EIAscreen is used initially, and only GDH positive specimens arefurther investigated by a NAAT in a two-step algorithm, or in athree-step approach a toxin EIA is also performed prior to aNAAT. While an algorithmic approach saves on costs, theprevalence of GDH positive, non-toxigenic C. difficile in thepopulation may be high, leading to unnecessary secondaryNAATs. In addition, as previously mentioned, EIAs are gener-ally less sensitive than NAAT for C. difficile detection. Thuslaboratories must monitor strains currently in circulation andadjust their testing approach accordingly. There is no clearoptimal approach and testing guidelines vary from country tocountry.29,30,40

Impact of diagnostic method on patient outcome

While there are many supporters of NAATs alone for C.difficile detection, there are still arguments for the applicationof toxin detection. Some believe that at least two detectionmethods are required either in parallel or series for C. difficiledetection. While cell culture cytotoxicity assay and toxigenicculture are still considered reference methods, they are notoptimal for rapid detection.41 One large English study foundthat mortality was significantly higher in patients foundpositive by cytotoxin assay compared to a cytotoxin/toxigenicculture negative group (odds ratio 1.61, 95% confidence inter-val 1.12–2.31).15 No significant increase in mortality wasapparent when toxigenic C. difficile alone was detected. Theseresults have been replicated, showing that patients with apositive NAAT but no detectable toxin have less severe andshorter-lived symptoms, lower bacterial load, are less likely todevelop complications and have lower all-cause mortalitycompared with those who had toxin detected directly.42–44

Future diagnostic methods

In the future, C. difficile diagnostics could possibly includedetection of not only C. difficile but also a biomarker correlat-ing with current disease state. Faecal lactoferrin,45–47 faecalcalprotectin48,49 and various cytokines17 have been identifiedas possible targets. An optimal adjunct assay has not yet beenestablished but these assays may be recommended in tandemwith NAATs in the future.

C. DIFFICILE TYPING METHODS

Since its emergence as a pathogen in the late 1970s, a widearray of methods has been used to type C. difficile. Initially,

typing was mainly restricted to phenotypic methods due to thetechnological limitations of the time. These phenotypicmethods included immunoblotting,50 bacteriocin/bacterio-phage susceptibility schemes,51,52 antibiograms, pyrolysis,53

SDS-PAGE of whole cell proteins,54 and serogrouping.55 Plas-mid profiling, one of the few molecular methods from thatperiod, was largely unsuccessful due to the apparent scarcity ofthese extrachromosomal elements.56

Current typing methods in common usage

To our knowledge, routine typing of C. difficile is not per-formed anywhere in the world currently on a national or evenregional scale. The UK has the most active program, with PCRribotyping available on request via the Clostridium difficileRibotyping Network (CDRN) based in Leeds. Otherwise typingremains the purview of research laboratories.

There are several molecular typing methods currently usedfor typing C. difficile: pulsed-field gel electrophoresis (PFGE),multilocus VNTR analysis (MLVA), PCR ribotyping, restric-tion endonuclease analysis (REA) and multilocus sequencingtyping (MLST). In addition to these, a plethora of othermolecular methods57,58 are used infrequently or have beenabandoned. These include: plasmid profiling, arbitrarily-primed PCR,59 repetitive element palindromic PCR,60 ampli-fied fragment length polymorphism (AFLP),61 toxinotyping,62

S-layer protein A sequence typing (slpAST)63 and restrictionfragment length polymorphism (RFLP) analysis of the fliCgene.64

PFGEPFGE involves the digestion of chromosomal DNA using a rarecutting endonuclease, and the separation of those fragments byelectrophoresis to produce a fingerprint. Third generationPFGE utilises a contour-clamped homogenous electric field(CHEF) to separate large fragments efficiently. Despite expens-ive setup costs, and complicated and slow protocols, PFGE byCHEF became the gold standard for typing most pathogens inthe 1990s, largely due to the creation of PulseNet(www.cdc.gov/pulsenet/) in the USA. Attempts to apply PFGEto typing C. difficile encountered serious difficulties though,with many isolates non-typeable due to the sensitivity of theirDNA to degradation. Despite these issues, PFGE was adoptedas the primary typing method for C. difficile by both the CDCsin Canada and the USA, although neither typed many isolates.A modified PFGE protocol that largely resolved the problem ofnon-typeable isolates was eventually designed,65 but by thenother methods had largely supplanted PFGE in many regions ofthe world. Typically SmaI is the restriction endonuclease usedfor C. difficile PFGE, although others are used occasionally.Clostridium difficile pulsetypes use the prefix NAP (NorthAmerican Pulsetype) followed by a number indicating thepulsetype group, and letter for the subgroup (i.e., NAP1b).

Restriction endonuclease analysisRestriction endonuclease analysis (REA) also involves the diges-tion of the chromosomal DNA using a rare-cutting enzyme, but incontrast to PFGE, the digestion fragments are separated bystandard electrophoresis, either on agarose or polyacrylamidegels. REA using HindIII is the most common typing method inthe USA, being the method of choice for some researchgroups.66,67 The fingerprints are time consuming and difficultto interpret though, due to the sheer number of bands. Due to its


http://www.cdc.gov/pulsenet/


continual use by Dale Gerding, a pioneer in the field of C. difficileresearch, who has an extensive library of REA typed strains,much of the typing data from the USA is in this format.

PCR ribotypingPCR ribotyping is the amplification of the 16S–23S rRNAintergenic spacer region (ISR) using primers that target the 3’end of the 16S rRNA gene, and the 5’ end of the 23S rRNAgene. Currently, it is the most commonly used molecular typingtechnique for C. difficile. The C. difficile genome containsseveral 16S–23S rRNA operons which vary in both theirnumber and in the size of their intergenic regions. Thusamplification of these regions followed by electrophoreticseparation generates a fingerprint. PCR ribotyping is not tobe confused with standard ribotyping, a technique that involvesthe amplification of the 16S rRNA gene and subsequentrestriction-fragment length polymorphisms (RFLP) analysis.68

Gurtler first identified the potential of the C. difficile rRNAISR for typing,69 but O’Neill et al.70 designed the simplifiedprotocol that is currently in use. Using the O’Neill primers,PCR products, usually 250–600 bp in size, are separated byslab agarose gel electrophoresis, or more recentlycapillary electrophoresis.

The original library of C. difficile PCR ribotypes was estab-lished at the then English and Welsh Public Health LaboratoryService Anaerobe Reference Unit (ARU) in Cardiff, Wales.71

The current Public Health England C. difficile RibotypingNetwork (CDRN), comprises several referral laboratories inEngland and Northern Ireland, and the central library, and hasbeen operational since 1996.57 The central library recentlymoved to Leeds General Infirmary (UK), and all >800 ribo-types were reprocessed using sequencer-based capillary elec-trophoresis. The nomenclature for PCR ribotypes is a tripledigit number sometimes with the prefix UK (i.e., UK 001) todifferentiate this system from other similar systems. As capil-lary electrophoresis allows better resolution of patterns, it hasresulted in the separation of some existing types, and PCRribotype profiles assigned from comparison against the newcapillary electrophoresis are often given the prefix CE (i.e., CE001). PCR ribotyping by capillary electrophoresis also allowsfaster visualisation of profiles, as well as exchange of electronicfiles between laboratories that use compatible platforms, mak-ing it more portable than gel electrophoresis resolution, andPFGE and REA.

PCR ribotyping of C. difficile has been performed in WesternAustralia since 2006, initially using slab agarose gel electro-phoresis, but more recently by capillary electrophoresis. Thereare currently more than 400 PCR ribotypes in the WA library.The nomenclature used for these is the prefix QX followed bythree digits (i.e., QX 001) when the corresponding UK type isnot known.

An additional set of PCR ribotyping primers with differentbinding positions is in limited use by some groups.72 Theseprimers typically produce the same number of bands as finger-prints generated using the O’Neill primers, but are slightlysmaller in size. A web service, WEBRIBO (https://webribo.ages.at/), allows users to upload sequencer-generated finger-prints and be assigned, but is unaffiliated with the UK library.73

MLVAMLVA targets variable number tandem repeats (VNTR) thatare distributed throughout the genome as a typing method.

Multiple tandem repeat (TR) loci are targeted by PCR, and thenumber of copies per loci established by sequencing or frag-ment analysis. Several methods have been published with thenumber of TR loci used ranging from 2 to 12.74–77 The greaterthe number of TR loci examined, the more discriminatory themethod. The UK CDRN currently offers MLVA analysis of C.difficile using seven TR loci, or an expanded panel of 12, as anenhanced service for investigating outbreaks.

MLSTMLST involves the examining of sequence variations in a set ofgenes, and the assignment of an allelic type to each gene to givean allelic profile. Usually housekeeping genes are used,although virulence genes are sometimes included. The firstMLST scheme for C. difficile used seven housekeeping genes(aroE, ddl, dutA, tpi, recA, gmk, and sodA) for the purpose ofanalysing evolutionary relationships.78 A later scheme devel-oped by Griffiths and colleagues specifically for epidemiolo-gical analysis, used a different set of seven housekeeping genes(adk, atpA, dxr, glyA, recA, sodA, and tpi).79 The Griffithsscheme has now been applied to a large collection of isolates inthe UK as a part of several epidemiological investigationprojects, and is part of the PubMLST site (http://pubmlst.org/). MLST has the benefit of being able to be performedin silico using next generation sequencing (NGS) data.

Typing in the future

As the costs associated with new sequencing technologiessteadily decline, it is inevitable that whole genome sequencing(WGS) will become the gold standard typing method of thefuture. Single nucleotide polymorphism (SNP) analysis ofWGS data involves examining the genes common betweenthe strains under investigation, essentially performing MLSTon a massive scale. This method has already been used foroutbreak investigations,25 and large epidemiological stu-dies.80,81 A recent study by Eyre and colleagues revealed thegreat diversity in strains which appeared to be epidemiologi-cally linked because of close hospital association, indicatingthat a broad reservoir of C. difficile exists.80 Some moleculartyping methods can be performed in silico using WGS data,such as MLVA and MLST. At the moment PCR ribotypingcannot be performed in silico because the 16S–23S rRNAoperon is a large repetitive element, but this may change as readlengths increase.

CONCLUSIONS

Detection and typing of C. difficile continue be challenging.The constant changes in circulating strains mean laboratoriescannot settle on a standard detection method without continuedmonitoring of C. difficile molecular epidemiology. WhileNAATs are highly sensitive, commercially-available testsare currently costly and a positive test could be due to toxigenicstrain colonisation. However, they do appear to have the benefitof more rapid detection of CDI patients, enabling quickerisolation and prevention of further transmission within hospi-tals. If indeed it is shown that the presence of toxin in stool is apre-requisite for a clinically significant positive test then theonus on diagnostic companies will be to develop more sensitivetoxin EIAs. As yet, an optimal test or testing algorithm remainselusive. Molecular typing methods vary from region to region.The advent of more highly resolved capillary electrophoresispatterns for PCR ribotyping and sequence-based analysis will


https://webribo.ages.at/

https://webribo.ages.at/

http://pubmlst.org/

http://pubmlst.org/


facilitate sharing of more data between surveillance labora-tories and allow more rapid identification of internationalmovements of strains and emergence of strains with increasedvirulence. Improved diagnostic and typing methods arerequired to reduce the incidence of CDI and outbreaks ofthe scale seen in recent decades.

ADDENDUM

Since this review was submitted several studies have beenpublished examining the BioFire FilmArray gastrointestinalpanel which targets multiple bacterial, viral and parasitic causesof gastroenteritis, including C. difficile. For example, Busset al. showed that the sensitivity and specificity for detectingC. difficile versus other PCR-based methods were 98.8% (95%CI 95.7–99.9%) and 97.1% (95% CI 96.0–97.9%), respec-tively.113 While the use of multiplexed assays has manyattractions, there are likely to be impediments to their imple-mentation such as cost and the issue of how to deal with theincreased numbers of positive tests that will arise when diar-rhoeal stools are ‘screened’ for multiple pathogens simul-taneously with a highly sensitive test.114

Acknowledgements: Deirdre Collins is a recipient of anInternational Postgraduate Scholarship and an AustralianPostgraduate Award from The University of Western Australia.

Conflicts of interest and sources of funding: The authors statethat there are no conflicts of interest to disclose.

Address for correspondence: Professor Thomas V. Riley, Microbiology and

Immunology, School of Pathology and Laboratory Medicine, Queen Elizabeth II

Medical Centre, Nedlands, WA 6009, Australia. E-mail: thomas.riley@uwa.

edu.au

References1. Freeman J, Bauer MP, Baines SD, et al. The changing epidemiology of

Clostridium difficile infections. Clin Microbiol Rev 2010; 23: 529–49.2. Bartlett JG. Clinical practice. Antibiotic-associated diarrhea. N Engl J

Med 2002; 346: 334–9.3. Ackermann G, Thomalla S, Ackermann F, et al. Prevalence and char-

acteristics of bacteria and host factors in an outbreak situation of anti-biotic-associated diarrhoea. J Med Microbiol 2005; 54: 149–53.

4. Khanna S, Pardi DS, Aronson SL, et al. The epidemiology of community-acquired Clostridium difficile infection: a population-based study. Am JGastroenterol 2012; 107: 89–95.

5. Slimings C, Armstrong P, Beckingham WD, et al. Increasing incidence ofClostridium difficile infection, Australia, 2011-2012. Med J Aust 2014;200: 272–6.

6. Heinlen L, Ballard JD. Clostridium difficile infection. Am J Med Sci 2010;340: 247–52.

7. Rupnik M. Clostridium difficile toxinotyping. Methods Mol Biol 2010;646: 67–76.

8. Rupnik M, Kato N, Grabnar M, et al. New types of toxin A-negative, toxinB-positive strains among Clostridium difficile isolates from Asia. J ClinMicrobiol 2003; 41: 1118–25.

9. Goorhuis A, Bakker D, Corver J, et al. Emergence of Clostridium difficileinfection due to a new hypervirulent strain, polymerase chain reactionribotype 078. Clin Infect Dis 2008; 47: 1162–70.

10. Goorhuis A, Debast SB, van Leengoed LA, et al. Clostridium difficile PCRribotype 078: an emerging strain in humans and in pigs? J Clin Microbiol2008; 46: 1157; (author reply 58).

11. McDonald LC, Killgore GE, Thompson A, et al. An epidemic, toxin gene-variant strain of Clostridium difficile. N Engl J Med 2005; 353: 2433–41.

12. Bowman RA, Riley TV. Laboratory diagnosis of Clostridium difficile-associated diarrhoea. Eur J Clin Microbiol Infect Dis 1988; 7: 476–84.

13. Wilkins TD, Lyerly DM. Clostridium difficile testing: after 20 years, stillchallenging. J Clin Microbiol 2003; 41: 531–4.

14. Lyerly DM, Barroso LA, Wilkins TD. Identification of the latex test-reactive protein of Clostridium difficile as glutamate dehydrogenase.J Clin Microbiol 1991; 29: 2639–42.

15. Planche TD, Davies KA, Coen PG, et al. Differences in outcomeaccording to Clostridium difficile testing method: a prospective multi-centre diagnostic validation study of C difficile infection. Lancet Infect Dis2013; 13: 936–45.

16. Brecher SM, Novak-Weekley SM, Nagy E. Laboratory diagnosis ofClostridium difficile infections: there is light at the end of the colon. ClinInfect Dis 2013; 57: 1175–81.

17. Burnham CA, Carroll KC. Diagnosis of Clostridium difficile infection: anongoing conundrum for clinicians and for clinical laboratories. ClinMicrobiol Rev 2013; 26: 604–30.

18. Kato N, Ou CY, Kato H, et al. Identification of toxigenic Clostridiumdifficile by the polymerase chain reaction. J Clin Microbiol 1991; 29: 33–7.

19. Wren B, Clayton C, Tabaqchali S. Rapid identification of toxigenicClostridium difficile by polymerase chain reaction. Lancet 1990; 335: 423.

20. Gumerlock PH, Tang YJ, Weiss JB, et al. Specific detection of toxigenicstrains of Clostridium difficile in stool specimens. J Clin Microbiol 1993;31: 507–11.

21. MacCannell DR, Louie TJ, Gregson DB, et al. Molecular analysis ofClostridium difficile PCR ribotype 027 isolates from Eastern and WesternCanada. J Clin Microbiol 2006; 44: 2147–52.

22. Drudy D, Kyne L, O’Mahony R, et al. GyrA mutations in fluoroquinolone-resistant Clostridium difficile PCR-027. Emerg Infect Dis 2007; 13: 504–5.

23. Drudy D, Harnedy N, Fanning S, et al. Emergence and control offluoroquinolone-resistant, toxin A-negative, toxin B-positive Clostridiumdifficile. Infect Control Hosp Epidemiol 2007; 28: 932–40.

24. Collins DA, Hawkey PM, Riley TV. Epidemiology of Clostridium difficileinfection in Asia. Antimicrob Resist Infect Control 2013; 2: 21.

25. Lim SK, Stuart RL, Mackin KE, et al. Emergence of a ribotype 244strain of Clostridium difficile associated with severe disease andrelated to the epidemic ribotype 027 strain. Clin Infect Dis 2014; 58:1723–30.

26. Kok J, Wang Q, Thomas LC, et al. Presumptive identification ofClostridium difficile strain 027/NAP1/BI on Cepheid Xpert: interpret withcaution. J Clin Microbiol 2011; 49: 3719–21.

27. Androga G, McGovern A, Elliott B, et al. Evaluation of the CepheidXpert(R) C. difficile/Epi and Meridian Bioscience illumigene(R) C.difficile for detecting Clostridium difficile ribotype 033 strains. J ClinMicrobiol 2014. DOI 10.1128/JCM.03297-14.

28. Dubberke ER, Han Z, Bobo L, et al. Impact of clinical symptoms oninterpretation of diagnostic assays for Clostridium difficile infections.J Clin Microbiol 2011; 49: 2887–93.

29. Cohen Stuart H, Gerding Dale N, Johnson S, et al. Clinical practiceguidelines for Clostridium difficile infection in adults: 2010 Update bythe Society for Healthcare Epidemiology of America (SHEA) and theInfectious Diseases Society of America (IDSA). Infect Control HospEpidemiol 2010; 31: 431–55.

30. Cheng AC, Ferguson JK, Richards MJ, et al. Australasian Society forInfectious Diseases guidelines for the diagnosis and treatment ofClostridium difficile infection. Med J Aust 2011; 194: 353–8.

31. McDonald LC, Coignard B, Dubberke E, et al. Recommendations forsurveillance of Clostridium difficile-associated disease. Infect ControlHosp Epidemiol 2007; 28: 140–5.

32. Gould CV, Edwards JR, Cohen J, et al. Effect of nucleic acid amplificationtesting on population-based incidence rates of Clostridium difficile infec-tion. Clin Infect Dis 2013; 57: 1304–7.

33. Goldenberg SD. Public reporting of Clostridium difficile and improve-ments in diagnostic tests. Infect Control Hosp Epidemiol 2011; 32: 1231–2; (author reply 33).

34. Tenover FC, Novak-Weekley S, Woods CW, et al. Impact of strain type ondetection of toxigenic Clostridium difficile: comparison of moleculardiagnostic and enzyme immunoassay approaches. J Clin Microbiol2010; 48: 3719–24.

35. Goldenberg SD, Gumban M, Hall A, et al. Lack of effect of strain type ondetection of toxigenic Clostridium difficile by glutamate dehydrogenaseand polymerase chain reaction. Diagn Microbiol Infect Dis 2011; 70:417–9.

36. Peterson LR, Robicsek A. Does my patient have Clostridium difficileinfection? Ann Intern Med 2009; 151: 176–9.

37. Tenover FC, Baron EJ, Peterson LR, et al. Laboratory diagnosis ofClostridium difficile infection: can molecular amplification methods moveus out of uncertainty? J Mol Diagn 2011; 13: 573–82.

38. Larson AM, Fung AM, Fang FC. Evaluation of tcdB real-time PCR in athree-step diagnostic algorithm for detection of toxigenic Clostridiumdifficile. J Clin Microbiol 2010; 48: 124–30.

39. Babady NE, Stiles J, Ruggiero P, et al. Evaluation of the Cepheid XpertClostridium difficile Epi assay for diagnosis of Clostridium difficile



infection and typing of the NAP1 strain at a cancer hospital. J ClinMicrobiol 2010; 48: 4519–24.

40. Crobach MJ, Dekkers OM, Wilcox MH, et al. European Society ofClinical Microbiology and Infectious Diseases (ESCMID): data reviewand recommendations for diagnosing Clostridium difficile-infection(CDI). Clin Microbiol Infect 2009; 15: 1053–66.

41. Wilcox MH, Planche T, Fang FC, et al. What is the current role ofalgorithmic approaches for diagnosis of Clostridium difficile infection?J Clin Microbiol 2010; 48: 4347–53.

42. Baker I, Leeming JP, Reynolds R, et al. Clinical relevance of a positivemolecular test in the diagnosis of Clostridium difficile infection. J HospInfect 2013; 84: 311–5.

43. Polage CR, Chin DL, Leslie JL, et al. Outcomes in patients tested forClostridium difficile toxins. Diagn Microbiol Infect Dis 2012; 74: 369–73.

44. Longtin Y, Trottier S, Brochu G, et al. Impact of the type of diagnosticassay on Clostridium difficile infection and complication rates in amandatory reporting program. Clin Infect Dis 2013; 56: 67–73.

45. Miller JR, Barrett LJ, Kotloff K, et al. A rapid test for infectious andinflammatory enteritis. Arch Intern Med 1994; 154: 2660–4.

46. Steiner TS, Flores CA, Pizarro TT, et al. Fecal lactoferrin, interleukin-1beta, and interleukin-8 are elevated in patients with severe Clostridiumdifficile colitis. Clin Diagn Lab Immunol 1997; 4: 719–22.

47. Pawlowski SW, Archbald-Pannone L, Carman RJ, et al. Elevatedlevels of intestinal inflammation in Clostridium difficile infection asso-ciated with fluoroquinolone-resistant C. difficile. J Hosp Infect 2009; 73:185–7.

48. Shastri YM, Bergis D, Povse N, et al. Prospective multicenter studyevaluating fecal calprotectin in adult acute bacterial diarrhea. Am J Med2008; 121: 1099–106.

49. LaSala PR, Ekhmimi T, Hill AK, et al. Quantitative fecal lactoferrin intoxin-positive and toxin-negative Clostridium difficile specimens. J ClinMicrobiol 2013; 51: 311–3.

50. Poxton IR, Byrne MD. Immunological analysis of the EDTA-solubleantigens of Clostridium difficile and related species. J Gen Microbiol1981; 122: 41–6.

51. Mahony DE, Clow J, Atkinson L, et al. Development and application of amultiple typing system for Clostridium difficile. Appl Environ Microbiol1991; 57: 1873–9.

52. Sell TL, Schaberg DR, Fekety FR. Bacteriophage and bacteriocin typingscheme for Clostridium difficile. J Clin Microbiol 1983; 17: 1148–52.

53. Cartmill TD, Orr K, Freeman R, et al. Nosocomial infection withClostridium difficile investigated by pyrolysis mass spectrometry.J Med Microbiol 1992; 37: 352–6.

54. Ogunsola FT, Ryley HC, Duerden BI. Sodium dodecyl sulfate-polyacry-lamide gel electrophoresis analysis of EDTA-extracted cell-surface pro-tein antigens is a simple and reproducible method for typing Clostridiumdifficile. Clin Infect Dis 1995; 20 (Suppl 2): S327–30.

55. Delmee M, Homel M, Wauters G. Serogrouping of Clostridium difficilestrains by slide agglutination. J Clin Microbiol 1985; 21: 323–7.

56. Mulligan ME, Peterson LR, Kwok RY, et al. Immunoblots and plasmidfingerprints compared with serotyping and polyacrylamide gel electro-phoresis for typing Clostridium difficile. J Clin Microbiol 1988; 26: 41–6.

57. Brazier JS. Typing of Clostridium difficile. Clin Microbiol Infect 2001; 7:428–31.

58. Kuijper EJ, van den Berg RJ, Brazier JS. Comparison of molecular typingmethods applied to Clostridium difficile. Methods Mol Biol 2009; 551:159–71.

59. Barbut F, Mario N, Frottier J, et al. Use of the arbitrary primer polymerasechain reaction for investigating an outbreak of Clostridium difficile-associated diarrhea in AIDS patients. Eur J Clin Microbiol Infect Dis1993; 12: 794–5.

60. Church DL, Chow BL, Lloyd T, et al. Evaluation of automated repetitive-sequence-based PCR (DiversiLab) compared to PCR ribotyping for rapidmolecular typing of community- and nosocomial-acquired Clostridiumdifficile. Diagn Microbiol Infect Dis 2011; 70: 183–90.

61. Klaassen CH, van Haren HA, Horrevorts AM. Molecular fingerprinting ofClostridium difficile isolates: pulsed-field gel electrophoresis versusamplified fragment length polymorphism. J Clin Microbiol 2002; 40:101–4.

62. Rupnik M, Avesani V, Janc M, et al. A novel toxinotyping scheme andcorrelation of toxinotypes with serogroups of Clostridium difficile isolates.J Clin Microbiol 1998; 36: 2240–7.

63. Eidhin DN, Ryan AW, Doyle RM, et al. Sequence and phylogeneticanalysis of the gene for surface layer protein, slpA, from 14 PCR ribotypesof Clostridium difficile. J Med Microbiol 2006; 55: 69–83.

64. Tasteyre A, Karjalainen T, Avesani V, et al. Phenotypic and genotypicdiversity of the flagellin gene (fliC) among Clostridium difficile isolatesfrom different serogroups. J Clin Microbiol 2000; 38: 3179–86.

65. Alonso R, Martin A, Pelaez T, et al. An improved protocol for pulsed-fieldgel electrophoresis typing of Clostridium difficile. J Med Microbiol 2005;54: 155–7.

66. Belmares J, Johnson S, Parada JP, et al. Molecular epidemiology ofClostridium difficile over the course of 10 years in a tertiary care hospital.Clin Infect Dis 2009; 49: 1141–7.

67. Tenover FC, Akerlund T, Gerding DN, et al. Comparison of strain typingresults for Clostridium difficile isolates from North America. J ClinMicrobiol 2011; 49: 1831–7.

68. Bowman RA, O’Neill GL, Riley TV. Non-radioactive restriction fragmentlength polymorphism (RFLP) typing of Clostridium difficile. FEMSMicrobiol Lett 1991; 63: 269–72.

69. Gurtler V. Typing of Clostridium difficile strains by PCR-amplification ofvariable length 16S-23S rDNA spacer regions. J Gen Microbiol 1993; 139:3089–97.

70. O’Neill GL, Ogunsola FT, Brazier JS, et al. Modification of a PCRribotyping method for application as a routine typing scheme for Clos-tridium difficile. Anaerobe 1996; 2: 205–9.

71. Stubbs SLJ, Brazier JS, O’Neill GL, et al. PCR targeted to the 16S-23SrRNA gene intergenic spacer region of Clostridium difficile and construc-tion of a library consisting of 116 different PCR ribotypes. J ClinMicrobiol 1999; 37: 461–3.

72. Bidet P, Barbut F, Lalande V, et al. Development of a new PCR-ribotypingmethod for Clostridium difficile based on ribosomal RNA gene sequen-cing. FEMS Microbiol Lett 1999; 175: 261–6.

73. Indra A, Huhulescu S, Schneeweis M, et al. Characterization of Clos-tridium difficile isolates using capillary gel electrophoresis-based PCRribotyping. J Med Microbiol 2008; 57: 1377–82.

74. Marsh JW, O’Leary MM, Shutt KA, et al. Multilocus variable-numbertandem-repeat analysis for investigation of Clostridium difficile transmis-sion in hospitals. J Clin Microbiol 2006; 44: 2558–66.

75. Fawley WN, Wilcox MH. An enhanced DNA fingerprinting service toinvestigate potential Clostridium difficile infection case clusters sharingthe same PCR ribotype. J Clin Microbiol 2011; 49: 4333–7.

76. Manzoor SE, Tanner HE, Marriott CL, et al. Extended multilocus vari-able-number tandem-repeat analysis of Clostridium difficile correlatesexactly with ribotyping and enables identification of hospital transmission.J Clin Microbiol 2011; 49: 3523–30.

77. van den Berg RJ, Schaap I, Templeton KE, et al. Typing and subtyping ofClostridium difficile isolates by using multiple-locus variable-numbertandem-repeat analysis. J Clin Microbiol 2007; 45: 1024–8.

78. Lemee L, Dhalluin A, Pestel-Caron M, et al. Multilocus sequence typinganalysis of human and animal Clostridium difficile isolates of varioustoxigenic types. J Clin Microbiol 2004; 42: 2609–17.

79. Griffiths D, Fawley W, Kachrimanidou M, et al. Multilocus sequencetyping of Clostridium difficile. J Clin Microbiol 2010; 48: 770–8.

80. Eyre DW, Cule ML, Wilson DJ, et al. Diverse sources of C. difficileinfection identified on whole-genome sequencing. N Engl J Med 2013;369: 1195–205.

81. Eyre DW, Fawley WN, Best EL, et al. Comparison of multilocus variable-number tandem-repeat analysis and whole-genome sequencing for inves-tigation of Clostridium difficile transmission. J Clin Microbiol 2013; 51:4141–9.

82. Dalpke AH, Hofko M, Zorn M, et al. Evaluation of the fully automated BDMAX Cdiff and Xpert C. difficile assays for direct detection of Clos-tridium difficile in stool specimens. J Clin Microbiol 2013; 51: 1906–8.

83. Le Guern R, Herwegh S, Grandbastien B, et al. Evaluation of a newmolecular test, the BD Max Cdiff, for detection of toxigenic Clostridiumdifficile in fecal samples. J Clin Microbiol 2012; 50: 3089–90.

84. Leitner E, Einetter M, Grisold AJ, et al. Evaluation of the BD MAX Cdiffassay for the detection of the toxin B gene of Clostridium difficile out offaecal specimens. Diagn Microbiol Infect Dis 2013; 76: 390–1.

85. Putsathit P, Morgan J, Bradford D, et al. Evaluation of the BD Max Cdiffassay for the detection of toxigenic Clostridium difficile in human stoolspecimens. Pathology 2015; 47: 165–8.

86. Beck ET, Buchan BW, Riebe KM, et al. Multicenter evaluation of theQuidel Lyra Direct C. difficile nucleic acid amplification assay. J ClinMicrobiol 2014; 52: 1998–2002.

87. Deak E, Miller SA, Humphries RM. Comparison of illumigene, Simplexa,and AmpliVue Clostridium difficile molecular assays for diagnosis ofC. difficile infection. J Clin Microbiol 2014; 52: 960–3.

88. Gilbreath JJ, Verma P, Abbott AN, et al. Comparison of the VerigeneClostridium difficile, Simplexa C. difficile Universal Direct, BD MAXCdiff, and Xpert C. difficile assays for the detection of toxigenicC. difficile. Diagn Microbiol Infect Dis 2014; 80: 13–8.

89. Tojo M, Nagamatsu M, Hayakawa K, et al. Evaluation of an automatedrapid diagnostic test for detection of Clostridium difficile. PLoS One 2014;9: e106102.

90. Carroll KC, Buchan BW, Tan S, et al. Multicenter evaluation of theVerigene Clostridium difficile nucleic acid assay. J Clin Microbiol 2013;51: 4120–5.

91. Denys GA. Portrait Toxigenic Clostridium difficile assay, an isothermalamplification assay detects toxigenic C. difficile in clinical stool speci-mens. Exp Rev Mol Diagn 2014; 14: 17–26.



92. Buchan BW, Mackey TL, Daly JA, et al. Multicenter clinical evaluationof the Portrait Toxigenic C. difficile assay for detection of toxigenicClostridium difficile strains in clinical stool specimens. J Clin Microbiol2012; 50: 3932–6.

93. Hong G, Park KS, Ki CS, et al. Evaluation of the illumigene C. difficileassay for toxigenic Clostridium difficile detection: a prospective study of302 consecutive clinical fecal samples. Diagn Microbiol Infect Dis 2014;80: 177–80.

94. Hart J, Putsathit P, Knight DR, et al. Clostridium difficile infectiondiagnosis in a paediatric population: comparison of methodologies. EurJ Clin Microbiol Infect Dis 2014; 33: 1555–64.

95. Soh YS, Yang JJ, You E, et al. Comparison of two molecular methods fordetecting toxigenic Clostridium difficile. Ann Clin Lab Sci 2014; 44: 27–31.

96. McElgunn CJ, Pereira CR, Parham NJ, et al. A low complexity rapidmolecular method for detection of Clostridium difficile in stool. PLoS One2014; 9: e83808.

97. Noren T, Unemo M, Magnusson C, et al. Evaluation of the rapidloop-mediated isothermal amplification assay illumigene for diagnosisof Clostridium difficile in an outbreak situation. APMIS 2014; 122: 155–60.

98. Couturier B, Schlaberg R, Konzak C, et al. TcdA as a diagnostic target in aloop-mediated amplification assay for detecting toxigenic Clostridiumdifficile. J Clin Lab Anal 2013; 27: 171–6.

99. Walkty A, Lagace-Wiens PR, Manickam K, et al. Evaluation of analgorithmic approach in comparison with the illumigene assay forlaboratory diagnosis of Clostridium difficile infection. J Clin Microbiol2013; 51: 1152–7.

100. Bamber AI, Fitzsimmons K, Cunniffe JG, et al. Diagnosis of Clostridiumdifficile-associated disease: examination of multiple algorithms usingtoxin EIA, glutamate dehydrogenase EIA and loop-mediated isothermalamplification. Br J Biomed Sci 2012; 69: 112–8.

101. Bruins MJ, Verbeek E, Wallinga JA, et al. Evaluation of three enzymeimmunoassays and a loop-mediated isothermal amplification test for thelaboratory diagnosis of Clostridium difficile infection. Eur J Clin Micro-biol Infect Dis 2012; 31: 3035–9.

102. Viala C, Le Monnier A, Maataoui N, et al. Comparison of commercialmolecular assays for toxigenic Clostridium difficile detection in stools: BDGeneOhm Cdiff, XPert C. difficile and illumigene C. difficile. J MicrobiolMethods 2012; 90: 83–5.

103. Pancholi P, Kelly C, Raczkowski M, et al. Detection of toxigenicClostridium difficile: comparison of the cell culture neutralization, XpertC. difficile, Xpert C. difficile/Epi, and illumigene C. difficile assays. J ClinMicrobiol 2012; 50: 1331–5.

104. Boyanton BL Jr, Sural P, Loomis CR, et al. Loop-mediated isothermalamplification compared to real-time PCR and enzyme immunoassay fortoxigenic Clostridium difficile detection. J Clin Microbiol 2012; 50: 640–5.

105. Lalande V, Barrault L, Wadel S, et al. Evaluation of a loop-mediatedisothermal amplification assay for diagnosis of Clostridium difficileinfections. J Clin Microbiol 2011; 49: 2714–6.

106. Shin S, Kim M, Kim M, et al. Evaluation of the Xpert Clostridium difficileassay for the diagnosis of Clostridium difficile infection. Ann Lab Med2012; 32: 355–8.

107. Chapin KC, Dickenson RA, Wu F, et al. Comparison of five assays fordetection of Clostridium difficile toxin. J Mol Diagn 2011; 13: 395–400.

108. Zidaric V, Kevorkijan BK, Oresic N, et al. Comparison of two commercialmolecular tests for the detection of Clostridium difficile in the routinediagnostic laboratory. J Med Microbiol 2011; 60: 1131–6.

109. Selvaraju SB, Gripka M, Estes K, et al. Detection of toxigenic Clostridiumdifficile in pediatric stool samples: an evaluation of Quik Check CompleteAntigen assay, BD GeneOhm Cdiff PCR, and ProGastro Cd PCR assays.Diagn Microbiol Infect Dis 2011; 71: 224–9.

110. Knetsch CW, Bakker D, de Boer RF, et al. Comparison of real-time PCRtechniques to cytotoxigenic culture methods for diagnosing Clostridiumdifficile infection. J Clin Microbiol 2011; 49: 227–31.

111. Terhes G, Urban E, Soki J, et al. Comparison of a rapid molecular method,the BD GeneOhm Cdiff assay, to the most frequently used laboratory testsfor detection of toxin-producing Clostridium difficile in diarrheal feces.J Clin Microbiol 2009; 47: 3478–81.

112. Stamper PD, Alcabasa R, Aird D, et al. Comparison of a commercial real-time PCR assay for tcdB detection to a cell culture cytotoxicity assay andtoxigenic culture for direct detection of toxin-producing Clostridiumdifficile in clinical samples. J Clin Microbiol 2009; 47: 373–8.

113. Buss SN, Leber A, Chapin K, et al. Multicenter evaluation of the BioFireFilmArrayTM gastrointestinal panel for the etiologic diagnosis of infec-tious gastroenteritis. J Clin Microbiol 2015; 53: 915–25.

114. Stockmann C, Rogatcheva M, Harrel B, et al. How well does physicianselection of microbiologic tests identify Clostridium difficile and otherpathogens in paediatric diarrhoea? Insights using multiplex PCR-baseddetection. Clin Microbiol Infect 2015; 21: . 179.e9–15.


Original article

Incorrect diagnosis of Clostridium difficile infection in a universityhospital in Japan

Nobuaki Mori a, 1, Sadako Yoshizawa b, 1, Tomoo Saga a, Yoshikazu Ishii a, b,Hinako Murakami c, Morihiro Iwata c, Deirdre A. Collins d, Thomas V. Riley d, e,Kazuhiro Tateda a, b, c, *

a Department of Microbiology and Infectious Diseases, Toho University School of Medicine, Tokyo, Japanb Department of Infection Control, Toho University Medical Center, Omori Hospital, Tokyo, Japanc Laboratory Microbiological Section, Toho University Medical Center, Omori Hospital, Tokyo, Japand Department of Microbiology and Immunology, School of Pathology and Laboratory Medicine, The University of Western Australia, Perth, Australiae Division of Microbiology and Infectious Diseases, PathWest Laboratory Medicine, Queen Elizabeth II Medical Centre, Perth, Australia


Article history:Received 29 January 2015Received in revised form22 June 2015Accepted 27 June 2015Available online 6 July 2015

Keywords:Clostridium difficile infectionBinary toxinPCR ribotypeEpidemiologyIncorrect diagnosis

a b s t r a c t

Physicians often fail to suspect Clostridium difficile infection (CDI) and many microbiology laboratoriesuse suboptimal diagnostic techniques. To estimate the extent of and reasons for incorrect diagnosis ofCDI in Japan, we investigated toxigenic C. difficile isolated from all stool culture samples and clinicalcourse. Over a 12-month period in 2010, all stool culture samples (n ¼ 975) submitted from inpatients ina university hospital in Japan were cultured for C. difficile and routine microbiological testing was con-ducted. In total, 177 C. difficile isolates were recovered, and 127 isolates were toxigenic. Among the toxin-A-positive/toxin-B-positive isolates, 12 were also positive for the binary toxin gene. However, clinicallyimportant ribotypes, such as 027 and 078, were not identified. A total of 58 (45.7%) cases with toxigenicC. difficile had unformed stool, and the incidence CDI was 1.6 cases per 10,000 patient-days. Of these 58cases, 40 were not diagnosed in routine testing due to a lack of clinical suspicion (24.1%, 14/58) or anegative C. difficile toxin assay result (44.8%, 26/58). A stool toxin assay was performed in 54 patients(78.2%, 54/69) who did not have unformed stool. The present study demonstrated that a significantnumber of CDI cases in Japan might be overlooked or misdiagnosed in clinical practice due to a lack ofclinical suspicion and limitations of microbiological testing for CDI in Japan. Providing education topromote awareness of CDI among physicians is important to improve the accuracy of diagnosis in Japan.

© 2015, Japanese Society of Chemotherapy and The Japanese Association for Infectious Diseases.Published by Elsevier Ltd. All rights reserved.

1. Introduction

Clostridium difficile is a leading cause of hospital- andantimicrobial-associated diarrhea. Following reports of outbreaksof a hypervirulent strain of C. difficile in Canada and the UnitedStates, there has been a marked global increase in the reportedincidence of C. difficile infection (CDI) and associated mortality overthe last decade [1].

Nationwide surveillance of CDI has already reported in theUnited States and Europe [2,3]. A definition for CDI has been pro-posed by the Center for Disease Control and Prevention and Euro-pean Centre for Disease Prevention and Control (ECDC), whichincludes the presence of 3 or more unformed stools per day incombination with a positive toxin test or isolation of toxin-producing C. difficile [4,5]. This definition is commonly used inepidemiological studies involving CDI. However, the diagnosis ofCDI remains an important clinical problem. Previous reports fromEurope suggest that underdiagnosis of CDI may occur due to lack ofclinical suspicion and request, or the low sensitivity of C. difficiletoxin assays when compared to themolecular detection of toxins instool and cultivation of toxigenic C. difficile on selective media.

Recently, epidemiological data about the incidence, variation,and frequency of toxigenic CDI have gradually been reported in

* Corresponding author. 5-21-16 Ohmori-nishi, Ohtaku, Tokyo 143-8540, Japan.Tel.: þ81 03 5763 6702.

E-mail address: [email protected] (K. Tateda).1 N.M. and S.Y. contributed equally to this work.


Journal of Infection and Chemotherapy

journal homepage: http: / /www.elsevier .com/locate/ j ic

http://dx.doi.org/10.1016/j.jiac.2015.06.0091341-321X/© 2015, Japanese Society of Chemotherapy and The Japanese Association for Infectious Diseases. Published by Elsevier Ltd. All rights reserved.

J Infect Chemother 21 (2015) 718e722


http://crossmark.crossref.org/dialog/?doi=10.1016/j.jiac.2015.06.009&domain=pdf

www.sciencedirect.com/science/journal/1341321X

http://www.elsevier.com/locate/jic

http://dx.doi.org/10.1016/j.jiac.2015.06.009



Japan [6e9]. However, a multi-institutional study is problematic inthat it included a low number of registered CDI cases [7]. A retro-spective cohort study of CDI in a tertiary hospital suggests that therelatively low incidence of CDI might be due to the use of relativelyinsensitive tests and relatively low diagnostic densities [8]. The trueburden of CDI in Japan remains unknown.

To estimate the extent of and reasons for incorrect diagnosis ofCDI, we investigated the toxigenic C. difficile isolated from all stoolculture samples in the Toho University Medical Center, OmoriHospital in Japan. The molecular and clinical characteristics of CDIcases in the hospital were also investigated.

2. Materials and methods

2.1. Study design

A retrospective analysis of a stool culture database was con-ducted at Toho University Medical Center, Omori Hospital andapproved by the ethical committee of the Toho University School ofMedicine (No. 24018). Our hospital is a 972-bed university-affiliated, tertiary care teaching hospital located in Tokyo, Japan.

In our hospital, stool culture samples with any gastrointestinalsymptoms have been examined using culture of C. difficile, irre-spective of the physician's request, in addition to routine microbi-ological testing. We investigated the stool culture samples isolatedfrom inpatients over a 12-month period in 2010, but excluded frompatients aged 2 years or less. A PCR assay of the toxin gene wasconducted for C. difficile isolates. We investigated the molecularcharacterization of the C. difficile isolates and the patient's clinicaldata including whether the C. difficile toxin assay was performed.

2.2. Isolation of C. difficile

To isolate C. difficile from the stool culture samples, acycloserine-cefoxitin-mannitol agar plate (Nissui Pharmaceutical,Co., Ltd. Tokyo, Japan) was inoculated and incubated anaerobically.Suspected colonies based on colonial morphology with typicalGram stain appearance were further examined for production of L-proline aminopeptidase (PRO-AN test; Mitsubishi Chemical Co.Tokyo, Japan). Isolates were confirmed as being C. difficile by thepresence of the triose phosphate isomerase (tpi) gene (a C. difficilehousekeeping gene), and then stored at �80 �C until further mo-lecular analysis.

2.3. Molecular characterization of C. difficile isolates

For genomic DNA, bacterial pellets were prepared from anaer-obic cultivation of C. difficile in a brain-heart infusion agar withyeast extract, containing 10% L-cysteine and 10% taurocholic acid(BHIS) for 48 h at 35 �C. DNA extraction was performed using acommercial kit (Wizard Genomic DNA purification kit; Promega,Madison, WI), according to the manufacturer's instructions.

The toxin-A (tcdA), toxin-B (tcdB), and binary toxin (cdtA andcdtB) genes were detected using a PCR assay with primer sets, aspreviously described [10]. PCR ribotyping was performed using themethod of Stubbs et al. [11], with some modifications. PCR ribo-typing reaction products were concentrated using a Min-Elute PCRpurification kit and run on a QIAxcel capillary electrophoresisplatform (Qiagen, Hilden, Germany). Visualization of PCR productswas performed using QIAxcel ScreenGel software. PCR ribotypingbanding patterns were identified by comparing banding patternswith a reference library consisting of a collection of referencestrains from the ECDC, a collection of the most prevalent PCRribotypes currently circulating in Australia (B. Elliott, unpublisheddata), and a selection of binary toxin-positive strains. Interpretation

of the capillary electrophoresis data (PCR ribotyping banding pat-terns) was performed using a dendrogram and cluster analysisusing the Dice coefficient within the BioNumerics software packagev.6.5 (Applied Maths, Sint-Martens-Latem, Belgium). Isolates thatcould not be identified with the available reference library weredesignated with internal nomenclature.

We investigated the tcdC gene and antimicrobial susceptibilityin isolates that were positive for the binary toxin gene. The entiretcdC gene was amplified using primer sets C1 and C2, as previouslydescribed [12]. Amplified PCR products were purified using QIA-quick PCR Purification Kit (Qiagen, Tokyo, Japan), according to themanufacturer's instructions. Sequencing of PCR products was per-formed by FASMAC Co., Ltd. (Tokyo, Japan). The sequence data werecompared to the published tcdC sequence of C. difficile 630 andribotype 027. Antimicrobial susceptibility tests for metronidazole,vancomycin and moxifloxacin were performed by using the Etest(AB Biodisk, Solna, Sweden) on Centers for Disease Controlanaerobe blood agar (BD, Tokyo, Japan), as previously described[13]. The minimum inhibitory concentrations values were deter-mined 48 h after incubation. According to the Clinical and Labo-ratory Standards Institute (CLSI) guidelines, the followingbreakpoints were used for metronidazole: � 8 mg/mL (susceptible),16 mg/mL (intermediate) and >16 mg/mL (resistant). There are noestablished CLSI breakpoints for vancomycin and moxifloxacin;therefore, the following values were used: �2 mg/mL (susceptible),4 mg/mL (intermediate) and >4 mg/mL (resistant), as previouslydescribed [14].

To confirm the toxin production of C. difficile, the isolates thatwere positive for at least one of the toxin genes were re-culturedanaerobically on BHIS agar (48 h, 35 �C), and then transferred tothe BHIS broth for further anaerobic incubation (72 h, 35 �C). Thecultured broth was tested using an enzyme immune assay (EIA)-based rapid combination test kit (C. DIFF Quick Check Complete;Techlab, Inc., Blacksburg, VA), according to the manufacturer'sinstructions.

2.4. Clinical data collection

The clinical data of CDI patients were collected as follows: age,sex, the date of admission, the date of isolation of C. difficile, theunderlying disease (inflammatory bowel diseases, heart failure,respiratory failure, chronic kidney disease, diabetes mellitus, ma-lignancy, cerebrovascular disease), any surgery prior to developingCDI, the use of antibiotics or gastric acid suppressors prior todevelopment of CDI, laboratory data at the time of diagnosis (whiteblood cell counts [WBC], serum albumin, serum creatinine, and C-reactive protein), the drug used to treat CDI, outcome, and whetheror not the C. difficile toxin assay (Uniquick; Kanto Chemical Co., Inc.Tokyo, Japan) was performed.

2.5. Definition

For the purpose of this study, toxigenic C. difficilewas defined asthe C. difficile with the presence of any toxin gene and toxin pro-duction detected by EIA-based rapid test kit. CDI was defined as thepresence of diarrhea and toxigenic C. difficile isolate. Health carefacility-onset CDI was defined as symptom onset >48 h fromadmission. Severe CDI was classified according to the criteria of Zaret al. [15] Briefly, patients with more than 2 points were classifiedas severe CDI. One point each was given for age >60 years, tem-perature >38.3 �C, serum albumin level <2.5 mg/dL, or WBC count>15,000 cell/mm3 within 48 h of diagnosis of CDI. Two points weregiven for endoscopic evidence of pseudomembranous colitis ortreatment in the intensive care unit due to CDI.

N. Mori et al. / J Infect Chemother 21 (2015) 718e722 719

2.6. Statistical analyses

The incidence of CDI was expressed per 10,000 patient-days.Testing densities were calculated using the number of toxin as-says tested per 10,000 patient-days. We calculated summary sta-tistics for demographic and clinical characteristics. Statisticalanalyses were performed using GraphPad Prism 5 software(GraphPad, Inc., La Jolla, CA).

3. Results

3.1. Molecular characterization of toxigenic C. difficile

During the 12-month study period, 975 stool culture sampleswere examined and a total of 177 nonduplicate isolates of C. difficilewere recovered. Of the 177 isolates,124 (70.0%) were toxin-A/toxin-B positive (AþBþ), and 3 strains (1.7%) were toxin-A negative/toxin-B positive (A-Bþ). Among the AþBþ isolates, 12 strains (6.8%,12/177) were positive for the binary toxin gene. The prevalence rateof toxigenic C. difficile in all stool culture samples was 13.0% (127/975). Further examination of toxin production of C. difficilewith anytoxin gene was conducted in culture medium. All examined isolatesthat were positive for the toxin gene exhibited a positive toxin testin culture medium, suggesting that the detected toxin genes werefunctional.

A total of 111 isolates was available for PCR ribotyping analysis.The most common ribotype was 369 at 21.6% (24/111), followed by018 at 10.8% (12/111), and the 014/020 group and 002 at 9.9% (11/111) each. These top four ribotypes comprised 52.2% of the total.Although we found 12 isolates possessing the binary toxin gene,there were no ribotype 027 isolates. Of 12 isolates, ribotype 127comprised five, ribotype 131 comprised two, ribotypes QX046, andQX060 comprised one each, and three were unknown. No isolate

was resistant to moxifloxacin, metronidazole, and vancomycin. Todetect the deletion mutation, sequence analysis of the tcdC genewas performed in 11 strains (the DNA sample was not amplified inone strain). There were several mutations, including an 18-bpdeletion at positions 330e347 (2 isolates) and a 39-bp deletion atpositions 341e379 (6 isolates). A deletion at position 117 was notfound in isolates with the binary toxin gene.

3.2. CDI diagnosis

To estimate the extent of and reasons for incorrect diagnosis ofCDI, we investigated the clinical course of patients with toxigenicC. difficile. Fig. 1 provides information about the clinical course ofpatients with toxigenic C. difficile, including the information of stoolform and whether the physicians requested a toxin assay. Of 127toxigenic C. difficile isolates, 58 (45.7%) were unformed stool spec-imens.We identified 58 of the patients as CDI; of these, a stool toxinassaywas performed in 44 (75.9%, 44/58) at the physician's request.Positive results for toxigenic C. difficile were found in 18 of thesecases (40.9%, 18/44). On the other hand, stool culture samples from69 patients who had any gastrointestinal symptoms did not haveunformed stool. However, a stool toxin assay was performed in 54(78.2%, 54/69) at the physician's request. These results suggest thatmany patients with any gastrointestinal symptoms went under-diagnosed or weremisdiagnosed, owing to lack of clinical suspicionand request of the needed tests, or low sensitivity of the diagnostictest.

3.3. Clinical characteristic of CDI patients

Demographic and clinical characteristics of CDI patients areshown in Table 1. Nine patients (15.5%) were diagnosed within 48 hafter admission. The mean age was 67.5 years old (range, 4e94

Fig. 1. Clinical course of 127 cases with toxigenic C. difficile.

N. Mori et al. / J Infect Chemother 21 (2015) 718e722720

years), and 30 (51.7%) were women. CDI patients spent a median of26.9 days (range; 1e199) in the hospital prior to CDI development.The incidence of healthcare facility-onset CDI was 1.6 cases per10,000 patient-days. Testing densities for C. difficile were 37 pa-tients tested per 10,000 patient-days.

Among these patients, 53 (91.4%, 53/58) had received antimi-crobials, and 36 (62.0%, 36/58) had received gastric acid suppres-sors (H2 blocker and proton pump inhibitor) in the 60 dayspreceeding the CDI development. Although 29 patients (50.0%)fulfilled the criteria according to Zar et al. for severe CDI, no patientsunderwent CDI-related surgery. A recurrence was observed in 3(5.2%) patients within 60 days following the onset of symptoms.The 30-day all-cause mortality was 6.9% (4/58).

4. Discussion

The current study highlights the possible incorrect diagnosis ofCDI relating to a lack of clinical recognition and limitations in toxindetection methodology; moreover, it documents the characteristicsof CDI in a university hospital in Japan.

The diagnosis of CDI remains a challenge to microbiology lab-oratories. A significant number of clinical laboratories are currentlyutilizing EIAs for C. difficile toxin detection for routine microbio-logical testing, primarily due to their ease of use. However, recentdata clearly demonstrated that C. difficile toxin assays are lesssensitive than molecular detection methods and cultivation oftoxigenic C. difficile. Optimum laboratory diagnosis of CDI dependson testing relevant patients at the correct time with suitable tests[16]. An investigation of the underdiagnosis of CDI in hospitals in 20European countries showed that 23% of CDI patients were not

diagnosed because of an absence of clinical suspicion [16]. A seriesof laboratories in Spain on a single day performed C. difficile cul-tures on all unformed stool specimens that were provided, irre-spective of the type of request [17]. They detected toxigenic isolatesin 45 specimens from 730 patients, and concluded that two out ofevery three episodes were underdiagnosed or were misdiagnosed,owing to insensitive diagnostic tests (19.0%) or lack of clinicalsuspicion and request (47.6%). Furthermore, a subsequent Spanishstudy showed that half of CDI episodes were not diagnosed despitevarious interventions were undertaken [18]. Similar findings arepresented here, with 69.0% of CDI cases undiagnosed owing tonegative C. difficile toxin assay results (44.9%) or a lack of clinicalsuspicions (24.1%). An educational program to raise awareness ofCDI among physicians and more sensitive tests might optimize thediagnosis of CDI. A stool toxin assay was performed in 54 patients(78.2%, 54/69) who did not have unformed stool. The diagnostic testmight be inappropriately performed. However, 17 patients who didnot have unformed stool had gastrointestinal symptoms and pos-itive results in the toxin assay and toxigenic cultures. Whether thepatients subsequently develop CDI was unknown. Therefore, CDImay not be diagnosed accurately due to the difficulty in correctlyevaluating diarrheal symptoms relating to the disease [19,20].

The incidence of healthcare facility-onset CDI in our hospitalwas estimated at 1.6 cases per 10,000 patient-days, and the 30-daymortality was 6.9% during this study. Recently, ECDIS reported thatthe incidence of CDI in 34 European countries was 4.1 (weightedmean) per 10,000 patient-days per hospitals (range, 0.0e36.3), andCDI was related to 40% (40/101) of the deaths reported during thefollow-up period [3]. National Healthcare Safety Network datasuggest that in US hospitals, hospital-onset CDI cases occurred in anaverage of 6e8 cases per 10,000 patient-days [2]. Data from Asiancountries are limited. Healthcare facility-onset CDI incidence was3.11 cases per 10,000 patient-days in a Japanese tertiary care center[8]. Overall, these results may indicate the incidence of CDI in Japanis lower than in Western countries. However, CDI rates may beinfluenced by a variety of factors, including prevalent strains in aparticular area, the definition of disease used, testing densities,microbiological testing methodology, and trends in the use of an-tibiotics, age of patients, underlying diseases, hospital size, andstudy design. It may therefore be inappropriate to simply comparethe number of cases per 10,000 patient-days from each study.

Although we identified the isolates with binary toxin gene,ribotype 027 strains were not found and they had no mutations atposition 117 in the tcdC gene in this study. The appearance andspread of the epidemic “hypervirulent” ribotype 027 strain radi-cally altered the epidemiological pattern of CDI in North Americaand Europe [21]. This strain has been associated with severe dis-ease, severe outcome, and death within 14 days [22]. Fortunately inJapan, there have been few reports of CDI cases associated withribotype 027 strains [23e25]. These ribotype 027 strains in Japanwhich were susceptible to fluoroquinolones were similar to the“historic ribotype 027 isolates” found in the United States in 1990except a single case report [1]. The CDI cases with binary toxinwerenot necessarily associated with poor outcome. During the studyperiod, there was no CDI outbreak. Therefore, there was noobserved increase in any particular strain. Ribotype 018 was thesecond most common ribotype (13.8%) isolated from CDI patients.This ribotype has been previously recognized in Japan as slpAsequence typing smz and appears to have persisted as the mostcommon ribotype for over a decade [26]. This strain may be one ofthe endemic strains in Japan.

The strength of this study is the fact that all stool samplesvoluntarily submitted from inpatients with gastrointestinal symp-tomswere examined using culture of C. difficile and testing for toxingenes. Thus, approximately 1000 stool specimens were

Table 1Demographic and clinical characteristic of CDI patients.

Patient characteristics n ¼ 58 (%)

Age 67.5 ± 20.7Women 30 (51.7)Underlying diseaseHeart failure 20 (34.4)Respiratory failure 11 (19.0)Diabetes Mellitus 26 (44.8)Inflammatory bowel disease 15 (25.9)Chronic kidney disease 20 (34.4)Cerebral vascular disease 16 (27.6)Malignacy 29 (50.0)

Surgery in the past 60 days 14 (24.1)Prior exposure of antibiotics in the past 60 days 53 (91.4)Prior exposure of gastric acid suppressors in the past 60 days 36 (62,0)Days of administration before CDI development 26.9 ± 36.7Blood temperature (�C) 38.1 ± 0.8Laboratory dataWhite blood cell (cell/mm3) 8955 ± 4842Serum albumin (mg/dL) 2.6 ± 0.7Serum creatinine (mg/dL) 1.11 ± 1.1C-reactive protein 5.74 ± 5.84

RibotypeUK369 9 (15.5)UK018 8 (13.8)UK002 6 (10.3)UK127 4 (6.9)QX029 3 (5.2)UK014/020, UK056, UK012, UK001/271 2 (3.4)UK015/193, QX161, UK131, QX026, QX013, QX133, UK014,QX049

1 (1.7)

Drug used to treat CDIMetronidazole 3 (5.2)Vancomycin 23 (39.7)

Severe CDI (by the criteria of Zar et al.) 29 (50.0)CDI recurrence 3 (5.2)The 30-day all-cause of death 4 (6.9)

N. Mori et al. / J Infect Chemother 21 (2015) 718e722 721

prospectively examined. However, there are limitations in thepresent data. Our study was conducted in only one facility over ashort time period, restricting the ability to generalize the results tothe Japanese population as a whole. Nonetheless, the majoroutbreak strains that are prevalent in North America or Europe,including 027 and 078, were not present in our hospital. We couldnot investigate patients whose fecal samples could not be collectedbecause of an absence of clinical suspicion even if the patients haddiarrhea. Thus, a prospective study is necessary to evaluate thispopulation.

In conclusion, the present study demonstrated that a significantnumber of CDI cases might be overlooked or misdiagnosed inclinical practice due to a lack of clinical suspicion and limitations ofmicrobiological testing for CDI in Japan. The incidence of CDI in thisstudy was lower than that in North America or Europe, but thestrains were different from those found elsewhere. Providing ed-ucation to promote awareness of CDI among physicians is impor-tant to improve the accuracy of diagnosis in Japan.

Conflict of interests

None.

Acknowledgment

Wewould like to thank the laboratory staff of the Department ofMicrobiology and Infectious Diseases, Toho University School ofMedicine, for their assistance and advice.

References

[1] McDonald LC, Killgore GE, Thompson A, Owens RC, Kazakova SV, Sambol SP,et al. An epidemic, toxin gene-variant strain of Clostridium difficile. N Engl JMed 2005;353:2433e41.

[2] Lessa FC, Gould CV, McDonald LC. Current status of Clostridium difficileinfection epidemiology. Clin Infecti Dis 2012;55:S65e70.

[3] Bauer M, Notermans D, Van Benthem B. Clostridium difficile infection inEurope: a hospital-based survey. Lancet 2011;377:63e73.

[4] Kuijper EJ, Coignard B, Tull P. ESCMID study group for Clostridium difficile, EUmember states, European Centre for disease prevention and control. Emer-gence of Clostridium difficile-associated disease in North America and Europe.Clin Microbiol Infect 2006;12(Suppl 6):2e18.

[5] McDonald LC, Coignard B, Dubberke E, Song X, Horan T, Kutty PK, et al. Rec-ommendations for surveillance of Clostridium difficileeassociated disease.Infect Control Hosp Epidemiol 2007;28:140e5.

[6] Iwashima Y, Nakamura A, Kato H, Kato H, Wakimoto Y, Wakiyama N, et al.A retrospective study of the epidemiology of Clostridium difficile infection at auniversity hospital in Japan: genotypic features of the isolates and clinicalcharacteristics of the patients. J Infect Chemother 2010;16:329e33.

[7] Takahashi M, Mori N, Bito S. Multi-institution case-control and cohort study ofrisk factors for the development and mortality of Clostridium difficile infectionsin Japan. BMJ Open 2014;4:e005665.

[8] Honda H, Yamazaki A, Sato Y, Dubberke ER. Incidence and mortality associ-ated with Clostridium difficile infection at a Japanese tertiary care center.Anaerobe 2014;25:5e10.

[9] Kuwata Y, Tanimoto S, Sawabe E, Shima M, Takahashi Y, Ushizawa H, et al.Molecular epidemiology and antimicrobial susceptibility of Clostridium difficileisolated from a university teaching hospital in Japan. Eur J Clin MicrobiolInfect Dis 2015;34:763e72.

[10] Lemee L, Dhalluin A, Testelin S, Mattrat MA, Maillard K, Lemeland JF, et al.Multiplex PCR targeting tpi (triose phosphate isomerase), tcdA (Toxin A), andtcdB (Toxin B) genes for toxigenic culture of Clostridium difficile. J ClinMicrobiol 2004;42:5710e4.

[11] Stubbs SL, Brazier JS, O'Neill GL, Duerden BI. PCR targeted to the 16S-23S rRNAgene intergenic spacer region of Clostridium difficile and construction of a li-brary consisting of 116 different PCR ribotypes. J Clin Microbiol 1999;37:461e3.

[12] Spigaglia P, Mastrantonio P. Molecular analysis of the pathogenicity locus andpolymorphism in the putative negative regulator of toxin production (TcdC)among Clostridium difficile clinical isolates. J Clin Microbiol 2002;40:3470e5.

[13] Wong SS, Woo PC, Luk WK, Yuen KY. Susceptibility testing of Clostridiumdifficile against metronidazole and vancomycin by disk diffusion and Etest.Diagn Microbiol Infect Dis 1999;34:1e6.

[14] Muto CA, Pokrywka M, Shutt K, Mendelsohn AB, Nouri K, Posey K, et al. A largeoutbreak of Clostridium difficile-associated disease with an unexpected pro-portion of deaths and colectomies at a teaching hospital following increasedfluoroquinolone use. Infect Control Hosp Epidemiol 2005;26:273e80.

[15] Zar FA, Bakkanagari SR, Moorthi KM, Davis MB. A comparison of vancomycinand metronidazole for the treatment of Clostridium difficile-associated diar-rhea, stratified by disease severity. Clin Infect Dis 2007;45:302e7.

[16] Davies KA, Longshaw CM, Davis GL, Bouza E, Barbut F, Barna Z, et al. Under-diagnosis of Clostridium difficile across Europe: the European, multicentre,prospective, biannual, point-prevalence study of Clostridium difficile infectionin hospitalised patients with diarrhoea (EUCLID). Lancet Infect Dis 2014;14:1208e19.

[17] Alcal�a L, Martin A, Marín M, S�anchez-Somolinos M, Catal�an P, Pel�aez T, et al.The undiagnosed cases of Clostridium difficile infection in a whole nation:where is the problem? Clin Microbiol Infect 2012;18:E204e13.

[18] Alcal�a L, Reigadas E, Marín M, Martin A, Catal�an P, Bouza E, et al. Impact ofclinical awareness and diagnostic tests on the underdiagnosis of Clostridiumdifficile infection. Eur J Clin Microbiol Infect Dis 2015, Apr 24. [Epub ahead ofprint].

[19] Peterson LR, Manson RU, Paule SM, Hacek DM, Robicsek A, Thomson RB, et al.Detection of toxigenic Clostridium difficile in stool samples by real-time po-lymerase chain reaction for the diagnosis of C. difficile-associated diarrhea.Clin Infect Dise 2007;45:1152e60.

[20] Dubberke ER, Han Z, Bobo L, Hink T, Lawrence B, Copper S, et al. Impact ofclinical symptoms on interpretation of diagnostic assays for Clostridiumdifficile infections. J Clin Microbiol 2011;49:2887e93.

[21] Louie T, Miller M, Mullane K. Fidaxomicin versus vancomycin for Clostridiumdifficile infection. N Engl J Med 2011;364:422e31.

[22] See I, Mu Y, Cohen J, Beldavs ZG, Winston LG, Dumyati G, et al. NAP1 straintype predicts outcomes from Clostridium difficile infection. Clin Infect Dis2014;58:1394e400.

[23] Kato H, Ito Y, van den Berg RJ, Kuijper EJ, Arakawa Y. First isolation of Clos-tridium difficile 027 in Japan. Euro Surveill 2006;12:E070111e3.

[24] Sawabe E, Kato H, Osawa K, Chida T, Tojo N, Arakawa Y, et al. Molecularanalysis of Clostridium difficile at a university teaching hospital in Japan: a shiftin the predominant type over a five-year period. Eur J Clin Microbiol Infect Dis2007;26:695e703.

[25] Nakamura I, Yamaguchi T, Tsukimori A, Sato A, Fukushima S, Mizuno Y, et al.Fiulminant colitis from Clostridium difficile infection, the epidemic strainribotype 027, in Japan. J Infect Dis 2014;20:380e3.

[26] Collins DA, Hawkey PM, Riley TV. Epidemiology of Clostridium difficile infec-tion in Asia. Antimicrob Resist Infect Control 2013;2:21.

N. Mori et al. / J Infect Chemother 21 (2015) 718e722722

http://refhub.elsevier.com/S1341-321X(15)00153-1/sref1


















































































































Routine detection of Clostridium difficile in Western Australia

Deirdre A. Collins a, *, Thomas V. Riley a, b

a Microbiology and Immunology, School of Pathology and Laboratory Medicine, The University of Western Australia, Perth, Australiab Department of Microbiology, PathWest Laboratory Medicine, WA, Perth, Australia


Article history:Received 30 July 2015Received in revised form6 November 2015Accepted 18 November 2015Available online 22 November 2015

Keywords:Clostridium difficile infectionEpidemiologyCommunity associated CDIPaediatric CDIDiagnosis

a b s t r a c t

Despite increasing infection rates, Clostridium difficile is not currently routinely tested for in all diarrhoealfaecal specimens in Australia. In July 2014, all diarrhoeal specimens submitted to a diagnostic laboratoryin Western Australia were surveyed to determine the true prevalence of C. difficile. In total, 1010 diar-rhoeal non-duplicate faecal specimens were received during the month. Testing for C. difficile wasrequested, or the criteria for a C. difficile investigation were met, for 678 specimens which were inves-tigated by PCR for the tcdB gene using the BD MAX platform, followed by toxigenic culture on PCR-positive samples. The remaining 332 specimens, with either no C. difficile test request or the criteria for aC. difficile investigation were not met, were examined by toxigenic culture. All isolates were PCR ribo-typed. C. difficilewas the most commonly detected diarrhoeal pathogen among all specimens. The overallprevalence of C. difficile in all 1010 specimens was 6.4%; 7.2% in the routinely tested group, and 4.8% in thenon-requested group. The proportion of non-requested positive detections among all cases was 24.6%.Community-onset infection was present in 50.8% of all cases. The median age of all CDI cases was 60.0years and the age range in CDI patients in the routine group was 0.6e96.6 years (median 72.7 years),compared to 0.2e2.3 years (median 0.8 years) in the non-requested group. The most common ribotype(RT) found was RT 014/020 (34.1% in the routine group, 43.8% in the non-requested group), followed byRTs 002, 056, 005 and 018. While the routine testing group and the non-requested group differedmarkedly in age and patient classification, C. difficile was the most common cause of diarrhoea in hos-pitals and the community in Western Australia. The significance of finding C. difficile in the communitypaediatric population requires further study.

© 2015 Elsevier Ltd. All rights reserved.

1. Introduction

Clostridium difficile is the most common cause of healthcare-associated infection in the developed world [1]. C. difficile infec-tion (CDI) is caused by C. difficilewhich generally produces toxins A(enterotoxin) and B (cytotoxin) and, occasionally, a binary toxin(CDT). CDI manifests as mild to severe diarrhoea following inges-tion of spores when the endogenous gut microflora have beenreduced or depleted, usually by antimicrobial use. Advanced ageand hospitalisation are additional significant risk factors for CDI [2].The incidence and severity of CDI has increased worldwide in thepast 20 years, with the emergence of epidemic strains such asribotype (RT) 027 which caused major outbreaks of CDI in North

America and Europe [3]. More recently an increase in community-associated (CA)-CDI and/or community-onset (CO)-CDI has beennoted internationally, where cases have been younger and lesslikely to have used antimicrobials [4]. CA-CDI is considered where acase has not been hospitalised in the 12 weeks prior to onset oftheir symptoms, and onset occurs before or within 48 h of hospi-talisation, whereas CO-CDI refers to an episode of CDI which hascommenced in the community [5].

To date, CDI has not been extensively studied in Australia. In arecent nationwide study the incidence of hospital-identified CDIincreased significantly from3.2/10,000 bed days (BD) in 2011 to 4.0/10,000 BD in 2012. Among these cases 26%were CA-CDI [6]. This risecannot be completely explained by increased testing or alteredtesting practices, and other Australian studies have indicated thatthe greatest proportion of the increased national incidence isattributable to CA-CDI [7]. Recently, a new strain of C. difficile (RT244) emerged to cause severe CA-CDI in Australia and New Zealandin 2011 and 2012 [8,9]. This strain rose to become the 3rd mostcommon RT circulating in the state of Queensland (and presumably

* Corresponding author. School of Pathology and Laboratory Medicine (M504)Faculty of Medicine, Dentistry and Health Sciences, The University of WesternAustralia, 35 Stirling Highway, Crawley, WA 6009, Australia.

E-mail address: [email protected] (D.A. Collins).


Anaerobe

journal homepage: www.elsevier .com/locate/anaerobe

http://dx.doi.org/10.1016/j.anaerobe.2015.11.0071075-9964/© 2015 Elsevier Ltd. All rights reserved.

Anaerobe 37 (2016) 34e37

Delta:1_given name

Delta:1_surname


http://crossmark.crossref.org/dialog/?doi=10.1016/j.anaerobe.2015.11.007&domain=pdf

www.sciencedirect.com/science/journal/10759964

http://www.elsevier.com/locate/anaerobe

http://dx.doi.org/10.1016/j.anaerobe.2015.11.007



thewhole of Australia) in early 2012 [10]. RT 244 is closely related tothe RT 027 strain. Both strains produce CDT, which is thought tocontribute to more severe disease and greater mortality [8], how-ever, RT 244 has predominantly been identified in patients with CO-CDI [11]. Whole genome sequencing analysis of outbreak isolatesfrom across Australia suggested most likely a single source for theoutbreak [11]. This source must lie somewhere in the community,possibly in food production animals [12e14] or food [15,16].

In Western Australia (WA), testing for CDI is not mandatory indiarrhoeal cases. Many outpatient clinics and community medicalpractices do not necessarily consider CDI in diarrhoeal patients.Point-prevalence studies conducted in Spain in 2008 and 2013found that under-diagnosis of CDI occurred in 66% and 51%,respectively, of diarrhoeal patients [17,18]. A multi-country surveyacross Europe between September 2011 and August 2013 showedthat 23% of CDI cases were missed [19]. These missed diagnoseswere due to either lack of clinical suspicion leading to no testrequest, or insufficient laboratory testing.

The aims of this study were to survey testing requests for CDIreceived by a busy diagnostic laboratory in WA. This would enableus to determine in which cases testing for CDI does not occur, thetrue prevalence of CDI inWA, andwhether CDI is being identified incommunity patients in WA.

2. Methods

All faecal specimens submitted for testing to the PathWest Lab-oratory Medicine (WA) enteric diagnostic laboratory serving sitesacross thewhole state ofWAwere surveyed for 1month in July 2014.Diarrhoeal (loose or watery consistency, assuming the shape of itscontainer) specimenswhere CDI testingwas specifically requested orroutinely performed (routine group) [20] were processed as normalby the diagnostic laboratory. The criteria for routine testing were adiarrhoeal specimen received from any hospital in-patient, from anyhospital Emergency Department patient, residents of long term carefacilities or patients aged>65 years attending outpatient (OP) clinics.Paediatric (<15 years) inpatient or OP specimens were tested if theywere haematology/oncology patients, or if recent antibiotic use orpseudomembranous colitis were documented on the request.Routine testingwas performed by PCR for the tcdB gene using the BDMAX™ Cdiff assay (BD diagnostics, Franklin Lakes, NJ) [21], aswell asan array of tests for bacterial, parasitic and viral causes of diarrhoeaand gastroenteritis. C. difficile-positive routine specimens were sub-sequently cultured on ChromID™ C. difficile agar (bioMerieux, Marcyl'Etoile, France) [22]. Any diarrhoeal specimens received where CDIwas not requested or routinely performed (non-requested group)were examined byculture onChromID™, followedby PCR for tcdBonisolate DNA (toxigenic culture) [23].

All isolates of C. difficile were PCR ribotyped [24]. The overallprevalence of tcdB-positive C. difficilewas calculated. Patients weregrouped according to whether they were hospital inpatients (IP),attending OP clinics or day surgeries based in hospitals (OP) or acommunity-based general practice (GP), and defined as HO-CDI ifthey were an IP, or CO-CDI if they attended an OP or GP clinic. Theoverall prevalence and % females were compared with the equiv-alent proportions identified by routine testing by the c2 test. Me-dian ages were compared using the Mann Whitney U test.Statistical calculations were performed using GraphPad InStat 3,considering p < 0.05 as a significant result.

3. Results

3.1. Prevalence of C. difficile

In total, 1010 non-duplicate diarrhoeal specimens were received

during the study period. Of these, 678 were processed routinely,resulting in 49 (7.2%) C. difficile PCR-positive specimens, 41 of whichwere positive by culture (Table 1). The remaining 332 specimenswere examined by toxigenic culture resulting in 16 (4.8%) positivesamples.

In the routine group, a potential second pathogen in addition toC. difficile was identified in three cases; Blastocystis hominis in twoand Salmonella montevideo in one. Among the C. difficile-negativeroutine specimens Campylobacter spp. was detected in 14 (2.1%),B. hominis in 13 (1.9%), Salmonella spp. in 11 (1.6%), Aeromonas spp.in four (0.1%), Giardia in four (0.1%), and other parasitic and bac-terial agents of diarrhoea in nine specimens. A second pathogenwas identified in 10 non-requested C. difficile-positive specimens;Salmonella spp. in five, Campylobacter jejuni in two, rotavirus in twoand Giardia in one specimen. In the non-requested group,B. hominiswas detected in 25 (7.5%) specimens, Giardia in 14 (4.2%),Campylobacter spp. in 11 (3.3%), Salmonella spp. in seven (2.1%),rotavirus in five (1.5%), Cryptosporidium in four (1.2%), norovirus inthree (0.1%), and other parasitic and bacterial agents in eightspecimens.

The overall prevalence of C. difficile among all specimens wascalculated as 6.4% (Table 1). The prevalence in GP specimens washigher (15.4%) in the routine group than in the non-requestedgroup (1.0%). The overall prevalence among IPs was 7.2%, and 6.9%and 4.6% in OP and GP specimens, respectively. The proportion ofnon-requested positive specimens among all C. difficile-positivespecimens was 24.6%.

The median age of C. difficile-positive cases was 60.0 (inter-quartile range 1.5e75.7) years. In the non-requested group, themedian age was 0.8 (IQR 0.6e1.3) years, while in the routine groupit was 72.7 (IQR 54.8e76.3) (Table 2). Females accounted for 64.6%of cases overall.

3.2. Sources of testing requests

The majority of tests in the routine group were performed on IPspecimens (417, 61.5%), while in the non-requested group the ma-jority of tests were performed on GP specimens (208, 62.7%). IP CDIcases accounted for 59.2% of all routine cases, and 49.2% overall ofcases. GP cases contributed to 18.5% of all cases overall, while OPcases accounted for the remaining 32.3% (Table 1).

3.3. Molecular epidemiology

The most common RT identified was RT 014/020 (34.1% inroutine group, 43.8% in non-requested), followed by 002, 056, 005and 018 (Fig. 1). RTs 002, 056 and 018 were identified in routinelytested patients only and, in addition, RT 018 was only found in GPpatients. RTs 103 and 247 were found in OP patients only.

4. Discussion

C. difficile was the most common diarrhoeal pathogen detectedduring the study period. By screening all diarrhoeal specimens, thenumber of C. difficile-positive specimens increased from 49 to 65,but the overall prevalence of C. difficile was lower (6.4%) than thatidentified by routine testing (7.2%, Table 1). The prevalence wassimilar to that found in Spain in 2013 (6.0%) [17]. Previously, theprevalence of toxigenic C. difficile was 7.2% in a survey fromDecember 2013eJanuary 2014 on routine testing in the samediagnostic laboratory [25]. This suggests the prevalence identifiedby routine testing in WA remained constant.

CDI was most common in females (64.6%) and the overall me-dian age in cases was 60.0 years (IQR 1.5e75.7 years). These char-acteristics correlate with many other epidemiological surveys of

D.A. Collins, T.V. Riley / Anaerobe 37 (2016) 34e37 35

CDI worldwide [3,26]. Our case age range was lower than the re-sults of point prevalence surveys in Spain and across Europe, wherethe median ages of patients were 62.7 years (IQR 32.6e78.4) and64.0 (IQR 35.0e78.0) years. In a similar 3 month survey in theNetherlands in 2008, the median age was 69 years [27].

While the results found here are similar to those found in theEuropean surveys, the characteristics of cases of CDI missed havenot previously been described. In our study, missed CDI casesoccurred in children only, and 32.3% of all CDI cases were aged <20years. In both the European and the Netherlands prevalence sur-veys, 18% of CDI cases identified were aged <20 years [19,27], butthe age characteristics of missed cases were not described.

A previous study in WA in 2011e2012 found a high prevalenceof C. difficile (45%) among diarrhoeal paediatric patients [28].Several reports from the USA have shown increasing rates of CDIamong hospitalised paediatric patients [29e31]. Whether C. difficiletruly causes infection in children is continually debated due to highcolonization rates in infants and a high proportion of cases of co-infection with other diarrhoeal pathogens [32]. It has been sug-gested that co-infection with another pathogen may increase thelikelihood of isolating C. difficile by altering the normal gut flora[33]. We found co-infectionwith another diarrhoeal pathogen in 10

cases. While there is some disagreement whether C. difficile causesinfections or merely colonizes the gut, particularly in very youngchildren, CDI may need to be considered in these paediatric cases. Arecent study from the USA [34] reported that children diagnosedwith CA-CDI by PCR often lacked traditional risk factors and couldnot be confirmed as having CDI using different testing methods,suggesting overdiagnosis of CDI in children with CO diarrhoea.These findings show that testing practices for children with diar-rhoea may require re-evaluation to include only children who mayplausibly have CDI. This would most likely be the case in patientswith recent antimicrobial use.

Overall 50.8% of cases were CO-CDI (OP or GP; Table 1). Amongthe 33 CO cases, 13 cases were not identified by routine testing. Thehigh proportion of CO cases highlights the need for a high level ofclinical suspicion for CDI in community patients. GPs appearedrarely to consider C. difficile in diarrhoeal patients given the lowproportion of requests for C. difficile testing, however, the actualprevalence of CDI was indeed low among GP specimens (4.6%). Onthe other hand, the highest prevalence of C. difficile was identifiedin the routinely processed specimens from GPs (15.4%) indicatingGPs do request C. difficile testing for patients who are most likely tohave CDI. The low proportion of GP requests may reflect either a

Table 1Prevalence of C. difficile among routine and non-requested groups, and determination of true (overall) prevalence.

Routine Non-requested Overall

Source Number ofspecimens

Prevalence(%)

Proportion of all cases(%)

Number ofspecimens

Prevalence(%)


Prevalence(%)


IP 417 7.0 59.2 29 10.3 18.8 7.2 49.2OP 209 5.7 24.5 95 9.5 56.2 6.9 32.3GP 52 15.4 16.3 208 1.9 25.0 4.6* 18.5Total 678 7.2 100.0 332 4.8 100.0 6.4 100.0

*p < 0.05 for comparison of overall prevalence with routine prevalence by c2.

Table 2Demographic characteristics of C. difficile cases by request source and study group.

Group Routine (n ¼ 49) Non-requested (n ¼ 16) Overall

Source Median age (interquartile range) % Female Median age (interquartile range) % Female Median age (interquartile range) % Female

IP 73.4 (58.0e76.2) 75.9 1.0 (0.8e1.0) 66.6 70.8 (54.2e75.8) 75.0OP 57.9 (20.7e77.5) 50.0 0.7 (0.6e1.2) 55.6 1.9 (0.7e64.5) 52.4GP 74.2 (60.5e76.1) 62.5 1.2 (0.8e2.3) 100.0 58.8 (1.5e75.7) 58.3Total 72.7 (54.8e76.3) 67.3 0.8 (0.6e1.3) 55.6 60.0 (1.5e75.7)* 64.6

*p < 0.05 for comparison of overall median age with routine median age by Mann Whitney U test.

Fig. 1. Ribotypes of C. difficile isolates by request source and study group. Routine group n ¼ 41; Non-requested group n ¼ 16. RT014/020 was most common, followed by 002, 056,018 and 005.

D.A. Collins, T.V. Riley / Anaerobe 37 (2016) 34e3736

lack of consideration of CDI in diarrhoeal cases, or perhaps a lack ofawareness of the recent high incidence of CA-CDI in Australia [6].While the prevalence was low among GP patient specimens, wehave nevertheless recently recommended vigilance among GPs andcommunity health centres for CDI in community patients [35].

RT 014/020 was most commonly isolated, followed by 002 and056. RT 014/020 is the most commonly identified RT in Australia[28,36], and features prominently in Europe, Asia and NorthAmerica, as does RT 002 [26,37,38]. In Europe, RT 056 was amongthe most frequent RTs and was associated with complicated CDIoutcomes [37]. RT 056 was also among the top four RTs in WA in asurvey in 2012 [36]. RTs 027 and 244 were not identified among thecases in this study. The molecular epidemiology did not differgreatly in non-requested cases compared to routine cases.

Given the high proportion of CO-CDI cases found, communitysources of C. difficile need to be identified. One likely source is foodanimals and their meat products. RT 014 was the most frequentlyidentified RT in Australian pigs in a survey in 2014 [39], and RT 056was among the most common RTs found in Australian calves atslaughter in 2012 [14]. Many studies overseas have also identified C.difficile in meat products and fresh vegetables [15,40e42], however,to date C. difficile has not been reported in Australian meat or otherfood products. Improvement in isolation protocols should allowmore broad-scale testing to be performed to provide a better pic-ture of possible contamination in Australian foods.

In this study we have determined that C. difficile is likely themost common cause of diarrhoea in WA, frequently affectingcommunity patients. We recommend increasing awareness andclinical suspicion in community health providers, particularly GPs[35]. With the recent increases in incidence of CDI in Australia [6]and emergence of new virulent strains [11], ongoing surveillanceand awareness are required to control CDI in Australia.

Acknowledgements

Deirdre Collins is a recipient of an International PostgraduateResearch Scholarship and an Australian Postgraduate Award.

References

[1] Centers for Disease Control and Prevention (CDC), Antibiotic ResistanceThreats in the United States, 2013, Atlanta 2013.

[2] D.A. Leffler, J.T. Lamont, Clostridium difficile infection, N. Engl. J. Med. 372(2015) 1539e1548.

[3] J. Freeman, M.P. Bauer, S.D. Baines, et al., The changing epidemiology ofClostridium difficile infections, Clin. Microbiol. Rev. 23 (2010) 529e549.

[4] S. Khanna, D.S. Pardi, S.L. Aronson, et al., The epidemiology of community-acquired Clostridium difficile infection: a population-based study, Am. J. Gas-troenterol. 107 (2012) 89e95.

[5] L.C. McDonald, B. Coignard, E. Dubberke, et al., Recommendations for sur-veillance of Clostridium difficile-associated disease, Infect. Control Hosp. Epi-demiol. 28 (2007) 140e145.

[6] C. Slimings, P. Armstrong, W.D. Beckingham, et al., Increasing incidence ofClostridium difficile infection, Australia, 2011-2012, Med. J. Aust. 200 (2014)272e276.

[7] B.G. Mitchell, F. Wilson, A. McGregor, An Increase in Community Onset Clos-tridium difficile Infection: A Population-based Study 17, Healthcare Infection,Tasmania, Australia, 2012, pp. 127e132.

[8] S.K. Lim, R.L. Stuart, K.E. Mackin, et al., Emergence of a ribotype 244 strain ofClostridium difficile associated with severe disease and related to the epidemicribotype 027 strain, Clin. Infect. Dis. 58 (2014) 1723e1730.

[9] M. De Almeida, H. Heffernan, A. Dervan, et al., Severe Clostridium difficileinfection in New Zealand associated with an emerging strain, PCR-ribotype244, N. Z. Med. J. 126 (2013) 9e14.

[10] C.A. Huber, L. Hall, N.F. Foster, et al., Surveillance snapshot of Clostridiumdifficile infection in hospitals across Queensland detects binary toxin pro-ducing ribotype UK 244, Commun. Dis. Intell. Q. Rep. 38 (2014) E279eE284.

[11] D.W. Eyre, L. Tracey, B. Elliot, et al., Emergence and spread of predominantlycommunity-onset Clostridium difficile PCR ribotype 244 infection in Australia,2010 to 2012, Eurosurveillance 20 (2015) 21059.

[12] D.R. Knight, T.V. Riley, Prevalence of gastrointestinal Clostridium difficile car-riage in Australian sheep and lambs, Appl. Environ. Microbiol. 79 (2013)

5689e5692.[13] D.R. Knight, M.M. Squire, T.V. Riley, Laboratory detection of Clostridium difficile

in piglets in Australia, J. Clin. Microbiol. 52 (2014) 3856e3862.[14] D.R. Knight, S. Thean, P. Putsathit, et al., Cross-sectional study reveals high

prevalence of Clostridium difficile non-PCR ribotype 078 strains in Australianveal calves at slaughter, Appl. Environ. Microbiol. 79 (2013) 2630e2635.

[15] D.S. Metcalf, M.C. Costa, W.M.V. Dew, et al., Clostridium difficile in vegetables,Canada, Lett. Appl. Microbiol. 51 (2010) 600e602.

[16] D. Metcalf, R.J. Reid-Smith, B.P. Avery, et al., Prevalence of Clostridium difficilein retail pork, Can. Vet. J. 51 (2010) 873e876.

[17] L. Alcala, E. Reigadas, M. Marin, et al., Impact of clinical awareness anddiagnostic tests on the underdiagnosis of Clostridium difficile infection, Eur. J.Clin. Microbiol. Infect. Dis. 34 (2015) 1515e1525.

[18] L. Alcala, A. Martin, M. Marin, et al., The undiagnosed cases of Clostridiumdifficile infection in a whole nation: where is the problem? Clin. Microbiol.Infect. 18 (2012) E204eE213.

[19] K.A. Davies, C.M. Longshaw, G.L. Davis, et al., Underdiagnosis of Clostridiumdifficile across Europe: the European, multicentre, prospective, biannual,point-prevalence study of Clostridium difficile infection in hospitalised pa-tients with diarrhoea (EUCLID), Lancet Infect. Dis. 14 (2014) 1208e1219.

[20] R.A. Bowman, T.V. Riley, Routine culturing for Clostridium difficile? Pathology16 (1984) 240e242.

[21] R. Le Guern, S. Herwegh, B. Grandbastien, et al., Evaluation of a new moleculartest, the BD Max Cdiff, for detection of toxigenic Clostridium difficile in fecalsamples, J. Clin. Microbiol. 50 (2012) 3089e3090.

[22] K.C. Carson, L.V. Boseiwaqa, S.K. Thean, et al., Isolation of Clostridium difficilefrom faecal specimensea comparison of ChromID C. difficile agar andcycloserine-cefoxitin-fructose agar, J. Med. Microbiol. 62 (2013) 1423e1427.

[23] H. Kato, N. Kato, K. Watanabe, et al., Identification of toxin A-negative, toxin B-positive Clostridium difficile by PCR, J. Clin. Microbiol. 36 (1998) 2178e2182.

[24] G.L. O'Neill, F.T. Ogunsola, J.S. Brazier, et al., Modification of a PCR ribotypingmethod for application as a routine typing scheme for Clostridium difficile,Anaerobe 2 (1996) 205e209.

[25] P. Putsathit, J. Morgan, D. Bradford, et al., Evaluation of the BD Max Cdiff assayfor the detection of toxigenic Clostridium difficile in human stool specimens,Pathology 47 (2015) 165e168.

[26] D.A. Collins, P.M. Hawkey, T.V. Riley, Epidemiology of Clostridium difficileinfection in Asia, Antimicrob. Resist Infect. Control 2 (2013) 21.

[27] M.P. Bauer, D. Veenendaal, L. Verhoef, et al., Clinical and microbiologicalcharacteristics of community-onset Clostridium difficile infection in TheNetherlands, Clin. Microbiol. Infect. 15 (2009) 1087e1092.

[28] J. Hart, P. Putsathit, D.R. Knight, et al., Clostridium difficile infection diagnosis ina paediatric population: comparison of methodologies, Eur. J. Clin. Microbiol.Infect. Dis. 33 (2014) 1555e1564.

[29] J. Kim, S.A. Smathers, P. Prasad, et al., Epidemiological features of Clostridiumdifficile-associated disease among inpatients at children's hospitals in theUnited States, 2001-2006, Pediatrics 122 (2008) 1266e1270.

[30] M.D. Zilberberg, G.S. Tillotson, C. McDonald, Clostridium difficile infectionsamong hospitalized children, United States, 1997-2006, Emerg. Infect. Dis. 16(2010) 604e609.

[31] C.M. Nylund, A. Goudie, J.M. Garza, et al., Clostridium difficile infection inhospitalized children in the United States, Arch. Pediatr. Adolesc. Med. 165(2011) 451e457.

[32] H. de Graaf, S. Pai, D.A. Burns, et al., Co-infection as a confounder for the role ofClostridium difficile infection in children with diarrhoea: a summary of theliterature, Eur. J. Clin. Microbiol. Infect. Dis. 34 (2015) 1281e1287.

[33] E. Falsen, B. Kaijser, L. Nehls, et al., Clostridium difficile in relation to entericbacterial pathogens, J. Clin. Microbiol. 12 (1980) 297e300.

[34] L.K. Kociolek, S.J. Patel, X. Zheng, et al., Clinical and microbiologic assessmentof cases of pediatric community-associated Clostridium difficile infection re-veals opportunities for improved testing decisions, Pediatr. Infect. Dis. J.(2015), http://dx.doi.org/10.1097/INF.0000000000000954.

[35] L. Tracey, A. Kirke, P. Armstrong, et al., From the hospital to the home - the riseand rise of Clostridium difficile infection, Aust. Fam. Physician 44 (2015)712e717.

[36] N.F. Foster, D.A. Collins, S.L. Ditchburn, et al., Epidemiology of Clostridiumdifficile infection in two tertiary-care hospitals in Perth, Western Australia: across-sectional study, New Microbes New Infect. 2 (2014) 64e71.

[37] M.P. Bauer, D.W. Notermans, B.H. van Benthem, et al., Clostridium difficileinfection in Europe: a hospital-based survey, Lancet 377 (2011) 63e73.

[38] I.A. Tickler, R.V. Goering, J.D. Whitmore, et al., Strain types and antimicrobialresistance patterns of Clostridium difficile isolates from the United States:2011-2013, Antimicrob. Agents Chemother. 58 (2014) 4214e4218.

[39] D.R. Knight, M.M. Squire, T.V. Riley, Nationwide surveillance study of Clos-tridium difficile in Australian neonatal pigs shows high prevalence and het-erogeneity of PCR ribotypes, Appl. Environ. Microbiol. 81 (2015) 119e123.

[40] A. Rodriguez-Palacios, S. Borgmann, T.R. Kline, et al., Clostridium difficile infoods and animals: history and measures to reduce exposure, Anim. HealthRes. Rev. 14 (2013) 1e19.

[41] E. de Boer, A. Zwartkruis-Nahuis, A.E. Heuvelink, et al., Prevalence of Clos-tridium difficile in retailed meat in The Netherlands, Int. J. Food Microbiol. 144(2011) 561e564.

[42] J.S. Weese, B.P. Avery, J. Rousseau, et al., Detection and enumeration of Clos-tridium difficile spores in retail beef and pork, Appl. Environ. Microbiol. 75(2009) 5009e5011.

D.A. Collins, T.V. Riley / Anaerobe 37 (2016) 34e37 37

http://refhub.elsevier.com/S1075-9964(15)30083-4/sref1

























































































































http://dx.doi.org/10.1097/INF.0000000000000954
































M I C R O B I O L O G Y

Laboratory-based surveillance of Clostridium difficilecirculating in Australia, September – November 2010

ALLEN C. CHENG1,2, DEIRDRE A. COLLINS

3, BRIONY ELLIOTT3,

JOHN K. FERGUSON4,5, DAVID L. PATERSON6, SARA THEAN7

AND

THOMAS V. RILEY3,7

1Department of Epidemiology and Preventive Medicine, Monash University, 2InfectiousDiseases Unit, Alfred Hospital, Melbourne, Vic, 3School of Pathology and LaboratoryMedicine, University of Western Australia, WA, 4Infection Prevention Service, Hunter NewEngland Health, 5Hunter New England Health and University of Newcastle, John HunterHospital, Newcastle, NSW, 6University of Queensland Centre for Clinical Research, RoyalBrisbane and Women’s Hospital, Brisbane, Qld, and 7Department of Microbiology, PathWestLaboratory Medicine (WA), Perth, WA, Australia

SummaryClostridium difficile rose in prominence in the early 2000swith large-scale outbreaks of a particular binary toxin-positive strain, ribotype 027, in North America andEurope. In Australia outbreaks of the same scale had notand have not been seen. A survey of C. difficile acrossAustralia was performed for 1 month in 2010. A collectionof 330 C. difficile isolates from all States and Territoriesexcept Victoria and the Northern Territory was amassed.PCR ribotyping revealed a diverse array of strains. Ribo-types 014/020 (30.0%) and 002 (11.8%) were mostcommon, followed by 054 (4.2%), 056 (3.9%), 070 (3.6%)and 005 (3.3%). The collection also contained few binarytoxin positive strains, namely 027 (0.9%), 078 (0.3%), 244(0.3%), 251 (0.3%) and 127 (0.3%). The survey highlightsthe need for vigilance for emerging strains in Australia, andgives an overview of the molecular epidemiology ofC. difficile in Australia prior to an increase in incidencenoted from mid-2011.

Key words: Clostridium difficile; ribotype; epidemiology; surveillance; mo-lecular typing; Australia.

Received 18 August, revised 5 November, accepted 11 November 2015Available online 2 March 2016

INTRODUCTIONAn epidemic strain ofClostridium difficile [PCR ribotype (RT)027] was first identified in Quebec Province, Canada, in 2005,as a cause of hospital outbreaks of severe infection with highmortality rates.1 Retrospective analyses suggested this straincaused outbreaks across North America dating back to 2000.The organism later spread to Europe and cases have now beendescribed in Asia and Central America.2 Increased toxin A andB production by C. difficile RT 027, as well as the presence ofan additional binary toxin (CDT), may be responsible for itsincreased virulence,3 however, fluoroquinolone resistance is

likely to have contributed to its spread.4 Infection with thisstrain leads more often to severe disease, more recurrences anda greater risk of death.1

There has been concern in Australia because of the lack ofsuitable surveillance systems to detect the entry of epidemicC. difficile into this country.5,6 The first infected patient withRT 027 in Australia was reported in 2009 in WesternAustralia (WA), but the infection was thought to have beenacquired in North America.7 The first case of C. difficile RT027 infection thought to have been acquired in Australia wasreported in early 2011 (although detected at the beginning of2010) in a case from Melbourne, Victoria.8 The strain wasidentified after clinicians alerted the laboratory to the severityof the infection and the possibility of a ‘hyper-virulent’ strain,and molecular strain typing identified C. difficile RT 027. Ofconcern, several other cases were subsequently detected atthe same hospital, other hospitals and a nursing home inMelbourne. In late 2010, a cluster of cases of RT 027infection was discovered in North Sydney, New South Wales(NSW).9 The outbreaks of RT 027 in Victoria appear to haveoriginated from a single introduction into the country fromNorth America.10

Ongoing surveillance, including monitoring of changes inmolecular epidemiology, is required to provide informationfor clinicians and to inform infection prevention in-terventions. A recommendation from the Australian Com-mission on Safety and Quality in Healthcare for hospitalsurveillance programs in all States and Territories to monitorC. difficile11 was approved by Australian Health Ministers inNovember 2008. All States and Territories have implementedthis recommendation. A significant increase in both hospital-acquired CDI (HA-CDI) and community-acquired (CA-CDI)in Australia during 2011–2012 was identified throughcollation of hospital surveillance data.12 In this study, wedescribe the molecular epidemiology of C. difficile infection(CDI), and the relative frequency of epidemic strains inAustralia in late 2010 prior to the increases in CDI reportedfor 2011. As such, this analysis provides baseline results forfuture comparisons.

Print ISSN 0031-3025/Online ISSN 1465-3931 © 2016 Royal College of Pathologists of Australasia. Published by Elsevier B.V. All rights reserved.DOI: http://dx.doi.org/10.1016/j.pathol.2016.02.005

Pathology (April 2016) 48(3), pp. 257–260

http://crossmark.crossref.org/dialog/?doi=10.1016/j.pathol.2016.02.005&domain=pdf

http://dx.doi.org/10.1016/j.pathol.2016.02.005

METHODSStudy design

This laboratory-based survey was performed for 1 month betweenSeptember and November, 2010. Isolates of C. difficile from patientsdeveloping diarrhoea in hospital or presenting with diarrhoea to a hospitalor in the community were collected in participating diagnostic laboratoriesacross all States and Territories except the Northern Territory and Victoria.One laboratory participated in the Australian Capital Territory (ACT), fivein NSW, three in Queensland (QLD), one in South Australia (SA), one inTasmania (TAS) and one in WA. Most of these laboratories provideddiagnostic services to public hospitals. No change to current testing stra-tegies operating at the participating laboratories was proposed. Participatinglaboratories routinely performed culture for C. difficile or cultured anyspecimen positive by a screening test for inclusion in the isolate collection.This may have been as part of primary screening or following positive rapidtests. If toxin detection tests were performed on isolates, then both toxinpositive and negative isolates were referred. Isolates from duplicate speci-mens taken within 7 days were excluded. No patient demographic or clinicaldata were collected.Clostridium difficile isolates or specimens were transported to a central

reference laboratory [PathWest Laboratory Medicine (WA), Nedlands, WA]in either Robertson’s cooked meat (RCM) medium, thioglycollate broth or asspore suspensions on swabs in transport medium. Results of ribotyping werereported back to the referring hospitals/laboratories and/or local department ofhealth.

Clostridium difficile culture and molecular analysis

Clostridium difficile isolates or specimens were cultured on cycloserinecefoxitin fructose agar plates (PathWest). All plates were incubated in ananaerobic chamber (Don Whitley Scientific, Australia) for 48 h at 37�C, inan atmosphere containing 80% nitrogen, 10% hydrogen and 10% carbondioxide. Clostridium difficile was identified on the basis of characteristiccolony morphology (yellow, ground glass appearance) and odour (horsedung smell). The identity of doubtful isolates was confirmed by Gramstain, latex agglutination test kit (Oxoid, UK)13 and/or species specificPCR.14

PCR ribotyping was performed for all isolates according to the method ofO’Neill et al.15 with some modifications. Amplification was performed in a 50mL reaction volume with 1× reaction buffer, 4 mM MgCl2, 100 mM of eachdNTP, 0.4 mM of each primer, 20 mg/mL BSA, 3.75 U AmpliTaq Gold DNApolymerase (Applied Biosystems, USA), and 10 mL of DNA template. PCRwas carried out with the following thermal cycler program: an initial dena-turation step of 95�C for 1 min, followed by 35 cycles of 94�C for 1 min,55�C for 1 min and 72�C for 2 min, with a final extension step of 72�C for7 min. PCR products were purified with the MinElute PCR Purification Kit(Qiagen, Germany). Capillary gel electrophoresis of PCR products wasperformed using a QIAxcel instrument (Qiagen) with 15 bp/1 kb alignmentmarker and patterns were compared to known RTs of C. difficile. Ribotypingprofiles were analysed with Bionumerics (version 7.5; Applied Maths,Belgium) and compared with a collection of reference strains. Isolates werealso characterised for toxigenic properties using PCR reactions for the tcdA,tcdB, cdtA and cdtB genes.16,17

RESULTSCDI incidence rates

The number of viable isolates of C. difficile collected and anestimate of the incidence rate per 100,000 population areshown in Table 1. Person time was calculated by dividing thepopulation (per year) for each state by 12 to give person-months. The overall national incidence rate was calculatedat 23.8/100,000 person months. The incidence rate washighest in WA at 30.2/100,000 and lowest in SA at 16.8/100,000.

Molecular epidemiology

The distribution of C. difficile RTs throughout Australia isshown in Table 2. The 10 most common RTs comprised67.3% of the total number. More than 60 RTs were repre-sented in the remaining 31.8% of isolates, many with onlyone representative strain. The two most common RTs wereRT 14/020 (30.0%) and RT 002 (11.8%), followed by RT054 (4.2%), RT 056 (3.9%) and RT 070 (3.6%). SeveralCDT-positive isolates were detected. These were three RT027 (0.9%) isolates identified in NSW, one RT 078 (0.3%)isolate in NSW, one RT 127 (0.3%) in NSW, one RT 251(0.3%) in NSW and one RT 244 (0.3%) isolate in Qld.

DISCUSSIONIn this study, we found a significant number of confirmedcases of CDI from participating laboratories. The crudeincidence rate of CDI at 23.8/100,000 (Table 1) represents anunderestimate of the true population rate, as not all labora-tories in participating States and Territories referred isolatesfor the survey. Rates also varied between jurisdictions, due inpart to variation in numbers of participating laboratories.However, in the ACT and TAS, where the participatinglaboratories service the entire population of these jurisdic-tions, ascertainment is likely to be close to complete. A USstudy estimated the national incidence rate in 2011 to be 51.9/100,000 population for CA-CDI and 95.3/100,000 for HA-CDI.18 In Canada, incidence rates increased from a baselinerate of 35.6/100,000 population in 1991 to 156.3/100,000 in200319 and in Germany rates increased from 1.7–3.8/100,000 population in 2003 to 14.8/100,000 in 2006.20 Whilethe Australian incidence rate appeared to be lower than thoseidentified in North America, following this study thenationwide incidence of hospital-identified CDI in Australiaincreased from 3.25/10,000 patient days (PD) in 2011 to4.03/10,000 PD in 2012.12

Table 1 Number of isolates of C. difficile received from participating State or Territory

Jurisdiction Population Person months of surveillance Number of isolates Rate per 100,000 population

NSW 7,253,400 604,450 154 25.5QLD 4,532,300 377,692 76 20.1WA 2,306,200 192,183 58 30.2SA 1,647,800 137,317 23 16.8TASa 508,500 42,375 10 23.6ACTa 359,700 29,975 9 30.0Australia 16,607,900 1,383,992 330 23.8

a Jurisdictions with complete ascertainment.

258 CHENG et al. Pathology (2016), 48(3), April

This is the first systematic study of the molecular epide-miology of CDI across Australia. Previously, the scope andprevalence of epidemic strains of C. difficile in Australianhospitals were not known. The survey revealed that RTs 014/020 and 002 were most common, similar to European andNorth American isolate collections.21,22 RT 014 is the mostcommon RT of C. difficile in most typing studies worldwide.As part of a study that evaluated a new antimicrobial treat-ment for CDI, Cheknis et al. typed 24 Australian C. difficileisolates recovered in the mid-2000s by restriction endonu-clease analysis (REA) and found that two-thirds of them wereuncommon types (compared to isolates from North Americaand Europe). However, a quarter of the isolates belonged toREA group Y, a group that corresponds to PCR RTs 014 and020.23

The isolation of eight RT 018 strains in NSW is notable.RT 018 is highly prevalent in Asia, particularly in Japan.24 Itappeared in Korea in the early 2000s25 and also emerged inItaly at around the same time and caused widespread severedisease and a number of outbreaks.26 A recent study of ItalianC. difficile strains showed RT 018 was the most prevalent RToverall and was highly transmissible, accounting for 95.7%of secondary CDI cases within hospitals.27 This apparentlyhigh transmissibility may be related to enhanced virulence,and surveillance for this strain in Australia appears warrantedto guard against increasing RT 018 infection rates in thefuture. Another common strain in Asia, RT 017, was isolatedonly once in our collection (0.3%). Given the high frequencyof travel between Australia and its close neighbours in Asia,the rarity of RT 017 among the Australian collection wasunexpected.RT 056 was among the five most common strains in this

collection. It was also isolated frequently in the Europeansurvey of C. difficile in hospital patients,21 where it wasassociated with complicated disease outcomes. A recentstudy found RT 056 was commonly isolated from thegastrointestinal tracts of Australian veal calves at slaughter,28

meaning it may have potential to enter the food chain inAustralia, and possibly be exported to other countries as well.The relationship between human RT 056 strains and animalRT 056 strains is currently being explored with wholegenome sequencing.

The prevalence of CDT positive strains, particularly RTs027 and 078, in this collection was low. Nonetheless, thestudy highlighted the need for vigilance for the appearance ofvirulent or ‘hyper-virulent’ strains. Three RT 027 strains weredetected, all from NSW, and all from Sydney. AlthoughVictorian laboratories did not participate in the study, smallnumbers of cases of infection due to RT 027 associated withhospitals and nursing homes were reported in 2010.8 Also ofsome concern was the detection of two RT 078 strains inNSW. RT 078 is commonly found in production animalsoutside Australia. Many pig and cattle herds in the USA andEurope (up to 95%) are infected with C. difficile RT 078.29

Most animal isolates of C. difficile produce CDT and RT078 is a strain that, like RT 027, also produces more toxins Aand B, as well as CDT. RT 078 was the third most commonRT of C. difficile isolated from human infections in Europewhen our study was performed.21 This is a strain that requiresmonitoring in Australia. Despite the fact that Australianproduction animals do not appear to harbour this RT,28 in-fections are still occurring in humans Australia-wide (T. V.Riley, unpublished) suggesting, possibly, infection from asource emanating from outside Australia.Interestingly, this survey includes the earliest known

detection of RT 244 in Australia. In 2011 and later, RT 244went on to cause outbreaks of disease across Australia andNew Zealand.30–32 In a similar snapshot survey of CDI inQld in 2012, it was the third most common RT in circulationafter 002 and 014/020.33 RT 244 has been associated withhigher mortality rates and increased severity of disease.32 Ifmore virulent strains were identified more quickly, infectioncontrol measures could be re-enforced to reduce the spread ofthese strains of C. difficile.There are several limitations to this study. Systematic bias

may result in underestimation of the true incidence of diseasedue to under-recognition of disease by clinicians, under-testing by laboratories, particularly of community speci-mens, the use of diagnostics with imperfect sensitivity and asmall proportion of non-viable isolates. Estimates of inci-dence rates per population are approximate as the majority ofthe isolates came from hospital-identified CDI. We wereunable to access details of hospitalisations or patients’ lengthof stay as we did not receive details of source hospitals with

Table 2 Distribution of C. difficile ribotypes in participating States and Territories

Ribotype State/Territory n (%) Australia

NSW QLD WA SA TAS/ACT n (%)

014/020 44 (28.6) 18 (23.7) 26 (44.8) 6 (26.1) 5 (26.3) 99 (30.0)002 24 (15.6) 7 (9.2) 5 (8.6) 2 (8.7) 1 (5.3) 39 (11.8)054 7 (4.5) 1 (1.3) 1 (1.7) 3 (13.0) 2 (10.5) 14 (4.2)056 4 (2.6) 7 (9.2) 1 (1.7) 0 1 (5.3) 13 (3.9)070 5 (3.2) 3 (3.9) 1 (1.7) 2 (8.7) 1 (5.3) 12 (3.6)005 2 (1.3) 1 (1.3) 5 (8.6) 2 (8.7) 1 (5.3) 11 (3.3)012 7 (4.5) 1 (1.3) 1 (1.7) 0 0 9 (2.7)018 8 (5.2) 1 (1.3) 0 0 0 9 (2.7)046 5 (3.2) 3 (1.3) 0 0 0 8 (2.4)103 2 (1.3) 2 (2.6) 1 (1.7) 2 (8.7) 1 (5.3) 8 (2.4)Other 45 (29.2) 31 (40.8) 17 (29.3) 5 (21.7) 7 (36.8) 105 (31.8)NT 1 (0.6) 1 (1.3) 0 1 (4.3) 0 3 (0.9)Total 154 76 58 23 19 330

NT, not typeable.

CLOSTRIDIUM DIFFICILE SURVEILLANCE IN AUSTRALIA 259

isolates in the study. Also we could not collect clinical ordemographic data, and thus were unable to comment on theclinical significance of these strains. This, and previouslypublished work, suggests that there is only a low prevalenceof the epidemic strains RTs 027 and 078 in Australia.However, ongoing surveillance is required to monitor theincidence of CDI in the country. Further studies are needed todefine the clinical profile, risk factors (including antibiotics,animal reservoirs and long term care facilities), and optimalinfection control measures in hospitals and other healthcarefacilities.This study and the subsequent emergence of RT 24430–33

have highlighted the need to raise awareness of C. difficile inAustralia. Continued evaluation of the molecular epidemi-ology of CDI will have implications for antibiotic steward-ship programs (particularly use of quinolone antibiotics),isolation and infection control programs, and the need for andscope of ongoing surveillance of severe CDI. This systematicstudy of the molecular epidemiology of CDI in Australiaprovides a baseline to evaluate future changes in straindistribution.

Conflicts of interest and sources of funding: This studywas made possible by the financial support of the AustralianCommission on Safety and Quality in Health Care. We aregrateful to Dr Marilyn Cruikshank for her support. The au-thors state that there are no conflicts of interest to disclose.

Address for correspondence: Prof Thomas V. Riley, School of Pathologyand Laboratory Medicine (M504), Faculty of Medicine, Dentistry and HealthSciences, The University of Western Australia, 35 Stirling Highway,Crawley, WA 6009, Australia. E-mail: [email protected]

References1. Loo VG, Poirier L, Miller MA, et al. A predominantly clonal multi-

institutional outbreak of Clostridium difficile-associated diarrhea withhigh morbidity and mortality. N Engl J Med 2005; 353: 2442–9.

2. Clements AC, Magalhaes RJ, Tatem AJ, et al. Clostridium difficile PCRribotype 027: assessing the risks of further worldwide spread. LancetInfect Dis 2010; 10: 395–404.

3. Warny M, Pepin J, Fang A, et al. Toxin production by an emergingstrain of Clostridium difficile associated with outbreaks of severe diseasein North America and Europe. Lancet 2005; 366: 1079–84.

4. Pepin J, Saheb N, Coulombe MA, et al. Emergence of fluoroquinolonesas the predominant risk factor for Clostridium difficile-associated diar-rhea: a cohort study during an epidemic in Quebec. Clin Infect Dis 2005;41: 1254–60.

5. Riley TV. Epidemic Clostridium difficile.Med J Aust 2006; 185: 133–4.6. Riley TV. Is Clostridium difficile a threat to Australia’s biosecurity?

Med J Aust 2009; 190: 661–2.7. Riley TV, Thean S, Hool G, et al. First Australian isolation of epidemic

Clostridium difficile PCR ribotype 027. Med J Aust 2009; 190: 706–8.8. Richards M, Knox J, Elliott B, et al. Severe infection with Clostridium

difficile PCR ribotype 027 acquired in Melbourne, Australia. Med J Aust2011; 194: 369–71.

9. New South Wales Department of Health. Health Testing Finds NewBacteria Strain.2010; cited 12 Aug 2015. http://www.health.nsw.gov.au/news/Pages/20101217_01.aspx

10. He M, Miyajima F, Roberts P, et al. Emergence and global spread ofepidemic healthcare-associated Clostridium difficile. Nat Genet 2013;45: 109–13.

11. McGregor A, Riley T, Gessel HV. Clostridium difficile associated dis-ease. In: Cruikshank M, Ferguson J, editors. Reducing Harm to Patientsfrom Healthcare Associated Infection: The Role of Surveillance.Sydney: Australian Commission on Safety and Quality in Healthcare,2008; 171–87.

12. Slimings C, Armstrong P, Beckingham WD, et al. Increasing incidenceof Clostridium difficile infection, Australia, 2011–2012. Med J Aust2014; 200: 272–6.

13. Bowman RA, Arrow SA, Riley TV. Latex particle agglutination fordetecting and identifying Clostridium difficile. J Clin Pathol 1986; 39:212–4.

14. Gumerlock PH, Tang YJ, Meyers FJ, et al. Use of the polymerase chainreaction for the specific and direct detection of Clostridium difficile inhuman feces. Rev Infect Dis 1991; 13: 1053–60.

15. O’Neill GL, Ogunsola FT, Brazier JS, et al. Modification of a PCRribotyping method for application as a routine typing scheme forClostridium difficile. Anaerobe 1996; 2: 205–9.

16. Kato H, Kato N, Watanabe K, et al. Identification of toxin A-negative,toxin B-positive Clostridium difficile by PCR. J Clin Microbiol 1998;36: 2178–82.

17. Stubbs S, Rupnik M, Gibert M, et al. Production of actin-specific ADP-ribosyltransferase (binary toxin) by strains of Clostridium difficile.FEMS Microbiol Lett 2000; 186: 307–12.

18. Lessa FC, Mu Y, Bamberg WM, et al. Burden of Clostridium difficileinfection in the United States. N Engl J Med 2015; 372: 825–34.

19. Pepin J, Valiquette L, Alary ME, et al. Clostridium difficile-associateddiarrhea in a region of Quebec from 1991 to 2003: a changing pattern ofdisease severity. CMAJ 2004; 171: 466–72.

20. Burckhardt F, Friedrich A, Beier D, et al. Clostridium difficile surveil-lance trends, Saxony, Germany. Emerg Infect Dis 2008; 14: 691–2.

21. Bauer MP, Notermans DW, van Benthem BH, et al. Clostridium difficileinfection in Europe: a hospital-based survey. Lancet 2011; 377: 63–73.

22. Tickler IA, Goering RV, Whitmore JD, et al. Strain types and antimi-crobial resistance patterns of Clostridium difficile isolates from theUnited States, 2011 to 2013. Antimicrob Agents Chemother 2014; 58:4214–8.

23. Cheknis AK, Sambol SP, Davidson DM, et al. Distribution of Clos-tridium difficile strains from a North American, European and Australiantrial of treatment for C. difficile infections: 2005–2007. Anaerobe 2009;15: 230–3.

24. Collins DA, Hawkey PM, Riley TV. Epidemiology of Clostridiumdifficile infection in Asia. Antimicrob Resist Infect Control 2013; 2: 21.

25. Kim J, Kang JO, Kim H, et al. Epidemiology of Clostridium difficileinfections in a tertiary-care hospital in Korea. Clin Microbiol Infect2013; 19: 521–7.

26. Spigaglia P, Barbanti F, Dionisi AM, et al. Clostridium difficile isolatesresistant to fluoroquinolones in Italy: emergence of PCR ribotype 018.J Clin Microbiol 2010; 48: 2892–6.

27. Baldan R, Trovato A, Bianchini V, et al. Clostridium difficile PCRribotype 018, a successful epidemic genotype. J Clin Microbiol 2015;53: 2575–80.

28. Knight DR, Thean S, Putsathit P, et al. Cross-sectional study revealshigh prevalence of Clostridium difficile non-PCR ribotype 078 strains inAustralian veal calves at slaughter. Appl Environ Microbiol 2013; 79:2630–5.

29. Keel K, Brazier JS, Post KW, et al. Prevalence of PCR ribotypes amongClostridium difficile isolates from pigs, calves, and other species. J ClinMicrobiol 2007; 45: 1963–4.

30. Eyre DW, Tracey L, Elliott B, et al. Emergence and spread of pre-dominantly community-onset Clostridium difficile PCR ribotype 244infection in Australia, 2010 to 2012. Euro Surveill 2015; 20: 21059.

31. De Almeida MN, Heffernan H, Dervan A, et al. Severe Clostridiumdifficile infection in New Zealand associated with an emerging strain,PCR-ribotype 244. N Z Med J 2013; 126: 9–14.

32. Lim SK, Stuart RL, Mackin KE, et al. Emergence of a ribotype 244strain of Clostridium difficile associated with severe disease and relatedto the epidemic ribotype 027 strain. Clin Infect Dis 2014; 58: 1723–30.

33. Huber CA, Hall L, Foster NF, et al. Surveillance snapshot of Clos-tridium difficile infection in hospitals across Queensland detects binarytoxin producing ribotype UK 244. Commun Dis Intell Q Rep 2014; 38:E279–84.

260 CHENG et al. Pathology (2016), 48(3), April























http://www.health.nsw.gov.au/news/Pages/20101217_01.aspx

http://www.health.nsw.gov.au/news/Pages/20101217_01.aspx











































































Epidemiology of Clostridium difficile infection in the Asia- Pacific … · Epidemiology of...

Documents

Transcript of Epidemiology of Clostridium difficile infection in the Asia- Pacific … · Epidemiology of...